Take SST-only branch device into account
Subsystems (2)
cover Take SST-only branch device into account
Noticed this while testing 4 ports MST hub from StarTech.com. While plugging in and display a MST monitor(Dell U2417H), change the MST feature to off from OSD. Monitor then can't display anymore.
After analyzing, found out that the CSN reports the specific port from Device with MST Branching Unit (Message_Capability_Status=1 && Peer_Device_Type=010) to SST-only Branch Device ((Message_Capability_Status=0 && Peer_Device_Type=010). Current code expects peer device as MST branch device and can handle sideband message once the peer device type is declared as Peer_Device_Type=010. But for SST-only Branch Device, its Peer_Device_Type is 010 and can't handle sideband message. For this case, we should follow the same handling procedure as the case in DP_PEER_DEVICE_DP_LEGACY_CONV & DP_PEER_DEVICE_SST_SINK.
This series is trying to add the SST-only branch device case in and remain the original processing logic in current code.
Wayne Lin (2):
drm/dp_mst: Add a function to determine the mst end device drm/dp_mst: Handle SST-only branch device case
drivers/gpu/drm/drm_dp_mst_topology.c | 147 +++++++++++++++----------- 1 file changed, 84 insertions(+), 63 deletions(-)
Signature · 6 lines
2.17.1 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 1/2 drm/dp_mst: Add a function to determine the mst end device
[Why]
For later usage convenience, add the function drm_dp_mst_is_dp_mst_end_device() to decide whether a peer device connected to a DFP is mst end device. Which also indicates if the peer device is capable of handling message or not.
--- drivers/gpu/drm/drm_dp_mst_topology.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)
drivers/gpu/drm/drm_dp_mst_topology.c +16 -0
@@ -1918,6 +1918,22 @@ static u8 drm_dp_calculate_rad(struct drm_dp_mst_port *port,return parent_lct +1 ; } +static bool drm_dp_mst_is_dp_mst_end_device (u8 pdt,bool mcs) +{ +switch (pdt) { +case DP_PEER_DEVICE_DP_LEGACY_CONV : +case DP_PEER_DEVICE_SST_SINK : +return true; +case DP_PEER_DEVICE_MST_BRANCHING : +/* For sst branch device */ +if (!mcs) +return true; + +return false; + } +return true; +} +static int drm_dp_port_set_pdt (struct drm_dp_mst_port *port,u8 new_pdt) {struct drm_dp_mst_topology_mgr *mgr = port->mgr;
Signature · 6 lines
2.17.1 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
Re: [PATCH 1/2] drm/dp_mst: Add a function to determine the mst end device
This patch series looks awesome so far, thank you for the great work! This patch looks great, I think we should just squash it into the next patch though since we don't use this function until then.
Quoted 40 lines · Wayne Lin
[Why]
For later usage convenience, add the functiondrm_dp_mst_is_dp_mst_end_device()to decide whether a peer device connected to a DFP is mst end device. Which also indicates if the peer device is capable of handling message or not.Signed-off-by Wayne Lin <[email protected]>--- drivers/gpu/drm/drm_dp_mst_topology.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)drivers/gpu/drm/drm_dp_mst_topology.c +16
-0@@ -1918,6 +1918,22 @@ static u8 drm_dp_calculate_rad(struct drm_dp_mst_port *port, return parent_lct + 1; } +static bool drm_dp_mst_is_dp_mst_end_device (u8 pdt,bool mcs) +{ +switch (pdt) { +case DP_PEER_DEVICE_DP_LEGACY_CONV : +case DP_PEER_DEVICE_SST_SINK : +return true; +case DP_PEER_DEVICE_MST_BRANCHING : +/* For sst branch device */ +if (!mcs) +return true; + +return false; + } +return true; +} +static int drm_dp_port_set_pdt (struct drm_dp_mst_port *port,u8 new_pdt) {struct
drm_dp_mst_topology_mgr*mgr =port->mgr;
Signature · 7 lines
Cheers, Lyude Paul _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
RE: [PATCH 1/2] drm/dp_mst: Add a function to determine the mst end device
[AMD Public Use]
Quoted 10 lines
-----Original Message-----
From: Lyude Paul <[email protected]>
Sent: Wednesday, January 15, 2020 5:16 AM
To: Lin, Wayne <[email protected]>; [email protected]; [email protected]
Cc: Kazlauskas, Nicholas <[email protected]>; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 1/2] drm/dp_mst: Add a function to determine the mst end device
Quoted 4 lines
This patch series looks awesome so far, thank you for the great work! This patch looks great, I think we should just squash it into the next patch though since we don't use this function until then.
Thanks for your time.
I will squash it in the new version.
