Re: [Intel-gfx] black screen if sna & TearFree
Top-posting seems appropriate to say that I think this has been fixed in latest git and possibly 2.20.6. I will report back if not. Thank you, Grant >>> 'xset dpms force off; sleep 1; xset dpms force on' does indeed trigger >>> the error. I've attached my Xorg.0.log.old which includes the >>> segfault backtrace. I'm like a newborn baby when it comes to >>> compiling and Gentoo doesn't have a debug USE flag for >>> xf86-video-intel or xorg-server so I can't help there. Is there a >>> different package I should check for a debug USE flag? >> >> In the Xorg.log, the first error appears to be a failed modeset after >> resume: >> >> [ 5131.319] (II) AIGLX: Resuming AIGLX clients after VT switch >> [ 5131.319] (II) intel(0): switch to mode 1366x768 on crtc 3 (pipe 0) >> [ 5131.319] (EE) intel(0): failed to set mode: Invalid argument >> [ 5131.319] (WW) intel(0): failed to restore desired modes on VT switch >> >> That is going to leave the screens blank. And the subsequent segfault >> suggests that I don't have sufficient safeguards to handle that failure. >> Can you please run 'addr2line -e >> /usr/lib64/xorg/modules/drivers/intel_drv.so 0xb9ed7 0x68517 0x5aa2e' ? > > I reproduced the segfault on my other (identical) laptop and got: > > [ 10865.673] 3: /usr/lib64/xorg/modules/drivers/intel_drv.so > (0x7f5cf32f5000+0x39ba5) [0x7f5cf332eba5] > [ 10865.673] 4: /usr/lib64/xorg/modules/drivers/intel_drv.so > (0x7f5cf32f5000+0x47015) [0x7f5cf333c015] > [ 10865.673] 5: /usr/lib64/xorg/modules/drivers/intel_drv.so > (0x7f5cf32f5000+0x4bc29) [0x7f5cf3340c29] > [ 10865.673] 6: /usr/lib64/xorg/modules/drivers/intel_drv.so > (0x7f5cf32f5000+0x57708) [0x7f5cf334c708] > > so I changed the command: > > # addr2line -e /usr/lib64/xorg/modules/drivers/intel_drv.so 0x39ba5 > 0x47015 0x4bc29 0x57708 > ??:0 > ??:0 > ??:0 > ??:0 > # > >> To solve what happened to prevent the modeswitch we need to have the >> dmesg, preferrably with drm.debug=6 (echo 6 > >> /sys/modules/drm/parameters/debug). > > I've attached kern.log since the segfault. It looks like there's some > good stuff in there. > >> Compiling a debug driver is pretty straightforward. >> $ git clone git://anongit.freedesktop.org/xorg/driver/xf86-video-intel >> $ cd xf86-video-intel >> $ ./autogen.sh --prefix=/usr --enable-debug=full >> $ make && sudo make install >> >> This will overwrite the system driver and so not need any modification >> to config files. (It is possible to install into local and change >> /etc/X11/xorg.conf to load the driver from there.) So just reinstall >> your distro package to revert. >> -Chris > > If it's still necessary for me to do so, please tell me how to install > into a folder (/usr/local ?) that I can delete later to completely > wipe out the install. > > Thanks, > Grant ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] black screen if sna & TearFree
>>> 'xset dpms force off; sleep 1; xset dpms force on' does indeed trigger >>> the error. I've attached my Xorg.0.log.old which includes the >>> segfault backtrace. I'm like a newborn baby when it comes to >>> compiling and Gentoo doesn't have a debug USE flag for >>> xf86-video-intel or xorg-server so I can't help there. Is there a >>> different package I should check for a debug USE flag? >> >> In the Xorg.log, the first error appears to be a failed modeset after >> resume: >> >> [ 5131.319] (II) AIGLX: Resuming AIGLX clients after VT switch >> [ 5131.319] (II) intel(0): switch to mode 1366x768 on crtc 3 (pipe 0) >> [ 5131.319] (EE) intel(0): failed to set mode: Invalid argument >> [ 5131.319] (WW) intel(0): failed to restore desired modes on VT switch >> >> That is going to leave the screens blank. And the subsequent segfault >> suggests that I don't have sufficient safeguards to handle that failure. >> Can you please run 'addr2line -e >> /usr/lib64/xorg/modules/drivers/intel_drv.so 0xb9ed7 