Kobo: Yet another attempt at working around the hangs on the latest NXP boards #10771
Conversation
This effectively reverts the relevant part of koreader#10083 It never actually helped with the issue at hand, and had subtly detrimental UX effects in some cases.
On some boards, fast refresh artifacts could get "baked in" the working buffer, and left unchanged by REAGL or partial updates. A flashing update takes care of that, so, just flash the keyboard on close.
It trips a useless refresh otherwise.
timeouts, so, back to stupid hacks it is.
|
Oh, and this also re-enables the jump marker on affected devices ;). Sadly, as described in the comments and the base PR, this didn't help with the wait_for ioctl timeouts, so we've kept our crappy workaround. If you do manually disable it, you will get a slightly different behavior than before, in that it'll now (again, on affected devices) do those ioctls in the same weird pairs as Nickel. FWIW, it took me over 2k page turns to get a couple of timeouts (back to back, things start to go wonky quick after the first one).... And then, like, 15 page turns in Nickel after a reboot, lol >_<". |
There was a problem hiding this comment.
Reviewed 8 of 8 files at r1, all commit messages.
Reviewable status:complete! all files reviewed, all discussions resolved (waiting on @Frenzie, @pazos, and @poire-z)
Thank you so much, I just got a kobo a few weeks ago and put koreader on immediately, never touched the stock FW. But after the freezing I attempted stock (patched) FW, but I still missed koreader. How do I get this update on my device like yesterday? |
|
If you have an existing installation, just switch to the Dev channel and do an OTA update ;). Otherwise, you can grab a current nighly build: https://build.koreader.rocks/download/nightly/ |
I wanted to come back and say that I did this about 3 weeks ago, and in that time I only had 2 freezes, which only required me to tap the power button once to put it to sleep, and then I was able to wake it, and start reading. contrast this with the previous behavoir which had me doing a ahrd reset almost daily. So thank you! I am currently still on dev branch, has this fix been merged to the normal release? Basically, can I switch my branch back to stable now, was this fix implemented into the stable release? nvm, read the release notes haha |
On Kobo GloHD, this is supposed to be an issue with autostandby, cf #10133 - no really a freeze, but a touch panel issue (as the power button does work). |
|
On which device? IIRC, none of the affected devices have an IR grid, which I think was the culprit on the Glo HD, so that's... mildly weird. (I mean, autostandby can appear to send a few input events to the void on some devices, so it's not entirely out of the question; I definitely didn't test autostandby thoroughly because it doesn't make much sense on most of these devices). |
|
Re: #10558 The Nia does NOT actually provide the necessary sysfs knob (at least according to the published kernel sources), so, that's yet another strike against it making any kind of sense as a viable device ;o). |
writeToSysfsto base (i.e.,ffi/util), as we need it there (and it actually makes more sense there anyway ;p).Re #8414, #9806, #10558
This change is