Refactor and clean up codes of mst
Subsystems (7)
cover Refactor and clean up codes of mst
This patch set is mainly trying to organize the mst code today a bit. Like to clarify and organize the sequence of mst payload allocation and removement.And also clean up some redundant codes today.
The main refactor one is the patch
"drm/mst: Refactor the flow for payload allocation/removement" which is adding a new enum variable in stuct drm_dp_mst_atomic_payload to represent the status of paylad alloction, and then handle the payload accordingly. Besides, rename some drm mst fnctions to better express the behind idea.
The other two patches are mainly to clean up unnecessary codes.
Wayne Lin (3):
drm/mst: delete unnecessary case in drm_dp_add_payload_part2() drm/mst: Refactor the flow for payload allocation/removement drm/mst: adjust the function drm_dp_remove_payload_part2()
.../amd/display/amdgpu_dm/amdgpu_dm_helpers.c | 60 +----- drivers/gpu/drm/display/drm_dp_mst_topology.c | 189 +++++++++++------- drivers/gpu/drm/i915/display/intel_dp_mst.c | 13 +- drivers/gpu/drm/nouveau/dispnv50/disp.c | 11 +- include/drm/display/drm_dp_mst_helper.h | 22 +- 5 files changed, 158 insertions(+), 137 deletions(-)
Signature · 1 line
2.37.3
PATCH 1/3 drm/mst: delete unnecessary case in drm_dp_add_payload_part2()
[Why]
There is no need to consider payload->delete case since we won't call drm_dp_add_payload_part2() to create a payload when we're about to remove it.
[How]
Delete unnecessary case to simplify the code.
--- drivers/gpu/drm/display/drm_dp_mst_topology.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-)
drivers/gpu/drm/display/drm_dp_mst_topology.c +2 -6
@@ -3411,12 +3411,8 @@ int drm_dp_add_payload_part2(struct drm_dp_mst_topology_mgr *mgr, ret =drm_dp_create_payload_step2 (mgr, payload);if (ret <0 ) { -if (!payload->delete) -drm_err (mgr->dev,"Step 2 of creating MST payload for %p failed: %d\n" , - payload->port, ret); -else -drm_dbg_kms (mgr->dev,"Step 2 of removing MST payload for %p failed: %d\n" , - payload->port, ret); +drm_err (mgr->dev,"Step 2 of creating MST payload for %p failed: %d\n" , + payload->port, ret); }return ret;
Signature · 1 line
2.37.3
PATCH 2/3 drm/mst: Refactor the flow for payload allocation/removement
In reply to Wayne Lin[Why]
Today, the allocation/deallocation steps and status is a bit unclear.
For instance, payload->vc_start_slot = -1 stands for "the failure of updating DPCD payload ID table" and can also represent as "payload is not allocated yet". These two cases should be handled differently and hence better to distinguish them for better understanding.
[How]
Define enumeration - ALLOCATION_LOCAL, ALLOCATION_DFP and ALLOCATION_REMOTE to distinguish different allocation status. Adjust the code to handle different status accordingly for better understanding the sequence of payload allocation and payload removement.
For payload creation, the procedure should look like this:
DRM part 1:
* step 1 - update sw mst mgr variables to add a new payload
* step 2 - add payload at immediate DFP DPCD payload table
Driver:
* Add new payload in HW and sync up with DFP by sending ACT
DRM Part 2:
* Send ALLOCATE_PAYLOAD sideband message to allocate bandwidth along the virtual channel.
And as for payload removement, the procedure should look like this:
DRM part 1:
* step 1 - Send ALLOCATE_PAYLOAD sideband message to release bandwidth along the virtual channel
* step 2 - Clear payload allocation at immediate DFP DPCD payload table
Driver:
* Remove the payload in HW and sync up with DFP by sending ACT
DRM part 2:
* update sw mst mgr variables to remove the payload
Note that it's fine to fail when communicate with the branch device connected at immediate downstrean-facing port, but updating variables of SW mst mgr and HW configuration should be conducted anyway. That's because it's under commit_tail and we need to complete the HW programming.
--- .../amd/display/amdgpu_dm/amdgpu_dm_helpers.c | 20 ++- drivers/gpu/drm/display/drm_dp_mst_topology.c | 159 +++++++++++------- drivers/gpu/drm/i915/display/intel_dp_mst.c | 18 +- drivers/gpu/drm/nouveau/dispnv50/disp.c | 15 +- include/drm/display/drm_dp_mst_helper.h | 23 ++- 5 files changed, 152 insertions(+), 83 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c +12 -8
@@ -219,7 +219,7 @@ static void dm_helpers_construct_old_payload(/* Set correct time_slots/PBN of old payload. * other fields (delete & dsc_enabled) in * struct drm_dp_mst_atomic_payload are don't care fields -* while calling drm_dp_remove_payload() +* while calling drm_dp_remove_payload_part2() */ for (i =0 ; i < current_link_table.stream_count; i++) { dc_alloc =
@@ -262,13 +262,12 @@ bool dm_helpers_dp_mst_write_payload_allocation_table( mst_mgr = &aconnector->mst_root->mst_mgr; mst_state =to_drm_dp_mst_topology_state (mst_mgr->base.state); - -/* It's OK for this to fail */ new_payload =drm_atomic_get_mst_payload_state (mst_state, aconnector->mst_output_port);if (enable) { target_payload = new_payload; +/* It's OK for this to fail */ drm_dp_add_payload_part1 (mst_mgr, mst_state, new_payload); }else {/* construct old payload by VCPI*/
@@ -276,7 +275,7 @@ bool dm_helpers_dp_mst_write_payload_allocation_table( new_payload, &old_payload); target_payload = &old_payload; -drm_dp_remove_payload (mst_mgr, mst_state, &old_payload, new_payload); +drm_dp_remove_payload_part1 (mst_mgr, mst_state, new_payload); }/* mst_mgr->->payloads are VC payload notify MST branch using DPCD or
@@ -342,7 +341,7 @@ bool dm_helpers_dp_mst_send_payload_allocation(struct amdgpu_dm_connector *aconnector;struct drm_dp_mst_topology_state *mst_state;struct drm_dp_mst_topology_mgr *mst_mgr; -struct drm_dp_mst_atomic_payload *payload; +struct drm_dp_mst_atomic_payload *new_payload, *old_payload;enum mst_progress_status set_flag =MST_ALLOCATE_NEW_PAYLOAD ;enum mst_progress_status clr_flag =MST_CLEAR_ALLOCATED_PAYLOAD ;int ret =0 ;
@@ -355,15 +354,20 @@ bool dm_helpers_dp_mst_send_payload_allocation( mst_mgr = &aconnector->mst_root->mst_mgr; mst_state =to_drm_dp_mst_topology_state (mst_mgr->base.state); - payload =drm_atomic_get_mst_payload_state (mst_state, aconnector->mst_output_port); + new_payload =drm_atomic_get_mst_payload_state (mst_state, aconnector->mst_output_port);if (!enable) { set_flag =MST_CLEAR_ALLOCATED_PAYLOAD ; clr_flag =MST_ALLOCATE_NEW_PAYLOAD ; } -if (enable) - ret =drm_dp_add_payload_part2 (mst_mgr, mst_state->base.state, payload); +if (enable) { + ret =drm_dp_add_payload_part2 (mst_mgr, mst_state->base.state, new_payload); + }else { +dm_helpers_construct_old_payload (stream->link, mst_state->pbn_div, + new_payload, old_payload); +drm_dp_remove_payload_part2 (mst_mgr, mst_state, old_payload, new_payload); + }if (ret) { amdgpu_dm_set_mst_status(&aconnector->mst_status,
drivers/gpu/drm/display/drm_dp_mst_topology.c +100 -59
284 lines not on this page · raw message
drivers/gpu/drm/i915/display/intel_dp_mst.c +12 -6
@@ -557,12 +557,8 @@ static void intel_mst_disable_dp(struct intel_atomic_state *state,struct intel_dp *intel_dp = &dig_port->dp;struct intel_connector *connector =to_intel_connector (old_conn_state->connector); -struct drm_dp_mst_topology_state *old_mst_state = -drm_atomic_get_old_mst_topology_state (&state->base, &intel_dp->mst_mgr);struct drm_dp_mst_topology_state *new_mst_state =drm_atomic_get_new_mst_topology_state (&state->base, &intel_dp->mst_mgr); -const struct drm_dp_mst_atomic_payload *old_payload = -drm_atomic_get_mst_payload_state (old_mst_state, connector->port);struct drm_dp_mst_atomic_payload *new_payload =drm_atomic_get_mst_payload_state (new_mst_state, connector->port);struct drm_i915_private *i915 =to_i915 (connector->base.dev);
@@ -572,8 +568,7 @@ static void intel_mst_disable_dp(struct intel_atomic_state *state,intel_hdcp_disable (intel_mst->connector); -drm_dp_remove_payload (&intel_dp->mst_mgr, new_mst_state, - old_payload, new_payload); +drm_dp_remove_payload_part1 (&intel_dp->mst_mgr, new_mst_state, new_payload);intel_audio_codec_disable (encoder, old_crtc_state, old_conn_state); }
@@ -588,6 +583,14 @@ static void intel_mst_post_disable_dp(struct intel_atomic_state *state,struct intel_dp *intel_dp = &dig_port->dp;struct intel_connector *connector =to_intel_connector (old_conn_state->connector); +struct drm_dp_mst_topology_state *old_mst_state = +drm_atomic_get_old_mst_topology_state (&state->base, &intel_dp->mst_mgr); +struct drm_dp_mst_topology_state *new_mst_state = +drm_atomic_get_new_mst_topology_state (&state->base, &intel_dp->mst_mgr); +const struct drm_dp_mst_atomic_payload *old_payload = +drm_atomic_get_mst_payload_state (old_mst_state, connector->port); +struct drm_dp_mst_atomic_payload *new_payload = +drm_atomic_get_mst_payload_state (new_mst_state, connector->port);struct drm_i915_private *dev_priv =to_i915 (connector->base.dev);bool last_mst_stream;
@@ -608,6 +611,9 @@ static void intel_mst_post_disable_dp(struct intel_atomic_state *state,wait_for_act_sent (encoder, old_crtc_state); +drm_dp_remove_payload_part2 (&intel_dp->mst_mgr, new_mst_state, + old_payload, new_payload); +intel_ddi_disable_transcoder_func (old_crtc_state);if (DISPLAY_VER (dev_priv) >=9 )
drivers/gpu/drm/nouveau/dispnv50/disp.c +10 -5
@@ -882,21 +882,26 @@ struct nouveau_encoder *nv50_real_outp(struct drm_encoder *encoder)static void nv50_msto_cleanup (struct drm_atomic_state *state, -struct drm_dp_mst_topology_state *mst_state, +struct drm_dp_mst_topology_state *new_mst_state,struct drm_dp_mst_topology_mgr *mgr,struct nv50_msto *msto) {struct nouveau_drm *drm =nouveau_drm (msto->encoder.dev); -struct drm_dp_mst_atomic_payload *payload = -drm_atomic_get_mst_payload_state (mst_state, msto->mstc->port); +struct drm_dp_mst_atomic_payload *new_payload = +drm_atomic_get_mst_payload_state (new_mst_state, msto->mstc->port); +struct drm_dp_mst_topology_state *old_mst_state = +drm_atomic_get_old_mst_topology_state (state, mgr); +const struct drm_dp_mst_atomic_payload *old_payload = +drm_atomic_get_mst_payload_state (old_mst_state, msto->mstc->port);NV_ATOMIC (drm,"%s: msto cleanup\n" , msto->encoder.name);if (msto->disabled) { msto->mstc =NULL ; msto->disabled = false; +drm_dp_remove_payload_part2 (mgr, new_mst_state, old_payload, new_payload); }else if (msto->enabled) { -drm_dp_add_payload_part2 (mgr, state, payload); +drm_dp_add_payload_part2 (mgr, state, new_payload); msto->enabled = false; } }
@@ -922,7 +927,7 @@ nv50_msto_prepare(struct drm_atomic_state *state,// TODO: Figure out if we want to do a better job of handling VCPI allocation failures here? if (msto->disabled) { -drm_dp_remove_payload (mgr, mst_state, old_payload, payload); +drm_dp_remove_payload_part1 (mgr, mst_state, payload);nvif_outp_dp_mst_vcpi (&mstm->outp->outp, msto->head->base.index,0 ,0 ,0 ,0 ); }else {
include/drm/display/drm_dp_mst_helper.h +18 -5
@@ -46,6 +46,13 @@ struct drm_dp_mst_topology_ref_history { };#endif /* IS_ENABLED(CONFIG_DRM_DEBUG_DP_MST_TOPOLOGY_REFS) */ +enum drm_dp_mst_payload_allocation { +DRM_DP_MST_PAYLOAD_ALLOCATION_NONE , +DRM_DP_MST_PAYLOAD_ALLOCATION_LOCAL , +DRM_DP_MST_PAYLOAD_ALLOCATION_DFP , +DRM_DP_MST_PAYLOAD_ALLOCATION_REMOTE , +}; +struct drm_dp_mst_branch ;/**
@@ -537,7 +544,7 @@ struct drm_dp_mst_atomic_payload {* drm_dp_mst_atomic_wait_for_dependencies() has been called, which will ensure the * previous MST states payload start slots have been copied over to the new state. Note * that a new start slot won't be assigned/removed from this payload until -* drm_dp_add_payload_part1()/drm_dp_remove_payload() have been called. +* drm_dp_add_payload_part1()/drm_dp_remove_payload_part2() have been called. * * Acquire the MST modesetting lock, and then wait for any pending MST-related commits to * get committed to hardware by calling drm_crtc_commit_wait() on each of the * &drm_crtc_commit structs in &drm_dp_mst_topology_state.commit_deps.
@@ -564,6 +571,9 @@ struct drm_dp_mst_atomic_payload {/** @dsc_enabled: Whether or not this payload has DSC enabled */ bool dsc_enabled : 1; +/** @payload_allocation_status: The allocation status of this payload */ +enum drm_dp_mst_payload_allocation payload_allocation_status; +/** @next: The list node for this payload */ struct list_head next; };
@@ -842,10 +852,13 @@ int drm_dp_add_payload_part1(struct drm_dp_mst_topology_mgr *mgr,int drm_dp_add_payload_part2 (struct drm_dp_mst_topology_mgr *mgr,struct drm_atomic_state *state,struct drm_dp_mst_atomic_payload *payload); -void drm_dp_remove_payload (struct drm_dp_mst_topology_mgr *mgr, -struct drm_dp_mst_topology_state *mst_state, -const struct drm_dp_mst_atomic_payload *old_payload, -struct drm_dp_mst_atomic_payload *new_payload); +void drm_dp_remove_payload_part1 (struct drm_dp_mst_topology_mgr *mgr, +struct drm_dp_mst_topology_state *mst_state, +struct drm_dp_mst_atomic_payload *payload); +void drm_dp_remove_payload_part2 (struct drm_dp_mst_topology_mgr *mgr, +struct drm_dp_mst_topology_state *mst_state, +const struct drm_dp_mst_atomic_payload *old_payload, +struct drm_dp_mst_atomic_payload *new_payload);int drm_dp_check_act_status (struct drm_dp_mst_topology_mgr *mgr);
Signature · 1 line
2.37.3
kernel test robot
build botRe: [PATCH 2/3] drm/mst: Refactor the flow for payload allocation/removement
kernel test robot noticed the following build warnings:
Hi Wayne,
kernel test robot noticed the following build warnings:
[auto build test WARNING on drm-misc/drm-misc-next]
[also build test WARNING on drm-intel/for-linux-next-fixes drm/drm-next linus/master v6.5-rc4 next-20230804] [cannot apply to drm-intel/for-linux-next]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Wayne-Lin/drm-mst-delete-unnecessary-case-in-drm_dp_add_payload_part2/20230804-142405 base: git://anongit.freedesktop.org/drm/drm-misc drm-misc-next patch link: https://lore.kernel.org/r/20230804062029.5686-3-Wayne.Lin%40amd.com patch subject: [PATCH 2/3] drm/mst: Refactor the flow for payload allocation/removement config: microblaze-randconfig-r021-20230731 (https://download.01.org/0day-ci/archive/20230804/[email protected]/config) compiler: microblaze-linux-gcc (GCC) 12.3.0 reproduce: (https://download.01.org/0day-ci/archive/20230804/[email protected]/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/
All warnings (new ones prefixed by >>):
drivers/gpu/drm/nouveau/dispnv50/disp.c: In function 'nv50_msto_prepare':
>> drivers/gpu/drm/nouveau/dispnv50/disp.c:919:53: warning: variable 'old_payload' set but not used [-Wunused-but-set-variable]
919 | struct drm_dp_mst_atomic_payload *payload, *old_payload;
| ^~~~~~~~~~~vim +/old_payload +919 drivers/gpu/drm/nouveau/dispnv50/disp.c
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 908
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 909 static void
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 910 nv50_msto_prepare(struct drm_atomic_state *state,
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 911 struct drm_dp_mst_topology_state *mst_state,
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 912 struct drm_dp_mst_topology_mgr *mgr,
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 913 struct nv50_msto *msto)
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 914 {
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 915 struct nouveau_drm *drm = nouveau_drm(msto->encoder.dev);
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 916 struct nv50_mstc *mstc = msto->mstc;
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 917 struct nv50_mstm *mstm = mstc->mstm;
8fb3e25c3dd1a2 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2023-06-13 918 struct drm_dp_mst_topology_state *old_mst_state;
8fb3e25c3dd1a2 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2023-06-13 @919 struct drm_dp_mst_atomic_payload *payload, *old_payload;
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 920
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 921 NV_ATOMIC(drm, "%s: msto prepare\n", msto->encoder.name);
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 922
8fb3e25c3dd1a2 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2023-06-13 923 old_mst_state = drm_atomic_get_old_mst_topology_state(state, mgr);
8fb3e25c3dd1a2 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2023-06-13 924
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 925 payload = drm_atomic_get_mst_payload_state(mst_state, mstc->port);
8fb3e25c3dd1a2 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2023-06-13 926 old_payload = drm_atomic_get_mst_payload_state(old_mst_state, mstc->port);
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 927
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 928 // TODO: Figure out if we want to do a better job of handling VCPI allocation failures here?
