2
0
mirror of https://github.com/koreader/koreader synced 2024-11-16 06:12:56 +00:00
koreader/frontend/device/kobo/device.lua

1537 lines
57 KiB
Lua
Raw Normal View History

local Generic = require("device/generic/device")
local Geom = require("ui/geometry")
local UIManager
local WakeupMgr = require("device/wakeupmgr")
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
local time = require("ui/time")
local ffiUtil = require("ffi/util")
local lfs = require("libs/libkoreader-lfs")
local logger = require("logger")
local util = require("util")
local _ = require("gettext")
2021-07-21 20:20:08 +00:00
-- We're going to need a few <linux/fb.h> & <linux/input.h> constants...
local ffi = require("ffi")
local C = ffi.C
require("ffi/linux_fb_h")
2021-07-21 20:20:08 +00:00
require("ffi/linux_input_h")
require("ffi/posix_h")
local function yes() return true end
local function no() return false end
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local function NOP() return end
local function koboEnableWifi(toggle)
Various Wi-Fi QoL improvements (#6424) * Revamped most actions that require an internet connection to a new/fixed backend that allows forwarding the initial action and running it automatically once connected. (i.e., it'll allow you to set "Action when Wi-Fi is off" to "turn_on", and whatch stuff connect and do what you wanted automatically without having to re-click anywhere instead of showing you a Wi-Fi prompt and then not doing anything without any other feedback). * Speaking of, fixed the "turn_on" beforeWifi action to, well, actually work. It's no longer marked as experimental. * Consistently use "Wi-Fi" everywhere. * On Kobo/Cervantes/Sony, implemented a "Kill Wi-Fi connection when inactive" system that will automatically disconnect from Wi-Fi after sustained *network* inactivity (i.e., you can keep reading, it'll eventually turn off on its own). This should be smart and flexible enough not to murder Wi-Fi while you need it, while still not keeping it uselessly on and murdering your battery. (i.e., enable that + turn Wi-Fi on when off and enjoy never having to bother about Wi-Fi ever again). * Made sending `NetworkConnected` / `NetworkDisconnected` events consistent (they were only being sent... sometimes, which made relying on 'em somewhat problematic). * restoreWifiAsync is now only run when really needed (i.e., we no longer stomp on an existing working connection just for the hell of it). * We no longer attempt to kill a bogus non-existent Wi-Fi connection when going to suspend, we only do it when it's actually needed. * Every method of enabling Wi-Fi will now properly tear down Wi-Fi on failure, instead of leaving it in an undefined state. * Fixed an issue in the fancy crash screen on Kobo/reMarkable that could sometime lead to the log excerpt being missing. * Worked-around a number of sneaky issues related to low-level Wi-Fi/DHCP/DNS handling on Kobo (see the lengthy comments [below](https://github.com/koreader/koreader/pull/6424#issuecomment-663881059) for details). Fix #6421 Incidentally, this should also fix the inconsistencies experienced re: Wi-Fi behavior in Nickel when toggling between KOReader and Nickel (use NM/KFMon, and run a current FW for best results). * For developers, this involves various cleanups around NetworkMgr and NetworkListener. Documentation is in-line, above the concerned functions.
2020-07-27 01:39:06 +00:00
if toggle == true then
logger.info("Kobo Wi-Fi: enabling Wi-Fi")
os.execute("./enable-wifi.sh")
else
Various Wi-Fi QoL improvements (#6424) * Revamped most actions that require an internet connection to a new/fixed backend that allows forwarding the initial action and running it automatically once connected. (i.e., it'll allow you to set "Action when Wi-Fi is off" to "turn_on", and whatch stuff connect and do what you wanted automatically without having to re-click anywhere instead of showing you a Wi-Fi prompt and then not doing anything without any other feedback). * Speaking of, fixed the "turn_on" beforeWifi action to, well, actually work. It's no longer marked as experimental. * Consistently use "Wi-Fi" everywhere. * On Kobo/Cervantes/Sony, implemented a "Kill Wi-Fi connection when inactive" system that will automatically disconnect from Wi-Fi after sustained *network* inactivity (i.e., you can keep reading, it'll eventually turn off on its own). This should be smart and flexible enough not to murder Wi-Fi while you need it, while still not keeping it uselessly on and murdering your battery. (i.e., enable that + turn Wi-Fi on when off and enjoy never having to bother about Wi-Fi ever again). * Made sending `NetworkConnected` / `NetworkDisconnected` events consistent (they were only being sent... sometimes, which made relying on 'em somewhat problematic). * restoreWifiAsync is now only run when really needed (i.e., we no longer stomp on an existing working connection just for the hell of it). * We no longer attempt to kill a bogus non-existent Wi-Fi connection when going to suspend, we only do it when it's actually needed. * Every method of enabling Wi-Fi will now properly tear down Wi-Fi on failure, instead of leaving it in an undefined state. * Fixed an issue in the fancy crash screen on Kobo/reMarkable that could sometime lead to the log excerpt being missing. * Worked-around a number of sneaky issues related to low-level Wi-Fi/DHCP/DNS handling on Kobo (see the lengthy comments [below](https://github.com/koreader/koreader/pull/6424#issuecomment-663881059) for details). Fix #6421 Incidentally, this should also fix the inconsistencies experienced re: Wi-Fi behavior in Nickel when toggling between KOReader and Nickel (use NM/KFMon, and run a current FW for best results). * For developers, this involves various cleanups around NetworkMgr and NetworkListener. Documentation is in-line, above the concerned functions.
2020-07-27 01:39:06 +00:00
logger.info("Kobo Wi-Fi: disabling Wi-Fi")
os.execute("./disable-wifi.sh")
end
end
-- checks if standby is available on the device
local function checkStandby()
logger.dbg("Kobo: checking if standby is possible ...")
local f = io.open("/sys/power/state")
if not f then
return no
end
local mode = f:read()
logger.dbg("Kobo: available power states:", mode)
if mode and mode:find("standby") then
logger.dbg("Kobo: standby state is supported")
return yes
end
logger.dbg("Kobo: standby state is unsupported")
return no
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
-- Return the highest core number
local function getCPUCount()
local fd = io.open("/sys/devices/system/cpu/possible", "re")
if fd then
local str = fd:read("*line")
fd:close()
-- Format is n-N, where n is the first core, and N the last (e.g., 0-3)
return tonumber(str:match("%d+$")) or 1
else
return 1
end
end
local function getCPUGovernor(knob)
local fd = io.open(knob, "re")
if fd then
local str = fd:read("*line")
fd:close()
-- If we're currently using the userspace governor, fudge that to conservative, as we won't ever standby with Wi-Fi on.
-- (userspace is only used on i.MX5 for DVFS shenanigans when Wi-Fi is enabled)
if str == "userspace" then
str = "conservative"
end
return str
else
return nil
end
end
local function getRTCName()
local fd = io.open("/sys/class/rtc/rtc0/name", "re")
if fd then
local str = fd:read("*line")
fd:close()
return str
else
return nil
end
end
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local Kobo = Generic:extend{
model = "Kobo",
isKobo = yes,
isTouchDevice = yes, -- all of them are
hasOTAUpdates = yes,
hasFastWifiStatusQuery = yes,
hasWifiManager = yes,
canStandby = no, -- will get updated by checkStandby()
canReboot = yes,
canPowerOff = yes,
canSuspend = yes,
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
supportsScreensaver = yes,
-- most Kobos are MT-capable
hasMultitouch = yes,
-- most Kobos have X/Y switched for the touch screen
touch_switch_xy = true,
-- most Kobos have also mirrored X coordinates
touch_mirrored_x = true,
-- enforce portrait mode on Kobos
--- @note: In practice, the check that is used for in ffi/framebuffer is no longer relevant,
--- since, in almost every case, we enforce a hardware Portrait rotation via fbdepth on startup by default ;).
--- We still want to keep it in case an unfortunate soul on an older device disables the bitdepth switch...
isAlwaysPortrait = yes,
Enable HW dithering in a few key places (#4541) * Enable HW dithering on supported devices (Clara HD, Forma; Oasis 2, PW4) * FileManager and co. (where appropriate, i.e., when covers are shown) * Book Status * Reader, where appropriate: * CRe: on pages whith image content (for over 7.5% of the screen area, should hopefully leave stuff like bullet points or small scene breaks alone). * Other engines: on user-request (in the gear tab of the bottom menu), via the new "Dithering" knob (will only appear on supported devices). * ScreenSaver * ImageViewer * Minimize repaints when flash_ui is enabled (by, almost everywhere, only repainting the flashing element, and not the toplevel window which hosts it). (The first pass of this involved fixing a few Button instances whose show_parent was wrong, in particular, chevrons in the FM & TopMenu). * Hunted down a few redundant repaints (unneeded setDirty("all") calls), either by switching the widget to nil when only a refresh was needed, and not a repaint, or by passing the appropritate widget to setDirty. (Note to self: Enable *verbose* debugging to catch broken setDirty calls via its post guard). There were also a few instances of 'em right behind a widget close. * Don't repaint the underlying widget when initially showing TopMenu & ConfigDialog. We unfortunately do need to do it when switching tabs, because of their variable heights. * On Kobo, disabled the extra and completely useless full refresh before suspend/reboot/poweroff, as well as on resume. No more double refreshes! * Fix another debug guard in Kobo sysfs_light * Switch ImageWidget & ImageViewer mostly to "ui" updates, which will be better suited to image content pretty much everywhere, REAGL or not. PS: (Almost :100: commits! :D)
2019-02-07 00:14:37 +00:00
-- we don't need an extra refreshFull on resume, thank you very much.
needsScreenRefreshAfterResume = no,
2023-01-18 08:15:33 +00:00
-- some devices have coloured frontlights
hasNaturalLight = no,
hasNaturalLightMixer = no,
-- HW inversion is generally safe on Kobo, except on a few boards/kernels
canHWInvert = yes,
home_dir = "/mnt/onboard",
canToggleMassStorage = yes,
-- New devices *may* be REAGL-aware, but generally don't expect explicit REAGL requests, default to not.
