Messages in this thread | | | Date | Sat, 3 Aug 2013 20:49:04 -0400 | From | Theodore Ts'o <> | Subject | Re: [PATCH 1/1] EXT4: LazyInit Mount Bug Fix |
| |
On Mon, Jul 15, 2013 at 12:14:18PM +0530, Nitin Singla wrote: > - sbi->s_itb_per_group = sbi->s_inodes_per_group / > - sbi->s_inodes_per_block; > + sbi->s_itb_per_group = DIV_ROUND_UP(sbi->s_inodes_per_group, > + sbi->s_inodes_per_block);
This would only matter if s_inodes_per_group is not a multiple of s_inodes_per_block. Which is never supposed to happen; mke2fs doesn't create file systems like this.
Ancient Android build systems, before the bug was fixed, did do this in the past, but that was a long time ago. Where did this file system come from?
I could apply this patch, but you should be warned that there may be other bugs hiding here for file systems like this, both in the kernel and in e2fsprogs.
- Ted
| |