Quoted 35 lines · Wayne Lin
Quoted 34 lines
[Why]
For later usage convenience, add the functiondrm_dp_mst_is_dp_mst_end_device()to decide whether a peer device connected to a DFP is mst end device. Which also indicates if the peer device is capable of handling message or not.Signed-off-by Wayne Lin <[email protected]>--- drivers/gpu/drm/drm_dp_mst_topology.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)drivers/gpu/drm/drm_dp_mst_topology.c +13
-0@@ -1918,6 +1918,22 @@ static u8 drm_dp_calculate_rad(structdrm_dp_mst_port *port, return parent_lct + 1; } +static bool drm_dp_mst_is_dp_mst_end_device (u8 pdt,bool mcs) { +switch (pdt) { +case DP_PEER_DEVICE_DP_LEGACY_CONV : +case DP_PEER_DEVICE_SST_SINK : +return true; +case DP_PEER_DEVICE_MST_BRANCHING : +/* For sst branch device */ +if (!mcs) +return true; + +return false; + } +return true;
Quoted 8 lines
drivers/gpu/drm/drm_dp_mst_topology.c static u8 drm_dp_calculate_rad(struct drm_dp_mst_port *port, · line 1939
Quoted 5 lines
drivers/gpu/drm/drm_dp_mst_topology.c +2
-0+} +static int drm_dp_port_set_pdt (struct drm_dp_mst_port *port,u8 new_pdt) {struct drm_dp_mst_topology_mgr *mgr = port->mgr;--
Cheers,
Lyude Paul
--
Best regards,
Wayne Lin
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 2/2 drm/dp_mst: Handle SST-only branch device case
In reply to Wayne Lin[Why]
While handling LINK_ADDRESS reply, current code expects a peer device can handle sideband message once the peer device type is reported as DP_PEER_DEVICE_MST_BRANCHING. However, when the connected device is a SST branch case, it can't handle the sideband message(MST_CAP=0 in DPCD 00021h).
Current code will try to send LINK_ADDRESS to SST branch device and end up with message timeout and monitor can't display normally. As the result of that, we should take SST branch device into account.
[How]
According to DP 1.4 spec, we can use Peer_Device_Type as DP_PEER_DEVICE_MST_BRANCHING and Message_Capability_Status as 0 to indicate peer device as a SST-only branch device.
Fix following:
- Take SST-only branch device case into account in drm_dp_port_set_pdt() and add a new parameter 'new_mcs'. Take sst branch device case as the same case as DP_PEER_DEVICE_DP_LEGACY_CONV and DP_PEER_DEVICE_SST_SINK. All original handling logics remain.
- Take SST-only branch device case into account in drm_dp_mst_port_add_connector().
- Fix some parts in drm_dp_mst_handle_link_address_port() to have SST branch device case into consideration.
- Fix the arguments of drm_dp_port_set_pdt() in drm_dp_mst_handle_conn_stat().
- Have SST branch device also report connector_status_connected when the ddps is true
in drm_dp_mst_detect_port()
- Fix the arguments of drm_dp_port_set_pdt() in drm_dp_delayed_destroy_port()
--- drivers/gpu/drm/drm_dp_mst_topology.c | 131 +++++++++++++------------- 1 file changed, 68 insertions(+), 63 deletions(-)
drivers/gpu/drm/drm_dp_mst_topology.c +68 -63
@@ -1934,73 +1934,74 @@ static bool drm_dp_mst_is_dp_mst_end_device(u8 pdt, bool mcs)return true; } -static int drm_dp_port_set_pdt (struct drm_dp_mst_port *port,u8 new_pdt) +static int +drm_dp_port_set_pdt (struct drm_dp_mst_port *port,u8 new_pdt, +bool new_mcs) {struct drm_dp_mst_topology_mgr *mgr = port->mgr;struct drm_dp_mst_branch *mstb;u8 rad[8 ], lct;int ret =0 ; -if (port->pdt == new_pdt) +if (port->pdt == new_pdt && port->mcs == new_mcs)return 0 ;/* Teardown the old pdt, if there is one */ -switch (port->pdt) { -case DP_PEER_DEVICE_DP_LEGACY_CONV : -case DP_PEER_DEVICE_SST_SINK : -/* -* If the new PDT would also have an i2c bus, don't bother -* with reregistering it -*/ -if (new_pdt ==DP_PEER_DEVICE_DP_LEGACY_CONV || - new_pdt ==DP_PEER_DEVICE_SST_SINK ) { - port->pdt = new_pdt; -return 0 ; - } +if (port->pdt !=DP_PEER_DEVICE_NONE ) { +if (drm_dp_mst_is_dp_mst_end_device (port->pdt, port->mcs)) { +/* +* If the new PDT would also have an i2c bus, +* don't bother with reregistering it +*/ +if (new_pdt !