linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCHv2] fpga: stratix10-soc: remove the pre-set reconfiguration condition
@ 2020-05-15 14:35 richard.gong
  2020-05-29 13:15 ` Richard Gong
  0 siblings, 1 reply; 5+ messages in thread
From: richard.gong @ 2020-05-15 14:35 UTC (permalink / raw)
  To: mdf; +Cc: linux-fpga, linux-kernel, dinguyen, richard.gong, Richard Gong

From: Richard Gong <richard.gong@intel.com>

The reconfiguration mode is pre-set by driver as the full reconfiguration.
As a result, user have to change code and recompile the drivers if he or
she wants to perform a partial reconfiguration. Removing the pre-set
reconfiguration condition so that user can select full or partial
reconfiguration via overlay device tree without recompiling the drivers.

Also add an error message if the configuration request is failure.

Signed-off-by: Richard Gong <richard.gong@intel.com>
---
v2: define and use constant values
---
 drivers/fpga/stratix10-soc.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/drivers/fpga/stratix10-soc.c b/drivers/fpga/stratix10-soc.c
index 44b7c56..4d52a80 100644
--- a/drivers/fpga/stratix10-soc.c
+++ b/drivers/fpga/stratix10-soc.c
@@ -14,9 +14,13 @@
 /*
  * FPGA programming requires a higher level of privilege (EL3), per the SoC
  * design.
+ * SoC firmware supports full and partial reconfiguration.
  */
 #define NUM_SVC_BUFS	4
 #define SVC_BUF_SIZE	SZ_512K
+#define FULL_RECONFIG_FLAG	0
+#define PARTIAL_RECONFIG_FLAG	1
+
 
 /* Indicates buffer is in use if set */
 #define SVC_BUF_LOCK	0
@@ -182,12 +186,12 @@ static int s10_ops_write_init(struct fpga_manager *mgr,
 	uint i;
 	int ret;
 
-	ctype.flags = 0;
 	if (info->flags & FPGA_MGR_PARTIAL_RECONFIG) {
 		dev_dbg(dev, "Requesting partial reconfiguration.\n");
-		ctype.flags |= BIT(COMMAND_RECONFIG_FLAG_PARTIAL);
+		ctype.flags = PARTIAL_RECONFIG_FLAG;
 	} else {
 		dev_dbg(dev, "Requesting full reconfiguration.\n");
+		ctype.flags = FULL_RECONFIG_FLAG;
 	}
 
 	reinit_completion(&priv->status_return_completion);
@@ -210,6 +214,7 @@ static int s10_ops_write_init(struct fpga_manager *mgr,
 
 	ret = 0;
 	if (!test_and_clear_bit(SVC_STATUS_OK, &priv->status)) {
+		dev_err(dev, "RECONFIG_REQUEST failed\n");
 		ret = -ETIMEDOUT;
 		goto init_done;
 	}
-- 
2.7.4


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCHv2] fpga: stratix10-soc: remove the pre-set reconfiguration condition
  2020-05-15 14:35 [PATCHv2] fpga: stratix10-soc: remove the pre-set reconfiguration condition richard.gong
@ 2020-05-29 13:15 ` Richard Gong
  2020-05-29 16:18   ` Russ Weight
  2020-05-31 19:49   ` Moritz Fischer
  0 siblings, 2 replies; 5+ messages in thread
From: Richard Gong @ 2020-05-29 13:15 UTC (permalink / raw)
  To: mdf; +Cc: linux-fpga, linux-kernel, dinguyen, Richard Gong

Hi Moritz,

Sorry for asking.

When you get chance, can you review my version 2 patch submitted on 
05/15/20?

