[PATCHv2] zram: use notify_free to account all free notifications

[PATCHv2] zram: use notify_free to account all free notifications

[Sergey Senozhatsky]

Hello,

This patch change the way zram uses `notify_free' stat attr. Currently,
this attr accounts only swap free notifications, iow used only when zram
device is a swap device. The patch preserves the existing behaviour and
additionally makes `notify_free' universal, so it's also used by zram to
account REQ_DISCARDs when zram is not used as a swap device.

Please refer to:
https://lkml.org/lkml/2014/9/3/1115

V2:
- fix accounting error found by Minchan Kim.

Please revert 2cbf6f7bea1048ab5 (zram: add num_{discard_req, discarded}
for discard stat) and apply this patch.

Sergey Senozhatsky (1):
  zram: use notify_free to account all free notifications

 Documentation/ABI/testing/sysfs-block-zram | 13 ++++++++-----
 drivers/block/zram/zram_drv.c              |  2 +-
 2 files changed, 9 insertions(+), 6 deletions(-)

-- 
2.1.0.134.gb1ae451

[PATCHv2] zram: use notify_free to account all free notifications

[Sergey Senozhatsky]

notify_free device attribute accounts the number of slot free notifications
and internally represents the number of zram_free_page() calls. Slot free
notifications are sent only when device is used as a swap device, hence
notify_free is used only for swap devices. Since f4659d8e620d08 (zram:
support REQ_DISCARD) ZRAM handles yet another one free notification (also
via zram_free_page() call) -- REQ_DISCARD requests, which are sent by a
filesystem, whenever some data blocks are discarded. However, there is no
way to know the number of notifications in the latter case.

Change zram_free_page() to return a bool status, indicating if zs_free()
has happened. So we can use `notify_free' to account the number of pages
freed by zram_bio_discard() and zram_slot_free_notify().

This means that depending on usage scenario `notify_free' represents:
 a) the number of pages freed because of slot free notifications, which is
   equal to the number of swap_slot_free_notify() calls, so there is no
   behaviour change

 b) the number of pages freed because of REQ_DISCARD notifications

Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
---
 Documentation/ABI/testing/sysfs-block-zram | 13 ++++++++-----
 drivers/block/zram/zram_drv.c              | 12 +++++++-----
 2 files changed, 15 insertions(+), 10 deletions(-)

diff --git a/Documentation/ABI/testing/sysfs-block-zram b/Documentation/ABI/testing/sysfs-block-zram
index b13dc99..a6148ea 100644
--- a/Documentation/ABI/testing/sysfs-block-zram
+++ b/Documentation/ABI/testing/sysfs-block-zram
@@ -77,11 +77,14 @@ What:		/sys/block/zram<id>/notify_free
 Date:		August 2010
 Contact:	Nitin Gupta <ngupta@vflare.org>
 Description:
-		The notify_free file is read-only and specifies the number of
-		swap slot free notifications received by this device. These
-		notifications are sent to a swap block device when a swap slot
-		is freed. This statistic is applicable only when this disk is
-		being used as a swap disk.
+		The notify_free file is read-only. Depending on device usage
+		scenario it may account a) the number of pages freed because
+		of swap slot free notifications or b) the number of pages freed
+		because of REQ_DISCARD requests sent by bio. The former ones
+		are sent to a swap block device when a swap slot is freed, which
+		implies that this disk is being used as a swap disk. The latter
+		ones are sent by filesystem mounted with discard option,
+		whenever some data blocks are getting discarded.
 
 What:		/sys/block/zram<id>/zero_pages
 Date:		August 2010
diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index d78b245..03d11d5 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -388,7 +388,7 @@ static void handle_zero_page(struct bio_vec *bvec)
  * caller should hold this table index entry's bit_spinlock to
  * indicate this index entry is accessing.
  */
-static void zram_free_page(struct zram *zram, size_t index)
+static bool zram_free_page(struct zram *zram, size_t index)
 {
 	struct zram_meta *meta = zram->meta;
 	unsigned long handle = meta->table[index].handle;
@@ -402,7 +402,7 @@ static void zram_free_page(struct zram *zram, size_t index)
 			zram_clear_flag(meta, index, ZRAM_ZERO);
 			atomic64_dec(&zram->stats.zero_pages);
 		}
-		return;
+		return false;
 	}
 
 	zs_free(meta->mem_pool, handle);
@@ -413,6 +413,7 @@ static void zram_free_page(struct zram *zram, size_t index)
 