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 929 if (msto->disabled) {
c9e8c7f37133c0 drivers/gpu/drm/nouveau/dispnv50/disp.c Wayne Lin 2023-08-04 930 drm_dp_remove_payload_part1(mgr, mst_state, payload);
8c7d980da9ba3e drivers/gpu/drm/nouveau/dispnv50/disp.c Ben Skeggs 2022-06-01 931
8c7d980da9ba3e drivers/gpu/drm/nouveau/dispnv50/disp.c Ben Skeggs 2022-06-01 932 nvif_outp_dp_mst_vcpi(&mstm->outp->outp, msto->head->base.index, 0, 0, 0, 0);
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 933 } else {
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 934 if (msto->enabled)
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 935 drm_dp_add_payload_part1(mgr, mst_state, payload);
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 936
8c7d980da9ba3e drivers/gpu/drm/nouveau/dispnv50/disp.c Ben Skeggs 2022-06-01 937 nvif_outp_dp_mst_vcpi(&mstm->outp->outp, msto->head->base.index,
8c7d980da9ba3e drivers/gpu/drm/nouveau/dispnv50/disp.c Ben Skeggs 2022-06-01 938 payload->vc_start_slot, payload->time_slots,
8c7d980da9ba3e drivers/gpu/drm/nouveau/dispnv50/disp.c Ben Skeggs 2022-06-01 939 payload->pbn, payload->time_slots * mst_state->pbn_div);
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 940 }
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 941 }
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 942 Signature · 2 lines
0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki
kernel test robot
build botRe: [PATCH 2/3] drm/mst: Refactor the flow for payload allocation/removement
In reply to Wayne Linkernel test robot noticed the following build warnings:
Hi Wayne,
kernel test robot noticed the following build warnings:
[auto build test WARNING on drm-misc/drm-misc-next]
[also build test WARNING on drm-intel/for-linux-next-fixes drm/drm-next linus/master v6.5-rc4 next-20230804] [cannot apply to drm-intel/for-linux-next]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Wayne-Lin/drm-mst-delete-unnecessary-case-in-drm_dp_add_payload_part2/20230804-142405 base: git://anongit.freedesktop.org/drm/drm-misc drm-misc-next patch link: https://lore.kernel.org/r/20230804062029.5686-3-Wayne.Lin%40amd.com patch subject: [PATCH 2/3] drm/mst: Refactor the flow for payload allocation/removement config: s390-randconfig-r044-20230731 (https://download.01.org/0day-ci/archive/20230804/[email protected]/config) compiler: clang version 17.0.0 (https://github.com/llvm/llvm-project.git 4a5ac14ee968ff0ad5d2cc1ffa0299048db4c88a) reproduce: (https://download.01.org/0day-ci/archive/20230804/[email protected]/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/
All warnings (new ones prefixed by >>):
In file included from drivers/gpu/drm/nouveau/dispnv50/disp.c:24:
In file included from drivers/gpu/drm/nouveau/dispnv50/disp.h:4:
In file included from drivers/gpu/drm/nouveau/include/nvif/mem.h:3:
In file included from drivers/gpu/drm/nouveau/include/nvif/mmu.h:3:
In file included from drivers/gpu/drm/nouveau/include/nvif/object.h:4:
In file included from drivers/gpu/drm/nouveau/include/nvif/os.h:8:
In file included from include/linux/pci.h:39:
In file included from include/linux/io.h:13:
In file included from arch/s390/include/asm/io.h:75:
include/asm-generic/io.h:547:31: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
547 | val = __raw_readb(PCI_IOBASE + addr);
| ~~~~~~~~~~ ^
include/asm-generic/io.h:560:61: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
560 | val = __le16_to_cpu((__le16 __force)__raw_readw(PCI_IOBASE + addr));
| ~~~~~~~~~~ ^
include/uapi/linux/byteorder/big_endian.h:37:59: note: expanded from macro '__le16_to_cpu'
37 | #define __le16_to_cpu(x) __swab16((__force __u16)(__le16)(x))
| ^
include/uapi/linux/swab.h:102:54: note: expanded from macro '__swab16'
102 | #define __swab16(x) (__u16)__builtin_bswap16((__u16)(x))
| ^
In file included from drivers/gpu/drm/nouveau/dispnv50/disp.c:24:
In file included from drivers/gpu/drm/nouveau/dispnv50/disp.h:4:
In file included from drivers/gpu/drm/nouveau/include/nvif/mem.h:3:
In file included from drivers/gpu/drm/nouveau/include/nvif/mmu.h:3:
In file included from drivers/gpu/drm/nouveau/include/nvif/object.h:4:
In file included from drivers/gpu/drm/nouveau/include/nvif/os.h:8:
In file included from include/linux/pci.h:39:
In file included from include/linux/io.h:13:
In file included from arch/s390/include/asm/io.h:75:
include/asm-generic/io.h:573:61: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
573 | val = __le32_to_cpu((__le32 __force)__raw_readl(PCI_IOBASE + addr));
| ~~~~~~~~~~ ^
include/uapi/linux/byteorder/big_endian.h:35:59: note: expanded from macro '__le32_to_cpu'
35 | #define __le32_to_cpu(x) __swab32((__force __u32)(__le32)(x))
| ^
include/uapi/linux/swab.h:115:54: note: expanded from macro '__swab32'
115 | #define __swab32(x) (__u32)__builtin_bswap32((__u32)(x))
| ^
In file included from drivers/gpu/drm/nouveau/dispnv50/disp.c:24:
In file included from drivers/gpu/drm/nouveau/dispnv50/disp.h:4:
In file included from drivers/gpu/drm/nouveau/include/nvif/mem.h:3:
In file included from drivers/gpu/drm/nouveau/include/nvif/mmu.h:3:
In file included from drivers/gpu/drm/nouveau/include/nvif/object.h:4:
In file included from drivers/gpu/drm/nouveau/include/nvif/os.h:8:
In file included from include/linux/pci.h:39:
In file included from include/linux/io.h:13:
In file included from arch/s390/include/asm/io.h:75:
include/asm-generic/io.h:584:33: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
584 | __raw_writeb(value, PCI_IOBASE + addr);
| ~~~~~~~~~~ ^
include/asm-generic/io.h:594:59: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
594 | __raw_writew((u16 __force)cpu_to_le16(value), PCI_IOBASE + addr);
| ~~~~~~~~~~ ^
include/asm-generic/io.h:604:59: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
604 | __raw_writel((u32 __force)cpu_to_le32(value), PCI_IOBASE + addr);
| ~~~~~~~~~~ ^
include/asm-generic/io.h:692:20: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
692 | readsb(PCI_IOBASE + addr, buffer, count);
| ~~~~~~~~~~ ^
include/asm-generic/io.h:700:20: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
700 | readsw(PCI_IOBASE + addr, buffer, count);
| ~~~~~~~~~~ ^
include/asm-generic/io.h:708:20: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
708 | readsl(PCI_IOBASE + addr, buffer, count);
| ~~~~~~~~~~ ^
include/asm-generic/io.h:717:21: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
717 | writesb(PCI_IOBASE + addr, buffer, count);
| ~~~~~~~~~~ ^
include/asm-generic/io.h:726:21: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
726 | writesw(PCI_IOBASE + addr, buffer, count);
| ~~~~~~~~~~ ^
include/asm-generic/io.h:735:21: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
735 | writesl(PCI_IOBASE + addr, buffer, count);
| ~~~~~~~~~~ ^
>> drivers/gpu/drm/nouveau/dispnv50/disp.c:919:46: warning: variable 'old_payload' set but not used [-Wunused-but-set-variable]
919 | struct drm_dp_mst_atomic_payload *payload, *old_payload;
| ^13 warnings generated.
vim +/old_payload +919 drivers/gpu/drm/nouveau/dispnv50/disp.c
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 908
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 909 static void
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 910 nv50_msto_prepare(struct drm_atomic_state *state,
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 911 struct drm_dp_mst_topology_state *mst_state,
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 912 struct drm_dp_mst_topology_mgr *mgr,
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 913 struct nv50_msto *msto)
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 914 {
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 915 struct nouveau_drm *drm = nouveau_drm(msto->encoder.dev);
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 916 struct nv50_mstc *mstc = msto->mstc;
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 917 struct nv50_mstm *mstm = mstc->mstm;
8fb3e25c3dd1a2 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2023-06-13 918 struct drm_dp_mst_topology_state *old_mst_state;
8fb3e25c3dd1a2 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2023-06-13 @919 struct drm_dp_mst_atomic_payload *payload, *old_payload;
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 920
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 921 NV_ATOMIC(drm, "%s: msto prepare\n", msto->encoder.name);
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 922
8fb3e25c3dd1a2 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2023-06-13 923 old_mst_state = drm_atomic_get_old_mst_topology_state(state, mgr);
8fb3e25c3dd1a2 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2023-06-13 924
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 925 payload = drm_atomic_get_mst_payload_state(mst_state, mstc->port);
8fb3e25c3dd1a2 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2023-06-13 926 old_payload = drm_atomic_get_mst_payload_state(old_mst_state, mstc->port);
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 927
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 928 // TODO: Figure out if we want to do a better job of handling VCPI allocation failures here?
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 929 if (msto->disabled) {
c9e8c7f37133c0 drivers/gpu/drm/nouveau/dispnv50/disp.c Wayne Lin 2023-08-04 930 drm_dp_remove_payload_part1(mgr, mst_state, payload);
8c7d980da9ba3e drivers/gpu/drm/nouveau/dispnv50/disp.c Ben Skeggs 2022-06-01 931
8c7d980da9ba3e drivers/gpu/drm/nouveau/dispnv50/disp.c Ben Skeggs 2022-06-01 932 nvif_outp_dp_mst_vcpi(&mstm->outp->outp, msto->head->base.index, 0, 0, 0, 0);
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 933 } else {
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 934 if (msto->enabled)
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 935 drm_dp_add_payload_part1(mgr, mst_state, payload);
4d07b0bc403403 drivers/gpu/drm/nouveau/dispnv50/disp.c Lyude Paul 2022-08-17 936
8c7d980da9ba3e drivers/gpu/drm/nouveau/dispnv50/disp.c Ben Skeggs 2022-06-01 937 nvif_outp_dp_mst_vcpi(&mstm->outp->outp, msto->head->base.index,
8c7d980da9ba3e drivers/gpu/drm/nouveau/dispnv50/disp.c Ben Skeggs 2022-06-01 938 payload->vc_start_slot, payload->time_slots,
8c7d980da9ba3e drivers/gpu/drm/nouveau/dispnv50/disp.c Ben Skeggs 2022-06-01 939 payload->pbn, payload->time_slots * mst_state->pbn_div);
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 940 }
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 941 }
f479c0ba4a170a drivers/gpu/drm/nouveau/nv50_display.c Ben Skeggs 2016-11-04 942 Signature · 2 lines
0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki
PATCH 3/3 drm/mst: adjust the function drm_dp_remove_payload_part2()
In reply to Wayne Lin[Why]
Now in drm_dp_remove_payload_part2(), it utilizes the time slot number of the payload in old state to represent the one in the payload table at the moment.
It would be better to clarify the idea by using the latest allocated time slot number for the port at the moment instead and which info is already included in new mst_state. By this, we can also remove redundant workaround for amdgpu driver.
[How]
Remove "old_payload" input of drm_dp_remove_payload_part2() and get the latest number of allocated time slot for the port from new mst_state instead.
