* blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly
@ 2018-05-16 2:54 胡海
0 siblings, 0 replies; 7+ messages in thread
From: 胡海 @ 2018-05-16 2:54 UTC (permalink / raw)
To: axboe; +Cc: linux-block, hch
[-- Attachment #1: Type: text/plain, Size: 1080 bytes --]
Author: huhai <huhai@kylinos.cn>
Date: Wed May 16 10:34:22 2018 +0800
blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly
For sq make_request instances, we should issue sync request directly too,
otherwise it will break down the semantics of sync request,
the current code logic is to send synchronous requests asynchronously.
Signed-off-by: huhai <huhai@kylinos.cn>
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 5629f18..fcf2f16 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -1771,7 +1771,7 @@ static blk_qc_t blk_mq_make_request(struct request_queue *q, struct bio *bio)
blk_mq_try_issue_directly(data.hctx, same_queue_rq,
&cookie);
}
- } else if (q->nr_hw_queues > 1 && is_sync) {
+ } else if (is_sync) {
blk_mq_put_ctx(data.ctx);
blk_mq_bio_to_request(rq, bio);
blk_mq_try_issue_directly(data.hctx, rq, &cookie);
[-- Attachment #2: Type: text/html, Size: 1922 bytes --]
^ permalink raw reply related [flat|nested] 7+ messages in thread* blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly
@ 2018-05-16 3:43 胡海
0 siblings, 0 replies; 7+ messages in thread
From: 胡海 @ 2018-05-16 3:43 UTC (permalink / raw)
To: axboe; +Cc: linux-block, hch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^ permalink raw reply [flat|nested] 7+ messages in thread
* blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly
@ 2018-05-16 3:42 胡海
2018-05-16 14:15 ` Jens Axboe
0 siblings, 1 reply; 7+ messages in thread
From: 胡海 @ 2018-05-16 3:42 UTC (permalink / raw)
To: linux-block
QXV0aG9yOiBodWhhaSA8aHVoYWlAa3lsaW5vcy5jbj4NCkRhdGU6ICAgV2VkIE1heSAxNiAx
MDozNDoyMiAyMDE4ICswODAwDQoNCiAgICBibGstbXE6IGZvciBzeW5jIGNhc2UsIHdoZXRo
ZXIgaXQgaXMgbXEgb3Igc3EgbWFrZV9yZXF1ZXN0IGluc3RhbmNlcywgd2Ugc2hvdWxkIHNl
bmQgdGhlIHJlcXVlc3QgZGlyZWN0bHkNCiAgICANCiAgICBGb3Igc3EgbWFrZV9yZXF1ZXN0
IGluc3RhbmNlcywgd2Ugc2hvdWxkIGlzc3VlIHN5bmMgcmVxdWVzdCBkaXJlY3RseSB0b28s
IA0KICAgIG90aGVyd2lzZSBpdCB3aWxsIGJyZWFrIGRvd24gdGhlIHNlbWFudGljcyBvZiBz
eW5jIHJlcXVlc3QsIA0KICAgIHRoZSBjdXJyZW50IGNvZGUgbG9naWMgaXMgdG8gc2VuZCBz
eW5jaHJvbm91cyByZXF1ZXN0cyBhc3luY2hyb25vdXNseS4NCiAgICANCiAgICBTaWduZWQt
b2ZmLWJ5OiBodWhhaSA8aHVoYWlAa3lsaW5vcy5jbj4NCg0KZGlmZiAtLWdpdCBhL2Jsb2Nr
L2Jsay1tcS5jIGIvYmxvY2svYmxrLW1xLmMNCmluZGV4IDU2MjlmMTguLmZjZjJmMTYgMTAw
NjQ0DQotLS0gYS9ibG9jay9ibGstbXEuYw0KKysrIGIvYmxvY2svYmxrLW1xLmMNCkBAIC0x
NzcxLDcgKzE3NzEsNyBAQCBzdGF0aWMgYmxrX3FjX3QgYmxrX21xX21ha2VfcmVxdWVzdChz
dHJ1Y3QgcmVxdWVzdF9xdWV1ZSAqcSwgc3RydWN0IGJpbyAqYmlvKQ0KICAgICAgICAgICAg
ICAgICAgICAgICAgYmxrX21xX3RyeV9pc3N1ZV9kaXJlY3RseShkYXRhLmhjdHgsIHNhbWVf
cXVldWVfcnEsDQogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgJmNv
b2tpZSk7DQogICAgICAgICAgICAgICAgfQ0KLSAgICAgICB9IGVsc2UgaWYgKHEtPm5yX2h3
X3F1ZXVlcyA+IDEgJiYgaXNfc3luYykgew0KKyAgICAgICB9IGVsc2UgaWYgKGlzX3N5bmMp
IHsNCiAgICAgICAgICAgICAgICBibGtfbXFfcHV0X2N0eChkYXRhLmN0eCk7DQogICAgICAg
ICAgICAgICAgYmxrX21xX2Jpb190b19yZXF1ZXN0KHJxLCBiaW8pOw0KICAgICAgICAgICAg
ICAgIGJsa19tcV90cnlfaXNzdWVfZGlyZWN0bHkoZGF0YS5oY3R4LCBycSwgJmNvb2tpZSk7
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly
2018-05-16 3:42 胡海
@ 2018-05-16 14:15 ` Jens Axboe
0 siblings, 0 replies; 7+ messages in thread
From: Jens Axboe @ 2018-05-16 14:15 UTC (permalink / raw)
To: 胡海, linux-block
On 5/15/18 9:42 PM, 胡海 wrote:
> Author: huhai <huhai@kylinos.cn>
> Date: Wed May 16 10:34:22 2018 +0800
>
> blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly
>
> For sq make_request instances, we should issue sync request directly too,
> otherwise it will break down the semantics of sync request,
> the current code logic is to send synchronous requests asynchronously.
