* [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
@ 2011-12-26 6:01 ` b29237
0 siblings, 0 replies; 9+ messages in thread
From: b29237 @ 2011-12-26 6:01 UTC (permalink / raw)
To: iws, vinod.koul, dan.j.williams, linuxppc-dev, linux-kernel; +Cc: Forrest shi
From: Forrest shi <b29237@freescale.com>
dma status check function fsl_tx_status is heavily called in
a tight loop and the desc lock in fsl_tx_status contended by
the dma status update function. this caused the dma performance
degrades much.
this patch releases the lock in the fsl_tx_status function, and
introduce the smp_mb() to avoid possible memory inconsistency.
Signed-off-by: Forrest Shi <xuelin.shi@freescale.com>
---
drivers/dma/fsldma.c | 6 +-----
1 files changed, 1 insertions(+), 5 deletions(-)
diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c
index 8a78154..008fb5e 100644
--- a/drivers/dma/fsldma.c
+++ b/drivers/dma/fsldma.c
@@ -986,15 +986,11 @@ static enum dma_status fsl_tx_status(struct dma_chan *dchan,
struct fsldma_chan *chan = to_fsl_chan(dchan);
dma_cookie_t last_complete;
dma_cookie_t last_used;
- unsigned long flags;
-
- spin_lock_irqsave(&chan->desc_lock, flags);
last_complete = chan->completed_cookie;
+ smp_mb();
last_used = dchan->cookie;
- spin_unlock_irqrestore(&chan->desc_lock, flags);
-
dma_set_tx_state(txstate, last_complete, last_used, 0);
return dma_async_is_complete(cookie, last_complete, last_used);
}
--
1.7.0.4
^ permalink raw reply related [flat|nested] 9+ messages in thread* [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
@ 2011-12-26 6:01 ` b29237
0 siblings, 0 replies; 9+ messages in thread
From: b29237 @ 2011-12-26 6:01 UTC (permalink / raw)
To: iws, vinod.koul, dan.j.williams, linuxppc-dev, linux-kernel; +Cc: Forrest shi
From: Forrest shi <b29237@freescale.com>
dma status check function fsl_tx_status is heavily called in
a tight loop and the desc lock in fsl_tx_status contended by
the dma status update function. this caused the dma performance
degrades much.
this patch releases the lock in the fsl_tx_status function, and
introduce the smp_mb() to avoid possible memory inconsistency.
Signed-off-by: Forrest Shi <xuelin.shi@freescale.com>
---
drivers/dma/fsldma.c | 6 +-----
1 files changed, 1 insertions(+), 5 deletions(-)
diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c
index 8a78154..008fb5e 100644
--- a/drivers/dma/fsldma.c
+++ b/drivers/dma/fsldma.c
@@ -986,15 +986,11 @@ static enum dma_status fsl_tx_status(struct dma_chan *dchan,
struct fsldma_chan *chan = to_fsl_chan(dchan);
dma_cookie_t last_complete;
dma_cookie_t last_used;
- unsigned long flags;
-
- spin_lock_irqsave(&chan->desc_lock, flags);
last_complete = chan->completed_cookie;
+ smp_mb();
last_used = dchan->cookie;
- spin_unlock_irqrestore(&chan->desc_lock, flags);
-
dma_set_tx_state(txstate, last_complete, last_used, 0);
return dma_async_is_complete(cookie, last_complete, last_used);
}
--
1.7.0.4
^ permalink raw reply related [flat|nested] 9+ messages in thread* RE: [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
2011-12-26 6:01 ` b29237
@ 2012-01-11 7:54 ` Shi Xuelin-B29237
-1 siblings, 0 replies; 9+ messages in thread
From: Shi Xuelin-B29237 @ 2012-01-11 7:54 UTC (permalink / raw)
To: iws@ovro.caltech.edu
Cc: vinod.koul@intel.com, dan.j.williams@intel.com,
linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org
Hello Iris,
As we discussed in the previous patch, I add one smp_mb() in fsl_tx_status.
In my testing with iozone, this smp_mb() could cause 1%~2% performance degr=
adation.
Anyway it is acceptable for me. Do you have any other comments?
