diff mbox

[V2,2/2] ubifs: use dirty_writeback_interval value for wbuf timer

Message ID 20160920083615.31099-2-zajec5@gmail.com
State Accepted
Headers show

Commit Message

Rafał Miłecki Sept. 20, 2016, 8:36 a.m. UTC
From: Rafał Miłecki <rafal@milecki.pl>

Right now wbuf timer has hardcoded timeouts and there is no place for
manual adjustments. Some projects / cases many need that though. Few
file systems allow doing that by respecting dirty_writeback_interval
that can be set using sysctl (dirty_writeback_centisecs).

Lowering dirty_writeback_interval could be some way of dealing with user
space apps lacking proper fsyncs. This is definitely *not* a perfect
solution but we don't have ideal (user space) world. There were already
advanced discussions on this matter, mostly when ext4 was introduced and
it wasn't behaving as ext3. Anyway, the final decision was to add some
hacks to the ext4, as trying to fix whole user space or adding new API
was pointless.

We can't (and shouldn't?) just follow ext4. We can't e.g. sync on close
as this would cause too many commits and flash wearing. On the other
hand we still should allow some trade-off between -o sync and default
wbuf timeout. Respecting dirty_writeback_interval should allow some sane
cutomizations if used warily.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
---
V2: Fix delta calculation and make it more clear. Thanks Boris.
---
 fs/ubifs/io.c    | 8 ++++----
 fs/ubifs/ubifs.h | 4 ----
 2 files changed, 4 insertions(+), 8 deletions(-)

Comments

Boris Brezillon Sept. 20, 2016, 8:40 a.m. UTC | #1
On Tue, 20 Sep 2016 10:36:15 +0200
Rafał Miłecki <zajec5@gmail.com> wrote:

> From: Rafał Miłecki <rafal@milecki.pl>
> 
> Right now wbuf timer has hardcoded timeouts and there is no place for
> manual adjustments. Some projects / cases many need that though. Few
> file systems allow doing that by respecting dirty_writeback_interval
> that can be set using sysctl (dirty_writeback_centisecs).
> 
> Lowering dirty_writeback_interval could be some way of dealing with user
> space apps lacking proper fsyncs. This is definitely *not* a perfect
> solution but we don't have ideal (user space) world. There were already
> advanced discussions on this matter, mostly when ext4 was introduced and
> it wasn't behaving as ext3. Anyway, the final decision was to add some
> hacks to the ext4, as trying to fix whole user space or adding new API
> was pointless.
> 
> We can't (and shouldn't?) just follow ext4. We can't e.g. sync on close
> as this would cause too many commits and flash wearing. On the other
> hand we still should allow some trade-off between -o sync and default
> wbuf timeout. Respecting dirty_writeback_interval should allow some sane
> cutomizations if used warily.
> 
> Signed-off-by: Rafał Miłecki <rafal@milecki.pl>

Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>

> ---
> V2: Fix delta calculation and make it more clear. Thanks Boris.
> ---
>  fs/ubifs/io.c    | 8 ++++----
>  fs/ubifs/ubifs.h | 4 ----
>  2 files changed, 4 insertions(+), 8 deletions(-)
> 
> diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c
> index 4d6ce4a..3be2890 100644
> --- a/fs/ubifs/io.c
> +++ b/fs/ubifs/io.c
> @@ -452,11 +452,11 @@ static enum hrtimer_restart wbuf_timer_callback_nolock(struct hrtimer *timer)
>   */
>  static void new_wbuf_timer_nolock(struct ubifs_wbuf *wbuf)
>  {
> -	ktime_t softlimit = ktime_set(WBUF_TIMEOUT_SOFTLIMIT, 0);
> -	unsigned long long delta;
> +	ktime_t softlimit = ms_to_ktime(dirty_writeback_interval * 10);
> +	unsigned long long delta = dirty_writeback_interval;
>  
> -	delta = WBUF_TIMEOUT_HARDLIMIT - WBUF_TIMEOUT_SOFTLIMIT;
> -	delta *= 1000000000ULL;
> +	/* centi to milli, milli to nano, then 10% */
> +	delta *= 10ULL * NSEC_PER_MSEC / 10ULL;
>  
>  	ubifs_assert(!hrtimer_active(&wbuf->timer));
>  	ubifs_assert(delta <= ULONG_MAX);
> diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
> index 11bc8fa..26e6340 100644
> --- a/fs/ubifs/ubifs.h
> +++ b/fs/ubifs/ubifs.h
> @@ -83,10 +83,6 @@
>   */
>  #define BGT_NAME_PATTERN "ubifs_bgt%d_%d"
>  
> -/* Write-buffer synchronization timeout interval in seconds */
> -#define WBUF_TIMEOUT_SOFTLIMIT 3
> -#define WBUF_TIMEOUT_HARDLIMIT 5
> -
>  /* Maximum possible inode number (only 32-bit inodes are supported now) */
>  #define MAX_INUM 0xFFFFFFFF
>
diff mbox

Patch

diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c
index 4d6ce4a..3be2890 100644
--- a/fs/ubifs/io.c
+++ b/fs/ubifs/io.c
@@ -452,11 +452,11 @@  static enum hrtimer_restart wbuf_timer_callback_nolock(struct hrtimer *timer)
  */
 static void new_wbuf_timer_nolock(struct ubifs_wbuf *wbuf)
 {
-	ktime_t softlimit = ktime_set(WBUF_TIMEOUT_SOFTLIMIT, 0);
-	unsigned long long delta;
+	ktime_t softlimit = ms_to_ktime(dirty_writeback_interval * 10);
+	unsigned long long delta = dirty_writeback_interval;
 
-	delta = WBUF_TIMEOUT_HARDLIMIT - WBUF_TIMEOUT_SOFTLIMIT;
-	delta *= 1000000000ULL;
+	/* centi to milli, milli to nano, then 10% */
+	delta *= 10ULL * NSEC_PER_MSEC / 10ULL;
 
 	ubifs_assert(!hrtimer_active(&wbuf->timer));
 	ubifs_assert(delta <= ULONG_MAX);
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
index 11bc8fa..26e6340 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -83,10 +83,6 @@ 
  */
 #define BGT_NAME_PATTERN "ubifs_bgt%d_%d"
 
-/* Write-buffer synchronization timeout interval in seconds */
-#define WBUF_TIMEOUT_SOFTLIMIT 3
-#define WBUF_TIMEOUT_HARDLIMIT 5
-
 /* Maximum possible inode number (only 32-bit inodes are supported now) */
 #define MAX_INUM 0xFFFFFFFF