[PATCH] drm/amdgpu: extend MAX_KIQ_REG_TRY to 1000

[PATCH] drm/amdgpu: extend MAX_KIQ_REG_TRY to 1000

[Jiawei Gu]

Extend retry times of KIQ to avoid starvation situation caused by
long time full access of GPU by other VFs.

Signed-off-by: Jiawei Gu <Jiawei.Gu@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 639db32c1383..e0c797a5f739 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -288,7 +288,7 @@ enum amdgpu_kiq_irq {
 
 #define MAX_KIQ_REG_WAIT       5000 /* in usecs, 5ms */
 #define MAX_KIQ_REG_BAILOUT_INTERVAL   5 /* in msecs, 5ms */
-#define MAX_KIQ_REG_TRY 80 /* 20 -> 80 */
+#define MAX_KIQ_REG_TRY 1000
 
 int amdgpu_device_ip_set_clockgating_state(void *dev,
 					   enum amd_ip_block_type block_type,
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH] drm/amdgpu: extend MAX_KIQ_REG_TRY to 1000

[Christian König]

Am 08.02.21 um 06:45 schrieb Jiawei Gu:
> Extend retry times of KIQ to avoid starvation situation caused by
> long time full access of GPU by other VFs.

In what units is that? We also need the KIQ during interrupt handling 
and that looks like *way* to big for that.

Christian.

>
> Signed-off-by: Jiawei Gu <Jiawei.Gu@amd.com>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> index 639db32c1383..e0c797a5f739 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> @@ -288,7 +288,7 @@ enum amdgpu_kiq_irq {
>   
>   #define MAX_KIQ_REG_WAIT       5000 /* in usecs, 5ms */
>   #define MAX_KIQ_REG_BAILOUT_INTERVAL   5 /* in msecs, 5ms */
> -#define MAX_KIQ_REG_TRY 80 /* 20 -> 80 */
> +#define MAX_KIQ_REG_TRY 1000
>   
>   int amdgpu_device_ip_set_clockgating_state(void *dev,
>   					   enum amd_ip_block_type block_type,

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

RE: [PATCH] drm/amdgpu: extend MAX_KIQ_REG_TRY to 1000

[Gu, JiaWei (Will)]

[AMD Official Use Only - Internal Distribution Only]

Hi Christian,

That's how many times it would retry before give up.
And we always skip this retry routine if we are in interrupt, so it's fine for interrupt condition.

Best regards,
Jiawei

-----Original Message-----
From: Christian König <ckoenig.leichtzumerken@gmail.com> 
Sent: Monday, February 8, 2021 5:28 PM
To: Gu, JiaWei (Will) <JiaWei.Gu@amd.com>; amd-gfx@lists.freedesktop.org
Cc: Deng, Emily <Emily.Deng@amd.com>
Subject: Re: [PATCH] drm/amdgpu: extend MAX_KIQ_REG_TRY to 1000

Am 08.02.21 um 06:45 schrieb Jiawei Gu:
> Extend retry times of KIQ to avoid starvation situation caused by long 
> time full access of GPU by other VFs.

In what units is that? We also need the KIQ during interrupt handling and that looks like *way* to big for that.

Christian.

>
> Signed-off-by: Jiawei Gu <Jiawei.Gu@amd.com>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h 
> b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> index 639db32c1383..e0c797a5f739 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> @@ -288,7 +288,7 @@ enum amdgpu_kiq_irq {
>   
>   #define MAX_KIQ_REG_WAIT       5000 /* in usecs, 5ms */
>   #define MAX_KIQ_REG_BAILOUT_INTERVAL   5 /* in msecs, 5ms */
> -#define MAX_KIQ_REG_TRY 80 /* 20 -> 80 */
> +#define MAX_KIQ_REG_TRY 1000
>   
>   int amdgpu_device_ip_set_clockgating_state(void *dev,
>   					   enum amd_ip_block_type block_type,
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH] drm/amdgpu: extend MAX_KIQ_REG_TRY to 1000

[Christian König]

Hi Jiawei,

ok in this case it's fine with me.

Just please also get a reviewed-by from somebody which has more KIQ 
background than I have.

Thanks,
Christian.