Thanks,
Forrest
-----Original Message-----
From: Shi Xuelin-B29237=20
Sent: 2011=1B$BG/=1B(B12=1B$B7n=1B(B26=1B$BF|=1B(B 14:01
To: iws@ovro.caltech.edu; vinod.koul@intel.com; dan.j.williams@intel.com; l=
inuxppc-dev@lists.ozlabs.org; linux-kernel@vger.kernel.org
Cc: Shi Xuelin-B29237
Subject: [PATCH] fsldma: fix performance degradation by optimizing spinlock=
use.
From: Forrest shi <b29237@freescale.com>
dma status check function fsl_tx_status is heavily called in
a tight loop and the desc lock in fsl_tx_status contended by
the dma status update function. this caused the dma performance
degrades much.
this patch releases the lock in the fsl_tx_status function, and
introduce the smp_mb() to avoid possible memory inconsistency.
Signed-off-by: Forrest Shi <xuelin.shi@freescale.com>
---
drivers/dma/fsldma.c | 6 +-----
1 files changed, 1 insertions(+), 5 deletions(-)
diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c index 8a78154..008=
fb5e 100644
--- a/drivers/dma/fsldma.c
+++ b/drivers/dma/fsldma.c
@@ -986,15 +986,11 @@ static enum dma_status fsl_tx_status(struct dma_chan =
*dchan,
struct fsldma_chan *chan =3D to_fsl_chan(dchan);
dma_cookie_t last_complete;
dma_cookie_t last_used;
- unsigned long flags;
-
- spin_lock_irqsave(&chan->desc_lock, flags);
=20
last_complete =3D chan->completed_cookie;
+ smp_mb();
last_used =3D dchan->cookie;
=20
- spin_unlock_irqrestore(&chan->desc_lock, flags);
-
dma_set_tx_state(txstate, last_complete, last_used, 0);
return dma_async_is_complete(cookie, last_complete, last_used); }
--
1.7.0.4
^ permalink raw reply [flat|nested] 9+ messages in thread* RE: [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
@ 2012-01-11 7:54 ` Shi Xuelin-B29237
0 siblings, 0 replies; 9+ messages in thread
From: Shi Xuelin-B29237 @ 2012-01-11 7:54 UTC (permalink / raw)
To: iws@ovro.caltech.edu
Cc: vinod.koul@intel.com, dan.j.williams@intel.com,
linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org
Hello Iris,
As we discussed in the previous patch, I add one smp_mb() in fsl_tx_status.
In my testing with iozone, this smp_mb() could cause 1%~2% performance degradation.
Anyway it is acceptable for me. Do you have any other comments?
Thanks,
Forrest
-----Original Message-----
From: Shi Xuelin-B29237
Sent: 2011年12月26日 14:01
To: iws@ovro.caltech.edu; vinod.koul@intel.com; dan.j.williams@intel.com; linuxppc-dev@lists.ozlabs.org; linux-kernel@vger.kernel.org
Cc: Shi Xuelin-B29237
Subject: [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
From: Forrest shi <b29237@freescale.com>
dma status check function fsl_tx_status is heavily called in
a tight loop and the desc lock in fsl_tx_status contended by
the dma status update function. this caused the dma performance
degrades much.
this patch releases the lock in the fsl_tx_status function, and
introduce the smp_mb() to avoid possible memory inconsistency.
Signed-off-by: Forrest Shi <xuelin.shi@freescale.com>
---
drivers/dma/fsldma.c | 6 +-----
1 files changed, 1 insertions(+), 5 deletions(-)
diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c index 8a78154..008fb5e 100644
--- a/drivers/dma/fsldma.c
+++ b/drivers/dma/fsldma.c
@@ -986,15 +986,11 @@ static enum dma_status fsl_tx_status(struct dma_chan *dchan,
struct fsldma_chan *chan = to_fsl_chan(dchan);
dma_cookie_t last_complete;
dma_cookie_t last_used;
- unsigned long flags;
-
- spin_lock_irqsave(&chan->desc_lock, flags);
last_complete = chan->completed_cookie;
+ smp_mb();
last_used = dchan->cookie;
- spin_unlock_irqrestore(&chan->desc_lock, flags);
-
dma_set_tx_state(txstate, last_complete, last_used, 0);
return dma_async_is_complete(cookie, last_complete, last_used); }
--
1.7.0.4
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
2012-01-11 7:54 ` Shi Xuelin-B29237
(?)
@ 2012-01-11 16:13 ` Ira W. Snyder
2012-04-26 2:03 ` Shi Xuelin-B29237
-1 siblings, 1 reply; 9+ messages in thread
From: Ira W. Snyder @ 2012-01-11 16:13 UTC (permalink / raw)
To: Shi Xuelin-B29237
Cc: vinod.koul@intel.com, dan.j.williams@intel.com,
linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org
On Wed, Jan 11, 2012 at 07:54:55AM +0000, Shi Xuelin-B29237 wrote:
> Hello Iris,
>
> As we discussed in the previous patch, I add one smp_mb() in fsl_tx_status.
> In my testing with iozone, this smp_mb() could cause 1%~2% performance degradation.
> Anyway it is acceptable for me. Do you have any other comments?
>
This patch looks fine to me.
Ira
> -----Original Message-----
> From: Shi Xuelin-B29237
> Sent: 2011年12月26日 14:01
> To: iws@ovro.caltech.edu; vinod.koul@intel.com; dan.j.williams@intel.com; linuxppc-dev@lists.ozlabs.org; linux-kernel@vger.kernel.org
> Cc: Shi Xuelin-B29237
> Subject: [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
>
> From: Forrest shi <b29237@freescale.com>
>
> dma status check function fsl_tx_status is heavily called in
> a tight loop and the desc lock in fsl_tx_status contended by
> the dma status update function. this caused the dma performance
> degrades much.
>
> this patch releases the lock in the fsl_tx_status function, and
> introduce the smp_mb() to avoid possible memory inconsistency.
>
> Signed-off-by: Forrest Shi <xuelin.shi@freescale.com>
> ---
> drivers/dma/fsldma.c | 6 +-----
> 1 files changed, 1 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c index 8a78154..008fb5e 100644
> --- a/drivers/dma/fsldma.c
> +++ b/drivers/dma/fsldma.c
> @@ -986,15 +986,11 @@ static enum dma_status fsl_tx_status(struct dma_chan *dchan,
> struct fsldma_chan *chan = to_fsl_chan(dchan);
> dma_cookie_t last_complete;
> dma_cookie_t last_used;
> - unsigned long flags;
> -
> - spin_lock_irqsave(&chan->desc_lock, flags);
>
> last_complete = chan->completed_cookie;
> + smp_mb();
> last_used = dchan->cookie;
>
> - spin_unlock_irqrestore(&chan->desc_lock, flags);
> -
> dma_set_tx_state(txstate, last_complete, last_used, 0);
> return dma_async_is_complete(cookie, last_complete, last_used); }
> --
> 1.7.0.4
>
>
^ permalink raw reply [flat|nested] 9+ messages in thread
* RE: [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
2012-01-11 16:13 ` Ira W. Snyder
@ 2012-04-26 2:03 ` Shi Xuelin-B29237
0 siblings, 0 replies; 9+ messages in thread
From: Shi Xuelin-B29237 @ 2012-04-26 2:03 UTC (permalink / raw)
To: dan.j.williams@intel.com
Cc: vinod.koul@intel.com, Li Yang-R58472,
linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org,
Ira W. Snyder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^ permalink raw reply [flat|nested] 9+ messages in thread
* RE: [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
@ 2012-04-26 2:03 ` Shi Xuelin-B29237
0 siblings, 0 replies; 9+ messages in thread
From: Shi Xuelin-B29237 @ 2012-04-26 2:03 UTC (permalink / raw)
To: dan.j.williams@intel.com
Cc: vinod.koul@intel.com, linuxppc-dev@lists.ozlabs.org,
linux-kernel@vger.kernel.org, Ira W. Snyder, Li Yang-R58472
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="utf-8", Size: 2550 bytes --]
Hi Dan Williams,
Do you have any comment about this patch?
Thanks,
Forrest
-----Original Message-----
From: Ira W. Snyder [mailto:iws@ovro.caltech.edu]
Sent: 2012年1月12日 0:14
To: Shi Xuelin-B29237
Cc: vinod.koul@intel.com; dan.j.williams@intel.com; linuxppc-dev@lists.ozlabs.org; linux-kernel@vger.kernel.org
Subject: Re: [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
On Wed, Jan 11, 2012 at 07:54:55AM +0000, Shi Xuelin-B29237 wrote:
> Hello Iris,
>
> As we discussed in the previous patch, I add one smp_mb() in fsl_tx_status.
> In my testing with iozone, this smp_mb() could cause 1%~2% performance degradation.
> Anyway it is acceptable for me. Do you have any other comments?
>
This patch looks fine to me.
Ira
> -----Original Message-----
> From: Shi Xuelin-B29237
> Sent: 2011年12月26日 14:01
> To: iws@ovro.caltech.edu; vinod.koul@intel.com; dan.j.williams@intel.com; linuxppc-dev@lists.ozlabs.org; linux-kernel@vger.kernel.org
> Cc: Shi Xuelin-B29237
> Subject: [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
>
> From: Forrest shi <b29237@freescale.com>
>
> dma status check function fsl_tx_status is heavily called in
> a tight loop and the desc lock in fsl_tx_status contended by
> the dma status update function. this caused the dma performance
> degrades much.
>
> this patch releases the lock in the fsl_tx_status function, and
> introduce the smp_mb() to avoid possible memory inconsistency.
>
> Signed-off-by: Forrest Shi <xuelin.shi@freescale.com>
> ---
> drivers/dma/fsldma.c | 6 +-----
> 1 files changed, 1 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c index 8a78154..008fb5e 100644
> --- a/drivers/dma/fsldma.c
> +++ b/drivers/dma/fsldma.c
> @@ -986,15 +986,11 @@ static enum dma_status fsl_tx_status(struct dma_chan *dchan,
> struct fsldma_chan *chan = to_fsl_chan(dchan);
> dma_cookie_t last_complete;
> dma_cookie_t last_used;
> - unsigned long flags;
> -
> - spin_lock_irqsave(&chan->desc_lock, flags);
>
> last_complete = chan->completed_cookie;
> + smp_mb();
> last_used = dchan->cookie;
>
> - spin_unlock_irqrestore(&chan->desc_lock, flags);
> -
> dma_set_tx_state(txstate, last_complete, last_used, 0);
> return dma_async_is_complete(cookie, last_complete, last_used); }
> --
> 1.7.0.4
>
>
ÿôèº{.nÇ+‰·Ÿ®‰†+%ŠËÿ±éݶ\x17¥Šwÿº{.nÇ+‰·¥Š{±þG«éÿŠ{ayº\x1dʇڙë,j\a¢f£¢·hšïêÿ‘êçz_è®\x03(階ŠÝ¢j"ú\x1a¶^[m§ÿÿ¾\a«þG«éÿ¢¸?™¨èÚ&£ø§~á¶iO•æ¬z·švØ^\x14\x04\x1a¶^[m§ÿÿÃ\fÿ¶ìÿ¢¸?–I¥
^ permalink raw reply [flat|nested] 9+ messages in thread* RE: [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
2012-04-26 2:03 ` Shi Xuelin-B29237
@ 2012-04-26 2:59 ` Li Yang-R58472
-1 siblings, 0 replies; 9+ messages in thread
From: Li Yang-R58472 @ 2012-04-26 2:59 UTC (permalink / raw)
To: Shi Xuelin-B29237, dan.j.williams@intel.com, vinod.koul@intel.com
Cc: linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org,
Ira W. Snyder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^ permalink raw reply [flat|nested] 9+ messages in thread
* RE: [PATCH] fsldma: fix performance degradation by optimizing spinlock use.
@ 2012-04-26 2:59 ` Li Yang-R58472
0 siblings, 0 replies; 9+ messages in thread
From: Li Yang-R58472 @ 2012-04-26 2:59 UTC (permalink / raw)
To: Shi Xuelin-B29237, dan.j.williams@intel.com, vinod.koul@intel.com
Cc: linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org,
Ira W. Snyder
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="utf-8", Size: 3208 bytes --]
Hi Dan and Vinod,
Please help to pick this patch up. It addresses a great performance regression for RAID offloading.
Acked-by: Li Yang <leoli@freescale.com>
- Leo
> -----Original Message-----
> From: Shi Xuelin-B29237
> Sent: Thursday, April 26, 2012 10:04 AM
> To: dan.j.williams@intel.com
> Cc: vinod.koul@intel.com; linuxppc-dev@lists.ozlabs.org; linux-
> kernel@vger.kernel.org; Ira W. Snyder; Li Yang-R58472
> Subject: RE: [PATCH] fsldma: fix performance degradation by optimizing
> spinlock use.
