Subject : Very bad regression from 2.6.33 as of 1600f9def Submitter : Alex Elsayed <eternaleye@gmail.com> Date : 2010-04-29 2:28 Message-ID : loom.20100429T041908-663@post.gmane.org References : http://marc.info/?l=linux-kernel&m=127250825306178&w=2 This entry is being used for tracking a regression from 2.6.33. Please don't close it until the problem is fixed in the mainline.
Apparently, not a regression, but a "different .config" problem. Removing from the list.
Rafael, do you know what .config option is causing this? I've been unable to find it, so I can't build any new kernels. (Currently running 2.6.33, which is less than optimal.)
Not at the first sight, but you have some unusual storage-related .config options set. I'd start from these. [Also please attach the diff between the working and non-working .config from 2.6.33 to this bug entry.]
Created attachment 26756 [details] Diff between 2.6.33-zen0 config (working) and 2.6.33 vanilla config (broken)
Created attachment 26757 [details] Diff between 2.6.33-zen0 config (working) and 2.6.34-zen0 config (broken)
I attached diffs against a more recent version of the same (specialized) kernel as well as against vanilla 2.6.33, both of which have the same failure mode on boot.
A "git bisect" would be more useful.
Tried that. Even going back to revisions prior to the one that worked give the problem when I input the working config and 'make oldconfig'