[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20110203144800.GA21160@redhat.com>
Date: Thu, 3 Feb 2011 09:48:00 -0500
From: Mike Snitzer <snitzer@...hat.com>
To: Tejun Heo <tj@...nel.org>
Cc: Jens Axboe <jaxboe@...ionio.com>, tytso@....edu, djwong@...ibm.com,
shli@...nel.org, neilb@...e.de, adilger.kernel@...ger.ca,
jack@...e.cz, linux-kernel@...r.kernel.org, kmannth@...ibm.com,
cmm@...ibm.com, linux-ext4@...r.kernel.org, rwheeler@...hat.com,
hch@....de, josef@...hat.com, jmoyer@...hat.com, vgoyal@...hat.com
Subject: [PATCH v4 1/2] block: skip elevator data initialization for flush
requests
Set REQ_SORTED, in the @rw_flags passed to the request allocator, for
all requests that may be put on IO scheduler. REQ_SORTED is not set for
flush requests because they are never put on the IO scheduler.
Skip elevator data initialization during request allocation if
REQ_SORTED is not set in @rw_flags passed to get_request().
Signed-off-by: Mike Snitzer <snitzer@...hat.com>
Acked-by: Vivek Goyal <vgoyal@...hat.com>
---
block/blk-core.c | 33 ++++++++++++++++++++++++++++-----
1 file changed, 28 insertions(+), 5 deletions(-)
v4: fixed bug where REQ_SORTED wasn't cleared on entry to get_request
- and Jens, yes I agree this is still a hack
Index: linux-2.6/block/blk-core.c
===================================================================
--- linux-2.6.orig/block/blk-core.c
+++ linux-2.6/block/blk-core.c
@@ -754,6 +754,9 @@ static void freed_request(struct request
/*
* Get a free request, queue_lock must be held.
+ * @rw_flags: may be overloaded to convey additional request features;
+ * any overloaded feature flags must be cleared immediately.
+ *
* Returns NULL on failure, with queue_lock held.
* Returns !NULL on success, with queue_lock *not held*.
*/
@@ -764,7 +767,11 @@ static struct request *get_request(struc
struct request_list *rl = &q->rq;
struct io_context *ioc = NULL;
const bool is_sync = rw_is_sync(rw_flags) != 0;
- int may_queue, priv;
+ const bool init_elv_data = !!(rw_flags & REQ_SORTED);
+ int may_queue, priv = 0;
+
+ if (init_elv_data)
+ rw_flags &= ~REQ_SORTED;
may_queue = elv_may_queue(q, rw_flags);
if (may_queue == ELV_MQUEUE_NO)
@@ -808,9 +815,14 @@ static struct request *get_request(struc
rl->count[is_sync]++;
rl->starved[is_sync] = 0;
- priv = !test_bit(QUEUE_FLAG_ELVSWITCH, &q->queue_flags);
- if (priv)
- rl->elvpriv++;
+ /*
+ * Only initialize elevator data if IO scheduler may be used.
+ */
+ if (init_elv_data) {
+ priv = !test_bit(QUEUE_FLAG_ELVSWITCH, &q->queue_flags);
+ if (priv)
+ rl->elvpriv++;
+ }
if (blk_queue_io_stat(q))
rw_flags |= REQ_IO_STAT;
@@ -1197,6 +1209,7 @@ static int __make_request(struct request
const unsigned short prio = bio_prio(bio);
const bool sync = !!(bio->bi_rw & REQ_SYNC);
const bool unplug = !!(bio->bi_rw & REQ_UNPLUG);
+ const bool flush = !!(bio->bi_rw & (REQ_FLUSH | REQ_FUA));
const unsigned long ff = bio->bi_rw & REQ_FAILFAST_MASK;
int where = ELEVATOR_INSERT_SORT;
int rw_flags;
@@ -1210,7 +1223,7 @@ static int __make_request(struct request
spin_lock_irq(q->queue_lock);
- if (bio->bi_rw & (REQ_FLUSH | REQ_FUA)) {
+ if (flush) {
where = ELEVATOR_INSERT_FLUSH;
goto get_rq;
}
@@ -1293,6 +1306,16 @@ get_rq:
rw_flags |= REQ_SYNC;
/*
+ * Set REQ_SORTED for all requests that may be put on IO scheduler.
+ * The request allocator's IO scheduler initialization will be skipped
+ * if REQ_SORTED is not set -- elv_set_request() is avoided so that
+ * that the allocated request's elevator_private pointers are not
+ * initialized and that space can be used by flush request data.
+ */
+ if (!flush)
+ rw_flags |= REQ_SORTED;
+
+ /*
* Grab a free request. This is might sleep but can not fail.
* Returns with the queue unlocked.
*/
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists