Description of problem: I get this message from a kickstart install (from HD) of FC3 rc2: *** glibc detected *** double free or corruption: 0x000000000079d4e5 Then the installation aborts. VT3 says: [...] * Found hd stage2 * Copying /tmp/loopimage//Fedora/base/stage2.img in RAM as stage 2 * copied 77369344 bytes to /tmp/ramfs/stage2.img (complete) * mntloop loop1 on /mnt/runtime as /tmp/ramfs/stage2.img fd is 14 * umounting loopback /tmp/loopimage loop0 Version-Release number of selected component (if applicable): anaconda-10.1.0.2-1.x86_64.rpm How reproducible: 100% Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Also happens with FC3-3.4.
Created attachment 105939 [details] ks-auto.cfg Here is the kickstart file I'm using.
Same happens with RC5, except the address is 0x79d565. Can anyone else reproduce this, or is it just me?
Hmm, I might have found the problem -- testing a fix.
Created attachment 106064 [details] anaconda-free.patch Here is the fix. We are passing ptr+5 to free() instead of ptr.
I've checked in this fix to HEAD and to fc3-branch.
Thanks Tim. Probably won't get into FC3 unless we do a respin at this point (workaround is to remove /dev from your device name)
*** Bug 139488 has been marked as a duplicate of this bug. ***