On Thu, 10 May 2018 15:37:54 +1000 Daniel Kasak <d.j.kasak...@gmail.com> said:
> It's more likely that you've received a kernel upgrade that's broken > resuming from suspension. I've have the bitter experience of a long line of > integrated Intel GPUs, and they have NEVER been stable resuming from > suspension for me. Since the meltdown / spectre patches landed, that seems > to have gotten worse, but it might just be that Intel broke things for > older GPUs when adding support for newer ones in 4.15 / 4.16. It's a bit of > a stretch to say that this instability resuming from suspension makes your > laptop "not usable for work any more". Anyway, the moral of the story is to > buy quality parts with stable drivers. I guarantee you that I'll never buy > a laptop with an integrated intel GPU ( unfortunately my workplaces keep > giving them to me ). If you want to rage at someone for your instability, > go to intel-...@lists.freedesktop.org and tell them about it ... just as information, suspend and resume are "invisible" to userspace. i need to explain this. there is nothing special a wm or compositor or nay x app has to or should do during suspend or resume at all. it's "handled for you by the kernel" basically. yes - e is involved VISUALLY by offering keybindings or acpi bindings (listening to key and acpi events) and presenting a gui or initiate a suspend, where e would mostly ask systemd "please suspend now". it happens to first fade the screen (and backlight) to black.off before asking this of systemd. and in reverse it detects it has resumed from suspend (totally in userspace with no special kernel access) and does the reverse (fades back in). if rendering HANGS and e is stuck e.g. in a gl call this absolutely is not something e can do much or anything about. e itself doesnt even do gl calls - it's abstracted through evas so all e does it move, show, hide and otherwise change the properties of objects. e has done this fading in and out now for like over 5+ years. this is nothing new. it's actually done by the theme when it's signalled. more recently we also fade/modify the x cursor too, but in principal it's all valid userspace activity just before and after a suspend/resume cycle. now as i said. i have sometimes seen e stop rendering. my checks seem to show it just stops calling evas_render () well e doesnt call this. ecore_evas does). i suspect something pauses rendering voluntarily then doesn't unpause (this can be done) but i have yet to catch this often enough to put in debugging to find out why. i wouldn't cal this hung because e's loop is still working and processing events etc. so ctl+alt+end gets it back to running. i call it "stop rendering" as opposed to hung because e can still process events etc. the reason i detail this is, this is where problems get identified. either they are issues to farm off to mesa or kernel or they are issues in efl or e, and then figuring out where the issue happens and how etc. is important, but that requires people to help. the more people provide data, the more likely it is to be finally found. things like "it only happens after long suspends". figuring out how long is "long" is important. which driver, mesa. kernel versions are effected. which laptops? what usage patterns? when my lpatop suspends it's normally for short periods (maybe an hour or 2 or 3 at most, often a few seconds or minutes). if i am not going to use it for a while i shut it down (for me i press power button twice and presto). it boots in only about 8 seconds so i'm happy with that. but i can try and reproduce the behaviour that people say produces the issue. > Dan > > On Thu, May 10, 2018 at 3:11 PM, Marc MERLIN <marc_...@merlins.org> wrote: > > > On Thu, May 10, 2018 at 03:04:03PM +1000, Daniel Kasak wrote: > > > This is probably a driver issue, and not E. > > > > That what Raster said when I first reported this, I believe him, but at the > > end of the day, it doesn't matter. > > 1) E changed to do new more fancy stuff I didn't need. > > 2) It triggers bugs in Xorg or the intel driver, or god knows what. > > 3) My laptop becomes not usable for work anymore > > ... > > 5) !profit > > > > I realize that Raster wants to keep it fun for himself and add more > > features > > and eye candy, I don't blame him. > > But as an end user, I really don't need this, I just need an E that will > > work without any extra features compared to the many many so many features > > it already has, many I probably don't even know. > > Stablizing a branch is not fun work, but it's essential to software, > > especially something as "mustn't fail" and important as a WM. > > > > Marc > > -- > > "A mouse is a device used to point at the xterm you want to type in" - > > A.S.R. > > Microsoft is to operating systems .... > > .... what McDonalds is to gourmet > > cooking > > Home page: http://marc.merlins.org/ > > > > ------------------------------------------------------------ > > ------------------ > > Check out the vibrant tech community on one of the world's most > > engaging tech sites, Slashdot.org! http://sdm.link/slashdot > > _______________________________________________ > > enlightenment-users mailing list > > enlightenment-users@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/enlightenment-users > > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, Slashdot.org! http://sdm.link/slashdot > _______________________________________________ > enlightenment-users mailing list > enlightenment-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/enlightenment-users > -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- Carsten Haitzler - ras...@rasterman.com ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
