Hi. I cannot compile a 2.6.38 kernel (and 2.6.37 too) with Nouveau support. I always get the error drivers/built-in.o: In function `nouveau_acpi_edid': (.text+0xd6a00): undefined reference to `acpi_video_get_edid' I have googled, found discussions like that - http://us.generation-nt.com/answer/patch-nouveau-build-regression-undefined-reference-acpi-video-get-edid-help-200299821.html - but it didn't help. Chí-Thanh Christopher Nguyễn said ( http://bugs.gentoo.org/show_bug.cgi?id=357763 ) that this error shouldn't happen, "unless I messed around manually in .config or neglected to run make oldconfig". But I didn't edit .config manually. I created my .config by `zcat /proc/config.gz` with the current Minimal CD running, then `make menuconfig`. Please help. Reproducible: Always Steps to Reproduce: 1. Take my .config (or `zcat /proc/config.gz > /usr/src/linux/.config` with the current Minimal CD running) 2. make menuconfig 3. make && make modules_install Actual Results: Kernel is not compiled because of error drivers/built-in.o: In function `nouveau_acpi_edid': (.text+0xd6a00): undefined reference to `acpi_video_get_edid' Expected Results: Kernel is not compiled
Created attachment 277567 [details] My kernel 2.6.38 .config
To be exact, 1. Take my .config (or `zcat /proc/config.gz > /usr/src/linux/.config` with the current Minimal CD running) 2. make menuconfig __(and turn on Nouveau support, if you ran zcat)__ 3. make && make modules_install
It is possible that nouveau has not specified all dependencies properly in the Kconfig. Find out which kernel option is behind acpi_video_get_edid (presumably ACPI_VIDEO) and make sure that it is enabled built-in. Your problem may be related to http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=c42988012ad9c1807b7c7a5ff855cd630094989b and the fact that you have CONFIG_VIDEO_OUTPUT_CONTROL=m
(In reply to comment #3) > It is possible that nouveau has not specified all dependencies properly in the > Kconfig. Find out which kernel option is behind acpi_video_get_edid (presumably > ACPI_VIDEO) and make sure that it is enabled built-in. > > Your problem may be related to > http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=c42988012ad9c1807b7c7a5ff855cd630094989b > and the fact that you have CONFIG_VIDEO_OUTPUT_CONTROL=m I think that this is indeed what causes the build failure. I'd suggest reporting the issue upstream (either at bugzilla.kernel.org or directly at the nouveau mailing list [1]). In the meantime, setting CONFIG_VIDEO_OUTPUT_CONTROL to y will most likely resolve the build failure. If you report this upstream, please report back at this bug with the URL of the upstream bug report.
Thank you for your replies. I'll be able to check if setting CONFIG_VIDEO_OUTPUT_CONTROL to y will resolve my problem, tonight. I have reported this issue upstream. The link is https://bugzilla.kernel.org/show_bug.cgi?id=38052
Created attachment 277761 [details, diff] CONFIG_ACPI_VIDEO build failure patch I think this could resolve the build failure. I'll post it to the upstream bug too, and at the nouveau mailing list. Btw, could you add kernel@gentoo.org at the CC address at bugzilla.kernel.org? Thanks.
>Btw, could you add kernel@gentoo.org at the CC address at bugzilla.kernel.org? Done.
It works! Thanks a lot! *Best Regards*
Should we repoke on the mailing list? http://lists.freedesktop.org/archives/nouveau/2011-June/008585.html
Closing, not sure if this issue still exists or this patch or something similar was ever committed.
can we get this patch into the tree and re-open this bug as the build failure is happening in the 3.2.12 gentoo sources? Ran genkernel twice and happened both times. I could get around it if I were to disable ACPI but then the monitor would never sleep. I'll stick to the nvidia-drivers until this is fixed.
We only commit patches from upstream. This would need to be accepted into the upstream repositories, and at that point, we can include in gentoo-sources. Q. Will you apply [some kernel patch]? If it's a serious and non-intrusive bugfix that has been accepted into the upstream development kernel, then usually yes (please open a bug). Otherwise, generally the answer is no - instead of getting new features added to our kernel, we'd prefer you to contact the authors of that code and get them to finish off their work and submit it to the mainline kernel so that everyone can benefit.
(In reply to comment #12) > > Q. Will you apply [some kernel patch]? > If it's a serious and non-intrusive bugfix that has been accepted into the > upstream development kernel, then usually yes (please open a bug). > Otherwise, generally the answer is no - instead of getting new features > added to our kernel, we'd prefer you to contact the authors of that code and > get them to finish off their work and submit it to the mainline kernel so > that everyone can benefit. That is fine if that is policy, but would you please reconsider opening the bug so that others who are smart enough to google find an open bug. I can confirm that disabling the nouveau driver in the staging area solves the issue
that would be fine.
As a workaround, can you try to build with CONFIG_VIDEO_OUTPUT_CONTROL=y, CONFIG_ACPI_VIDEO=y
(In reply to comment #15) > As a workaround, can you try to build with > > CONFIG_VIDEO_OUTPUT_CONTROL=y, > CONFIG_ACPI_VIDEO=y Will Test later on this weekend Updating System after 6 months of no updates takes awhile, please leave in test request atatus
(In reply to comment #15) > As a workaround, can you try to build with > > CONFIG_VIDEO_OUTPUT_CONTROL=y, > CONFIG_ACPI_VIDEO=y The above workaround does fix things here, but I believe the bug should stay open until there is a proper fix. All that's needed to reproduce this bug is: 1. genkernel --menuconfig all 2. set CONFIG_DRM=y 3. set CONFIG_DRM_NOUVEAU=y 4. save/exit/build Lots of users are still hitting this bug.
(In reply to comment #13) > That is fine if that is policy, but would you please reconsider opening the > bug so that others who are smart enough to google find an open bug. They won't find it, regardless of its state. https://bugs.gentoo.org/robots.txt (In reply to comment #17) > Lots of users are still hitting this bug. Not much people are on CC. Is this bug still present in the latest kernels?