diff mbox series

fpga: altera-cvp: Increase the credit timeout

Message ID 20220221141127.3765-1-tien.sung.ang@intel.com (mailing list archive)
State New
Headers show
Series fpga: altera-cvp: Increase the credit timeout | expand

Commit Message

Ang, Tien Sung Feb. 21, 2022, 2:11 p.m. UTC
From: Ang Tien Sung <tien.sung.ang@intel.com>

Increase the timeout for SDM (Secure device manager) data credits from
20ms to 40ms. Internal stress tests running at 500 loops failed with the
current timeout of 20ms. At the start of a FPGA configuration, the CVP
host driver reads the transmit credits from SDM. It then sends bitstream
FPGA data to SDM based on the total credits. Each credit allows the
CVP host driver to send 4kBytes of data. There are situations whereby,
the SDM did not respond in time during testing.

Signed-off-by: Ang Tien Sung <tien.sung.ang@intel.com>
---
 drivers/fpga/altera-cvp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Xu Yilun Feb. 25, 2022, 4:02 p.m. UTC | #1
On Mon, Feb 21, 2022 at 10:11:27PM +0800, tien.sung.ang@intel.com wrote:
> From: Ang Tien Sung <tien.sung.ang@intel.com>
> 
> Increase the timeout for SDM (Secure device manager) data credits from
> 20ms to 40ms. Internal stress tests running at 500 loops failed with the
> current timeout of 20ms. At the start of a FPGA configuration, the CVP
> host driver reads the transmit credits from SDM. It then sends bitstream
> FPGA data to SDM based on the total credits. Each credit allows the
> CVP host driver to send 4kBytes of data. There are situations whereby,
> the SDM did not respond in time during testing.
> 
> Signed-off-by: Ang Tien Sung <tien.sung.ang@intel.com>
> ---
>  drivers/fpga/altera-cvp.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/fpga/altera-cvp.c b/drivers/fpga/altera-cvp.c
> index 4ffb9da537d8..5295ff90482b 100644
> --- a/drivers/fpga/altera-cvp.c
> +++ b/drivers/fpga/altera-cvp.c
> @@ -52,7 +52,7 @@
>  /* V2 Defines */
>  #define VSE_CVP_TX_CREDITS		0x49	/* 8bit */
>  
> -#define V2_CREDIT_TIMEOUT_US		20000
> +#define V2_CREDIT_TIMEOUT_US		40000
>  #define V2_CHECK_CREDIT_US		10
>  #define V2_POLL_TIMEOUT_US		1000000
>  #define V2_USER_TIMEOUT_US		500000

Acked-by: Xu Yilun <yilun.xu@intel.com>

> -- 
> 2.25.1
Wu, Hao Feb. 28, 2022, 9:54 a.m. UTC | #2
> Subject: Re: [PATCH] fpga: altera-cvp: Increase the credit timeout
> 
> On Mon, Feb 21, 2022 at 10:11:27PM +0800, tien.sung.ang@intel.com wrote:
> > From: Ang Tien Sung <tien.sung.ang@intel.com>
> >
> > Increase the timeout for SDM (Secure device manager) data credits from
> > 20ms to 40ms. Internal stress tests running at 500 loops failed with the
> > current timeout of 20ms. At the start of a FPGA configuration, the CVP
> > host driver reads the transmit credits from SDM. It then sends bitstream
> > FPGA data to SDM based on the total credits. Each credit allows the
> > CVP host driver to send 4kBytes of data. There are situations whereby,
> > the SDM did not respond in time during testing.

Why this timeout is changed to 40ms not 50ms or 60ms?
What's the time we expected from SDM?
I hope we don't need to change this again.

