From: Dan Williams <dan.j.williams@intel.com>
To: neilb@suse.de
Cc: linux-raid@vger.kernel.org, Arjan van de Ven <arjan@linux.intel.com>
Subject: [PATCH v2 9/9] md/raid456: distribute raid processing over multiple cores
Date: Mon, 31 Aug 2009 09:41:50 -0700 [thread overview]
Message-ID: <20090831164150.1696.83369.stgit@dwillia2-linux.ch.intel.com> (raw)
In-Reply-To: <20090831163914.1696.55782.stgit@dwillia2-linux.ch.intel.com>
Now that the resources to handle stripe_head operations are allocated
percpu it is possible for raid5d to distribute stripe handling over
multiple cores. This conversion also adds a call to cond_resched() in
the non-multicore case to prevent one core from getting monopolized for
raid operations.
Cc: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
drivers/md/Kconfig | 11 +++++++++++
drivers/md/raid5.c | 36 ++++++++++++++++++++++++++++++++++--
2 files changed, 45 insertions(+), 2 deletions(-)
diff --git a/drivers/md/Kconfig b/drivers/md/Kconfig
index abb8636..09c0c6e 100644
--- a/drivers/md/Kconfig
+++ b/drivers/md/Kconfig
@@ -154,6 +154,17 @@ config MD_RAID456
If unsure, say Y.
+config MULTICORE_RAID456
+ bool "RAID-4/RAID-5/RAID-6 Multicore processing (EXPERIMENTAL)"
+ depends on MD_RAID456
+ depends on SMP
+ depends on EXPERIMENTAL
+ ---help---
+ Enable the raid456 module to dispatch per-stripe raid operations to a
+ thread pool.
+
+ If unsure, say N.
+
config MD_RAID6_PQ
tristate
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 7c22e19..364ea37 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -47,6 +47,7 @@
#include <linux/kthread.h>
#include <linux/raid/pq.h>
#include <linux/async_tx.h>
+#include <linux/async.h>
#include <linux/seq_file.h>
#include <linux/cpu.h>
#include "md.h"
@@ -4314,6 +4315,36 @@ static int retry_aligned_read(raid5_conf_t *conf, struct bio *raid_bio)
return handled;
}
+#ifdef CONFIG_MULTICORE_RAID456
+static void __process_stripe(void *param, async_cookie_t cookie)
+{
+ struct stripe_head *sh = param;
+
+ handle_stripe(sh);
+ release_stripe(sh);
+}
+
+static void process_stripe(struct stripe_head *sh, struct list_head *domain)
+{
+ async_schedule_domain(__process_stripe, sh, domain);
+}
+
+static void synchronize_stripe_processing(struct list_head *domain)
+{
+ async_synchronize_full_domain(domain);
+}
+#else
+static void process_stripe(struct stripe_head *sh, struct list_head *domain)
+{
+ handle_stripe(sh);
+ release_stripe(sh);
+ cond_resched();
+}
+
+static void synchronize_stripe_processing(struct list_head *domain)
+{
+}
+#endif
/*
@@ -4328,6 +4359,7 @@ static void raid5d(mddev_t *mddev)
struct stripe_head *sh;
raid5_conf_t *conf = mddev_to_conf(mddev);
int handled;
+ LIST_HEAD(raid_domain);
pr_debug("+++ raid5d active\n");
@@ -4364,8 +4396,7 @@ static void raid5d(mddev_t *mddev)
spin_unlock_irq(&conf->device_lock);
handled++;
- handle_stripe(sh);
- release_stripe(sh);
+ process_stripe(sh, &raid_domain);
spin_lock_irq(&conf->device_lock);
}
@@ -4373,6 +4404,7 @@ static void raid5d(mddev_t *mddev)
spin_unlock_irq(&conf->device_lock);
+ synchronize_stripe_processing(&raid_domain);
async_tx_issue_pending_all();
unplug_slaves(mddev);
next prev parent reply other threads:[~2009-08-31 16:41 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-08-31 16:41 [PATCH v2 0/9] Asynchronous raid6 acceleration (part 3 of 3) Dan Williams
2009-08-31 16:41 ` [PATCH v2 1/9] md/raid5: factor out mark_uptodate from ops_complete_compute5 Dan Williams
2009-08-31 16:41 ` [PATCH v2 2/9] md/raid6: asynchronous raid6 operations Dan Williams
2009-09-15 5:32 ` Neil Brown
2009-08-31 16:41 ` [PATCH v2 3/9] md/raid5, 6: common schedule_reconstruction for raid5/6 Dan Williams
2009-08-31 16:41 ` [PATCH v2 4/9] md/raid6: asynchronous handle_stripe_fill6 Dan Williams
2009-08-31 16:41 ` [PATCH v2 5/9] md/raid6: asynchronous handle_stripe_dirtying6 Dan Williams
2009-08-31 16:41 ` [PATCH v2 6/9] md/raid6: asynchronous handle_parity_check6 Dan Williams
2009-08-31 16:41 ` [PATCH v2 7/9] md/raid6: asynchronous handle_stripe6 Dan Williams
2009-09-15 5:26 ` Neil Brown
2009-09-15 8:42 ` Dan Williams
2009-08-31 16:41 ` [PATCH v2 8/9] md/raid6: remove synchronous infrastructure Dan Williams
2009-08-31 16:41 ` Dan Williams [this message]
2009-08-31 17:23 ` [PATCH v2 0/9] Asynchronous raid6 acceleration (part 3 of 3) kwick
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=20090831164150.1696.83369.stgit@dwillia2-linux.ch.intel.com \
--to=dan.j.williams@intel.com \
--cc=arjan@linux.intel.com \
--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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).