isREAGL = no,
-- Mark 7 devices sport an updated driver.
isMk7 = no,
-- MXCFB_WAIT_FOR_UPDATE_COMPLETE ioctls are generally reliable
hasReliableMxcWaitFor = yes,
-- Sunxi devices require a completely different fb backend...
isSunxi = no,
-- On sunxi, "native" panel layout used to compute the G2D rotation handle (e.g., deviceQuirks.nxtBootRota in FBInk).
boot_rota = nil,
-- Standard sysfs path to the battery directory
battery_sysfs = "/sys/class/power_supply/mc13892_bat",
-- Stable path to the NTX input device
ntx_dev = "/dev/input/event0",
-- Stable path to the Touch input device
touch_dev = "/dev/input/event1",
-- Stable path to the Power Button input device
power_dev = nil,
2021-07-21 20:20:08 +00:00
-- Event code to use to detect contact pressure
pressure_event = nil,
-- Device features multiple CPU cores
isSMP = no,
-- Device supports "eclipse" waveform modes (i.e., optimized for nightmode).
hasEclipseWfm = no,
-- Device ships with various hardware revisions under the same device code, requirign automatic hardware detection...
automagic_sysfs = false,
unexpected_wakeup_count = 0,
}
local KoboTrilogyA = Kobo:extend{
model = "Kobo_trilogy_A",
-- Unlike its B brethren, this one doesn't do the weird translation dance when ABS_PRESSURE is 0...
hasKeys = yes,
hasMultitouch = no,
}
-- Kobo Touch B:
local KoboTrilogyB = Kobo:extend{
model = "Kobo_trilogy_B",
touch_kobo_mk3_protocol = true,
hasKeys = yes,
hasMultitouch = no,
}
-- Kobo Touch C:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboTrilogyC = Kobo:extend{
model = "Kobo_trilogy_C",
2014-12-26 11:45:09 +00:00
hasKeys = yes,
hasMultitouch = no,
}
-- Kobo Mini:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboPixie = Kobo:extend{
model = "Kobo_pixie",
display_dpi = 200,
hasMultitouch = no,
-- bezel:
viewport = Geom:new{x=0, y=2, w=596, h=794},
}
2016-09-13 05:50:07 +00:00
-- Kobo Aura One:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboDaylight = Kobo:extend{
2016-09-13 05:50:07 +00:00
model = "Kobo_daylight",
hasFrontlight = yes,
touch_phoenix_protocol = true,
display_dpi = 300,
hasNaturalLight = yes,
2018-10-27 20:49:18 +00:00
frontlight_settings = {
frontlight_white = "/sys/class/backlight/lm3630a_led1b",
frontlight_red = "/sys/class/backlight/lm3630a_led1a",
frontlight_green = "/sys/class/backlight/lm3630a_ledb",
},
2016-09-13 05:50:07 +00:00
}
-- Kobo Aura H2O:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboDahlia = Kobo:extend{
model = "Kobo_dahlia",
canToggleChargingLED = yes,
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
led_uses_channel_3 = true,
hasFrontlight = yes,
touch_phoenix_protocol = true,
-- There's no slot 0, the first finger gets assigned slot 1, and the second slot 2.
-- NOTE: Could be queried at runtime via EVIOCGABS on C.ABS_MT_TRACKING_ID (minimum field).
-- Used to be handled via an adjustTouchAlyssum hook that just mangled ABS_MT_TRACKING_ID values.
main_finger_slot = 1,
display_dpi = 265,
-- the bezel covers the top 11 pixels:
viewport = Geom:new{x=0, y=11, w=1080, h=1429},
}
-- Kobo Aura HD:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboDragon = Kobo:extend{
model = "Kobo_dragon",
hasFrontlight = yes,
hasMultitouch = no,
display_dpi = 265,
}
-- Kobo Glo:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboKraken = Kobo:extend{
model = "Kobo_kraken",
hasFrontlight = yes,
hasMultitouch = no,
display_dpi = 212,
}
-- Kobo Aura:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboPhoenix = Kobo:extend{
model = "Kobo_phoenix",
hasFrontlight = yes,
touch_phoenix_protocol = true,
display_dpi = 212,
-- The bezel covers 10 pixels at the bottom:
viewport = Geom:new{x=0, y=0, w=758, h=1014},
-- NOTE: AFAICT, the Aura was the only one explicitly requiring REAGL requests...
isREAGL = yes,
-- NOTE: May have a buggy kernel, according to the nightmode hack...
canHWInvert = no,
}
-- Kobo Aura H2O2:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboSnow = Kobo:extend{
model = "Kobo_snow",
hasFrontlight = yes,
touch_snow_protocol = true,
touch_mirrored_x = false,
display_dpi = 265,
hasNaturalLight = yes,
frontlight_settings = {
frontlight_white = "/sys/class/backlight/lm3630a_ledb",
frontlight_red = "/sys/class/backlight/lm3630a_led",
frontlight_green = "/sys/class/backlight/lm3630a_leda",
},
}
-- Kobo Aura H2O2, Rev2:
--- @fixme Check if the Clara fix actually helps here... (#4015)
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboSnowRev2 = Kobo:extend{
model = "Kobo_snow_r2",
isMk7 = yes,
hasFrontlight = yes,
touch_snow_protocol = true,
display_dpi = 265,
hasNaturalLight = yes,
frontlight_settings = {
frontlight_white = "/sys/class/backlight/lm3630a_ledb",
frontlight_red = "/sys/class/backlight/lm3630a_leda",
},
}
-- Kobo Aura second edition:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboStar = Kobo:extend{
model = "Kobo_star",
hasFrontlight = yes,
touch_phoenix_protocol = true,
display_dpi = 212,
}
-- Kobo Aura second edition, Rev 2:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboStarRev2 = Kobo:extend{
model = "Kobo_star_r2",
isMk7 = yes,
hasFrontlight = yes,
touch_phoenix_protocol = true,
display_dpi = 212,
}
2015-05-15 21:46:16 +00:00
-- Kobo Glo HD:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboAlyssum = Kobo:extend{
2015-05-15 21:46:16 +00:00
model = "Kobo_alyssum",
hasFrontlight = yes,
2015-05-16 01:57:00 +00:00
touch_phoenix_protocol = true,
main_finger_slot = 1,
2015-05-15 21:46:16 +00:00
display_dpi = 300,
}
-- Kobo Touch 2.0:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboPika = Kobo:extend{
model = "Kobo_pika",
touch_phoenix_protocol = true,
main_finger_slot = 1,
}
-- Kobo Clara HD:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboNova = Kobo:extend{
model = "Kobo_nova",
isMk7 = yes,
canToggleChargingLED = yes,
hasFrontlight = yes,
touch_snow_protocol = true,
display_dpi = 300,
hasNaturalLight = yes,
frontlight_settings = {
frontlight_white = "/sys/class/backlight/mxc_msp430.0/brightness",
frontlight_mixer = "/sys/class/backlight/lm3630a_led/color",
nl_min = 0,
nl_max = 10,
nl_inverted = true,
},
}
-- Kobo Forma:
-- NOTE: Right now, we enforce Portrait orientation on startup to avoid getting touch coordinates wrong,
-- no matter the rotation we were started from (c.f., platform/kobo/koreader.sh).
-- NOTE: For the FL, assume brightness is WO, and actual_brightness is RO!
-- i.e., we could have a real KoboPowerD:frontlightIntensityHW() by reading actual_brightness ;).
-- NOTE: Rotation events *may* not be enabled if Nickel has never been brought up in that power cycle.
-- i.e., this will affect KSM users.
-- c.f., https://github.com/koreader/koreader/pull/4414#issuecomment-449652335
-- There's also a CM_ROTARY_ENABLE command, but which seems to do as much nothing as the STATUS one...
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboFrost = Kobo:extend{
model = "Kobo_frost",
isMk7 = yes,
canToggleChargingLED = yes,
hasFrontlight = yes,
hasKeys = yes,
hasGSensor = yes,
touch_snow_protocol = true,
display_dpi = 300,
hasNaturalLight = yes,
frontlight_settings = {
frontlight_white = "/sys/class/backlight/mxc_msp430.0/brightness",
frontlight_mixer = "/sys/class/backlight/tlc5947_bl/color",
-- Warmth goes from 0 to 10 on the device's side (our own internal scale is still normalized to [0...100])
-- NOTE: Those three extra keys are *MANDATORY* if frontlight_mixer is set!
nl_min = 0,
nl_max = 10,
nl_inverted = true,
},
}
2019-09-11 17:40:36 +00:00
-- Kobo Libra:
-- NOTE: Assume the same quirks as the Forma apply.
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboStorm = Kobo:extend{
2019-09-11 17:40:36 +00:00
model = "Kobo_storm",
isMk7 = yes,
canToggleChargingLED = yes,
2019-09-11 17:40:36 +00:00
hasFrontlight = yes,
hasKeys = yes,
hasGSensor = yes,
2019-09-11 17:40:36 +00:00
touch_snow_protocol = true,
display_dpi = 300,
hasNaturalLight = yes,
frontlight_settings = {
frontlight_white = "/sys/class/backlight/mxc_msp430.0/brightness",
frontlight_mixer = "/sys/class/backlight/lm3630a_led/color",
2019-09-11 17:40:36 +00:00
-- Warmth goes from 0 to 10 on the device's side (our own internal scale is still normalized to [0...100])
-- NOTE: Those three extra keys are *MANDATORY* if frontlight_mixer is set!
nl_min = 0,
nl_max = 10,
nl_inverted = true,
},
-- NOTE: The Libra apparently suffers from a mysterious issue where completely innocuous WAIT_FOR_UPDATE_COMPLETE ioctls
-- will mysteriously fail with a timeout (5s)...
2023-01-18 08:15:33 +00:00
-- This obviously leads to *terrible* user experience, so, until more is understood about the issue,
-- bypass this ioctl on this device.