Regards,
Richard

On 5/15/20 9:35 AM, richard.gong@linux.intel.com wrote:
> From: Richard Gong <richard.gong@intel.com>
> 
> The reconfiguration mode is pre-set by driver as the full reconfiguration.
> As a result, user have to change code and recompile the drivers if he or
> she wants to perform a partial reconfiguration. Removing the pre-set
> reconfiguration condition so that user can select full or partial
> reconfiguration via overlay device tree without recompiling the drivers.
> 
> Also add an error message if the configuration request is failure.
> 
> Signed-off-by: Richard Gong <richard.gong@intel.com>
> ---
> v2: define and use constant values
> ---
>   drivers/fpga/stratix10-soc.c | 9 +++++++--
>   1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/fpga/stratix10-soc.c b/drivers/fpga/stratix10-soc.c
> index 44b7c56..4d52a80 100644
> --- a/drivers/fpga/stratix10-soc.c
> +++ b/drivers/fpga/stratix10-soc.c
> @@ -14,9 +14,13 @@
>   /*
>    * FPGA programming requires a higher level of privilege (EL3), per the SoC
>    * design.
> + * SoC firmware supports full and partial reconfiguration.
>    */
>   #define NUM_SVC_BUFS	4
>   #define SVC_BUF_SIZE	SZ_512K
> +#define FULL_RECONFIG_FLAG	0
> +#define PARTIAL_RECONFIG_FLAG	1
> +
>   
>   /* Indicates buffer is in use if set */
>   #define SVC_BUF_LOCK	0
> @@ -182,12 +186,12 @@ static int s10_ops_write_init(struct fpga_manager *mgr,
>   	uint i;
>   	int ret;
>   
> -	ctype.flags = 0;
>   	if (info->flags & FPGA_MGR_PARTIAL_RECONFIG) {
>   		dev_dbg(dev, "Requesting partial reconfiguration.\n");
> -		ctype.flags |= BIT(COMMAND_RECONFIG_FLAG_PARTIAL);
> +		ctype.flags = PARTIAL_RECONFIG_FLAG;
>   	} else {
>   		dev_dbg(dev, "Requesting full reconfiguration.\n");
> +		ctype.flags = FULL_RECONFIG_FLAG;
>   	}
>   
>   	reinit_completion(&priv->status_return_completion);
> @@ -210,6 +214,7 @@ static int s10_ops_write_init(struct fpga_manager *mgr,
>   
>   	ret = 0;
>   	if (!test_and_clear_bit(SVC_STATUS_OK, &priv->status)) {
> +		dev_err(dev, "RECONFIG_REQUEST failed\n");
>   		ret = -ETIMEDOUT;
>   		goto init_done;
>   	}
> 

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCHv2] fpga: stratix10-soc: remove the pre-set reconfiguration condition
  2020-05-29 13:15 ` Richard Gong
@ 2020-05-29 16:18   ` Russ Weight
  2020-05-31 19:49   ` Moritz Fischer
  1 sibling, 0 replies; 5+ messages in thread
From: Russ Weight @ 2020-05-29 16:18 UTC (permalink / raw)
  To: Richard Gong, mdf; +Cc: linux-fpga, linux-kernel, dinguyen, Richard Gong

Minor comment below.

