From: Dan Williams <dan.j.williams@intel.com>
To: neilb@suse.de
Cc: linux-raid@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: [PATCH v3 00/14] Asynchronous raid6 acceleration (part 1 of 3)
Date: Sat, 29 Aug 2009 19:30:00 -0700 [thread overview]
Message-ID: <20090830022908.5934.72142.stgit@dwillia2-linux.ch.intel.com> (raw)
Extend the async_tx api to include routines for offloading raid6 operations to
hardware. This is part1, part2 contains raid6 extensions to the iop-adma
driver, and part3 contains the changes to make the md/raid6 driver
asynchronous.
The combined series is available at:
git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx.git raid6
The changes since v2 [1] are:
1/ Moved various stripe handling resources to a percpu allocation.
Specifically the 'scribble' buffer (for passing and converting address
lists) and the raid6 spare page have been moved.
2/ Added 4-disk and 5-disk special cases to the raid6 recovery code as
they are special with respect to offload engines. The recovery
algorithms use a zero page in the place of the dead data pages. When
the async_tx api sees the raid6_empty_zero_page as a source it simply
drops it in the asynchronous path. This can result, for example, in a
PQ operation with a source count of zero in the case of recovering 2
data disks in a 4-disk array. This does not map to hardware so the
recovery code refactors the operation in these special cases. The
raid6test module has been updated to hit these cases.
I can carry this in async_tx.git/next, or let me know if you would
rather pull this into md.git/for-next.
Thanks,
Dan
[1]: http://marc.info/?l=linux-raid&m=124269482227272&w=2
Note, I know that "md/raid6: release spare page at ->stop()" is already
upstream, but I have not rebased the tree to preserve the testing that
has gone into this series to date.
---
Dan Williams (14):
async_tx: rename zero_sum to val
async_tx: kill ASYNC_TX_DEP_ACK flag
async_tx: structify submission arguments, add scribble
async_xor: permit callers to pass in a 'dma/page scribble' region
md/raid6: release spare page at ->stop()
md/raid6: move the spare page to a percpu allocation
md/raid5,6: add percpu scribble region for buffer lists
async_tx: add sum check flags
async_tx: kill needless module_{init|exit}
async_tx: remove walk of tx->parent chain in dma_wait_for_async_tx
async_tx: add support for asynchronous GF multiplication
async_tx: add support for asynchronous RAID6 recovery operations
dmatest: add pq support
async_tx: raid6 recovery self test
Documentation/crypto/async-tx-api.txt | 75 +++-
arch/arm/include/asm/hardware/iop3xx-adma.h | 5
arch/arm/mach-iop13xx/include/mach/adma.h | 12 -
arch/arm/mach-iop13xx/setup.c | 10 -
arch/arm/plat-iop/adma.c | 2
crypto/async_tx/Kconfig | 9 +
crypto/async_tx/Makefile | 3
crypto/async_tx/async_memcpy.c | 39 +-
crypto/async_tx/async_memset.c | 38 +-
crypto/async_tx/async_pq.c | 388 +++++++++++++++++++++++
crypto/async_tx/async_raid6_recov.c | 448 +++++++++++++++++++++++++++
crypto/async_tx/async_tx.c | 83 ++---
crypto/async_tx/async_xor.c | 199 ++++++------
crypto/async_tx/raid6test.c | 241 ++++++++++++++
drivers/dma/Kconfig | 2
drivers/dma/dmaengine.c | 53 +--
drivers/dma/dmatest.c | 26 ++
drivers/dma/iop-adma.c | 40 +-
drivers/md/Kconfig | 13 +
drivers/md/raid5.c | 452 ++++++++++++++++++---------
drivers/md/raid5.h | 22 +
include/linux/async_tx.h | 126 ++++++--
include/linux/dmaengine.h | 116 ++++++-
23 files changed, 1907 insertions(+), 495 deletions(-)
create mode 100644 crypto/async_tx/async_pq.c
create mode 100644 crypto/async_tx/async_raid6_recov.c
create mode 100644 crypto/async_tx/raid6test.c
next reply other threads:[~2009-08-30 2:30 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-08-30 2:30 Dan Williams [this message]
2009-08-30 2:30 ` [PATCH v3 01/14] async_tx: rename zero_sum to val Dan Williams
2009-08-30 2:30 ` [PATCH v3 02/14] async_tx: kill ASYNC_TX_DEP_ACK flag Dan Williams
2009-08-30 2:30 ` [PATCH v3 03/14] async_tx: structify submission arguments, add scribble Dan Williams
2009-08-30 2:30 ` [PATCH v3 04/14] async_xor: permit callers to pass in a 'dma/page scribble' region Dan Williams
2009-08-30 2:30 ` [PATCH v3 05/14] md/raid6: release spare page at ->stop() Dan Williams
2009-08-30 2:30 ` [PATCH v3 06/14] md/raid6: move the spare page to a percpu allocation Dan Williams
2009-08-30 2:30 ` [PATCH v3 07/14] md/raid5, 6: add percpu scribble region for buffer lists Dan Williams
2009-08-30 2:30 ` [PATCH v3 08/14] async_tx: add sum check flags Dan Williams
2009-08-30 2:30 ` [PATCH v3 09/14] async_tx: kill needless module_{init|exit} Dan Williams
2009-08-30 2:30 ` [PATCH v3 10/14] async_tx: remove walk of tx->parent chain in dma_wait_for_async_tx Dan Williams
2009-08-30 2:30 ` [PATCH v3 11/14] async_tx: add support for asynchronous GF multiplication Dan Williams
2009-08-30 2:31 ` [PATCH v3 12/14] async_tx: add support for asynchronous RAID6 recovery operations Dan Williams
2009-08-30 2:31 ` [PATCH v3 13/14] dmatest: add pq support Dan Williams
2009-08-30 2:31 ` [PATCH v3 14/14] async_tx: raid6 recovery self test Dan Williams
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20090830022908.5934.72142.stgit@dwillia2-linux.ch.intel.com \
--to=dan.j.williams@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-raid@vger.kernel.org \
--cc=neilb@suse.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.