-- c.f., https://github.com/koreader/koreader/issues/7340
hasReliableMxcWaitFor = no,
2019-09-11 17:40:36 +00:00
}
-- Kobo Nia:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboLuna = Kobo:extend{
model = "Kobo_luna",
isMk7 = yes,
canToggleChargingLED = yes,
hasFrontlight = yes,
touch_phoenix_protocol = true,
display_dpi = 212,
hasReliableMxcWaitFor = no, -- Board is similar to the Libra 2, but it's such an unpopular device that reports are scarce.
-- Handle the HW revision w/ a BD71828 PMIC
automagic_sysfs = true,
}
-- Kobo Elipsa
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboEuropa = Kobo:extend{
model = "Kobo_europa",
isSunxi = yes,
hasEclipseWfm = yes,
canToggleChargingLED = yes,
led_uses_channel_3 = true,
hasFrontlight = yes,
hasGSensor = yes,
2021-07-21 20:20:08 +00:00
pressure_event = C.ABS_MT_PRESSURE,
display_dpi = 227,
boot_rota = C.FB_ROTATE_CCW,
battery_sysfs = "/sys/class/power_supply/battery",
ntx_dev = "/dev/input/by-path/platform-ntx_event0-event",
touch_dev = "/dev/input/by-path/platform-0-0010-event",
isSMP = yes,
}
-- Kobo Sage
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboCadmus = Kobo:extend{
model = "Kobo_cadmus",
isSunxi = yes,
hasEclipseWfm = yes,
canToggleChargingLED = yes,
led_uses_channel_3 = true,
hasFrontlight = yes,
hasKeys = yes,
hasGSensor = yes,
pressure_event = C.ABS_MT_PRESSURE,
display_dpi = 300,
hasNaturalLight = yes,
frontlight_settings = {
frontlight_white = "/sys/class/backlight/mxc_msp430.0/brightness",
frontlight_mixer = "/sys/class/leds/aw99703-bl_FL1/color",
-- Warmth goes from 0 to 10 on the device's side (our own internal scale is still normalized to [0...100])
-- NOTE: Those three extra keys are *MANDATORY* if frontlight_mixer is set!
nl_min = 0,
nl_max = 10,
nl_inverted = false,
--- @note: The Sage natively ramps when setting the frontlight intensity.
--- A side-effect of this behavior is that if you queue a series of intensity changes ending at 0,
--- it won't ramp *at all*, and jump straight to zero.
--- So we delay the final ramp off step to prevent (both) the native and our ramping from being optimized out
ramp_off_delay = 0.5,
},
boot_rota = C.FB_ROTATE_CW,
battery_sysfs = "/sys/class/power_supply/battery",
hasAuxBattery = yes,
aux_battery_sysfs = "/sys/class/misc/cilix",
ntx_dev = "/dev/input/by-path/platform-ntx_event0-event",
touch_dev = "/dev/input/by-path/platform-0-0010-event",
isSMP = yes,
-- Much like the Libra 2, there are at least two different HW revisions, with different PMICs...
automagic_sysfs = true,
}
-- Kobo Libra 2:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboIo = Kobo:extend{
model = "Kobo_io",
isMk7 = yes,
hasEclipseWfm = yes,
canToggleChargingLED = yes,
hasFrontlight = yes,
hasKeys = yes,
hasGSensor = yes,
pressure_event = C.ABS_MT_PRESSURE,
touch_mirrored_x = false,
display_dpi = 300,
hasNaturalLight = yes,
frontlight_settings = {
frontlight_white = "/sys/class/backlight/mxc_msp430.0/brightness",
-- Warmth goes from 0 to 10 on the device's side (our own internal scale is still normalized to [0...100])
-- NOTE: Those three extra keys are *MANDATORY* if frontlight_mixer is set!
nl_min = 0,
nl_max = 10,
nl_inverted = true,
},
-- It would appear that the Libra 2 inherited its ancestor's quirks, and more...
-- c.f., https://github.com/koreader/koreader/issues/8414 & https://github.com/koreader/koreader/issues/8664
hasReliableMxcWaitFor = no,
-- NOTE: There are at least two hardware revisions of this device (*without* a device code change, this time),
-- with *significant* hardware changes, so we'll handle this by making the sysfs path discovery automagic.
-- c.f., https://github.com/koreader/koreader/issues/9218
automagic_sysfs = true,
}
-- Kobo Clara 2E:
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
local KoboGoldfinch = Kobo:extend{
model = "Kobo_goldfinch",
isMk7 = yes,
hasEclipseWfm = yes,
canToggleChargingLED = yes,
led_uses_channel_3 = true,
hasFrontlight = yes,
display_dpi = 300,
hasNaturalLight = yes,
frontlight_settings = {
frontlight_white = "/sys/class/backlight/mxc_msp430.0/brightness",
frontlight_mixer = "/sys/class/leds/aw99703-bl_FL1/color",
nl_min = 0,
nl_max = 10,
nl_inverted = true,
},
battery_sysfs = "/sys/class/power_supply/battery",
power_dev = "/dev/input/by-path/platform-bd71828-pwrkey-event",
-- Board is eerily similar to the Libra 2, which, unfortunately, means it's also buggy as hell...
-- c.f., https://github.com/koreader/koreader/issues/9552#issuecomment-1293000313
hasReliableMxcWaitFor = no,
}
function Kobo:setupChargingLED()
if G_reader_settings:nilOrTrue("enable_charging_led") then
if self:hasAuxBattery() and self.powerd:isAuxBatteryConnected() then
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
self:toggleChargingLED(self.powerd:isAuxCharging() and not self.powerd:isAuxCharged())
else
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
self:toggleChargingLED(self.powerd:isCharging() and not self.powerd:isCharged())
end
end
end
function Kobo:getKeyRepeat()
-- Sanity check (mostly for the testsuite's benefit...)
if not self.ntx_fd then
return false
end
self.key_repeat = ffi.new("unsigned int[?]", C.REP_CNT)
if C.ioctl(self.ntx_fd, C.EVIOCGREP, self.key_repeat) < 0 then
local err = ffi.errno()
logger.warn("Device:getKeyRepeat: EVIOCGREP ioctl failed:", ffi.string(C.strerror(err)))
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
return false
else
logger.dbg("Key repeat is set up to repeat every", self.key_repeat[C.REP_PERIOD], "ms after a delay of", self.key_repeat[C.REP_DELAY], "ms")
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
return true
end
end
function Kobo:disableKeyRepeat()
-- NOTE: LuaJIT zero inits, and PERIOD == 0 with DELAY == 0 disables repeats ;).
local key_repeat = ffi.new("unsigned int[?]", C.REP_CNT)
if C.ioctl(self.ntx_fd, C.EVIOCSREP, key_repeat) < 0 then
local err = ffi.errno()
logger.warn("Device:disableKeyRepeat: EVIOCSREP ioctl failed:", ffi.string(C.strerror(err)))
end
end
function Kobo:restoreKeyRepeat()
if C.ioctl(self.ntx_fd, C.EVIOCSREP, self.key_repeat) < 0 then
local err = ffi.errno()
logger.warn("Device:restoreKeyRepeat: EVIOCSREP ioctl failed:", ffi.string(C.strerror(err)))
end
end
function Kobo:init()
-- Check if we need to disable MXCFB_WAIT_FOR_UPDATE_COMPLETE ioctls...
local mxcfb_bypass_wait_for
if G_reader_settings:has("mxcfb_bypass_wait_for") then
mxcfb_bypass_wait_for = G_reader_settings:isTrue("mxcfb_bypass_wait_for")
else
mxcfb_bypass_wait_for = not self:hasReliableMxcWaitFor()
end
if self:isSunxi() then
self.screen = require("ffi/framebuffer_sunxi"):new{
device = self,
debug = logger.dbg,
is_always_portrait = self.isAlwaysPortrait(),
mxcfb_bypass_wait_for = mxcfb_bypass_wait_for,
boot_rota = self.boot_rota,
}
-- Sunxi means no HW inversion :(
self.canHWInvert = no
else
self.screen = require("ffi/framebuffer_mxcfb"):new{
device = self,
debug = logger.dbg,
is_always_portrait = self.isAlwaysPortrait(),
mxcfb_bypass_wait_for = mxcfb_bypass_wait_for,
}
if self.screen.fb_bpp == 32 then
-- Ensure we decode images properly, as our framebuffer is BGRA...
logger.info("Enabling Kobo @ 32bpp BGR tweaks")
self.hasBGRFrameBuffer = yes
end
end
-- Automagic sysfs discovery
if self.automagic_sysfs then
-- Battery
if util.pathExists("/sys/class/power_supply/battery") then
-- Newer devices (circa sunxi)
self.battery_sysfs = "/sys/class/power_supply/battery"
else
self.battery_sysfs = "/sys/class/power_supply/mc13892_bat"
end
-- Frontlight
if self:hasNaturalLight() then
if util.fileExists("/sys/class/leds/aw99703-bl_FL1/color") then
-- HWConfig FL_PWM is AW99703x2
self.frontlight_settings.frontlight_mixer = "/sys/class/leds/aw99703-bl_FL1/color"
elseif util.fileExists("/sys/class/backlight/lm3630a_led/color") then
-- HWConfig FL_PWM is LM3630
self.frontlight_settings.frontlight_mixer = "/sys/class/backlight/lm3630a_led/color"
elseif util.fileExists("/sys/class/backlight/tlc5947_bl/color") then
-- HWConfig FL_PWM is TLC5947
self.frontlight_settings.frontlight_mixer = "/sys/class/backlight/tlc5947_bl/color"
end
end
-- Touch panel input
if util.fileExists("/dev/input/by-path/platform-1-0010-event") then
-- Elan (HWConfig TouchCtrl is ekth6) on i2c bus 1
self.touch_dev = "/dev/input/by-path/platform-1-0010-event"
elseif util.fileExists("/dev/input/by-path/platform-0-0010-event") then
-- Elan (HWConfig TouchCtrl is ekth6) on i2c bus 0
self.touch_dev = "/dev/input/by-path/platform-0-0010-event"
else
self.touch_dev = "/dev/input/event1"
end
-- Physical buttons & synthetic NTX events
if util.fileExists("/dev/input/by-path/platform-gpio-keys-event") then
-- Libra 2 w/ a BD71828 PMIC
self.ntx_dev = "/dev/input/by-path/platform-gpio-keys-event"
elseif util.fileExists("/dev/input/by-path/platform-ntx_event0-event") then
-- sunxi & Mk. 7
self.ntx_dev = "/dev/input/by-path/platform-ntx_event0-event"
elseif util.fileExists("/dev/input/by-path/platform-mxckpd-event") then
-- circa Mk. 5 i.MX
self.ntx_dev = "/dev/input/by-path/platform-mxckpd-event"
else
self.ntx_dev = "/dev/input/event0"
end
-- Power button (this usually ends up in ntx_dev, except with some PMICs)
if util.fileExists("/dev/input/by-path/platform-bd71828-pwrkey-event") then
-- Libra 2 & Nia w/ a BD71828 PMIC
self.power_dev = "/dev/input/by-path/platform-bd71828-pwrkey-event"
elseif util.fileExists("/dev/input/by-path/platform-bd71828-pwrkey.4.auto-event") then
-- Sage w/ a BD71828 PMIC
self.power_dev = "/dev/input/by-path/platform-bd71828-pwrkey.4.auto-event"
end
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
-- NOTE: i.MX5 devices have a wonky RTC that doesn't like alarms set further away that UINT16_MAX seconds from now...