Thanks
Hao

> >
> > Signed-off-by: Ang Tien Sung <tien.sung.ang@intel.com>
> > ---
> >  drivers/fpga/altera-cvp.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/fpga/altera-cvp.c b/drivers/fpga/altera-cvp.c
> > index 4ffb9da537d8..5295ff90482b 100644
> > --- a/drivers/fpga/altera-cvp.c
> > +++ b/drivers/fpga/altera-cvp.c
> > @@ -52,7 +52,7 @@
> >  /* V2 Defines */
> >  #define VSE_CVP_TX_CREDITS		0x49	/* 8bit */
> >
> > -#define V2_CREDIT_TIMEOUT_US		20000
> > +#define V2_CREDIT_TIMEOUT_US		40000
> >  #define V2_CHECK_CREDIT_US		10
> >  #define V2_POLL_TIMEOUT_US		1000000
> >  #define V2_USER_TIMEOUT_US		500000
> 
> Acked-by: Xu Yilun <yilun.xu@intel.com>
> 
> > --
> > 2.25.1
Ang, Tien Sung March 4, 2022, 8:44 a.m. UTC | #3
Hi Hao,

We have tested the 40ms change with 500 loops and deem this satisfying
based on the recommendations by the Intel FPGA SDM HW team.
SDM was expected to respond within  a maximum time of 20ms +-error margin. 
However, there were cases this weren't the case during tests. 
After many rounds of tests, they conclude that 40ms is doable. 
And, this is  double the previous time-out of 20ms. 
There is a bigger margin added this time. 
Overall, we don't intend to increase the time-out any further. 
Thanks
Tien Sung

-----Original Message-----
From: Wu, Hao <hao.wu@intel.com> 
Sent: Monday, 28 February, 2022 5:54 PM
To: Xu, Yilun <yilun.xu@intel.com>; Ang, Tien Sung <tien.sung.ang@intel.com>
Cc: Moritz Fischer <mdf@kernel.org>; Tom Rix <trix@redhat.com>; linux-fpga@vger.kernel.org
Subject: RE: [PATCH] fpga: altera-cvp: Increase the credit timeout

> Subject: Re: [PATCH] fpga: altera-cvp: Increase the credit timeout
> 
> On Mon, Feb 21, 2022 at 10:11:27PM +0800, tien.sung.ang@intel.com wrote:
> > From: Ang Tien Sung <tien.sung.ang@intel.com>
> >
> > Increase the timeout for SDM (Secure device manager) data credits 
> > from 20ms to 40ms. Internal stress tests running at 500 loops failed 
> > with the current timeout of 20ms. At the start of a FPGA 
> > configuration, the CVP host driver reads the transmit credits from 
> > SDM. It then sends bitstream FPGA data to SDM based on the total 
> > credits. Each credit allows the CVP host driver to send 4kBytes of 
> > data. There are situations whereby, the SDM did not respond in time during testing.

Why this timeout is changed to 40ms not 50ms or 60ms?
What's the time we expected from SDM?
I hope we don't need to change this again.

Thanks
Hao

> >
> > Signed-off-by: Ang Tien Sung <tien.sung.ang@intel.com>
> > ---
> >  drivers/fpga/altera-cvp.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/fpga/altera-cvp.c b/drivers/fpga/altera-cvp.c 
> > index 4ffb9da537d8..5295ff90482b 100644
> > --- a/drivers/fpga/altera-cvp.c
> > +++ b/drivers/fpga/altera-cvp.c
> > @@ -52,7 +52,7 @@
> >  /* V2 Defines */
> >  #define VSE_CVP_TX_CREDITS		0x49	/* 8bit */
> >
> > -#define V2_CREDIT_TIMEOUT_US		20000
> > +#define V2_CREDIT_TIMEOUT_US		40000
> >  #define V2_CHECK_CREDIT_US		10
> >  #define V2_POLL_TIMEOUT_US		1000000
> >  #define V2_USER_TIMEOUT_US		500000
> 
> Acked-by: Xu Yilun <yilun.xu@intel.com>
> 
> > --
> > 2.25.1
Wu, Hao March 11, 2022, 2:55 a.m. UTC | #4
> -----Original Message-----
> From: Ang, Tien Sung <tien.sung.ang@intel.com>
> Sent: Friday, March 4, 2022 4:45 PM
> To: Wu, Hao <hao.wu@intel.com>; Xu, Yilun <yilun.xu@intel.com>
> Cc: Moritz Fischer <mdf@kernel.org>; Tom Rix <trix@redhat.com>; linux-
> fpga@vger.kernel.org
> Subject: RE: [PATCH] fpga: altera-cvp: Increase the credit timeout
> 
> Hi Hao,
> 
> We have tested the 40ms change with 500 loops and deem this satisfying
> based on the recommendations by the Intel FPGA SDM HW team.
> SDM was expected to respond within  a maximum time of 20ms +-error margin.
> However, there were cases this weren't the case during tests.
> After many rounds of tests, they conclude that 40ms is doable.
> And, this is  double the previous time-out of 20ms.
> There is a bigger margin added this time.
> Overall, we don't intend to increase the time-out any further.