=DP_PEER_DEVICE_NONE && +drm_dp_mst_is_dp_mst_end_device (new_pdt, new_mcs)) { + port->pdt = new_pdt; + port->mcs = new_mcs; +return 0 ; + } -/* remove i2c over sideband */ -drm_dp_mst_unregister_i2c_bus (&port->aux); -break ; -case DP_PEER_DEVICE_MST_BRANCHING : -mutex_lock (&mgr->lock); -drm_dp_mst_topology_put_mstb (port->mstb); - port->mstb =NULL ; -mutex_unlock (&mgr->lock); -break ; +/* remove i2c over sideband */ +drm_dp_mst_unregister_i2c_bus (&port->aux); + }else { +mutex_lock (&mgr->lock); +drm_dp_mst_topology_put_mstb (port->mstb); + port->mstb =NULL ; +mutex_unlock (&mgr->lock); + } } port->pdt = new_pdt; -switch (port->pdt) { -case DP_PEER_DEVICE_DP_LEGACY_CONV : -case DP_PEER_DEVICE_SST_SINK : -/* add i2c over sideband */ - ret =drm_dp_mst_register_i2c_bus (&port->aux); -break ; + port->mcs = new_mcs; -case DP_PEER_DEVICE_MST_BRANCHING : - lct =drm_dp_calculate_rad (port, rad); - mstb =drm_dp_add_mst_branch_device (lct, rad); -if (!mstb) { - ret = -ENOMEM ; -DRM_ERROR ("Failed to create MSTB for port %p" , port); - goto out; - } +if (port->pdt !=DP_PEER_DEVICE_NONE ) { +if (drm_dp_mst_is_dp_mst_end_device (port->pdt, port->mcs)) { +/* add i2c over sideband */ + ret =drm_dp_mst_register_i2c_bus (&port->aux); + }else { + lct =drm_dp_calculate_rad (port, rad); + mstb =drm_dp_add_mst_branch_device (lct, rad); +if (!mstb) { + ret = -ENOMEM ; +DRM_ERROR ("Failed to create MSTB for port %p" , + port); + goto out; + } -mutex_lock (&mgr->lock); - port->mstb = mstb; - mstb->mgr = port->mgr; - mstb->port_parent = port; +mutex_lock (&mgr->lock); + port->mstb = mstb; + mstb->mgr = port->mgr; + mstb->port_parent = port; -/* -* Make sure this port's memory allocation stays -* around until its child MSTB releases it -*/ -drm_dp_mst_get_port_malloc (port); -mutex_unlock (&mgr->lock); +/* +* Make sure this port's memory allocation stays +* around until its child MSTB releases it +*/ +drm_dp_mst_get_port_malloc (port); +mutex_unlock (&mgr->lock); -/* And make sure we send a link address for this */ - ret =1 ; -break ; +/* And make sure we send a link address for this */ + ret =1 ; + } } out:
@@ -2153,12 +2154,12 @@ drm_dp_mst_port_add_connector(struct drm_dp_mst_branch *mstb, goto error; } -if ((port->pdt ==DP_PEER_DEVICE_DP_LEGACY_CONV || - port->pdt ==DP_PEER_DEVICE_SST_SINK ) && - port->port_num >=DP_MST_LOGICAL_PORT_0 ) { - port->cached_edid =drm_get_edid (port->connector, - &port->aux.ddc); -drm_connector_set_tile_property (port->connector); +if (port->pdt !=DP_PEER_DEVICE_NONE ) { +if (drm_dp_mst_is_dp_mst_end_device (port->pdt, port->mcs)) { + port->cached_edid =drm_get_edid (port->connector, + &port->aux.ddc); +drm_connector_set_tile_property (port->connector); + } } mgr->cbs->register_connector (port->connector);
@@ -2223,6 +2224,7 @@ drm_dp_mst_handle_link_address_port(struct drm_dp_mst_branch *mstb,struct drm_dp_mst_port *port;int old_ddps =0 , ret;u8 new_pdt =DP_PEER_DEVICE_NONE ; +bool new_mcs =0 ;bool created = false, send_link_addr = false, changed = false; port =drm_dp_get_port (mstb, port_msg->port_number);
@@ -2267,7 +2269,7 @@ drm_dp_mst_handle_link_address_port(struct drm_dp_mst_branch *mstb, port->input = port_msg->input_port;if (!port->input) new_pdt = port_msg->peer_device_type; - port->mcs = port_msg->mcs; + new_mcs = port_msg->mcs; port->ddps = port_msg->ddps; port->ldps = port_msg->legacy_device_plug_status; port->dpcd_rev = port_msg->dpcd_revision;
@@ -2295,7 +2297,7 @@ drm_dp_mst_handle_link_address_port(struct drm_dp_mst_branch *mstb, } } - ret =drm_dp_port_set_pdt (port, new_pdt); + ret =drm_dp_port_set_pdt (port, new_pdt, new_mcs);if (ret ==1 ) { send_link_addr = true; }else if (ret <0 ) {
@@ -2309,7 +2311,8 @@ drm_dp_mst_handle_link_address_port(struct drm_dp_mst_branch *mstb,* we're coming out of suspend. In this case, always resend the link * address if there's an MSTB on this port */ -if (!created && port->pdt ==DP_PEER_DEVICE_MST_BRANCHING ) +if (!created && port->pdt ==DP_PEER_DEVICE_MST_BRANCHING && + port->mcs) send_link_addr = true;if (port->connector)
@@ -2346,6 +2349,7 @@ drm_dp_mst_handle_conn_stat(struct drm_dp_mst_branch *mstb,struct drm_dp_mst_port *port;int old_ddps, old_input, ret, i;u8 new_pdt; +bool new_mcs;bool dowork = false, create_connector = false; port =drm_dp_get_port (mstb, conn_stat->port_number);