--- .../amd/display/amdgpu_dm/amdgpu_dm_helpers.c | 70 ++++--------------- drivers/gpu/drm/display/drm_dp_mst_topology.c | 32 ++++++--- drivers/gpu/drm/i915/display/intel_dp_mst.c | 7 +- drivers/gpu/drm/nouveau/dispnv50/disp.c | 6 +- include/drm/display/drm_dp_mst_helper.h | 9 ++- 5 files changed, 40 insertions(+), 84 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c +13 -57
@@ -203,40 +203,6 @@ void dm_helpers_dp_update_branch_info( const struct dc_link *link) {} -static void dm_helpers_construct_old_payload ( -struct dc_link *link, -int pbn_per_slot, -struct drm_dp_mst_atomic_payload *new_payload, -struct drm_dp_mst_atomic_payload *old_payload) -{ -struct link_mst_stream_allocation_table current_link_table = - link->mst_stream_alloc_table; -struct link_mst_stream_allocation *dc_alloc; -int i; - - *old_payload = *new_payload; - -/* Set correct time_slots/PBN of old payload. -* other fields (delete & dsc_enabled) in -* struct drm_dp_mst_atomic_payload are don't care fields -* while calling drm_dp_remove_payload_part2() -*/ -for (i =0 ; i < current_link_table.stream_count; i++) { - dc_alloc = - ¤t_link_table.stream_allocations[i]; - -if (dc_alloc->vcp_id == new_payload->vcpi) { - old_payload->time_slots = dc_alloc->slot_count; - old_payload->pbn = dc_alloc->slot_count * pbn_per_slot; -break ; - } - } - -/* make sure there is an old payload*/ -ASSERT (i != current_link_table.stream_count); - -} -/* * Writes payload allocation table in immediate downstream device. */
@@ -248,7 +214,7 @@ bool dm_helpers_dp_mst_write_payload_allocation_table( {struct amdgpu_dm_connector *aconnector;struct drm_dp_mst_topology_state *mst_state; -struct drm_dp_mst_atomic_payload *target_payload, *new_payload, old_payload; +struct drm_dp_mst_atomic_payload *payload;struct drm_dp_mst_topology_mgr *mst_mgr; aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
@@ -262,27 +228,20 @@ bool dm_helpers_dp_mst_write_payload_allocation_table( mst_mgr = &aconnector->mst_root->mst_mgr; mst_state =to_drm_dp_mst_topology_state (mst_mgr->base.state); - new_payload =drm_atomic_get_mst_payload_state (mst_state, aconnector->mst_output_port); - -if (enable) { - target_payload = new_payload; + payload =drm_atomic_get_mst_payload_state (mst_state, aconnector->mst_output_port); +if (enable)/* It's OK for this to fail */ -drm_dp_add_payload_part1 (mst_mgr, mst_state, new_payload); - }else { -/* construct old payload by VCPI*/ -dm_helpers_construct_old_payload (stream->link, mst_state->pbn_div, - new_payload, &old_payload); - target_payload = &old_payload; +drm_dp_add_payload_part1 (mst_mgr, mst_state, payload); +else -drm_dp_remove_payload_part1 (mst_mgr, mst_state, new_payload); - } +drm_dp_remove_payload_part1 (mst_mgr, mst_state, payload);/* mst_mgr->->payloads are VC payload notify MST branch using DPCD or * AUX message. The sequence is slot 1-63 allocated sequence for each * stream. AMD ASIC stream slot allocation should follow the same * sequence. copy DRM MST allocation to dc */ -fill_dc_mst_payload_table_from_drm (stream->link, enable, target_payload, proposed_table); +fill_dc_mst_payload_table_from_drm (stream->link, enable, payload, proposed_table);return true; }
@@ -341,7 +300,7 @@ bool dm_helpers_dp_mst_send_payload_allocation(struct amdgpu_dm_connector *aconnector;struct drm_dp_mst_topology_state *mst_state;struct drm_dp_mst_topology_mgr *mst_mgr; -struct drm_dp_mst_atomic_payload *new_payload, *old_payload; +struct drm_dp_mst_atomic_payload *payload;enum mst_progress_status set_flag =MST_ALLOCATE_NEW_PAYLOAD ;enum mst_progress_status clr_flag =MST_CLEAR_ALLOCATED_PAYLOAD ;int ret =0 ;
@@ -354,20 +313,17 @@ bool dm_helpers_dp_mst_send_payload_allocation( mst_mgr = &aconnector->mst_root->mst_mgr; mst_state =to_drm_dp_mst_topology_state (mst_mgr->base.state); - new_payload =drm_atomic_get_mst_payload_state (mst_state, aconnector->mst_output_port); + payload =drm_atomic_get_mst_payload_state (mst_state, aconnector->mst_output_port);if (!enable) { set_flag =MST_CLEAR_ALLOCATED_PAYLOAD ; clr_flag =MST_ALLOCATE_NEW_PAYLOAD ; } -if (enable) { - ret =drm_dp_add_payload_part2 (mst_mgr, mst_state->base.state, new_payload); - }else { -dm_helpers_construct_old_payload (stream->link, mst_state->pbn_div, - new_payload, old_payload); -drm_dp_remove_payload_part2 (mst_mgr, mst_state, old_payload, new_payload); - } +if (enable) + ret =drm_dp_add_payload_part2 (mst_mgr, mst_state->base.state, payload); +else +drm_dp_remove_payload_part2 (mst_mgr, mst_state, payload);if (ret) { amdgpu_dm_set_mst_status(&aconnector->mst_status,
drivers/gpu/drm/display/drm_dp_mst_topology.c +21 -11
@@ -3382,8 +3382,7 @@ EXPORT_SYMBOL(drm_dp_remove_payload_part1);* drm_dp_remove_payload_part2() - Remove an MST payload locally * @mgr: Manager to use. * @mst_state: The MST atomic state -* @old_payload: The payload with its old state -* @new_payload: The payload with its latest state +* @payload: The payload with its latest state * * Updates the starting time slots of all other payloads which would have been shifted towards * the start of the payload ID table as a result of removing a payload. Driver should call this
@@ -3392,25 +3391,36 @@ EXPORT_SYMBOL(drm_dp_remove_payload_part1);*/ void drm_dp_remove_payload_part2 (struct drm_dp_mst_topology_mgr *mgr,struct drm_dp_mst_topology_state *mst_state, -const struct drm_dp_mst_atomic_payload *old_payload, -struct drm_dp_mst_atomic_payload *new_payload) +struct drm_dp_mst_atomic_payload *payload) {struct drm_dp_mst_atomic_payload *pos; +u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; + +/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - payload->vc_start_slot;/* Remove local payload allocation */ list_for_each_entry (pos, &mst_state->payloads, next) { -if (pos != new_payload && pos->vc_start_slot > new_payload->vc_start_slot) - pos->vc_start_slot -= old_payload->time_slots; +if (pos != payload && pos->vc_start_slot > payload->vc_start_slot) + pos->vc_start_slot -= time_slots_to_remove; } - new_payload->vc_start_slot =-1 ; + payload->vc_start_slot =-1 ; mgr->payload_count--; - mgr->next_start_slot -= old_payload->time_slots; + mgr->next_start_slot -= time_slots_to_remove; -if (new_payload->delete) -drm_dp_mst_put_port_malloc (new_payload->port); +if (payload->delete) +drm_dp_mst_put_port_malloc (payload->port); - new_payload->payload_allocation_status =DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; + payload->payload_allocation_status =DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; }EXPORT_SYMBOL (drm_dp_remove_payload_part2);/**
drivers/gpu/drm/i915/display/intel_dp_mst.c +1 -6
@@ -583,12 +583,8 @@ static void intel_mst_post_disable_dp(struct intel_atomic_state *state,struct intel_dp *intel_dp = &dig_port->dp;struct intel_connector *connector =to_intel_connector (old_conn_state->connector); -struct drm_dp_mst_topology_state *old_mst_state = -drm_atomic_get_old_mst_topology_state (&state->base, &intel_dp->mst_mgr);struct drm_dp_mst_topology_state *new_mst_state =drm_atomic_get_new_mst_topology_state (&state->base, &intel_dp->mst_mgr); -const struct drm_dp_mst_atomic_payload *old_payload = -drm_atomic_get_mst_payload_state (old_mst_state, connector->port);struct drm_dp_mst_atomic_payload *new_payload =drm_atomic_get_mst_payload_state (new_mst_state, connector->port);struct drm_i915_private *dev_priv =to_i915 (connector->base.dev);
@@ -611,8 +607,7 @@ static void intel_mst_post_disable_dp(struct intel_atomic_state *state,wait_for_act_sent (encoder, old_crtc_state); -drm_dp_remove_payload_part2 (&intel_dp->mst_mgr, new_mst_state, - old_payload, new_payload); +drm_dp_remove_payload_part2 (&intel_dp->mst_mgr, new_mst_state, new_payload);intel_ddi_disable_transcoder_func (old_crtc_state);
drivers/gpu/drm/nouveau/dispnv50/disp.c +1 -5
@@ -889,17 +889,13 @@ nv50_msto_cleanup(struct drm_atomic_state *state,struct nouveau_drm *drm =nouveau_drm (msto->encoder.dev);struct drm_dp_mst_atomic_payload *new_payload =drm_atomic_get_mst_payload_state (new_mst_state, msto->mstc->port); -struct drm_dp_mst_topology_state *old_mst_state = -drm_atomic_get_old_mst_topology_state (state, mgr); -const struct drm_dp_mst_atomic_payload *old_payload = -drm_atomic_get_mst_payload_state (old_mst_state, msto->mstc->port);NV_ATOMIC (drm,"%s: msto cleanup\n" , msto->encoder.name);if (msto->disabled) { msto->mstc =NULL ; msto->disabled = false; -drm_dp_remove_payload_part2 (mgr, new_mst_state, old_payload, new_payload); +drm_dp_remove_payload_part2 (mgr, new_mst_state, new_payload); }else if (msto->enabled) {drm_dp_add_payload_part2 (mgr, state, new_payload); msto->enabled = false;
include/drm/display/drm_dp_mst_helper.h +4 -5
@@ -853,12 +853,11 @@ int drm_dp_add_payload_part2(struct drm_dp_mst_topology_mgr *mgr,struct drm_atomic_state *state, struct drm_dp_mst_atomic_payload *payload);void drm_dp_remove_payload_part1 (struct drm_dp_mst_topology_mgr *mgr, -struct drm_dp_mst_topology_state *mst_state, -struct drm_dp_mst_atomic_payload *payload); +struct drm_dp_mst_topology_state *mst_state, +struct drm_dp_mst_atomic_payload *payload);void drm_dp_remove_payload_part2 (struct drm_dp_mst_topology_mgr *mgr, -struct drm_dp_mst_topology_state *mst_state, -const struct drm_dp_mst_atomic_payload *old_payload, -struct drm_dp_mst_atomic_payload *new_payload); +struct drm_dp_mst_topology_state *mst_state, +struct drm_dp_mst_atomic_payload *payload);int drm_dp_check_act_status (struct drm_dp_mst_topology_mgr *mgr);
Signature · 1 line
2.37.3
Re: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
Quoted 26 lines · Wayne Lin
[...]
drivers/gpu/drm/display/drm_dp_mst_topology.c +4
-4@@ -3382,8 +3382,7 @@ EXPORT_SYMBOL(drm_dp_remove_payload_part1);* drm_dp_remove_payload_part2() - Remove an MST payload locally * @mgr: Manager to use. * @mst_state: The MST atomic state -* @old_payload: The payload with its old state -* @new_payload: The payload with its latest state +* @payload: The payload with its latest state * * Updates the starting time slots of all other payloads which would have been shifted towards * the start of the payload ID table as a result of removing a payload. Driver should call this @@ -3392,25 +3391,36 @@ EXPORT_SYMBOL(drm_dp_remove_payload_part1);*/ void drm_dp_remove_payload_part2 (struct drm_dp_mst_topology_mgr *mgr,struct drm_dp_mst_topology_state *mst_state, -const struct drm_dp_mst_atomic_payload *old_payload, -struct drm_dp_mst_atomic_payload *new_payload) +struct drm_dp_mst_atomic_payload *payload) {struct drm_dp_mst_atomic_payload *pos; +u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot;
drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1); · line 3408
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +10
-0+ +/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - payload->vc_start_slot;
Imo, the intuitive way would be to pass the old payload state to this function - which already contains the required time_slots param - and refactor things instead moving vc_start_slot from the payload state to mgr suggested by Ville earlier.
--Imre
Quoted 22 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +7
-7/* Remove local payload allocation */ list_for_each_entry (pos, &mst_state->payloads, next) { -if (pos != new_payload && pos->vc_start_slot > new_payload->vc_start_slot) - pos->vc_start_slot -= old_payload->time_slots; +if (pos != payload && pos->vc_start_slot > payload->vc_start_slot) + pos->vc_start_slot -= time_slots_to_remove; } - new_payload->vc_start_slot =-1 ; + payload->vc_start_slot =-1 ; mgr->payload_count--; - mgr->next_start_slot -= old_payload->time_slots; + mgr->next_start_slot -= time_slots_to_remove; -if (new_payload->delete) -drm_dp_mst_put_port_malloc (new_payload->port); +if (payload->delete) +drm_dp_mst_put_port_malloc (payload->port); - new_payload->payload_allocation_status =DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; + payload->payload_allocation_status =DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; }
RE: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
[AMD Official Use Only - General]
Quoted 64 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 4, 2023 11:32 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 7 lines · Wayne Lin
[...]
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0@@ -3382,8 +3382,7 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 8 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-2* drm_dp_remove_payload_part2() - Remove an MST payload locally * @mgr: Manager to use. * @mst_state: The MST atomic state -* @old_payload: The payload with its old state -* @new_payload: The payload with its latest state +* @payload: The payload with its latest state * * Updates the starting time slots of all other payloads which would have been shifted towards
Quoted 2 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0* the start of the payload ID table as a result of removing a payload. Driver should call this @@ -3392,25 +3391,36 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 2 lines
*/
void drm_dp_remove_payload_part2(structdrm_dp_mst_topology_mgr*mgr,
Quoted 1 line
struct
drm_dp_mst_topology_state
*mst_state,Quoted 1 line
- const struct
drm_dp_mst_atomic_payload
*old_payload,Quoted 1 line
- struct
drm_dp_mst_atomic_payload
*new_payload)Quoted 1 line
+ struct
drm_dp_mst_atomic_payload*payload)
Quoted 5 lines
{ struct drm_dp_mst_atomic_payload *pos;drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-0+u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; +drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +10
-0+/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - +payload->vc_start_slot;Imo, the intuitive way would be to pass the old payload state to this function - which already contains the required
time_slotsparam - and refactor things instead movingvc_start_slotfrom the payload state to mgr suggested by Ville earlier.
Quoted 1 line
--Imre
Hi Imre,
Thanks for your feedback!
I understand it's functionally correct. But IMHO, it's still a bit conceptually different between the time slot in old state and the time slot in current payload table. My thought is the time slot at the moment when we are removing the payload would be a better choice. And with this, we can also simplify some codes. Especially remove workaround in amd driver. In fact, DRM mst code maintains the payload table and all the time slot info is in it already. We don't really have to pass a new parameter.
Quoted 17 lines
Quoted 17 lines
/* Remove local payload allocation */ list_for_each_entry(pos, &mst_state->payloads, next) {- if (pos !=
new_payload&&pos->vc_start_slot> new_payload-vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-1- pos->vc_start_slot -= old_payload->time_slots; + if (pos != payload && pos->vc_start_slot > payload-
vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-2+ pos->vc_start_slot -= time_slots_to_remove; } - new_payload->vc_start_slot =-1 ; + payload->vc_start_slot =-1 ; mgr->payload_count--; - mgr->next_start_slot -= old_payload->time_slots; + mgr->next_start_slot -= time_slots_to_remove;
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 6 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-3-if (new_payload->delete) -drm_dp_mst_put_port_malloc (new_payload->port); +if (payload->delete) +drm_dp_mst_put_port_malloc (payload->port); - new_payload->payload_allocation_status =
DRM_DP_MST_PAYLOAD_ALLOCATION_NONE;Quoted 3 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-0+ payload->payload_allocation_status = +DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; }
--
Regards,
Wayne
Re: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
Quoted 68 lines · Wayne
[AMD Official Use Only - General]
Quoted 64 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 4, 2023 11:32 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 7 lines · Wayne Lin
[...]
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0@@ -3382,8 +3382,7 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 8 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-2* drm_dp_remove_payload_part2() - Remove an MST payload locally * @mgr: Manager to use. * @mst_state: The MST atomic state -* @old_payload: The payload with its old state -* @new_payload: The payload with its latest state +* @payload: The payload with its latest state * * Updates the starting time slots of all other payloads which would have been shifted towards
Quoted 2 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0* the start of the payload ID table as a result of removing a payload. Driver should call this @@ -3392,25 +3391,36 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 2 lines
*/
void drm_dp_remove_payload_part2(structdrm_dp_mst_topology_mgr*mgr,
Quoted 1 line
struct
drm_dp_mst_topology_state
*mst_state,Quoted 1 line
- const struct
drm_dp_mst_atomic_payload
*old_payload,Quoted 1 line
- struct
drm_dp_mst_atomic_payload
*new_payload)Quoted 1 line
+ struct
drm_dp_mst_atomic_payload*payload)
Quoted 5 lines
{ struct drm_dp_mst_atomic_payload *pos;drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-0+u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; +drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +10
-0+/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - +payload->vc_start_slot;Imo, the intuitive way would be to pass the old payload state to this function - which already contains the required
time_slotsparam - and refactor things instead movingvc_start_slotfrom the payload state to mgr suggested by Ville earlier.Quoted 1 line
--Imre
Quoted 7 lines
Hi Imre,
Thanks for your feedback!I understand it's functionally correct. But IMHO, it's still a bit conceptually different between the time slot in old state and the time slot in current payload table. My thought is the time slot at the moment when we are removing the payload would be a better choice.
Yes, they are different. The old state contains the time slot the payload was added with in a preceding commit and so the time slot value which should be used when removing the same payload in the current commit.
The new state contains a time slot value with which the payload will be added in the current commit and can be different than the one in the old state if the current commit has changed the payload size (meaning that the same atomic commit will first remove the payload using the time slot value in the old state and afterwards will add back the same payload using the time slot value in the new state).
Quoted 4 lines
And with this, we can also simplify some codes. Especially remove workaround in amd driver. In fact, DRM mst code maintains the payload table and all the time slot info is in it already. We don't really have to pass a new parameter.
I agree that drm_dp_remove_payload() could be simplified, but this should be done so that the drivers can pass the old payload state to it (without having to pass the new state). This would be possible if vc_start_slot was not tracked in the payload state (which is really not an atomic state that can be precomputed as all other atomic state), rather it would be tracked in struct drm_dp_mst_topology_mgr.
It looks like AMD has to reconstruct the old state in dm_helpers_construct_old_payload(). Could you explain why it couldn't instead just pass old_payload acquired by
old_mst_state = drm_atomic_get_old_mst_topology_state(); old_payload = drm_atomic_get_mst_payload_state(old_mst_state);
?
Quoted 30 lines
Quoted 16 lines
Quoted 16 lines
/* Remove local payload allocation */ list_for_each_entry(pos, &mst_state->payloads, next) {- if (pos !=
new_payload&&pos->vc_start_slot> new_payload-vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-1- pos->vc_start_slot -= old_payload->time_slots; + if (pos != payload && pos->vc_start_slot > payload-
vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-2+ pos->vc_start_slot -= time_slots_to_remove; } - new_payload->vc_start_slot =-1 ; + payload->vc_start_slot =-1 ; mgr->payload_count--; - mgr->next_start_slot -= old_payload->time_slots; + mgr->next_start_slot -= time_slots_to_remove;Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 6 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-3-if (new_payload->delete) -drm_dp_mst_put_port_malloc (new_payload->port); +if (payload->delete) +drm_dp_mst_put_port_malloc (payload->port); - new_payload->payload_allocation_status =
DRM_DP_MST_PAYLOAD_ALLOCATION_NONE;Quoted 3 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-0+ payload->payload_allocation_status = +DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; }--
Regards,
Wayne
RE: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
[AMD Official Use Only - General]
Quoted 94 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Tuesday, August 8, 2023 12:00 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 37 lines · Wayne
[AMD Official Use Only - General]
Quoted 35 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 4, 2023 11:32 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 7 lines · Wayne Lin
[...]
