On a 32bit machine, passing the negative unsigned long fe_len into percpu_counter_add (which expects s64) messes up the counters: [root@bear-05 linux-2.6.23-rc7]# df -h Filesystem Size Used Avail Use% Mounted on /dev/sdb7 97G -149P 149P - /mnt/test Patch below fixes this. Signed-off-by: Eric Sandeen <sandeen@xxxxxxxxxx> --- linux.orig/fs/ext4/mballoc.c 2007-10-25 14:47:25.048435451 -0500 +++ linux/fs/ext4/mballoc.c 2007-10-25 14:47:35.815160682 -0500 @@ -3005,7 +3005,7 @@ - ac->ac_b_ex.fe_len); gdp->bg_checksum = ext4_group_desc_csum(sbi, ac->ac_b_ex.fe_group, gdp); spin_unlock(sb_bgl_lock(sbi, ac->ac_b_ex.fe_group)); - percpu_counter_add(&sbi->s_freeblocks_counter, - ac->ac_b_ex.fe_len); + percpu_counter_sub(&sbi->s_freeblocks_counter, ac->ac_b_ex.fe_len); err = ext4_journal_dirty_metadata(handle, bitmap_bh); if (err) - To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
