[PATCH 1/2] drm/amdgpu: fix s3 resume back, uvd dpm randomly can't disable.

[PATCH 1/2] drm/amdgpu: fix s3 resume back, uvd dpm randomly can't disable.

[Rex Zhu]

the value of last_mclk_dpm_enable_mask will be changed if
other clients(vce,dal) trigger set power state between enable
and disable uvd dpm.

Change-Id: I5bdca2ff7d401243780c71c8d73fdc1e8febaae1
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/ci_dpm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
index 6ecfc48..966e5bf 100644
--- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
+++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
@@ -4081,7 +4081,7 @@ static int ci_enable_uvd_dpm(struct amdgpu_device *adev, bool enable)
 							  pi->dpm_level_enable_mask.mclk_dpm_enable_mask);
 		}
 	} else {
-		if (pi->last_mclk_dpm_enable_mask & 0x1) {
+		if (pi->uvd_enabled == true) {
 			pi->uvd_enabled = false;
 			pi->dpm_level_enable_mask.mclk_dpm_enable_mask |= 1;
 			amdgpu_ci_send_msg_to_smc_with_parameter(adev,
-- 
1.9.1

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

[PATCH 2/2] drm/amdgpu: when suspend, set boot state instand of disable dpm.

[Rex Zhu]

fix pm-hibernat bug, when suspend/resume, dpm start failed.

Change-Id: I03784f0049563b11721ebe905d9bea40ecc82170
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/ci_dpm.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
index 966e5bf..27b1fb4 100644
--- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
+++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
@@ -5276,7 +5276,7 @@ static int ci_dpm_enable(struct amdgpu_device *adev)
 	int ret;
 
 	if (amdgpu_ci_is_smc_running(adev))
-		return -EINVAL;
+		return 0;
 	if (pi->voltage_control != CISLANDS_VOLTAGE_CONTROL_NONE) {
 		ci_enable_voltage_control(adev);
 		ret = ci_construct_voltage_tables(adev);
@@ -6345,12 +6345,17 @@ static int ci_dpm_suspend(void *handle)
 
 	if (adev->pm.dpm_enabled) {
 		mutex_lock(&adev->pm.mutex);
-		/* disable dpm */
-		ci_dpm_disable(adev);
-		/* reset the power state */
-		adev->pm.dpm.current_ps = adev->pm.dpm.requested_ps = adev->pm.dpm.boot_ps;
+		amdgpu_irq_put(adev, &adev->pm.dpm.thermal.irq,
+			       AMDGPU_THERMAL_IRQ_LOW_TO_HIGH);
+		amdgpu_irq_put(adev, &adev->pm.dpm.thermal.irq,
+			       AMDGPU_THERMAL_IRQ_HIGH_TO_LOW);
+		adev->pm.dpm.user_state = POWER_STATE_TYPE_INTERNAL_BOOT;
+		adev->pm.dpm.state = POWER_STATE_TYPE_INTERNAL_BOOT;
 		mutex_unlock(&adev->pm.mutex);
+		amdgpu_pm_compute_clocks(adev);
+
 	}
+
 	return 0;
 }
 
@@ -6368,6 +6373,8 @@ static int ci_dpm_resume(void *handle)
 			adev->pm.dpm_enabled = false;
 		else
 			adev->pm.dpm_enabled = true;
+		adev->pm.dpm.user_state = POWER_STATE_TYPE_BALANCED;
+		adev->pm.dpm.state = POWER_STATE_TYPE_BALANCED;
 		mutex_unlock(&adev->pm.mutex);
 		if (adev->pm.dpm_enabled)
 			amdgpu_pm_compute_clocks(adev);
-- 
1.9.1

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

Re: [PATCH 1/2] drm/amdgpu: fix s3 resume back, uvd dpm randomly can't disable.

[Christian König]

[-- Attachment #1.1: Type: text/plain, Size: 1127 bytes --]

Am 21.10.2016 um 12:13 schrieb Rex Zhu:
> the value of last_mclk_dpm_enable_mask will be changed if
> other clients(vce,dal) trigger set power state between enable
> and disable uvd dpm.
>
> Change-Id: I5bdca2ff7d401243780c71c8d73fdc1e8febaae1
> Signed-off-by: Rex Zhu <Rex.Zhu-5C7GfCeVMHo@public.gmane.org>

Acked-by: Christian König <christian.koenig-5C7GfCeVMHo@public.gmane.org> for the series/./

> ---
>   drivers/gpu/drm/amd/amdgpu/ci_dpm.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
> index 6ecfc48..966e5bf 100644
> --- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
> +++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
> @@ -4081,7 +4081,7 @@ static int ci_enable_uvd_dpm(struct amdgpu_device *adev, bool enable)
>   							  pi->dpm_level_enable_mask.mclk_dpm_enable_mask);
>   		}
>   	} else {
> -		if (pi->last_mclk_dpm_enable_mask & 0x1) {
> +		if (pi->uvd_enabled == true) {
>   			pi->uvd_enabled = false;
>   			pi->dpm_level_enable_mask.mclk_dpm_enable_mask |= 1;
>   			amdgpu_ci_send_msg_to_smc_with_parameter(adev,



[-- Attachment #1.2: Type: text/html, Size: 1883 bytes --]

[-- Attachment #2: Type: text/plain, Size: 154 bytes --]

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

Re: [PATCH 1/2] drm/amdgpu: fix s3 resume back, uvd dpm randomly can't disable.

[Alex Deucher]

On Fri, Oct 21, 2016 at 6:13 AM, Rex Zhu <Rex.Zhu@amd.com> wrote:
> the value of last_mclk_dpm_enable_mask will be changed if
> other clients(vce,dal) trigger set power state between enable
> and disable uvd dpm.
>
> Change-Id: I5bdca2ff7d401243780c71c8d73fdc1e8febaae1
> Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

> ---
>  drivers/gpu/drm/amd/amdgpu/ci_dpm.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
> index 6ecfc48..966e5bf 100644
> --- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
> +++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
> @@ -4081,7 +4081,7 @@ static int ci_enable_uvd_dpm(struct amdgpu_device *adev, bool enable)
>                                                           pi->dpm_level_enable_mask.mclk_dpm_enable_mask);
>                 }
>         } else {
> -               if (pi->last_mclk_dpm_enable_mask & 0x1) {
> +               if (pi->uvd_enabled == true) {
>                         pi->uvd_enabled = false;
>                         pi->dpm_level_enable_mask.mclk_dpm_enable_mask |= 1;
>                         amdgpu_ci_send_msg_to_smc_with_parameter(adev,
> --
> 1.9.1
>
> _______________________________________________
> 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 2/2] drm/amdgpu: when suspend, set boot state instand of disable dpm.

[Alex Deucher]

On Fri, Oct 21, 2016 at 6:13 AM, Rex Zhu <Rex.Zhu@amd.com> wrote:
> fix pm-hibernat bug, when suspend/resume, dpm start failed.
>
> Change-Id: I03784f0049563b11721ebe905d9bea40ecc82170
> Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
> ---
>  drivers/gpu/drm/amd/amdgpu/ci_dpm.c | 17 ++++++++++++-----
>  1 file changed, 12 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
> index 966e5bf..27b1fb4 100644
> --- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
> +++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
> @@ -5276,7 +5276,7 @@ static int ci_dpm_enable(struct amdgpu_device *adev)
>         int ret;
>
>         if (amdgpu_ci_is_smc_running(adev))
> -               return -EINVAL;
> +               return 0;


I think this hunk should be a separate patch.  Also, should we
actually return here or just print a message?

>         if (pi->voltage_control != CISLANDS_VOLTAGE_CONTROL_NONE) {
>                 ci_enable_voltage_control(adev);
>                 ret = ci_construct_voltage_tables(adev);
> @@ -6345,12 +6345,17 @@ static int ci_dpm_suspend(void *handle)
>
>         if (adev->pm.dpm_enabled) {
>                 mutex_lock(&adev->pm.mutex);
> -               /* disable dpm */
> -               ci_dpm_disable(adev);
> -               /* reset the power state */
> -               adev->pm.dpm.current_ps = adev->pm.dpm.requested_ps = adev->pm.dpm.boot_ps;
> +               amdgpu_irq_put(adev, &adev->pm.dpm.thermal.irq,
> +                              AMDGPU_THERMAL_IRQ_LOW_TO_HIGH);
> +               amdgpu_irq_put(adev, &adev->pm.dpm.thermal.irq,
> +                              AMDGPU_THERMAL_IRQ_HIGH_TO_LOW);
> +               adev->pm.dpm.user_state = POWER_STATE_TYPE_INTERNAL_BOOT;
> +               adev->pm.dpm.state = POWER_STATE_TYPE_INTERNAL_BOOT;


Ideally we would save the selected state before forcing it to boot
mode in suspend and then restore the user state in resume.

Alex

>                 mutex_unlock(&adev->pm.mutex);
> +               amdgpu_pm_compute_clocks(adev);
> +
>         }
> +
>         return 0;
>  }
>
> @@ -6368,6 +6373,8 @@ static int ci_dpm_resume(void *handle)
>                         adev->pm.dpm_enabled = false;
>                 else
>                         adev->pm.dpm_enabled = true;
> +               adev->pm.dpm.user_state = POWER_STATE_TYPE_BALANCED;
> +               adev->pm.dpm.state = POWER_STATE_TYPE_BALANCED;
>                 mutex_unlock(&adev->pm.mutex);
>                 if (adev->pm.dpm_enabled)
>                         amdgpu_pm_compute_clocks(adev);
> --
> 1.9.1
>
> _______________________________________________
> 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