@@ -2377,7 +2381,6 @@ drm_dp_mst_handle_conn_stat(struct drm_dp_mst_branch *mstb, old_ddps = port->ddps; old_input = port->input; port->input = conn_stat->input_port; - port->mcs = conn_stat->message_capability_status; port->ldps = conn_stat->legacy_device_plug_status; port->ddps = conn_stat->displayport_device_plug_status;
@@ -2390,8 +2393,8 @@ drm_dp_mst_handle_conn_stat(struct drm_dp_mst_branch *mstb, } new_pdt = port->input ?DP_PEER_DEVICE_NONE : conn_stat->peer_device_type; - - ret =drm_dp_port_set_pdt (port, new_pdt); + new_mcs = conn_stat->message_capability_status; + ret =drm_dp_port_set_pdt (port, new_pdt, new_mcs);if (ret ==1 ) { dowork = true; }else if (ret <0 ) {
@@ -3958,6 +3961,8 @@ drm_dp_mst_detect_port(struct drm_connector *connector,switch (port->pdt) {case DP_PEER_DEVICE_NONE :case DP_PEER_DEVICE_MST_BRANCHING : +if (!port->mcs) + ret = connector_status_connected;break ;case DP_PEER_DEVICE_SST_SINK :
@@ -4597,7 +4602,7 @@ drm_dp_delayed_destroy_port(struct drm_dp_mst_port *port)if (port->connector) port->mgr->cbs->destroy_connector (port->mgr, port->connector); -drm_dp_port_set_pdt (port,DP_PEER_DEVICE_NONE ); +drm_dp_port_set_pdt (port,DP_PEER_DEVICE_NONE , port->mcs);drm_dp_mst_put_port_malloc (port); }
Signature · 6 lines
2.17.1 _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
Re: [PATCH 2/2] drm/dp_mst: Handle SST-only branch device case
Quoted 182 lines · Wayne Lin
[Why]
While handlingLINK_ADDRESSreply, current code expects a peer device can handle sideband message once the peer device type is reported asDP_PEER_DEVICE_MST_BRANCHING. However, when the connected device is a SST branch case, it can't handle the sidebandmessage(MST_CAP=0 in DPCD 00021h).Current code will try to send
LINK_ADDRESSto SST branch device and end up with message timeout and monitor can't display normally. As the result of that, we should take SST branch device into account.[How]
According to DP 1.4 spec, we can usePeer_Device_TypeasDP_PEER_DEVICE_MST_BRANCHINGandMessage_Capability_Statusas 0 to indicate peer device as a SST-only branch device.Fix following:
- Take SST-only branch device case into account indrm_dp_port_set_pdt()and add a new parameter 'new_mcs'. Take sst branch device case as the same case asDP_PEER_DEVICE_DP_LEGACY_CONVandDP_PEER_DEVICE_SST_SINK. All original handling logics remain.
- Take SST-only branch device case into account indrm_dp_mst_port_add_connector().
- Fix some parts indrm_dp_mst_handle_link_address_port()to have SST branch device case into consideration.
- Fix the arguments ofdrm_dp_port_set_pdt()indrm_dp_mst_handle_conn_stat().
- Have SST branch device also reportconnector_status_connectedwhen the ddps is true
indrm_dp_mst_detect_port()
- Fix the arguments ofdrm_dp_port_set_pdt()indrm_dp_delayed_destroy_port()Fixes c485e2c97dae ("drm/dp_mst: Refactor pdt setup/teardown, add morelocking")
Cc Ville Syrjälä <[email protected]>Cc Harry Wentland <[email protected]>Cc Lyude Paul <[email protected]>Signed-off-by Wayne Lin <[email protected]>--- drivers/gpu/drm/drm_dp_mst_topology.c | 131 +++++++++++++------------- 1 file changed, 68 insertions(+), 63 deletions(-)drivers/gpu/drm/drm_dp_mst_topology.c +51
-50@@ -1934,73 +1934,74 @@ static bool drm_dp_mst_is_dp_mst_end_device(u8 pdt, bool mcs)return true; } -static int drm_dp_port_set_pdt (struct drm_dp_mst_port *port,u8 new_pdt) +static int +drm_dp_port_set_pdt (struct drm_dp_mst_port *port,u8 new_pdt, +bool new_mcs) {struct drm_dp_mst_topology_mgr *mgr = port->mgr;struct drm_dp_mst_branch *mstb;u8 rad[8 ], lct;int ret =0 ; -if (port->pdt == new_pdt) +if (port->pdt == new_pdt && port->mcs == new_mcs)return 0 ;/* Teardown the old pdt, if there is one */ -switch (port->pdt) { -case DP_PEER_DEVICE_DP_LEGACY_CONV : -case DP_PEER_DEVICE_SST_SINK : -/* -* If the new PDT would also have an i2c bus, don't bother -* with reregistering it -*/ -if (new_pdt ==DP_PEER_DEVICE_DP_LEGACY_CONV || - new_pdt ==DP_PEER_DEVICE_SST_SINK ) { - port->pdt = new_pdt; -return 0 ; - } +if (port->pdt !=DP_PEER_DEVICE_NONE ) { +if (drm_dp_mst_is_dp_mst_end_device (port->pdt, port->mcs)) { +/* +* If the new PDT would also have an i2c bus, +* don't bother with reregistering it +*/ +if (new_pdt !