- Russ

On 5/29/20 6:15 AM, Richard Gong wrote:
> Hi Moritz,
>
> Sorry for asking.
>
> When you get chance, can you review my version 2 patch submitted on 
> 05/15/20?
>
> Regards,
> Richard
>
> On 5/15/20 9:35 AM, richard.gong@linux.intel.com wrote:
>> From: Richard Gong <richard.gong@intel.com>
>>
>> The reconfiguration mode is pre-set by driver as the full 
>> reconfiguration.
>> As a result, user have to change code and recompile the drivers if he or
>> she wants to perform a partial reconfiguration. Removing the pre-set
>> reconfiguration condition so that user can select full or partial
>> reconfiguration via overlay device tree without recompiling the drivers.
>>
>> Also add an error message if the configuration request is failure.
s/is failure/fails/
>>
>> Signed-off-by: Richard Gong <richard.gong@intel.com>
>> ---
>> v2: define and use constant values
>> ---
>>   drivers/fpga/stratix10-soc.c | 9 +++++++--
>>   1 file changed, 7 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/fpga/stratix10-soc.c b/drivers/fpga/stratix10-soc.c
>> index 44b7c56..4d52a80 100644
>> --- a/drivers/fpga/stratix10-soc.c
>> +++ b/drivers/fpga/stratix10-soc.c
>> @@ -14,9 +14,13 @@
>>   /*
>>    * FPGA programming requires a higher level of privilege (EL3), per 
>> the SoC
>>    * design.
>> + * SoC firmware supports full and partial reconfiguration.
>>    */
>>   #define NUM_SVC_BUFS    4
>>   #define SVC_BUF_SIZE    SZ_512K
>> +#define FULL_RECONFIG_FLAG    0
>> +#define PARTIAL_RECONFIG_FLAG    1
>> +
>>     /* Indicates buffer is in use if set */
>>   #define SVC_BUF_LOCK    0
>> @@ -182,12 +186,12 @@ static int s10_ops_write_init(struct 
>> fpga_manager *mgr,
>>       uint i;
>>       int ret;
>>   -    ctype.flags = 0;
>>       if (info->flags & FPGA_MGR_PARTIAL_RECONFIG) {
>>           dev_dbg(dev, "Requesting partial reconfiguration.\n");
>> -        ctype.flags |= BIT(COMMAND_RECONFIG_FLAG_PARTIAL);
>> +        ctype.flags = PARTIAL_RECONFIG_FLAG;
>>       } else {
>>           dev_dbg(dev, "Requesting full reconfiguration.\n");
>> +        ctype.flags = FULL_RECONFIG_FLAG;
>>       }
>> reinit_completion(&priv->status_return_completion);
>> @@ -210,6 +214,7 @@ static int s10_ops_write_init(struct fpga_manager 
>> *mgr,
>>         ret = 0;
>>       if (!test_and_clear_bit(SVC_STATUS_OK, &priv->status)) {
>> +        dev_err(dev, "RECONFIG_REQUEST failed\n");
>>           ret = -ETIMEDOUT;
>>           goto init_done;
>>       }
>>

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCHv2] fpga: stratix10-soc: remove the pre-set reconfiguration condition
  2020-05-29 13:15 ` Richard Gong
  2020-05-29 16:18   ` Russ Weight
@ 2020-05-31 19:49   ` Moritz Fischer
  2020-06-01 14:44     ` Richard Gong
  1 sibling, 1 reply; 5+ messages in thread
From: Moritz Fischer @ 2020-05-31 19:49 UTC (permalink / raw)
  To: Richard Gong; +Cc: mdf, linux-fpga, linux-kernel, dinguyen, Richard Gong

On Fri, May 29, 2020 at 08:15:15AM -0500, Richard Gong wrote:
> Hi Moritz,
> 
> Sorry for asking.
> 
> When you get chance, can you review my version 2 patch submitted on
> 05/15/20?
> 
> Regards,
> Richard
> 
> On 5/15/20 9:35 AM, richard.gong@linux.intel.com wrote:
> > From: Richard Gong <richard.gong@intel.com>
> > 
> > The reconfiguration mode is pre-set by driver as the full reconfiguration.
> > As a result, user have to change code and recompile the drivers if he or
> > she wants to perform a partial reconfiguration. Removing the pre-set
> > reconfiguration condition so that user can select full or partial
> > reconfiguration via overlay device tree without recompiling the drivers.