-- (c.f., WakeupMgr for more details).
-- NOTE: getRTCName is currently hardcoded to rtc0 (which is also WakeupMgr's default).
local dodgy_rtc = false
if getRTCName() == "pmic_rtc" then
-- This *should* match the 'RTC' (46) NTX HWConfig field being set to 'MSP430' (0).
dodgy_rtc = true
end
-- Detect the various CPU governor sysfs knobs...
if util.pathExists("/sys/devices/system/cpu/cpufreq/policy0") then
self.cpu_governor_knob = "/sys/devices/system/cpu/cpufreq/policy0/scaling_governor"
else
self.cpu_governor_knob = "/sys/devices/system/cpu/cpu0/cpufreq/scaling_governor"
end
self.default_cpu_governor = getCPUGovernor(self.cpu_governor_knob)
-- NOP unsupported methods
if not self.default_cpu_governor then
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
self.performanceCPUGovernor = NOP
self.defaultCPUGovernor = NOP
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
end
-- And while we're on CPU-related endeavors...
self.cpu_count = self:isSMP() and getCPUCount() or 1
-- NOP unsupported methods
if self.cpu_count == 1 then
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
self.enableCPUCores = NOP
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
end
-- Automagically set this so we never have to remember to do it manually ;p
if self:hasNaturalLight() and self.frontlight_settings and self.frontlight_settings.frontlight_mixer then
self.hasNaturalLightMixer = yes
end
-- Ditto
if self:isMk7() then
self.canHWDither = yes
end
self.powerd = require("device/kobo/powerd"):new{
device = self,
battery_sysfs = self.battery_sysfs,
aux_battery_sysfs = self.aux_battery_sysfs,
}
-- NOTE: For the Forma, with the buttons on the right, 193 is Top, 194 Bottom.
self.input = require("device/input"):new{
device = self,
event_map = {
[35] = "SleepCover", -- KEY_H, Elipsa
[59] = "SleepCover",
[90] = "LightButton",
[102] = "Home",
[116] = "Power",
[193] = "RPgBack",
[194] = "RPgFwd",
[331] = "Eraser",
[332] = "Highlighter",
},
event_map_adapter = {
SleepCover = function(ev)
if self.input:isEvKeyPress(ev) then
return "SleepCoverClosed"
elseif self.input:isEvKeyRelease(ev) then
return "SleepCoverOpened"
end
end,
LightButton = function(ev)
if self.input:isEvKeyRelease(ev) then
return "Light"
end
end,
2021-07-21 20:20:08 +00:00
},
main_finger_slot = self.main_finger_slot or 0,
pressure_event = self.pressure_event,
}
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
self.wakeup_mgr = WakeupMgr:new{
dodgy_rtc = dodgy_rtc,
}
-- Input handling on Kobo is a thing of nightmares, start by setting up the actual evdev handler...
self:setTouchEventHandler()
-- And then handle the extra shenanigans if necessary.
self:initEventAdjustHooks()
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
-- Let generic properly setup the standard stuff
-- (*after* we've set our input hooks, so that the viewport translation runs last).
2016-04-03 04:52:30 +00:00
Generic.init(self)
-- Various HW Buttons, Switches & Synthetic NTX events
self.ntx_fd = self.input.open(self.ntx_dev)
-- Dedicated Power Button input device (if any)
if self.power_dev then
self.input.open(self.power_dev)
end
-- Touch panel
self.input.open(self.touch_dev)
-- NOTE: On devices with a gyro, there may be a dedicated input device outputting the raw accelerometer data
-- (3-Axis Orientation/Motion Detection).
-- We skip it because we don't need it (synthetic rotation change events are sent to the main ntx input device),
-- and it's usually *extremely* verbose, so it'd just be a waste of processing power.
-- fake_events is only used for usb plug & charge events so far (generated via uevent, c.f., input/iput-kobo.h in base).
-- NOTE: usb hotplug event is also available in /tmp/nickel-hardware-status (... but only when Nickel is running ;p)
2016-10-22 07:04:37 +00:00
self.input.open("fake_events")
2016-04-03 04:52:30 +00:00
-- See if the device supports key repeat
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
if not self:getKeyRepeat() then
-- NOP unsupported methods
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
self.disableKeyRepeat = NOP
self.restoreKeyRepeat = NOP
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
end
-- Detect the NTX charging LED sysfs knob
if util.pathExists("/sys/devices/platform/ntx_led/lit") then
self.ntx_lit_sysfs_knob = "/sys/devices/platform/ntx_led/lit"
elseif util.pathExists("/sys/devices/platform/pmic_light.1/lit") then
self.ntx_lit_sysfs_knob = "/sys/devices/platform/pmic_light.1/lit"
else
self.canToggleChargingLED = no
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
-- NOP unsupported methods
if not self:canToggleChargingLED() then
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
self.toggleChargingLED = NOP
self.setupChargingLED = NOP
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
end
-- We have no way of querying the current state of the charging LED, so, start from scratch.
-- Much like Nickel, start by turning it off.
self:toggleChargingLED(false)
self:setupChargingLED()
-- Only enable a single core on startup
self:enableCPUCores(1)
self.canStandby = checkStandby()
if self.canStandby() and (self:isMk7() or self:isSunxi()) then
self.canPowerSaveWhileCharging = yes
end
-- Check if the device has a Neonode IR grid (to tone down the chatter on resume ;)).
if lfs.attributes("/sys/devices/virtual/input/input1/neocmd", "mode") == "file" then
-- As found on (at least), the Aura H2O
self.hasIRGridSysfsKnob = "/sys/devices/virtual/input/input1/neocmd"
elseif lfs.attributes("/sys/devices/platform/imx-i2c.1/i2c-1/1-0050/neocmd", "mode") == "file" then
-- As found on (at least) the Glo HD (c.f., https://github.com/koreader/koreader/pull/9377#issuecomment-1213544478)
self.hasIRGridSysfsKnob = "/sys/devices/platform/imx-i2c.1/i2c-1/1-0050/neocmd"
end
2016-04-03 04:52:30 +00:00
end
2017-09-18 17:04:36 +00:00
function Kobo:setDateTime(year, month, day, hour, min, sec)
if hour == nil or min == nil then return true end
2017-09-18 17:04:36 +00:00
local command
if year and month and day then
command = string.format("date -s '%d-%d-%d %d:%d:%d'", year, month, day, hour, min, sec)
else
command = string.format("date -s '%d:%d'",hour, min)
end
if os.execute(command) == 0 then
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
os.execute("hwclock -u -w")
return true
else
return false
end
end
function Kobo:initNetworkManager(NetworkMgr)
function NetworkMgr:turnOffWifi(complete_callback)
Various Wi-Fi QoL improvements (#6424) * Revamped most actions that require an internet connection to a new/fixed backend that allows forwarding the initial action and running it automatically once connected. (i.e., it'll allow you to set "Action when Wi-Fi is off" to "turn_on", and whatch stuff connect and do what you wanted automatically without having to re-click anywhere instead of showing you a Wi-Fi prompt and then not doing anything without any other feedback). * Speaking of, fixed the "turn_on" beforeWifi action to, well, actually work. It's no longer marked as experimental. * Consistently use "Wi-Fi" everywhere. * On Kobo/Cervantes/Sony, implemented a "Kill Wi-Fi connection when inactive" system that will automatically disconnect from Wi-Fi after sustained *network* inactivity (i.e., you can keep reading, it'll eventually turn off on its own). This should be smart and flexible enough not to murder Wi-Fi while you need it, while still not keeping it uselessly on and murdering your battery. (i.e., enable that + turn Wi-Fi on when off and enjoy never having to bother about Wi-Fi ever again). * Made sending `NetworkConnected` / `NetworkDisconnected` events consistent (they were only being sent... sometimes, which made relying on 'em somewhat problematic). * restoreWifiAsync is now only run when really needed (i.e., we no longer stomp on an existing working connection just for the hell of it). * We no longer attempt to kill a bogus non-existent Wi-Fi connection when going to suspend, we only do it when it's actually needed. * Every method of enabling Wi-Fi will now properly tear down Wi-Fi on failure, instead of leaving it in an undefined state. * Fixed an issue in the fancy crash screen on Kobo/reMarkable that could sometime lead to the log excerpt being missing. * Worked-around a number of sneaky issues related to low-level Wi-Fi/DHCP/DNS handling on Kobo (see the lengthy comments [below](https://github.com/koreader/koreader/pull/6424#issuecomment-663881059) for details). Fix #6421 Incidentally, this should also fix the inconsistencies experienced re: Wi-Fi behavior in Nickel when toggling between KOReader and Nickel (use NM/KFMon, and run a current FW for best results). * For developers, this involves various cleanups around NetworkMgr and NetworkListener. Documentation is in-line, above the concerned functions.
