On Mon, Aug 26, 2019 at 06:40:24PM +0300, Nikolay Borisov wrote: > >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=204371 > >> Fixes: 69d2480456d1 ("btrfs: use copy_page for copying pages instead of memcpy") > >> Cc: stable@xxxxxxxxxxxxxxx > >> Signed-off-by: Christophe Leroy <christophe.leroy@xxxxxx> > >> --- > >> v2: Using kmem_cache instead of get_zeroed_page() in order to benefit from SLAB debugging features like redzone. > > > > I'll take this version, thanks. Though I'm not happy about the allocator > > behaviour. The kmem cache based fix can be backported independently to > > 4.19 regardless of the SL*B fixes. > > > >> +extern struct kmem_cache *btrfs_bitmap_cachep; > > > > I've renamed the cache to btrfs_free_space_bitmap_cachep > > > > Reviewed-by: David Sterba <dsterba@xxxxxxxx> > > Isn't this obsoleted by > > '[PATCH v2 0/2] guarantee natural alignment for kmalloc()' ? Yeah, but this would add maybe another whole dev cycle to merge and release. The reporters of the bug seem to care enough to identify the problem and propose the fix so I feel like adding the btrfs-specific fix now is a little favor we can afford. The bug is reproduced on an architecture that's not widely tested so from practical POV I think this adds more coverage which is desirable.
