[Public]
Thanks, Lyude!
Comments inline.
> -----Original Message-----
> From: Lyude Paul <ly...@redhat.com>
> Sent: Saturday, November 5, 2022 7:59 AM
> To: amd-gfx@lists.freedesktop.org
> Cc: Wentland, Harry <harry.wentl...@amd.com>; sta...@vger.kernel.org;
> Li, Sun peng (Leo) <sunpeng...@amd.com>; Siqueira, Rodrigo
> <rodrigo.sique...@amd.com>; Deucher, Alexander
> <alexander.deuc...@amd.com>; Koenig, Christian
> <christian.koe...@amd.com>; Pan, Xinhui <xinhui....@amd.com>; David
> Airlie <airl...@gmail.com>; Daniel Vetter <dan...@ffwll.ch>; Kazlauskas,
> Nicholas <nicholas.kazlaus...@amd.com>; Pillai, Aurabindo
> <aurabindo.pil...@amd.com>; Li, Roman <roman...@amd.com>; Zuo, Jerry
> <jerry....@amd.com>; Wu, Hersen <hersenxs...@amd.com>; Lin, Wayne
> <wayne....@amd.com>; Thomas Zimmermann <tzimmerm...@suse.de>;
> Mahfooz, Hamza <hamza.mahf...@amd.com>; Hung, Alex
> <alex.h...@amd.com>; Francis, David <david.fran...@amd.com>; Mikita
> Lipski <mikita.lip...@amd.com>; Liu, Wenjing <wenjing....@amd.com>;
> open list:DRM DRIVERS <dri-de...@lists.freedesktop.org>; open list <linux-
> ker...@vger.kernel.org>
> Subject: [PATCH 1/2] drm/amdgpu/mst: Stop ignoring error codes and
> deadlocking
>
> It appears that amdgpu makes the mistake of completely ignoring the return
> values from the DP MST helpers, and instead just returns a simple true/false.
> In this case, it seems to have come back to bite us because as a result of
> simply returning false from compute_mst_dsc_configs_for_state(), amdgpu
> had no way of telling when a deadlock happened from these helpers. This
> could definitely result in some kernel splats.
>
> Signed-off-by: Lyude Paul <ly...@redhat.com>
> Fixes: 8c20a1ed9b4f ("drm/amd/display: MST DSC compute fair share")
> Cc: Harry Wentland <harry.wentl...@amd.com>
> Cc: <sta...@vger.kernel.org> # v5.6+
> ---
> .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 18 +--
> .../display/amdgpu_dm/amdgpu_dm_mst_types.c | 107 ++++++++++------
> --
> .../display/amdgpu_dm/amdgpu_dm_mst_types.h | 12 +-
> 3 files changed, 73 insertions(+), 64 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> index 0db2a88cd4d7b..6f76b2c84cdb5 100644
> --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> @@ -6462,7 +6462,7 @@ static int
> dm_update_mst_vcpi_slots_for_dsc(struct drm_atomic_state *state,
> struct drm_connector_state *new_con_state;
> struct amdgpu_dm_connector *aconnector;
> struct dm_connector_state *dm_conn_state;
> - int i, j;
> + int i, j, ret;
> int vcpi, pbn_div, pbn, slot_num = 0;
>
> for_each_new_connector_in_state(state, connector,
> new_con_state, i) { @@ -6509,8 +6509,11 @@ static int
> dm_update_mst_vcpi_slots_for_dsc(struct drm_atomic_state *state,
> dm_conn_state->pbn = pbn;
> dm_conn_state->vcpi_slots = slot_num;
>
> - drm_dp_mst_atomic_enable_dsc(state, aconnector-
> >port, dm_conn_state->pbn,
> - false);
> + ret = drm_dp_mst_atomic_enable_dsc(state,
> aconnector->port,
> + dm_conn_state-
> >pbn, false);
> + if (ret != 0)
> + return ret;
> +
> continue;
> }
>
> @@ -9523,10 +9526,9 @@ static int amdgpu_dm_atomic_check(struct
> drm_device *dev,
>
> #if defined(CONFIG_DRM_AMD_DC_DCN)
> if (dc_resource_is_dsc_encoding_supported(dc)) {
> - if (!pre_validate_dsc(state, &dm_state, vars)) {
> - ret = -EINVAL;
> + ret = pre_validate_dsc(state, &dm_state, vars);
> + if (ret != 0)
> goto fail;
> - }
> }
> #endif
>
> @@ -9621,9 +9623,9 @@ static int amdgpu_dm_atomic_check(struct
> drm_device *dev,
> }
>
> #if defined(CONFIG_DRM_AMD_DC_DCN)
> - if (!compute_mst_dsc_configs_for_state(state, dm_state-
> >context, vars)) {
> + ret = compute_mst_dsc_configs_for_state(state, dm_state-
> >context, vars);
> + if (ret) {
>
> DRM_DEBUG_DRIVER("compute_mst_dsc_configs_for_state()
> failed\n");
> - ret = -EINVAL;