2020-07-27 01:39:06 +00:00
self:releaseIP()
koboEnableWifi(false)
if complete_callback then
complete_callback()
end
end
function NetworkMgr:turnOnWifi(complete_callback)
Various Wi-Fi QoL improvements (#6424) * Revamped most actions that require an internet connection to a new/fixed backend that allows forwarding the initial action and running it automatically once connected. (i.e., it'll allow you to set "Action when Wi-Fi is off" to "turn_on", and whatch stuff connect and do what you wanted automatically without having to re-click anywhere instead of showing you a Wi-Fi prompt and then not doing anything without any other feedback). * Speaking of, fixed the "turn_on" beforeWifi action to, well, actually work. It's no longer marked as experimental. * Consistently use "Wi-Fi" everywhere. * On Kobo/Cervantes/Sony, implemented a "Kill Wi-Fi connection when inactive" system that will automatically disconnect from Wi-Fi after sustained *network* inactivity (i.e., you can keep reading, it'll eventually turn off on its own). This should be smart and flexible enough not to murder Wi-Fi while you need it, while still not keeping it uselessly on and murdering your battery. (i.e., enable that + turn Wi-Fi on when off and enjoy never having to bother about Wi-Fi ever again). * Made sending `NetworkConnected` / `NetworkDisconnected` events consistent (they were only being sent... sometimes, which made relying on 'em somewhat problematic). * restoreWifiAsync is now only run when really needed (i.e., we no longer stomp on an existing working connection just for the hell of it). * We no longer attempt to kill a bogus non-existent Wi-Fi connection when going to suspend, we only do it when it's actually needed. * Every method of enabling Wi-Fi will now properly tear down Wi-Fi on failure, instead of leaving it in an undefined state. * Fixed an issue in the fancy crash screen on Kobo/reMarkable that could sometime lead to the log excerpt being missing. * Worked-around a number of sneaky issues related to low-level Wi-Fi/DHCP/DNS handling on Kobo (see the lengthy comments [below](https://github.com/koreader/koreader/pull/6424#issuecomment-663881059) for details). Fix #6421 Incidentally, this should also fix the inconsistencies experienced re: Wi-Fi behavior in Nickel when toggling between KOReader and Nickel (use NM/KFMon, and run a current FW for best results). * For developers, this involves various cleanups around NetworkMgr and NetworkListener. Documentation is in-line, above the concerned functions.
2020-07-27 01:39:06 +00:00
koboEnableWifi(true)
self:reconnectOrShowNetworkMenu(complete_callback)
end
NetworkMgr: Refine isConnected check (#10098) Much easier to deal with thanks to the cleanup work done in #10062 ;). * `carrier` is set to 1 as soon as the device is *administratively* up (in practice, as soon as we run `ifconfig up`). This is perfectly fine for `isWifiOn`, but absolutely not for `isConnected`, because we are not, actually, connected to *anything*, no attempt at associating has even been made at that point. Besides being semantically wrong, in practice, this will horribly break the connectivity check, because it expects that `isConnected` means we can talk to at least the LAN. * Delving into the Linux docs reveals that `operstate` looks like a better candidate, as it reflects *operational status*; for Wi-Fi, that means associated and successfully authenticated. That's... closer, but still not it, because we still don't have an IP, so we technically can't talk to anything other than the AP. * So, I've brought out the big guns (`getifaddrs`), and replicated a bit of code that I already use in the USBNetwork hack on Kindle, to detect whether we actually have an IP assigned. (Other approaches, like `/proc/net/route`, may not be entirely fool-proof, and/or get complicated when IPv6 enters the fray (which it does, on Kobo, Mk. 8+ devices are IPv6-enabled)). TL;DR: Bunch of C via ffi, and `isConnected` now returns true only when the device is operationally up *and* we have an IP assigned. Pulls in https://github.com/koreader/koreader-base/pull/1579 & https://github.com/koreader/lj-wpaclient/pull/10
2023-02-10 22:47:08 +00:00
local net_if = os.getenv("INTERFACE") or "eth0"
Various Wi-Fi QoL improvements (#6424) * Revamped most actions that require an internet connection to a new/fixed backend that allows forwarding the initial action and running it automatically once connected. (i.e., it'll allow you to set "Action when Wi-Fi is off" to "turn_on", and whatch stuff connect and do what you wanted automatically without having to re-click anywhere instead of showing you a Wi-Fi prompt and then not doing anything without any other feedback). * Speaking of, fixed the "turn_on" beforeWifi action to, well, actually work. It's no longer marked as experimental. * Consistently use "Wi-Fi" everywhere. * On Kobo/Cervantes/Sony, implemented a "Kill Wi-Fi connection when inactive" system that will automatically disconnect from Wi-Fi after sustained *network* inactivity (i.e., you can keep reading, it'll eventually turn off on its own). This should be smart and flexible enough not to murder Wi-Fi while you need it, while still not keeping it uselessly on and murdering your battery. (i.e., enable that + turn Wi-Fi on when off and enjoy never having to bother about Wi-Fi ever again). * Made sending `NetworkConnected` / `NetworkDisconnected` events consistent (they were only being sent... sometimes, which made relying on 'em somewhat problematic). * restoreWifiAsync is now only run when really needed (i.e., we no longer stomp on an existing working connection just for the hell of it). * We no longer attempt to kill a bogus non-existent Wi-Fi connection when going to suspend, we only do it when it's actually needed. * Every method of enabling Wi-Fi will now properly tear down Wi-Fi on failure, instead of leaving it in an undefined state. * Fixed an issue in the fancy crash screen on Kobo/reMarkable that could sometime lead to the log excerpt being missing. * Worked-around a number of sneaky issues related to low-level Wi-Fi/DHCP/DNS handling on Kobo (see the lengthy comments [below](https://github.com/koreader/koreader/pull/6424#issuecomment-663881059) for details). Fix #6421 Incidentally, this should also fix the inconsistencies experienced re: Wi-Fi behavior in Nickel when toggling between KOReader and Nickel (use NM/KFMon, and run a current FW for best results). * For developers, this involves various cleanups around NetworkMgr and NetworkListener. Documentation is in-line, above the concerned functions.
2020-07-27 01:39:06 +00:00
function NetworkMgr:getNetworkInterfaceName()
return net_if
end
NetworkMgr:setWirelessBackend(
"wpa_supplicant", {ctrl_interface = "/var/run/wpa_supplicant/" .. net_if})
function NetworkMgr:obtainIP()
os.execute("./obtain-ip.sh")
end
function NetworkMgr:releaseIP()
os.execute("./release-ip.sh")
end
function NetworkMgr:restoreWifiAsync()
os.execute("./restore-wifi-async.sh")
end
NetworkMgr.isWifiOn = NetworkMgr.sysfsWifiOn
NetworkMgr: Refine isConnected check (#10098) Much easier to deal with thanks to the cleanup work done in #10062 ;). * `carrier` is set to 1 as soon as the device is *administratively* up (in practice, as soon as we run `ifconfig up`). This is perfectly fine for `isWifiOn`, but absolutely not for `isConnected`, because we are not, actually, connected to *anything*, no attempt at associating has even been made at that point. Besides being semantically wrong, in practice, this will horribly break the connectivity check, because it expects that `isConnected` means we can talk to at least the LAN. * Delving into the Linux docs reveals that `operstate` looks like a better candidate, as it reflects *operational status*; for Wi-Fi, that means associated and successfully authenticated. That's... closer, but still not it, because we still don't have an IP, so we technically can't talk to anything other than the AP. * So, I've brought out the big guns (`getifaddrs`), and replicated a bit of code that I already use in the USBNetwork hack on Kindle, to detect whether we actually have an IP assigned. (Other approaches, like `/proc/net/route`, may not be entirely fool-proof, and/or get complicated when IPv6 enters the fray (which it does, on Kobo, Mk. 8+ devices are IPv6-enabled)). TL;DR: Bunch of C via ffi, and `isConnected` now returns true only when the device is operationally up *and* we have an IP assigned. Pulls in https://github.com/koreader/koreader-base/pull/1579 & https://github.com/koreader/lj-wpaclient/pull/10
2023-02-10 22:47:08 +00:00
NetworkMgr.isConnected = NetworkMgr.ifHasAnAddress
-- Usually handled in NetworkMgr:init, but we'll need it *now*
NetworkMgr.interface = net_if
-- Kill Wi-Fi if NetworkMgr:isWifiOn() and NOT NetworkMgr:isConnected()
-- (i.e., if the launcher left the Wi-Fi in an inconsistent state: modules loaded, but no route to gateway).
if NetworkMgr:isWifiOn() and not NetworkMgr:isConnected() then
logger.info("Kobo Wi-Fi: Left in an inconsistent state by launcher!")
NetworkMgr:turnOffWifi()
end
end
function Kobo:setTouchEventHandler()
if self.touch_snow_protocol then
self.input.snow_protocol = true
elseif self.touch_phoenix_protocol then
self.input.handleTouchEv = self.input.handleTouchEvPhoenix
elseif not self:hasMultitouch() then
self.input.handleTouchEv = self.input.handleTouchEvLegacy
if self.touch_kobo_mk3_protocol then
self.input.touch_kobo_mk3_protocol = true
end
end
-- Accelerometer
if self:hasGSensor() then
self.input.handleMiscEv = function(this, ev)
-- As generated by gyroTranslation below
if ev.code == C.MSC_GYRO then
return this:handleGyroEv(ev)
end
end
end
end
-- HAL for gyro orientation switches (NTX's EV_MSC:MSC_RAW w/ custom values to EV_MSC:MSC_GYRO w/ our own custom values)
local function gyroTranslation(ev)
-- c.f., include/uapi/linux/input.h,
-- implementations in drivers/hwmon/mma8x5x.c & drivers/input/touchscreen/kx122.c
local MSC_RAW_GSENSOR_PORTRAIT_DOWN = 0x17
local MSC_RAW_GSENSOR_PORTRAIT_UP = 0x18
local MSC_RAW_GSENSOR_LANDSCAPE_RIGHT = 0x19
local MSC_RAW_GSENSOR_LANDSCAPE_LEFT = 0x1a
-- Not that we care about those, but they are reported, and accurate ;).
