Some fixes for things I spotted after trying to look at a bug Jerry Zuo
mentioned to me. Didn't manage to reproduce the bug! But I found these.
Cc: Jerry Zuo <[email protected]>
Lyude Paul (2):
drm/amd/dm: Don't forget to attach MST encoders
drm/amd/dm: Understand why attaching path/tile properties are needed
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
--
2.19.1
Drive-by fix, this is bound to cause problems somewhere.
Signed-off-by: Lyude Paul <[email protected]>
Cc: Jerry Zuo <[email protected]>
---
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index d02c32a1039c..0cca1809fdcd 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -342,6 +342,8 @@ dm_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr,
master->connector_id);
aconnector->mst_encoder = dm_dp_create_fake_mst_encoder(master);
+ drm_connector_attach_encoder(&aconnector->base,
+ &aconnector->mst_encoder->base);
/*
* TODO: understand why this one is needed
--
2.19.1
Path property is used for userspace to know what MST connector goes to
what actual DRM DisplayPort connector, the tiling property is for tiling
configurations. Not sure what else there is to figure out.
Signed-off-by: Lyude Paul <[email protected]>
---
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c | 3 ---
1 file changed, 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index 0cca1809fdcd..1b0d209d8367 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -345,9 +345,6 @@ dm_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr,
drm_connector_attach_encoder(&aconnector->base,
&aconnector->mst_encoder->base);
- /*
- * TODO: understand why this one is needed
- */
drm_object_attach_property(
&connector->base,
dev->mode_config.path_property,
--
2.19.1
Reviewed-by: Jerry (Fangzhi) Zuo <[email protected]>
The change fixed MST + SST daisy chain and S3 scenarios. The issue shows huge delay in MST + SST daisy chain, and soft hang in S3 resume.
The aux sequence is changed by failed iteration search in drm_connector_for_each_possible_encoder().
The failure of searching for the best encoder for the connector due to the miss of attached encoder in the process of adding MST connector.?The iteration search takes time to push drm_dp_send_enum_path_resources() aux transaction after the mode probe, and causes conflict to drm_dp_mst_i2c_xfer(), leading to the aux transaction timeout.
-----Original Message-----
From: Lyude Paul <[email protected]>
Sent: November 16, 2018 6:25 PM
To: [email protected]
Cc: Zuo, Jerry <[email protected]>; Wentland, Harry <[email protected]>; Li, Sun peng (Leo) <[email protected]>; Deucher, Alexander <[email protected]>; Koenig, Christian <[email protected]>; Zhou, David(ChunMing) <[email protected]>; David Airlie <[email protected]>; Li, Roman <[email protected]>; S, Shirish <[email protected]>; Daniel Vetter <[email protected]>; [email protected]; [email protected]
Subject: [PATCH 1/2] drm/amd/dm: Don't forget to attach MST encoders
Drive-by fix, this is bound to cause problems somewhere.
Signed-off-by: Lyude Paul <[email protected]>
Cc: Jerry Zuo <[email protected]>
---
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index d02c32a1039c..0cca1809fdcd 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -342,6 +342,8 @@ dm_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr,
master->connector_id);
aconnector->mst_encoder = dm_dp_create_fake_mst_encoder(master);
+ drm_connector_attach_encoder(&aconnector->base,
+ &aconnector->mst_encoder->base);
/*
* TODO: understand why this one is needed
--
2.19.1
Cool! If it did actually fix those problems, would you mind making sure this
gets Cc'd to stable when it gets pushed upstream?
On Mon, 2018-11-19 at 15:00 +0000, Zuo, Jerry wrote:
> Reviewed-by: Jerry (Fangzhi) Zuo <[email protected]>
>
> The change fixed MST + SST daisy chain and S3 scenarios. The issue shows
> huge delay in MST + SST daisy chain, and soft hang in S3 resume.
>
> The aux sequence is changed by failed iteration search in
> drm_connector_for_each_possible_encoder().
> The failure of searching for the best encoder for the connector due to the
> miss of attached encoder in the process of adding MST connector. The
> iteration search takes time to push drm_dp_send_enum_path_resources() aux
> transaction after the mode probe, and causes conflict to
> drm_dp_mst_i2c_xfer(), leading to the aux transaction timeout.
>
> -----Original Message-----
> From: Lyude Paul <[email protected]>
> Sent: November 16, 2018 6:25 PM
> To: [email protected]
> Cc: Zuo, Jerry <[email protected]>; Wentland, Harry <[email protected]>
> ; Li, Sun peng (Leo) <[email protected]>; Deucher, Alexander <
> [email protected]>; Koenig, Christian <[email protected]>;
> Zhou, David(ChunMing) <[email protected]>; David Airlie <[email protected]>
> ; Li, Roman <[email protected]>; S, Shirish <[email protected]>; Daniel
> Vetter <[email protected]>; [email protected];
> [email protected]
> Subject: [PATCH 1/2] drm/amd/dm: Don't forget to attach MST encoders
>
> Drive-by fix, this is bound to cause problems somewhere.
