* [PATCH v2 2/3] drm/amdgpu: convert code name to ip version for noretry set
2022-03-01 15:01 [PATCH v2 1/3] drm/amdgpu: move amdgpu_gmc_noretry_set after ip_versions populated Yifan Zhang
@ 2022-03-01 15:01 ` Yifan Zhang
2022-03-01 15:24 ` Paul Menzel
2022-03-02 7:47 ` Christian König
2022-03-01 15:01 ` [PATCH v2 3/3] drm/amdgpu: set noretry=1 for gc 10.3.6 Yifan Zhang
` (2 subsequent siblings)
3 siblings, 2 replies; 9+ messages in thread
From: Yifan Zhang @ 2022-03-01 15:01 UTC (permalink / raw)
To: amd-gfx; +Cc: Alexander.Deucher, Yifan Zhang, Ray.Huang, christian.koenig
Use IP version rather than codename for noretry set.
Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 17 ++++++++---------
1 file changed, 8 insertions(+), 9 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index ab75e189bc0b..fbc22b7b6315 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -559,14 +559,14 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
{
struct amdgpu_gmc *gmc = &adev->gmc;
- switch (adev->asic_type) {
- case CHIP_VEGA10:
- case CHIP_VEGA20:
- case CHIP_ARCTURUS:
- case CHIP_ALDEBARAN:
- case CHIP_BEIGE_GOBY:
- case CHIP_YELLOW_CARP:
- case CHIP_RENOIR:
+ switch (adev->ip_versions[GC_HWIP][0]) {
+ case IP_VERSION(9, 0, 1):
+ case IP_VERSION(9, 4, 0):
+ case IP_VERSION(9, 4, 1):
+ case IP_VERSION(9, 4, 2):
+ case IP_VERSION(10, 3, 5):
+ case IP_VERSION(10, 3, 3):
+ case IP_VERSION(9, 3, 0):
/*
* noretry = 0 will cause kfd page fault tests fail
* for some ASICs, so set default to 1 for these ASICs.
@@ -576,7 +576,6 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
else
gmc->noretry = amdgpu_noretry;
break;
- case CHIP_RAVEN:
default:
/* Raven currently has issues with noretry
* regardless of what we decide for other
--
2.25.1
^ permalink raw reply related [flat|nested] 9+ messages in thread* Re: [PATCH v2 2/3] drm/amdgpu: convert code name to ip version for noretry set
2022-03-01 15:01 ` [PATCH v2 2/3] drm/amdgpu: convert code name to ip version for noretry set Yifan Zhang
@ 2022-03-01 15:24 ` Paul Menzel
2022-03-02 7:46 ` Christian König
2022-03-02 7:47 ` Christian König
1 sibling, 1 reply; 9+ messages in thread
From: Paul Menzel @ 2022-03-01 15:24 UTC (permalink / raw)
To: Yifan Zhang; +Cc: alexander.deucher, Ray.Huang, christian.koenig, amd-gfx
Dear Yifan,
Thank you for your patch.
Am 01.03.22 um 16:01 schrieb Yifan Zhang:
> Use IP version rather than codename for noretry set.
Why?
> Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 17 ++++++++---------
> 1 file changed, 8 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
> index ab75e189bc0b..fbc22b7b6315 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
> @@ -559,14 +559,14 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
> {
> struct amdgpu_gmc *gmc = &adev->gmc;
>
> - switch (adev->asic_type) {
> - case CHIP_VEGA10:
> - case CHIP_VEGA20:
> - case CHIP_ARCTURUS:
> - case CHIP_ALDEBARAN:
> - case CHIP_BEIGE_GOBY:
> - case CHIP_YELLOW_CARP:
> - case CHIP_RENOIR:
> + switch (adev->ip_versions[GC_HWIP][0]) {
> + case IP_VERSION(9, 0, 1):
> + case IP_VERSION(9, 4, 0):
> + case IP_VERSION(9, 4, 1):
> + case IP_VERSION(9, 4, 2):
> + case IP_VERSION(10, 3, 5):
> + case IP_VERSION(10, 3, 3):
> + case IP_VERSION(9, 3, 0):
I think, sorting these entries might be useful.
Should the names be added as comments for those not having them memorized?
> /*
> * noretry = 0 will cause kfd page fault tests fail
> * for some ASICs, so set default to 1 for these ASICs.
