On Mon, Apr 23, 2018 at 02:03:52AM -0400, Theodore Y. Ts'o wrote: > On Sun, Apr 22, 2018 at 10:54:23PM +0100, Ben Hutchings wrote: > > > offset = blk - group_first_block; > > > - if (!ext4_test_bit(EXT4_B2C(sbi, offset), bh->b_data)) > > > + if (offset < 0 || EXT4_B2C(sbi, offset) >= sb->s_blocksize || > > > + !ext4_test_bit(EXT4_B2C(sbi, offset), bh->b_data)) > > > > Isn't sb->s_blocksize a count of bytes? If so, doesn't that mean that > > we should be comparing with sb->s_blocksize * 8? > > Yes, nice catch, Ben! Can we temporarily drop this patch from the > stable queue until I can get a fix in? > > In practice this shouldn't be a problem because of the default ext4 > layout with flex_bg. But we should definitely get this fixed before > we let this flow into the stable kernel. Now dropped from all queues. When you submit the fix, can you also tag it for stable so I know to pick this one back up? thanks, greg k-h