diff mbox series

[v16,17/26] blk-zoned: Uninline functions that are not in the hot path

Message ID 20241119002815.600608-18-bvanassche@acm.org (mailing list archive)
State New
Headers show
Series Improve write performance for zoned UFS devices | expand

Commit Message

Bart Van Assche Nov. 19, 2024, 12:28 a.m. UTC
Apply the convention that is followed elsewhere in the block layer: only
declare functions inline if these are in the hot path.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 block/blk-zoned.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

Comments

Damien Le Moal Nov. 19, 2024, 7:55 a.m. UTC | #1
On 11/19/24 09:28, Bart Van Assche wrote:
> Apply the convention that is followed elsewhere in the block layer: only
> declare functions inline if these are in the hot path.
> 
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>

I do not really see the point... Anyway, looks OK.

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Bart Van Assche Nov. 19, 2024, 9:20 p.m. UTC | #2
On 11/18/24 11:55 PM, Damien Le Moal wrote:
> I do not really see the point... Anyway, looks OK.
Hi Damien,

One of the approaches I used to debug this patch series is to add
trace_printk() calls to track zwplug state changes. trace_printk()
reports the function name in front of the information in its argument
list. I noticed that the function name reported by trace_printk() is
incorrect for inlined functions. Hence this patch.

Thanks,

Bart.
Damien Le Moal Nov. 21, 2024, 3:36 a.m. UTC | #3
On 11/20/24 06:20, Bart Van Assche wrote:
> On 11/18/24 11:55 PM, Damien Le Moal wrote:
>> I do not really see the point... Anyway, looks OK.
> Hi Damien,
> 
> One of the approaches I used to debug this patch series is to add
> trace_printk() calls to track zwplug state changes. trace_printk()
> reports the function name in front of the information in its argument
> list. I noticed that the function name reported by trace_printk() is
> incorrect for inlined functions. Hence this patch.

Maybe add this explanation to the commit message ? It does make sense to no
inline these function with tracing.

> 
> Thanks,
> 
> Bart.
diff mbox series

Patch

diff --git a/block/blk-zoned.c b/block/blk-zoned.c
index 17fe40db1888..2b4783026450 100644
--- a/block/blk-zoned.c
+++ b/block/blk-zoned.c
@@ -558,8 +558,8 @@  static struct blk_zone_wplug *disk_get_and_lock_zone_wplug(struct gendisk *disk,
 	return zwplug;
 }
 
-static inline void blk_zone_wplug_bio_io_error(struct blk_zone_wplug *zwplug,
-					       struct bio *bio)
+static void blk_zone_wplug_bio_io_error(struct blk_zone_wplug *zwplug,
+					struct bio *bio)
 {
 	struct request_queue *q = zwplug->disk->queue;
 
@@ -580,8 +580,8 @@  static void disk_zone_wplug_abort(struct blk_zone_wplug *zwplug)
 		blk_zone_wplug_bio_io_error(zwplug, bio);
 }
 
-static inline void disk_zone_wplug_set_error(struct gendisk *disk,
-					     struct blk_zone_wplug *zwplug)
+static void disk_zone_wplug_set_error(struct gendisk *disk,
+				      struct blk_zone_wplug *zwplug)
 {
 	unsigned long flags;
 
@@ -607,8 +607,8 @@  static inline void disk_zone_wplug_set_error(struct gendisk *disk,
 	spin_unlock_irqrestore(&disk->zone_wplugs_lock, flags);
 }
 
-static inline void disk_zone_wplug_clear_error(struct gendisk *disk,
-					       struct blk_zone_wplug *zwplug)
+static void disk_zone_wplug_clear_error(struct gendisk *disk,
+					struct blk_zone_wplug *zwplug)
 {
 	unsigned long flags;
 
@@ -1597,7 +1597,7 @@  void disk_free_zone_resources(struct gendisk *disk)
 	disk->nr_zones = 0;
 }
 
-static inline bool disk_need_zone_resources(struct gendisk *disk)
+static bool disk_need_zone_resources(struct gendisk *disk)
 {
 	/*
 	 * All mq zoned devices need zone resources so that the block layer