It doesn't break the semantics of a sync request. The reason we don't do
it on single queue hardware is for efficiency reasons. On multiple queue
hardware, we can efficiently have multiple processes doing sync issue,
that isn't the case on single queue hardware.
--
Jens Axboe
^ permalink raw reply [flat|nested] 7+ messages in thread
* blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly
@ 2018-05-16 3:24 胡海
0 siblings, 0 replies; 7+ messages in thread
From: 胡海 @ 2018-05-16 3:24 UTC (permalink / raw)
To: axboe; +Cc: linux-block
[-- Attachment #1: Type: text/plain, Size: 1048 bytes --]
Author: huhai <huhai@kylinos.cn>
Date: Wed May 16 10:34:22 2018 +0800
blk-mq: for sync case, whether it is mq or sq make_request instances,
we should send the request directly
For sq make_request instances, we should issue sync request directly
too,
otherwise it will break down the semantics of sync request,
the current code logic is to send synchronous requests asynchronously.
Signed-off-by: huhai <huhai@kylinos.cn>
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 5629f18..fcf2f16 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -1771,7 +1771,7 @@ static blk_qc_t blk_mq_make_request(struct
request_queue *q, struct bio *bio)
blk_mq_try_issue_directly(data.hctx, same_queue_rq,
&cookie);
}
- } else if (q->nr_hw_queues > 1 && is_sync) {
+ } else if (is_sync) {
blk_mq_put_ctx(data.ctx);
blk_mq_bio_to_request(rq, bio);
blk_mq_try_issue_directly(data.hctx, rq, &cookie);
[-- Attachment #2: Type: text/html, Size: 1340 bytes --]
^ permalink raw reply related [flat|nested] 7+ messages in thread* blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly
@ 2018-05-16 2:56 胡海
0 siblings, 0 replies; 7+ messages in thread
From: 胡海 @ 2018-05-16 2:56 UTC (permalink / raw)
To: axboe; +Cc: linux-block
[-- Attachment #1: Type: text/plain, Size: 1080 bytes --]
Author: huhai <huhai@kylinos.cn>
Date: Wed May 16 10:34:22 2018 +0800
blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly
For sq make_request instances, we should issue sync request directly too,
otherwise it will break down the semantics of sync request,
the current code logic is to send synchronous requests asynchronously.
Signed-off-by: huhai <huhai@kylinos.cn>
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 5629f18..fcf2f16 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -1771,7 +1771,7 @@ static blk_qc_t blk_mq_make_request(struct request_queue *q, struct bio *bio)
blk_mq_try_issue_directly(data.hctx, same_queue_rq,
&cookie);
}
- } else if (q->nr_hw_queues > 1 && is_sync) {
+ } else if (is_sync) {
blk_mq_put_ctx(data.ctx);
blk_mq_bio_to_request(rq, bio);
blk_mq_try_issue_directly(data.hctx, rq, &cookie);
[-- Attachment #2: Type: text/html, Size: 1882 bytes --]
^ permalink raw reply related [flat|nested] 7+ messages in thread* blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly
@ 2018-05-16 2:44 胡海
0 siblings, 0 replies; 7+ messages in thread
From: 胡海 @ 2018-05-16 2:44 UTC (permalink / raw)
To: axboe; +Cc: linux-block, hch
[-- Attachment #1: Type: text/plain, Size: 1080 bytes --]
Author: huhai <huhai@kylinos.cn>
Date: Wed May 16 10:34:22 2018 +0800
blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly
For sq make_request instances, we should issue sync request directly too,
otherwise it will break down the semantics of sync request,
the current code logic is to send synchronous requests asynchronously.
Signed-off-by: huhai <huhai@kylinos.cn>
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 5629f18..fcf2f16 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -1771,7 +1771,7 @@ static blk_qc_t blk_mq_make_request(struct request_queue *q, struct bio *bio)
blk_mq_try_issue_directly(data.hctx, same_queue_rq,
&cookie);
}
- } else if (q->nr_hw_queues > 1 && is_sync) {
+ } else if (is_sync) {
blk_mq_put_ctx(data.ctx);
blk_mq_bio_to_request(rq, bio);
blk_mq_try_issue_directly(data.hctx, rq, &cookie);
[-- Attachment #2: Type: text/html, Size: 2002 bytes --]
^ permalink raw reply related [flat|nested] 7+ messages in thread
end of thread, other threads:[~2018-05-16 14:15 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-05-16 2:54 blk-mq: for sync case, whether it is mq or sq make_request instances, we should send the request directly 胡海
-- strict thread matches above, loose matches on Subject: below --
2018-05-16 3:43 胡海
2018-05-16 3:42 胡海
2018-05-16 14:15 ` Jens Axboe
2018-05-16 3:24 胡海
2018-05-16 2:56 胡海
2018-05-16 2:44 胡海
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).