blk-cgroup cleanups

blk-cgroup cleanups

[Christoph Hellwig]

Hi Tejun and Jens,

this series has a bunch of blk-cgroup cleanups and preparation
for preparing to make blk-cgroup gendisk based.  Another series
for the next merge window will follow for the real changes that
include refcounting updates.

Diffstat:
 block/blk-cgroup.c         |  186 +++++++++++++++++----------------------------
 block/blk-cgroup.h         |   68 ++++------------
 block/blk-iocost.c         |   37 ++++----
 block/blk-iolatency.c      |    5 -
 block/blk-ioprio.c         |    8 -
 block/blk-ioprio.h         |    8 -
 block/blk-sysfs.c          |    2 
 block/blk-throttle.c       |   13 ++-
 block/blk-throttle.h       |   16 +--
 block/blk.h                |    4 
 block/genhd.c              |    7 -
 include/linux/blk-cgroup.h |    5 -
 mm/swapfile.c              |    2 
 13 files changed, 148 insertions(+), 213 deletions(-)

[PATCH 01/17] blk-cgroup: fix error unwinding in blkcg_init_queue

[Christoph Hellwig]

When blk_throtl_init fails, we need to call blk_ioprio_exit.  Switch to
proper goto based unwinding to fix this.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 869af9d72bcf8..3a88f8c011d27 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -1297,17 +1297,18 @@ int blkcg_init_queue(struct request_queue *q)
 
 	ret = blk_throtl_init(q);
 	if (ret)
-		goto err_destroy_all;
+		goto err_ioprio_exit;
 
 	ret = blk_iolatency_init(q);
-	if (ret) {
-		blk_throtl_exit(q);
-		blk_ioprio_exit(q);
-		goto err_destroy_all;
-	}
+	if (ret)
+		goto err_throtl_exit;
 
 	return 0;
 
+err_throtl_exit:
+	blk_throtl_exit(q);
+err_ioprio_exit:
+	blk_ioprio_exit(q);
 err_destroy_all:
 	blkg_destroy_all(q);
 	return ret;
-- 
2.30.2

Re: [PATCH 01/17] blk-cgroup: fix error unwinding in blkcg_init_queue

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:45PM +0200, Christoph Hellwig wrote:
> When blk_throtl_init fails, we need to call blk_ioprio_exit.  Switch to
> proper goto based unwinding to fix this.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c | 13 +++++++------
>  1 file changed, 7 insertions(+), 6 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index 869af9d72bcf8..3a88f8c011d27 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -1297,17 +1297,18 @@ int blkcg_init_queue(struct request_queue *q)
>  
>  	ret = blk_throtl_init(q);
>  	if (ret)
> -		goto err_destroy_all;
> +		goto err_ioprio_exit;
>  
>  	ret = blk_iolatency_init(q);
> -	if (ret) {
> -		blk_throtl_exit(q);
> -		blk_ioprio_exit(q);
> -		goto err_destroy_all;
> -	}
> +	if (ret)
> +		goto err_throtl_exit;
>  
>  	return 0;
>  
> +err_throtl_exit:
> +	blk_throtl_exit(q);
> +err_ioprio_exit:
> +	blk_ioprio_exit(q);
>  err_destroy_all:
>  	blkg_destroy_all(q);
>  	return ret;
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 02/17] blk-cgroup: remove blk_queue_root_blkg

[Christoph Hellwig]

Just open code it in the only caller and drop the unused !BLK_CGROUP
stub.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c |  3 +--
 block/blk-cgroup.h | 13 -------------
 2 files changed, 1 insertion(+), 15 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 3a88f8c011d27..4180de4cbb3e1 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -915,8 +915,7 @@ static void blkcg_fill_root_iostats(void)
 	class_dev_iter_init(&iter, &block_class, NULL, &disk_type);
 	while ((dev = class_dev_iter_next(&iter))) {
 		struct block_device *bdev = dev_to_bdev(dev);
-		struct blkcg_gq *blkg =
-			blk_queue_root_blkg(bdev_get_queue(bdev));
+		struct blkcg_gq *blkg = bdev->bd_disk->queue->root_blkg;
 		struct blkg_iostat tmp;
 		int cpu;
 		unsigned long flags;
diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
index d2724d1dd7c9b..c1fb00a1dfc03 100644
--- a/block/blk-cgroup.h
+++ b/block/blk-cgroup.h
@@ -268,17 +268,6 @@ static inline struct blkcg_gq *blkg_lookup(struct blkcg *blkcg,
 	return __blkg_lookup(blkcg, q, false);
 }
 
-/**
- * blk_queue_root_blkg - return blkg for the (blkcg_root, @q) pair
- * @q: request_queue of interest
- *
- * Lookup blkg for @q at the root level. See also blkg_lookup().
- */
-static inline struct blkcg_gq *blk_queue_root_blkg(struct request_queue *q)
-{
-	return q->root_blkg;
-}
-
 /**
  * blkg_to_pdata - get policy private data
  * @blkg: blkg of interest
@@ -507,8 +496,6 @@ struct blkcg {
 };
 
 static inline struct blkcg_gq *blkg_lookup(struct blkcg *blkcg, void *key) { return NULL; }
-static inline struct blkcg_gq *blk_queue_root_blkg(struct request_queue *q)
-{ return NULL; }
 static inline int blkcg_init_queue(struct request_queue *q) { return 0; }
 static inline void blkcg_exit_queue(struct request_queue *q) { }
 static inline int blkcg_policy_register(struct blkcg_policy *pol) { return 0; }
-- 
2.30.2

Re: [PATCH 02/17] blk-cgroup: remove blk_queue_root_blkg

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:46PM +0200, Christoph Hellwig wrote:
> Just open code it in the only caller and drop the unused !BLK_CGROUP
> stub.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c |  3 +--
>  block/blk-cgroup.h | 13 -------------
>  2 files changed, 1 insertion(+), 15 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index 3a88f8c011d27..4180de4cbb3e1 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -915,8 +915,7 @@ static void blkcg_fill_root_iostats(void)
>  	class_dev_iter_init(&iter, &block_class, NULL, &disk_type);
>  	while ((dev = class_dev_iter_next(&iter))) {
>  		struct block_device *bdev = dev_to_bdev(dev);
> -		struct blkcg_gq *blkg =
> -			blk_queue_root_blkg(bdev_get_queue(bdev));
> +		struct blkcg_gq *blkg = bdev->bd_disk->queue->root_blkg;
>  		struct blkg_iostat tmp;
>  		int cpu;
>  		unsigned long flags;
> diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
> index d2724d1dd7c9b..c1fb00a1dfc03 100644
> --- a/block/blk-cgroup.h
> +++ b/block/blk-cgroup.h
> @@ -268,17 +268,6 @@ static inline struct blkcg_gq *blkg_lookup(struct blkcg *blkcg,
>  	return __blkg_lookup(blkcg, q, false);
>  }
>  
> -/**
> - * blk_queue_root_blkg - return blkg for the (blkcg_root, @q) pair
> - * @q: request_queue of interest
> - *
> - * Lookup blkg for @q at the root level. See also blkg_lookup().
> - */
> -static inline struct blkcg_gq *blk_queue_root_blkg(struct request_queue *q)
> -{
> -	return q->root_blkg;
> -}
> -
>  /**
>   * blkg_to_pdata - get policy private data
>   * @blkg: blkg of interest
> @@ -507,8 +496,6 @@ struct blkcg {
>  };
>  
>  static inline struct blkcg_gq *blkg_lookup(struct blkcg *blkcg, void *key) { return NULL; }
> -static inline struct blkcg_gq *blk_queue_root_blkg(struct request_queue *q)
> -{ return NULL; }
>  static inline int blkcg_init_queue(struct request_queue *q) { return 0; }
>  static inline void blkcg_exit_queue(struct request_queue *q) { }
>  static inline int blkcg_policy_register(struct blkcg_policy *pol) { return 0; }
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 03/17] blk-cgroup: remove open coded blkg_lookup instances

[Christoph Hellwig]

Use blkg_lookup instead of open coding it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c | 6 +++---
 block/blk-cgroup.h | 8 ++++----
 2 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 4180de4cbb3e1..b9a1dcee5a244 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -324,7 +324,7 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
 
 	/* link parent */
 	if (blkcg_parent(blkcg)) {
-		blkg->parent = __blkg_lookup(blkcg_parent(blkcg), q, false);
+		blkg->parent = blkg_lookup(blkcg_parent(blkcg), q);
 		if (WARN_ON_ONCE(!blkg->parent)) {
 			ret = -ENODEV;
 			goto err_put_css;
@@ -412,7 +412,7 @@ static struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
 		struct blkcg_gq *ret_blkg = q->root_blkg;
 
 		while (parent) {
-			blkg = __blkg_lookup(parent, q, false);
+			blkg = blkg_lookup(parent, q);
 			if (blkg) {
 				/* remember closest blkg */
 				ret_blkg = blkg;
@@ -724,7 +724,7 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
 		struct blkcg_gq *new_blkg;
 
 		parent = blkcg_parent(blkcg);
-		while (parent && !__blkg_lookup(parent, q, false)) {
+		while (parent && !blkg_lookup(parent, q)) {
 			pos = parent;
 			parent = blkcg_parent(parent);
 		}
diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
index c1fb00a1dfc03..30396cad50e9a 100644
--- a/block/blk-cgroup.h
+++ b/block/blk-cgroup.h
@@ -362,8 +362,8 @@ static inline void blkg_put(struct blkcg_gq *blkg)
  */
 #define blkg_for_each_descendant_pre(d_blkg, pos_css, p_blkg)		\
 	css_for_each_descendant_pre((pos_css), &(p_blkg)->blkcg->css)	\
-		if (((d_blkg) = __blkg_lookup(css_to_blkcg(pos_css),	\
-					      (p_blkg)->q, false)))
+		if (((d_blkg) = blkg_lookup(css_to_blkcg(pos_css),	\
+					    (p_blkg)->q)))
 
 /**
  * blkg_for_each_descendant_post - post-order walk of a blkg's descendants
@@ -377,8 +377,8 @@ static inline void blkg_put(struct blkcg_gq *blkg)
  */
 #define blkg_for_each_descendant_post(d_blkg, pos_css, p_blkg)		\
 	css_for_each_descendant_post((pos_css), &(p_blkg)->blkcg->css)	\
-		if (((d_blkg) = __blkg_lookup(css_to_blkcg(pos_css),	\
-					      (p_blkg)->q, false)))
+		if (((d_blkg) = blkg_lookup(css_to_blkcg(pos_css),	\
+					    (p_blkg)->q)))
 
 bool __blkcg_punt_bio_submit(struct bio *bio);
 
-- 
2.30.2

Re: [PATCH 03/17] blk-cgroup: remove open coded blkg_lookup instances

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:47PM +0200, Christoph Hellwig wrote:
> Use blkg_lookup instead of open coding it.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c | 6 +++---
>  block/blk-cgroup.h | 8 ++++----
>  2 files changed, 7 insertions(+), 7 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index 4180de4cbb3e1..b9a1dcee5a244 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -324,7 +324,7 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
>  
>  	/* link parent */
>  	if (blkcg_parent(blkcg)) {
> -		blkg->parent = __blkg_lookup(blkcg_parent(blkcg), q, false);
> +		blkg->parent = blkg_lookup(blkcg_parent(blkcg), q);
>  		if (WARN_ON_ONCE(!blkg->parent)) {
>  			ret = -ENODEV;
>  			goto err_put_css;
> @@ -412,7 +412,7 @@ static struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
>  		struct blkcg_gq *ret_blkg = q->root_blkg;
>  
>  		while (parent) {
> -			blkg = __blkg_lookup(parent, q, false);
> +			blkg = blkg_lookup(parent, q);
>  			if (blkg) {
>  				/* remember closest blkg */
>  				ret_blkg = blkg;
> @@ -724,7 +724,7 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
>  		struct blkcg_gq *new_blkg;
>  
>  		parent = blkcg_parent(blkcg);
> -		while (parent && !__blkg_lookup(parent, q, false)) {
> +		while (parent && !blkg_lookup(parent, q)) {
>  			pos = parent;
>  			parent = blkcg_parent(parent);
>  		}
> diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
> index c1fb00a1dfc03..30396cad50e9a 100644
> --- a/block/blk-cgroup.h
> +++ b/block/blk-cgroup.h
> @@ -362,8 +362,8 @@ static inline void blkg_put(struct blkcg_gq *blkg)
>   */
>  #define blkg_for_each_descendant_pre(d_blkg, pos_css, p_blkg)		\
>  	css_for_each_descendant_pre((pos_css), &(p_blkg)->blkcg->css)	\
> -		if (((d_blkg) = __blkg_lookup(css_to_blkcg(pos_css),	\
> -					      (p_blkg)->q, false)))
> +		if (((d_blkg) = blkg_lookup(css_to_blkcg(pos_css),	\
> +					    (p_blkg)->q)))
>  
>  /**
>   * blkg_for_each_descendant_post - post-order walk of a blkg's descendants
> @@ -377,8 +377,8 @@ static inline void blkg_put(struct blkcg_gq *blkg)
>   */
>  #define blkg_for_each_descendant_post(d_blkg, pos_css, p_blkg)		\
>  	css_for_each_descendant_post((pos_css), &(p_blkg)->blkcg->css)	\
> -		if (((d_blkg) = __blkg_lookup(css_to_blkcg(pos_css),	\
> -					      (p_blkg)->q, false)))
> +		if (((d_blkg) = blkg_lookup(css_to_blkcg(pos_css),	\
> +					    (p_blkg)->q)))
>  
>  bool __blkcg_punt_bio_submit(struct bio *bio);
>  
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

Re: [PATCH 03/17] blk-cgroup: remove open coded blkg_lookup instances

[Tejun Heo]

On Wed, Sep 21, 2022 at 08:04:47PM +0200, Christoph Hellwig wrote:
> Use blkg_lookup instead of open coding it.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

For patches 1-3:

 Acked-by: Tejun Heo <tj@kernel.org>

Thanks.