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0@@ -3382,8 +3382,7 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 9 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-2* drm_dp_remove_payload_part2() - Remove an MST payload locally * @mgr: Manager to use. * @mst_state: The MST atomic state -* @old_payload: The payload with its old state -* @new_payload: The payload with its latest state +* @payload: The payload with its latest state * * Updates the starting time slots of all other payloads which would have
been shifted towards
Quoted 2 lines
* the start of the payload ID table as a result of removing a
payload. Driver should call this @@ -3392,25 +3391,36 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 2 lines
*/
void drm_dp_remove_payload_part2(struct
drm_dp_mst_topology_mgrQuoted 32 lines
Quoted 30 lines
*mgr,
Quoted 1 line
struct
drm_dp_mst_topology_state
*mst_state,Quoted 1 line
- const struct
drm_dp_mst_atomic_payload
*old_payload,Quoted 1 line
- struct
drm_dp_mst_atomic_payload
*new_payload)Quoted 1 line
+ struct
drm_dp_mst_atomic_payload*payload)
Quoted 5 lines
{ struct drm_dp_mst_atomic_payload *pos;drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-0+u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; +drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +10
-0+/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - +payload->vc_start_slot;Imo, the intuitive way would be to pass the old payload state to this function - which already contains the required
time_slotsparam
- and refactor things instead movingvc_start_slotfrom the payload state to mgr suggested by Ville earlier.Quoted 1 line
--Imre
Quoted 7 lines
Hi Imre,
Thanks for your feedback!I understand it's functionally correct. But IMHO, it's still a bit conceptually different between the time slot in old state and the time slot in current payload table. My thought is the time slot at the moment when we are removing the payload would be a better choice.
Yes, they are different. The old state contains the time slot the payload was added with in a preceding commit and so the time slot value which should be used when removing the same payload in the current commit.
Quoted 7 lines
The new state contains a time slot value with which the payload will be added in the current commit and can be different than the one in the old state if the current commit has changed the payload size (meaning that the same atomic commit will first remove the payload using the time slot value in the old state and afterwards will add back the same payload using the time slot value in the new state).
Appreciate your time, Imre!
Yes I understand, so I'm not using the number of the time slot in the new state. I'm referring to the start slot instead which is updated during every allocation and removement at current commit.
Like what you said, current commit manipulation could be a mix of allocations and removements for the payload. My thought is, conceptually, looking up the latest number of time slot is a better choice rather than the one in old state. It's relatively intuitive to me since we are removing payload from current payload table and which changes since last preceding commit till the moment we're deleting the payload. Although it's unreasonable that these values are different.
Quoted 5 lines
Quoted 4 lines
And with this, we can also simplify some codes. Especially remove workaround in amd driver. In fact, DRM mst code maintains the payload table and all the time slot info is in it already. We don't really have to pass a new parameter.
Quoted 7 lines
I agree that
drm_dp_remove_payload()could be simplified, but this should be done so that the drivers can pass the old payload state to it (without having to pass the new state). This would be possible ifvc_start_slotwas not tracked in the payload state (which is really not an atomic state that can be precomputed as all other atomic state), rather it would be tracked in structdrm_dp_mst_topology_mgr.
So the reason I chose to pass the new state is like what I mentioned above. I would prefer to carry the latest updated payload table instead which is in the new state. And I agree with the explanation for the vc_start_slot and that's also my thought at the beginning. It could be a refactor later, but no matter the start slot is put into payload state or the topology manager I would prefer to refer to the latest payload table rather than the number of time slot in the old state.
Quoted 8 lines
It looks like AMD has to reconstruct the old state in
dm_helpers_construct_old_payload(). Could you explain why it couldn't instead just passold_payloadacquired by
old_mst_state=drm_atomic_get_old_mst_topology_state();old_payload=drm_atomic_get_mst_payload_state(old_mst_state);?
AMD doesn't pass the drm old state to the stage while HW is deleting the payload. The reason is that HW payload table is known during HW programming procedure, so the payload removement is based on the table at the moment. AMD expected the current number of time slot is also already maintained in drm layer.
Again, thanks for your feedback Imre!
Quoted 23 lines
Quoted 23 lines
Quoted 23 lines
Quoted 23 lines
/* Remove local payload allocation */ list_for_each_entry(pos, &mst_state->payloads, next) {- if (pos !=
new_payload&&pos->vc_start_slot> new_payload-vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-1- pos->vc_start_slot -= old_payload->time_slots; + if (pos != payload && pos->vc_start_slot > payload-
vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +5
-5+ pos->vc_start_slot -= time_slots_to_remove; } - new_payload->vc_start_slot =-1 ; + payload->vc_start_slot =-1 ; mgr->payload_count--; - mgr->next_start_slot -= old_payload->time_slots; + mgr->next_start_slot -= time_slots_to_remove; -if (new_payload->delete) -drm_dp_mst_put_port_malloc (new_payload->port); +if (payload->delete) +drm_dp_mst_put_port_malloc (payload->port); - new_payload->payload_allocation_status =
Quoted 8 lines
Quoted 8 lines
Quoted 4 lines
DRM_DP_MST_PAYLOAD_ALLOCATION_NONE;Quoted 3 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-0+ payload->payload_allocation_status = +DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; }--
Regards,
Wayne
--
Regards,
Wayne
Re: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
Quoted 146 lines · Wayne
[AMD Official Use Only - General]
Quoted 94 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Tuesday, August 8, 2023 12:00 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 37 lines · Wayne
[AMD Official Use Only - General]
Quoted 35 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 4, 2023 11:32 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 7 lines · Wayne Lin
[...]
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0@@ -3382,8 +3382,7 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 9 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-2* drm_dp_remove_payload_part2() - Remove an MST payload locally * @mgr: Manager to use. * @mst_state: The MST atomic state -* @old_payload: The payload with its old state -* @new_payload: The payload with its latest state +* @payload: The payload with its latest state * * Updates the starting time slots of all other payloads which would have
been shifted towards
Quoted 2 lines
* the start of the payload ID table as a result of removing a
payload. Driver should call this @@ -3392,25 +3391,36 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 2 lines
*/
void drm_dp_remove_payload_part2(struct
drm_dp_mst_topology_mgrQuoted 32 lines
Quoted 30 lines
*mgr,
Quoted 1 line
struct
drm_dp_mst_topology_state
*mst_state,Quoted 1 line
- const struct
drm_dp_mst_atomic_payload
*old_payload,Quoted 1 line
- struct
drm_dp_mst_atomic_payload
*new_payload)Quoted 1 line
+ struct
drm_dp_mst_atomic_payload*payload)
Quoted 5 lines
{ struct drm_dp_mst_atomic_payload *pos;drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-0+u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; +drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +10
-0+/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - +payload->vc_start_slot;Imo, the intuitive way would be to pass the old payload state to this function - which already contains the required
time_slotsparam
- and refactor things instead movingvc_start_slotfrom the payload state to mgr suggested by Ville earlier.Quoted 1 line
--Imre
Quoted 7 lines
Hi Imre,
Thanks for your feedback!I understand it's functionally correct. But IMHO, it's still a bit conceptually different between the time slot in old state and the time slot in current payload table. My thought is the time slot at the moment when we are removing the payload would be a better choice.
Yes, they are different. The old state contains the time slot the payload was added with in a preceding commit and so the time slot value which should be used when removing the same payload in the current commit.
Quoted 7 lines
The new state contains a time slot value with which the payload will be added in the current commit and can be different than the one in the old state if the current commit has changed the payload size (meaning that the same atomic commit will first remove the payload using the time slot value in the old state and afterwards will add back the same payload using the time slot value in the new state).
Appreciate your time, Imre!
Yes I understand, so I'm not using the number of the time slot in the new state. I'm referring to the start slot instead which is updated during every allocation and removement at current commit.
Like what you said, current commit manipulation could be a mix of allocations and removements for the payload. My thought is, conceptually, looking up the latest number of time slot is a better choice rather than the one in old state. It's relatively intuitive to me since we are removing payload from current payload table and which changes since last preceding commit till the moment we're deleting the payload. Although it's unreasonable that these values are different.
Quoted 5 lines
Quoted 4 lines
And with this, we can also simplify some codes. Especially remove workaround in amd driver. In fact, DRM mst code maintains the payload table and all the time slot info is in it already. We don't really have to pass a new parameter.
Quoted 7 lines
I agree that
drm_dp_remove_payload()could be simplified, but this should be done so that the drivers can pass the old payload state to it (without having to pass the new state). This would be possible ifvc_start_slotwas not tracked in the payload state (which is really not an atomic state that can be precomputed as all other atomic state), rather it would be tracked in structdrm_dp_mst_topology_mgr.So the reason I chose to pass the new state is like what I mentioned above. I would prefer to carry the latest updated payload table instead which is in the new state. And I agree with the explanation for the
vc_start_slotand that's also my thought at the beginning. It could be a refactor later, but no matter the start slot is put into payload state or the topology manager I would prefer to refer to the latest payload table rather than the number of time slot in the old state.Quoted 8 lines
It looks like AMD has to reconstruct the old state in
dm_helpers_construct_old_payload(). Could you explain why it couldn't instead just passold_payloadacquired by
old_mst_state=drm_atomic_get_old_mst_topology_state();old_payload=drm_atomic_get_mst_payload_state(old_mst_state);?
Quoted 7 lines
AMD doesn't pass the drm old state to the stage while HW is deleting the payload. The reason is that HW payload table is known during HW programming procedure, so the payload removement is based on the table at the moment.
AMD expected the current number of time slot is also already maintained in drm layer.
Yes, both of the above are maintained by the drm layer, but it also means it doesn't really need to recalculate time_slots_to_remove as done in this patch, since that info is already available in the old payload state.
Afaics the AMD driver calls properly
drm_atomic_helper_commit() -> drm_atomic_helper_swap_state()
after a commit, so that all the payloads it added should be tracked now as the old payload state.
So could you confirm what is the old_payload->time_slots value (which you get with the above functions) at the point of removing this payload and if it's not the time_slots value this same payload was actually added with previously, why this is so (via some example sequence)?
Thanks.
Quoted 37 lines
Again, thanks for your feedback Imre!
Quoted 23 lines
Quoted 23 lines
Quoted 23 lines
Quoted 23 lines
/* Remove local payload allocation */ list_for_each_entry(pos, &mst_state->payloads, next) {- if (pos !=
new_payload&&pos->vc_start_slot> new_payload-vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-1- pos->vc_start_slot -= old_payload->time_slots; + if (pos != payload && pos->vc_start_slot > payload-
vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +5
-5+ pos->vc_start_slot -= time_slots_to_remove; } - new_payload->vc_start_slot =-1 ; + payload->vc_start_slot =-1 ; mgr->payload_count--; - mgr->next_start_slot -= old_payload->time_slots; + mgr->next_start_slot -= time_slots_to_remove; -if (new_payload->delete) -drm_dp_mst_put_port_malloc (new_payload->port); +if (payload->delete) +drm_dp_mst_put_port_malloc (payload->port); - new_payload->payload_allocation_status =Quoted 8 lines
Quoted 8 lines
Quoted 4 lines
DRM_DP_MST_PAYLOAD_ALLOCATION_NONE;Quoted 3 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-0+ payload->payload_allocation_status = +DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; }--
Regards,
Wayne--
Regards,
Wayne
RE: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
[AMD Official Use Only - General]
Quoted 184 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Saturday, August 19, 2023 1:46 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 37 lines · Wayne
[AMD Official Use Only - General]
Quoted 35 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Tuesday, August 8, 2023 12:00 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 23 lines · Wayne
[AMD Official Use Only - General]
Quoted 21 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 4, 2023 11:32 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 7 lines · Wayne Lin
[...]
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0@@ -3382,8 +3382,7 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 1 line
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0* drm_dp_remove_payload_part2() - Remove an MST payload locally
Quoted 48 lines
Quoted 48 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-2* @mgr: Manager to use. * @mst_state: The MST atomic state -* @old_payload: The payload with its old state -* @new_payload: The payload with its latest state +* @payload: The payload with its latest state * * Updates the starting time slots of all other payloads which would have
Quoted 6 lines
Quoted 6 lines
been shifted towards
Quoted 2 lines
* the start of the payload ID table as a result of removing
a payload. Driver should call this @@ -3392,25 +3391,36 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 2 lines
*/
void drm_dp_remove_payload_part2(struct
drm_dp_mst_topology_mgrQuoted 33 lines
Quoted 32 lines
*mgr,
Quoted 1 line
struct
drm_dp_mst_topology_state
*mst_state,Quoted 1 line
- const struct
drm_dp_mst_atomic_payload
*old_payload,Quoted 1 line
- struct
drm_dp_mst_atomic_payload
*new_payload)Quoted 1 line
+ struct
drm_dp_mst_atomic_payload*payload)
Quoted 5 lines
{ struct drm_dp_mst_atomic_payload *pos;drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-0+u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; +drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +10
-0+/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - +payload->vc_start_slot;Imo, the intuitive way would be to pass the old payload state to this function - which already contains the required
time_slotsparam
- and refactor things instead movingvc_start_slotfrom the payload state to mgr suggested by Ville earlier.--Imre
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Hi Imre,
Thanks for your feedback!I understand it's functionally correct. But IMHO, it's still a bit conceptually different between the time slot in old state and the time slot in current payload table. My thought is the time slot at the moment when we are removing the payload would be a better
choice.
Quoted 4 lines
Quoted 4 lines
Yes, they are different. The old state contains the time slot the payload was added with in a preceding commit and so the time slot value which should be used when removing the same payload in the
current commit.
Quoted 11 lines
Quoted 8 lines
The new state contains a time slot value with which the payload will be added in the current commit and can be different than the one in the old state if the current commit has changed the payload size (meaning that the same atomic commit will first remove the payload using the time slot value in the old state and afterwards will add back the same payload using the time slot value in the new state).
Appreciate your time, Imre!
Yes I understand, so I'm not using the number of the time slot in the new
state.
Quoted 6 lines
I'm referring to the start slot instead which is updated during every allocation and removement at current commit.
Like what you said, current commit manipulation could be a mix of allocations and removements for the payload. My thought is, conceptually, looking up the latest number of time slot is a better choice
rather than the one in old state.
Quoted 18 lines
It's relatively intuitive to me since we are removing payload from current payload table and which changes since last preceding commit till the moment we're deleting the payload. Although it's unreasonable that these values are different.
Quoted 5 lines
Quoted 4 lines
And with this, we can also simplify some codes. Especially remove workaround in amd driver. In fact, DRM mst code maintains the payload table and all the time slot info is in it already. We don't really have to pass a new parameter.
Quoted 7 lines
I agree that
drm_dp_remove_payload()could be simplified, but this should be done so that the drivers can pass the old payload state to it (without having to pass the new state). This would be possible ifvc_start_slotwas not tracked in the payload state (which is really not an atomic state that can be precomputed as all other atomic state), rather it would be tracked in structdrm_dp_mst_topology_mgr.Quoted 6 lines
So the reason I chose to pass the new state is like what I mentioned above. I would prefer to carry the latest updated payload table instead which is in the new state. And I agree with the explanation for the
vc_start_slotand that's also my thought at the beginning. It could be a refactor later, but no matter the start slot is put into payload state or the topology manager I would prefer to refer to the latestpayload table rather than the number of time slot in the old state.
Quoted 10 lines
Quoted 9 lines
It looks like AMD has to reconstruct the old state in
dm_helpers_construct_old_payload(). Could you explain why it couldn't instead just passold_payloadacquired by
old_mst_state=drm_atomic_get_old_mst_topology_state();old_payload=drm_atomic_get_mst_payload_state(old_mst_state);?
Quoted 7 lines
AMD doesn't pass the drm old state to the stage while HW is deleting the payload. The reason is that HW payload table is known during HW programming procedure, so the payload removement is based on the table at the moment.
AMD expected the current number of time slot is also already maintained in drm layer.
Yes, both of the above are maintained by the drm layer, but it also means it doesn't really need to recalculate
time_slots_to_removeas done in this patch, since that info is already available in the old payload state.Afaics the AMD driver calls properly
drm_atomic_helper_commit()->drm_atomic_helper_swap_state()after a commit, so that all the payloads it added should be tracked now as the old payload state.
