Red Hat Bugzilla – Bug 202715
Loop in __find_get_block_slow after building a Kadischi FC6T2 LiveCD
Last modified: 2015-01-04 17:28:20 EST
Description of problem:
After building a LiveCD using Kadischi using a FC6T2 stock disc as a repository,
I boot the CD image in Qemu and/or on normal hardware and get an endless loop in
The loop looks like so:
__find_get_block_slow() failed. block=18446744071562067968, b_blocknr=2147483648
device blocksize: 4096
Version-Release number of selected component (if applicable):
Build LiveCD with Kadischi, boot teh CD under Qemu or normal hardware.
Steps to Reproduce:
1. kadischi --text /path/to/FC6T2/repository -f /tmp/fc6t2live.iso
Loop in __find_getblock_slow() kernel problem.
Should boot into a Live system.
Relevant information includes:
We already load cdrom.ko and ide-cd.ko which is used to allow find_live_cd.c in
Kadischi find and mount the CD in question after finding a /.livecd file on the
CD. This may or may not be a problem in find_live_cd.c or the kernel it's self
but I want to get more info before making an assumption.
cdrom.ko and ide-cd.ko were a part of the kernel before, and this didn't happen.
This could be a cause?
Also: Kadischi uses an Ext2 gzipped image as initrd, is this a problem either?
As The loop ends up happenning right after finding a compressed RAMDISK image at
block 0 during kernel bootstrap.
I happenned to be able to just get a peek at what it says before the loop and
after finding a compressed RAMDISK at block 0, and it does say:
Block size too small for device
I will try with a larger block size to Kadischi during making the initrd, and
report back results before pursuing this incedent as a bug. Thanks.
Adding mke2fs -b 4096 _does_ rectify the situation in Kadischi LiveCDs when
making the initrd..
This is using a initrd at size 10MB.
If this is right, please close this bug. If not I can provide any additional
information needed to rectify this issue.
the initrd has to be 4kb blocksize.