>
> Signed-off-by: Lyude Paul <[email protected]>
> Cc: Jerry Zuo <[email protected]>
> ---
> drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> index d02c32a1039c..0cca1809fdcd 100644
> --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> @@ -342,6 +342,8 @@ dm_dp_add_mst_connector(struct drm_dp_mst_topology_mgr
> *mgr,
> master->connector_id);
>
> aconnector->mst_encoder = dm_dp_create_fake_mst_encoder(master);
> + drm_connector_attach_encoder(&aconnector->base,
> + &aconnector->mst_encoder->base);
>
> /*
> * TODO: understand why this one is needed
--
Cheers,
Lyude Paul
Sure, I'll do it.
-----Original Message-----
From: Lyude Paul <[email protected]>
Sent: November 19, 2018 1:52 PM
To: Zuo, Jerry <[email protected]>; [email protected]
Cc: Wentland, Harry <[email protected]>; Li, Sun peng (Leo) <[email protected]>; Deucher, Alexander <[email protected]>; Koenig, Christian <[email protected]>; Zhou, David(ChunMing) <[email protected]>; David Airlie <[email protected]>; Li, Roman <[email protected]>; S, Shirish <[email protected]>; Daniel Vetter <[email protected]>; [email protected]; [email protected]
Subject: Re: [PATCH 1/2] drm/amd/dm: Don't forget to attach MST encoders
Cool! If it did actually fix those problems, would you mind making sure this gets Cc'd to stable when it gets pushed upstream?
On Mon, 2018-11-19 at 15:00 +0000, Zuo, Jerry wrote:
> Reviewed-by: Jerry (Fangzhi) Zuo <[email protected]>
>
> The change fixed MST + SST daisy chain and S3 scenarios. The issue
> shows huge delay in MST + SST daisy chain, and soft hang in S3 resume.
>
> The aux sequence is changed by failed iteration search in
> drm_connector_for_each_possible_encoder().
> The failure of searching for the best encoder for the connector due to
> the miss of attached encoder in the process of adding MST connector.
> The iteration search takes time to push
> drm_dp_send_enum_path_resources() aux transaction after the mode
> probe, and causes conflict to drm_dp_mst_i2c_xfer(), leading to the aux transaction timeout.
>
> -----Original Message-----
> From: Lyude Paul <[email protected]>
> Sent: November 16, 2018 6:25 PM
> To: [email protected]
> Cc: Zuo, Jerry <[email protected]>; Wentland, Harry
> <[email protected]> ; Li, Sun peng (Leo) <[email protected]>;
> Deucher, Alexander < [email protected]>; Koenig, Christian
> <[email protected]>; Zhou, David(ChunMing)
> <[email protected]>; David Airlie <[email protected]> ; Li, Roman
> <[email protected]>; S, Shirish <[email protected]>; Daniel Vetter
> <[email protected]>; [email protected];
> [email protected]
> Subject: [PATCH 1/2] drm/amd/dm: Don't forget to attach MST encoders
>
> Drive-by fix, this is bound to cause problems somewhere.
>
> Signed-off-by: Lyude Paul <[email protected]>
> Cc: Jerry Zuo <[email protected]>
> ---
> drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git
> a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> index d02c32a1039c..0cca1809fdcd 100644
> --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> @@ -342,6 +342,8 @@ dm_dp_add_mst_connector(struct
> drm_dp_mst_topology_mgr *mgr,
> master->connector_id);
>
> aconnector->mst_encoder = dm_dp_create_fake_mst_encoder(master);
> + drm_connector_attach_encoder(&aconnector->base,
> + &aconnector->mst_encoder->base);
>
> /*
> * TODO: understand why this one is needed
--
Cheers,
Lyude Paul
-----Original Message-----
From: Lyude Paul <[email protected]>
Sent: November 16, 2018 6:25 PM
To: [email protected]
Cc: Wentland, Harry <[email protected]>; Li, Sun peng (Leo) <[email protected]>; Deucher, Alexander <[email protected]>; Koenig, Christian <[email protected]>; Zhou, David(ChunMing) <[email protected]>; David Airlie <[email protected]>; Zuo, Jerry <[email protected]>; Li, Roman <[email protected]>; Cheng, Tony <[email protected]>; Daniel Vetter <[email protected]>; S, Shirish <[email protected]>; [email protected]; [email protected]
Subject: [PATCH 2/2] drm/amd/dm: Understand why attaching path/tile properties are needed
Path property is used for userspace to know what MST connector goes to what actual DRM DisplayPort connector, the tiling property is for tiling configurations. Not sure what else there is to figure out.
Signed-off-by: Lyude Paul <[email protected]>
Reviewed-by: Jerry (Fangzhi) Zuo <[email protected]>
---
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c | 3 ---
1 file changed, 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index 0cca1809fdcd..1b0d209d8367 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -345,9 +345,6 @@ dm_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr,
drm_connector_attach_encoder(&aconnector->base,
&aconnector->mst_encoder->base);
- /*
- * TODO: understand why this one is needed
- */
drm_object_attach_property(
&connector->base,
dev->mode_config.path_property,
--
2.19.1