Quoted 6 lines
So could you confirm what is the
old_payload->time_slotsvalue (which you get with the above functions) at the point of removing this payload and if it's not thetime_slotsvalue this same payload was actually added with previously, why this is so (via some example sequence)?Thanks.
Hi Imre,
I'm not saying that the time_slots carried in the old state is wrong within amd driver. But just amd driver doesn't carry the drm state to the step when it's removing the payload, since the info is already in its hardware and drm used to maintain the info in the drm layer. So the patch is trying to get the behavior of this helper function back to what it used to be.
And the main reason that I want to change the pass in parameter is like what I mentioned previously. The commit manipulation could be a mix of allocations and removements for the payload. And in the spec, it also introduces examples to reduce or increase the payload allocation. Although we're not using reduction/increment today, it implicitly imposes the limitation to use them before calling the removement helper function with the old state as the passed in parameter. So I also want to remove the dependency by this patch. Would like to know your thoughts about this.
Thanks, Imre!
Quoted 33 lines
Quoted 33 lines
Again, thanks for your feedback Imre!
Quoted 30 lines
Quoted 30 lines
Quoted 29 lines
Quoted 15 lines
/* Remove local payload allocation */ list_for_each_entry(pos, &mst_state->payloads, next) {- if (pos !=
new_payload&&pos->vc_start_slot> new_payload-vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-1- pos->vc_start_slot -= old_payload->time_slots; +if (pos != payload && pos->vc_start_slot > + payload-
vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-1+ pos->vc_start_slot -= + time_slots_to_remove; } - new_payload->vc_start_slot =-1 ; + payload->vc_start_slot =-1 ;drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-4mgr->payload_count--; - mgr->next_start_slot -= old_payload->time_slots; + mgr->next_start_slot -= time_slots_to_remove; -if (new_payload->delete) -drm_dp_mst_put_port_malloc (new_payload->port); +if (payload->delete) +drm_dp_mst_put_port_malloc (payload->port); - new_payload->payload_allocation_status =
DRM_DP_MST_PAYLOAD_ALLOCATION_NONE;Quoted 3 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-0+ payload->payload_allocation_status = +DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; }
Quoted 7 lines
Quoted 7 lines
Quoted 3 lines
Quoted 3 lines
--
Regards,
Wayne--
Regards,
Wayne
--
Regards,
Wayne
Re: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
Quoted 191 lines · Wayne
[AMD Official Use Only - General]
Quoted 189 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Saturday, August 19, 2023 1:46 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 37 lines · Wayne
[AMD Official Use Only - General]
Quoted 35 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Tuesday, August 8, 2023 12:00 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 23 lines · Wayne
[AMD Official Use Only - General]
Quoted 21 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 4, 2023 11:32 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 7 lines · Wayne Lin
[...]
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0@@ -3382,8 +3382,7 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 1 line
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0* drm_dp_remove_payload_part2() - Remove an MST payload locally
Quoted 48 lines
Quoted 48 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-2* @mgr: Manager to use. * @mst_state: The MST atomic state -* @old_payload: The payload with its old state -* @new_payload: The payload with its latest state +* @payload: The payload with its latest state * * Updates the starting time slots of all other payloads which would have
Quoted 6 lines
Quoted 6 lines
been shifted towards
Quoted 2 lines
* the start of the payload ID table as a result of removing
a payload. Driver should call this @@ -3392,25 +3391,36 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 2 lines
*/
void drm_dp_remove_payload_part2(struct
drm_dp_mst_topology_mgrQuoted 33 lines
Quoted 32 lines
*mgr,
Quoted 1 line
struct
drm_dp_mst_topology_state
*mst_state,Quoted 1 line
- const struct
drm_dp_mst_atomic_payload
*old_payload,Quoted 1 line
- struct
drm_dp_mst_atomic_payload
*new_payload)Quoted 1 line
+ struct
drm_dp_mst_atomic_payload*payload)
Quoted 5 lines
{ struct drm_dp_mst_atomic_payload *pos;drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-0+u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; +drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +10
-0+/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - +payload->vc_start_slot;Imo, the intuitive way would be to pass the old payload state to this function - which already contains the required
time_slotsparam
- and refactor things instead movingvc_start_slotfrom the payload state to mgr suggested by Ville earlier.--Imre
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Hi Imre,
Thanks for your feedback!I understand it's functionally correct. But IMHO, it's still a bit conceptually different between the time slot in old state and the time slot in current payload table. My thought is the time slot at the moment when we are removing the payload would be a better
choice.
Quoted 4 lines
Quoted 4 lines
Yes, they are different. The old state contains the time slot the payload was added with in a preceding commit and so the time slot value which should be used when removing the same payload in the
current commit.
Quoted 11 lines
Quoted 8 lines
The new state contains a time slot value with which the payload will be added in the current commit and can be different than the one in the old state if the current commit has changed the payload size (meaning that the same atomic commit will first remove the payload using the time slot value in the old state and afterwards will add back the same payload using the time slot value in the new state).
Appreciate your time, Imre!
Yes I understand, so I'm not using the number of the time slot in the new
state.
Quoted 6 lines
I'm referring to the start slot instead which is updated during every allocation and removement at current commit.
Like what you said, current commit manipulation could be a mix of allocations and removements for the payload. My thought is, conceptually, looking up the latest number of time slot is a better choice
rather than the one in old state.
Quoted 18 lines
It's relatively intuitive to me since we are removing payload from current payload table and which changes since last preceding commit till the moment we're deleting the payload. Although it's unreasonable that these values are different.
Quoted 5 lines
Quoted 4 lines
And with this, we can also simplify some codes. Especially remove workaround in amd driver. In fact, DRM mst code maintains the payload table and all the time slot info is in it already. We don't really have to pass a new parameter.
Quoted 7 lines
I agree that
drm_dp_remove_payload()could be simplified, but this should be done so that the drivers can pass the old payload state to it (without having to pass the new state). This would be possible ifvc_start_slotwas not tracked in the payload state (which is really not an atomic state that can be precomputed as all other atomic state), rather it would be tracked in structdrm_dp_mst_topology_mgr.Quoted 6 lines
So the reason I chose to pass the new state is like what I mentioned above. I would prefer to carry the latest updated payload table instead which is in the new state. And I agree with the explanation for the
vc_start_slotand that's also my thought at the beginning. It could be a refactor later, but no matter the start slot is put into payload state or the topology manager I would prefer to refer to the latestpayload table rather than the number of time slot in the old state.
Quoted 10 lines
Quoted 9 lines
It looks like AMD has to reconstruct the old state in
dm_helpers_construct_old_payload(). Could you explain why it couldn't instead just passold_payloadacquired by
old_mst_state=drm_atomic_get_old_mst_topology_state();old_payload=drm_atomic_get_mst_payload_state(old_mst_state);?
Quoted 7 lines
AMD doesn't pass the drm old state to the stage while HW is deleting the payload. The reason is that HW payload table is known during HW programming procedure, so the payload removement is based on the table at the moment.
AMD expected the current number of time slot is also already maintained in drm layer.
Yes, both of the above are maintained by the drm layer, but it also means it doesn't really need to recalculate
time_slots_to_removeas done in this patch, since that info is already available in the old payload state.Afaics the AMD driver calls properly
drm_atomic_helper_commit()->drm_atomic_helper_swap_state()after a commit, so that all the payloads it added should be tracked now as the old payload state.
So could you confirm what is the
old_payload->time_slotsvalue (which you get with the above functions) at the point of removing this payload and if it's not thetime_slotsvalue this same payload was actually added with previously, why this is so (via some example sequence)?
Quoted 7 lines
Quoted 1 line
Thanks.
Hi Imre,
I'm not saying that thetime_slotscarried in the old state is wrong within amd driver. But just amd driver doesn't carry the drm state to the step when it's removing the payload, since the info is already in its hardware and drm used to maintain the info in the drm layer.
Hm, in
dm_helpers_dp_mst_write_payload_allocation_table()
the
mst_state = to_drm_dp_mst_topology_state(mst_mgr->base.state);
used as the new state doesn't look ok to me. The above is the MST object's current SW state after an atomic commit/swap, but this isn't the new state a driver should program to the HW or pass to the drm helper functions. The object's current state could be ahead of what the driver should program to the HW, if the driver properly implements commit pipelining (so at the point you're programming a state you'd have already future commits computed and queued up, each queued-up commit with its own state).
So instead of the above mst_state the driver should pass down the drm_atomic_state to dm_helpers_dp_mst_write_payload_allocation_table() which should use that instead to get the MST state it should program or pass to the drm layer.
Quoted 2 lines
So the patch is trying to get the behavior of this helper function back to what it used to be.
The behavior before was actually broken and one reason for that was a confusion about what's stored in the new payload state. It's a mixture of old/current state (vc_start_slot) and new state (time_slots). So I don't think it's a good idea to add even more dependency on this semantic.
I think the right solution for what this patch is about - remove the need for dm_helpers_construct_old_payload(), would be to pass down drm_atomic_state to dm_helpers_dp_mst_write_payload_allocation_table()
based on the above.
Quoted 9 lines
And the main reason that I want to change the pass in parameter is like what I mentioned previously. The commit manipulation could be a mix of allocations and removements for the payload. And in the spec, it also introduces examples to reduce or increase the payload allocation. Although we're not using reduction/increment today, it implicitly imposes the limitation to use them before calling the removement helper function with the old state as the passed in parameter. So I also want to remove the dependency by this patch. Would like to know your thoughts about this.
It's not clear what would be the benefit of changing a payload without removing it first. In any case, I think the right direction for the driver would be to meet the requirement to use the proper atomic state with the drm helper functions (not the object's current SW state) as described above.
Quoted 45 lines
Thanks, Imre!
Quoted 33 lines
Quoted 33 lines
Again, thanks for your feedback Imre!
Quoted 30 lines
Quoted 30 lines
Quoted 29 lines
Quoted 15 lines
/* Remove local payload allocation */ list_for_each_entry(pos, &mst_state->payloads, next) {- if (pos !=
new_payload&&pos->vc_start_slot> new_payload-vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-1- pos->vc_start_slot -= old_payload->time_slots; +if (pos != payload && pos->vc_start_slot > + payload-
vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-1+ pos->vc_start_slot -= + time_slots_to_remove; } - new_payload->vc_start_slot =-1 ; + payload->vc_start_slot =-1 ;drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-4mgr->payload_count--; - mgr->next_start_slot -= old_payload->time_slots; + mgr->next_start_slot -= time_slots_to_remove; -if (new_payload->delete) -drm_dp_mst_put_port_malloc (new_payload->port); +if (payload->delete) +drm_dp_mst_put_port_malloc (payload->port); - new_payload->payload_allocation_status =
DRM_DP_MST_PAYLOAD_ALLOCATION_NONE;Quoted 3 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-0+ payload->payload_allocation_status = +DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; }Quoted 7 lines
Quoted 7 lines
Quoted 3 lines
Quoted 3 lines
--
Regards,
Wayne--
Regards,
Wayne--
Regards,
Wayne
RE: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
[AMD Official Use Only - General]
Quoted 270 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 25, 2023 9:56 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 126 lines · Wayne
[AMD Official Use Only - General]
Quoted 124 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Saturday, August 19, 2023 1:46 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 38 lines · Wayne
[AMD Official Use Only - General]
Quoted 36 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Tuesday, August 8, 2023 12:00 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 24 lines · Wayne
[AMD Official Use Only - General]
Quoted 22 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 4, 2023 11:32 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 7 lines · Wayne Lin
[...]
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0@@ -3382,8 +3382,7 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 1 line
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0* drm_dp_remove_payload_part2() - Remove an MST payload locally
Quoted 51 lines
Quoted 51 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-2* @mgr: Manager to use. * @mst_state: The MST atomic state -* @old_payload: The payload with its old state -* @new_payload: The payload with its latest state +* @payload: The payload with its latest state * * Updates the starting time slots of all other payloads which would have
Quoted 7 lines
Quoted 7 lines
been shifted towards
Quoted 3 lines
* the start of the payload ID table as a result of
removing a payload. Driver should call this @@ -3392,25 +3391,36 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 2 lines
*/
void drm_dp_remove_payload_part2(struct
drm_dp_mst_topology_mgrQuoted 35 lines
Quoted 34 lines
*mgr,
Quoted 2 lines
struct
drm_dp_mst_topology_state
*mst_state,Quoted 1 line
- const struct
drm_dp_mst_atomic_payload
*old_payload,Quoted 1 line
- struct
drm_dp_mst_atomic_payload
*new_payload)Quoted 2 lines
+ struct
+drm_dp_mst_atomic_payload*payload)
Quoted 5 lines
{ struct drm_dp_mst_atomic_payload *pos;drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-0+u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; +drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +10
-0+/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - +payload->vc_start_slot;Imo, the intuitive way would be to pass the old payload state to this function - which already contains the required
time_slotsparam
- and refactor things instead movingvc_start_slotfrom the payload state to mgr suggested by Ville earlier.--Imre
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Hi Imre,
Thanks for your feedback!I understand it's functionally correct. But IMHO, it's still a bit conceptually different between the time slot in old state and the time slot in current payload table. My thought is the time slot at the moment when we are removing the payload would be a better
choice.
Quoted 5 lines
Quoted 5 lines
Yes, they are different. The old state contains the time slot the payload was added with in a preceding commit and so the time slot value which should be used when removing the same payload in the
current commit.
Quoted 7 lines
Quoted 7 lines
The new state contains a time slot value with which the payload will be added in the current commit and can be different than the one in the old state if the current commit has changed the payload size (meaning that the same atomic commit will first remove the payload using the time slot value in the old state and afterwards will add back the same payload using the time slot value
in the new state).
Quoted 24 lines
Quoted 24 lines
Quoted 5 lines
Quoted 1 line
Appreciate your time, Imre!
Yes I understand, so I'm not using the number of the time slot in the new
state.
Quoted 7 lines
I'm referring to the start slot instead which is updated during every allocation and removement at current commit.
Like what you said, current commit manipulation could be a mix of allocations and removements for the payload. My thought is, conceptually, looking up the latest number of time slot is a better choice
rather than the one in old state.
Quoted 10 lines
It's relatively intuitive to me since we are removing payload from current payload table and which changes since last preceding commit till the moment we're deleting the payload. Although it's unreasonable that these values are different.
Quoted 5 lines
Quoted 4 lines
And with this, we can also simplify some codes. Especially remove workaround in amd driver. In fact, DRM mst code maintains the payload table and all the time slot info is in it already. We don't really have to pass a new parameter.
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
I agree that
drm_dp_remove_payload()could be simplified, but this should be done so that the drivers can pass the old payload state to it (without having to pass the new state). This would be possible ifvc_start_slotwas not tracked in the payload state (which is really not an atomic state that can be precomputed as all other atomic state), rather it would be tracked instruct
drm_dp_mst_topology_mgr.Quoted 25 lines
Quoted 25 lines
Quoted 9 lines
Quoted 1 line
So the reason I chose to pass the new state is like what I mentioned above. I would prefer to carry the latest updated payload table instead which is in the new state. And I agree with the explanation for the
vc_start_slotand that's also my thought at the beginning. It could be a refactor later, but no matter the start slot is put into payload state or the topology manager I would prefer to refer to the latestpayload table rather than the number of time slot in the old state.
Quoted 15 lines
Quoted 9 lines
It looks like AMD has to reconstruct the old state in
dm_helpers_construct_old_payload(). Could you explain why it couldn't instead just passold_payloadacquired by
old_mst_state=drm_atomic_get_old_mst_topology_state();old_payload=drm_atomic_get_mst_payload_state(old_mst_state);?
AMD doesn't pass the drm old state to the stage while HW is deleting the payload. The reason is that HW payload table is known during HW programming procedure, so the payload removement is based on the table at the moment.
Quoted 6 lines
Quoted 6 lines
Quoted 2 lines
AMD expected the current number of time slot is also already maintained in drm layer.
Yes, both of the above are maintained by the drm layer, but it also means it doesn't really need to recalculate
time_slots_to_removeas done in this patch, since that info is already available in the old payloadstate.
Quoted 5 lines
Quoted 5 lines
Afaics the AMD driver calls properly
drm_atomic_helper_commit()->drm_atomic_helper_swap_state()Quoted 7 lines
Quoted 7 lines
after a commit, so that all the payloads it added should be tracked now as the old payload state.
So could you confirm what is the
old_payload->time_slotsvalue (which you get with the above functions) at the point of removing this payload and if it's not thetime_slotsvalue this same payload was actually added with previously, why this is so (via some examplesequence)?
Quoted 5 lines
Quoted 2 lines
Thanks.
Hi Imre,
I'm not saying that thetime_slotscarried in the old state is wrong withinamd driver.
Quoted 3 lines
But just amd driver doesn't carry the drm state to the step when it's removing the payload, since the info is already in its hardware and drm used to maintain the info in the drm layer.
Hm, in
dm_helpers_dp_mst_write_payload_allocation_table()the
mst_state=to_drm_dp_mst_topology_state(mst_mgr->base.state);used as the new state doesn't look ok to me. The above is the MST object's current SW state after an atomic commit/swap, but this isn't the new state a driver should program to the HW or pass to the drm helper functions. The object's current state could be ahead of what the driver should program to the HW, if the driver properly implements commit pipelining (so at the point you're programming a state you'd have already future commits computed and queued up, each queued-up commit with its own state).
So instead of the above
mst_statethe driver should pass down thedrm_atomic_statetodm_helpers_dp_mst_write_payload_allocation_table()which should use that instead to get the MST state it should program or pass to the drm layer.Quoted 2 lines
So the patch is trying to get the behavior of this helper function back to what it used to be.
The behavior before was actually broken and one reason for that was a confusion about what's stored in the new payload state. It's a mixture of old/current state (
vc_start_slot) and new state (time_slots). So I don't think it's a good idea to add even more dependency on this semantic.I think the right solution for what this patch is about - remove the need for
dm_helpers_construct_old_payload(), would be to pass downdrm_atomic_statetodm_helpers_dp_mst_write_payload_allocation_table()
based on the above.Quoted 9 lines
And the main reason that I want to change the pass in parameter is like what I mentioned previously. The commit manipulation could be a mix of allocations and removements for the payload. And in the spec, it also introduces examples to reduce or increase the payload allocation. Although we're not using reduction/increment today, it implicitly imposes the limitation to use them before calling the removement helper function with the old state as the passed in parameter. So I also want to remove the dependency by this patch. Would like to know your thoughts about this.
Quoted 4 lines
It's not clear what would be the benefit of changing a payload without removing it first. In any case, I think the right direction for the driver would be to meet the requirement to use the proper atomic state with the drm helper functions (not the object's current SW state) as described above.
Hi Imre,
Thanks for pointing that out and that was also one of my plan to refactor! But I would take that as another patch to follow up and would like to align with you the idea for this helper function itself more.
My concern is referring to the old state to remove the time slot is just not generic right for what the helper function shall accomplish. Even it doesn't bring benefit as you see, having old state as the input imposes limitation to drivers using it which is the downside that it brought. The limitation is like what I tried to explain before. For an instance, if one driver has compressed streams allocated in the previous commit and wants to disable these streams at the current commit. It handles and accomplishes the commit into two steps that firstly to disable dsc engine only (which increases the time slot), and next disable the streams. Under this design, the old state can't represent the exact number of time slot that it want's to remove at the moment. For this case, I don't think drm_dp_remove_payload_part2 should block the driver to use it since the driver doesn't do things wrong. Conversely, it's due to the helper function constrains the driver to use an inappropriate input.
And with or without my patch, the current payload allocation table (i.e. vc_start_slot In new state) and the time slot that this new state is going to set (i.e. time_slots in new state) are already both in the new state. I think the patch doesn't add irrelevant input for removing a payload as this helper function should do, because the time slot it should remove is the exact one in the payload table now, not the one captured in previous commit. So in contrast, refer to old state time slot is a bit confusing to me because it's not generic right. Would like to know your thought on this point. Appreciate your time!
Quoted 4 lines
Quoted 4 lines
Thanks, Imre!
Quoted 1 line
Quoted 43 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Again, thanks for your feedback Imre!
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
/* Remove local payload allocation */ list_for_each_entry(pos, &mst_state->payloads, next) {- if (pos !=
new_payload&&pos->vc_start_slot>new_payload-
Quoted 36 lines
Quoted 26 lines
Quoted 26 lines
Quoted 26 lines
Quoted 26 lines
Quoted 25 lines
Quoted 11 lines
vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-1- pos->vc_start_slot -= old_payload->time_slots; +if (pos != payload && pos->vc_start_slot > + payload-
vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-1+ pos->vc_start_slot -= + time_slots_to_remove; } - new_payload->vc_start_slot =-1 ; + payload->vc_start_slot =-1 ;drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-4mgr->payload_count--; - mgr->next_start_slot -= old_payload->time_slots; + mgr->next_start_slot -= time_slots_to_remove; -if (new_payload->delete) -drm_dp_mst_put_port_malloc (new_payload->port); +if (payload->delete) +drm_dp_mst_put_port_malloc (payload->port); - new_payload->payload_allocation_status =
DRM_DP_MST_PAYLOAD_ALLOCATION_NONE;Quoted 3 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-0+ payload->payload_allocation_status = +DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; }Quoted 7 lines
Quoted 7 lines
Quoted 3 lines
Quoted 3 lines
--
Regards,
Wayne--
Regards,
Wayne--
Regards,
Wayne
--
Regards,
Wayne
Re: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
Quoted 296 lines · Wayne
[AMD Official Use Only - General]
Quoted 270 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 25, 2023 9:56 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 126 lines · Wayne
[AMD Official Use Only - General]
Quoted 124 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Saturday, August 19, 2023 1:46 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 38 lines · Wayne
[AMD Official Use Only - General]
Quoted 36 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Tuesday, August 8, 2023 12:00 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 24 lines · Wayne
[AMD Official Use Only - General]
Quoted 22 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 4, 2023 11:32 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 7 lines · Wayne Lin
[...]
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0@@ -3382,8 +3382,7 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 1 line
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0* drm_dp_remove_payload_part2() - Remove an MST payload locally
Quoted 51 lines
Quoted 51 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-2* @mgr: Manager to use. * @mst_state: The MST atomic state -* @old_payload: The payload with its old state -* @new_payload: The payload with its latest state +* @payload: The payload with its latest state * * Updates the starting time slots of all other payloads which would have
Quoted 7 lines
Quoted 7 lines
been shifted towards
Quoted 3 lines
* the start of the payload ID table as a result of
removing a payload. Driver should call this @@ -3392,25 +3391,36 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 2 lines
*/
void drm_dp_remove_payload_part2(struct
drm_dp_mst_topology_mgrQuoted 35 lines
Quoted 34 lines
*mgr,
Quoted 2 lines
struct
drm_dp_mst_topology_state
*mst_state,Quoted 1 line
- const struct
drm_dp_mst_atomic_payload
*old_payload,Quoted 1 line
- struct
drm_dp_mst_atomic_payload
*new_payload)Quoted 2 lines
+ struct
+drm_dp_mst_atomic_payload*payload)
Quoted 5 lines
{ struct drm_dp_mst_atomic_payload *pos;drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-0+u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; +drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +10
-0+/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - +payload->vc_start_slot;Imo, the intuitive way would be to pass the old payload state to this function - which already contains the required
time_slotsparam
- and refactor things instead movingvc_start_slotfrom the payload state to mgr suggested by Ville earlier.--Imre
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Hi Imre,
Thanks for your feedback!I understand it's functionally correct. But IMHO, it's still a bit conceptually different between the time slot in old state and the time slot in current payload table. My thought is the time slot at the moment when we are removing the payload would be a better
choice.
Quoted 5 lines
Quoted 5 lines
Yes, they are different. The old state contains the time slot the payload was added with in a preceding commit and so the time slot value which should be used when removing the same payload in the
current commit.
Quoted 7 lines
Quoted 7 lines
The new state contains a time slot value with which the payload will be added in the current commit and can be different than the one in the old state if the current commit has changed the payload size (meaning that the same atomic commit will first remove the payload using the time slot value in the old state and afterwards will add back the same payload using the time slot value
in the new state).
Quoted 24 lines
Quoted 24 lines
Quoted 5 lines
Quoted 1 line
Appreciate your time, Imre!
Yes I understand, so I'm not using the number of the time slot in the new
state.
Quoted 7 lines
I'm referring to the start slot instead which is updated during every allocation and removement at current commit.
Like what you said, current commit manipulation could be a mix of allocations and removements for the payload. My thought is, conceptually, looking up the latest number of time slot is a better choice
rather than the one in old state.
Quoted 10 lines
It's relatively intuitive to me since we are removing payload from current payload table and which changes since last preceding commit till the moment we're deleting the payload. Although it's unreasonable that these values are different.
Quoted 5 lines
Quoted 4 lines
And with this, we can also simplify some codes. Especially remove workaround in amd driver. In fact, DRM mst code maintains the payload table and all the time slot info is in it already. We don't really have to pass a new parameter.
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
I agree that
drm_dp_remove_payload()could be simplified, but this should be done so that the drivers can pass the old payload state to it (without having to pass the new state). This would be possible ifvc_start_slotwas not tracked in the payload state (which is really not an atomic state that can be precomputed as all other atomic state), rather it would be tracked instruct
drm_dp_mst_topology_mgr.Quoted 25 lines
Quoted 25 lines
Quoted 9 lines
Quoted 1 line
So the reason I chose to pass the new state is like what I mentioned above. I would prefer to carry the latest updated payload table instead which is in the new state. And I agree with the explanation for the
vc_start_slotand that's also my thought at the beginning. It could be a refactor later, but no matter the start slot is put into payload state or the topology manager I would prefer to refer to the latestpayload table rather than the number of time slot in the old state.
Quoted 15 lines
Quoted 9 lines
It looks like AMD has to reconstruct the old state in
dm_helpers_construct_old_payload(). Could you explain why it couldn't instead just passold_payloadacquired by
old_mst_state=drm_atomic_get_old_mst_topology_state();old_payload=drm_atomic_get_mst_payload_state(old_mst_state);?
AMD doesn't pass the drm old state to the stage while HW is deleting the payload. The reason is that HW payload table is known during HW programming procedure, so the payload removement is based on the table at the moment.
Quoted 6 lines
Quoted 6 lines
Quoted 2 lines
AMD expected the current number of time slot is also already maintained in drm layer.
Yes, both of the above are maintained by the drm layer, but it also means it doesn't really need to recalculate
time_slots_to_removeas done in this patch, since that info is already available in the old payloadstate.
Quoted 5 lines
Quoted 5 lines
Afaics the AMD driver calls properly
drm_atomic_helper_commit()->drm_atomic_helper_swap_state()Quoted 7 lines
Quoted 7 lines
after a commit, so that all the payloads it added should be tracked now as the old payload state.
So could you confirm what is the
old_payload->time_slotsvalue (which you get with the above functions) at the point of removing this payload and if it's not thetime_slotsvalue this same payload was actually added with previously, why this is so (via some examplesequence)?
Quoted 5 lines
Quoted 2 lines
Thanks.
Hi Imre,
I'm not saying that thetime_slotscarried in the old state is wrong withinamd driver.
Quoted 3 lines
But just amd driver doesn't carry the drm state to the step when it's removing the payload, since the info is already in its hardware and drm used to maintain the info in the drm layer.
Hm, in
dm_helpers_dp_mst_write_payload_allocation_table()the
mst_state=to_drm_dp_mst_topology_state(mst_mgr->base.state);used as the new state doesn't look ok to me. The above is the MST object's current SW state after an atomic commit/swap, but this isn't the new state a driver should program to the HW or pass to the drm helper functions. The object's current state could be ahead of what the driver should program to the HW, if the driver properly implements commit pipelining (so at the point you're programming a state you'd have already future commits computed and queued up, each queued-up commit with its own state).
So instead of the above
mst_statethe driver should pass down thedrm_atomic_statetodm_helpers_dp_mst_write_payload_allocation_table()which should use that instead to get the MST state it should program or pass to the drm layer.Quoted 2 lines
So the patch is trying to get the behavior of this helper function back to what it used to be.
The behavior before was actually broken and one reason for that was a confusion about what's stored in the new payload state. It's a mixture of old/current state (
vc_start_slot) and new state (time_slots). So I don't think it's a good idea to add even more dependency on this semantic.I think the right solution for what this patch is about - remove the need for
dm_helpers_construct_old_payload(), would be to pass downdrm_atomic_statetodm_helpers_dp_mst_write_payload_allocation_table()
based on the above.Quoted 9 lines
And the main reason that I want to change the pass in parameter is like what I mentioned previously. The commit manipulation could be a mix of allocations and removements for the payload. And in the spec, it also introduces examples to reduce or increase the payload allocation. Although we're not using reduction/increment today, it implicitly imposes the limitation to use them before calling the removement helper function with the old state as the passed in parameter. So I also want to remove the dependency by this patch. Would like to know your thoughts about this.
Quoted 4 lines
It's not clear what would be the benefit of changing a payload without removing it first. In any case, I think the right direction for the driver would be to meet the requirement to use the proper atomic state with the drm helper functions (not the object's current SW state) as described above.
Hi Imre,
Thanks for pointing that out and that was also one of my plan to refactor! But I would take that as another patch to follow up and would like to align with you the idea for this helper function itself more.
My concern is referring to the old state to remove the time slot is just not generic right for what the helper function shall accomplish. Even it doesn't bring benefit as you see, having old state as the input imposes limitation to drivers using it which is the downside that it brought. The limitation is like what I tried to explain before. For an instance, if one driver has compressed streams allocated in the previous commit and wants to disable these streams at the current commit. It handles and accomplishes the commit into two steps that firstly to disable dsc engine only (which increases the time slot), and next disable the streams. Under this design, the old state can't represent the exact number of time slot that it want's to remove at the moment. For this case, I don't think
drm_dp_remove_payload_part2should block the driver to use it since the driver doesn't do things wrong. Conversely, it's due to the helper function constrains the driver to use an inappropriate input.
Quoted 8 lines
And with or without my patch, the current payload allocation table (i.e.
vc_start_slotIn new state) and the time slot that this new state is going to set (i.e.time_slotsin new state) are already both in the new state. I think the patch doesn't add irrelevant input for removing a payload as this helper function should do, because the time slot it should remove is the exact one in the payload table now, not the one captured in previous commit. So in contrast, refer to old state time slot is a bit confusing to me because it's not generic right. Would like to know your thought on this point. Appreciate your time!
I'd like to be sure that the payload is removed with the size it was added with in the previous commit and as I wrote above not depend for this on the new payload state with a mixture of old/current/new states. Based on that I'd be ok for instance with a new
int drm_dp_remove_port_payload(mgr, mst_state, port)
function which looks up / removes the payload with the time_slots calculated based on the payload table as in your patch and returns the calculated time_slots.
The AMD driver could call the above function and the current drm_dp_remove_payload(mgr, mst_state, old_payload) function would be
time_slots = drm_dp_remove_port_payload(mgr, mst_state, old_payload->port); WARN_ON(time_slots != old_payload->time_slots);
--Imre
RE: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
[Public]
Quoted 349 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Saturday, September 9, 2023 3:18 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 104 lines · Wayne
[AMD Official Use Only - General]
Quoted 102 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 25, 2023 9:56 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 90 lines · Wayne
[AMD Official Use Only - General]
Quoted 88 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Saturday, August 19, 2023 1:46 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 41 lines · Wayne
[AMD Official Use Only - General]
Quoted 39 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Tuesday, August 8, 2023 12:00 AM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 26 lines · Wayne
[AMD Official Use Only - General]
Quoted 24 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 4, 2023 11:32 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 8 lines · Wayne Lin
[...]
diff --gita/drivers/gpu/drm/display/drm_dp_mst_topology.cb/drivers/gpu/drm/display/drm_dp_mst_topology.c
indexe04f87ff755a..4270178f95f6100644drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0@@ -3382,8 +3382,7 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 2 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0* drm_dp_remove_payload_part2() - Remove an MST payload
locally
Quoted 34 lines
Quoted 34 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
* @mgr: Manager to use.