--[[
local MSC_RAW_GSENSOR_BACK = 0x1b
local MSC_RAW_GSENSOR_FRONT = 0x1c
--]]
if ev.value == MSC_RAW_GSENSOR_PORTRAIT_UP then
-- i.e., UR
ev.code = C.MSC_GYRO
ev.value = C.DEVICE_ROTATED_UPRIGHT
elseif ev.value == MSC_RAW_GSENSOR_LANDSCAPE_RIGHT then
-- i.e., CW
ev.code = C.MSC_GYRO
ev.value = C.DEVICE_ROTATED_CLOCKWISE
elseif ev.value == MSC_RAW_GSENSOR_PORTRAIT_DOWN then
-- i.e., UD
ev.code = C.MSC_GYRO
ev.value = C.DEVICE_ROTATED_UPSIDE_DOWN
elseif ev.value == MSC_RAW_GSENSOR_LANDSCAPE_LEFT then
-- i.e., CCW
ev.code = C.MSC_GYRO
ev.value = C.DEVICE_ROTATED_COUNTER_CLOCKWISE
end
end
function Kobo:initEventAdjustHooks()
-- Build a single composite hook, to avoid duplicated branches...
local koboInputMangling
-- NOTE: touch_switch_xy is *always* true, but not touch_mirrored_x...
if self.touch_switch_xy and self.touch_mirrored_x then
local max_x = self.screen:getWidth() - 1
koboInputMangling = function(this, ev)
if ev.type == C.EV_ABS then
this:adjustABS_SwitchAxesAndMirrorX(ev, max_x)
elseif ev.type == C.EV_MSC and ev.code == C.MSC_RAW then
gyroTranslation(ev)
end
end
elseif self.touch_switch_xy and not self.touch_mirrored_x then
koboInputMangling = function(this, ev)
if ev.type == C.EV_ABS then
this:adjustABS_SwitchXY(ev)
elseif ev.type == C.EV_MSC and ev.code == C.MSC_RAW then
gyroTranslation(ev)
end
end
end
if koboInputMangling then
self.input:registerEventAdjustHook(koboInputMangling)
end
end
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
local function getCodeName()
-- Try to get it from the env first
local codename = os.getenv("PRODUCT")
-- If that fails, run the script ourselves
if not codename then
local std_out = io.popen("/bin/kobo_config.sh 2>/dev/null", "re")
if std_out then
codename = std_out:read("*line")
std_out:close()
end
end
return codename
end
function Kobo:getFirmwareVersion()
local version_file = io.open("/mnt/onboard/.kobo/version", "re")
if not version_file then
self.firmware_rev = "none"
return
end
local version_str = version_file:read("*line")
version_file:close()
local i = 1
for field in version_str:gmatch("([^,]+)") do
if i == 3 then
self.firmware_rev = field
break
end
i = i + 1
end
end
local function getProductId()
-- Try to get it from the env first (KSM only)
local product_id = os.getenv("MODEL_NUMBER")
-- If that fails, devise it ourselves
if not product_id then
local version_file = io.open("/mnt/onboard/.kobo/version", "re")
if not version_file then
return "000"
end
local version_str = version_file:read("*line")
version_file:close()
product_id = string.sub(version_str, -3, -1)
end
return product_id
end
-- NOTE: We overload this to make sure checkUnexpectedWakeup doesn't trip *before* the newly scheduled suspend
function Kobo:rescheduleSuspend()
UIManager:unschedule(self.suspend)
UIManager:unschedule(self.checkUnexpectedWakeup)
UIManager:scheduleIn(self.suspend_wait_timeout, self.suspend, self)
end
function Kobo:checkUnexpectedWakeup()
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- Just in case another event like SleepCoverClosed also scheduled a suspend
UIManager:unschedule(self.suspend)
-- The proximity window is rather large, because we're scheduled to run 15 seconds after resuming,
-- so we're already guaranteed to be at least 15s away from the alarm ;).
if self.wakeup_mgr:isWakeupAlarmScheduled() and self.wakeup_mgr:wakeupAction(30) then
-- Assume we want to go back to sleep after running the scheduled action
-- (Kobo:resume will unschedule this on an user-triggered resume).
logger.info("Kobo suspend: scheduled wakeup; the device will go back to sleep in 30s.")
-- We need significant leeway for the poweroff action to send out close events to all requisite widgets,
-- since we don't actually want to suspend behind its back ;).
UIManager:scheduleIn(30, self.suspend, self)
else
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- We've hit an early resume, assume this is unexpected (as we only run if Kobo:resume hasn't already).
logger.dbg("Kobo suspend: checking unexpected wakeup number", self.unexpected_wakeup_count)
if self.unexpected_wakeup_count > 20 then
-- If we've failed to put the device back to sleep over 20 consecutive times, we give up.
-- Broadcast a specific event, so that AutoSuspend can pick up the baton...
local Event = require("ui/event")
UIManager:broadcastEvent(Event:new("UnexpectedWakeupLimit"))
return
end
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
logger.err("Kobo suspend: putting device back to sleep after", self.unexpected_wakeup_count, "unexpected wakeups.")
self:suspend()
end
end
--- The function to put the device into standby, with enabled touchscreen.
-- max_duration ... maximum time for the next standby, can wake earlier (e.g. Tap, Button ...)
function Kobo:standby(max_duration)
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
-- NOTE: Switch to the performance CPU governor, in order to speed up the resume process so as to lower its latency cost...
-- (It won't have any impact on power efficiency *during* suspend, so there's not really any drawback).
self:performanceCPUGovernor()
--[[
-- On most devices, attempting to PM with a Wi-Fi module loaded will horribly crash the kernel, so, don't?
-- NOTE: Much like suspend, our caller should ensure this never happens, hence this being commented out ;).
if NetworkMgr:isWifiOn() then
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
-- AutoSuspend relies on NetworkMgr:getWifiState to prevent this, so, if we ever trip this, it's a bug ;).
logger.err("Kobo standby: cannot standby with Wi-Fi modules loaded! (NetworkMgr is confused: this is a bug)")
return
end
--]]
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- We don't really have anything to schedule, we just need an alarm out of WakeupMgr ;).
local function standby_alarm()
end
if max_duration then
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
self.wakeup_mgr:addTask(max_duration, standby_alarm)
end
logger.dbg("Kobo standby: asking to enter standby . . .")
local standby_time = time.boottime_or_realtime_coarse()
local ret = util.writeToSysfs("standby", "/sys/power/state")
self.last_standby_time = time.boottime_or_realtime_coarse() - standby_time
self.total_standby_time = self.total_standby_time + self.last_standby_time
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
if ret then
logger.dbg("Kobo standby: zZz zZz zZz zZz... And woke up!")
if G_reader_settings:isTrue("pm_debug_entry_failure") then
-- NOTE: This is a debug option where we coopt the charging LED, hence us not using setupChargingLED here.
-- (It's called on resume anyway).
self:toggleChargingLED(false)
end
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
else
logger.warn("Kobo standby: the kernel refused to enter standby!")
if G_reader_settings:isTrue("pm_debug_entry_failure") then
self:toggleChargingLED(true)
end
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
end
if max_duration then
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- NOTE: We don't actually care about discriminating exactly *why* we woke up,
-- and our scheduled wakeup action is a NOP anyway,
-- so we can just drop the task instead of doing things the right way like suspend ;).
-- This saves us some pointless RTC shenanigans, so, everybody wins.
--[[
-- There's no scheduling shenanigans like in suspend, so the proximity window can be much tighter...
if self.wakeup_mgr:isWakeupAlarmScheduled() and self.wakeup_mgr:wakeupAction(5) then
-- We tripped the standby alarm, UIManager will be able to run whatever was actually scheduled,
-- and AutoSuspend will handle going back to standby if necessary.
logger.dbg("Kobo standby: tripped rtc wake alarm")
end
--]]
self.wakeup_mgr:removeTasks(nil, standby_alarm)
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
-- And restore the standard CPU scheduler once we're done dealing with the wakeup event.
UIManager:tickAfterNext(self.defaultCPUGovernor, self)
end
function Kobo:suspend()
logger.info("Kobo suspend: going to sleep . . .")
UIManager:unschedule(self.checkUnexpectedWakeup)
-- NOTE: Sleep as little as possible here, sleeping has a tendency to make
-- everything mysteriously hang...
-- Depending on device/FW version, some kernels do not support
-- wakeup_count, account for that
--
-- NOTE: ... and of course, it appears to be broken, which probably
-- explains why nickel doesn't use this facility...
-- (By broken, I mean that the system wakes up right away).
-- So, unless that changes, unconditionally disable it.
--[[
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
local f, re, err_msg, err_code
local has_wakeup_count = false
f = io.open("/sys/power/wakeup_count", "re")
if f ~= nil then
f:close()
has_wakeup_count = true
end
-- Clear the kernel ring buffer... (we're missing a proper -C flag...)