>
> Hi Dan Williams,
>
> Do you have any comment about this patch?
>
> Thanks,
> Forrest
>
> -----Original Message-----
> From: Ira W. Snyder [mailto:iws@ovro.caltech.edu]
> Sent: 2012年1月12日 0:14
> To: Shi Xuelin-B29237
> Cc: vinod.koul@intel.com; dan.j.williams@intel.com; linuxppc-
> dev@lists.ozlabs.org; linux-kernel@vger.kernel.org
> Subject: Re: [PATCH] fsldma: fix performance degradation by optimizing
> spinlock use.
>
> On Wed, Jan 11, 2012 at 07:54:55AM +0000, Shi Xuelin-B29237 wrote:
> > Hello Iris,
> >
> > As we discussed in the previous patch, I add one smp_mb() in
> fsl_tx_status.
> > In my testing with iozone, this smp_mb() could cause 1%~2% performance
> degradation.
> > Anyway it is acceptable for me. Do you have any other comments?
> >
>
> This patch looks fine to me.
>
> Ira
>
> > -----Original Message-----
> > From: Shi Xuelin-B29237
> > Sent: 2011年12月26日 14:01
> > To: iws@ovro.caltech.edu; vinod.koul@intel.com;
> dan.j.williams@intel.com; linuxppc-dev@lists.ozlabs.org; linux-
> kernel@vger.kernel.org
> > Cc: Shi Xuelin-B29237
> > Subject: [PATCH] fsldma: fix performance degradation by optimizing
> spinlock use.
> >
> > From: Forrest shi <b29237@freescale.com>
> >
> > dma status check function fsl_tx_status is heavily called in
> > a tight loop and the desc lock in fsl_tx_status contended by
> > the dma status update function. this caused the dma performance
> > degrades much.
> >
> > this patch releases the lock in the fsl_tx_status function, and
> > introduce the smp_mb() to avoid possible memory inconsistency.
> >
> > Signed-off-by: Forrest Shi <xuelin.shi@freescale.com>
> > ---
> > drivers/dma/fsldma.c | 6 +-----
> > 1 files changed, 1 insertions(+), 5 deletions(-)
> >
> > diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c index
> 8a78154..008fb5e 100644
> > --- a/drivers/dma/fsldma.c
> > +++ b/drivers/dma/fsldma.c
> > @@ -986,15 +986,11 @@ static enum dma_status fsl_tx_status(struct
> dma_chan *dchan,
> > struct fsldma_chan *chan = to_fsl_chan(dchan);
> > dma_cookie_t last_complete;
> > dma_cookie_t last_used;
> > - unsigned long flags;
> > -
> > - spin_lock_irqsave(&chan->desc_lock, flags);
> >
> > last_complete = chan->completed_cookie;
> > + smp_mb();
> > last_used = dchan->cookie;
> >
> > - spin_unlock_irqrestore(&chan->desc_lock, flags);
> > -
> > dma_set_tx_state(txstate, last_complete, last_used, 0);
> > return dma_async_is_complete(cookie, last_complete, last_used); }
> > --
> > 1.7.0.4
> >
> >
ÿôèº{.nÇ+‰·Ÿ®‰†+%ŠËÿ±éݶ\x17¥Šwÿº{.nÇ+‰·¥Š{±þG«éÿŠ{ayº\x1dʇڙë,j\a¢f£¢·hšïêÿ‘êçz_è®\x03(階ŠÝ¢j"ú\x1a¶^[m§ÿÿ¾\a«þG«éÿ¢¸?™¨èÚ&£ø§~á¶iO•æ¬z·švØ^\x14\x04\x1a¶^[m§ÿÿÃ\fÿ¶ìÿ¢¸?–I¥
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2012-04-26 2:59 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-12-26 6:01 [PATCH] fsldma: fix performance degradation by optimizing spinlock use b29237
2011-12-26 6:01 ` b29237
2012-01-11 7:54 ` Shi Xuelin-B29237
2012-01-11 7:54 ` Shi Xuelin-B29237
2012-01-11 16:13 ` Ira W. Snyder
2012-04-26 2:03 ` Shi Xuelin-B29237
2012-04-26 2:03 ` Shi Xuelin-B29237
2012-04-26 2:59 ` Li Yang-R58472
2012-04-26 2:59 ` Li Yang-R58472
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.