* @mst_state: The MST atomic state
- * @old_payload: The payload with its old state
- * @new_payload: The payload with its latest state
+ * @payload: The payload with its latest state *
* Updates the starting time slots of all other
payloads which would haveQuoted 8 lines
Quoted 8 lines
been shifted towards
Quoted 4 lines
* the start of the payload ID table as a result of
removing a payload. Driver should call this @@ -3392,25
+3391,36 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 2 lines
*/
void drm_dp_remove_payload_part2(struct
drm_dp_mst_topology_mgrQuoted 17 lines
Quoted 17 lines
*mgr,
Quoted 2 lines
struct
drm_dp_mst_topology_state
*mst_state,Quoted 1 line
- const struct
drm_dp_mst_atomic_payload
*old_payload,Quoted 1 line
- struct
drm_dp_mst_atomic_payload
*new_payload)Quoted 2 lines
+ struct
+drm_dp_mst_atomic_payload*payload)
Quoted 6 lines
{ struct drm_dp_mst_atomic_payload *pos;drivers/gpu/drm/display/drm_dp_mst_topology.c +4
-0+u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; + +/* Find the current allocated time slot number of the payload */
Quoted 141 lines
Quoted 141 lines
Quoted 34 lines
Quoted 34 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 17 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 9 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +9
-0+list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - +payload->vc_start_slot;Imo, the intuitive way would be to pass the old payload state to this function - which already contains the required
time_slotsparam
- and refactor things instead movingvc_start_slotfrom the payload state to mgr suggested by Ville earlier.--Imre
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Hi Imre,
Thanks for your feedback!I understand it's functionally correct. But IMHO, it's still a bit conceptually different between the time slot in old state and the time slot in current payload table. My thought is the time slot at the moment when we are removing the payload would be a better
choice.
Quoted 5 lines
Quoted 5 lines
Yes, they are different. The old state contains the time slot the payload was added with in a preceding commit and so the time slot value which should be used when removing the same payload in the
current commit.
Quoted 1 line
Quoted 1 line
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
The new state contains a time slot value with which the payload will be added in the current commit and can be different than the one in the old state if the current commit has changed the payload size (meaning that the same atomic commit will first remove the payload using the time slot value in the old state and afterwards will add back the same payload using the time slot value
in the new state).
Quoted 24 lines
Quoted 24 lines
Quoted 5 lines
Quoted 1 line
Appreciate your time, Imre!
Yes I understand, so I'm not using the number of the time slot in the new
state.
Quoted 7 lines
I'm referring to the start slot instead which is updated during every allocation and removement at current commit.
Like what you said, current commit manipulation could be a mix of allocations and removements for the payload. My thought is, conceptually, looking up the latest number of time slot is a better choice
rather than the one in old state.
Quoted 10 lines
It's relatively intuitive to me since we are removing payload from current payload table and which changes since last preceding commit till the moment we're deleting the payload. Although it's unreasonable that these values are different.
Quoted 5 lines
Quoted 4 lines
And with this, we can also simplify some codes. Especially remove workaround in amd driver. In fact, DRM mst code maintains the payload table and all the time slot info is in it already. We don't really have to pass a new parameter.
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
I agree that
drm_dp_remove_payload()could be simplified, but this should be done so that the drivers can pass the old payload state to it (without having to pass the new state). This would be possible ifvc_start_slotwas not tracked in the payload state (which is really not an atomic state that can be precomputed as all other atomic state), rather it would be tracked instruct
drm_dp_mst_topology_mgr.Quoted 26 lines
Quoted 26 lines
Quoted 9 lines
Quoted 1 line
So the reason I chose to pass the new state is like what I mentioned above. I would prefer to carry the latest updated payload table instead which is in the new state. And I agree with the explanation for the
vc_start_slotand that's also my thought at the beginning. It could be a refactor later, but no matter the start slot is put into payload state or the topology manager I would prefer to refer to the latestpayload table rather than the number of time slot in the old state.
Quoted 16 lines
Quoted 10 lines
It looks like AMD has to reconstruct the old state in
dm_helpers_construct_old_payload(). Could you explain why it couldn't instead just passold_payloadacquired by
old_mst_state=drm_atomic_get_old_mst_topology_state();old_payload=drm_atomic_get_mst_payload_state(old_mst_state);?
AMD doesn't pass the drm old state to the stage while HW is deleting the payload. The reason is that HW payload table is known during HW programming procedure, so the payload removement is based on the table at the moment.
Quoted 7 lines
Quoted 7 lines
Quoted 2 lines
AMD expected the current number of time slot is also already maintained in drm layer.
Yes, both of the above are maintained by the drm layer, but it also means it doesn't really need to recalculate
time_slots_to_removeas done in this patch, since that info is already available in the old payloadstate.
Quoted 5 lines
Quoted 5 lines
Afaics the AMD driver calls properly
drm_atomic_helper_commit()->drm_atomic_helper_swap_state()Quoted 8 lines
Quoted 8 lines
after a commit, so that all the payloads it added should be tracked now as the old payload state.
So could you confirm what is the
old_payload->time_slotsvalue (which you get with the above functions) at the point of removing this payload and if it's not thetime_slotsvalue this same payload was actually added with previously, why this is so (via some examplesequence)?
Quoted 6 lines
Quoted 2 lines
Thanks.
Hi Imre,
I'm not saying that thetime_slotscarried in the old state is wrong withinamd driver.
Quoted 3 lines
But just amd driver doesn't carry the drm state to the step when it's removing the payload, since the info is already in its hardware and drm used to maintain the info in the drm layer.
Hm, in
dm_helpers_dp_mst_write_payload_allocation_table()the
mst_state=to_drm_dp_mst_topology_state(mst_mgr->base.state);Quoted 7 lines
Quoted 7 lines
used as the new state doesn't look ok to me. The above is the MST object's current SW state after an atomic commit/swap, but this isn't the new state a driver should program to the HW or pass to the drm helper functions. The object's current state could be ahead of what the driver should program to the HW, if the driver properly implements commit pipelining (so at the point you're programming a state you'd have already future commits computed and queued up, each
queued-up commit with its own state).
Quoted 7 lines
Quoted 7 lines
So instead of the above
mst_statethe driver should pass down thedrm_atomic_statetodm_helpers_dp_mst_write_payload_allocation_table()
which should use that instead to get the MST state it should program or pass to the drm layer.Quoted 7 lines
Quoted 7 lines
Quoted 2 lines
So the patch is trying to get the behavior of this helper function back to what it used to be.
The behavior before was actually broken and one reason for that was a confusion about what's stored in the new payload state. It's a mixture of old/current state (
vc_start_slot) and new state (time_slots). So I don't think it's a good idea to add even more dependencyon this semantic.
Quoted 7 lines
Quoted 7 lines
I think the right solution for what this patch is about - remove the need for
dm_helpers_construct_old_payload(), would be to pass downdrm_atomic_statetodm_helpers_dp_mst_write_payload_allocation_table()
based on the above.Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
And the main reason that I want to change the pass in parameter is like what I mentioned previously. The commit manipulation could be a mix of allocations and removements for the payload. And in the spec, it also introduces examples to reduce or increase the payload allocation. Although we're not using reduction/increment today, it implicitly imposes the limitation to use them before calling the removement helper function with the old state as the passed in parameter. So I also want to remove the dependency by this
patch.
Quoted 6 lines
Quoted 6 lines
Quoted 1 line
Would like to know your thoughts about this.
It's not clear what would be the benefit of changing a payload without removing it first. In any case, I think the right direction for the driver would be to meet the requirement to use the proper atomic state with the drm helper functions (not the object's current SW
state) as described above.
Quoted 11 lines
Hi Imre,
Thanks for pointing that out and that was also one of my plan to refactor! But I would take that as another patch to follow up and would like to align with you the idea for this helper function itself more.
My concern is referring to the old state to remove the time slot is just not generic right for what the helper function shall accomplish. Even it doesn't bring benefit as you see, having old state as the input imposes limitation to drivers using it which is the downside that it
brought. The limitation is like what I tried to explain before.
Quoted 8 lines
For an instance, if one driver has compressed streams allocated in the previous commit and wants to disable these streams at the current commit. It handles and accomplishes the commit into two steps that firstly to disable dsc engine only (which increases the time slot), and next disable the streams. Under this design, the old state can't represent the exact number of time slot that it want's to remove at the moment. For this case, I don't think
drm_dp_remove_payload_part2should block the driver to use it since the driver doesn't do things wrong.Conversely, it's due to the helper function constrains the driver to use an inappropriate input.
Quoted 8 lines
And with or without my patch, the current payload allocation table (i.e.
vc_start_slotIn new state) and the time slot that this new state is going to set (i.e.time_slotsin new state) are already both in the new state. I think the patch doesn't add irrelevant input for removing a payload as this helper function should do, because the time slot it should remove is the exact one in the payload table now, not the one captured in previous commit. So in contrast, refer to old state timeslot is a bit confusing to me because it's not generic right. Would like to know your thought on this point.
Quoted 1 line
Appreciate your time!
I'd like to be sure that the payload is removed with the size it was added with in the previous commit and as I wrote above not depend for this on the new payload state with a mixture of old/current/new states.
Based on that I'd be ok for instance with a newint
drm_dp_remove_port_payload(mgr, mst_state, port)function which looks up / removes the payload with the
time_slotscalculated based on the payload table as in your patch and returns the calculatedtime_slots.
Quoted 8 lines
The AMD driver could call the above function and the current
drm_dp_remove_payload(mgr, mst_state, old_payload)function would be
time_slots= drm_dp_remove_port_payload(mgr,mst_state,old_payload->port); WARN_ON(time_slots != old_payload->time_slots);--Imre
Sorry but I might not fully understand what you suggested here. Would like to know if you agree on referring to the time slot number of the payload table at the moment is better then referring old_payload->time_slots for drm_dp_remove_payload()? If you agree on that, my patch actually is just replacing old_payload->time_slots with the more appropriate one. Not adding mixture of old/current but replacing the old with the current one. And like what I explained in previous example, when calling drm_dp_remove_payload(), the time slot number to be removed shouldn't be constrained to the one in previous commit. The number in the payload table when we're about to remove the payload might be a better choice. Could you elaborate more what's the mixture that this patch is adding on, please?
As for the changing suggestion, are you suggesting to create a new function drm_dp_remove_port_payload() to wrap up the calculation in my patch? If so, I think that's the consensus to use current time slot number to replace the one in old_payload. Therefore, it doesn't have to pass old_payload to drm_dp_remove_port_payload(), and "WARN_ON(time_slots != old_payload->time_slots);" is not appropriate as for the example that I gave previously.
Thanks for helping me out here.
--
Regards,
Wayne
Re: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
Quoted 21 lines · Wayne
[Public]
[...]Quoted 19 lines
I'd like to be sure that the payload is removed with the size it was added with in the previous commit and as I wrote above not depend for this on the new payload state with a mixture of old/current/new states.
Based on that I'd be ok for instance with a newint
drm_dp_remove_port_payload(mgr, mst_state, port)function which looks up / removes the payload with the
time_slotscalculated based on the payload table as in your patch and returns the calculatedtime_slots.The AMD driver could call the above function and the current
drm_dp_remove_payload(mgr, mst_state, old_payload)function would be
time_slots= drm_dp_remove_port_payload(mgr,mst_state,old_payload->port); WARN_ON(time_slots != old_payload->time_slots);
Quoted 8 lines
Quoted 1 line
--Imre
Sorry but I might not fully understand what you suggested here. Would like to know if you agree on referring to the time slot number of the payload table at the moment is better then referring
old_payload->time_slotsfordrm_dp_remove_payload()? If you agree on that, my patch actually is just replacingold_payload->time_slotswith the more appropriate one. Not adding mixture of old/current but replacing the old with the current one.
The new_payload state contains a mixture of old/current/new state at the moment and this patch adds more dependency on that, recalculating the old payload size from that state. For i915 this recalculation isn't needed, the size is already available in the old payload state.
Quoted 6 lines
And like what I explained in previous example, when calling
drm_dp_remove_payload(), the time slot number to be removed shouldn't be constrained to the one in previous commit. The number in the payload table when we're about to remove the payload might be a better choice. Could you elaborate more what's the mixture that this patch is adding on, please?
Quoted 6 lines
As for the changing suggestion, are you suggesting to create a new function
drm_dp_remove_port_payload()to wrap up the calculation in my patch? If so, I think that's the consensus to use current time slot number to replace the one inold_payload. Therefore, it doesn't have to passold_payloadtodrm_dp_remove_port_payload(), and "WARN_ON(time_slots != old_payload->time_slots);" is not appropriate as for the example that I gave previously.
I meant something like the following:
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c +3 -4
@@ -343,7 +343,7 @@ bool dm_helpers_dp_mst_send_payload_allocation(struct amdgpu_dm_connector *aconnector;struct drm_dp_mst_topology_state *mst_state;struct drm_dp_mst_topology_mgr *mst_mgr; -struct drm_dp_mst_atomic_payload *new_payload, *old_payload; +struct drm_dp_mst_atomic_payload *new_payload;enum mst_progress_status set_flag =MST_ALLOCATE_NEW_PAYLOAD ;enum mst_progress_status clr_flag =MST_CLEAR_ALLOCATED_PAYLOAD ;int ret =0 ;
@@ -366,9 +366,8 @@ bool dm_helpers_dp_mst_send_payload_allocation(if (enable) { ret =drm_dp_add_payload_part2 (mst_mgr, mst_state->base.state, new_payload); }else { -dm_helpers_construct_old_payload (stream->link, mst_state->pbn_div, - new_payload, old_payload); -drm_dp_remove_payload_part2 (mst_mgr, mst_state, old_payload, new_payload); +drm_dp_remove_current_payload_part2 (mst_mgr, mst_state->base.state, + aconnector->mst_output_port); }if (ret) {
drivers/gpu/drm/display/drm_dp_mst_topology.c +41 -8
78 lines not on this page · raw message
drivers/gpu/drm/i915/display/intel_dp_mst.c +1 -10
@@ -576,14 +576,6 @@ static void intel_mst_post_disable_dp(struct intel_atomic_state *state,struct intel_dp *intel_dp = &dig_port->dp;struct intel_connector *connector =to_intel_connector (old_conn_state->connector); -struct drm_dp_mst_topology_state *old_mst_state = -drm_atomic_get_old_mst_topology_state (&state->base, &intel_dp->mst_mgr); -struct drm_dp_mst_topology_state *new_mst_state = -drm_atomic_get_new_mst_topology_state (&state->base, &intel_dp->mst_mgr); -const struct drm_dp_mst_atomic_payload *old_payload = -drm_atomic_get_mst_payload_state (old_mst_state, connector->port); -struct drm_dp_mst_atomic_payload *new_payload = -drm_atomic_get_mst_payload_state (new_mst_state, connector->port);struct drm_i915_private *dev_priv =to_i915 (connector->base.dev);bool last_mst_stream;
@@ -604,8 +596,7 @@ static void intel_mst_post_disable_dp(struct intel_atomic_state *state,wait_for_act_sent (encoder, old_crtc_state); -drm_dp_remove_payload_part2 (&intel_dp->mst_mgr, new_mst_state, - old_payload, new_payload); +drm_dp_remove_payload_part2 (&intel_dp->mst_mgr, &state->base, connector->port);intel_ddi_disable_transcoder_func (old_crtc_state);
drivers/gpu/drm/nouveau/dispnv50/disp.c +1 -5
@@ -889,17 +889,13 @@ nv50_msto_cleanup(struct drm_atomic_state *state,struct nouveau_drm *drm =nouveau_drm (msto->encoder.dev);struct drm_dp_mst_atomic_payload *new_payload =drm_atomic_get_mst_payload_state (new_mst_state, msto->mstc->port); -struct drm_dp_mst_topology_state *old_mst_state = -drm_atomic_get_old_mst_topology_state (state, mgr); -const struct drm_dp_mst_atomic_payload *old_payload = -drm_atomic_get_mst_payload_state (old_mst_state, msto->mstc->port);NV_ATOMIC (drm,"%s: msto cleanup\n" , msto->encoder.name);if (msto->disabled) { msto->mstc =NULL ; msto->disabled = false; -drm_dp_remove_payload_part2 (mgr, new_mst_state, old_payload, new_payload); +drm_dp_remove_payload_part2 (mgr, state, msto->mstc->port); }else if (msto->enabled) {drm_dp_add_payload_part2 (mgr, state, new_payload); msto->enabled = false;
include/drm/display/drm_dp_mst_helper.h +5 -3
@@ -856,9 +856,11 @@ void drm_dp_remove_payload_part1(struct drm_dp_mst_topology_mgr *mgr,struct drm_dp_mst_topology_state *mst_state, struct drm_dp_mst_atomic_payload *payload);void drm_dp_remove_payload_part2 (struct drm_dp_mst_topology_mgr *mgr, -struct drm_dp_mst_topology_state *mst_state, -const struct drm_dp_mst_atomic_payload *old_payload, -struct drm_dp_mst_atomic_payload *new_payload); +struct drm_atomic_state *state, +struct drm_dp_mst_port *port); +int drm_dp_remove_current_payload_part2 (struct drm_dp_mst_topology_mgr *mgr, +struct drm_atomic_state *state, +struct drm_dp_mst_port *port);int drm_dp_check_act_status (struct drm_dp_mst_topology_mgr *mgr);
Quoted 1 line
Thanks for helping me out here.
