linux-stable.git/fs/bio.c, branch linux-3.14.y

bio: return EINTR if copying to user space got interrupted

2016-03-09T23:33:25+00:00

commit 2d99b55d378c996b9692a0c93dd25f4ed5d58934 upstream.

Commit 35dc248383bbab0a7203fca4d722875bc81ef091 introduced a check for
current->mm to see if we have a user space context and only copies data
if we do. Now if an IO gets interrupted by a signal data isn't copied
into user space any more (as we don't have a user space context) but
user space isn't notified about it.

This patch modifies the behaviour to return -EINTR from bio_uncopy_user()
to notify userland that a signal has interrupted the syscall, otherwise
it could lead to a situation where the caller may get a buffer with
no data returned.

This can be reproduced by issuing SG_IO ioctl()s in one thread while
constantly sending signals to it.

Fixes: 35dc248 [SCSI] sg: Fix user memory corruption when SG_IO is interrupted by a signal
Signed-off-by: Johannes Thumshirn 
Signed-off-by: Hannes Reinecke 
Cc: stable@vger.kernel.org # v.3.11+
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

block: Do a full clone when splitting discard bios

2015-08-03T16:29:58+00:00

commit f3f5da624e0a891c34d8cd513c57f1d9b0c7dadc upstream.

This fixes a data corruption bug when using discard on top of MD linear,
raid0 and raid10 personalities.

Commit 20d0189b1012 "block: Introduce new bio_split()" permits sharing
the bio_vec between the two resulting bios. That is fine for read/write
requests where the bio_vec is immutable. For discards, however, we need
to be able to attach a payload and update the bio_vec so the page can
get mapped to a scatterlist entry. Therefore the bio_vec can not be
shared when splitting discards and we must do a full clone.

Signed-off-by: Martin K. Petersen 
Reported-by: Seunguk Shin 
Tested-by: Seunguk Shin 
Cc: Seunguk Shin 
Cc: Jens Axboe 
Cc: Kent Overstreet 
Reviewed-by: Christoph Hellwig 
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

block: Fix cloning of discard/write same bios

2014-02-11T15:40:45+00:00

Immutable biovecs changed the way bio segments are treated in such a way that
bio_for_each_segment() cannot now do what we want for discard/write same bios,
since bi_size means something completely different for them.

Fortunately discard and write same bios never have more than a single biovec, so
bio_for_each_segment() is unnecessary and not terribly meaningful for them, but
we still have to special case them in a few places.

Signed-off-by: Kent Overstreet 
Tested-by: Richard W.M. Jones 
Signed-off-by: Jens Axboe

Revert "block: Warn and free bio if bi_end_io is not set"

2014-01-08T21:14:22+00:00

This reverts commit 95d4403889acbd98e06d41a255df76452210996a.

The patch is broken for on-stack bios, amongst other things.

block: Warn and free bio if bi_end_io is not set

2014-01-08T16:39:49+00:00

In bio_endio if bio doesn't have bi_end_io (should be an error case),
we set bio to NULL and continue silently without freeing the bio. It
would be good to have a WARN and free the bio to avoid memory leak.

Signed-off-by: Muthukumar Ratty 
Signed-off-by: Jens Axboe

block: Kill bio_pair_split()

2013-11-24T06:33:57+00:00

Signed-off-by: Kent Overstreet 
Cc: Jens Axboe

block: Introduce new bio_split()

2013-11-24T06:33:57+00:00

The new bio_split() can split arbitrary bios - it's not restricted to
single page bios, like the old bio_split() (previously renamed to
bio_pair_split()). It also has different semantics - it doesn't allocate
a struct bio_pair, leaving it up to the caller to handle completions.

Then convert the existing bio_pair_split() users to the new bio_split()
- and also nvme, which was open coding bio splitting.

(We have to take that BUG_ON() out of bio_integrity_trim() because this
bio_split() needs to use it, and there's no reason it has to be used on
bios marked as cloned; BIO_CLONED doesn't seem to have clearly
documented semantics anyways.)

Signed-off-by: Kent Overstreet 
Cc: Jens Axboe 
Cc: Martin K. Petersen 
Cc: Matthew Wilcox 
Cc: Keith Busch 
Cc: Vishal Verma 
Cc: Jiri Kosina 
Cc: Neil Brown

block: Rename bio_split() -> bio_pair_split()

2013-11-24T06:33:56+00:00

This is prep work for introducing a more general bio_split().

Signed-off-by: Kent Overstreet 
Cc: Jens Axboe 
Cc: NeilBrown 
Cc: Alasdair Kergon 
Cc: Lars Ellenberg 
Cc: Peter Osterlund 
Cc: Sage Weil

block: Generic bio chaining

2013-11-24T06:33:56+00:00

This adds a generic mechanism for chaining bio completions. This is
going to be used for a bio_split() replacement, and it turns out to be
very useful in a fair amount of driver code - a fair number of drivers
were implementing this in their own roundabout ways, often painfully.

Note that this means it's no longer to call bio_endio() more than once
on the same bio! This can cause problems for drivers that save/restore
bi_end_io. Arguably they shouldn't be saving/restoring bi_end_io at all
- in all but the simplest cases they'd be better off just cloning the
bio, and immutable biovecs is making bio cloning cheaper. But for now,
we add a bio_endio_nodec() for these cases.

Signed-off-by: Kent Overstreet 
Cc: Jens Axboe

block: Remove bi_idx hacks

2013-11-24T06:33:55+00:00

Now that drivers have been converted to the new bvec_iter primitives,
there's no need to trim the bvec before we submit it; and we can't trim
it once we start sharing bvecs.

It used to be that passing a partially completed bio (i.e. one with
nonzero bi_idx) to generic_make_request() was a dangerous thing -
various drivers would choke on such things. But with immutable biovecs
and our new bio splitting that shares the biovecs, submitting partially
completed bios has to work (and should work, now that all the drivers
have been completed to the new primitives)

Signed-off-by: Kent Overstreet 
Cc: Jens Axboe