0x68517 0x5aa2e' ? > > I reproduced the segfault on my other (identical) laptop and got: > > [ 10865.673] 3: /usr/lib64/xorg/modules/drivers/intel_drv.so > (0x7f5cf32f5000+0x39ba5) [0x7f5cf332eba5] > [ 10865.673] 4: /usr/lib64/xorg/modules/drivers/intel_drv.so > (0x7f5cf32f5000+0x47015) [0x7f5cf333c015] > [ 10865.673] 5: /usr/lib64/xorg/modules/drivers/intel_drv.so > (0x7f5cf32f5000+0x4bc29) [0x7f5cf3340c29] > [ 10865.673] 6: /usr/lib64/xorg/modules/drivers/intel_drv.so > (0x7f5cf32f5000+0x57708) [0x7f5cf334c708] > > so I changed the command: > > # addr2line -e /usr/lib64/xorg/modules/drivers/intel_drv.so 0x39ba5 > 0x47015 0x4bc29 0x57708 > ??:0 > ??:0 > ??:0 > ??:0 > # > >> To solve what happened to prevent the modeswitch we need to have the >> dmesg, preferrably with drm.debug=6 (echo 6 > >> /sys/modules/drm/parameters/debug). > > I've attached kern.log since the segfault. It looks like there's some > good stuff in there. > >> Compiling a debug driver is pretty straightforward. >> $ git clone git://anongit.freedesktop.org/xorg/driver/xf86-video-intel >> $ cd xf86-video-intel >> $ ./autogen.sh --prefix=/usr --enable-debug=full >> $ make && sudo make install >> >> This will overwrite the system driver and so not need any modification >> to config files. (It is possible to install into local and change >> /etc/X11/xorg.conf to load the driver from there.) So just reinstall >> your distro package to revert. >> -Chris > > If it's still necessary for me to do so, please tell me how to install > into a folder (/usr/local ?) that I can delete later to completely > wipe out the install. > > Thanks, > Grant Hi, have you had a chance to look into this? - Grant ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] black screen if sna & TearFree
On Thu, 16 Aug 2012 04:19:26 -0700, Grant wrote: > 'xset dpms force off; sleep 1; xset dpms force on' does indeed trigger > the error. I've attached my Xorg.0.log.old which includes the > segfault backtrace. I'm like a newborn baby when it comes to > compiling and Gentoo doesn't have a debug USE flag for > xf86-video-intel or xorg-server so I can't help there. Is there a > different package I should check for a debug USE flag? In the Xorg.log, the first error appears to be a failed modeset after resume: [ 5131.319] (II) AIGLX: Resuming AIGLX clients after VT switch [ 5131.319] (II) intel(0): switch to mode 1366x768 on crtc 3 (pipe 0) [ 5131.319] (EE) intel(0): failed to set mode: Invalid argument [ 5131.319] (WW) intel(0): failed to restore desired modes on VT switch That is going to leave the screens blank. And the subsequent segfault suggests that I don't have sufficient safeguards to handle that failure. Can you please run 'addr2line -e /usr/lib64/xorg/modules/drivers/intel_drv.so 0xb9ed7 0x68517 0x5aa2e' ? To solve what happened to prevent the modeswitch we need to have the dmesg, preferrably with drm.debug=6 (echo 6 > /sys/modules/drm/parameters/debug). Compiling a debug driver is pretty straightforward. $ git clone git://anongit.freedesktop.org/xorg/driver/xf86-video-intel $ cd xf86-video-intel $ ./autogen.sh --prefix=/usr --enable-debug=full $ make && sudo make install This will overwrite the system driver and so not need any modification to config files. (It is possible to install into local and change /etc/X11/xorg.conf to load the driver from there.) So just reinstall your distro package to revert. -Chris -- Chris Wilson, Intel Open Source Technology Centre ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] black screen if sna & TearFree