 	meta->table[index].handle = 0;
 	zram_set_obj_size(meta, index, 0);
+	return true;
 }
 
 static int zram_decompress_page(struct zram *zram, char *mem, u32 index)
@@ -696,7 +697,8 @@ static void zram_bio_discard(struct zram *zram, u32 index,
 
 	while (n >= PAGE_SIZE) {
 		bit_spin_lock(ZRAM_ACCESS, &meta->table[index].value);
-		zram_free_page(zram, index);
+		if (zram_free_page(zram, index))
+			atomic64_inc(&zram->stats.notify_free);
 		bit_spin_unlock(ZRAM_ACCESS, &meta->table[index].value);
 		index++;
 		n -= PAGE_SIZE;
@@ -936,9 +938,9 @@ static void zram_slot_free_notify(struct block_device *bdev,
 	meta = zram->meta;
 
 	bit_spin_lock(ZRAM_ACCESS, &meta->table[index].value);
-	zram_free_page(zram, index);
+	if (zram_free_page(zram, index))
+		atomic64_inc(&zram->stats.notify_free);
 	bit_spin_unlock(ZRAM_ACCESS, &meta->table[index].value);
-	atomic64_inc(&zram->stats.notify_free);
 }
 
 static const struct block_device_operations zram_devops = {
-- 
2.1.0.251.ga182987

Re: [PATCHv2] zram: use notify_free to account all free notifications

[Minchan Kim]

Hi Sergey,

On Sat, Sep 13, 2014 at 12:52:14PM +0900, Sergey Senozhatsky wrote:
> notify_free device attribute accounts the number of slot free notifications
> and internally represents the number of zram_free_page() calls. Slot free
> notifications are sent only when device is used as a swap device, hence
> notify_free is used only for swap devices. Since f4659d8e620d08 (zram:
> support REQ_DISCARD) ZRAM handles yet another one free notification (also
> via zram_free_page() call) -- REQ_DISCARD requests, which are sent by a
> filesystem, whenever some data blocks are discarded. However, there is no
> way to know the number of notifications in the latter case.
> 
> Change zram_free_page() to return a bool status, indicating if zs_free()
> has happened. So we can use `notify_free' to account the number of pages
> freed by zram_bio_discard() and zram_slot_free_notify().
> 
> This means that depending on usage scenario `notify_free' represents:
>  a) the number of pages freed because of slot free notifications, which is
>    equal to the number of swap_slot_free_notify() calls, so there is no
>    behaviour change
> 
>  b) the number of pages freed because of REQ_DISCARD notifications
> 
> Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
> ---
>  Documentation/ABI/testing/sysfs-block-zram | 13 ++++++++-----
>  drivers/block/zram/zram_drv.c              | 12 +++++++-----
>  2 files changed, 15 insertions(+), 10 deletions(-)
> 
> diff --git a/Documentation/ABI/testing/sysfs-block-zram b/Documentation/ABI/testing/sysfs-block-zram
> index b13dc99..a6148ea 100644
> --- a/Documentation/ABI/testing/sysfs-block-zram
> +++ b/Documentation/ABI/testing/sysfs-block-zram
> @@ -77,11 +77,14 @@ What:		/sys/block/zram<id>/notify_free
>  Date:		August 2010
>  Contact:	Nitin Gupta <ngupta@vflare.org>
>  Description:
> -		The notify_free file is read-only and specifies the number of
> -		swap slot free notifications received by this device. These
> -		notifications are sent to a swap block device when a swap slot
> -		is freed. This statistic is applicable only when this disk is
> -		being used as a swap disk.
> +		The notify_free file is read-only. Depending on device usage
> +		scenario it may account a) the number of pages freed because
> +		of swap slot free notifications or b) the number of pages freed
> +		because of REQ_DISCARD requests sent by bio. The former ones
> +		are sent to a swap block device when a swap slot is freed, which
> +		implies that this disk is being used as a swap disk. The latter
> +		ones are sent by filesystem mounted with discard option,
> +		whenever some data blocks are getting discarded.
>  
>  What:		/sys/block/zram<id>/zero_pages
>  Date:		August 2010
> diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
> index d78b245..03d11d5 100644
> --- a/drivers/block/zram/zram_drv.c
> +++ b/drivers/block/zram/zram_drv.c
> @@ -388,7 +388,7 @@ static void handle_zero_page(struct bio_vec *bvec)
>   * caller should hold this table index entry's bit_spinlock to
>   * indicate this index entry is accessing.
>   */
> -static void zram_free_page(struct zram *zram, size_t index)
> +static bool zram_free_page(struct zram *zram, size_t index)
>  {
>  	struct zram_meta *meta = zram->meta;
>  	unsigned long handle = meta->table[index].handle;
> @@ -402,7 +402,7 @@ static void zram_free_page(struct zram *zram, size_t index)
>  			zram_clear_flag(meta, index, ZRAM_ZERO);
>  			atomic64_dec(&zram->stats.zero_pages);
>  		}
> -		return;
> +		return false;
>  	}
>  
>  	zs_free(meta->mem_pool, handle);
> @@ -413,6 +413,7 @@ static void zram_free_page(struct zram *zram, size_t index)
>  
>  	meta->table[index].handle = 0;
>  	zram_set_obj_size(meta, index, 0);
> +	return true;
>  }