--dmesg -c >/dev/null
-- Go to sleep
local curr_wakeup_count
if has_wakeup_count then
curr_wakeup_count = "$(cat /sys/power/wakeup_count)"
logger.dbg("Kobo suspend: Current WakeUp count:", curr_wakeup_count)
end
-]]
-- NOTE: Sets gSleep_Mode_Suspend to 1. Used as a flag throughout the
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- kernel to suspend/resume various subsystems
-- c.f., state_extended_store @ kernel/power/main.c
local ret = util.writeToSysfs("1", "/sys/power/state-extended")
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
if ret then
logger.dbg("Kobo suspend: successfully asked the kernel to put subsystems to sleep")
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
else
logger.warn("Kobo suspend: the kernel refused to flag subsystems for suspend!")
end
ffiUtil.sleep(2)
logger.dbg("Kobo suspend: waited for 2s because of reasons...")
os.execute("sync")
logger.dbg("Kobo suspend: synced FS")
--[[
if has_wakeup_count then
f = io.open("/sys/power/wakeup_count", "we")
if not f then
logger.err("cannot open /sys/power/wakeup_count")
return false
end
re, err_msg, err_code = f:write(tostring(curr_wakeup_count), "\n")
logger.dbg("Kobo suspend: wrote WakeUp count:", curr_wakeup_count)
if not re then
logger.err("Kobo suspend: failed to write WakeUp count:",
err_msg,
err_code)
end
f:close()
end
--]]
logger.dbg("Kobo suspend: asking for a suspend to RAM . . .")
local suspend_time = time.boottime_or_realtime_coarse()
ret = util.writeToSysfs("mem", "/sys/power/state")
-- NOTE: At this point, we *should* be in suspend to RAM, as such,
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- execution should only resume on wakeup...
self.last_suspend_time = time.boottime_or_realtime_coarse() - suspend_time
self.total_suspend_time = self.total_suspend_time + self.last_suspend_time
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
if ret then
logger.info("Kobo suspend: ZzZ ZzZ ZzZ... And woke up!")
if G_reader_settings:isTrue("pm_debug_entry_failure") then
self:toggleChargingLED(false)
end
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
else
logger.warn("Kobo suspend: the kernel refused to enter suspend!")
-- Reset state-extended back to 0 since we are giving up.
util.writeToSysfs("0", "/sys/power/state-extended")
if G_reader_settings:isTrue("pm_debug_entry_failure") then
self:toggleChargingLED(true)
end
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
end
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- NOTE: Ideally, we'd need a way to warn the user that suspending
-- gloriously failed at this point...
-- We can safely assume that just from a non-zero return code, without
-- looking at the detailed stderr message
-- (most of the failures we'll see are -EBUSY anyway)
-- For reference, when that happens to nickel, it appears to keep retrying
-- to wakeup & sleep ad nauseam,
-- which is where the non-sensical 1 -> mem -> 0 loop idea comes from...
-- cf. nickel_suspend_strace.txt for more details.
-- NOTE: On recent enough kernels, with debugfs enabled and mounted, see also
-- /sys/kernel/debug/suspend_stats & /sys/kernel/debug/wakeup_sources
--[[
if has_wakeup_count then
logger.dbg("wakeup count: $(cat /sys/power/wakeup_count)")
end
-- Print tke kernel log since our attempt to sleep...
--dmesg -c
--]]
-- NOTE: We unflag /sys/power/state-extended in Kobo:resume() to keep
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- things tidy and easier to follow
-- Kobo:resume() will reset unexpected_wakeup_count and unschedule the check to signal a sane wakeup.
self.unexpected_wakeup_count = self.unexpected_wakeup_count + 1
logger.dbg("Kobo suspend: scheduling unexpected wakeup guard")
UIManager:scheduleIn(15, self.checkUnexpectedWakeup, self)
end
function Kobo:resume()
logger.dbg("Kobo resume: clean up after wakeup")
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- Reset unexpected_wakeup_count ASAP
self.unexpected_wakeup_count = 0
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- Unschedule the checkUnexpectedWakeup shenanigans.
UIManager:unschedule(self.checkUnexpectedWakeup)
UIManager:unschedule(self.suspend)
-- Now that we're up, unflag subsystems for suspend...
-- NOTE: Sets gSleep_Mode_Suspend to 0. Used as a flag throughout the
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- kernel to suspend/resume various subsystems
-- cf. kernel/power/main.c @ L#207
-- Among other things, this sets up the wakeup pins (e.g., resume on input).
local ret = util.writeToSysfs("0", "/sys/power/state-extended")
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
if ret then
logger.dbg("Kobo resume: successfully asked the kernel to resume subsystems")
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
else
logger.warn("Kobo resume: the kernel refused to flag subsystems for resume!")
end
-- HACK: wait a bit (0.1 sec) for the kernel to catch up
ffiUtil.usleep(100000)
if self.hasIRGridSysfsKnob then
-- cf. #1862, I can reliably break IR touch input on resume...
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
-- cf. also #1943 for the rationale behind applying this workaround in every case...
-- c.f., neo_ctl @ drivers/input/touchscreen/zforce_i2c.c,
-- basically, a is wakeup (for activate), d is sleep (for deactivate), and we don't care about s (set res),
-- and l (led signal level, actually a NOP on NTX kernels).
util.writeToSysfs("a", self.hasIRGridSysfsKnob)
end
-- A full suspend may have toggled the LED off.
self:setupChargingLED()
end
function Kobo:usbPlugOut()
-- Rewind the unexpected wakeup counter, since we're no longer charging, meaning power savings are now critical again ;).
-- NOTE: We don't reset it to 0 because, semantically, only resume should ever be allowed to do so.
if self.unexpected_wakeup_count > 0 then
self.unexpected_wakeup_count = 1
end
end
function Kobo:saveSettings()
-- save frontlight state to G_reader_settings (and NickelConf if needed)
self.powerd:saveSettings()
end
function Kobo:powerOff()
-- Much like Nickel itself, disable the RTC alarm before powering down.
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
self.wakeup_mgr:unsetWakeupAlarm()
if self:isSunxi() then
-- On sunxi, apparently, we *do* go through init
os.execute("sleep 1 && poweroff &")
else
-- Then shut down without init's help
os.execute("sleep 1 && poweroff -f &")
end
end
function Kobo:reboot()
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
os.execute("sleep 1 && reboot &")
end
function Kobo:toggleChargingLED(toggle)
-- We have no way of querying the current state from the HW!
if toggle == nil then
return
end
-- Don't do anything if the state is already correct
-- NOTE: What happens to the LED when attempting/successfully entering PM is... kind of a mess.
-- On a H2O, even *attempting* to enter PM will kill the light (and it'll stay off).
-- On a Forma, a failed attempt will *not* affect the light, but a successful one *will* kill it,
-- be that standby or suspend, but it'll be restored on wakeup...
-- On sunxi, PM appears to have zero effect on the LED.
if self.charging_led_state == toggle then
return
end
self.charging_led_state = toggle
logger.dbg("Kobo: Turning the charging LED", toggle and "on" or "off")
-- NOTE: While most/all Kobos actually have a charging LED, and it can usually be fiddled with in a similar fashion,
-- we've seen *extremely* weird behavior in the past when playing with it on older devices (c.f., #5479).
-- In fact, Nickel itself doesn't provide this feature on said older devices
-- (when it does, it's an option in the Energy saving settings),
-- which is why we also limit ourselves to "true" on devices where this was tested.
-- c.f., drivers/misc/ntx_misc_light.c
local fd = C.open(self.ntx_lit_sysfs_knob, bit.bor(C.O_WRONLY, C.O_CLOEXEC)) -- procfs/sysfs, we shouldn't need O_TRUNC
if fd == -1 then
logger.err("Cannot open file `" .. self.ntx_lit_sysfs_knob .. "`:", ffi.string(C.strerror(ffi.errno())))
return false
end
-- c.f., strace -fittTvyy -e trace=ioctl,file,signal,ipc,desc -s 256 -o /tmp/nickel.log -p $(pidof -s nickel) &
-- This was observed on a Forma, so I'm mildly hopeful that it's safe on other Mk. 7 devices ;).
-- NOTE: ch stands for channel, cur for current, dc for duty cycle. c.f., the driver source.
if toggle == true then
-- NOTE: Technically, Nickel forces a toggle off before that, too.
-- But since we do that on startup, it shouldn't be necessary here...
for ch = self.led_uses_channel_3 and 3 or 4, 4 do
C.write(fd, "ch " .. tostring(ch), 4)
C.write(fd, "cur 1", 5)
C.write(fd, "dc 63", 5)
end
else
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
for ch = 3, 5 do
C.write(fd, "ch " .. tostring(ch), 4)
C.write(fd, "cur 1", 5)
C.write(fd, "dc 0", 4)
Clarify our OOP semantics across the codebase (#9586) Basically: * Use `extend` for class definitions * Use `new` for object instantiations That includes some minor code cleanups along the way: * Updated `Widget`'s docs to make the semantics clearer. * Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283) * Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass). * Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events. * Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier. * Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references. * ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak). * Terminal: Make sure the shell is killed on plugin teardown. * InputText: Fix Home/End/Del physical keys to behave sensibly. * InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...). * OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of. * ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed! * Kobo: Minor code cleanups.
2022-10-06 00:14:48 +00:00
end
end
C.close(fd)
end
-- Return the highest core number
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
function Kobo:getCPUCount()
local fd = io.open("/sys/devices/system/cpu/possible", "re")
if fd then
local str = fd:read("*line")
fd:close()
-- Format is n-N, where n is the first core, and N the last (e.g., 0-3)
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
return tonumber(str:match("%d+$")) or 1
else
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
return 1
end
end
function Kobo:enableCPUCores(amount)
-- CPU0 is *always* online ;).
for n = 1, self.cpu_count do
local path = "/sys/devices/system/cpu/cpu" .. n .. "/online"
local up
if n >= amount then
up = "0"
else
up = "1"
end
util.writeToSysfs(up, path)
end
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
function Kobo:performanceCPUGovernor()
util.writeToSysfs("performance", self.cpu_governor_knob)
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
end
function Kobo:defaultCPUGovernor()
util.writeToSysfs(self.default_cpu_governor, self.cpu_governor_knob)
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
end
function Kobo:isStartupScriptUpToDate()
-- Compare the hash of the *active* script (i.e., the one in /tmp) to the *potential* one (i.e., the one in KOREADER_DIR)
local current_script = "/tmp/koreader.sh"
local new_script = os.getenv("KOREADER_DIR") .. "/" .. "koreader.sh"
local md5 = require("ffi/MD5")
return md5.sumFile(current_script) == md5.sumFile(new_script)
end
function Kobo:UIManagerReady(uimgr)
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager = uimgr
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
function Kobo:setEventHandlers(uimgr)
-- We do not want auto suspend procedure to waste battery during
-- suspend. So let's unschedule it when suspending, and restart it after
-- resume. Done via the plugin's onSuspend/onResume handlers.