Can you help me understand? See comment below, I'm not sure how this
change changes the behavior.
> > 
> > Also add an error message if the configuration request is failure.
> > 
> > Signed-off-by: Richard Gong <richard.gong@intel.com>
> > ---
> > v2: define and use constant values
> > ---
> >   drivers/fpga/stratix10-soc.c | 9 +++++++--
> >   1 file changed, 7 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/fpga/stratix10-soc.c b/drivers/fpga/stratix10-soc.c
> > index 44b7c56..4d52a80 100644
> > --- a/drivers/fpga/stratix10-soc.c
> > +++ b/drivers/fpga/stratix10-soc.c
> > @@ -14,9 +14,13 @@
> >   /*
> >    * FPGA programming requires a higher level of privilege (EL3), per the SoC
> >    * design.
> > + * SoC firmware supports full and partial reconfiguration.
Consider:
"The SoC firmware supports full and partial reconfiguration."
> >    */
> >   #define NUM_SVC_BUFS	4
> >   #define SVC_BUF_SIZE	SZ_512K
> > +#define FULL_RECONFIG_FLAG	0
> > +#define PARTIAL_RECONFIG_FLAG	1
> > +
> >   /* Indicates buffer is in use if set */
> >   #define SVC_BUF_LOCK	0
> > @@ -182,12 +186,12 @@ static int s10_ops_write_init(struct fpga_manager *mgr,
> >   	uint i;
> >   	int ret;
> > -	ctype.flags = 0;
> >   	if (info->flags & FPGA_MGR_PARTIAL_RECONFIG) {
> >   		dev_dbg(dev, "Requesting partial reconfiguration.\n");
> > -		ctype.flags |= BIT(COMMAND_RECONFIG_FLAG_PARTIAL);
> > +		ctype.flags = PARTIAL_RECONFIG_FLAG;
> >   	} else {
> >   		dev_dbg(dev, "Requesting full reconfiguration.\n");
> > +		ctype.flags = FULL_RECONFIG_FLAG;
> >   	}
Am I missing something here: Doesn't this do the same as before?

Before:
If info->flags & FPGA_MGR_PARTIAL_RECONFIG -> ctype.flags = 0 |
BIT(COMMAND_RECONFIG_FLAG_PARTIAL) -> 1
and ctype->flags = FULL_RECONFIG -> 0 else.

Now:
If info->flags & FPGA_MGR_PARTIAL_RECONFIG -> ctype.flags = PARTIAL_RECONFIG_FLAG -> 1
ctype->flags = FULL_REECONFIG_FLAG -> 0 else.

Am I missing something here? If I don't set the flag for partial
reconfig I'd end up with full reconfiguration in both cases?
If I do set the flag, I get partial reconfiguration in both cases?

> >   	reinit_completion(&priv->status_return_completion);
> > @@ -210,6 +214,7 @@ static int s10_ops_write_init(struct fpga_manager *mgr,
> >   	ret = 0;
> >   	if (!test_and_clear_bit(SVC_STATUS_OK, &priv->status)) {
> > +		dev_err(dev, "RECONFIG_REQUEST failed\n");
> >   		ret = -ETIMEDOUT;
> >   		goto init_done;
> >   	}
> > 

Thanks,
Moritz

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCHv2] fpga: stratix10-soc: remove the pre-set reconfiguration condition
  2020-05-31 19:49   ` Moritz Fischer
@ 2020-06-01 14:44     ` Richard Gong
  0 siblings, 0 replies; 5+ messages in thread
From: Richard Gong @ 2020-06-01 14:44 UTC (permalink / raw)
  To: Moritz Fischer; +Cc: linux-fpga, linux-kernel, dinguyen, Richard Gong

Hi Moritz,


On 5/31/20 2:49 PM, Moritz Fischer wrote:
> On Fri, May 29, 2020 at 08:15:15AM -0500, Richard Gong wrote:
>> Hi Moritz,
>>
>> Sorry for asking.
>>
>> When you get chance, can you review my version 2 patch submitted on
>> 05/15/20?
>>
>> Regards,
>> Richard
>>
>> On 5/15/20 9:35 AM, richard.gong@linux.intel.com wrote:
>>> From: Richard Gong <richard.gong@intel.com>
>>>
>>> The reconfiguration mode is pre-set by driver as the full reconfiguration.
>>> As a result, user have to change code and recompile the drivers if he or
>>> she wants to perform a partial reconfiguration. Removing the pre-set
>>> reconfiguration condition so that user can select full or partial
>>> reconfiguration via overlay device tree without recompiling the drivers.
> 
> Can you help me understand? See comment below, I'm not sure how this
> change changes the behavior.