For just stat accounting, adding return value for fast path?
I don't think it's not a important stat at the cost of adding more overhead
in fastpath. If you have a strong reason, I will do that. Otherwise,
please, don't touch fast path and just account it regardless of real freeing.
I should have said before resending.
Sorry for bothering you. :)

>  
>  static int zram_decompress_page(struct zram *zram, char *mem, u32 index)
> @@ -696,7 +697,8 @@ static void zram_bio_discard(struct zram *zram, u32 index,
>  
>  	while (n >= PAGE_SIZE) {
>  		bit_spin_lock(ZRAM_ACCESS, &meta->table[index].value);
> -		zram_free_page(zram, index);
> +		if (zram_free_page(zram, index))
> +			atomic64_inc(&zram->stats.notify_free);
>  		bit_spin_unlock(ZRAM_ACCESS, &meta->table[index].value);
>  		index++;
>  		n -= PAGE_SIZE;
> @@ -936,9 +938,9 @@ static void zram_slot_free_notify(struct block_device *bdev,
>  	meta = zram->meta;
>  
>  	bit_spin_lock(ZRAM_ACCESS, &meta->table[index].value);
> -	zram_free_page(zram, index);
> +	if (zram_free_page(zram, index))
> +		atomic64_inc(&zram->stats.notify_free);
>  	bit_spin_unlock(ZRAM_ACCESS, &meta->table[index].value);
> -	atomic64_inc(&zram->stats.notify_free);
>  }
>  
>  static const struct block_device_operations zram_devops = {
> -- 
> 2.1.0.251.ga182987
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

-- 
Kind regards,
Minchan Kim

Re: [PATCHv2] zram: use notify_free to account all free notifications

[Sergey Senozhatsky]

Hi Minchan,

On (09/15/14 08:22), Minchan Kim wrote:
> Hi Sergey,
> 
> On Sat, Sep 13, 2014 at 12:52:14PM +0900, Sergey Senozhatsky wrote:
> > notify_free device attribute accounts the number of slot free notifications
> > and internally represents the number of zram_free_page() calls. Slot free
> > notifications are sent only when device is used as a swap device, hence
> > notify_free is used only for swap devices. Since f4659d8e620d08 (zram:
> > support REQ_DISCARD) ZRAM handles yet another one free notification (also
> > via zram_free_page() call) -- REQ_DISCARD requests, which are sent by a
> > filesystem, whenever some data blocks are discarded. However, there is no
> > way to know the number of notifications in the latter case.
> > 
> > Change zram_free_page() to return a bool status, indicating if zs_free()
> > has happened. So we can use `notify_free' to account the number of pages
> > freed by zram_bio_discard() and zram_slot_free_notify().
> > 
> > This means that depending on usage scenario `notify_free' represents:
> >  a) the number of pages freed because of slot free notifications, which is
> >    equal to the number of swap_slot_free_notify() calls, so there is no
> >    behaviour change
> > 
> >  b) the number of pages freed because of REQ_DISCARD notifications
> > 
> > Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
> > ---
> >  Documentation/ABI/testing/sysfs-block-zram | 13 ++++++++-----
> >  drivers/block/zram/zram_drv.c              | 12 +++++++-----
> >  2 files changed, 15 insertions(+), 10 deletions(-)
> > 
> > diff --git a/Documentation/ABI/testing/sysfs-block-zram b/Documentation/ABI/testing/sysfs-block-zram
> > index b13dc99..a6148ea 100644
> > --- a/Documentation/ABI/testing/sysfs-block-zram
> > +++ b/Documentation/ABI/testing/sysfs-block-zram
> > @@ -77,11 +77,14 @@ What:		/sys/block/zram<id>/notify_free
> >  Date:		August 2010
> >  Contact:	Nitin Gupta <ngupta@vflare.org>
> >  Description:
> > -		The notify_free file is read-only and specifies the number of
> > -		swap slot free notifications received by this device. These
> > -		notifications are sent to a swap block device when a swap slot
> > -		is freed. This statistic is applicable only when this disk is
> > -		being used as a swap disk.
> > +		The notify_free file is read-only. Depending on device usage
> > +		scenario it may account a) the number of pages freed because
> > +		of swap slot free notifications or b) the number of pages freed
> > +		because of REQ_DISCARD requests sent by bio. The former ones
> > +		are sent to a swap block device when a swap slot is freed, which
> > +		implies that this disk is being used as a swap disk. The latter
> > +		ones are sent by filesystem mounted with discard option,
> > +		whenever some data blocks are getting discarded.
> >  
> >  What:		/sys/block/zram<id>/zero_pages
> >  Date:		August 2010
> > diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
> > index d78b245..03d11d5 100644
> > --- a/drivers/block/zram/zram_drv.c
> > +++ b/drivers/block/zram/zram_drv.c
> > @@ -388,7 +388,7 @@ static void handle_zero_page(struct bio_vec *bvec)
> >   * caller should hold this table index entry's bit_spinlock to
> >   * indicate this index entry is accessing.
> >   */
> > -static void zram_free_page(struct zram *zram, size_t index)
> > +static bool zram_free_page(struct zram *zram, size_t index)
> >  {
> >  	struct zram_meta *meta = zram->meta;
> >  	unsigned long handle = meta->table[index].handle;
> > @@ -402,7 +402,7 @@ static void zram_free_page(struct zram *zram, size_t index)
> >  			zram_clear_flag(meta, index, ZRAM_ZERO);
> >  			atomic64_dec(&zram->stats.zero_pages);
> >  		}
> > -		return;
> > +		return false;
> >  	}
> >  
> >  	zs_free(meta->mem_pool, handle);
> > @@ -413,6 +413,7 @@ static void zram_free_page(struct zram *zram, size_t index)
> >  
> >  	meta->table[index].handle = 0;
> >  	zram_set_obj_size(meta, index, 0);
> > +	return true;
> >  }
> 
> For just stat accounting, adding return value for fast path?
> I don't think it's not a important stat at the cost of adding more overhead
> in fastpath. If you have a strong reason, I will do that. Otherwise,
> please, don't touch fast path and just account it regardless of real freeing.

yes, that was the reason -- we can show just a number of notifications
or a number of notifications that forced zram to take some actions (free memory).
my target was the second case. [[besides, a single mov could not dramatically
impact the perfomance; we do much heavier staff when we actually free the
memory]].

> I should have said before resending.
> Sorry for bothering you. :)

no problem, resent v3.
hoping this time it'll work out for everyone :)

	-ss

> 
> >  
> >  static int zram_decompress_page(struct zram *zram, char *mem, u32 index)
> > @@ -696,7 +697,8 @@ static void zram_bio_discard(struct zram *zram, u32 index,
> >  
> >  	while (n >= PAGE_SIZE) {
> >  		bit_spin_lock(ZRAM_ACCESS, &meta->table[index].value);
> > -		zram_free_page(zram, index);
> > +		if (zram_free_page(zram, index))
> > +			atomic64_inc(&zram->stats.notify_free);
> >  		bit_spin_unlock(ZRAM_ACCESS, &meta->table[index].value);
> >  		index++;
> >  		n -= PAGE_SIZE;
> > @@ -936,9 +938,9 @@ static void zram_slot_free_notify(struct block_device *bdev,
> >  	meta = zram->meta;
> >  
> >  	bit_spin_lock(ZRAM_ACCESS, &meta->table[index].value);
> > -	zram_free_page(zram, index);
> > +	if (zram_free_page(zram, index))
> > +		atomic64_inc(&zram->stats.notify_free);
> >  	bit_spin_unlock(ZRAM_ACCESS, &meta->table[index].value);
> > -	atomic64_inc(&zram->stats.notify_free);
> >  }
> >  
> >  static const struct block_device_operations zram_devops = {
> > -- 
> > 2.1.0.251.ga182987
> > 
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at  http://www.tux.org/lkml/
> 
> -- 
> Kind regards,
> Minchan Kim
>