> @@ -576,7 +576,6 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
> else
> gmc->noretry = amdgpu_noretry;
> break;
> - case CHIP_RAVEN:
Why remove this?
> default:
> /* Raven currently has issues with noretry
> * regardless of what we decide for other
Kind regards,
Paul
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCH v2 2/3] drm/amdgpu: convert code name to ip version for noretry set
2022-03-01 15:24 ` Paul Menzel
@ 2022-03-02 7:46 ` Christian König
2022-03-02 7:52 ` Paul Menzel
0 siblings, 1 reply; 9+ messages in thread
From: Christian König @ 2022-03-02 7:46 UTC (permalink / raw)
To: Paul Menzel, Yifan Zhang
Cc: alexander.deucher, Ray.Huang, christian.koenig, amd-gfx
Am 01.03.22 um 16:24 schrieb Paul Menzel:
> Dear Yifan,
>
>
> Thank you for your patch.
>
> Am 01.03.22 um 16:01 schrieb Yifan Zhang:
>> Use IP version rather than codename for noretry set.
>
> Why?
Why not?
>
>> Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
>> ---
>> drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 17 ++++++++---------
>> 1 file changed, 8 insertions(+), 9 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
>> index ab75e189bc0b..fbc22b7b6315 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
>> @@ -559,14 +559,14 @@ void amdgpu_gmc_noretry_set(struct
>> amdgpu_device *adev)
>> {
>> struct amdgpu_gmc *gmc = &adev->gmc;
>> - switch (adev->asic_type) {
>> - case CHIP_VEGA10:
>> - case CHIP_VEGA20:
>> - case CHIP_ARCTURUS:
>> - case CHIP_ALDEBARAN:
>> - case CHIP_BEIGE_GOBY:
>> - case CHIP_YELLOW_CARP:
>> - case CHIP_RENOIR:
>> + switch (adev->ip_versions[GC_HWIP][0]) {
>> + case IP_VERSION(9, 0, 1):
>> + case IP_VERSION(9, 4, 0):
>> + case IP_VERSION(9, 4, 1):
>> + case IP_VERSION(9, 4, 2):
>> + case IP_VERSION(10, 3, 5):
>> + case IP_VERSION(10, 3, 3):
>> + case IP_VERSION(9, 3, 0):
>
> I think, sorting these entries might be useful.
>
> Should the names be added as comments for those not having them
> memorized?
>
>> /*
>> * noretry = 0 will cause kfd page fault tests fail
>> * for some ASICs, so set default to 1 for these ASICs.
>> @@ -576,7 +576,6 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device
>> *adev)
>> else
>> gmc->noretry = amdgpu_noretry;
>> break;
>> - case CHIP_RAVEN:
>
> Why remove this?
>
>> default:
>> /* Raven currently has issues with noretry
>> * regardless of what we decide for other
>
>
> Kind regards,
>
> Paul
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCH v2 2/3] drm/amdgpu: convert code name to ip version for noretry set
2022-03-02 7:46 ` Christian König
@ 2022-03-02 7:52 ` Paul Menzel
0 siblings, 0 replies; 9+ messages in thread
From: Paul Menzel @ 2022-03-02 7:52 UTC (permalink / raw)
To: Christian König, Yifan Zhang
Cc: alexander.deucher, ray.huang, christian.koenig, amd-gfx
Dear Christian,
Am 02.03.22 um 08:46 schrieb Christian König:
> Am 01.03.22 um 16:24 schrieb Paul Menzel:
[…]
>> Am 01.03.22 um 16:01 schrieb Yifan Zhang:
>>> Use IP version rather than codename for noretry set.
>>
>> Why?
>
> Why not?
No idea as the commit message currently does not document the motivation.