-- 
tejun

[PATCH 04/17] blk-cgroup: cleanup the blkg_lookup family of functions

[Christoph Hellwig]

Add a fully inlined blkg_lookup as the extra two checks aren't going
to generated a lot more code vs the call to the slowpath routine, and
open code the hint update in the two callers that care.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c | 38 +++++++++++++++-----------------------
 block/blk-cgroup.h | 39 ++++++++++++---------------------------
 2 files changed, 27 insertions(+), 50 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index b9a1dcee5a244..d1216760d0255 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -263,29 +263,13 @@ static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct request_queue *q,
 	return NULL;
 }
 
-struct blkcg_gq *blkg_lookup_slowpath(struct blkcg *blkcg,
-				      struct request_queue *q, bool update_hint)
+static void blkg_update_hint(struct blkcg *blkcg, struct blkcg_gq *blkg)
 {
-	struct blkcg_gq *blkg;
-
-	/*
-	 * Hint didn't match.  Look up from the radix tree.  Note that the
-	 * hint can only be updated under queue_lock as otherwise @blkg
-	 * could have already been removed from blkg_tree.  The caller is
-	 * responsible for grabbing queue_lock if @update_hint.
-	 */
-	blkg = radix_tree_lookup(&blkcg->blkg_tree, q->id);
-	if (blkg && blkg->q == q) {
-		if (update_hint) {
-			lockdep_assert_held(&q->queue_lock);
-			rcu_assign_pointer(blkcg->blkg_hint, blkg);
-		}
-		return blkg;
-	}
+	lockdep_assert_held(&blkg->q->queue_lock);
 
-	return NULL;
+	if (blkcg != &blkcg_root && blkg != rcu_dereference(blkcg->blkg_hint))
+		rcu_assign_pointer(blkcg->blkg_hint, blkg);
 }
-EXPORT_SYMBOL_GPL(blkg_lookup_slowpath);
 
 /*
  * If @new_blkg is %NULL, this function tries to allocate a new one as
@@ -397,9 +381,11 @@ static struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
 		return blkg;
 
 	spin_lock_irqsave(&q->queue_lock, flags);
-	blkg = __blkg_lookup(blkcg, q, true);
-	if (blkg)
+	blkg = blkg_lookup(blkcg, q);
+	if (blkg) {
+		blkg_update_hint(blkcg, blkg);
 		goto found;
+	}
 
 	/*
 	 * Create blkgs walking down from blkcg_root to @blkcg, so that all
@@ -621,12 +607,18 @@ static struct blkcg_gq *blkg_lookup_check(struct blkcg *blkcg,
 					  const struct blkcg_policy *pol,
 					  struct request_queue *q)
 {
+	struct blkcg_gq *blkg;
+
 	WARN_ON_ONCE(!rcu_read_lock_held());
 	lockdep_assert_held(&q->queue_lock);
 
 	if (!blkcg_policy_enabled(q, pol))
 		return ERR_PTR(-EOPNOTSUPP);
-	return __blkg_lookup(blkcg, q, true /* update_hint */);
+
+	blkg = blkg_lookup(blkcg, q);
+	if (blkg)
+		blkg_update_hint(blkcg, blkg);
+	return blkg;
 }
 
 /**
diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
index 30396cad50e9a..91b7ae0773be6 100644
--- a/block/blk-cgroup.h
+++ b/block/blk-cgroup.h
@@ -178,8 +178,6 @@ struct blkcg_policy {
 extern struct blkcg blkcg_root;
 extern bool blkcg_debug_stats;
 
-struct blkcg_gq *blkg_lookup_slowpath(struct blkcg *blkcg,
-				      struct request_queue *q, bool update_hint);
 int blkcg_init_queue(struct request_queue *q);
 void blkcg_exit_queue(struct request_queue *q);
 
@@ -227,22 +225,21 @@ static inline bool bio_issue_as_root_blkg(struct bio *bio)
 }
 
 /**
- * __blkg_lookup - internal version of blkg_lookup()
+ * blkg_lookup - lookup blkg for the specified blkcg - q pair
  * @blkcg: blkcg of interest
  * @q: request_queue of interest
- * @update_hint: whether to update lookup hint with the result or not
  *
- * This is internal version and shouldn't be used by policy
- * implementations.  Looks up blkgs for the @blkcg - @q pair regardless of
- * @q's bypass state.  If @update_hint is %true, the caller should be
- * holding @q->queue_lock and lookup hint is updated on success.
+ * Lookup blkg for the @blkcg - @q pair.
+
+ * Must be called in a RCU critical section.
  */
-static inline struct blkcg_gq *__blkg_lookup(struct blkcg *blkcg,
-					     struct request_queue *q,
-					     bool update_hint)
+static inline struct blkcg_gq *blkg_lookup(struct blkcg *blkcg,
+					   struct request_queue *q)
 {
 	struct blkcg_gq *blkg;
 
+	WARN_ON_ONCE(!rcu_read_lock_held());
+
 	if (blkcg == &blkcg_root)
 		return q->root_blkg;
 
@@ -250,22 +247,10 @@ static inline struct blkcg_gq *__blkg_lookup(struct blkcg *blkcg,
 	if (blkg && blkg->q == q)
 		return blkg;
 
-	return blkg_lookup_slowpath(blkcg, q, update_hint);
-}
-
-/**
- * blkg_lookup - lookup blkg for the specified blkcg - q pair
- * @blkcg: blkcg of interest
- * @q: request_queue of interest
- *
- * Lookup blkg for the @blkcg - @q pair.  This function should be called
- * under RCU read lock.
- */
-static inline struct blkcg_gq *blkg_lookup(struct blkcg *blkcg,
-					   struct request_queue *q)
-{
-	WARN_ON_ONCE(!rcu_read_lock_held());
-	return __blkg_lookup(blkcg, q, false);
+	blkg = radix_tree_lookup(&blkcg->blkg_tree, q->id);
+	if (blkg && blkg->q != q)
+		blkg = NULL;
+	return blkg;
 }
 
 /**
-- 
2.30.2

Re: [PATCH 04/17] blk-cgroup: cleanup the blkg_lookup family of functions

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:48PM +0200, Christoph Hellwig wrote:
> Add a fully inlined blkg_lookup as the extra two checks aren't going
> to generated a lot more code vs the call to the slowpath routine, and

s/generated/generate/

> open code the hint update in the two callers that care.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c | 38 +++++++++++++++-----------------------
>  block/blk-cgroup.h | 39 ++++++++++++---------------------------
>  2 files changed, 27 insertions(+), 50 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index b9a1dcee5a244..d1216760d0255 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -263,29 +263,13 @@ static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct request_queue *q,
>  	return NULL;
>  }
>  
> -struct blkcg_gq *blkg_lookup_slowpath(struct blkcg *blkcg,
> -				      struct request_queue *q, bool update_hint)
> +static void blkg_update_hint(struct blkcg *blkcg, struct blkcg_gq *blkg)
>  {
> -	struct blkcg_gq *blkg;
> -
> -	/*
> -	 * Hint didn't match.  Look up from the radix tree.  Note that the
> -	 * hint can only be updated under queue_lock as otherwise @blkg
> -	 * could have already been removed from blkg_tree.  The caller is
> -	 * responsible for grabbing queue_lock if @update_hint.
> -	 */
> -	blkg = radix_tree_lookup(&blkcg->blkg_tree, q->id);
> -	if (blkg && blkg->q == q) {
> -		if (update_hint) {
> -			lockdep_assert_held(&q->queue_lock);
> -			rcu_assign_pointer(blkcg->blkg_hint, blkg);
> -		}
> -		return blkg;
> -	}
> +	lockdep_assert_held(&blkg->q->queue_lock);
>  
> -	return NULL;
> +	if (blkcg != &blkcg_root && blkg != rcu_dereference(blkcg->blkg_hint))
> +		rcu_assign_pointer(blkcg->blkg_hint, blkg);
>  }
> -EXPORT_SYMBOL_GPL(blkg_lookup_slowpath);
>  
>  /*
>   * If @new_blkg is %NULL, this function tries to allocate a new one as
> @@ -397,9 +381,11 @@ static struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
>  		return blkg;
>  
>  	spin_lock_irqsave(&q->queue_lock, flags);
> -	blkg = __blkg_lookup(blkcg, q, true);
> -	if (blkg)
> +	blkg = blkg_lookup(blkcg, q);
> +	if (blkg) {
> +		blkg_update_hint(blkcg, blkg);
>  		goto found;
> +	}
>  
>  	/*
>  	 * Create blkgs walking down from blkcg_root to @blkcg, so that all
> @@ -621,12 +607,18 @@ static struct blkcg_gq *blkg_lookup_check(struct blkcg *blkcg,
>  					  const struct blkcg_policy *pol,
>  					  struct request_queue *q)
>  {
> +	struct blkcg_gq *blkg;
> +
>  	WARN_ON_ONCE(!rcu_read_lock_held());
>  	lockdep_assert_held(&q->queue_lock);
>  
>  	if (!blkcg_policy_enabled(q, pol))
>  		return ERR_PTR(-EOPNOTSUPP);
> -	return __blkg_lookup(blkcg, q, true /* update_hint */);
> +
> +	blkg = blkg_lookup(blkcg, q);
> +	if (blkg)
> +		blkg_update_hint(blkcg, blkg);
> +	return blkg;
>  }
>  
>  /**
> diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
> index 30396cad50e9a..91b7ae0773be6 100644
> --- a/block/blk-cgroup.h
> +++ b/block/blk-cgroup.h
> @@ -178,8 +178,6 @@ struct blkcg_policy {
>  extern struct blkcg blkcg_root;
>  extern bool blkcg_debug_stats;
>  
> -struct blkcg_gq *blkg_lookup_slowpath(struct blkcg *blkcg,
> -				      struct request_queue *q, bool update_hint);
>  int blkcg_init_queue(struct request_queue *q);
>  void blkcg_exit_queue(struct request_queue *q);
>  
> @@ -227,22 +225,21 @@ static inline bool bio_issue_as_root_blkg(struct bio *bio)
>  }
>  
>  /**
> - * __blkg_lookup - internal version of blkg_lookup()
> + * blkg_lookup - lookup blkg for the specified blkcg - q pair
>   * @blkcg: blkcg of interest
>   * @q: request_queue of interest
> - * @update_hint: whether to update lookup hint with the result or not
>   *
> - * This is internal version and shouldn't be used by policy
> - * implementations.  Looks up blkgs for the @blkcg - @q pair regardless of
> - * @q's bypass state.  If @update_hint is %true, the caller should be
> - * holding @q->queue_lock and lookup hint is updated on success.
> + * Lookup blkg for the @blkcg - @q pair.
> +
> + * Must be called in a RCU critical section.
>   */
> -static inline struct blkcg_gq *__blkg_lookup(struct blkcg *blkcg,
> -					     struct request_queue *q,
> -					     bool update_hint)
> +static inline struct blkcg_gq *blkg_lookup(struct blkcg *blkcg,
> +					   struct request_queue *q)
>  {
>  	struct blkcg_gq *blkg;
>  
> +	WARN_ON_ONCE(!rcu_read_lock_held());
> +
>  	if (blkcg == &blkcg_root)
>  		return q->root_blkg;
>  
> @@ -250,22 +247,10 @@ static inline struct blkcg_gq *__blkg_lookup(struct blkcg *blkcg,
>  	if (blkg && blkg->q == q)
>  		return blkg;
>  
> -	return blkg_lookup_slowpath(blkcg, q, update_hint);
> -}
> -
> -/**
> - * blkg_lookup - lookup blkg for the specified blkcg - q pair
> - * @blkcg: blkcg of interest
> - * @q: request_queue of interest
> - *
> - * Lookup blkg for the @blkcg - @q pair.  This function should be called
> - * under RCU read lock.
> - */
> -static inline struct blkcg_gq *blkg_lookup(struct blkcg *blkcg,
> -					   struct request_queue *q)
> -{
> -	WARN_ON_ONCE(!rcu_read_lock_held());
> -	return __blkg_lookup(blkcg, q, false);
> +	blkg = radix_tree_lookup(&blkcg->blkg_tree, q->id);
> +	if (blkg && blkg->q != q)
> +		blkg = NULL;
> +	return blkg;
>  }
>  
>  /**
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

Re: [PATCH 04/17] blk-cgroup: cleanup the blkg_lookup family of functions

[Tejun Heo]

On Wed, Sep 21, 2022 at 08:04:48PM +0200, Christoph Hellwig wrote:
> Add a fully inlined blkg_lookup as the extra two checks aren't going
> to generated a lot more code vs the call to the slowpath routine, and
> open code the hint update in the two callers that care.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Tejun Heo <tj@kernel.org>

Thanks.

-- 
tejun

[PATCH 05/17] blk-cgroup: remove blkg_lookup_check

[Christoph Hellwig]

The combinations of an error check with an ERR_PTR return and a lookup
with a NULL return leads to ugly handling of the return values in the
callers.  Just open coding the check and the lookup is much simpler.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c | 36 ++++++++++--------------------------
 1 file changed, 10 insertions(+), 26 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index d1216760d0255..1306112d76486 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -602,25 +602,6 @@ u64 __blkg_prfill_u64(struct seq_file *sf, struct blkg_policy_data *pd, u64 v)
 }
 EXPORT_SYMBOL_GPL(__blkg_prfill_u64);
 
-/* Performs queue bypass and policy enabled checks then looks up blkg. */
-static struct blkcg_gq *blkg_lookup_check(struct blkcg *blkcg,
-					  const struct blkcg_policy *pol,
-					  struct request_queue *q)
-{
-	struct blkcg_gq *blkg;
-
-	WARN_ON_ONCE(!rcu_read_lock_held());
-	lockdep_assert_held(&q->queue_lock);
-
-	if (!blkcg_policy_enabled(q, pol))
-		return ERR_PTR(-EOPNOTSUPP);
-
-	blkg = blkg_lookup(blkcg, q);
-	if (blkg)
-		blkg_update_hint(blkcg, blkg);
-	return blkg;
-}
-
 /**
  * blkcg_conf_open_bdev - parse and open bdev for per-blkg config update
  * @inputp: input string pointer
@@ -697,14 +678,16 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
 	rcu_read_lock();
 	spin_lock_irq(&q->queue_lock);
 
-	blkg = blkg_lookup_check(blkcg, pol, q);
-	if (IS_ERR(blkg)) {
-		ret = PTR_ERR(blkg);
+	if (!blkcg_policy_enabled(q, pol)) {
+		ret = -EOPNOTSUPP;
 		goto fail_unlock;
 	}
 
-	if (blkg)
+	blkg = blkg_lookup(blkcg, q);
+	if (blkg) {
+		blkg_update_hint(blkcg, blkg);
 		goto success;
+	}
 
 	/*
 	 * Create blkgs walking down from blkcg_root to @blkcg, so that all
@@ -740,14 +723,15 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
 		rcu_read_lock();
 		spin_lock_irq(&q->queue_lock);
 
-		blkg = blkg_lookup_check(pos, pol, q);
-		if (IS_ERR(blkg)) {
-			ret = PTR_ERR(blkg);
+		if (!blkcg_policy_enabled(q, pol)) {
 			blkg_free(new_blkg);
+			ret = -EOPNOTSUPP;
 			goto fail_preloaded;
 		}
 
+		blkg = blkg_lookup(pos, q);
 		if (blkg) {
+			blkg_update_hint(pos, blkg);
 			blkg_free(new_blkg);
 		} else {
 			blkg = blkg_create(pos, q, new_blkg);
-- 
2.30.2

Re: [PATCH 05/17] blk-cgroup: remove blkg_lookup_check

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:49PM +0200, Christoph Hellwig wrote:
> The combinations of an error check with an ERR_PTR return and a lookup
> with a NULL return leads to ugly handling of the return values in the
> callers.  Just open coding the check and the lookup is much simpler.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c | 36 ++++++++++--------------------------
>  1 file changed, 10 insertions(+), 26 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index d1216760d0255..1306112d76486 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -602,25 +602,6 @@ u64 __blkg_prfill_u64(struct seq_file *sf, struct blkg_policy_data *pd, u64 v)
>  }
>  EXPORT_SYMBOL_GPL(__blkg_prfill_u64);
>  
> -/* Performs queue bypass and policy enabled checks then looks up blkg. */
> -static struct blkcg_gq *blkg_lookup_check(struct blkcg *blkcg,
> -					  const struct blkcg_policy *pol,
> -					  struct request_queue *q)
> -{
> -	struct blkcg_gq *blkg;
> -
> -	WARN_ON_ONCE(!rcu_read_lock_held());
> -	lockdep_assert_held(&q->queue_lock);
> -
> -	if (!blkcg_policy_enabled(q, pol))
> -		return ERR_PTR(-EOPNOTSUPP);
> -
> -	blkg = blkg_lookup(blkcg, q);
> -	if (blkg)
> -		blkg_update_hint(blkcg, blkg);
> -	return blkg;
> -}
> -
>  /**
>   * blkcg_conf_open_bdev - parse and open bdev for per-blkg config update
>   * @inputp: input string pointer
> @@ -697,14 +678,16 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
>  	rcu_read_lock();
>  	spin_lock_irq(&q->queue_lock);
>  
> -	blkg = blkg_lookup_check(blkcg, pol, q);
> -	if (IS_ERR(blkg)) {
> -		ret = PTR_ERR(blkg);
> +	if (!blkcg_policy_enabled(q, pol)) {
> +		ret = -EOPNOTSUPP;
>  		goto fail_unlock;
>  	}
>  
> -	if (blkg)
> +	blkg = blkg_lookup(blkcg, q);
> +	if (blkg) {
> +		blkg_update_hint(blkcg, blkg);
>  		goto success;
> +	}
>  
>  	/*
>  	 * Create blkgs walking down from blkcg_root to @blkcg, so that all
> @@ -740,14 +723,15 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
>  		rcu_read_lock();
>  		spin_lock_irq(&q->queue_lock);
>  
> -		blkg = blkg_lookup_check(pos, pol, q);
> -		if (IS_ERR(blkg)) {
> -			ret = PTR_ERR(blkg);
> +		if (!blkcg_policy_enabled(q, pol)) {
>  			blkg_free(new_blkg);
> +			ret = -EOPNOTSUPP;
>  			goto fail_preloaded;
>  		}
>  
> +		blkg = blkg_lookup(pos, q);
>  		if (blkg) {
> +			blkg_update_hint(pos, blkg);
>  			blkg_free(new_blkg);
>  		} else {
>  			blkg = blkg_create(pos, q, new_blkg);
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

Re: [PATCH 05/17] blk-cgroup: remove blkg_lookup_check

[Tejun Heo]

Hello,

On Wed, Sep 21, 2022 at 08:04:49PM +0200, Christoph Hellwig wrote:
> The combinations of an error check with an ERR_PTR return and a lookup
> with a NULL return leads to ugly handling of the return values in the
> callers.  Just open coding the check and the lookup is much simpler.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Tejun Heo <tj@kernel.org>

but please look below.

>  /**
>   * blkcg_conf_open_bdev - parse and open bdev for per-blkg config update
>   * @inputp: input string pointer
> @@ -697,14 +678,16 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
>  	rcu_read_lock();
>  	spin_lock_irq(&q->queue_lock);
>  
> -	blkg = blkg_lookup_check(blkcg, pol, q);
> -	if (IS_ERR(blkg)) {
> -		ret = PTR_ERR(blkg);
> +	if (!blkcg_policy_enabled(q, pol)) {
> +		ret = -EOPNOTSUPP;
>  		goto fail_unlock;
>  	}
>  
> -	if (blkg)
> +	blkg = blkg_lookup(blkcg, q);
> +	if (blkg) {
> +		blkg_update_hint(blkcg, blkg);
>  		goto success;
> +	}
>  
>  	/*
>  	 * Create blkgs walking down from blkcg_root to @blkcg, so that all
> @@ -740,14 +723,15 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
>  		rcu_read_lock();
>  		spin_lock_irq(&q->queue_lock);
>  
> -		blkg = blkg_lookup_check(pos, pol, q);
> -		if (IS_ERR(blkg)) {
> -			ret = PTR_ERR(blkg);
> +		if (!blkcg_policy_enabled(q, pol)) {
>  			blkg_free(new_blkg);
> +			ret = -EOPNOTSUPP;
>  			goto fail_preloaded;
>  		}
>  
> +		blkg = blkg_lookup(pos, q);
>  		if (blkg) {
> +			blkg_update_hint(pos, blkg);
>  			blkg_free(new_blkg);

I don't think conf_prep needs to update the hint in the first place, so we
can just do blkg_lookup()'s and drop the blkg_update_hint() calls.

Thanks.

-- 
tejun

[PATCH 06/17] blk-cgroup: pass a gendisk to blkcg_init_queue and blkcg_exit_queue

[Christoph Hellwig]

Pass the gendisk to blkcg_init_disk and blkcg_exit_disk as part of moving
the blk-cgroup infrastructure to be gendisk based.  Also remove the
rather pointless kerneldoc comments for these internal functions with a
single caller each.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c | 25 +++++--------------------
 block/blk-cgroup.h |  8 ++++----
 block/genhd.c      |  5 +++--
 3 files changed, 12 insertions(+), 26 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 1306112d76486..4ca6933a7c3f5 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -1230,18 +1230,9 @@ static int blkcg_css_online(struct cgroup_subsys_state *css)
 	return 0;
 }
 
-/**
- * blkcg_init_queue - initialize blkcg part of request queue
- * @q: request_queue to initialize
- *
- * Called from blk_alloc_queue(). Responsible for initializing blkcg
- * part of new request_queue @q.
- *
- * RETURNS:
- * 0 on success, -errno on failure.
- */
-int blkcg_init_queue(struct request_queue *q)
+int blkcg_init_disk(struct gendisk *disk)
 {
+	struct request_queue *q = disk->queue;
 	struct blkcg_gq *new_blkg, *blkg;
 	bool preloaded;
 	int ret;
@@ -1294,16 +1285,10 @@ int blkcg_init_queue(struct request_queue *q)
 	return PTR_ERR(blkg);
 }
 
-/**
- * blkcg_exit_queue - exit and release blkcg part of request_queue
- * @q: request_queue being released
- *
- * Called from blk_exit_queue().  Responsible for exiting blkcg part.
- */
-void blkcg_exit_queue(struct request_queue *q)
+void blkcg_exit_disk(struct gendisk *disk)
 {
-	blkg_destroy_all(q);
-	blk_throtl_exit(q);
+	blkg_destroy_all(disk->queue);
+	blk_throtl_exit(disk->queue);
 }
 
 static void blkcg_bind(struct cgroup_subsys_state *root_css)
diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
index 91b7ae0773be6..aa2b286bc825f 100644
--- a/block/blk-cgroup.h
+++ b/block/blk-cgroup.h
@@ -178,8 +178,8 @@ struct blkcg_policy {
 extern struct blkcg blkcg_root;
 extern bool blkcg_debug_stats;
 
-int blkcg_init_queue(struct request_queue *q);
-void blkcg_exit_queue(struct request_queue *q);
+int blkcg_init_disk(struct gendisk *disk);
+void blkcg_exit_disk(struct gendisk *disk);
 
 /* Blkio controller policy registration */
 int blkcg_policy_register(struct blkcg_policy *pol);
@@ -481,8 +481,8 @@ struct blkcg {
 };
 
 static inline struct blkcg_gq *blkg_lookup(struct blkcg *blkcg, void *key) { return NULL; }
-static inline int blkcg_init_queue(struct request_queue *q) { return 0; }
-static inline void blkcg_exit_queue(struct request_queue *q) { }
+static inline int blkcg_init_disk(struct gendisk *disk) { return 0; }
+static inline void blkcg_exit_disk(struct gendisk *disk) { }
 static inline int blkcg_policy_register(struct blkcg_policy *pol) { return 0; }
 static inline void blkcg_policy_unregister(struct blkcg_policy *pol) { }
 static inline int blkcg_activate_policy(struct request_queue *q,
diff --git a/block/genhd.c b/block/genhd.c
index d36fabf0abc1f..f1af045fac2fe 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -1150,7 +1150,8 @@ static void disk_release(struct device *dev)
 	    !test_bit(GD_ADDED, &disk->state))
 		blk_mq_exit_queue(disk->queue);
 
-	blkcg_exit_queue(disk->queue);
+	blkcg_exit_disk(disk);
+
 	bioset_exit(&disk->bio_split);
 
 	disk_release_events(disk);
@@ -1363,7 +1364,7 @@ struct gendisk *__alloc_disk_node(struct request_queue *q, int node_id,
 	if (xa_insert(&disk->part_tbl, 0, disk->part0, GFP_KERNEL))
 		goto out_destroy_part_tbl;
 
-	if (blkcg_init_queue(q))
+	if (blkcg_init_disk(disk))
 		goto out_erase_part0;
 
 	rand_initialize_disk(disk);
-- 
2.30.2

Re: [PATCH 06/17] blk-cgroup: pass a gendisk to blkcg_init_queue and blkcg_exit_queue

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:50PM +0200, Christoph Hellwig wrote:
> Pass the gendisk to blkcg_init_disk and blkcg_exit_disk as part of moving
> the blk-cgroup infrastructure to be gendisk based.  Also remove the
> rather pointless kerneldoc comments for these internal functions with a
> single caller each.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c | 25 +++++--------------------
>  block/blk-cgroup.h |  8 ++++----
>  block/genhd.c      |  5 +++--
>  3 files changed, 12 insertions(+), 26 deletions(-)

This creates a stale reference to blkcg_exit_queue() in comment of
block/blk-core.c which needs to be adapted, I think:

--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -223,7 +223,7 @@ const char *blk_status_to_str(blk_status_t status)
  *
  *     This function does not cancel any asynchronous activity arising
  *     out of elevator or throttling code. That would require elevator_exit()
- *     and blkcg_exit_queue() to be called with queue lock initialized.
+ *     and blkcg_exit_disk() to be called with queue lock initialized.
  *
  */
 void blk_sync_queue(struct request_queue *q)

--
Otherwise
Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index 1306112d76486..4ca6933a7c3f5 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -1230,18 +1230,9 @@ static int blkcg_css_online(struct cgroup_subsys_state *css)
>  	return 0;
>  }
>  
> -/**
> - * blkcg_init_queue - initialize blkcg part of request queue
> - * @q: request_queue to initialize
> - *
> - * Called from blk_alloc_queue(). Responsible for initializing blkcg
> - * part of new request_queue @q.
> - *
> - * RETURNS:
> - * 0 on success, -errno on failure.
> - */
> -int blkcg_init_queue(struct request_queue *q)
> +int blkcg_init_disk(struct gendisk *disk)
>  {
> +	struct request_queue *q = disk->queue;
>  	struct blkcg_gq *new_blkg, *blkg;
>  	bool preloaded;
>  	int ret;
> @@ -1294,16 +1285,10 @@ int blkcg_init_queue(struct request_queue *q)
>  	return PTR_ERR(blkg);
>  }
>  
> -/**
> - * blkcg_exit_queue - exit and release blkcg part of request_queue
> - * @q: request_queue being released
> - *
> - * Called from blk_exit_queue().  Responsible for exiting blkcg part.
> - */
> -void blkcg_exit_queue(struct request_queue *q)
> +void blkcg_exit_disk(struct gendisk *disk)
>  {
> -	blkg_destroy_all(q);
> -	blk_throtl_exit(q);
> +	blkg_destroy_all(disk->queue);
> +	blk_throtl_exit(disk->queue);
>  }
>  
>  static void blkcg_bind(struct cgroup_subsys_state *root_css)
> diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
> index 91b7ae0773be6..aa2b286bc825f 100644
> --- a/block/blk-cgroup.h
> +++ b/block/blk-cgroup.h
> @@ -178,8 +178,8 @@ struct blkcg_policy {
>  extern struct blkcg blkcg_root;
>  extern bool blkcg_debug_stats;
>  
> -int blkcg_init_queue(struct request_queue *q);
> -void blkcg_exit_queue(struct request_queue *q);
> +int blkcg_init_disk(struct gendisk *disk);
> +void blkcg_exit_disk(struct gendisk *disk);
>  
>  /* Blkio controller policy registration */
>  int blkcg_policy_register(struct blkcg_policy *pol);
> @@ -481,8 +481,8 @@ struct blkcg {
>  };
>  
>  static inline struct blkcg_gq *blkg_lookup(struct blkcg *blkcg, void *key) { return NULL; }
> -static inline int blkcg_init_queue(struct request_queue *q) { return 0; }
> -static inline void blkcg_exit_queue(struct request_queue *q) { }
> +static inline int blkcg_init_disk(struct gendisk *disk) { return 0; }
> +static inline void blkcg_exit_disk(struct gendisk *disk) { }
>  static inline int blkcg_policy_register(struct blkcg_policy *pol) { return 0; }
>  static inline void blkcg_policy_unregister(struct blkcg_policy *pol) { }
>  static inline int blkcg_activate_policy(struct request_queue *q,
> diff --git a/block/genhd.c b/block/genhd.c
> index d36fabf0abc1f..f1af045fac2fe 100644
> --- a/block/genhd.c
> +++ b/block/genhd.c
> @@ -1150,7 +1150,8 @@ static void disk_release(struct device *dev)
>  	    !test_bit(GD_ADDED, &disk->state))
>  		blk_mq_exit_queue(disk->queue);
>  
> -	blkcg_exit_queue(disk->queue);
> +	blkcg_exit_disk(disk);
> +
>  	bioset_exit(&disk->bio_split);
>  
>  	disk_release_events(disk);
> @@ -1363,7 +1364,7 @@ struct gendisk *__alloc_disk_node(struct request_queue *q, int node_id,
>  	if (xa_insert(&disk->part_tbl, 0, disk->part0, GFP_KERNEL))
>  		goto out_destroy_part_tbl;
>  
> -	if (blkcg_init_queue(q))
> +	if (blkcg_init_disk(disk))
>  		goto out_erase_part0;
>  
>  	rand_initialize_disk(disk);
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 07/17] blk-ioprio: pass a gendisk to blk_ioprio_init and blk_ioprio_exit

[Christoph Hellwig]

Pass the gendisk to blk_ioprio_init and blk_ioprio_exit as part of moving
the blk-cgroup infrastructure to be gendisk based.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c | 4 ++--
 block/blk-ioprio.c | 8 ++++----
 block/blk-ioprio.h | 8 ++++----
 3 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 4ca6933a7c3f5..89974fd0db3da 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -1257,7 +1257,7 @@ int blkcg_init_disk(struct gendisk *disk)
 	if (preloaded)
 		radix_tree_preload_end();
 
-	ret = blk_ioprio_init(q);
+	ret = blk_ioprio_init(disk);
 	if (ret)
 		goto err_destroy_all;
 
@@ -1274,7 +1274,7 @@ int blkcg_init_disk(struct gendisk *disk)
 err_throtl_exit:
 	blk_throtl_exit(q);
 err_ioprio_exit:
-	blk_ioprio_exit(q);
+	blk_ioprio_exit(disk);
 err_destroy_all:
 	blkg_destroy_all(q);
 	return ret;
diff --git a/block/blk-ioprio.c b/block/blk-ioprio.c
index c00060a02c6ef..8bb6b8eba4cee 100644
--- a/block/blk-ioprio.c
+++ b/block/blk-ioprio.c
@@ -202,14 +202,14 @@ void blkcg_set_ioprio(struct bio *bio)
 		bio->bi_ioprio = prio;
 }
 
-void blk_ioprio_exit(struct request_queue *q)
+void blk_ioprio_exit(struct gendisk *disk)
 {
-	blkcg_deactivate_policy(q, &ioprio_policy);
+	blkcg_deactivate_policy(disk->queue, &ioprio_policy);
 }
 
-int blk_ioprio_init(struct request_queue *q)
+int blk_ioprio_init(struct gendisk *disk)
 {
-	return blkcg_activate_policy(q, &ioprio_policy);
+	return blkcg_activate_policy(disk->queue, &ioprio_policy);
 }
 
 static int __init ioprio_init(void)
diff --git a/block/blk-ioprio.h b/block/blk-ioprio.h
index 5a1eb550e178c..b6afb8e80de05 100644
--- a/block/blk-ioprio.h
+++ b/block/blk-ioprio.h
@@ -9,15 +9,15 @@ struct request_queue;
 struct bio;
 
 #ifdef CONFIG_BLK_CGROUP_IOPRIO
-int blk_ioprio_init(struct request_queue *q);
-void blk_ioprio_exit(struct request_queue *q);
+int blk_ioprio_init(struct gendisk *disk);
+void blk_ioprio_exit(struct gendisk *disk);
 void blkcg_set_ioprio(struct bio *bio);
 #else
-static inline int blk_ioprio_init(struct request_queue *q)
+static inline int blk_ioprio_init(struct gendisk *disk)
 {
 	return 0;
 }
-static inline void blk_ioprio_exit(struct request_queue *q)
+static inline void blk_ioprio_exit(struct gendisk *disk)
 {
 }
 static inline void blkcg_set_ioprio(struct bio *bio)
-- 
2.30.2

Re: [PATCH 07/17] blk-ioprio: pass a gendisk to blk_ioprio_init and blk_ioprio_exit

