| Summary: | [2.6.26 regression] sata_nv hang on boot when SMP/HIGHMEM enabled | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | M Baker <zerovian> |
| Component: | [OLD] Core system | Assignee: | Gentoo Kernel Bug Wranglers and Kernel Maintainers <kernel> |
| Status: | RESOLVED WORKSFORME | ||
| Severity: | critical | CC: | andrew.bugs |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | x86 | ||
| OS: | Linux | ||
| Whiteboard: | linux-2.6.26-regression | ||
| Package list: | Runtime testing required: | --- | |
| Attachments: |
Config for 2.6.28-rc9 which hangs on boot
.config for 2.6.25-r6 which does NOT hang on boot lspci -vvvv output for gentoo kernel 2.6.25-r6 git bisect log for 2.5.25 config used for git bisect git bisect showing single commit that hangs system /var/log/messages from booting machine |
||
|
Description
M Baker
2008-12-23 00:45:48 UTC
Correction to steps: Steps to Reproduce: 1. boot machine with kernel 2.6.26 or 2.6.27 2. 3. Is this reproducible with vanilla sources? If so, do you want to try doing git bisect using mainline git repository to figure out which patch triggered this problem? Tried with vanilla-sources-2.6.28-rc9 and that hangs as well. I can do a git bisect if you point me to some instructions. Certainly, here are a couple of useful guides: http://www.kernel.org/doc/local/git-quick.html http://www.reactivated.net/weblog/archives/2006/01/using-git-bisect-to-find-buggy-kernel-patches/ Did you try to build sata_nv in kernel (not as a module)? Can you provide us with the boot logs? Also post your .config please. (In reply to comment #5) > Did you try to build sata_nv in kernel (not as a module)? > Can you provide us with the boot logs? > Also post your .config please. > Which boot log files do you want and against which kernel? (In reply to comment #5) > Did you try to build sata_nv in kernel (not as a module)? > Can you provide us with the boot logs? > Also post your .config please. > Yes, I tried built into kernel...no change. (In reply to comment #4) > Certainly, here are a couple of useful guides: > http://www.kernel.org/doc/local/git-quick.html > http://www.reactivated.net/weblog/archives/2006/01/using-git-bisect-to-find-buggy-kernel-patches/ > This will take some time. Thanks for the isntructions and Merry Christmas to you. Created attachment 176310 [details]
Config for 2.6.28-rc9 which hangs on boot
Can you attach the .config file of the 2.6.25 kernel since we know it works there?Thanks Created attachment 176317 [details]
.config for 2.6.25-r6 which does NOT hang on boot
Could you please boot 2.6.25 kernel and give us an lspci -vvv output? Thanks Created attachment 176927 [details]
lspci -vvvv output for gentoo kernel 2.6.25-r6
Created attachment 176929 [details]
git bisect log for 2.5.25
git bisect reports 50be4917ee70218f59e04dec029121b97fb9cb3d is first bad checkin that hangs
Created attachment 176931 [details]
config used for git bisect
I used alternate .config (attached) to build the git bisect log due to compile errors. Attached is the one I used which turns off all the unused file systems (reiserfs, dlock, ). Used this to reproduce hang/no hang during bisect. It is baseically exactly what genkernel outputs except for file systems turned off.
Note that the git bisect log is only partial since I had to restart it a few times in the middle due to various compile errors and other occasional hang in various bisect builds. It does narrow down to the specific commit though. Not really, it narrowed it down to a merge commit (the parent of many others), and a large merge at that. How strange. I'll try and look at the log in more detail. I went through git-bisect again this time starting with the known bad one and it narrowed it down to a different commit. My mistake or gits...who knows. Anywhere attached is new bisect log Created attachment 177933 [details]
git bisect showing single commit that hangs system
(In reply to comment #19) > Created an attachment (id=177933) [edit] > git bisect showing single commit that hangs system > Great. That helps a lot Could you please try the latest vanilla-sources (2.6.28) and verify that the problem still exists? Thank you Very much still present in vanilla-sources-2.6.28. (In reply to comment #21) > Very much still present in vanilla-sources-2.6.28. > Could you provide us with the last few lines of booting output just before the hang? Also, the dmesg output of a working kernel (2.6.25?) would be really appreciated, to see which message would come normally after the hang. Created attachment 178316 [details]
/var/log/messages from booting machine
Last few messages Scanning for pcmcia_core...loaded Scanning for sata_promise...loaded Scanning for sata_sil...loaded Scanning for sata_sil24...loaded Scanning for sata_svw...loaded Scanning for sata_nv...<hang> I have also had problems with nv_sata and the 2.6.26 and 2.6.27 kernels (see my report in http://forums.gentoo.org/viewtopic-t-727145-highlight-.html) I don't know if my report is related to your bug or not. In summary: (1) going from gentoo-sources-2.6.24-r8 to 2.6.26-r4 and 2.6.27-r7 more or less broke my sata_nv. (broken dmesg in the forum report.) (2) By running mint linux with the 2.6.27 kernel and studying lsmod, I was able to configure 2.6.27-r7 by just adding a few items to my old configuration so that it ran quiet satisfactorily. (3) 2.6.26-r4 failed to work when I used the configuration obtained from (2). I don't know whether any of this is of interest to you or not. If it is, let me know if you want any more info. (In reply to comment #25) > I have also had problems with nv_sata and the 2.6.26 and 2.6.27 kernels > (see my report in http://forums.gentoo.org/viewtopic-t-727145-highlight-.html) > I don't know if my report is related to your bug or not. In summary: Then please file a new bug. It is very frustrating to deal with 2 problems on 1 bug, so please help us avoid the potential frustration if you are in doubt! :) (In reply to comment #26) > (In reply to comment #25) > > I have also had problems with nv_sata and the 2.6.26 and 2.6.27 kernels > > (see my report in http://forums.gentoo.org/viewtopic-t-727145-highlight-.html) > > I don't know if my report is related to your bug or not. In summary: > > Then please file a new bug. It is very frustrating to deal with 2 problems on 1 > bug, so please help us avoid the potential frustration if you are in doubt! :) > Please ignore my comment 25. Thanks for explaining this too me and sorry to have bothered you. I thought (mistakenly) that the additional info might be helpful to you. Actually I don't view that I have a problem as I have configured a 2.7.27 kernel that works just fine, even though I can't use make oldconfig to configure it. I won't be filing a new bug as from my perspective everything is now fine. I have tried this again with 2.6.29-rc4. It doesn't matter if SMP or high mem is turned on or off, it still hangs. I think the SMP/high mem thing is a red herring and I didn't test it right the first time around. Tried again with 2.6.29-gentoo-r1 and it failed. I then pulled the plugs on both hard-drives and switched to different SATA ports on motherboard....now it works fine. They were previously on ports 0 and 1, now on like 6 and 7. This bug can be closed since this is no longer a problem for me. I will reopen if it starts failing again. |