RE: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
[Public]
Quoted 41 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Tuesday, September 12, 2023 7:19 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()Quoted 21 lines · Wayne
[Public]
[...]Quoted 19 lines
I'd like to be sure that the payload is removed with the size it was added with in the previous commit and as I wrote above not depend for this on the new payload state with a mixture of old/current/new states. Based on that I'd be ok for instance with a new
int
drm_dp_remove_port_payload(mgr, mst_state, port)function which looks up / removes the payload with the
time_slotscalculated based on the payload table as in your patch and returns the calculatedtime_slots.The AMD driver could call the above function and the current
drm_dp_remove_payload(mgr, mst_state, old_payload)function would be
time_slots= drm_dp_remove_port_payload(mgr,mst_state,old_payload->port); WARN_ON(time_slots != old_payload->time_slots);Quoted 8 lines
Quoted 1 line
--Imre
Sorry but I might not fully understand what you suggested here. Would like to know if you agree on referring to the time slot number of the payload table at the moment is better then referring
old_payload->time_slotsfordrm_dp_remove_payload()? If you agree on that, my patch actually is just replacingold_payload->time_slotswith the more appropriate one. Not adding mixture of old/current but replacing
Quoted 7 lines
the old with the current one.
The
new_payloadstate contains a mixture of old/current/new state at the moment and this patch adds more dependency on that, recalculating the old payload size from that state. For i915 this recalculation isn't needed, the size is already available in the old payload state.
I see. Thanks, Imre. So it's about the idea to have another patch to extract things like vc_start_slot out of mst state.
Quoted 11 lines
Quoted 5 lines
And like what I explained in previous example, when calling
drm_dp_remove_payload(), the time slot number to be removed shouldn't be constrained to the one in previous commit. The number in the payload table when we're about to remove the payload might be a better choice. Could you elaborate more what's the mixture that this patch isadding on, please?
Quoted 5 lines
As for the changing suggestion, are you suggesting to create a new function
drm_dp_remove_port_payload()to wrap up the calculation in my patch? If so, I think that's the consensus to use current time slot number to replace
Quoted 7 lines
the one in
old_payload.Quoted 3 lines
Therefore, it doesn't have to pass
old_payloadtodrm_dp_remove_port_payload(), and "WARN_ON(time_slots != old_payload->time_slots);" is not appropriate as for the example that I gavepreviously.
I meant something like the following:
The change looks good to me. I'll update the patch. Thanks for the help.
Quoted 227 lines
diff --git
a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.cb/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.cindexcbef4ff28cd8a..0555433d8050b100644drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c +1
-1@@ -343,7 +343,7 @@ bool dm_helpers_dp_mst_send_payload_allocation(struct amdgpu_dm_connector *aconnector;struct drm_dp_mst_topology_state *mst_state;struct drm_dp_mst_topology_mgr *mst_mgr; -struct drm_dp_mst_atomic_payload *new_payload, *old_payload; +struct drm_dp_mst_atomic_payload *new_payload;enum mst_progress_status set_flag =MST_ALLOCATE_NEW_PAYLOAD ; enum mst_progress_status clr_flag =MST_CLEAR_ALLOCATED_PAYLOAD; int ret = 0;drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c +1
-3@@ -366,9 +366,8 @@ bool dm_helpers_dp_mst_send_payload_allocation(if (enable) { ret =drm_dp_add_payload_part2 (mst_mgr, mst_state- >base.state, new_payload); }else { -dm_helpers_construct_old_payload (stream->link, mst_state- >pbn_div, - new_payload, old_payload); -drm_dp_remove_payload_part2 (mst_mgr, mst_state, old_payload, new_payload); + drm_dp_remove_current_payload_part2(mst_mgr,
mst_state->base.state,
+ aconnector-Quoted 1 line
mst_output_port);} if (ret) {drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-2@@ -3382,37 +3382,70 @@EXPORT_SYMBOL (drm_dp_remove_payload_part1); * drm_dp_remove_payload_part2() - Remove an MST payload locally * @mgr: Manager to use. * @mst_state: The MST atomic state - * @old_payload: The payload with its old state - * @new_payload: The payload with its latest state + * @port: MST port * * Updates the starting time slots of all other payloads which would have beenshifted towards
* the start of the payload ID table as a result of removing a payload. Driver
should call this
* function whenever it removes a payload in its HW. It's independent to the
result of payload
* allocation/deallocation at branch devices along the virtual channel. */
-void drm_dp_remove_payload_part2(structdrm_dp_mst_topology_mgr*mgr,
- structdrm_dp_mst_topology_state*mst_state,
- const structdrm_dp_mst_atomic_payload*old_payload,
- structdrm_dp_mst_atomic_payload*new_payload)
+int drm_dp_remove_current_payload_part2(structdrm_dp_mst_topology_mgr*mgr,drivers/gpu/drm/display/drm_dp_mst_topology.c +19
-0+struct drm_atomic_state *state, + struct drm_dp_mst_port *port) { struct drm_dp_mst_atomic_payload *pos; +struct drm_dp_mst_topology_state *mst_state = +drm_atomic_get_new_mst_topology_state (state, mgr); +struct drm_dp_mst_atomic_payload *new_payload = +drm_atomic_get_mst_payload_state (mst_state, port); +int time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; + +/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != new_payload && + pos->vc_start_slot > new_payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - +new_payload->vc_start_slot;/* Remove local payload allocation */ list_for_each_entry(pos, &mst_state->payloads, next) { if (pos != new_payload && pos->vc_start_slot > new_payload-Quoted 1 line
vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-2- pos->vc_start_slot -= old_payload->time_slots; + pos->vc_start_slot -= time_slots_to_remove; } new_payload->vc_start_slot =-1 ; mgr->payload_count--; - mgr->next_start_slot -= old_payload->time_slots; + mgr->next_start_slot -= time_slots_to_remove;if (new_payload->delete)drm_dp_mst_put_port_malloc (new_payload->port); new_payload->payload_allocation_status =
DRM_DP_MST_PAYLOAD_ALLOCATION_NONE;drivers/gpu/drm/display/drm_dp_mst_topology.c +6
-0+ +return time_slots_to_remove; +} +EXPORT_SYMBOL (drm_dp_remove_current_payload_part2); + +void drm_dp_remove_payload_part2(struct drm_dp_mst_topology_mgr*mgr,
drivers/gpu/drm/display/drm_dp_mst_topology.c +10
-0+struct drm_atomic_state *state, + struct drm_dp_mst_port *port) +{ + struct drm_dp_mst_topology_state *old_mst_state = +drm_atomic_get_old_mst_topology_state (state, mgr); +struct drm_dp_mst_atomic_payload *old_payload = +drm_atomic_get_mst_payload_state (old_mst_state, port); +int time_slots; + + time_slots = drm_dp_remove_current_payload_part2(mgr, state,port);
drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-0+ +drm_WARN_ON (mgr->dev, time_slots != old_payload->time_slots); }EXPORT_SYMBOL (drm_dp_remove_payload_part2); +/** * drm_dp_add_payload_part2() - Execute payload update part 2 * @mgr: Manager to use. drivers/gpu/drm/i915/display/intel_dp_mst.c +0
-6@@ -576,14 +576,6 @@ static void intel_mst_post_disable_dp(structintel_atomic_state *state, struct intel_dp *intel_dp = &dig_port->dp;struct intel_connector *connector =to_intel_connector (old_conn_state->connector); -struct drm_dp_mst_topology_state *old_mst_state = -drm_atomic_get_old_mst_topology_state (&state->base, &intel_dp->mst_mgr); -struct drm_dp_mst_topology_state *new_mst_state = -drm_atomic_get_new_mst_topology_state (&state->base, &intel_dp->mst_mgr); -const struct drm_dp_mst_atomic_payload *old_payload = - drm_atomic_get_mst_payload_state(old_mst_state,
connector->port);drivers/gpu/drm/i915/display/intel_dp_mst.c +0
-2-struct drm_dp_mst_atomic_payload *new_payload = - drm_atomic_get_mst_payload_state(new_mst_state,connector->port); struct drm_i915_private *dev_priv = to_i915(connector->base.dev); bool last_mst_stream;drivers/gpu/drm/i915/display/intel_dp_mst.c +2
-2@@ -604,8 +596,7 @@ static void intel_mst_post_disable_dp(structintel_atomic_state *state,wait_for_act_sent (encoder ,old_crtc_state ); -drm_dp_remove_payload_part2 (&intel_dp->mst_mgr, new_mst_state, - old_payload, new_payload); +drm_dp_remove_payload_part2 (&intel_dp->mst_mgr, &state->base, +connector->port);intel_ddi_disable_transcoder_func (old_crtc_state);drivers/gpu/drm/nouveau/dispnv50/disp.c +0
-5@@ -889,17 +889,13 @@ nv50_msto_cleanup(struct drm_atomic_state *state, struct nouveau_drm *drm = nouveau_drm(msto->encoder.dev);struct drm_dp_mst_atomic_payload *new_payload =drm_atomic_get_mst_payload_state (new_mst_state, msto- >mstc->port); -struct drm_dp_mst_topology_state *old_mst_state = -drm_atomic_get_old_mst_topology_state (state, mgr); -const struct drm_dp_mst_atomic_payload *old_payload = -drm_atomic_get_mst_payload_state (old_mst_state, msto- >mstc->port);NV_ATOMIC (drm,"%s: msto cleanup\n" , msto->encoder.name);if (msto->disabled) { msto->mstc =NULL ; msto->disabled = false; - drm_dp_remove_payload_part2(mgr, new_mst_state,
old_payload,new_payload);
+ drm_dp_remove_payload_part2(mgr, state, msto->mstc-Quoted 1 line
port);
} else if (msto->enabled) { drm_dp_add_payload_part2(mgr, state, new_payload); msto->enabled = false;include/drm/display/drm_dp_mst_helper.h +0
-2@@ -856,9 +856,11 @@ void drm_dp_remove_payload_part1(structdrm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_topology_state *mst_state, struct drm_dp_mst_atomic_payload *payload); void drm_dp_remove_payload_part2(struct drm_dp_mst_topology_mgr *mgr, -struct drm_dp_mst_topology_state *mst_state, - const struct drm_dp_mst_atomic_payload
*old_payload,
- structdrm_dp_mst_atomic_payload*new_payload);include/drm/display/drm_dp_mst_helper.h +3
-0+struct drm_atomic_state *state, + struct drm_dp_mst_port *port); +int drm_dp_remove_current_payload_part2(struct
drm_dp_mst_topology_mgr*mgr,include/drm/display/drm_dp_mst_helper.h +2
-0+struct drm_atomic_state *state, + struct drm_dp_mst_port *port);int drm_dp_check_act_status (struct drm_dp_mst_topology_mgr *mgr);
Quoted 1 line
Quoted 1 line
Thanks for helping me out here.
--
Regards,
Wayne
Lyude Paul
maintainerRe: [PATCH 3/3] drm/mst: adjust the function drm_dp_remove_payload_part2()
In reply to Imre DeakQuoted 94 lines · Imre Deak
Quoted 68 lines · Wayne
[AMD Official Use Only - General]
Quoted 64 lines
-----Original Message-----
From: Imre Deak <[email protected]>
Sent: Friday, August 4, 2023 11:32 PM
To: Lin, Wayne <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Wentland, Harry <[email protected]>; Zuo, Jerry <[email protected]>
Subject: Re: [PATCH 3/3] drm/mst: adjust the functiondrm_dp_remove_payload_part2()drivers/gpu/drm/display/drm_dp_mst_topology.c
Quoted 7 lines · Wayne Lin
[...]
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0@@ -3382,8 +3382,7 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 8 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-2* drm_dp_remove_payload_part2() - Remove an MST payload locally * @mgr: Manager to use. * @mst_state: The MST atomic state -* @old_payload: The payload with its old state -* @new_payload: The payload with its latest state +* @payload: The payload with its latest state * * Updates the starting time slots of all other payloads which would have been shifted towards
Quoted 2 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +0
-0* the start of the payload ID table as a result of removing a payload. Driver should call this @@ -3392,25 +3391,36 @@
EXPORT_SYMBOL(drm_dp_remove_payload_part1);Quoted 2 lines
*/
void drm_dp_remove_payload_part2(structdrm_dp_mst_topology_mgr*mgr,
Quoted 1 line
struct
drm_dp_mst_topology_state
*mst_state,Quoted 1 line
- const struct
drm_dp_mst_atomic_payload
*old_payload,Quoted 1 line
- struct
drm_dp_mst_atomic_payload
*new_payload)Quoted 1 line
+ struct
drm_dp_mst_atomic_payload*payload)
Quoted 5 lines
{ struct drm_dp_mst_atomic_payload *pos;drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-0+u8 time_slots_to_remove; +u8 next_payload_vc_start = mgr->next_start_slot; +drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +10
-0+/* Find the current allocated time slot number of the payload */ +list_for_each_entry (pos, &mst_state->payloads, next) { +if (pos != payload && + pos->vc_start_slot > payload->vc_start_slot && + pos->vc_start_slot < next_payload_vc_start) + next_payload_vc_start = pos->vc_start_slot; + } + + time_slots_to_remove = next_payload_vc_start - +payload->vc_start_slot;Imo, the intuitive way would be to pass the old payload state to this function - which already contains the required
time_slotsparam - and refactor things instead movingvc_start_slotfrom the payload state to mgr suggested by Ville earlier.Quoted 1 line
--Imre
Quoted 7 lines
Hi Imre,
Thanks for your feedback!I understand it's functionally correct. But IMHO, it's still a bit conceptually different between the time slot in old state and the time slot in current payload table. My thought is the time slot at the moment when we are removing the payload would be a better choice.
Yes, they are different. The old state contains the time slot the payload was added with in a preceding commit and so the time slot value which should be used when removing the same payload in the current commit.
The new state contains a time slot value with which the payload will be added in the current commit and can be different than the one in the old state if the current commit has changed the payload size (meaning that the same atomic commit will first remove the payload using the time slot value in the old state and afterwards will add back the same payload using the time slot value in the new state).
Quoted 4 lines
And with this, we can also simplify some codes. Especially remove workaround in amd driver. In fact, DRM mst code maintains the payload table and all the time slot info is in it already. We don't really have to pass a new parameter.
Quoted 6 lines
I agree that
drm_dp_remove_payload()could be simplified, but this should be done so that the drivers can pass the old payload state to it (without having to pass the new state). This would be possible ifvc_start_slotwas not tracked in the payload state (which is really not an atomic state that can be precomputed as all other atomic state), rather it would be tracked in structdrm_dp_mst_topology_mgr.
JFYI too - I think I'm fine with us moving vc_start_slot elsewhere at this point ;)
Quoted 41 lines
It looks like AMD has to reconstruct the old state in
dm_helpers_construct_old_payload(). Could you explain why it couldn't instead just passold_payloadacquired by
old_mst_state=drm_atomic_get_old_mst_topology_state();old_payload=drm_atomic_get_mst_payload_state(old_mst_state);?
Quoted 30 lines
Quoted 16 lines
Quoted 16 lines
/* Remove local payload allocation */ list_for_each_entry(pos, &mst_state->payloads, next) {- if (pos !=
new_payload&&pos->vc_start_slot> new_payload-vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +1
-1- pos->vc_start_slot -= old_payload->time_slots; + if (pos != payload && pos->vc_start_slot > payload-
vc_start_slot)drivers/gpu/drm/display/drm_dp_mst_topology.c +3
-2+ pos->vc_start_slot -= time_slots_to_remove; } - new_payload->vc_start_slot =-1 ; + payload->vc_start_slot =-1 ; mgr->payload_count--; - mgr->next_start_slot -= old_payload->time_slots; + mgr->next_start_slot -= time_slots_to_remove;Quoted 10 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c EXPORT_SYMBOL(drm_dp_remove_payload_part1);
Quoted 6 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-3-if (new_payload->delete) -drm_dp_mst_put_port_malloc (new_payload->port); +if (payload->delete) +drm_dp_mst_put_port_malloc (payload->port); - new_payload->payload_allocation_status =
DRM_DP_MST_PAYLOAD_ALLOCATION_NONE;Quoted 3 lines
drivers/gpu/drm/display/drm_dp_mst_topology.c +2
-0+ payload->payload_allocation_status = +DRM_DP_MST_PAYLOAD_ALLOCATION_NONE ; }--
Regards,
Wayne
Signature · 3 lines
Cheers, Lyude Paul (she/her) Software Engineer at Red Hat