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:51PM +0200, Christoph Hellwig wrote:
> Pass the gendisk to blk_ioprio_init and blk_ioprio_exit as part of moving
> the blk-cgroup infrastructure to be gendisk based.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c | 4 ++--
>  block/blk-ioprio.c | 8 ++++----
>  block/blk-ioprio.h | 8 ++++----
>  3 files changed, 10 insertions(+), 10 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index 4ca6933a7c3f5..89974fd0db3da 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -1257,7 +1257,7 @@ int blkcg_init_disk(struct gendisk *disk)
>  	if (preloaded)
>  		radix_tree_preload_end();
>  
> -	ret = blk_ioprio_init(q);
> +	ret = blk_ioprio_init(disk);
>  	if (ret)
>  		goto err_destroy_all;
>  
> @@ -1274,7 +1274,7 @@ int blkcg_init_disk(struct gendisk *disk)
>  err_throtl_exit:
>  	blk_throtl_exit(q);
>  err_ioprio_exit:
> -	blk_ioprio_exit(q);
> +	blk_ioprio_exit(disk);
>  err_destroy_all:
>  	blkg_destroy_all(q);
>  	return ret;
> diff --git a/block/blk-ioprio.c b/block/blk-ioprio.c
> index c00060a02c6ef..8bb6b8eba4cee 100644
> --- a/block/blk-ioprio.c
> +++ b/block/blk-ioprio.c
> @@ -202,14 +202,14 @@ void blkcg_set_ioprio(struct bio *bio)
>  		bio->bi_ioprio = prio;
>  }
>  
> -void blk_ioprio_exit(struct request_queue *q)
> +void blk_ioprio_exit(struct gendisk *disk)
>  {
> -	blkcg_deactivate_policy(q, &ioprio_policy);
> +	blkcg_deactivate_policy(disk->queue, &ioprio_policy);
>  }
>  
> -int blk_ioprio_init(struct request_queue *q)
> +int blk_ioprio_init(struct gendisk *disk)
>  {
> -	return blkcg_activate_policy(q, &ioprio_policy);
> +	return blkcg_activate_policy(disk->queue, &ioprio_policy);
>  }
>  
>  static int __init ioprio_init(void)
> diff --git a/block/blk-ioprio.h b/block/blk-ioprio.h
> index 5a1eb550e178c..b6afb8e80de05 100644
> --- a/block/blk-ioprio.h
> +++ b/block/blk-ioprio.h
> @@ -9,15 +9,15 @@ struct request_queue;
>  struct bio;
>  
>  #ifdef CONFIG_BLK_CGROUP_IOPRIO
> -int blk_ioprio_init(struct request_queue *q);
> -void blk_ioprio_exit(struct request_queue *q);
> +int blk_ioprio_init(struct gendisk *disk);
> +void blk_ioprio_exit(struct gendisk *disk);
>  void blkcg_set_ioprio(struct bio *bio);
>  #else
> -static inline int blk_ioprio_init(struct request_queue *q)
> +static inline int blk_ioprio_init(struct gendisk *disk)
>  {
>  	return 0;
>  }
> -static inline void blk_ioprio_exit(struct request_queue *q)
> +static inline void blk_ioprio_exit(struct gendisk *disk)
>  {
>  }
>  static inline void blkcg_set_ioprio(struct bio *bio)
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 08/17] blk-iolatency: pass a gendisk to blk_iolatency_init

[Christoph Hellwig]

Pass the gendisk to blk_iolatency_init as part of moving the blk-cgroup
infrastructure to be gendisk based.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c    | 2 +-
 block/blk-iolatency.c | 3 ++-
 block/blk.h           | 4 ++--
 3 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 89974fd0db3da..82a117ff54de5 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -1265,7 +1265,7 @@ int blkcg_init_disk(struct gendisk *disk)
 	if (ret)
 		goto err_ioprio_exit;
 
-	ret = blk_iolatency_init(q);
+	ret = blk_iolatency_init(disk);
 	if (ret)
 		goto err_throtl_exit;
 
diff --git a/block/blk-iolatency.c b/block/blk-iolatency.c
index e285152345a20..c6f61fe88b875 100644
--- a/block/blk-iolatency.c
+++ b/block/blk-iolatency.c
@@ -756,8 +756,9 @@ static void blkiolatency_enable_work_fn(struct work_struct *work)
 	}
 }
 
-int blk_iolatency_init(struct request_queue *q)
+int blk_iolatency_init(struct gendisk *disk)
 {
+	struct request_queue *q = disk->queue;
 	struct blk_iolatency *blkiolat;
 	struct rq_qos *rqos;
 	int ret;
diff --git a/block/blk.h b/block/blk.h
index d7142c4d2fefb..361db83147c6f 100644
--- a/block/blk.h
+++ b/block/blk.h
@@ -389,9 +389,9 @@ static inline struct bio *blk_queue_bounce(struct bio *bio,
 }
 
 #ifdef CONFIG_BLK_CGROUP_IOLATENCY
-extern int blk_iolatency_init(struct request_queue *q);
+int blk_iolatency_init(struct gendisk *disk);
 #else
-static inline int blk_iolatency_init(struct request_queue *q) { return 0; }
+static int blk_iolatency_init(struct gendisk *disk) { return 0 };
 #endif
 
 #ifdef CONFIG_BLK_DEV_ZONED
-- 
2.30.2

Re: [PATCH 08/17] blk-iolatency: pass a gendisk to blk_iolatency_init

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:52PM +0200, Christoph Hellwig wrote:
> Pass the gendisk to blk_iolatency_init as part of moving the blk-cgroup
> infrastructure to be gendisk based.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c    | 2 +-
>  block/blk-iolatency.c | 3 ++-
>  block/blk.h           | 4 ++--
>  3 files changed, 5 insertions(+), 4 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index 89974fd0db3da..82a117ff54de5 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -1265,7 +1265,7 @@ int blkcg_init_disk(struct gendisk *disk)
>  	if (ret)
>  		goto err_ioprio_exit;
>  
> -	ret = blk_iolatency_init(q);
> +	ret = blk_iolatency_init(disk);
>  	if (ret)
>  		goto err_throtl_exit;
>  
> diff --git a/block/blk-iolatency.c b/block/blk-iolatency.c
> index e285152345a20..c6f61fe88b875 100644
> --- a/block/blk-iolatency.c
> +++ b/block/blk-iolatency.c
> @@ -756,8 +756,9 @@ static void blkiolatency_enable_work_fn(struct work_struct *work)
>  	}
>  }
>  
> -int blk_iolatency_init(struct request_queue *q)
> +int blk_iolatency_init(struct gendisk *disk)
>  {
> +	struct request_queue *q = disk->queue;
>  	struct blk_iolatency *blkiolat;
>  	struct rq_qos *rqos;
>  	int ret;
> diff --git a/block/blk.h b/block/blk.h
> index d7142c4d2fefb..361db83147c6f 100644
> --- a/block/blk.h
> +++ b/block/blk.h
> @@ -389,9 +389,9 @@ static inline struct bio *blk_queue_bounce(struct bio *bio,
>  }
>  
>  #ifdef CONFIG_BLK_CGROUP_IOLATENCY
> -extern int blk_iolatency_init(struct request_queue *q);
> +int blk_iolatency_init(struct gendisk *disk);
>  #else
> -static inline int blk_iolatency_init(struct request_queue *q) { return 0; }
> +static int blk_iolatency_init(struct gendisk *disk) { return 0 };
>  #endif
>  
>  #ifdef CONFIG_BLK_DEV_ZONED
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

Re: [PATCH 08/17] blk-iolatency: pass a gendisk to blk_iolatency_init

[Jens Axboe]

> diff --git a/block/blk.h b/block/blk.h
> index d7142c4d2fefb..361db83147c6f 100644
> --- a/block/blk.h
> +++ b/block/blk.h
> @@ -389,9 +389,9 @@ static inline struct bio *blk_queue_bounce(struct bio *bio,
>  }
>  
>  #ifdef CONFIG_BLK_CGROUP_IOLATENCY
> -extern int blk_iolatency_init(struct request_queue *q);
> +int blk_iolatency_init(struct gendisk *disk);
>  #else
> -static inline int blk_iolatency_init(struct request_queue *q) { return 0; }
> +static int blk_iolatency_init(struct gendisk *disk) { return 0 };
>  #endif

This is missing both an 'inline' and a semicolon... I fixed it up while
applying.

-- 
Jens Axboe

[PATCH 09/17] blk-iocost: simplify ioc_name

[Christoph Hellwig]

Just directly dereference the disk name instead of going through multiple
hoops to find the same value.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-iocost.c | 14 +++++---------
 1 file changed, 5 insertions(+), 9 deletions(-)

diff --git a/block/blk-iocost.c b/block/blk-iocost.c
index 7936e5f5821c7..cba9d3ad58e16 100644
--- a/block/blk-iocost.c
+++ b/block/blk-iocost.c
@@ -664,17 +664,13 @@ static struct ioc *q_to_ioc(struct request_queue *q)
 	return rqos_to_ioc(rq_qos_id(q, RQ_QOS_COST));
 }
 
-static const char *q_name(struct request_queue *q)
-{
-	if (blk_queue_registered(q))
-		return kobject_name(q->kobj.parent);
-	else
-		return "<unknown>";
-}
-
 static const char __maybe_unused *ioc_name(struct ioc *ioc)
 {
-	return q_name(ioc->rqos.q);
+	struct gendisk *disk = ioc->rqos.q->disk;
+
+	if (!disk)
+		return "<unknown>";
+	return disk->disk_name;
 }
 
 static struct ioc_gq *pd_to_iocg(struct blkg_policy_data *pd)
-- 
2.30.2

Re: [PATCH 09/17] blk-iocost: simplify ioc_name

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:53PM +0200, Christoph Hellwig wrote:
> Just directly dereference the disk name instead of going through multiple
> hoops to find the same value.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-iocost.c | 14 +++++---------
>  1 file changed, 5 insertions(+), 9 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-iocost.c b/block/blk-iocost.c
> index 7936e5f5821c7..cba9d3ad58e16 100644
> --- a/block/blk-iocost.c
> +++ b/block/blk-iocost.c
> @@ -664,17 +664,13 @@ static struct ioc *q_to_ioc(struct request_queue *q)
>  	return rqos_to_ioc(rq_qos_id(q, RQ_QOS_COST));
>  }
>  
> -static const char *q_name(struct request_queue *q)
> -{
> -	if (blk_queue_registered(q))
> -		return kobject_name(q->kobj.parent);
> -	else
> -		return "<unknown>";
> -}
> -
>  static const char __maybe_unused *ioc_name(struct ioc *ioc)
>  {
> -	return q_name(ioc->rqos.q);
> +	struct gendisk *disk = ioc->rqos.q->disk;
> +
> +	if (!disk)
> +		return "<unknown>";
> +	return disk->disk_name;
>  }
>  
>  static struct ioc_gq *pd_to_iocg(struct blkg_policy_data *pd)
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 10/17] blk-iocost: pass a gendisk to blk_iocost_init

[Christoph Hellwig]

Pass the gendisk to blk_iocost_init as part of moving the blk-cgroup
infrastructure to be gendisk based.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-iocost.c | 7 ++++---
 block/blk.h        | 2 +-
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/block/blk-iocost.c b/block/blk-iocost.c
index cba9d3ad58e16..1e7bf0d353227 100644
--- a/block/blk-iocost.c
+++ b/block/blk-iocost.c
@@ -2828,8 +2828,9 @@ static struct rq_qos_ops ioc_rqos_ops = {
 	.exit = ioc_rqos_exit,
 };
 
-static int blk_iocost_init(struct request_queue *q)
+static int blk_iocost_init(struct gendisk *disk)
 {
+	struct request_queue *q = disk->queue;
 	struct ioc *ioc;
 	struct rq_qos *rqos;
 	int i, cpu, ret;
@@ -3178,7 +3179,7 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
 
 	ioc = q_to_ioc(bdev_get_queue(bdev));
 	if (!ioc) {
-		ret = blk_iocost_init(bdev_get_queue(bdev));
+		ret = blk_iocost_init(bdev->bd_disk);
 		if (ret)
 			goto err;
 		ioc = q_to_ioc(bdev_get_queue(bdev));
@@ -3345,7 +3346,7 @@ static ssize_t ioc_cost_model_write(struct kernfs_open_file *of, char *input,
 
 	ioc = q_to_ioc(bdev_get_queue(bdev));
 	if (!ioc) {
-		ret = blk_iocost_init(bdev_get_queue(bdev));
+		ret = blk_iocost_init(bdev->bd_disk);
 		if (ret)
 			goto err;
 		ioc = q_to_ioc(bdev_get_queue(bdev));
diff --git a/block/blk.h b/block/blk.h
index 361db83147c6f..8d5c7a6f52a66 100644
--- a/block/blk.h
+++ b/block/blk.h
@@ -391,7 +391,7 @@ static inline struct bio *blk_queue_bounce(struct bio *bio,
 #ifdef CONFIG_BLK_CGROUP_IOLATENCY
 int blk_iolatency_init(struct gendisk *disk);
 #else
-static int blk_iolatency_init(struct gendisk *disk) { return 0 };
+static int blk_iolatency_init(struct gendisk *disk) { return 0; }
 #endif
 
 #ifdef CONFIG_BLK_DEV_ZONED
-- 
2.30.2

Re: [PATCH 10/17] blk-iocost: pass a gendisk to blk_iocost_init

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:54PM +0200, Christoph Hellwig wrote:
> Pass the gendisk to blk_iocost_init as part of moving the blk-cgroup
> infrastructure to be gendisk based.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-iocost.c | 7 ++++---
>  block/blk.h        | 2 +-
>  2 files changed, 5 insertions(+), 4 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-iocost.c b/block/blk-iocost.c
> index cba9d3ad58e16..1e7bf0d353227 100644
> --- a/block/blk-iocost.c
> +++ b/block/blk-iocost.c
> @@ -2828,8 +2828,9 @@ static struct rq_qos_ops ioc_rqos_ops = {
>  	.exit = ioc_rqos_exit,
>  };
>  
> -static int blk_iocost_init(struct request_queue *q)
> +static int blk_iocost_init(struct gendisk *disk)
>  {
> +	struct request_queue *q = disk->queue;
>  	struct ioc *ioc;
>  	struct rq_qos *rqos;
>  	int i, cpu, ret;
> @@ -3178,7 +3179,7 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
>  
>  	ioc = q_to_ioc(bdev_get_queue(bdev));
>  	if (!ioc) {
> -		ret = blk_iocost_init(bdev_get_queue(bdev));
> +		ret = blk_iocost_init(bdev->bd_disk);
>  		if (ret)
>  			goto err;
>  		ioc = q_to_ioc(bdev_get_queue(bdev));
> @@ -3345,7 +3346,7 @@ static ssize_t ioc_cost_model_write(struct kernfs_open_file *of, char *input,
>  
>  	ioc = q_to_ioc(bdev_get_queue(bdev));
>  	if (!ioc) {
> -		ret = blk_iocost_init(bdev_get_queue(bdev));
> +		ret = blk_iocost_init(bdev->bd_disk);
>  		if (ret)
>  			goto err;
>  		ioc = q_to_ioc(bdev_get_queue(bdev));
> diff --git a/block/blk.h b/block/blk.h
> index 361db83147c6f..8d5c7a6f52a66 100644
> --- a/block/blk.h
> +++ b/block/blk.h
> @@ -391,7 +391,7 @@ static inline struct bio *blk_queue_bounce(struct bio *bio,
>  #ifdef CONFIG_BLK_CGROUP_IOLATENCY
>  int blk_iolatency_init(struct gendisk *disk);
>  #else
> -static int blk_iolatency_init(struct gendisk *disk) { return 0 };
> +static int blk_iolatency_init(struct gendisk *disk) { return 0; }
>  #endif
>  
>  #ifdef CONFIG_BLK_DEV_ZONED
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 11/17] blk-iocost: cleanup ioc_qos_write

[Christoph Hellwig]

