[dm-devel] dm-crypt: Fix error with too large bios
Mikulas Patocka
mpatocka at redhat.com
Mon Aug 15 17:03:37 UTC 2016
On Sat, 13 Aug 2016, Mike Snitzer wrote:
> [top-posting just because others went wild with it]
>
> I don't have a strong opinion but I just assumed the local dm-crypt
> workaround wasn't the way forward. I didn't stage it because Christoph
> disagreed with it:
> https://lkml.org/lkml/2016/6/1/456
> https://lkml.org/lkml/2016/6/1/477
>
> Also, this would appear to be a more generic fix:
> "block: make sure big bio is splitted into at most 256 bvecs
> https://lkml.org/lkml/2016/8/12/154
> (but Christoph disagrees there too, so the way forward isn't clear)
>
> Mike
On Wed, Jun 01 2016 at 9:44am -0400, Christoph Hellwig <hch at infradead.org> wrote:
> > > be dm-crypt.c. Maybe you've identified some indirect use of
> > > BIO_MAX_SIZE?
> >
> > I mean the recently introduced BIO_MAX_SIZE in -next tree:
> >
> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/commit/drivers/md/dm-crypt.c?id=4ed89c97b0706477b822ea2182827640c0cec486
>
> The crazy bcache bios striking back once again. I really think it's
> harmful having a _MAX value and then having a minor driver
> reinterpreting it and sending larger ones. Until we can lift the
> maximum limit in general nad have common code exercise it we really need
> to stop bcache from sending these instead of littering the tree with
> workarounds.
The bio_kmalloc function allocates bios with up to 1024 vector entries (as
opposed to bio_alloc and bio_alloc_bioset that has a limit of 256 vector
entries).
Device mapper is using bio_alloc_bioset with a bio set, so it is limited
to 256 vector entries, but other kernel users may use bio_kmalloc and
create larger bios.
So, if you don't want bios with more than 256 vector entries to exist, you
should impose this limit in bio_kmalloc (and fix all the callers that use
it).
Mikulas
More information about the dm-devel
mailing list