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.Suspend = function()
self:onPowerEvent("Suspend")
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.Resume = function()
self:onPowerEvent("Resume")
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.PowerPress = function()
-- Always schedule power off.
-- Press the power button for 2+ seconds to shutdown directly from suspend.
UIManager:scheduleIn(2, UIManager.poweroff_action)
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.PowerRelease = function()
if not UIManager._entered_poweroff_stage then
UIManager:unschedule(UIManager.poweroff_action)
-- resume if we were suspended
if self.screen_saver_mode then
if self.screen_saver_lock then
UIManager.event_handlers.Suspend()
else
UIManager.event_handlers.Resume()
end
else
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.Suspend()
end
end
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.Light = function()
self:getPowerDevice():toggleFrontlight()
end
-- USB plug events with a power-only charger
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.Charging = function()
self:_beforeCharging()
-- NOTE: Plug/unplug events will wake the device up, which is why we put it back to sleep.
if self.screen_saver_mode and not self.screen_saver_lock then
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.Suspend()
end
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.NotCharging = function()
-- We need to put the device into suspension, other things need to be done before it.
self:usbPlugOut()
self:_afterNotCharging()
if self.screen_saver_mode and not self.screen_saver_lock then
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.Suspend()
end
end
-- USB plug events with a data-aware host
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.UsbPlugIn = function()
self:_beforeCharging()
-- NOTE: Plug/unplug events will wake the device up, which is why we put it back to sleep.
if self.screen_saver_mode and not self.screen_saver_lock then
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.Suspend()
elseif not self.screen_saver_lock then
-- Potentially start an USBMS session
local MassStorage = require("ui/elements/mass_storage")
MassStorage:start()
end
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.UsbPlugOut = function()
-- We need to put the device into suspension, other things need to be done before it.
self:usbPlugOut()
self:_afterNotCharging()
if self.screen_saver_mode and not self.screen_saver_lock then
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.Suspend()
elseif not self.screen_saver_lock then
-- Potentially dismiss the USBMS ConfirmBox
local MassStorage = require("ui/elements/mass_storage")
MassStorage:dismiss()
end
end
-- Sleep Cover handling
if G_reader_settings:isTrue("ignore_power_sleepcover") then
-- NOTE: The hardware event itself will wake the kernel up if it's in suspend (:/).
-- Let the unexpected wakeup guard handle that.
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.SleepCoverClosed = nil
UIManager.event_handlers.SleepCoverOpened = nil
elseif G_reader_settings:isTrue("ignore_open_sleepcover") then
-- Just ignore wakeup events, and do NOT set is_cover_closed,
-- so device/generic/device will let us use the power button to wake ;).
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.SleepCoverClosed = function()
UIManager.event_handlers.Suspend()
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.SleepCoverOpened = function()
self.is_cover_closed = false
end
else
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.SleepCoverClosed = function()
self.is_cover_closed = true
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.Suspend()
end
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.SleepCoverOpened = function()
self.is_cover_closed = false
Assorted bag'o tweaks & fixes (#9569) * UIManager: Support more specialized update modes for corner-cases: * A2, which we'll use for the VirtualKeyboards keys (they'd... inadvertently switched to UI with the highlight refactor). * NO_MERGE variants of ui & partial (for sunxi). Use `[ui]` in ReaderHighlight's popup, because of a Sage kernel bug that could otherwise make it translucent, sometimes completely so (*sigh*). * UIManager: Assorted code cleanups & simplifications. * Logger & dbg: Unify logging style, and code cleanups. * SDL: Unbreak suspend/resume outside of the emulator (fix #9567). * NetworkMgr: Cache the network status, and allow it to be queried. (Used by AutoSuspend to avoid repeatedly poking the system when computing the standby schedule delay). * OneTimeMigration: Don't forget about `NETWORK_PROXY` & `STARDICT_DATA_DIR` when migrating `defaults.persistent.lua` (fix #9573) * WakeupMgr: Workaround an apparent limitation of the RTC found on i.MX5 Kobo devices, where setting a wakealarm further than UINT16_MAX seconds in the future would apparently overflow and wraparound... (fix #8039, many thanks to @yfede for the extensive deep-dive and for actually accurately pinpointing the issue!). * Kobo: Handle standby transitions at full CPU clock speeds, in order to limit the latency hit. * UIManager: Properly quit on reboot & exit. This ensures our exit code is preserved, as we exit on our own terms (instead of being killed by the init system). This is important on platforms where exit codes are semantically meaningful (e.g., Kobo). * UIManager: Speaking of reboot & exit, make sure the Screensaver shows in all circumstances (e.g., autoshutdown, re: #9542)), and that there aren't any extraneous refreshes triggered. (Additionally, fix a minor regression since #9448 about tracking this very transient state on Kobo & Cervantes). * Kindle: ID the upcoming Scribe. * Bump base (https://github.com/koreader/koreader-base/pull/1524)
2022-10-02 01:01:49 +00:00
UIManager.event_handlers.Resume()
end
end
end
-------------- device probe ------------
A host of low power states related tweaks (#9036) * Disable all non power management related input during suspend. (This prevents wonky touch events from being tripped when closing a sleep cover on an already-in-suspend device, among other things). * Kobo: Use our WakeupMgr instance, not the class. * WakupMgr: split `removeTask` in two: * `removeTask`, which *only* takes a queue index as input, and only removes a single task. Greatly simplifies the function (i.e., it's just a `table.remove`). * `removeTasks`, which takes an epoch or a cb ref, and removes *every* task that matches. * Both of these will also *always* re-schedule the next task (if any) on exit, since we can have multiple WakeupMgr tasks queued, but we can only have a single RTC wake alarm set ;). * `wakeupAction` now takes a `proximity` argument, which it passes on to its `validateWakeupAlarmByProximity` call, allowing call sites to avoir having to duplicate that call themselves when they want to use a custom proximity window. * `wakeupAction` now re-schedules the next task (if any) on exit. * Simplify `Kobo:checkUnexpectedWakeup`, by removing the duplicate `WakerupMgr:validateWakeupAlarmByProximity` call, now that we can pass a proximity window to `WakeuoMgr:wakeupAction`. * The various network activity timeouts are now halved when autostandby is enabled. * Autostandby: get rid of the dummy deadline_guard task, as it's no longer necessary since #9009. * UIManager: The previous change allows us to simplify `getNextTaskTimes` into a simpler `getNextTaskTime` variant, getting rid of a table & a loop. * ReaderFooter & ReaderHeader: Make sure we only perform a single refresh when exiting standby. * Kobo: Rewrite sysfs writes to use ANSI C via FFI instead of stdio via Lua, as it obscured some common error cases (e.g., EBUSY on /sys/power/state). * Kobo: Simplify `suspend`, now that we have sane error handling in sysfs writes. * Kobo.powerd: Change `isCharging` & `isAuxCharging` behavior to match the behavior of the NTX ioctl (i.e., Charging == Plugged-in). This has the added benefit of making the AutoSuspend checks behave sensibly in the "fully-charged but still plugged in" scenario (because being plugged in is enough to break PM on `!canPowerSaveWhileCharging` devices). * AutoSuspend: Disable our `AllowStandby` handler when auto standby is disabled, so as to not interfere with other modules using `UIManager:allowStandby` (fix #9038). * PowerD: Allow platforms to implement `isCharged`, indicating that the battery is full while still plugged in to a power source (battery icon becomes a power plug icon). * Kobo.powerd: Implement `isCharged`, and kill charging LEDs once battery is full. * Kindle.powerd: Implement `isCharged` on post-Wario devices. (`isCharging` is still true in that state, as it ought to).
2022-05-01 21:41:08 +00:00
local codename = getCodeName()
local product_id = getProductId()
if codename == "dahlia" then
return KoboDahlia
elseif codename == "dragon" then
return KoboDragon
elseif codename == "kraken" then
return KoboKraken
elseif codename == "phoenix" then
return KoboPhoenix
elseif codename == "trilogy" and product_id == "310" then
-- This is where things get interesting...
-- The early 'A' variant (the actual model name being N905, without any letter suffix, unlike the two other variants)
-- does *NOT* feature an internal SD card, and is manufactured in China instead of Taiwan... because it is *NOT* an NTX board.
2023-01-18 08:15:33 +00:00
-- cf. https://github.com/koreader/koreader/issues/9742
if os.getenv("PLATFORM") == "freescale" then
return KoboTrilogyA
else
return KoboTrilogyB
end
elseif codename == "trilogy" and product_id == "320" then
return KoboTrilogyC
elseif codename == "pixie" then
return KoboPixie
2015-05-15 21:46:16 +00:00
elseif codename == "alyssum" then
2015-05-16 01:57:00 +00:00
return KoboAlyssum
elseif codename == "pika" then
return KoboPika
elseif codename == "star" and product_id == "379" then
return KoboStarRev2
elseif codename == "star" then
return KoboStar
2016-09-13 05:50:07 +00:00
elseif codename == "daylight" then
return KoboDaylight
elseif codename == "snow" and product_id == "378" then
return KoboSnowRev2
elseif codename == "snow" then
return KoboSnow
elseif codename == "nova" then
return KoboNova
elseif codename == "frost" then
return KoboFrost
2019-09-11 17:40:36 +00:00
elseif codename == "storm" then
return KoboStorm
elseif codename == "luna" then
return KoboLuna
elseif codename == "europa" then
return KoboEuropa
elseif codename == "cadmus" then
return KoboCadmus
elseif codename == "io" then
return KoboIo
elseif codename == "goldfinch" then
return KoboGoldfinch
else
error("unrecognized Kobo model ".. codename .. " with device id " .. product_id)
end