>>> Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
>>> ---
>>> drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 17 ++++++++---------
>>> 1 file changed, 8 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
>>> index ab75e189bc0b..fbc22b7b6315 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
>>> @@ -559,14 +559,14 @@ void amdgpu_gmc_noretry_set(struct
>>> amdgpu_device *adev)
>>> {
>>> struct amdgpu_gmc *gmc = &adev->gmc;
>>> - switch (adev->asic_type) {
>>> - case CHIP_VEGA10:
>>> - case CHIP_VEGA20:
>>> - case CHIP_ARCTURUS:
>>> - case CHIP_ALDEBARAN:
>>> - case CHIP_BEIGE_GOBY:
>>> - case CHIP_YELLOW_CARP:
>>> - case CHIP_RENOIR:
>>> + switch (adev->ip_versions[GC_HWIP][0]) {
>>> + case IP_VERSION(9, 0, 1):
>>> + case IP_VERSION(9, 4, 0):
>>> + case IP_VERSION(9, 4, 1):
>>> + case IP_VERSION(9, 4, 2):
>>> + case IP_VERSION(10, 3, 5):
>>> + case IP_VERSION(10, 3, 3):
>>> + case IP_VERSION(9, 3, 0):
>>
>> I think, sorting these entries might be useful.
>>
>> Should the names be added as comments for those not having them
>> memorized?
>>
>>> /*
>>> * noretry = 0 will cause kfd page fault tests fail
>>> * for some ASICs, so set default to 1 for these ASICs.
>>> @@ -576,7 +576,6 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device
>>> *adev)
>>> else
>>> gmc->noretry = amdgpu_noretry;
>>> break;
>>> - case CHIP_RAVEN:
>>
>> Why remove this?
>>
>>> default:
>>> /* Raven currently has issues with noretry
>>> * regardless of what we decide for other
Kind regards,
Paul
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v2 2/3] drm/amdgpu: convert code name to ip version for noretry set
2022-03-01 15:01 ` [PATCH v2 2/3] drm/amdgpu: convert code name to ip version for noretry set Yifan Zhang
2022-03-01 15:24 ` Paul Menzel
@ 2022-03-02 7:47 ` Christian König
1 sibling, 0 replies; 9+ messages in thread
From: Christian König @ 2022-03-02 7:47 UTC (permalink / raw)
To: Yifan Zhang, amd-gfx; +Cc: Alexander.Deucher, Ray.Huang
Am 01.03.22 um 16:01 schrieb Yifan Zhang:
> Use IP version rather than codename for noretry set.
>
> Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 17 ++++++++---------
> 1 file changed, 8 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
> index ab75e189bc0b..fbc22b7b6315 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
> @@ -559,14 +559,14 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
> {
> struct amdgpu_gmc *gmc = &adev->gmc;
>
> - switch (adev->asic_type) {
> - case CHIP_VEGA10:
> - case CHIP_VEGA20:
> - case CHIP_ARCTURUS:
> - case CHIP_ALDEBARAN:
> - case CHIP_BEIGE_GOBY:
> - case CHIP_YELLOW_CARP:
> - case CHIP_RENOIR:
> + switch (adev->ip_versions[GC_HWIP][0]) {
> + case IP_VERSION(9, 0, 1):
> + case IP_VERSION(9, 4, 0):
> + case IP_VERSION(9, 4, 1):
> + case IP_VERSION(9, 4, 2):
> + case IP_VERSION(10, 3, 5):
> + case IP_VERSION(10, 3, 3):
> + case IP_VERSION(9, 3, 0):
Maybe sort those?
Apart from that Acked-by: Christian König <christian.koenig@amd.com>
Regards,
Christian.
> /*
> * noretry = 0 will cause kfd page fault tests fail
> * for some ASICs, so set default to 1 for these ASICs.
> @@ -576,7 +576,6 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
> else
> gmc->noretry = amdgpu_noretry;
> break;
> - case CHIP_RAVEN:
> default:
> /* Raven currently has issues with noretry
> * regardless of what we decide for other
^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH v2 3/3] drm/amdgpu: set noretry=1 for gc 10.3.6
2022-03-01 15:01 [PATCH v2 1/3] drm/amdgpu: move amdgpu_gmc_noretry_set after ip_versions populated Yifan Zhang
2022-03-01 15:01 ` [PATCH v2 2/3] drm/amdgpu: convert code name to ip version for noretry set Yifan Zhang
@ 2022-03-01 15:01 ` Yifan Zhang
2022-03-01 15:11 ` [PATCH v2 1/3] drm/amdgpu: move amdgpu_gmc_noretry_set after ip_versions populated Alex Deucher
2022-03-02 8:49 ` Huang Rui
3 siblings, 0 replies; 9+ messages in thread
From: Yifan Zhang @ 2022-03-01 15:01 UTC (permalink / raw)
To: amd-gfx; +Cc: Alexander.Deucher, Yifan Zhang, Ray.Huang, christian.koenig
this patch to set noretry=1 for gc 10.3.6.
Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index fbc22b7b6315..7c2a9555b7cc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -567,6 +567,7 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
case IP_VERSION(10, 3, 5):
case IP_VERSION(10, 3, 3):
case IP_VERSION(9, 3, 0):
+ case IP_VERSION(10, 3, 6):
/*
* noretry = 0 will cause kfd page fault tests fail
* for some ASICs, so set default to 1 for these ASICs.
--
2.25.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH v2 1/3] drm/amdgpu: move amdgpu_gmc_noretry_set after ip_versions populated
2022-03-01 15:01 [PATCH v2 1/3] drm/amdgpu: move amdgpu_gmc_noretry_set after ip_versions populated Yifan Zhang
2022-03-01 15:01 ` [PATCH v2 2/3] drm/amdgpu: convert code name to ip version for noretry set Yifan Zhang
2022-03-01 15:01 ` [PATCH v2 3/3] drm/amdgpu: set noretry=1 for gc 10.3.6 Yifan Zhang
@ 2022-03-01 15:11 ` Alex Deucher
2022-03-02 8:49 ` Huang Rui
3 siblings, 0 replies; 9+ messages in thread
From: Alex Deucher @ 2022-03-01 15:11 UTC (permalink / raw)
To: Yifan Zhang; +Cc: Deucher, Alexander, Huang Rui, Christian Koenig, amd-gfx list
Series is:
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
On Tue, Mar 1, 2022 at 10:02 AM Yifan Zhang <yifan1.zhang@amd.com> wrote:
>
> otherwise adev->ip_versions is still empty when amdgpu_gmc_noretry_set
> is called.
>
> Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index 6113ddc765a7..61a6a7920c76 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -1554,7 +1554,6 @@ static int amdgpu_device_check_arguments(struct amdgpu_device *adev)
>
> amdgpu_gmc_tmz_set(adev);
>
> - amdgpu_gmc_noretry_set(adev);
>
> return 0;
> }
> @@ -3641,6 +3640,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
> if (r)
> return r;
>
> + amdgpu_gmc_noretry_set(adev);
> /* Need to get xgmi info early to decide the reset behavior*/
> if (adev->gmc.xgmi.supported) {
> r = adev->gfxhub.funcs->get_xgmi_info(adev);
> --
> 2.25.1
>
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCH v2 1/3] drm/amdgpu: move amdgpu_gmc_noretry_set after ip_versions populated
2022-03-01 15:01 [PATCH v2 1/3] drm/amdgpu: move amdgpu_gmc_noretry_set after ip_versions populated Yifan Zhang
` (2 preceding siblings ...)
2022-03-01 15:11 ` [PATCH v2 1/3] drm/amdgpu: move amdgpu_gmc_noretry_set after ip_versions populated Alex Deucher
@ 2022-03-02 8:49 ` Huang Rui
3 siblings, 0 replies; 9+ messages in thread
From: Huang Rui @ 2022-03-02 8:49 UTC (permalink / raw)
To: Zhang, Yifan; +Cc: Deucher, Alexander, Koenig, Christian, amd-gfx
On Tue, Mar 01, 2022 at 11:01:19PM +0800, Zhang, Yifan wrote:
> otherwise adev->ip_versions is still empty when amdgpu_gmc_noretry_set
> is called.
>
> Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
Series are Reviewed-by: Huang Rui <ray.huang@amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index 6113ddc765a7..61a6a7920c76 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -1554,7 +1554,6 @@ static int amdgpu_device_check_arguments(struct amdgpu_device *adev)
>
> amdgpu_gmc_tmz_set(adev);
>
> - amdgpu_gmc_noretry_set(adev);
>
> return 0;
> }
> @@ -3641,6 +3640,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
> if (r)
> return r;
>
> + amdgpu_gmc_noretry_set(adev);
> /* Need to get xgmi info early to decide the reset behavior*/
> if (adev->gmc.xgmi.supported) {
> r = adev->gfxhub.funcs->get_xgmi_info(adev);
> --
> 2.25.1
>
^ permalink raw reply [flat|nested] 9+ messages in thread