Use a local disk variable instead of retrieving the disk and
request_queue over and over by various means.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-iocost.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/block/blk-iocost.c b/block/blk-iocost.c
index 1e7bf0d353227..b8e5f550aa5be 100644
--- a/block/blk-iocost.c
+++ b/block/blk-iocost.c
@@ -3167,6 +3167,7 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
 			     size_t nbytes, loff_t off)
 {
 	struct block_device *bdev;
+	struct gendisk *disk;
 	struct ioc *ioc;
 	u32 qos[NR_QOS_PARAMS];
 	bool enable, user;
@@ -3177,12 +3178,13 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
 	if (IS_ERR(bdev))
 		return PTR_ERR(bdev);
 
-	ioc = q_to_ioc(bdev_get_queue(bdev));
+	disk = bdev->bd_disk;
+	ioc = q_to_ioc(disk->queue);
 	if (!ioc) {
-		ret = blk_iocost_init(bdev->bd_disk);
+		ret = blk_iocost_init(disk);
 		if (ret)
 			goto err;
-		ioc = q_to_ioc(bdev_get_queue(bdev));
+		ioc = q_to_ioc(disk->queue);
 	}
 
 	spin_lock_irq(&ioc->lock);
@@ -3259,11 +3261,11 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
 	spin_lock_irq(&ioc->lock);
 
 	if (enable) {
-		blk_stat_enable_accounting(ioc->rqos.q);
-		blk_queue_flag_set(QUEUE_FLAG_RQ_ALLOC_TIME, ioc->rqos.q);
+		blk_stat_enable_accounting(disk->queue);
+		blk_queue_flag_set(QUEUE_FLAG_RQ_ALLOC_TIME, disk->queue);
 		ioc->enabled = true;
 	} else {
-		blk_queue_flag_clear(QUEUE_FLAG_RQ_ALLOC_TIME, ioc->rqos.q);
+		blk_queue_flag_clear(QUEUE_FLAG_RQ_ALLOC_TIME, disk->queue);
 		ioc->enabled = false;
 	}
 
-- 
2.30.2

Re: [PATCH 11/17] blk-iocost: cleanup ioc_qos_write

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:55PM +0200, Christoph Hellwig wrote:
> Use a local disk variable instead of retrieving the disk and
> request_queue over and over by various means.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-iocost.c | 14 ++++++++------
>  1 file changed, 8 insertions(+), 6 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-iocost.c b/block/blk-iocost.c
> index 1e7bf0d353227..b8e5f550aa5be 100644
> --- a/block/blk-iocost.c
> +++ b/block/blk-iocost.c
> @@ -3167,6 +3167,7 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
>  			     size_t nbytes, loff_t off)
>  {
>  	struct block_device *bdev;
> +	struct gendisk *disk;
>  	struct ioc *ioc;
>  	u32 qos[NR_QOS_PARAMS];
>  	bool enable, user;
> @@ -3177,12 +3178,13 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
>  	if (IS_ERR(bdev))
>  		return PTR_ERR(bdev);
>  
> -	ioc = q_to_ioc(bdev_get_queue(bdev));
> +	disk = bdev->bd_disk;
> +	ioc = q_to_ioc(disk->queue);
>  	if (!ioc) {
> -		ret = blk_iocost_init(bdev->bd_disk);
> +		ret = blk_iocost_init(disk);
>  		if (ret)
>  			goto err;
> -		ioc = q_to_ioc(bdev_get_queue(bdev));
> +		ioc = q_to_ioc(disk->queue);
>  	}
>  
>  	spin_lock_irq(&ioc->lock);
> @@ -3259,11 +3261,11 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
>  	spin_lock_irq(&ioc->lock);
>  
>  	if (enable) {
> -		blk_stat_enable_accounting(ioc->rqos.q);
> -		blk_queue_flag_set(QUEUE_FLAG_RQ_ALLOC_TIME, ioc->rqos.q);
> +		blk_stat_enable_accounting(disk->queue);
> +		blk_queue_flag_set(QUEUE_FLAG_RQ_ALLOC_TIME, disk->queue);
>  		ioc->enabled = true;
>  	} else {
> -		blk_queue_flag_clear(QUEUE_FLAG_RQ_ALLOC_TIME, ioc->rqos.q);
> +		blk_queue_flag_clear(QUEUE_FLAG_RQ_ALLOC_TIME, disk->queue);
>  		ioc->enabled = false;
>  	}
>  
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 12/17] blk-throttle: pass a gendisk to blk_throtl_init and blk_throtl_exit

[Christoph Hellwig]

Pass the gendisk to blk_throtl_init and blk_throtl_exit as part of moving
the blk-cgroup infrastructure to be gendisk based.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c   | 6 +++---
 block/blk-throttle.c | 7 +++++--
 block/blk-throttle.h | 8 ++++----
 3 files changed, 12 insertions(+), 9 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 82a117ff54de5..3dfd78f1312db 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -1261,7 +1261,7 @@ int blkcg_init_disk(struct gendisk *disk)
 	if (ret)
 		goto err_destroy_all;
 
-	ret = blk_throtl_init(q);
+	ret = blk_throtl_init(disk);
 	if (ret)
 		goto err_ioprio_exit;
 
@@ -1272,7 +1272,7 @@ int blkcg_init_disk(struct gendisk *disk)
 	return 0;
 
 err_throtl_exit:
-	blk_throtl_exit(q);
+	blk_throtl_exit(disk);
 err_ioprio_exit:
 	blk_ioprio_exit(disk);
 err_destroy_all:
@@ -1288,7 +1288,7 @@ int blkcg_init_disk(struct gendisk *disk)
 void blkcg_exit_disk(struct gendisk *disk)
 {
 	blkg_destroy_all(disk->queue);
-	blk_throtl_exit(disk->queue);
+	blk_throtl_exit(disk);
 }
 
 static void blkcg_bind(struct cgroup_subsys_state *root_css)
diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index 55f2d985cfbbd..9ca8ae0ae6683 100644
--- a/block/blk-throttle.c
+++ b/block/blk-throttle.c
@@ -2358,8 +2358,9 @@ void blk_throtl_bio_endio(struct bio *bio)
 }
 #endif
 
-int blk_throtl_init(struct request_queue *q)
+int blk_throtl_init(struct gendisk *disk)
 {
+	struct request_queue *q = disk->queue;
 	struct throtl_data *td;
 	int ret;
 
@@ -2401,8 +2402,10 @@ int blk_throtl_init(struct request_queue *q)
 	return ret;
 }
 
-void blk_throtl_exit(struct request_queue *q)
+void blk_throtl_exit(struct gendisk *disk)
 {
+	struct request_queue *q = disk->queue;
+
 	BUG_ON(!q->td);
 	del_timer_sync(&q->td->service_queue.pending_timer);
 	throtl_shutdown_wq(q);
diff --git a/block/blk-throttle.h b/block/blk-throttle.h
index 66b4292b1b92a..f75852a4e5337 100644
--- a/block/blk-throttle.h
+++ b/block/blk-throttle.h
@@ -168,14 +168,14 @@ static inline struct throtl_grp *blkg_to_tg(struct blkcg_gq *blkg)
  * Internal throttling interface
  */
 #ifndef CONFIG_BLK_DEV_THROTTLING
-static inline int blk_throtl_init(struct request_queue *q) { return 0; }
-static inline void blk_throtl_exit(struct request_queue *q) { }
+static inline int blk_throtl_init(struct gendisk *disk) { return 0; }
+static inline void blk_throtl_exit(struct gendisk *disk) { }
 static inline void blk_throtl_register_queue(struct request_queue *q) { }
 static inline bool blk_throtl_bio(struct bio *bio) { return false; }
 static inline void blk_throtl_cancel_bios(struct request_queue *q) { }
 #else /* CONFIG_BLK_DEV_THROTTLING */
-int blk_throtl_init(struct request_queue *q);
-void blk_throtl_exit(struct request_queue *q);
+int blk_throtl_init(struct gendisk *disk);
+void blk_throtl_exit(struct gendisk *disk);
 void blk_throtl_register_queue(struct request_queue *q);
 bool __blk_throtl_bio(struct bio *bio);
 void blk_throtl_cancel_bios(struct request_queue *q);
-- 
2.30.2

Re: [PATCH 12/17] blk-throttle: pass a gendisk to blk_throtl_init and blk_throtl_exit

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:56PM +0200, Christoph Hellwig wrote:
> Pass the gendisk to blk_throtl_init and blk_throtl_exit as part of moving
> the blk-cgroup infrastructure to be gendisk based.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c   | 6 +++---
>  block/blk-throttle.c | 7 +++++--
>  block/blk-throttle.h | 8 ++++----
>  3 files changed, 12 insertions(+), 9 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index 82a117ff54de5..3dfd78f1312db 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -1261,7 +1261,7 @@ int blkcg_init_disk(struct gendisk *disk)
>  	if (ret)
>  		goto err_destroy_all;
>  
> -	ret = blk_throtl_init(q);
> +	ret = blk_throtl_init(disk);
>  	if (ret)
>  		goto err_ioprio_exit;
>  
> @@ -1272,7 +1272,7 @@ int blkcg_init_disk(struct gendisk *disk)
>  	return 0;
>  
>  err_throtl_exit:
> -	blk_throtl_exit(q);
> +	blk_throtl_exit(disk);
>  err_ioprio_exit:
>  	blk_ioprio_exit(disk);
>  err_destroy_all:
> @@ -1288,7 +1288,7 @@ int blkcg_init_disk(struct gendisk *disk)
>  void blkcg_exit_disk(struct gendisk *disk)
>  {
>  	blkg_destroy_all(disk->queue);
> -	blk_throtl_exit(disk->queue);
> +	blk_throtl_exit(disk);
>  }
>  
>  static void blkcg_bind(struct cgroup_subsys_state *root_css)
> diff --git a/block/blk-throttle.c b/block/blk-throttle.c
> index 55f2d985cfbbd..9ca8ae0ae6683 100644
> --- a/block/blk-throttle.c
> +++ b/block/blk-throttle.c
> @@ -2358,8 +2358,9 @@ void blk_throtl_bio_endio(struct bio *bio)
>  }
>  #endif
>  
> -int blk_throtl_init(struct request_queue *q)
> +int blk_throtl_init(struct gendisk *disk)
>  {
> +	struct request_queue *q = disk->queue;
>  	struct throtl_data *td;
>  	int ret;
>  
> @@ -2401,8 +2402,10 @@ int blk_throtl_init(struct request_queue *q)
>  	return ret;
>  }
>  
> -void blk_throtl_exit(struct request_queue *q)
> +void blk_throtl_exit(struct gendisk *disk)
>  {
> +	struct request_queue *q = disk->queue;
> +
>  	BUG_ON(!q->td);
>  	del_timer_sync(&q->td->service_queue.pending_timer);
>  	throtl_shutdown_wq(q);
> diff --git a/block/blk-throttle.h b/block/blk-throttle.h
> index 66b4292b1b92a..f75852a4e5337 100644
> --- a/block/blk-throttle.h
> +++ b/block/blk-throttle.h
> @@ -168,14 +168,14 @@ static inline struct throtl_grp *blkg_to_tg(struct blkcg_gq *blkg)
>   * Internal throttling interface
>   */
>  #ifndef CONFIG_BLK_DEV_THROTTLING
> -static inline int blk_throtl_init(struct request_queue *q) { return 0; }
> -static inline void blk_throtl_exit(struct request_queue *q) { }
> +static inline int blk_throtl_init(struct gendisk *disk) { return 0; }
> +static inline void blk_throtl_exit(struct gendisk *disk) { }
>  static inline void blk_throtl_register_queue(struct request_queue *q) { }
>  static inline bool blk_throtl_bio(struct bio *bio) { return false; }
>  static inline void blk_throtl_cancel_bios(struct request_queue *q) { }
>  #else /* CONFIG_BLK_DEV_THROTTLING */
> -int blk_throtl_init(struct request_queue *q);
> -void blk_throtl_exit(struct request_queue *q);
> +int blk_throtl_init(struct gendisk *disk);
> +void blk_throtl_exit(struct gendisk *disk);
>  void blk_throtl_register_queue(struct request_queue *q);
>  bool __blk_throtl_bio(struct bio *bio);
>  void blk_throtl_cancel_bios(struct request_queue *q);
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 13/17] blk-throttle: pass a gendisk to blk_throtl_register_queue

[Christoph Hellwig]

Pass the gendisk to blk_throtl_register_queue as part of moving the
blk-cgroup infrastructure to be gendisk based.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-sysfs.c    | 2 +-
 block/blk-throttle.c | 3 ++-
 block/blk-throttle.h | 4 ++--
 3 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c
index e1f009aba6fd2..e71b3b43927c0 100644
--- a/block/blk-sysfs.c
+++ b/block/blk-sysfs.c
@@ -844,7 +844,7 @@ int blk_register_queue(struct gendisk *disk)
 
 	blk_queue_flag_set(QUEUE_FLAG_REGISTERED, q);
 	wbt_enable_default(q);
-	blk_throtl_register_queue(q);
+	blk_throtl_register(disk);
 
 	/* Now everything is ready and send out KOBJ_ADD uevent */
 	kobject_uevent(&q->kobj, KOBJ_ADD);
diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index 9ca8ae0ae6683..a9709d4af6f39 100644
--- a/block/blk-throttle.c
+++ b/block/blk-throttle.c
@@ -2415,8 +2415,9 @@ void blk_throtl_exit(struct gendisk *disk)
 	kfree(q->td);
 }
 
-void blk_throtl_register_queue(struct request_queue *q)
+void blk_throtl_register(struct gendisk *disk)
 {
+	struct request_queue *q = disk->queue;
 	struct throtl_data *td;
 	int i;
 
diff --git a/block/blk-throttle.h b/block/blk-throttle.h
index f75852a4e5337..7e217e613aee2 100644
--- a/block/blk-throttle.h
+++ b/block/blk-throttle.h
@@ -170,13 +170,13 @@ static inline struct throtl_grp *blkg_to_tg(struct blkcg_gq *blkg)
 #ifndef CONFIG_BLK_DEV_THROTTLING
 static inline int blk_throtl_init(struct gendisk *disk) { return 0; }
 static inline void blk_throtl_exit(struct gendisk *disk) { }
-static inline void blk_throtl_register_queue(struct request_queue *q) { }
+static inline void blk_throtl_register(struct gendisk *disk) { }
 static inline bool blk_throtl_bio(struct bio *bio) { return false; }
 static inline void blk_throtl_cancel_bios(struct request_queue *q) { }
 #else /* CONFIG_BLK_DEV_THROTTLING */
 int blk_throtl_init(struct gendisk *disk);
 void blk_throtl_exit(struct gendisk *disk);
-void blk_throtl_register_queue(struct request_queue *q);
+void blk_throtl_register(struct gendisk *disk);
 bool __blk_throtl_bio(struct bio *bio);
 void blk_throtl_cancel_bios(struct request_queue *q);
 static inline bool blk_throtl_bio(struct bio *bio)
-- 
2.30.2

Re: [PATCH 13/17] blk-throttle: pass a gendisk to blk_throtl_register_queue

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:57PM +0200, Christoph Hellwig wrote:
> Pass the gendisk to blk_throtl_register_queue as part of moving the
> blk-cgroup infrastructure to be gendisk based.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-sysfs.c    | 2 +-
>  block/blk-throttle.c | 3 ++-
>  block/blk-throttle.h | 4 ++--
>  3 files changed, 5 insertions(+), 4 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c
> index e1f009aba6fd2..e71b3b43927c0 100644
> --- a/block/blk-sysfs.c
> +++ b/block/blk-sysfs.c
> @@ -844,7 +844,7 @@ int blk_register_queue(struct gendisk *disk)
>  
>  	blk_queue_flag_set(QUEUE_FLAG_REGISTERED, q);
>  	wbt_enable_default(q);
> -	blk_throtl_register_queue(q);
> +	blk_throtl_register(disk);
>  
>  	/* Now everything is ready and send out KOBJ_ADD uevent */
>  	kobject_uevent(&q->kobj, KOBJ_ADD);
> diff --git a/block/blk-throttle.c b/block/blk-throttle.c
> index 9ca8ae0ae6683..a9709d4af6f39 100644
> --- a/block/blk-throttle.c
> +++ b/block/blk-throttle.c
> @@ -2415,8 +2415,9 @@ void blk_throtl_exit(struct gendisk *disk)
>  	kfree(q->td);
>  }
>  
> -void blk_throtl_register_queue(struct request_queue *q)
> +void blk_throtl_register(struct gendisk *disk)
>  {
> +	struct request_queue *q = disk->queue;
>  	struct throtl_data *td;
>  	int i;
>  
> diff --git a/block/blk-throttle.h b/block/blk-throttle.h
> index f75852a4e5337..7e217e613aee2 100644
> --- a/block/blk-throttle.h
> +++ b/block/blk-throttle.h
> @@ -170,13 +170,13 @@ static inline struct throtl_grp *blkg_to_tg(struct blkcg_gq *blkg)
>  #ifndef CONFIG_BLK_DEV_THROTTLING
>  static inline int blk_throtl_init(struct gendisk *disk) { return 0; }
>  static inline void blk_throtl_exit(struct gendisk *disk) { }
> -static inline void blk_throtl_register_queue(struct request_queue *q) { }
> +static inline void blk_throtl_register(struct gendisk *disk) { }
>  static inline bool blk_throtl_bio(struct bio *bio) { return false; }
>  static inline void blk_throtl_cancel_bios(struct request_queue *q) { }
>  #else /* CONFIG_BLK_DEV_THROTTLING */
>  int blk_throtl_init(struct gendisk *disk);
>  void blk_throtl_exit(struct gendisk *disk);
> -void blk_throtl_register_queue(struct request_queue *q);
> +void blk_throtl_register(struct gendisk *disk);
>  bool __blk_throtl_bio(struct bio *bio);
>  void blk_throtl_cancel_bios(struct request_queue *q);
>  static inline bool blk_throtl_bio(struct bio *bio)
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 14/17] blk-throttle: pass a gendisk to blk_throtl_cancel_bios

[Christoph Hellwig]

Pass the gendisk to blk_throtl_cancel_bios as part of moving the
blk-cgroup infrastructure to be gendisk based.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-throttle.c | 3 ++-
 block/blk-throttle.h | 4 ++--
 block/genhd.c        | 2 +-
 3 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index a9709d4af6f39..8c683d5050483 100644
--- a/block/blk-throttle.c
+++ b/block/blk-throttle.c
@@ -1723,8 +1723,9 @@ struct blkcg_policy blkcg_policy_throtl = {
 	.pd_free_fn		= throtl_pd_free,
 };
 
-void blk_throtl_cancel_bios(struct request_queue *q)
+void blk_throtl_cancel_bios(struct gendisk *disk)
 {
+	struct request_queue *q = disk->queue;
 	struct cgroup_subsys_state *pos_css;
 	struct blkcg_gq *blkg;
 
diff --git a/block/blk-throttle.h b/block/blk-throttle.h
index 7e217e613aee2..143991fd13228 100644
--- a/block/blk-throttle.h
+++ b/block/blk-throttle.h
@@ -172,13 +172,13 @@ static inline int blk_throtl_init(struct gendisk *disk) { return 0; }
 static inline void blk_throtl_exit(struct gendisk *disk) { }
 static inline void blk_throtl_register(struct gendisk *disk) { }
 static inline bool blk_throtl_bio(struct bio *bio) { return false; }
-static inline void blk_throtl_cancel_bios(struct request_queue *q) { }
+static inline void blk_throtl_cancel_bios(struct gendisk *disk) { }
 #else /* CONFIG_BLK_DEV_THROTTLING */
 int blk_throtl_init(struct gendisk *disk);
 void blk_throtl_exit(struct gendisk *disk);
 void blk_throtl_register(struct gendisk *disk);
 bool __blk_throtl_bio(struct bio *bio);
-void blk_throtl_cancel_bios(struct request_queue *q);
+void blk_throtl_cancel_bios(struct gendisk *disk);
 static inline bool blk_throtl_bio(struct bio *bio)
 {
 	struct throtl_grp *tg = blkg_to_tg(bio->bi_blkg);
diff --git a/block/genhd.c b/block/genhd.c
index f1af045fac2fe..d6a21803a57e2 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -626,7 +626,7 @@ void del_gendisk(struct gendisk *disk)
 	pm_runtime_set_memalloc_noio(disk_to_dev(disk), false);
 	device_del(disk_to_dev(disk));
 
-	blk_throtl_cancel_bios(disk->queue);
+	blk_throtl_cancel_bios(disk);
 
 	blk_sync_queue(q);
 	blk_flush_integrity();
-- 
2.30.2

Re: [PATCH 14/17] blk-throttle: pass a gendisk to blk_throtl_cancel_bios

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:58PM +0200, Christoph Hellwig wrote:
> Pass the gendisk to blk_throtl_cancel_bios as part of moving the
> blk-cgroup infrastructure to be gendisk based.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-throttle.c | 3 ++-
>  block/blk-throttle.h | 4 ++--
>  block/genhd.c        | 2 +-
>  3 files changed, 5 insertions(+), 4 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-throttle.c b/block/blk-throttle.c
> index a9709d4af6f39..8c683d5050483 100644
> --- a/block/blk-throttle.c
> +++ b/block/blk-throttle.c
> @@ -1723,8 +1723,9 @@ struct blkcg_policy blkcg_policy_throtl = {
>  	.pd_free_fn		= throtl_pd_free,
>  };
>  
> -void blk_throtl_cancel_bios(struct request_queue *q)
> +void blk_throtl_cancel_bios(struct gendisk *disk)
>  {
> +	struct request_queue *q = disk->queue;
>  	struct cgroup_subsys_state *pos_css;
>  	struct blkcg_gq *blkg;
>  
> diff --git a/block/blk-throttle.h b/block/blk-throttle.h
> index 7e217e613aee2..143991fd13228 100644
> --- a/block/blk-throttle.h
> +++ b/block/blk-throttle.h
> @@ -172,13 +172,13 @@ static inline int blk_throtl_init(struct gendisk *disk) { return 0; }
>  static inline void blk_throtl_exit(struct gendisk *disk) { }
>  static inline void blk_throtl_register(struct gendisk *disk) { }
>  static inline bool blk_throtl_bio(struct bio *bio) { return false; }
> -static inline void blk_throtl_cancel_bios(struct request_queue *q) { }
> +static inline void blk_throtl_cancel_bios(struct gendisk *disk) { }
>  #else /* CONFIG_BLK_DEV_THROTTLING */
>  int blk_throtl_init(struct gendisk *disk);
>  void blk_throtl_exit(struct gendisk *disk);
>  void blk_throtl_register(struct gendisk *disk);
>  bool __blk_throtl_bio(struct bio *bio);
> -void blk_throtl_cancel_bios(struct request_queue *q);
> +void blk_throtl_cancel_bios(struct gendisk *disk);
>  static inline bool blk_throtl_bio(struct bio *bio)
>  {
>  	struct throtl_grp *tg = blkg_to_tg(bio->bi_blkg);
> diff --git a/block/genhd.c b/block/genhd.c
> index f1af045fac2fe..d6a21803a57e2 100644
> --- a/block/genhd.c
> +++ b/block/genhd.c
> @@ -626,7 +626,7 @@ void del_gendisk(struct gendisk *disk)
>  	pm_runtime_set_memalloc_noio(disk_to_dev(disk), false);
>  	device_del(disk_to_dev(disk));
>  
> -	blk_throtl_cancel_bios(disk->queue);
> +	blk_throtl_cancel_bios(disk);
>  
>  	blk_sync_queue(q);
>  	blk_flush_integrity();
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 15/17] blk-cgroup: pass a gendisk to blkg_destroy_all

[Christoph Hellwig]

Pass the gendisk to blkg_destroy_all as part of moving the blk-cgroup
infrastructure to be gendisk based.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c | 13 ++++---------
 1 file changed, 4 insertions(+), 9 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 3dfd78f1312db..c2d5ca2eb92e5 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -462,14 +462,9 @@ static void blkg_destroy(struct blkcg_gq *blkg)
 	percpu_ref_kill(&blkg->refcnt);
 }
 
-/**
- * blkg_destroy_all - destroy all blkgs associated with a request_queue
- * @q: request_queue of interest
- *
- * Destroy all blkgs associated with @q.
- */
-static void blkg_destroy_all(struct request_queue *q)
+static void blkg_destroy_all(struct gendisk *disk)
 {
+	struct request_queue *q = disk->queue;
 	struct blkcg_gq *blkg, *n;
 	int count = BLKG_DESTROY_BATCH_SIZE;
 
@@ -1276,7 +1271,7 @@ int blkcg_init_disk(struct gendisk *disk)
 err_ioprio_exit:
 	blk_ioprio_exit(disk);
 err_destroy_all:
-	blkg_destroy_all(q);
+	blkg_destroy_all(disk);
 	return ret;
 err_unlock:
 	spin_unlock_irq(&q->queue_lock);
@@ -1287,7 +1282,7 @@ int blkcg_init_disk(struct gendisk *disk)
 
 void blkcg_exit_disk(struct gendisk *disk)
 {
-	blkg_destroy_all(disk->queue);
+	blkg_destroy_all(disk);
 	blk_throtl_exit(disk);
 }
 
-- 
2.30.2

Re: [PATCH 15/17] blk-cgroup: pass a gendisk to blkg_destroy_all

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:04:59PM +0200, Christoph Hellwig wrote:
> Pass the gendisk to blkg_destroy_all as part of moving the blk-cgroup
> infrastructure to be gendisk based.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c | 13 ++++---------
>  1 file changed, 4 insertions(+), 9 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index 3dfd78f1312db..c2d5ca2eb92e5 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -462,14 +462,9 @@ static void blkg_destroy(struct blkcg_gq *blkg)
>  	percpu_ref_kill(&blkg->refcnt);
>  }
>  
> -/**
> - * blkg_destroy_all - destroy all blkgs associated with a request_queue
> - * @q: request_queue of interest
> - *
> - * Destroy all blkgs associated with @q.
> - */
> -static void blkg_destroy_all(struct request_queue *q)
> +static void blkg_destroy_all(struct gendisk *disk)
>  {
> +	struct request_queue *q = disk->queue;
>  	struct blkcg_gq *blkg, *n;
>  	int count = BLKG_DESTROY_BATCH_SIZE;
>  
> @@ -1276,7 +1271,7 @@ int blkcg_init_disk(struct gendisk *disk)
>  err_ioprio_exit:
>  	blk_ioprio_exit(disk);
>  err_destroy_all:
> -	blkg_destroy_all(q);
> +	blkg_destroy_all(disk);
>  	return ret;
>  err_unlock:
>  	spin_unlock_irq(&q->queue_lock);
> @@ -1287,7 +1282,7 @@ int blkcg_init_disk(struct gendisk *disk)
>  
>  void blkcg_exit_disk(struct gendisk *disk)
>  {
> -	blkg_destroy_all(disk->queue);
> +	blkg_destroy_all(disk);
>  	blk_throtl_exit(disk);
>  }
>  
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 16/17] blk-cgroup: pass a gendisk to blkcg_schedule_throttle

[Christoph Hellwig]

Pass the gendisk to blkcg_schedule_throttle as part of moving the
blk-cgroup infrastructure to be gendisk based.  Remove the unused
!BLK_CGROUP stub while we're at it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c         | 8 +++++---
 block/blk-iocost.c         | 4 ++--
 block/blk-iolatency.c      | 2 +-
 include/linux/blk-cgroup.h | 5 ++---
 mm/swapfile.c              | 2 +-
 5 files changed, 11 insertions(+), 10 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index c2d5ca2eb92e5..fc82057db9629 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -1792,13 +1792,13 @@ void blkcg_maybe_throttle_current(void)
 
 /**
  * blkcg_schedule_throttle - this task needs to check for throttling
- * @q: the request queue IO was submitted on
+ * @gendisk: disk to throttle
  * @use_memdelay: do we charge this to memory delay for PSI
  *
  * This is called by the IO controller when we know there's delay accumulated
  * for the blkg for this task.  We do not pass the blkg because there are places
  * we call this that may not have that information, the swapping code for
- * instance will only have a request_queue at that point.  This set's the
+ * instance will only have a block_device at that point.  This set's the
  * notify_resume for the task to check and see if it requires throttling before
  * returning to user space.
  *
@@ -1807,8 +1807,10 @@ void blkcg_maybe_throttle_current(void)
  * throttle once.  If the task needs to be throttled again it'll need to be
  * re-set at the next time we see the task.
  */
-void blkcg_schedule_throttle(struct request_queue *q, bool use_memdelay)
+void blkcg_schedule_throttle(struct gendisk *disk, bool use_memdelay)
 {
+	struct request_queue *q = disk->queue;
+
 	if (unlikely(current->flags & PF_KTHREAD))
 		return;
 
diff --git a/block/blk-iocost.c b/block/blk-iocost.c
index b8e5f550aa5be..b0899ab214c41 100644
--- a/block/blk-iocost.c
+++ b/block/blk-iocost.c
@@ -2636,7 +2636,7 @@ static void ioc_rqos_throttle(struct rq_qos *rqos, struct bio *bio)
 	if (use_debt) {
 		iocg_incur_debt(iocg, abs_cost, &now);
 		if (iocg_kick_delay(iocg, &now))
-			blkcg_schedule_throttle(rqos->q,
+			blkcg_schedule_throttle(rqos->q->disk,
 					(bio->bi_opf & REQ_SWAP) == REQ_SWAP);
 		iocg_unlock(iocg, ioc_locked, &flags);
 		return;
@@ -2737,7 +2737,7 @@ static void ioc_rqos_merge(struct rq_qos *rqos, struct request *rq,
 	if (likely(!list_empty(&iocg->active_list))) {
 		iocg_incur_debt(iocg, abs_cost, &now);
 		if (iocg_kick_delay(iocg, &now))
-			blkcg_schedule_throttle(rqos->q,
+			blkcg_schedule_throttle(rqos->q->disk,
 					(bio->bi_opf & REQ_SWAP) == REQ_SWAP);
 	} else {
 		iocg_commit_bio(iocg, bio, abs_cost, cost);
diff --git a/block/blk-iolatency.c b/block/blk-iolatency.c
index c6f61fe88b875..571fa95aafe96 100644
--- a/block/blk-iolatency.c
+++ b/block/blk-iolatency.c
@@ -292,7 +292,7 @@ static void __blkcg_iolatency_throttle(struct rq_qos *rqos,
 	unsigned use_delay = atomic_read(&lat_to_blkg(iolat)->use_delay);
 
 	if (use_delay)
-		blkcg_schedule_throttle(rqos->q, use_memdelay);
+		blkcg_schedule_throttle(rqos->q->disk, use_memdelay);
 
 	/*
 	 * To avoid priority inversions we want to just take a slot if we are
diff --git a/include/linux/blk-cgroup.h b/include/linux/blk-cgroup.h
index 9f40dbc65f82c..dd5841a42c331 100644
--- a/include/linux/blk-cgroup.h
+++ b/include/linux/blk-cgroup.h
@@ -18,14 +18,14 @@
 
 struct bio;
 struct cgroup_subsys_state;
-struct request_queue;
+struct gendisk;
 
 #define FC_APPID_LEN              129
 
 #ifdef CONFIG_BLK_CGROUP
 extern struct cgroup_subsys_state * const blkcg_root_css;
 
