linux-kernel - [GIT PATCHSET block#for-linus] block: blk-map related fixes

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:	Wed,  1 Apr 2009 20:04:36 +0900
From:	Tejun Heo <tj@...nel.org>
To:	axboe@...nel.dk, bharrosh@...asas.com, linux-kernel@...r.kernel.org
Subject: [GIT PATCHSET block#for-linus] block: blk-map related fixes

Hi, Jens.

Upon ack, please pull from the following git tree.

 git://git.kernel.org/pub/scm/linux/kernel/git/tj/misc.git blk-map-fixes
 http://git.kernel.org/?p=linux/kernel/git/tj/misc.git;a=shortlog;h=blk-map-fixes

This patchset contains the following eight patches to fix subtle bugs
in scatterlist and blk-map and are candidates for 2.6.30 inclusion.

 0001-scatterlist-make-sure-sg_miter_next-doesn-t-retur.patch
 0002-block-fix-SG_IO-vector-request-data-length-handling.patch
 0003-block-fix-queue-bounce-limit-setting.patch
 0004-bio-actually-inline-inline-bvecs-into-bio.patch
 0005-bio-fix-bio_kmalloc.patch
 0006-bio-remove-size-segments-limit-on-bio_-copy-map-_-u.patch
 0007-blk-map-let-blk_rq_map_user_iov-support-null-mapp.patch
 0008-blk-map-reimplement-blk_rq_map_user-using-blk_rq_.patch

0001 fixes a subtle bug in sg mapping iterator where zero-size maps
were returned.  The bug isn't critical but can sometimes lead to
unexpected results (when testing for full consumption for example).

0002 is visible to userland.  It fixes block iovec SG_IOs such that
when sghdr->dxfer_len and the data length in the iovec disagrees, the
shorter one wins instead of failing as described in the SG_IO howto.

0003 is pretty critical.  The bug means that GFP_DMA is set in
q->bounce_gfp for many ATA controllers on x86_64 which leads to OOM if
large SG_IO is issued.

0004-0008 fixes bio_kmalloc(), removes bio-imposed segment and length
limit when using bio_kmalloc() and makes blk_rq_map_{user|kern}*()
functions always use single bio instead of chain of bios.  It's not
only simpler but also makes sure the resulting request falls inside
queue limits as a whole.  Previously, only each bio was checked for
limits, so the whole request could go over the limits.

This patchset is on top of the current linux-2.6-block#for-linus[1]
and contains the following changes.

 block/blk-map.c            |  119 +++------------------------------------------
 block/blk-settings.c       |   20 ++++---
 block/scsi_ioctl.c         |   13 ++++
 drivers/md/dm-io.c         |    4 -
 drivers/md/dm.c            |    4 -
 drivers/scsi/scsi_lib.c    |    3 -
 fs/bio.c                   |   97 +++++++++++++++++++-----------------
 fs/btrfs/extent_io.c       |    2 
 fs/ext4/extents.c          |    4 -
 fs/xfs/linux-2.6/xfs_buf.c |    2 
 include/linux/bio.h        |   16 +++---
 lib/scatterlist.c          |    9 ++-
 12 files changed, 109 insertions(+), 184 deletions(-)

Thanks.

--
tejun

[1] 714ed0cf62319b14dc327273a7339a9a199fe046
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/