Am 08.02.21 um 11:00 schrieb Gu, JiaWei (Will):
> [AMD Official Use Only - Internal Distribution Only]
>
> Hi Christian,
>
> That's how many times it would retry before give up.
> And we always skip this retry routine if we are in interrupt, so it's fine for interrupt condition.
>
> Best regards,
> Jiawei
>
> -----Original Message-----
> From: Christian König <ckoenig.leichtzumerken@gmail.com>
> Sent: Monday, February 8, 2021 5:28 PM
> To: Gu, JiaWei (Will) <JiaWei.Gu@amd.com>; amd-gfx@lists.freedesktop.org
> Cc: Deng, Emily <Emily.Deng@amd.com>
> Subject: Re: [PATCH] drm/amdgpu: extend MAX_KIQ_REG_TRY to 1000
>
> Am 08.02.21 um 06:45 schrieb Jiawei Gu:
>> Extend retry times of KIQ to avoid starvation situation caused by long
>> time full access of GPU by other VFs.
> In what units is that? We also need the KIQ during interrupt handling and that looks like *way* to big for that.
>
> Christian.
>
>> Signed-off-by: Jiawei Gu <Jiawei.Gu@amd.com>
>> ---
>>    drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 +-
>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>> index 639db32c1383..e0c797a5f739 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>> @@ -288,7 +288,7 @@ enum amdgpu_kiq_irq {
>>    
>>    #define MAX_KIQ_REG_WAIT       5000 /* in usecs, 5ms */
>>    #define MAX_KIQ_REG_BAILOUT_INTERVAL   5 /* in msecs, 5ms */
>> -#define MAX_KIQ_REG_TRY 80 /* 20 -> 80 */
>> +#define MAX_KIQ_REG_TRY 1000
>>    
>>    int amdgpu_device_ip_set_clockgating_state(void *dev,
>>    					   enum amd_ip_block_type block_type,
> _______________________________________________
> amd-gfx mailing list
> amd-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

RE: [PATCH] drm/amdgpu: extend MAX_KIQ_REG_TRY to 1000

[Deng, Emily]

[AMD Official Use Only - Internal Distribution Only]

Reviewed-by: Emily.Deng <Emily.Deng@amd.com>

>-----Original Message-----
>From: Christian König <ckoenig.leichtzumerken@gmail.com>
>Sent: Monday, February 8, 2021 6:05 PM
>To: Gu, JiaWei (Will) <JiaWei.Gu@amd.com>; Koenig, Christian
><Christian.Koenig@amd.com>; amd-gfx@lists.freedesktop.org
>Cc: Deng, Emily <Emily.Deng@amd.com>
>Subject: Re: [PATCH] drm/amdgpu: extend MAX_KIQ_REG_TRY to 1000
>
>Hi Jiawei,
>
>ok in this case it's fine with me.
>
>Just please also get a reviewed-by from somebody which has more KIQ
>background than I have.
>
>Thanks,
>Christian.
>
>Am 08.02.21 um 11:00 schrieb Gu, JiaWei (Will):
>> [AMD Official Use Only - Internal Distribution Only]
>>
>> Hi Christian,
>>
>> That's how many times it would retry before give up.
>> And we always skip this retry routine if we are in interrupt, so it's fine for
>interrupt condition.
>>
>> Best regards,
>> Jiawei
>>
>> -----Original Message-----
>> From: Christian König <ckoenig.leichtzumerken@gmail.com>
>> Sent: Monday, February 8, 2021 5:28 PM
>> To: Gu, JiaWei (Will) <JiaWei.Gu@amd.com>;
>> amd-gfx@lists.freedesktop.org
>> Cc: Deng, Emily <Emily.Deng@amd.com>
>> Subject: Re: [PATCH] drm/amdgpu: extend MAX_KIQ_REG_TRY to 1000
>>
>> Am 08.02.21 um 06:45 schrieb Jiawei Gu:
>>> Extend retry times of KIQ to avoid starvation situation caused by
>>> long time full access of GPU by other VFs.
>> In what units is that? We also need the KIQ during interrupt handling and
>that looks like *way* to big for that.
>>
>> Christian.
>>
>>> Signed-off-by: Jiawei Gu <Jiawei.Gu@amd.com>
>>> ---
>>>    drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 +-
>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>>> index 639db32c1383..e0c797a5f739 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>>> @@ -288,7 +288,7 @@ enum amdgpu_kiq_irq {
>>>
>>>    #define MAX_KIQ_REG_WAIT       5000 /* in usecs, 5ms */
>>>    #define MAX_KIQ_REG_BAILOUT_INTERVAL   5 /* in msecs, 5ms */
>>> -#define MAX_KIQ_REG_TRY 80 /* 20 -> 80 */
>>> +#define MAX_KIQ_REG_TRY 1000
>>>
>>>    int amdgpu_device_ip_set_clockgating_state(void *dev,
>>>       enum amd_ip_block_type
>block_type,
>> _______________________________________________
>> amd-gfx mailing list
>> amd-gfx@lists.freedesktop.org
>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flist
>> s.freedesktop.org%2Fmailman%2Flistinfo%2Famd-
>gfx&amp;data=04%7C01%7CEm
>>
>ily.Deng%40amd.com%7Cd3ba1ae698ec408da87f08d8cc18f7b1%7C3dd8961fe
>4884e
>>
>608e11a82d994e183d%7C0%7C0%7C637483754906300007%7CUnknown%7CT
>WFpbGZsb3
>>
>d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%
>3D%7
>>
>C1000&amp;sdata=%2FLrBdotkzcAepWTBazik9S9ah5ul48DvtCNKN3wYyQU%3
>D&amp;r
>> eserved=0

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx