Skip to contentSkip to conversation

blk-mq: Minor tweaks

PATCH In review Alexander Gordeev 3 participants 10 messages Sep 3, 2014 – Sep 5, 2014 Block layer

lore
Share & tools
lore
#0 lore raw

cover blk-mq: Minor tweaks

Hi Jens,

These are few changes to blk-mq/blk-mq-tag, nothing really serious.

Cc Jens Axboe <[email protected]>

Alexander Gordeev (3):
blk-mq: Cleanup blk_mq_tag_busy() and blk_mq_tag_idle()
blk-mq: Eliminate superfluous check of BLK_MQ_S_TAG_ACTIVE flag
blk-mq: Fix formula to calculate fair share of tags

 block/blk-mq-tag.c |  9 +++------
 block/blk-mq-tag.h | 17 ++++++++---------
 2 files changed, 11 insertions(+), 15 deletions(-)
Signature · 1 line
1.9.3
#1 lore raw

PATCH 1/3 blk-mq: Cleanup blk_mq_tag_busy() and blk_mq_tag_idle()

---
 block/blk-mq-tag.c |  4 +---
 block/blk-mq-tag.h | 17 ++++++++---------
 2 files changed, 9 insertions(+), 12 deletions(-)
block/blk-mq-tag.c +1 -3
index c1b9242..4953b64 100644
@@ -58,13 +58,11 @@ static inline void bt_index_atomic_inc(atomic_t *index)
 /*
  * If a previously inactive queue goes active, bump the active user count.
  */
-bool __blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
+void __blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
 {
 	if (!test_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state) &&
 	    !test_and_set_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state))
 		atomic_inc(&hctx->tags->active_queues);
-
-	return true;
 }
 
 /*
block/blk-mq-tag.h +8 -9
index 6206ed1..795ec3f 100644
@@ -66,23 +66,22 @@ enum {
 	BLK_MQ_TAG_MAX		= BLK_MQ_TAG_FAIL - 1,
 };
 
-extern bool __blk_mq_tag_busy(struct blk_mq_hw_ctx *);
+extern void __blk_mq_tag_busy(struct blk_mq_hw_ctx *);
 extern void __blk_mq_tag_idle(struct blk_mq_hw_ctx *);
 
 static inline bool blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
 {
-	if (!(hctx->flags & BLK_MQ_F_TAG_SHARED))
-		return false;
-
-	return __blk_mq_tag_busy(hctx);
+	if (hctx->flags & BLK_MQ_F_TAG_SHARED) {
+		__blk_mq_tag_busy(hctx);
+		return true;
+	}
+	return false;
 }
 
 static inline void blk_mq_tag_idle(struct blk_mq_hw_ctx *hctx)
 {
-	if (!(hctx->flags & BLK_MQ_F_TAG_SHARED))
-		return;
-
-	__blk_mq_tag_idle(hctx);
+	if (hctx->flags & BLK_MQ_F_TAG_SHARED)
+		__blk_mq_tag_idle(hctx);
 }
 
 #endif
Signature · 1 line
1.9.3
maintainer
#2 lore raw

Re: [PATCH 1/3] blk-mq: Cleanup blk_mq_tag_busy() and blk_mq_tag_idle()

Quoted 14 lines · Alexander Gordeev
---
 block/blk-mq-tag.c |  4 +---
 block/blk-mq-tag.h | 17 ++++++++---------
 2 files changed, 9 insertions(+), 12 deletions(-)
block/blk-mq-tag.c +0 -0
index c1b9242..4953b64 100644
@@ -58,13 +58,11 @@ static inline void bt_index_atomic_inc(atomic_t *index)
 /*
  * If a previously inactive queue goes active, bump the active user count.

block/blk-mq-tag.c static inline void bt_index_atomic_inc(atomic_t *index) · line 66

Quoted 10 lines
block/blk-mq-tag.c +1 -3
  */
-bool __blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
+void __blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
 {
 	if (!test_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state) &&
 	    !test_and_set_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state))
 		atomic_inc(&hctx->tags->active_queues);
-
-	return true;
 }

This is obviously a good idea.

