Red Hat Bugzilla – Bug 469611
endless loop in mkinitrd
Last modified: 2008-12-04 23:02:17 EST
Created attachment 322261 [details]
Patch that avoids searches in the wrong dirs and mismatched qpops, and thus the endless loop
Description of problem:
When qpushd fails because the named directory does not exist, we go happily ahead searching for stuff in the wrong directory, and then qpop what the caller had pushed. Enough of these pops and we end up back in /tmp, and entering an endless loop because $PWD will never end up matching /sys, no matter how many times we cd .. in findstoragedriverinsys.
I ran into this after a couple of USB bus resets left me with two failed slaves for the RAID device holding the root filesystem (in an encrypted physical volume), later re-added as the disk came back up:
md3 : active raid1 sde3 sdd3(F) sda3 sdb3 sdc3(F)
77650112 blocks [3/3] [UUU]
bitmap: 0/149 pages [0KB], 256KB chunk
It didn't matter that bug 465542 wasn't fixed yet, it still tried to descend into the components of md3, and failed at that.
The attached patch, along with the patch for bug 465542 (strictly speaking, probably not needed to fix the bug, but what the heck :-) fixes it.
Version-Release number of selected component (if applicable):
Should be fixed in 6.0.70-1 .
This bug appears to have been reported against 'rawhide' during the Fedora 10 development cycle.
Changing version to '10'.
More information and reason for this action is here:
Confirmed fixed in F-10 GOLD, thanks.