recurring problem, booting git-built to black screen with intel GPU
this is a copy of something i forwarded to the linux kernel mailing list, and i'd be interested if anyone else would test this and reproduce the problem: ===== start ===== on my ubuntu laptop (gateway NV79), a 2.6.37-rc7+ kernel i just built now boots to a black screen. no text mode, no virtual consoles, nothing. this is probably the best explanation: http://www.oblurb.com/ubuntu-dell-latitude-e6510-intel-gpu-blank-screen.html annoyingly, i'm currently running a 2.6.37-rc5+ hand-rolled kernel that works fine, but this more recent kernel doesn't, and a kernel *earlier* also doesn't work, so this issue seems to come and go. here's another extended conversation on the same topic: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/600453 here's the relevant output from lspci: 00:02.0 VGA compatible controller: Intel Corporation Core Processor Integrated Graphics Controller (rev 12) (prog-if 00 [VGA controller]) Subsystem: Acer Incorporated [ALI] Device 031c Flags: bus master, fast devsel, latency 0, IRQ 41 Memory at d0000000 (64-bit, non-prefetchable) [size=4M] Memory at c0000000 (64-bit, prefetchable) [size=256M] I/O ports at 3050 [size=8] Expansion ROM at <unassigned> [disabled] Capabilities: <access denied> Kernel driver in use: i915 Kernel modules: i915 if i use "nomodeset", i can boot to the command line and no further. i'm open to suggestions. ===== end ===== in short, a while back, building a kernel from the git repo failed to give me a working kernel. suddenly, as of 2.6.37-rc5, it worked fine and it's the kernel i'm running under right now: $ uname -r 2.6.37-rc5+ $ and now, as of 2.6.37-rc7, it's back to being borked. i'd love to have piles of free time to track this down but, before i do that, i'd be interested in having someone else try this and let me know if exactly the same thing happens for them. this is a a gateway NV79 laptop, intel core i3-330m, with intel integrated graphics controller. thoughts? rday p.s. it seems clear that the system doesn't simply hang. even after the screen goes black very early in the boot process, there is ample disk activity showing that the boot is continuing, and i eventually hear the little ubuntu "drum roll" that indicates i'm at the login screen. so it's just the video that's died. -- ======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA http://crashcourse.ca Twitter: http://twitter.com/rpjday LinkedIn: http://ca.linkedin.com/in/rpjday ========================================================================
Hi Robert, I have a tiny comment - inline - On Mon, Dec 27, 2010 at 4:18 PM, Robert P. J. Day <rpjday@crashcourse.ca> wrote:
this is a copy of something i forwarded to the linux kernel mailing list, and i'd be interested if anyone else would test this and reproduce the problem:
===== start =====
on my ubuntu laptop (gateway NV79), a 2.6.37-rc7+ kernel i just built now boots to a black screen. no text mode, no virtual consoles, nothing.
this is probably the best explanation:
http://www.oblurb.com/ubuntu-dell-latitude-e6510-intel-gpu-blank-screen.html
annoyingly, i'm currently running a 2.6.37-rc5+ hand-rolled kernel that works fine, but this more recent kernel doesn't, and a kernel *earlier* also doesn't work, so this issue seems to come and go.
here's another extended conversation on the same topic:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/600453
here's the relevant output from lspci:
00:02.0 VGA compatible controller: Intel Corporation Core Processor Integrated Graphics Controller (rev 12) (prog-if 00 [VGA controller]) Subsystem: Acer Incorporated [ALI] Device 031c Flags: bus master, fast devsel, latency 0, IRQ 41 Memory at d0000000 (64-bit, non-prefetchable) [size=4M] Memory at c0000000 (64-bit, prefetchable) [size=256M] I/O ports at 3050 [size=8] Expansion ROM at <unassigned> [disabled] Capabilities: <access denied> Kernel driver in use: i915 Kernel modules: i915
if i use "nomodeset", i can boot to the command line and no further. i'm open to suggestions.
===== end =====
in short, a while back, building a kernel from the git repo failed to give me a working kernel. suddenly, as of 2.6.37-rc5, it worked fine and it's the kernel i'm running under right now:
$ uname -r 2.6.37-rc5+ $
and now, as of 2.6.37-rc7, it's back to being borked. i'd love to have piles of free time to track this down but, before i do that, i'd be interested in having someone else try this and let me know if exactly the same thing happens for them.
this is a a gateway NV79 laptop, intel core i3-330m, with intel integrated graphics controller. thoughts?
rday
p.s. it seems clear that the system doesn't simply hang. even after the screen goes black very early in the boot process, there is ample disk activity showing that the boot is continuing, and i eventually hear the little ubuntu "drum roll" that indicates i'm at the login screen. so it's just the video that's died.