=DP_PEER_DEVICE_NONE && +drm_dp_mst_is_dp_mst_end_device (new_pdt, new_mcs)) { + port->pdt = new_pdt; + port->mcs = new_mcs; +return 0 ; + } -/* remove i2c over sideband */ -drm_dp_mst_unregister_i2c_bus (&port->aux); -break ; -case DP_PEER_DEVICE_MST_BRANCHING : -mutex_lock (&mgr->lock); -drm_dp_mst_topology_put_mstb (port->mstb); - port->mstb =NULL ; -mutex_unlock (&mgr->lock); -break ; +/* remove i2c over sideband */ +drm_dp_mst_unregister_i2c_bus (&port->aux); + }else { +mutex_lock (&mgr->lock); +drm_dp_mst_topology_put_mstb (port->mstb); + port->mstb =NULL ; +mutex_unlock (&mgr->lock); + } } port->pdt = new_pdt; -switch (port->pdt) { -case DP_PEER_DEVICE_DP_LEGACY_CONV : -case DP_PEER_DEVICE_SST_SINK : -/* add i2c over sideband */ - ret =drm_dp_mst_register_i2c_bus (&port->aux); -break ; + port->mcs = new_mcs; -case DP_PEER_DEVICE_MST_BRANCHING : - lct =drm_dp_calculate_rad (port, rad); - mstb =drm_dp_add_mst_branch_device (lct, rad); -if (!mstb) { - ret = -ENOMEM ; -DRM_ERROR ("Failed to create MSTB for port %p" , port); - goto out; - } +if (port->pdt !=DP_PEER_DEVICE_NONE ) { +if (drm_dp_mst_is_dp_mst_end_device (port->pdt, port->mcs)) { +/* add i2c over sideband */ + ret =drm_dp_mst_register_i2c_bus (&port->aux); + }else { + lct =drm_dp_calculate_rad (port, rad); + mstb =drm_dp_add_mst_branch_device (lct, rad); +if (!mstb) { + ret = -ENOMEM ; +DRM_ERROR ("Failed to create MSTB for port %p" , + port); + goto out; + } -mutex_lock (&mgr->lock); - port->mstb = mstb; - mstb->mgr = port->mgr; - mstb->port_parent = port; +mutex_lock (&mgr->lock); + port->mstb = mstb; + mstb->mgr = port->mgr; + mstb->port_parent = port; -/* -* Make sure this port's memory allocation stays -* around until its child MSTB releases it -*/ -drm_dp_mst_get_port_malloc (port); -mutex_unlock (&mgr->lock); +/* +* Make sure this port's memory allocation stays +* around until its child MSTB releases it +*/ +drm_dp_mst_get_port_malloc (port); +mutex_unlock (&mgr->lock); -/* And make sure we send a link address for this */ - ret =1 ; -break ; +/* And make sure we send a link address for this */ + ret =1 ; + } }out:
drivers/gpu/drm/drm_dp_mst_topology.c +0
-2@@ -2153,12 +2154,12 @@ drm_dp_mst_port_add_connector(structdrm_dp_mst_branch *mstb, goto error; } - if ((port->pdt == DP_PEER_DEVICE_DP_LEGACY_CONV || - port->pdt == DP_PEER_DEVICE_SST_SINK) &&
drivers/gpu/drm/drm_dp_mst_topology.c drm_dp_mst_port_add_connector(struct · line 2163
Quoted 10 lines
drivers/gpu/drm/drm_dp_mst_topology.c +6
-4- port->port_num >= DP_MST_LOGICAL_PORT_0) { - port->cached_edid =drm_get_edid (port->connector, - &port->aux.ddc); -drm_connector_set_tile_property (port->connector); +if (port->pdt !=DP_PEER_DEVICE_NONE ) { +if (drm_dp_mst_is_dp_mst_end_device (port->pdt, port->mcs)) { + port->cached_edid =drm_get_edid (port->connector, + &port->aux.ddc); +drm_connector_set_tile_property (port->connector); + }
I'd combine these two if statements here into one, otherwise this looks great. Thank you for all of the great fixes recently :)
Quoted 94 lines
drivers/gpu/drm/drm_dp_mst_topology.c +0
-0}drivers/gpu/drm/drm_dp_mst_topology.c +1
-0mgr->cbs->register_connector (port->connector);@@ -2223,6 +2224,7 @@ drm_dp_mst_handle_link_address_port(structdrm_dp_mst_branch *mstb, struct drm_dp_mst_port *port;int old_ddps =0 , ret;u8 new_pdt =DP_PEER_DEVICE_NONE ; +bool new_mcs =0 ;bool created = false, send_link_addr = false, changed = false;drivers/gpu/drm/drm_dp_mst_topology.c +1
-1port =drm_dp_get_port (mstb, port_msg->port_number);@@ -2267,7 +2269,7 @@ drm_dp_mst_handle_link_address_port(struct drm_dp_mst_branch *mstb, port->input = port_msg->input_port;if (!port->input) new_pdt = port_msg->peer_device_type; - port->mcs = port_msg->mcs; + new_mcs = port_msg->mcs; port->ddps = port_msg->ddps; port->ldps = port_msg->legacy_device_plug_status;drivers/gpu/drm/drm_dp_mst_topology.c +1
-1port->dpcd_rev = port_msg->dpcd_revision;@@ -2295,7 +2297,7 @@ drm_dp_mst_handle_link_address_port(structdrm_dp_mst_branch *mstb, } } - ret =drm_dp_port_set_pdt (port, new_pdt); + ret =drm_dp_port_set_pdt (port, new_pdt, new_mcs);if (ret ==1 ) { send_link_addr = true;drivers/gpu/drm/drm_dp_mst_topology.c +2
-1}else if (ret <0 ) {@@ -2309,7 +2311,8 @@ drm_dp_mst_handle_link_address_port(structdrm_dp_mst_branch *mstb, * we're coming out of suspend. In this case, always resend the link * address if there's an MSTB on this port */ -if (!created && port->pdt ==DP_PEER_DEVICE_MST_BRANCHING ) +if (!created && port->pdt ==DP_PEER_DEVICE_MST_BRANCHING && + port->mcs) send_link_addr = true;drivers/gpu/drm/drm_dp_mst_topology.c +1
-0if (port->connector)@@ -2346,6 +2349,7 @@ drm_dp_mst_handle_conn_stat(struct drm_dp_mst_branch *mstb, struct drm_dp_mst_port *port;int old_ddps, old_input, ret, i;u8 new_pdt; +bool new_mcs;bool dowork = false, create_connector = false;drivers/gpu/drm/drm_dp_mst_topology.c +2
-3port =drm_dp_get_port (mstb, conn_stat->port_number);@@ -2377,7 +2381,6 @@ drm_dp_mst_handle_conn_stat(struct drm_dp_mst_branch *mstb, old_ddps = port->ddps; old_input = port->input; port->input = conn_stat->input_port; - port->mcs = conn_stat->message_capability_status; port->ldps = conn_stat->legacy_device_plug_status; port->ddps = conn_stat->displayport_device_plug_status;@@ -2390,8 +2393,8 @@ drm_dp_mst_handle_conn_stat(struct drm_dp_mst_branch *mstb, } new_pdt = port->input ?DP_PEER_DEVICE_NONE : conn_stat- >peer_device_type; - - ret =drm_dp_port_set_pdt (port, new_pdt); + new_mcs = conn_stat->message_capability_status; + ret =drm_dp_port_set_pdt (port, new_pdt, new_mcs);if (ret ==1 ) {drivers/gpu/drm/drm_dp_mst_topology.c +2
-0dowork = true; }else if (ret <0 ) {@@ -3958,6 +3961,8 @@ drm_dp_mst_detect_port(struct drm_connector *connector, switch (port->pdt) {case DP_PEER_DEVICE_NONE :case DP_PEER_DEVICE_MST_BRANCHING : +if (!port->mcs) + ret = connector_status_connected;break ;drivers/gpu/drm/drm_dp_mst_topology.c +1
-1case DP_PEER_DEVICE_SST_SINK :@@ -4597,7 +4602,7 @@ drm_dp_delayed_destroy_port(struct drm_dp_mst_port *port)if (port->connector) port->mgr->cbs->destroy_connector (port->mgr, port->connector); -drm_dp_port_set_pdt (port,DP_PEER_DEVICE_NONE ); +drm_dp_port_set_pdt (port,DP_PEER_DEVICE_NONE , port->mcs);drm_dp_mst_put_port_malloc (port); }
Signature · 7 lines
Cheers, Lyude Paul _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
RE: [PATCH 2/2] drm/dp_mst: Handle SST-only branch device case
[AMD Public Use]
Quoted 213 lines
-----Original Message-----
From: Lyude Paul <[email protected]>
Sent: Wednesday, January 15, 2020 5:19 AM
To: Lin, Wayne <[email protected]>; [email protected]; [email protected]
Cc: Kazlauskas, Nicholas <[email protected]>; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>; Ville Syrjälä <[email protected]>; Wentland, Harry <[email protected]>
Subject: Re: [PATCH 2/2] drm/dp_mst: Handle SST-only branch device caseQuoted 4 lines · Wayne Lin
[Why]
While handlingLINK_ADDRESSreply, current code expects a peer device can handle sideband message once the peer device type is reported asDP_PEER_DEVICE_MST_BRANCHING. However, when the connected deviceis a
Quoted 12 lines
SST branch case, it can't handle the sideband
message(MST_CAP=0 in DPCD 00021h).Current code will try to send
LINK_ADDRESSto SST branch device and end up with message timeout and monitor can't display normally. As the result of that, we should take SST branch device into account.[How]
According to DP 1.4 spec, we can usePeer_Device_TypeasDP_PEER_DEVICE_MST_BRANCHINGandMessage_Capability_Statusas 0 to indicate peer device as a SST-only branch device.Quoted 4 lines
Fix following:
- Take SST-only branch device case into account indrm_dp_port_set_pdt()and add a new parameter 'new_mcs'. Take sst branch device case as the same case as
DP_PEER_DEVICE_DP_LEGACY_CONVdrivers/gpu/drm/drm_dp_mst_topology.c static bool
Quoted 31 lines
and
DP_PEER_DEVICE_SST_SINK. All original handling logics remain.
- Take SST-only branch device case into account indrm_dp_mst_port_add_connector().
- Fix some parts indrm_dp_mst_handle_link_address_port()to have SST branch device case into consideration.
- Fix the arguments ofdrm_dp_port_set_pdt()indrm_dp_mst_handle_conn_stat().