-void blkcg_schedule_throttle(struct request_queue *q, bool use_memdelay);
+void blkcg_schedule_throttle(struct gendisk *disk, bool use_memdelay);
 void blkcg_maybe_throttle_current(void);
 bool blk_cgroup_congested(void);
 void blkcg_pin_online(struct cgroup_subsys_state *blkcg_css);
@@ -39,7 +39,6 @@ struct cgroup_subsys_state *bio_blkcg_css(struct bio *bio);
 
 static inline void blkcg_maybe_throttle_current(void) { }
 static inline bool blk_cgroup_congested(void) { return false; }
-static inline void blkcg_schedule_throttle(struct request_queue *q, bool use_memdelay) { }
 static inline struct cgroup_subsys_state *bio_blkcg_css(struct bio *bio)
 {
 	return NULL;
diff --git a/mm/swapfile.c b/mm/swapfile.c
index 1fdccd2f1422e..82e62007881db 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -3655,7 +3655,7 @@ void __cgroup_throttle_swaprate(struct page *page, gfp_t gfp_mask)
 	plist_for_each_entry_safe(si, next, &swap_avail_heads[nid],
 				  avail_lists[nid]) {
 		if (si->bdev) {
-			blkcg_schedule_throttle(bdev_get_queue(si->bdev), true);
+			blkcg_schedule_throttle(si->bdev->bd_disk, true);
 			break;
 		}
 	}
-- 
2.30.2

Re: [PATCH 16/17] blk-cgroup: pass a gendisk to blkcg_schedule_throttle

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:05:00PM +0200, Christoph Hellwig wrote:
> Pass the gendisk to blkcg_schedule_throttle as part of moving the
> blk-cgroup infrastructure to be gendisk based.  Remove the unused
> !BLK_CGROUP stub while we're at it.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c         | 8 +++++---
>  block/blk-iocost.c         | 4 ++--
>  block/blk-iolatency.c      | 2 +-
>  include/linux/blk-cgroup.h | 5 ++---
>  mm/swapfile.c              | 2 +-
>  5 files changed, 11 insertions(+), 10 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index c2d5ca2eb92e5..fc82057db9629 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -1792,13 +1792,13 @@ void blkcg_maybe_throttle_current(void)
>  
>  /**
>   * blkcg_schedule_throttle - this task needs to check for throttling
> - * @q: the request queue IO was submitted on
> + * @gendisk: disk to throttle
>   * @use_memdelay: do we charge this to memory delay for PSI
>   *
>   * This is called by the IO controller when we know there's delay accumulated
>   * for the blkg for this task.  We do not pass the blkg because there are places
>   * we call this that may not have that information, the swapping code for
> - * instance will only have a request_queue at that point.  This set's the
> + * instance will only have a block_device at that point.  This set's the
>   * notify_resume for the task to check and see if it requires throttling before
>   * returning to user space.
>   *
> @@ -1807,8 +1807,10 @@ void blkcg_maybe_throttle_current(void)
>   * throttle once.  If the task needs to be throttled again it'll need to be
>   * re-set at the next time we see the task.
>   */
> -void blkcg_schedule_throttle(struct request_queue *q, bool use_memdelay)
> +void blkcg_schedule_throttle(struct gendisk *disk, bool use_memdelay)
>  {
> +	struct request_queue *q = disk->queue;
> +
>  	if (unlikely(current->flags & PF_KTHREAD))
>  		return;
>  
> diff --git a/block/blk-iocost.c b/block/blk-iocost.c
> index b8e5f550aa5be..b0899ab214c41 100644
> --- a/block/blk-iocost.c
> +++ b/block/blk-iocost.c
> @@ -2636,7 +2636,7 @@ static void ioc_rqos_throttle(struct rq_qos *rqos, struct bio *bio)
>  	if (use_debt) {
>  		iocg_incur_debt(iocg, abs_cost, &now);
>  		if (iocg_kick_delay(iocg, &now))
> -			blkcg_schedule_throttle(rqos->q,
> +			blkcg_schedule_throttle(rqos->q->disk,
>  					(bio->bi_opf & REQ_SWAP) == REQ_SWAP);
>  		iocg_unlock(iocg, ioc_locked, &flags);
>  		return;
> @@ -2737,7 +2737,7 @@ static void ioc_rqos_merge(struct rq_qos *rqos, struct request *rq,
>  	if (likely(!list_empty(&iocg->active_list))) {
>  		iocg_incur_debt(iocg, abs_cost, &now);
>  		if (iocg_kick_delay(iocg, &now))
> -			blkcg_schedule_throttle(rqos->q,
> +			blkcg_schedule_throttle(rqos->q->disk,
>  					(bio->bi_opf & REQ_SWAP) == REQ_SWAP);
>  	} else {
>  		iocg_commit_bio(iocg, bio, abs_cost, cost);
> diff --git a/block/blk-iolatency.c b/block/blk-iolatency.c
> index c6f61fe88b875..571fa95aafe96 100644
> --- a/block/blk-iolatency.c
> +++ b/block/blk-iolatency.c
> @@ -292,7 +292,7 @@ static void __blkcg_iolatency_throttle(struct rq_qos *rqos,
>  	unsigned use_delay = atomic_read(&lat_to_blkg(iolat)->use_delay);
>  
>  	if (use_delay)
> -		blkcg_schedule_throttle(rqos->q, use_memdelay);
> +		blkcg_schedule_throttle(rqos->q->disk, use_memdelay);
>  
>  	/*
>  	 * To avoid priority inversions we want to just take a slot if we are
> diff --git a/include/linux/blk-cgroup.h b/include/linux/blk-cgroup.h
> index 9f40dbc65f82c..dd5841a42c331 100644
> --- a/include/linux/blk-cgroup.h
> +++ b/include/linux/blk-cgroup.h
> @@ -18,14 +18,14 @@
>  
>  struct bio;
>  struct cgroup_subsys_state;
> -struct request_queue;
> +struct gendisk;
>  
>  #define FC_APPID_LEN              129
>  
>  #ifdef CONFIG_BLK_CGROUP
>  extern struct cgroup_subsys_state * const blkcg_root_css;
>  
> -void blkcg_schedule_throttle(struct request_queue *q, bool use_memdelay);
> +void blkcg_schedule_throttle(struct gendisk *disk, bool use_memdelay);
>  void blkcg_maybe_throttle_current(void);
>  bool blk_cgroup_congested(void);
>  void blkcg_pin_online(struct cgroup_subsys_state *blkcg_css);
> @@ -39,7 +39,6 @@ struct cgroup_subsys_state *bio_blkcg_css(struct bio *bio);
>  
>  static inline void blkcg_maybe_throttle_current(void) { }
>  static inline bool blk_cgroup_congested(void) { return false; }
> -static inline void blkcg_schedule_throttle(struct request_queue *q, bool use_memdelay) { }
>  static inline struct cgroup_subsys_state *bio_blkcg_css(struct bio *bio)
>  {
>  	return NULL;
> diff --git a/mm/swapfile.c b/mm/swapfile.c
> index 1fdccd2f1422e..82e62007881db 100644
> --- a/mm/swapfile.c
> +++ b/mm/swapfile.c
> @@ -3655,7 +3655,7 @@ void __cgroup_throttle_swaprate(struct page *page, gfp_t gfp_mask)
>  	plist_for_each_entry_safe(si, next, &swap_avail_heads[nid],
>  				  avail_lists[nid]) {
>  		if (si->bdev) {
> -			blkcg_schedule_throttle(bdev_get_queue(si->bdev), true);
> +			blkcg_schedule_throttle(si->bdev->bd_disk, true);
>  			break;
>  		}
>  	}
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

[PATCH 17/17] blk-cgroup: pass a gendisk to the blkg allocation helpers

[Christoph Hellwig]

Prepare for storing the blkcg information in the gendisk instead of
the request_queue.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-cgroup.c | 56 +++++++++++++++++++++++-----------------------
 1 file changed, 28 insertions(+), 28 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index fc82057db9629..94af5f3f3620b 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -202,19 +202,19 @@ static inline struct blkcg *blkcg_parent(struct blkcg *blkcg)
 /**
  * blkg_alloc - allocate a blkg
  * @blkcg: block cgroup the new blkg is associated with
- * @q: request_queue the new blkg is associated with
+ * @disk: gendisk the new blkg is associated with
  * @gfp_mask: allocation mask to use
  *
  * Allocate a new blkg assocating @blkcg and @q.
  */
-static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct request_queue *q,
+static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct gendisk *disk,
 				   gfp_t gfp_mask)
 {
 	struct blkcg_gq *blkg;
 	int i, cpu;
 
 	/* alloc and init base part */
-	blkg = kzalloc_node(sizeof(*blkg), gfp_mask, q->node);
+	blkg = kzalloc_node(sizeof(*blkg), gfp_mask, disk->queue->node);
 	if (!blkg)
 		return NULL;
 
@@ -225,10 +225,10 @@ static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct request_queue *q,
 	if (!blkg->iostat_cpu)
 		goto err_free;
 
-	if (!blk_get_queue(q))
+	if (!blk_get_queue(disk->queue))
 		goto err_free;
 
-	blkg->q = q;
+	blkg->q = disk->queue;
 	INIT_LIST_HEAD(&blkg->q_node);
 	spin_lock_init(&blkg->async_bio_lock);
 	bio_list_init(&blkg->async_bios);
@@ -243,11 +243,11 @@ static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct request_queue *q,
 		struct blkcg_policy *pol = blkcg_policy[i];
 		struct blkg_policy_data *pd;
 
-		if (!blkcg_policy_enabled(q, pol))
+		if (!blkcg_policy_enabled(disk->queue, pol))
 			continue;
 
 		/* alloc per-policy data and attach it to blkg */
-		pd = pol->pd_alloc_fn(gfp_mask, q, blkcg);
+		pd = pol->pd_alloc_fn(gfp_mask, disk->queue, blkcg);
 		if (!pd)
 			goto err_free;
 
@@ -275,17 +275,16 @@ static void blkg_update_hint(struct blkcg *blkcg, struct blkcg_gq *blkg)
  * If @new_blkg is %NULL, this function tries to allocate a new one as
  * necessary using %GFP_NOWAIT.  @new_blkg is always consumed on return.
  */
-static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
-				    struct request_queue *q,
+static struct blkcg_gq *blkg_create(struct blkcg *blkcg, struct gendisk *disk,
 				    struct blkcg_gq *new_blkg)
 {
 	struct blkcg_gq *blkg;
 	int i, ret;
 
-	lockdep_assert_held(&q->queue_lock);
+	lockdep_assert_held(&disk->queue->queue_lock);
 
 	/* request_queue is dying, do not create/recreate a blkg */
-	if (blk_queue_dying(q)) {
+	if (blk_queue_dying(disk->queue)) {
 		ret = -ENODEV;
 		goto err_free_blkg;
 	}
@@ -298,7 +297,7 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
 
 	/* allocate */
 	if (!new_blkg) {
-		new_blkg = blkg_alloc(blkcg, q, GFP_NOWAIT | __GFP_NOWARN);
+		new_blkg = blkg_alloc(blkcg, disk, GFP_NOWAIT | __GFP_NOWARN);
 		if (unlikely(!new_blkg)) {
 			ret = -ENOMEM;
 			goto err_put_css;
@@ -308,7 +307,7 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
 
 	/* link parent */
 	if (blkcg_parent(blkcg)) {
-		blkg->parent = blkg_lookup(blkcg_parent(blkcg), q);
+		blkg->parent = blkg_lookup(blkcg_parent(blkcg), disk->queue);
 		if (WARN_ON_ONCE(!blkg->parent)) {
 			ret = -ENODEV;
 			goto err_put_css;
@@ -326,10 +325,10 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
 
 	/* insert */
 	spin_lock(&blkcg->lock);
-	ret = radix_tree_insert(&blkcg->blkg_tree, q->id, blkg);
+	ret = radix_tree_insert(&blkcg->blkg_tree, disk->queue->id, blkg);
 	if (likely(!ret)) {
 		hlist_add_head_rcu(&blkg->blkcg_node, &blkcg->blkg_list);
-		list_add(&blkg->q_node, &q->blkg_list);
+		list_add(&blkg->q_node, &disk->queue->blkg_list);
 
 		for (i = 0; i < BLKCG_MAX_POLS; i++) {
 			struct blkcg_policy *pol = blkcg_policy[i];
@@ -358,19 +357,20 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
 /**
  * blkg_lookup_create - lookup blkg, try to create one if not there
  * @blkcg: blkcg of interest
- * @q: request_queue of interest
+ * @disk: gendisk of interest
  *
- * Lookup blkg for the @blkcg - @q pair.  If it doesn't exist, try to
+ * Lookup blkg for the @blkcg - @disk pair.  If it doesn't exist, try to
  * create one.  blkg creation is performed recursively from blkcg_root such
  * that all non-root blkg's have access to the parent blkg.  This function
- * should be called under RCU read lock and takes @q->queue_lock.
+ * should be called under RCU read lock and takes @disk->queue->queue_lock.
  *
  * Returns the blkg or the closest blkg if blkg_create() fails as it walks
  * down from root.
  */
 static struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
-		struct request_queue *q)
+		struct gendisk *disk)
 {
+	struct request_queue *q = disk->queue;
 	struct blkcg_gq *blkg;
 	unsigned long flags;
 
@@ -408,7 +408,7 @@ static struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
 			parent = blkcg_parent(parent);
 		}
 
-		blkg = blkg_create(pos, q, NULL);
+		blkg = blkg_create(pos, disk, NULL);
 		if (IS_ERR(blkg)) {
 			blkg = ret_blkg;
 			break;
@@ -652,6 +652,7 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
 	__acquires(rcu) __acquires(&bdev->bd_queue->queue_lock)
 {
 	struct block_device *bdev;
+	struct gendisk *disk;
 	struct request_queue *q;
 	struct blkcg_gq *blkg;
 	int ret;
@@ -659,8 +660,8 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
 	bdev = blkcg_conf_open_bdev(&input);
 	if (IS_ERR(bdev))
 		return PTR_ERR(bdev);
-
-	q = bdev_get_queue(bdev);
+	disk = bdev->bd_disk;
+	q = disk->queue;
 
 	/*
 	 * blkcg_deactivate_policy() requires queue to be frozen, we can grab
@@ -703,7 +704,7 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
 		spin_unlock_irq(&q->queue_lock);
 		rcu_read_unlock();
 
-		new_blkg = blkg_alloc(pos, q, GFP_KERNEL);
+		new_blkg = blkg_alloc(pos, disk, GFP_KERNEL);
 		if (unlikely(!new_blkg)) {
 			ret = -ENOMEM;
 			goto fail_exit_queue;
@@ -729,7 +730,7 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
 			blkg_update_hint(pos, blkg);
 			blkg_free(new_blkg);
 		} else {
-			blkg = blkg_create(pos, q, new_blkg);
+			blkg = blkg_create(pos, disk, new_blkg);
 			if (IS_ERR(blkg)) {
 				ret = PTR_ERR(blkg);
 				goto fail_preloaded;
@@ -1234,7 +1235,7 @@ int blkcg_init_disk(struct gendisk *disk)
 
 	INIT_LIST_HEAD(&q->blkg_list);
 
-	new_blkg = blkg_alloc(&blkcg_root, q, GFP_KERNEL);
+	new_blkg = blkg_alloc(&blkcg_root, disk, GFP_KERNEL);
 	if (!new_blkg)
 		return -ENOMEM;
 
@@ -1243,7 +1244,7 @@ int blkcg_init_disk(struct gendisk *disk)
 	/* Make sure the root blkg exists. */
 	/* spin_lock_irq can serve as RCU read-side critical section. */
 	spin_lock_irq(&q->queue_lock);
-	blkg = blkg_create(&blkcg_root, q, new_blkg);
+	blkg = blkg_create(&blkcg_root, disk, new_blkg);
 	if (IS_ERR(blkg))
 		goto err_unlock;
 	q->root_blkg = blkg;
@@ -1860,8 +1861,7 @@ static inline struct blkcg_gq *blkg_tryget_closest(struct bio *bio,
 	struct blkcg_gq *blkg, *ret_blkg = NULL;
 
 	rcu_read_lock();
-	blkg = blkg_lookup_create(css_to_blkcg(css),
-				  bdev_get_queue(bio->bi_bdev));
+	blkg = blkg_lookup_create(css_to_blkcg(css), bio->bi_bdev->bd_disk);
 	while (blkg) {
 		if (blkg_tryget(blkg)) {
 			ret_blkg = blkg;
-- 
2.30.2

Re: [PATCH 17/17] blk-cgroup: pass a gendisk to the blkg allocation helpers

[Andreas Herrmann]

On Wed, Sep 21, 2022 at 08:05:01PM +0200, Christoph Hellwig wrote:
> Prepare for storing the blkcg information in the gendisk instead of
> the request_queue.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  block/blk-cgroup.c | 56 +++++++++++++++++++++++-----------------------
>  1 file changed, 28 insertions(+), 28 deletions(-)

Reviewed-by: Andreas Herrmann <aherrmann@suse.de>

> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index fc82057db9629..94af5f3f3620b 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -202,19 +202,19 @@ static inline struct blkcg *blkcg_parent(struct blkcg *blkcg)
>  /**
>   * blkg_alloc - allocate a blkg
>   * @blkcg: block cgroup the new blkg is associated with
> - * @q: request_queue the new blkg is associated with
> + * @disk: gendisk the new blkg is associated with
>   * @gfp_mask: allocation mask to use
>   *
>   * Allocate a new blkg assocating @blkcg and @q.
>   */
> -static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct request_queue *q,
> +static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct gendisk *disk,
>  				   gfp_t gfp_mask)
>  {
>  	struct blkcg_gq *blkg;
>  	int i, cpu;
>  
>  	/* alloc and init base part */
> -	blkg = kzalloc_node(sizeof(*blkg), gfp_mask, q->node);
> +	blkg = kzalloc_node(sizeof(*blkg), gfp_mask, disk->queue->node);
>  	if (!blkg)
>  		return NULL;
>  
> @@ -225,10 +225,10 @@ static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct request_queue *q,
>  	if (!blkg->iostat_cpu)
>  		goto err_free;
>  
> -	if (!blk_get_queue(q))
> +	if (!blk_get_queue(disk->queue))
>  		goto err_free;
>  
> -	blkg->q = q;
> +	blkg->q = disk->queue;
>  	INIT_LIST_HEAD(&blkg->q_node);
>  	spin_lock_init(&blkg->async_bio_lock);
>  	bio_list_init(&blkg->async_bios);
> @@ -243,11 +243,11 @@ static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct request_queue *q,
>  		struct blkcg_policy *pol = blkcg_policy[i];
>  		struct blkg_policy_data *pd;
>  
> -		if (!blkcg_policy_enabled(q, pol))
> +		if (!blkcg_policy_enabled(disk->queue, pol))
>  			continue;
>  
>  		/* alloc per-policy data and attach it to blkg */
> -		pd = pol->pd_alloc_fn(gfp_mask, q, blkcg);
> +		pd = pol->pd_alloc_fn(gfp_mask, disk->queue, blkcg);
>  		if (!pd)
>  			goto err_free;
>  
> @@ -275,17 +275,16 @@ static void blkg_update_hint(struct blkcg *blkcg, struct blkcg_gq *blkg)
>   * If @new_blkg is %NULL, this function tries to allocate a new one as
>   * necessary using %GFP_NOWAIT.  @new_blkg is always consumed on return.
>   */
> -static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
> -				    struct request_queue *q,
> +static struct blkcg_gq *blkg_create(struct blkcg *blkcg, struct gendisk *disk,
>  				    struct blkcg_gq *new_blkg)
>  {
>  	struct blkcg_gq *blkg;
>  	int i, ret;
>  
> -	lockdep_assert_held(&q->queue_lock);
> +	lockdep_assert_held(&disk->queue->queue_lock);
>  
>  	/* request_queue is dying, do not create/recreate a blkg */
> -	if (blk_queue_dying(q)) {
> +	if (blk_queue_dying(disk->queue)) {
>  		ret = -ENODEV;
>  		goto err_free_blkg;
>  	}
> @@ -298,7 +297,7 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
>  
>  	/* allocate */
>  	if (!new_blkg) {
> -		new_blkg = blkg_alloc(blkcg, q, GFP_NOWAIT | __GFP_NOWARN);
> +		new_blkg = blkg_alloc(blkcg, disk, GFP_NOWAIT | __GFP_NOWARN);
>  		if (unlikely(!new_blkg)) {
>  			ret = -ENOMEM;
>  			goto err_put_css;
> @@ -308,7 +307,7 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
>  
>  	/* link parent */
>  	if (blkcg_parent(blkcg)) {
> -		blkg->parent = blkg_lookup(blkcg_parent(blkcg), q);
> +		blkg->parent = blkg_lookup(blkcg_parent(blkcg), disk->queue);
>  		if (WARN_ON_ONCE(!blkg->parent)) {
>  			ret = -ENODEV;
>  			goto err_put_css;
> @@ -326,10 +325,10 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
>  
>  	/* insert */
>  	spin_lock(&blkcg->lock);
> -	ret = radix_tree_insert(&blkcg->blkg_tree, q->id, blkg);
> +	ret = radix_tree_insert(&blkcg->blkg_tree, disk->queue->id, blkg);
>  	if (likely(!ret)) {
>  		hlist_add_head_rcu(&blkg->blkcg_node, &blkcg->blkg_list);
> -		list_add(&blkg->q_node, &q->blkg_list);
> +		list_add(&blkg->q_node, &disk->queue->blkg_list);
>  
>  		for (i = 0; i < BLKCG_MAX_POLS; i++) {
>  			struct blkcg_policy *pol = blkcg_policy[i];
> @@ -358,19 +357,20 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
>  /**
>   * blkg_lookup_create - lookup blkg, try to create one if not there
>   * @blkcg: blkcg of interest
> - * @q: request_queue of interest
> + * @disk: gendisk of interest
>   *
> - * Lookup blkg for the @blkcg - @q pair.  If it doesn't exist, try to
> + * Lookup blkg for the @blkcg - @disk pair.  If it doesn't exist, try to
>   * create one.  blkg creation is performed recursively from blkcg_root such
>   * that all non-root blkg's have access to the parent blkg.  This function
> - * should be called under RCU read lock and takes @q->queue_lock.
> + * should be called under RCU read lock and takes @disk->queue->queue_lock.
>   *
>   * Returns the blkg or the closest blkg if blkg_create() fails as it walks
>   * down from root.
>   */
>  static struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
> -		struct request_queue *q)
> +		struct gendisk *disk)
>  {
> +	struct request_queue *q = disk->queue;
>  	struct blkcg_gq *blkg;
>  	unsigned long flags;
>  
> @@ -408,7 +408,7 @@ static struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
>  			parent = blkcg_parent(parent);
>  		}
>  
> -		blkg = blkg_create(pos, q, NULL);
> +		blkg = blkg_create(pos, disk, NULL);
>  		if (IS_ERR(blkg)) {
>  			blkg = ret_blkg;
>  			break;
> @@ -652,6 +652,7 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
>  	__acquires(rcu) __acquires(&bdev->bd_queue->queue_lock)
>  {
>  	struct block_device *bdev;
> +	struct gendisk *disk;
>  	struct request_queue *q;
>  	struct blkcg_gq *blkg;
>  	int ret;
> @@ -659,8 +660,8 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
>  	bdev = blkcg_conf_open_bdev(&input);
>  	if (IS_ERR(bdev))
>  		return PTR_ERR(bdev);
> -
> -	q = bdev_get_queue(bdev);
> +	disk = bdev->bd_disk;
> +	q = disk->queue;
>  
>  	/*
>  	 * blkcg_deactivate_policy() requires queue to be frozen, we can grab
> @@ -703,7 +704,7 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
>  		spin_unlock_irq(&q->queue_lock);
>  		rcu_read_unlock();
>  
> -		new_blkg = blkg_alloc(pos, q, GFP_KERNEL);
> +		new_blkg = blkg_alloc(pos, disk, GFP_KERNEL);
>  		if (unlikely(!new_blkg)) {
>  			ret = -ENOMEM;
>  			goto fail_exit_queue;
> @@ -729,7 +730,7 @@ int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
>  			blkg_update_hint(pos, blkg);
>  			blkg_free(new_blkg);
>  		} else {
> -			blkg = blkg_create(pos, q, new_blkg);
> +			blkg = blkg_create(pos, disk, new_blkg);
>  			if (IS_ERR(blkg)) {
>  				ret = PTR_ERR(blkg);
>  				goto fail_preloaded;
> @@ -1234,7 +1235,7 @@ int blkcg_init_disk(struct gendisk *disk)
>  
>  	INIT_LIST_HEAD(&q->blkg_list);
>  
> -	new_blkg = blkg_alloc(&blkcg_root, q, GFP_KERNEL);
> +	new_blkg = blkg_alloc(&blkcg_root, disk, GFP_KERNEL);
>  	if (!new_blkg)
>  		return -ENOMEM;
>  
> @@ -1243,7 +1244,7 @@ int blkcg_init_disk(struct gendisk *disk)
>  	/* Make sure the root blkg exists. */
>  	/* spin_lock_irq can serve as RCU read-side critical section. */
>  	spin_lock_irq(&q->queue_lock);
> -	blkg = blkg_create(&blkcg_root, q, new_blkg);
> +	blkg = blkg_create(&blkcg_root, disk, new_blkg);
>  	if (IS_ERR(blkg))
>  		goto err_unlock;
>  	q->root_blkg = blkg;
> @@ -1860,8 +1861,7 @@ static inline struct blkcg_gq *blkg_tryget_closest(struct bio *bio,
>  	struct blkcg_gq *blkg, *ret_blkg = NULL;
>  
>  	rcu_read_lock();
> -	blkg = blkg_lookup_create(css_to_blkcg(css),
> -				  bdev_get_queue(bio->bi_bdev));
> +	blkg = blkg_lookup_create(css_to_blkcg(css), bio->bi_bdev->bd_disk);
>  	while (blkg) {
>  		if (blkg_tryget(blkg)) {
>  			ret_blkg = blkg;
> -- 
> 2.30.2
> 

-- 
Regards,
Andreas

SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nürnberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

Re: blk-cgroup cleanups

[Tejun Heo]

Hello,

On Wed, Sep 21, 2022 at 08:04:44PM +0200, Christoph Hellwig wrote:
> Hi Tejun and Jens,
> 
> this series has a bunch of blk-cgroup cleanups and preparation
> for preparing to make blk-cgroup gendisk based.  Another series
> for the next merge window will follow for the real changes that
> include refcounting updates.

I only had a minor comment on patch 5 which can be addressed later. For the
whole series,

Acked-by: Tejun Heo <tj@kernel.org>

Thanks.

-- 
tejun

Re: blk-cgroup cleanups

[Jens Axboe]

On Wed, 21 Sep 2022 20:04:44 +0200, Christoph Hellwig wrote:
> this series has a bunch of blk-cgroup cleanups and preparation
> for preparing to make blk-cgroup gendisk based.  Another series
> for the next merge window will follow for the real changes that
> include refcounting updates.
> 
> Diffstat:
>  block/blk-cgroup.c         |  186 +++++++++++++++++----------------------------
>  block/blk-cgroup.h         |   68 ++++------------
>  block/blk-iocost.c         |   37 ++++----
>  block/blk-iolatency.c      |    5 -
>  block/blk-ioprio.c         |    8 -
>  block/blk-ioprio.h         |    8 -
>  block/blk-sysfs.c          |    2
>  block/blk-throttle.c       |   13 ++-
>  block/blk-throttle.h       |   16 +--
>  block/blk.h                |    4
>  block/genhd.c              |    7 -
>  include/linux/blk-cgroup.h |    5 -
>  mm/swapfile.c              |    2
>  13 files changed, 148 insertions(+), 213 deletions(-)
> 
> [...]

Applied, thanks!

[01/17] blk-cgroup: fix error unwinding in blkcg_init_queue
        commit: 33dc62796cb657a633050138a86253fb2a553713
[02/17] blk-cgroup: remove blk_queue_root_blkg
        commit: 928f6f00a91ecbef6cb1fe59474831ceaf205290
[03/17] blk-cgroup: remove open coded blkg_lookup instances
        commit: 79fcc5be93e5b17a2a5d36553f7a5c1ad9e953b6
[04/17] blk-cgroup: cleanup the blkg_lookup family of functions
        commit: 4a69f325aa43847e0827fbfe4b3623307b0c9baa
[05/17] blk-cgroup: remove blkg_lookup_check
        commit: f753526e327bc849c445c084d0f374e992038ae9
[06/17] blk-cgroup: pass a gendisk to blkcg_init_queue and blkcg_exit_queue
        commit: 9823538fb7efe66ce987a1e4c0e0f3dc882623c4
[07/17] blk-ioprio: pass a gendisk to blk_ioprio_init and blk_ioprio_exit
        commit: b0dde3f5d628f76f461fb650e2cebfac3460cff6
[08/17] blk-iolatency: pass a gendisk to blk_iolatency_init
        commit: 16fac1b5912b778a30d8863dbc928bef25c8d307
[09/17] blk-iocost: simplify ioc_name
        commit: 9df3e65139b923dfe98f76b7057882c7afb2d3e4
[10/17] blk-iocost: pass a gendisk to blk_iocost_init
        commit: 57b64554977e28ab84d33d298032872a8047a557
[11/17] blk-iocost: cleanup ioc_qos_write
        commit: 3657647e33dff916a2d2d9df926d9bca3907d34f
[12/17] blk-throttle: pass a gendisk to blk_throtl_init and blk_throtl_exit
        commit: e13793bae65919cd3e6a7827f8d30f4dbb8584ee
[13/17] blk-throttle: pass a gendisk to blk_throtl_register_queue
        commit: 5f6dc7522ac2e1701c92f20b9a1a664736787728
[14/17] blk-throttle: pass a gendisk to blk_throtl_cancel_bios
        commit: cad9266abcef586aa95f6f4095781e3e55473f2a
[15/17] blk-cgroup: pass a gendisk to blkg_destroy_all
        commit: 00ad6991bbae116b7c83f68754edd6f4d5e65e01
[16/17] blk-cgroup: pass a gendisk to blkcg_schedule_throttle
        commit: de185b56e8a62822d4e1cdb3e068b38ca709aa47
[17/17] blk-cgroup: pass a gendisk to the blkg allocation helpers
        commit: 99e603874366be1115b40ecbc0e25847186d84ea

Best regards,
-- 
Jens Axboe

blk-cgroup cleanups

[Christoph Hellwig]

Hi all,

below are a couple of cleanups I came up with when trying to understand
the blk-cgroup code.

Re: blk-cgroup cleanups

[Tejun Heo]

On Thu, Jun 06, 2019 at 12:26:18PM +0200, Christoph Hellwig wrote:
> Hi all,
> 
> below are a couple of cleanups I came up with when trying to understand
> the blk-cgroup code.

For the whole patchset,

 Acked-by: Tejun Heo <tj@kernel.org>

Thanks.

-- 
tejun

Re: blk-cgroup cleanups

[Christoph Hellwig]

ping?

On Thu, Jun 06, 2019 at 12:26:18PM +0200, Christoph Hellwig wrote:
> Hi all,
> 
> below are a couple of cleanups I came up with when trying to understand
> the blk-cgroup code.
---end quoted text---

Re: blk-cgroup cleanups

[Jens Axboe]

On 6/6/19 4:26 AM, Christoph Hellwig wrote:
> Hi all,
> 
> below are a couple of cleanups I came up with when trying to understand
> the blk-cgroup code.

Applied, thanks.

-- 
Jens Axboe