Flag COMMAND_RECONFIG_FLAG_PARTIAL is defined in Intel service layer 
driver (include/linux/firmware/intel/stratix10-svc-client.h) and the 
default value is zero. It is obvious that COMMAND_RECONFIG_FLAG_PARTIAL 
should be set to 1 to support partial reconfiguration.

Please discard this FPGA patch, I will submit a patch on Intel service 
layer driver.

Regards,
Richard

>>>
>>> Also add an error message if the configuration request is failure.
>>>
>>> Signed-off-by: Richard Gong <richard.gong@intel.com>
>>> ---
>>> v2: define and use constant values
>>> ---
>>>    drivers/fpga/stratix10-soc.c | 9 +++++++--
>>>    1 file changed, 7 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/fpga/stratix10-soc.c b/drivers/fpga/stratix10-soc.c
>>> index 44b7c56..4d52a80 100644
>>> --- a/drivers/fpga/stratix10-soc.c
>>> +++ b/drivers/fpga/stratix10-soc.c
>>> @@ -14,9 +14,13 @@
>>>    /*
>>>     * FPGA programming requires a higher level of privilege (EL3), per the SoC
>>>     * design.
>>> + * SoC firmware supports full and partial reconfiguration.
> Consider:
> "The SoC firmware supports full and partial reconfiguration."
>>>     */
>>>    #define NUM_SVC_BUFS	4
>>>    #define SVC_BUF_SIZE	SZ_512K
>>> +#define FULL_RECONFIG_FLAG	0
>>> +#define PARTIAL_RECONFIG_FLAG	1
>>> +
>>>    /* Indicates buffer is in use if set */
>>>    #define SVC_BUF_LOCK	0
>>> @@ -182,12 +186,12 @@ static int s10_ops_write_init(struct fpga_manager *mgr,
>>>    	uint i;
>>>    	int ret;
>>> -	ctype.flags = 0;
>>>    	if (info->flags & FPGA_MGR_PARTIAL_RECONFIG) {
>>>    		dev_dbg(dev, "Requesting partial reconfiguration.\n");
>>> -		ctype.flags |= BIT(COMMAND_RECONFIG_FLAG_PARTIAL);
>>> +		ctype.flags = PARTIAL_RECONFIG_FLAG;
>>>    	} else {
>>>    		dev_dbg(dev, "Requesting full reconfiguration.\n");
>>> +		ctype.flags = FULL_RECONFIG_FLAG;
>>>    	}
> Am I missing something here: Doesn't this do the same as before?
> 
> Before:
> If info->flags & FPGA_MGR_PARTIAL_RECONFIG -> ctype.flags = 0 |
> BIT(COMMAND_RECONFIG_FLAG_PARTIAL) -> 1
> and ctype->flags = FULL_RECONFIG -> 0 else.
> 
> Now:
> If info->flags & FPGA_MGR_PARTIAL_RECONFIG -> ctype.flags = PARTIAL_RECONFIG_FLAG -> 1
> ctype->flags = FULL_REECONFIG_FLAG -> 0 else.
> 
> Am I missing something here? If I don't set the flag for partial
> reconfig I'd end up with full reconfiguration in both cases?
> If I do set the flag, I get partial reconfiguration in both cases?
> 
>>>    	reinit_completion(&priv->status_return_completion);
>>> @@ -210,6 +214,7 @@ static int s10_ops_write_init(struct fpga_manager *mgr,
>>>    	ret = 0;
>>>    	if (!test_and_clear_bit(SVC_STATUS_OK, &priv->status)) {
>>> +		dev_err(dev, "RECONFIG_REQUEST failed\n");
>>>    		ret = -ETIMEDOUT;
>>>    		goto init_done;
>>>    	}
>>>
> 
> Thanks,
> Moritz
> 

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2020-06-01 14:29 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-05-15 14:35 [PATCHv2] fpga: stratix10-soc: remove the pre-set reconfiguration condition richard.gong
2020-05-29 13:15 ` Richard Gong
2020-05-29 16:18   ` Russ Weight
2020-05-31 19:49   ` Moritz Fischer
2020-06-01 14:44     ` Richard Gong

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).