I see. Thanks.

Acked-by: Wu Hao <hao.wu@intel.com>

Hao

> Thanks
> Tien Sung
> 
> -----Original Message-----
> From: Wu, Hao <hao.wu@intel.com>
> Sent: Monday, 28 February, 2022 5:54 PM
> To: Xu, Yilun <yilun.xu@intel.com>; Ang, Tien Sung <tien.sung.ang@intel.com>
> Cc: Moritz Fischer <mdf@kernel.org>; Tom Rix <trix@redhat.com>; linux-
> fpga@vger.kernel.org
> Subject: RE: [PATCH] fpga: altera-cvp: Increase the credit timeout
> 
> > Subject: Re: [PATCH] fpga: altera-cvp: Increase the credit timeout
> >
> > On Mon, Feb 21, 2022 at 10:11:27PM +0800, tien.sung.ang@intel.com wrote:
> > > From: Ang Tien Sung <tien.sung.ang@intel.com>
> > >
> > > Increase the timeout for SDM (Secure device manager) data credits
> > > from 20ms to 40ms. Internal stress tests running at 500 loops failed
> > > with the current timeout of 20ms. At the start of a FPGA
> > > configuration, the CVP host driver reads the transmit credits from
> > > SDM. It then sends bitstream FPGA data to SDM based on the total
> > > credits. Each credit allows the CVP host driver to send 4kBytes of
> > > data. There are situations whereby, the SDM did not respond in time during
> testing.
> 
> Why this timeout is changed to 40ms not 50ms or 60ms?
> What's the time we expected from SDM?
> I hope we don't need to change this again.
> 
> Thanks
> Hao
> 
> > >
> > > Signed-off-by: Ang Tien Sung <tien.sung.ang@intel.com>
> > > ---
> > >  drivers/fpga/altera-cvp.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/fpga/altera-cvp.c b/drivers/fpga/altera-cvp.c
> > > index 4ffb9da537d8..5295ff90482b 100644
> > > --- a/drivers/fpga/altera-cvp.c
> > > +++ b/drivers/fpga/altera-cvp.c
> > > @@ -52,7 +52,7 @@
> > >  /* V2 Defines */
> > >  #define VSE_CVP_TX_CREDITS		0x49	/* 8bit */
> > >
> > > -#define V2_CREDIT_TIMEOUT_US		20000
> > > +#define V2_CREDIT_TIMEOUT_US		40000
> > >  #define V2_CHECK_CREDIT_US		10
> > >  #define V2_POLL_TIMEOUT_US		1000000
> > >  #define V2_USER_TIMEOUT_US		500000
> >
> > Acked-by: Xu Yilun <yilun.xu@intel.com>
> >
> > > --
> > > 2.25.1
diff mbox series

Patch

diff --git a/drivers/fpga/altera-cvp.c b/drivers/fpga/altera-cvp.c
index 4ffb9da537d8..5295ff90482b 100644
--- a/drivers/fpga/altera-cvp.c
+++ b/drivers/fpga/altera-cvp.c
@@ -52,7 +52,7 @@ 
 /* V2 Defines */
 #define VSE_CVP_TX_CREDITS		0x49	/* 8bit */
 
-#define V2_CREDIT_TIMEOUT_US		20000
+#define V2_CREDIT_TIMEOUT_US		40000
 #define V2_CHECK_CREDIT_US		10
 #define V2_POLL_TIMEOUT_US		1000000
 #define V2_USER_TIMEOUT_US		500000