Quoted 14 lines
block/blk-mq-tag.h +1 -1
index 6206ed1..795ec3f 100644
@@ -66,23 +66,22 @@ enum {
 	BLK_MQ_TAG_MAX		= BLK_MQ_TAG_FAIL - 1,
 };
 
-extern bool __blk_mq_tag_busy(struct blk_mq_hw_ctx *);
+extern void __blk_mq_tag_busy(struct blk_mq_hw_ctx *);
 extern void __blk_mq_tag_idle(struct blk_mq_hw_ctx *);
 
 static inline bool blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
 {

block/blk-mq-tag.h enum { · line 79

Quoted 10 lines
block/blk-mq-tag.h +5 -4
-	if (!(hctx->flags & BLK_MQ_F_TAG_SHARED))
-		return false;
-
-	return __blk_mq_tag_busy(hctx);
+	if (hctx->flags & BLK_MQ_F_TAG_SHARED) {
+		__blk_mq_tag_busy(hctx);
+		return true;
+	}
+	return false;
 }

The normal/fast path here is the flag NOT being set, which is why it was coded that way to put the fast path inline.

block/blk-mq-tag.h enum { · line 85

Quoted 10 lines
block/blk-mq-tag.h +2 -4
 static inline void blk_mq_tag_idle(struct blk_mq_hw_ctx *hctx)
 {
-	if (!(hctx->flags & BLK_MQ_F_TAG_SHARED))
-		return;
-
-	__blk_mq_tag_idle(hctx);
+	if (hctx->flags & BLK_MQ_F_TAG_SHARED)
+		__blk_mq_tag_idle(hctx);
 }

Ditto

Signature · 1 line
Jens Axboe
#3 lore raw

Re: [PATCH 1/3] blk-mq: Cleanup blk_mq_tag_busy() and blk_mq_tag_idle()

Quoted 1 line · On Wed, 03 Sep 2014 14:35:29 -0600 Jens Axboe <[email protected]> wrote:

On 09/03/2014 02:33 PM, Alexander Gordeev wrote:

<snip>

Quoted 29 lines
Quoted 14 lines
block/blk-mq-tag.h +1 -1
index 6206ed1..795ec3f 100644
@@ -66,23 +66,22 @@ enum {
 	BLK_MQ_TAG_MAX		= BLK_MQ_TAG_FAIL - 1,
 };
 
-extern bool __blk_mq_tag_busy(struct blk_mq_hw_ctx *);
+extern void __blk_mq_tag_busy(struct blk_mq_hw_ctx *);
 extern void __blk_mq_tag_idle(struct blk_mq_hw_ctx *);
 
 static inline bool blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
 {

block/blk-mq-tag.h enum { · line 79

Quoted 10 lines
block/blk-mq-tag.h +5 -4
-	if (!(hctx->flags & BLK_MQ_F_TAG_SHARED))
-		return false;
-
-	return __blk_mq_tag_busy(hctx);
+	if (hctx->flags & BLK_MQ_F_TAG_SHARED) {
+		__blk_mq_tag_busy(hctx);
+		return true;
+	}
+	return false;
 }

The normal/fast path here is the flag NOT being set, which is why it was coded that way to put the fast path inline.

Quoted 1 line
Quoted 11 lines

block/blk-mq-tag.h enum { · line 85

Quoted 9 lines

static inline void blk_mq_tag_idle(struct blk_mq_hw_ctx *hctx) {

block/blk-mq-tag.h +2 -4
-	if (!(hctx->flags & BLK_MQ_F_TAG_SHARED))
-		return;
-
-	__blk_mq_tag_idle(hctx);
+	if (hctx->flags & BLK_MQ_F_TAG_SHARED)
+		__blk_mq_tag_idle(hctx);
 }

Ditto

Shouldn't it just add unlikely() then? That way it's obvious what the common case is, instead of relying on convoluted code.

maintainer
#4 lore raw

Re: [PATCH 1/3] blk-mq: Cleanup blk_mq_tag_busy() and blk_mq_tag_idle()

Quoted 46 lines · Chuck Ebbert
Quoted 1 line · On Wed, 03 Sep 2014 14:35:29 -0600 Jens Axboe <[email protected]> wrote:

On 09/03/2014 02:33 PM, Alexander Gordeev wrote:

<snip>

Quoted 29 lines
Quoted 14 lines
block/blk-mq-tag.h +1 -1
index 6206ed1..795ec3f 100644
@@ -66,23 +66,22 @@ enum {
 	BLK_MQ_TAG_MAX		= BLK_MQ_TAG_FAIL - 1,
 };
 
-extern bool __blk_mq_tag_busy(struct blk_mq_hw_ctx *);
+extern void __blk_mq_tag_busy(struct blk_mq_hw_ctx *);
 extern void __blk_mq_tag_idle(struct blk_mq_hw_ctx *);
 
 static inline bool blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
 {

block/blk-mq-tag.h enum { · line 79

Quoted 10 lines
block/blk-mq-tag.h +5 -4
-	if (!(hctx->flags & BLK_MQ_F_TAG_SHARED))
-		return false;
-
-	return __blk_mq_tag_busy(hctx);
+	if (hctx->flags & BLK_MQ_F_TAG_SHARED) {
+		__blk_mq_tag_busy(hctx);
+		return true;
+	}
+	return false;
 }

The normal/fast path here is the flag NOT being set, which is why it was coded that way to put the fast path inline.

Quoted 1 line
Quoted 11 lines

block/blk-mq-tag.h enum { · line 85

Quoted 9 lines

static inline void blk_mq_tag_idle(struct blk_mq_hw_ctx *hctx) {

block/blk-mq-tag.h +2 -4
-	if (!(hctx->flags & BLK_MQ_F_TAG_SHARED))
-		return;
-
-	__blk_mq_tag_idle(hctx);
+	if (hctx->flags & BLK_MQ_F_TAG_SHARED)
+		__blk_mq_tag_idle(hctx);
 }

Ditto

Quoted 2 lines

Shouldn't it just add unlikely() then? That way it's obvious what the common case is, instead of relying on convoluted code.

It's a common construct. Besides, if you find a flag-not-set check convoluted, then I hope you are not programming anything I use. That's a bit of a straw man, imho.

Signature · 1 line
Jens Axboe
#5 lore raw

Re: [PATCH 1/3] blk-mq: Cleanup blk_mq_tag_busy() and blk_mq_tag_idle()

Quoted 47 lines · On Thu, 04 Sep 2014 19:30:18 -0600 Jens Axboe <[email protected]> wrote:
Quoted 46 lines · Chuck Ebbert
Quoted 1 line · On Wed, 03 Sep 2014 14:35:29 -0600 Jens Axboe <[email protected]> wrote:

On 09/03/2014 02:33 PM, Alexander Gordeev wrote:

<snip>

Quoted 40 lines
Quoted 14 lines
block/blk-mq-tag.h +1 -1
index 6206ed1..795ec3f 100644
@@ -66,23 +66,22 @@ enum {
 	BLK_MQ_TAG_MAX		= BLK_MQ_TAG_FAIL - 1,
 };
 
-extern bool __blk_mq_tag_busy(struct blk_mq_hw_ctx *);
+extern void __blk_mq_tag_busy(struct blk_mq_hw_ctx *);
 extern void __blk_mq_tag_idle(struct blk_mq_hw_ctx *);
 
 static inline bool blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
 {

block/blk-mq-tag.h enum { · line 79

Quoted 10 lines
block/blk-mq-tag.h +5 -4
-	if (!(hctx->flags & BLK_MQ_F_TAG_SHARED))
-		return false;
-
-	return __blk_mq_tag_busy(hctx);
+	if (hctx->flags & BLK_MQ_F_TAG_SHARED) {
+		__blk_mq_tag_busy(hctx);
+		return true;
+	}
+	return false;
 }

The normal/fast path here is the flag NOT being set, which is why it was coded that way to put the fast path inline.

block/blk-mq-tag.h enum { · line 85

Quoted 10 lines
block/blk-mq-tag.h +2 -4
 static inline void blk_mq_tag_idle(struct blk_mq_hw_ctx *hctx)
 {
-	if (!(hctx->flags & BLK_MQ_F_TAG_SHARED))
-		return;
-
-	__blk_mq_tag_idle(hctx);
+	if (hctx->flags & BLK_MQ_F_TAG_SHARED)
+		__blk_mq_tag_idle(hctx);
 }

Ditto

Quoted 7 lines
Quoted 2 lines

Shouldn't it just add unlikely() then? That way it's obvious what the common case is, instead of relying on convoluted code.