>>> > I'm using a Dell XPS 13 laptop on Gentoo and if I use: >>> > >>> > Option "AccelMethod" "sna" >>> > Option "TearFree" "true" >>> > >>> > my video tearing problem disappears, but if I close my laptop lid for >>> > a few minutes, it comes back up with a black screen, although I can >>> > get my session back if I switch to VT1 and then back to VT7. I can't >>> > find anything in the logs. I've tried xf86-video-intel-2.20.3 as well >>> > as the latest from git. The problem doesn't occur without the above >>> > two directives. >>> > >>> > Does anyone know how to diagnose or fix this? >>> > >>> > - Grant >>> >>> I can prevent this if I issue 'xset -dpms' before closing the laptop >>> lid, but xlockmore automatically re-enables dpms unfortunately. BTW, >>> this does not prevent the problem: >> >> The implication would seem to be that an action during dpms is going >> wrong. Does a simple 'xset dpms force off; sleep 1; xset dpms force on' >> trigger the error? If so can you please attach your Xorg.0.log and if >> you could compile with --enable-debug=full and send me the log >> reproducing the error that would be very helpful. >> >> Thanks, >> -Chris > > 'xset dpms force off; sleep 1; xset dpms force on' does indeed trigger > the error. I've attached my Xorg.0.log.old which includes the > segfault backtrace. I'm like a newborn baby when it comes to > compiling and Gentoo doesn't have a debug USE flag for > xf86-video-intel or xorg-server so I can't help there. Is there a > different package I should check for a debug USE flag? > > - Grant > > P.S. Sorry for the multiple sends. It looks like I also need to do 'xset s 0' because X also crashes if the screensaver comes on for more than a few seconds. Screen Saver: prefer blanking: yesallow exposures: yes timeout: 0cycle: 600 Maybe it has something to do with turning off the backlight? - Grant ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] black screen if sna & TearFree
>> > I'm using a Dell XPS 13 laptop on Gentoo and if I use: >> > >> > Option "AccelMethod" "sna" >> > Option "TearFree" "true" >> > >> > my video tearing problem disappears, but if I close my laptop lid for >> > a few minutes, it comes back up with a black screen, although I can >> > get my session back if I switch to VT1 and then back to VT7. I can't >> > find anything in the logs. I've tried xf86-video-intel-2.20.3 as well >> > as the latest from git. The problem doesn't occur without the above >> > two directives. >> > >> > Does anyone know how to diagnose or fix this? >> > >> > - Grant >> >> I can prevent this if I issue 'xset -dpms' before closing the laptop >> lid, but xlockmore automatically re-enables dpms unfortunately. BTW, >> this does not prevent the problem: > > The implication would seem to be that an action during dpms is going > wrong. Does a simple 'xset dpms force off; sleep 1; xset dpms force on' > trigger the error? If so can you please attach your Xorg.0.log and if > you could compile with --enable-debug=full and send me the log > reproducing the error that would be very helpful. > > Thanks, > -Chris 'xset dpms force off; sleep 1; xset dpms force on' does indeed trigger the error. I've attached my Xorg.0.log.old which includes the segfault backtrace. I'm like a newborn baby when it comes to compiling and Gentoo doesn't have a debug USE flag for xf86-video-intel or xorg-server so I can't help there. Is there a different package I should check for a debug USE flag? - Grant P.S. Sorry for the multiple sends. Xorg.0.log.old Description: Binary data ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] black screen if sna & TearFree
>> > I'm using a Dell XPS 13 laptop on Gentoo and if I use: >> > >> > Option "AccelMethod" "sna" >> > Option "TearFree" "true" >> > >> > my video tearing problem disappears, but if I close my laptop lid for >> > a few minutes, it comes back up with a black screen, although I can >> > get my session back if I switch to VT1 and then back to VT7. I can't >> > find anything in the logs. I've tried xf86-video-intel-2.20.3 as well >> > as the latest from git. The problem doesn't occur without the above >> > two directives. >> > >> > Does anyone know how to diagnose or fix this? >> > >> > - Grant >> >> I can prevent this if I issue 'xset -dpms' before closing