What parameters are being passed to the new kernel via Grub? I had the same issue (don't remember, but I think it was with one of the rc build of 2.6.34) and it could drill it down to a 'vga' parameter on the Grub command line. I removed everything (including any standard 'quiet' parameter which Ubuntu distro kernel has) and managed to boot with some vga value I can't recall (will search and post in case I can find it). I understand that you priority could be to find the solution to this bug - what I state here is just a stop gap solution in case you have not already figured that out. Apologies in case this is obvious to you.
--
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA http://crashcourse.ca
Twitter: http://twitter.com/rpjday LinkedIn: http://ca.linkedin.com/in/rpjday ========================================================================
_______________________________________________ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
- Shreyansh
On Mon, 27 Dec 2010, Shreyansh Jain wrote:
What parameters are being passed to the new kernel via Grub? I had the same issue (don't remember, but I think it was with one of the rc build of 2.6.34) and it could drill it down to a 'vga' parameter on the Grub command line. I removed everything (including any standard 'quiet' parameter which Ubuntu distro kernel has) and managed to boot with some vga value I can't recall (will search and post in case I can find it).
menuentry 'Ubuntu, with Linux 2.6.37-rc7+' --class ubuntu --class gnu-linux --class gnu --class os { recordfail insmod part_msdos insmod ext2 set root='(hd0,msdos1)' search --no-floppy --fs-uuid --set 9f59ebbf-6aab-4581-8234-81a092728571 linux /vmlinuz-2.6.37-rc7+ root=/dev/mapper/lynx-root ro quiet splash initrd /initrd.img-2.6.37-rc7+ } the grub entry for the working 2.6.37-rc5+ kernel is identical other than the kernel version references, of course. rday -- ======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA http://crashcourse.ca Twitter: http://twitter.com/rpjday LinkedIn: http://ca.linkedin.com/in/rpjday ========================================================================
Hi Robert, On Mon, Dec 27, 2010 at 6:47 PM, Robert P. J. Day <rpjday@crashcourse.ca> wrote:
On Mon, 27 Dec 2010, Shreyansh Jain wrote:
What parameters are being passed to the new kernel via Grub? I had the same issue (don't remember, but I think it was with one of the rc build of 2.6.34) and it could drill it down to a 'vga' parameter on the Grub command line. I removed everything (including any standard 'quiet' parameter which Ubuntu distro kernel has) and managed to boot with some vga value I can't recall (will search and post in case I can find it).
menuentry 'Ubuntu, with Linux 2.6.37-rc7+' --class ubuntu --class gnu-linux --class gnu --class os { recordfail insmod part_msdos insmod ext2 set root='(hd0,msdos1)' search --no-floppy --fs-uuid --set 9f59ebbf-6aab-4581-8234-81a092728571 linux /vmlinuz-2.6.37-rc7+ root=/dev/mapper/lynx-root ro quiet splash initrd /initrd.img-2.6.37-rc7+ }
the grub entry for the working 2.6.37-rc5+ kernel is identical other than the kernel version references, of course.
Can you please boot this kernel once again by modifying the command line at GRUB prompt and removing 'quiet', 'splash' keywords? I don't have a reason as to why this may work, just that in my case I removed them and I could see the bootup including the error dump (oops), if there was any.
rday
--
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA http://crashcourse.ca
Twitter: http://twitter.com/rpjday LinkedIn: http://ca.linkedin.com/in/rpjday ========================================================================
- Shreyansh
Hi :) Just to add what you have said below....based on my own experience (which is the issue is somewhat similar...but not strongly similar) On Mon, Dec 27, 2010 at 20:10, Shreyansh Jain <shrey.linux@gmail.com> wrote:
What parameters are being passed to the new kernel via Grub? I had the same issue (don't remember, but I think it was with one of the rc build of 2.6.34) and it could drill it down to a 'vga' parameter on the Grub command line. I removed everything (including any standard 'quiet' parameter which Ubuntu distro kernel has) and managed to boot with some vga value I can't recall (will search and post in case I can find it).
In my case, this is happen when I used Fedora 7 (or 9, couldn't recall perfectly) plus manually compiled 2.6.3{1,2}. Sometimes, X could start....in other occasions, it couldn't. I do this by simply entering run level 3 (console text mode, full networking) then executed "startx". But using stock kernel, it never happened. X log didn't reveal anything suspicious. But eventually I came to raw guess that it was due to DRI incompatibility between the X user space part versus kernel space. -- regards, Mulyadi Santosa Freelance Linux trainer and consultant blog: the-hydra.blogspot.com training: mulyaditraining.blogspot.com
participants (3)
-
Mulyadi Santosa -
Robert P. J. Day -
Shreyansh Jain