It's a common construct. Besides, if you find a flag-not-set check convoluted, then I hope you are not programming anything I use. That's a bit of a straw man, imho.

Sure, it's a common construct. But there's nothing there to prevent the optimizer from rearranging things any way it pleases. Nor is there anything keeping a human from doing the same. ;)

#6 lore raw

PATCH 2/3 blk-mq: Eliminate superfluous check of BLK_MQ_S_TAG_ACTIVE flag

In reply to Alexander Gordeev
---
 block/blk-mq-tag.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
block/blk-mq-tag.c +1 -2
index 4953b64..d1eb579 100644
@@ -60,8 +60,7 @@ static inline void bt_index_atomic_inc(atomic_t *index)
  */
 void __blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
 {
-	if (!test_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state) &&
-	    !test_and_set_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state))
+	if (!test_and_set_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state))
 		atomic_inc(&hctx->tags->active_queues);
 }
 
Signature · 1 line
1.9.3
maintainer
#7 lore raw

Re: [PATCH 2/3] blk-mq: Eliminate superfluous check of BLK_MQ_S_TAG_ACTIVE flag

Quoted 11 lines · Alexander Gordeev
---
 block/blk-mq-tag.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
block/blk-mq-tag.c +0 -0
index 4953b64..d1eb579 100644

block/blk-mq-tag.c static inline void bt_index_atomic_inc(atomic_t *index) · line 64

Quoted 7 lines
block/blk-mq-tag.c +1 -2
@@ -60,8 +60,7 @@ static inline void bt_index_atomic_inc(atomic_t *index)
  */
 void __blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
 {
-	if (!test_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state) &&
-	    !test_and_set_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state))
+	if (!test_and_set_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state))
 		atomic_inc(&hctx->tags->active_queues);

That's done on purpose as well, it's a lot faster to do the (non-atomic) check first and only do the expensive test and set if needed.

See 7fcbbaf18392 for similar construct and actual numbers on a real case.

Signature · 1 line
Jens Axboe
#8 lore raw

PATCH 3/3 blk-mq: Fix formula to calculate fair share of tags

In reply to Alexander Gordeev

Fair share of tags is number of tags diveded on number of users. Not sure why it is different.

---
 block/blk-mq-tag.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
block/blk-mq-tag.c +1 -1
index d1eb579..1b9c949 100644
@@ -127,7 +127,7 @@ static inline bool hctx_may_queue(struct blk_mq_hw_ctx *hctx,
 	/*
 	 * Allow at least some tags
 	 */
-	depth = max((bt->depth + users - 1) / users, 4U);
+	depth = max(bt->depth / users, 4U);
 	return atomic_read(&hctx->nr_active) < depth;
 }
 
Signature · 1 line
1.9.3
maintainer
#9 lore raw

Re: [PATCH 3/3] blk-mq: Fix formula to calculate fair share of tags

Quoted 14 lines · Alexander Gordeev

Fair share of tags is number of tags diveded on number of users. Not sure why it is different.

---
 block/blk-mq-tag.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
block/blk-mq-tag.c +0 -0
index d1eb579..1b9c949 100644

block/blk-mq-tag.c static inline bool hctx_may_queue(struct blk_mq_hw_ctx *hctx, · line 131

Quoted 6 lines
block/blk-mq-tag.c +1 -1
@@ -127,7 +127,7 @@ static inline bool hctx_may_queue(struct blk_mq_hw_ctx *hctx,
 	/*
 	 * Allow at least some tags
 	 */
-	depth = max((bt->depth + users - 1) / users, 4U);
+	depth = max(bt->depth / users, 4U);
 	return atomic_read(&hctx->nr_active) < depth;

It's normal rounding, you'll round down otherwise. Say you have a tag depth of 31 (SATA), and 4 active users. Your change would make that 7 tags per user, leaving 3 idle. If you round up, you end up with 8 tags instead. That will potentially just leave some sleeping for a new tag, but at least you'll exhaust the space.

Signature · 1 line
Jens Axboe