the laptop >> lid, but xlockmore automatically re-enables dpms unfortunately. BTW, >> this does not prevent the problem: > > The implication would seem to be that an action during dpms is going > wrong. Does a simple 'xset dpms force off; sleep 1; xset dpms force on' > trigger the error? If so can you please attach your Xorg.0.log and if > you could compile with --enable-debug=full and send me the log > reproducing the error that would be very helpful. > > Thanks, > -Chris 'xset dpms force off; sleep 1; xset dpms force on' does indeed trigger the error. I've attached my Xorg.0.log.old which includes the segfault backtrace. I'm like a newborn baby when it comes to compiling and Gentoo doesn't have a debug USE flag for xf86-video-intel or xorg-server so I can't help there. Is there a different package I should check for a debug USE flag? - Grant ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] black screen if sna & TearFree
On Wed, 15 Aug 2012 08:09:47 -0700, Grant wrote: > > I'm using a Dell XPS 13 laptop on Gentoo and if I use: > > > > Option "AccelMethod" "sna" > > Option "TearFree" "true" > > > > my video tearing problem disappears, but if I close my laptop lid for > > a few minutes, it comes back up with a black screen, although I can > > get my session back if I switch to VT1 and then back to VT7. I can't > > find anything in the logs. I've tried xf86-video-intel-2.20.3 as well > > as the latest from git. The problem doesn't occur without the above > > two directives. > > > > Does anyone know how to diagnose or fix this? > > > > - Grant > > I can prevent this if I issue 'xset -dpms' before closing the laptop > lid, but xlockmore automatically re-enables dpms unfortunately. BTW, > this does not prevent the problem: The implication would seem to be that an action during dpms is going wrong. Does a simple 'xset dpms force off; sleep 1; xset dpms force on' trigger the error? If so can you please attach your Xorg.0.log and if you could compile with --enable-debug=full and send me the log reproducing the error that would be very helpful. Thanks, -Chris -- Chris Wilson, Intel Open Source Technology Centre ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] black screen if sna & TearFree
> I'm using a Dell XPS 13 laptop on Gentoo and if I use: > > Option "AccelMethod" "sna" > Option "TearFree" "true" > > my video tearing problem disappears, but if I close my laptop lid for > a few minutes, it comes back up with a black screen, although I can > get my session back if I switch to VT1 and then back to VT7. I can't > find anything in the logs. I've tried xf86-video-intel-2.20.3 as well > as the latest from git. The problem doesn't occur without the above > two directives. > > Does anyone know how to diagnose or fix this? > > - Grant I can prevent this if I issue 'xset -dpms' before closing the laptop lid, but xlockmore automatically re-enables dpms unfortunately. BTW, this does not prevent the problem: DPMS (Energy Star): Standby: 0Suspend: 0Off: 0 DPMS is Enabled Monitor is On It has to be this: DPMS (Energy Star): Standby: 0Suspend: 0Off: 0 DPMS is Disabled I've tried different kernel versions as well as different versions of xf86-video-intel and xorg-server with the same result. - Grant ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
[Intel-gfx] black screen if sna & TearFree
I'm using a Dell XPS 13 laptop on Gentoo and if I use: Option "AccelMethod" "sna" Option "TearFree" "true" my video tearing problem disappears, but if I close my laptop lid for a few minutes, it comes back up with a black screen, although I can get my session back if I switch to VT1 and then back to VT7. I can't find anything in the logs. I've tried xf86-video-intel-2.20.3 as well as the latest from git. The problem doesn't occur without the above two directives. Does anyone know how to diagnose or fix this? - Grant ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx