On 12/28/06, Adam J. Richter <adam@xxxxxxxxxxxxx> wrote:
On Tue, 26 Dec 2006 06:36:45 +0800, Adam Richter wrote: >On Fri, 22 Dec 2006 14:56:20 +0100, Bartlomiej Zolnierkiewicz wrote: >>On 12/22/06, Adam J. Richter <adam@xxxxxxxxxxxxx> wrote: >>> I have a 666MHz Via Epia-ML motherboard (using a Via Eden, I >>> think). I am booting it from a CompactFlash card attached by an >>> IDE-CF adapter. Everything works fine under 2.6.18. Under 2.6.19, >>> 2.6.19.1 and 2.6.20-rc1, the master device on ide0 is not seen. >[...] >>>From the quick look the only IDE subsystem change in 2.6.19 >>which could be responsible (wild guess) for this is the commit: >>17c1033d331a430ce631805e15215e63b4cce764 >>[ patch attached in case you don't have git tree at hand ] >>Could you check if reverting it fixes the problem? >>If not it seems that the best way to proceed is git bisect. >I'll cc this message to linux-ide in case anyone else encounters the >problem and wants to be informed about it. >As mentioned in my previous off-list email, the reverting attached >patch did not fix the problem. >I now have a few more data points to narrow where the bug was introduced: >2.6.18.4 ok >2.6.18.6 ok >2.6.19-rc1 fails >I expect I will now work on setting up a git tree fed by kernel.org >and using git bisect as you suggest. >Thanks for your advice. I'll let you know when I have more news. Before pulling from the official git development tree, I tried the linux-2.6.18-git[n] snapshots. It looks the problem goes back as far as 2.6.18-git1. Here is a list of test results, for completeness. 2.6.18 ok 2.6.18.1 ok 2.6.18.4 ok 2.6.18.6 ok 2.6.18-git1 fails 2.6.18-git2 fails 2.6.18-git5 fails 2.6.18-git10 fails 2.6.19-rc1 fails 2.6.19 fails 2.6.20-rc1 fails 2.6.20-rc2 fails
Thanks for testing.
Now I guess will pull from the official git development tree and try using git bisect to narrow the problem further.
Yes, 2.6.18-git1 patch is still near 3MB large, more-over [bzolnier@trik adam]$ grep drivers/ide patch-2.6.18-git1 diff --git a/drivers/ide/ppc/pmac.c b/drivers/ide/ppc/pmac.c --- a/drivers/ide/ppc/pmac.c +++ b/drivers/ide/ppc/pmac.c so it looks like some non-IDE commit is affecting IDE driver. :? Bart - To unsubscribe from this list: send the line "unsubscribe linux-ide" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