- Have SST branch device also reportconnector_status_connectedwhen the ddps is true indrm_dp_mst_detect_port()
- Fix the arguments ofdrm_dp_port_set_pdt()indrm_dp_delayed_destroy_port()Fixes c485e2c97dae ("drm/dp_mst: Refactor pdt setup/teardown, addmore
locking")Cc Ville Syrjälä <[email protected]>Cc Harry Wentland <[email protected]>Cc Lyude Paul <[email protected]>Signed-off-by Wayne Lin <[email protected]>--- drivers/gpu/drm/drm_dp_mst_topology.c | 131 +++++++++++++------------- 1 file changed, 68 insertions(+), 63 deletions(-)drivers/gpu/drm/drm_dp_mst_topology.c +0
-0@@ -1934,73 +1934,74 @@ static booldrm_dp_mst_is_dp_mst_end_device(u8
Quoted 134 lines
drivers/gpu/drm/drm_dp_mst_topology.c +51
-50pdt, bool mcs)return true; } -static int drm_dp_port_set_pdt (struct drm_dp_mst_port *port,u8 new_pdt) +static int +drm_dp_port_set_pdt (struct drm_dp_mst_port *port,u8 new_pdt, +bool new_mcs) {struct drm_dp_mst_topology_mgr *mgr = port->mgr;struct drm_dp_mst_branch *mstb;u8 rad[8 ], lct;int ret =0 ; -if (port->pdt == new_pdt) +if (port->pdt == new_pdt && port->mcs == new_mcs)return 0 ;/* Teardown the old pdt, if there is one */ -switch (port->pdt) { -case DP_PEER_DEVICE_DP_LEGACY_CONV : -case DP_PEER_DEVICE_SST_SINK : -/* -* If the new PDT would also have an i2c bus, don't bother -* with reregistering it -*/ -if (new_pdt ==DP_PEER_DEVICE_DP_LEGACY_CONV || - new_pdt ==DP_PEER_DEVICE_SST_SINK ) { - port->pdt = new_pdt; -return 0 ; - } +if (port->pdt !=DP_PEER_DEVICE_NONE ) { +if (drm_dp_mst_is_dp_mst_end_device (port->pdt, port->mcs)) { +/* +* If the new PDT would also have an i2c bus, +* don't bother with reregistering it +*/ +if (new_pdt !=DP_PEER_DEVICE_NONE && +drm_dp_mst_is_dp_mst_end_device (new_pdt, new_mcs)) { + port->pdt = new_pdt; + port->mcs = new_mcs; +return 0 ; + } -/* remove i2c over sideband */ -drm_dp_mst_unregister_i2c_bus (&port->aux); -break ; -case DP_PEER_DEVICE_MST_BRANCHING : -mutex_lock (&mgr->lock); -drm_dp_mst_topology_put_mstb (port->mstb); - port->mstb =NULL ; -mutex_unlock (&mgr->lock); -break ; +/* remove i2c over sideband */ +drm_dp_mst_unregister_i2c_bus (&port->aux); + }else { +mutex_lock (&mgr->lock); +drm_dp_mst_topology_put_mstb (port->mstb); + port->mstb =NULL ; +mutex_unlock (&mgr->lock); + } } port->pdt = new_pdt; -switch (port->pdt) { -case DP_PEER_DEVICE_DP_LEGACY_CONV : -case DP_PEER_DEVICE_SST_SINK : -/* add i2c over sideband */ - ret =drm_dp_mst_register_i2c_bus (&port->aux); -break ; + port->mcs = new_mcs; -case DP_PEER_DEVICE_MST_BRANCHING : - lct =drm_dp_calculate_rad (port, rad); - mstb =drm_dp_add_mst_branch_device (lct, rad); -if (!mstb) { - ret = -ENOMEM ; -DRM_ERROR ("Failed to create MSTB for port %p" , port); - goto out; - } +if (port->pdt !=DP_PEER_DEVICE_NONE ) { +if (drm_dp_mst_is_dp_mst_end_device (port->pdt, port->mcs)) { +/* add i2c over sideband */ + ret =drm_dp_mst_register_i2c_bus (&port->aux); + }else { + lct =drm_dp_calculate_rad (port, rad); + mstb =drm_dp_add_mst_branch_device (lct, rad); +if (!mstb) { + ret = -ENOMEM ; +DRM_ERROR ("Failed to create MSTB for port %p" , + port); + goto out; + } -mutex_lock (&mgr->lock); - port->mstb = mstb; - mstb->mgr = port->mgr; - mstb->port_parent = port; +mutex_lock (&mgr->lock); + port->mstb = mstb; + mstb->mgr = port->mgr; + mstb->port_parent = port; -/* -* Make sure this port's memory allocation stays -* around until its child MSTB releases it -*/ -drm_dp_mst_get_port_malloc (port); -mutex_unlock (&mgr->lock); +/* +* Make sure this port's memory allocation stays +* around until its child MSTB releases it +*/ +drm_dp_mst_get_port_malloc (port); +mutex_unlock (&mgr->lock); -/* And make sure we send a link address for this */ - ret =1 ; -break ; +/* And make sure we send a link address for this */ + ret =1 ; + }drivers/gpu/drm/drm_dp_mst_topology.c +0
-2} out:@@ -2153,12 +2154,12 @@ drm_dp_mst_port_add_connector(structdrm_dp_mst_branch *mstb, goto error; } - if ((port->pdt == DP_PEER_DEVICE_DP_LEGACY_CONV || - port->pdt == DP_PEER_DEVICE_SST_SINK) &&drivers/gpu/drm/drm_dp_mst_topology.c drm_dp_mst_port_add_connector(struct · line 2163
Quoted 10 lines
drivers/gpu/drm/drm_dp_mst_topology.c +6
-4- port->port_num >= DP_MST_LOGICAL_PORT_0) { - port->cached_edid =drm_get_edid (port->connector, - &port->aux.ddc); -drm_connector_set_tile_property (port->connector); +if (port->pdt !=DP_PEER_DEVICE_NONE ) { +if (drm_dp_mst_is_dp_mst_end_device (port->pdt, port->mcs)) { + port->cached_edid =drm_get_edid (port->connector, + &port->aux.ddc); +drm_connector_set_tile_property (port->connector); + }I'd combine these two if statements here into one, otherwise this looks great. Thank you for all of the great fixes recently :)
Quoted 1 line
Reviewed-by Lyude Paul <[email protected]>
Thanks for your time!
I will fix it in the new version.
Thanks.
Quoted 92 lines
Quoted 53 lines
drivers/gpu/drm/drm_dp_mst_topology.c +0
-0}drivers/gpu/drm/drm_dp_mst_topology.c +1
-0mgr->cbs->register_connector (port->connector);@@ -2223,6 +2224,7 @@ drm_dp_mst_handle_link_address_port(structdrm_dp_mst_branch *mstb, struct drm_dp_mst_port *port;int old_ddps =0 , ret;u8 new_pdt =DP_PEER_DEVICE_NONE ; +bool new_mcs =0 ;bool created = false, send_link_addr = false, changed = false;port =
drm_dp_get_port(mstb, port_msg->port_number); @@ -2267,7
+2269,7 @@ drm_dp_mst_handle_link_address_port(structdrm_dp_mst_branch*mstb,drivers/gpu/drm/drm_dp_mst_topology.c +1
-1port->input = port_msg->input_port;if (!port->input) new_pdt = port_msg->peer_device_type; - port->mcs = port_msg->mcs; + new_mcs = port_msg->mcs; port->ddps = port_msg->ddps; port->ldps = port_msg->legacy_device_plug_status; port->dpcd_rev = port_msg->dpcd_revision; @@ -2295,7 +2297,7 @@drm_dp_mst_handle_link_address_port(struct
drm_dp_mst_branch*mstb,drivers/gpu/drm/drm_dp_mst_topology.c +3
-2} } - ret =drm_dp_port_set_pdt (port, new_pdt); + ret =drm_dp_port_set_pdt (port, new_pdt, new_mcs);if (ret ==1 ) { send_link_addr = true; }else if (ret <0 ) {@@ -2309,7 +2311,8 @@ drm_dp_mst_handle_link_address_port(structdrm_dp_mst_branch *mstb, * we're coming out of suspend. In this case, always resend the link * address if there's an MSTB on this port */ -if (!created && port->pdt ==DP_PEER_DEVICE_MST_BRANCHING ) +if (!created && port->pdt ==DP_PEER_DEVICE_MST_BRANCHING && + port->mcs) send_link_addr = true;drivers/gpu/drm/drm_dp_mst_topology.c +1
-0if (port->connector)@@ -2346,6 +2349,7 @@ drm_dp_mst_handle_conn_stat(structdrm_dp_mst_branch *mstb, struct drm_dp_mst_port *port;int old_ddps, old_input, ret, i;u8 new_pdt; +bool new_mcs;bool dowork = false, create_connector = false;Quoted 2 lines
port =
drm_dp_get_port(mstb, conn_stat->port_number); @@ -2377,7
+2381,6 @@ drm_dp_mst_handle_conn_stat(structdrm_dp_mst_branch*mstb,
Quoted 36 lines
drivers/gpu/drm/drm_dp_mst_topology.c +2
-3old_ddps = port->ddps; old_input = port->input; port->input = conn_stat->input_port; - port->mcs = conn_stat->message_capability_status; port->ldps = conn_stat->legacy_device_plug_status; port->ddps = conn_stat->displayport_device_plug_status;@@ -2390,8 +2393,8 @@ drm_dp_mst_handle_conn_stat(structdrm_dp_mst_branch *mstb, } new_pdt = port->input ?DP_PEER_DEVICE_NONE : conn_stat- >peer_device_type; - - ret =drm_dp_port_set_pdt (port, new_pdt); + new_mcs = conn_stat->message_capability_status; + ret =drm_dp_port_set_pdt (port, new_pdt, new_mcs);if (ret ==1 ) {drivers/gpu/drm/drm_dp_mst_topology.c +2
-0dowork = true; }else if (ret <0 ) {@@ -3958,6 +3961,8 @@ drm_dp_mst_detect_port(struct drm_connector *connector, switch (port->pdt) {case DP_PEER_DEVICE_NONE :case DP_PEER_DEVICE_MST_BRANCHING : +if (!port->mcs) + ret = connector_status_connected;break ;drivers/gpu/drm/drm_dp_mst_topology.c +0
-0case DP_PEER_DEVICE_SST_SINK :@@ -4597,7 +4602,7 @@ drm_dp_delayed_destroy_port(struct
drm_dp_mst_port
*port)drivers/gpu/drm/drm_dp_mst_topology.c +0
-0if (port->connector) port->mgr->cbs->destroy_connector (port->mgr, port->connector);
Quoted 8 lines
drivers/gpu/drm/drm_dp_mst_topology.c drm_dp_delayed_destroy_port(struct drm_dp_mst_port *port) · line 4607
Quoted 5 lines
drivers/gpu/drm/drm_dp_mst_topology.c +1
-1-drm_dp_port_set_pdt (port,DP_PEER_DEVICE_NONE ); +drm_dp_port_set_pdt (port,DP_PEER_DEVICE_NONE , port->mcs);drm_dp_mst_put_port_malloc (port); }--
Cheers,
Lyude Paul
--
Best regards,
Wayne Lin
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx