Summary: | xorg-x11-6.8.0 randomly blanks the screen | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Dale K Dicks <dale_d> |
Component: | Current packages | Assignee: | Gentoo X packagers <x11> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | bugreports, carlo, cryos, david.w.noon, flash3001, greg_g, jaime.f, m.debruijne, pasrospa, polynomial-c, rutger, tavin.cole, teidakankan, thomas.bettler, xero |
Priority: | High | Keywords: | Inclusion |
Version: | unspecified | ||
Hardware: | All | ||
OS: | All | ||
URL: | http://freedesktop.org/bugzilla/show_bug.cgi?id=792 | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
The /etc/X11/xorg.conf the bug happens with.
my current config patch from xfree patch to fix random screen blanking |
Description
Dale K Dicks
2004-04-15 06:50:12 UTC
I can confirm this bug on amd64, gentoo-dev-sources-2.6.5-r1 using nvidia-glx/nvidia-kernel. I attached my /etc/X11/xorg.conf. Created attachment 29661 [details]
The /etc/X11/xorg.conf the bug happens with.
I just remembered I actually had this happen on 4.3.99.901 (hasnt happened since). Im using nvidia-kernel-5336 what are you all using? nvidia-kernel-1.0.5332-r1 on amd64 in TwinView mode. two desktops over here with the same problem; media-video/nvidia-kernel-1.0.5336-r2 x11-base/xorg-x11-6.7.0 Option DPMS enabled x86 This only happens (for me) when xscreensaver is running and has the power management options turned on. I got the same troubles... xorg-x11, kde3.2.2, gentoo-dev-sources-2.6.5-r1, radeon (mobility R100 m6 ly), gentoo-dev-sources-2.6.5-r1+win4lin Created attachment 32339 [details]
my current config
Same problem here, noticed immediately after switching from xfree to xorg. I'm currently using nvidia-kernel-1.0.5336-r4 on my geforce 4 ti4200, mm-sources 2.6.7-rc3-mm2. Pretty annoying, especially when it happens while I'm playing a full screen game. Same problem with my setup ... Kernel 2.6.7 NTPL x11-base/xorg-x11-6.7.0-r1 media-video/nvidia-kernel-1.0.6106 DPMS enabled This is still happening to me and it's just as annoying as ever. It seems absolutely no progress has been made on this bug. See: http://freedesktop.org/bugzilla/show_bug.cgi?id=792 https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=120817 for more reports on the same bug Created attachment 37387 [details, diff] patch from xfree this bug also effected xfree 4.4.0 and was fixed. see: http://bugs.xfree86.org/show_bug.cgi?id=1252 However the patch to fix it is under the new xfree license..... attached is the relevant section of the patch... Here's the commit: http://cvsweb.xfree86.org/cvsweb/xc/programs/Xserver/os/WaitFor.c?r1=3.46 I haven't actually applied it to try it yet but I'd assume it works....I have no idea what kind of licensing mess would be involved in using this... I have this problem, too. Exactly _once_ 20 minutes after xorg-x11 started. I've been using that patch for probably almost a week now and haven't had this happen once since then...so it does work. We won't add anything under the new license, sorry. You'll have to find a fix in freedesktop.org bugzilla or CVS to get it into Gentoo. Any people who code and are interested in fixing this, don't look at that patch, because then you can't submit a fix. I figured as much, hopefully some progress will be made in finding an alternative. I have used the patch too, works a treat (once I sused out how to patch an ebuild!) I now have DPMS enabled with no annoying blank outs. Dave The same problem here. x11-xorg 6.7.99.904 Using ATI Radeon 9600 Icewm Pentium4 LCD monitor connected to DVI output. If any other info is required, just ask me. Here are my notes on how to patch the e-build, (probarbly teaching people to suck eggs !) but hope it helps ... ebuild /usr/portage/x11-base/xorg-x11/xorg-x11-6.7.0-r1.ebuild unpack cd /var/tmp/portage/xorg-x11-6.7.0-r1/work patch -p0 < /...path to patch.../dpms.patch ebuild /usr/portage/x11-base/xorg-x11/xorg-x11-6.7.0-r1.ebuild merge X.org 6.8.0 is now released and the WaitFor.c file hasn't been modified since April 23 so this problem has not been fixed as far as I can tell...plus the bug is still open over at X.org....I can't believe they'd release 6.8 without fixing a bug this annoying... Its been a majour pain in the neck for me. Until you pointed out the patch (Thanks for that :)), I was considering reverting to Xfree86, only package dependencies stopped me. What bug is open at X.org If there is a non-xfree 4.4 licensed patch we will put it into the patchset. the same bug as this one at X.org is still open. http://freedesktop.org/bugzilla/show_bug.cgi?id=792 Blerg, teach me for not reading the whole bug report, just me recent mail from it ... anyway from what I can see this is a result of enabling DPMS on your monitor but not having server side configuration specifying the actual times for the various DPMS stages. Basically we want a fix that disables the x11 server dpms timers if they haven't been specified. I must re-iterate what spyderous said though -- Do not read the patch dawes provided. well xscreensaver is what sets the times for my DPMS settings. I was under the impression that the problem is something to do with the timer not being reset properly or something like that.... Befor I patched xorg I tried enableing DPMS on my monitor & setting the server side configuration for the DPMS stages, unfortunatley it made no difference. I cannot re-verify this since I do not want to revert back to a non patched xorg :) Dave S This bug is also somewhat hard to test for due to the fact you have to wait around for it to occur...so even if you were able to revert back it's not even guaranteed you'd notice it right away... the patch earlier posted works. someone needs to rewrite it and submit it ot x-org, or even just rewirte it and submit it to Gentoo. who cares - as long as it gets fixed - its an annoying bug and should not have been there in the first place if they had tested properly. Since I added the following lines to my xorg.conf (Section "ServerFlags") the problem didn't occur any longer: Option "BlankTime" "10" Option "StandbyTime" "20" Option "SuspendTime" "0" Option "OffTime" "0" Before wildly applying patches you should maybe give it a try. :) Even if that does work it's not really solving the problem, it's avoiding it. In my case I want 10 blank time, 0 standby time, 0 suspend time, and 16 minutes off time. I set it with xscreensaver. It doesn't work properly, but it should, and used to. The patch fixes it. *** Bug 56443 has been marked as a duplicate of this bug. *** Jordan : The patch will NOT be added to portage or xorg-x11 until something else that does the job is re-written... Unless someone tells me otherwise im going assume specifying the actual DPMS server options (when not using the patch) resolves this... I understand the patch cannot be added. However, simply adding times to your file doesn't fix the real issue, you have to disable offtime and use standby time (or suspend?) instead. Essentially Offtime (possibly suspend?) is broken badly from this. From what it says in the xorg bug adding the times in xorg.conf and then changing them back with xset works fine. However that's a work around at best, not a fix, as I was trying to say above. It should just work. This bug seems to be a bit forgotten, and the one reported on freedesktop.org it's even older. Did you all stop using DPMS off? Or did you applied the Xfree patch? It would be nice to have it fixed in next release. Perhaps it hasn't been made sufficiently clear yet. so I'll say it in all caps. THE XFREE CODE IS LICENSE-CONTAMINATED AND WILL NOT BE USED. WE CAN'T ADD THIS UNTIL WE GET A VERSION UNDER A FRIENDLIER LICENSE. Thanks for your time. =) Created attachment 43761 [details, diff] patch to fix random screen blanking i wrote this patch to fix the problem. i did not look at the patch from xfree. i hereby release these 3 lines of code and 8 lines of indentation under a dual license, either BSD (http://opensource.org/licenses/bsd-license.php) or GPL (http://opensource.org/licenses/gpl-license.php) What's the status with this? Does there need to be any type of verification or can this finally be solved? I've notified upstream of Tevin's patch, and now we'll wait to see what happens there. Feel free to reopen this once something is committed upstream, if you'd like me to add a patch to ours. Sorry, Tavin. Not Tevin. can't the gentoo ebuild at least be fixed now? Could I get a few testers to confirm that this patch fixes the issue and doesn't introduce new ones? ebuild /usr/portage/x11-base/xorg-x11/xorg-x11-6.8.0-r1 unpack patch -d /var/tmp/portage/xorg-x11-6.8.0-r1/work -p1 < /path/to/patch ebuild /usr/portage/x11-base/xorg-x11/xorg-x11-6.8.0-r1 qmerge Hi, I applied the patch against xorg-x11-6.8.0-r3 on two ~x86 machines. I tested one machine for about 7 hours without catching the error again. Seems to work. Poly Hi everybody, I spoke with Donnie on the mail. He will add this patch to the portage tree if 5 people test this and report back that this patch works (or not). Right now 2 people have tested and reported back (Tavin and Lars). We can do better then that. I'm building x11-xorg-6.8.0-r3 with the supplied patch right now and will report back later. There are a lot of people on the cc-list, so if we have two more testers and the patch works every Gentoo user can benefit from this. Donnie forgot one step in the instruction, hereby a new instruction (I also replaced -r1 with -r3 which is marked stable now); ebuild /usr/portage/x11-base/xorg-x11/xorg-x11-6.8.0-r3.ebuild unpack patch -d /var/tmp/portage/xorg-x11-6.8.0-r1/work -p1 < /path/to/patch ebuild /usr/portage/x11-base/xorg-x11/xorg-x11-6.8.0-r3.ebuild install ebuild /usr/portage/x11-base/xorg-x11/xorg-x11-6.8.0-r3.ebuild qmerge oops, wrong copy/paste, patch also needs to point to -r3; instead of patch -d /var/tmp/portage/xorg-x11-6.8.0-r1/work -p1 < /path/to/patch type patch -d /var/tmp/portage/xorg-x11-6.8.0-r3/work -p1 < /path/to/patch install shouldn't be required before qmerge. If it is, that's a bug IMO. Install step was required when I applied the patch and built xorg-x11, I am using portage-2.0.51-r3. Seems to have built and installed fine - I will now test it for a few hours anyway to see if it has in fact fixed it. Hi, for all of you who don't like patching the stuff themselves I created an ebuild with the patch included. You can find it here: http://polynomial-c.homelinux.net/pub/gentoo/portage/x11-base/xorg-x11/ Please download all files in this dir and its subdir not only the ebuild and place all files in the same manner into your PORTDIR_OVERLAY. Poly seeing as the content of the patch is basically the same as the patch from xfree which I know works I'd assume the same is with this one. However just to be sure I will apply the patch and report on it in a few hours...hopefully we can get this closed for good. Hi, by having only a look at Tavin's patch I feared that the patch would be similar to the xfree one because the fix looks rather trivial to me though I'm not a coder. I hope the fix (in case it turns out to be working) won't be rejected beause of this. Poly Installed xorg-x11-6.8.0-r3 with the patch on two systems, works perfectly, I can control power settings from kde again and the screen doesn't blank randomly. Seeing as it is so trivial the fact that it's basically identicle I don't think is a problem as long as it was written from scratch. If it were something more complex, there may be multiple ways to fix it, but in this situation it doesn't seem to be the case. BTW for those who are lazy like me you can simple emerge xorg-x11 and patch the build while it's going, preferrably right after it finishes applying the other patches. I'm pretty sure nothing else touches WaitFor.c so you should be safe...Seeing as xorg-x11 is friggen huge there's no way it'll be compiling it until later so I've successfully done this a few times...plus it doesn't modify any makefiles or such, just a few lines of code. I have it compiling now... The patch is fine for me, too. I applied the patch to my dual head amd64 nvidia system, and have been testing for the last four hours or so. The screen would have usually blanked at least a couple of times, and it has not. The patch certainly seems to work for me. I don't know how long you would like to have this tested for though - it did appear quite random. I am using xorg-x11-6.8.0-r3 patched with patch 43761, compiled using gcc-3.4.2-r2 and using nvidia-kernel-1.0.6111-r3. all seems well here, been about 3 days using the new patch. Same here patch works just fine Going to add what actually got committed upstream: https://bugs.freedesktop.org/show_bug.cgi?id=1934 Fixed in 6.8.0-r4, patchset 0.2.11. |