> goto fail;
> }
>
> diff --git
> a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> index 6ff96b4bdda5c..30bc2e5058b70 100644
> --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> +++
> b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
> @@ -864,25 +864,25 @@ static bool try_disable_dsc(struct
> drm_atomic_state *state,
> return true;
> }
>
> -static bool compute_mst_dsc_configs_for_link(struct drm_atomic_state
> *state,
> - struct dc_state *dc_state,
> - struct dc_link *dc_link,
> - struct dsc_mst_fairness_vars *vars,
> - struct drm_dp_mst_topology_mgr
> *mgr,
> - int *link_vars_start_index)
> +static int compute_mst_dsc_configs_for_link(struct drm_atomic_state
> *state,
> + struct dc_state *dc_state,
> + struct dc_link *dc_link,
> + struct dsc_mst_fairness_vars *vars,
> + struct drm_dp_mst_topology_mgr
> *mgr,
> + int *link_vars_start_index)
> {
> struct dc_stream_state *stream;
> struct dsc_mst_fairness_params params[MAX_PIPES];
> struct amdgpu_dm_connector *aconnector;
> struct drm_dp_mst_topology_state *mst_state =
> drm_atomic_get_mst_topology_state(state, mgr);
> int count = 0;
> - int i, k;
> + int i, k, ret;
> bool debugfs_overwrite = false;
>
> memset(params, 0, sizeof(params));
>
> if (IS_ERR(mst_state))
> - return false;
> + return PTR_ERR(mst_state);
>
> mst_state->pbn_div = dm_mst_get_pbn_divider(dc_link); #if
> defined(CONFIG_DRM_AMD_DC_DCN) @@ -933,7 +933,7 @@ static bool
> compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
>
> if (count == 0) {
> ASSERT(0);
> - return true;
> + return 0;
> }
>
> /* k is start index of vars for current phy link used by mst hub */ @@
> -949,11 +949,14 @@ static bool compute_mst_dsc_configs_for_link(struct
> drm_atomic_state *state,
> vars[i + k].bpp_x16 = 0;
> if (drm_dp_atomic_find_time_slots(state, params[i].port-
> >mgr, params[i].port,
> vars[i + k].pbn) < 0)
> - return false;
> + return -EINVAL;
Should we also return the error code get from drm_dp_atomic_find_time_slots()
rather than
assigning a new one here?
> }
> - if (!drm_dp_mst_atomic_check(state) && !debugfs_overwrite) {
> + ret = drm_dp_mst_atomic_check(state);
> + if (ret == 0 && !debugfs_overwrite) {
> set_dsc_configs_from_fairness_vars(params, vars, count, k);
> - return true;
> + return 0;
> + } else if (ret == -EDEADLK) {
> + return ret;
I think we should return here whenever there is an error. Not just for EDEADLK
case.
> }
>
> /* Try max compression */
> @@ -964,29 +967,30 @@ static bool
> compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
> vars[i + k].bpp_x16 =
> params[i].bw_range.min_target_bpp_x16;
> if (drm_dp_atomic_find_time_slots(state,
> params[i].port->mgr,
> params[i].port, vars[i
> + k].pbn) < 0)
> - return false;
> + return -EINVAL;
Same as above.
> } else {
> vars[i + k].pbn =
> kbps_to_peak_pbn(params[i].bw_range.stream_kbps);
> vars[i + k].dsc_enabled = false;
> vars[i + k].bpp_x16 = 0;
> if (drm_dp_atomic_find_time_slots(state,
> params[i].port->mgr,
> params[i].port, vars[i
> + k].pbn) < 0)
> - return false;
> + return -EINVAL;
Same as above.
> }
> }
> - if (drm_dp_mst_atomic_check(state))
> - return false;
> + ret = drm_dp_mst_atomic_check(state);
> + if (ret != 0)
> + return ret;
>
> /* Optimize degree of compression */
> if (!increase_dsc_bpp(state, mst_state, dc_link, params, vars, count,
> k))
> - return false;
> + return -ENOSPC;
>
> if (!try_disable_dsc(state, dc_link, params, vars, count, k))
> - return false;
> + return -ENOSPC;
>
> set_dsc_configs_from_fairness_vars(params, vars, count, k);
>
> - return true;
> + return 0;
> }
>
> static bool is_dsc_need_re_compute(
> @@ -1087,15 +1091,16 @@ static bool is_dsc_need_re_compute(
> return is_dsc_need_re_compute;
> }
>
> -bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
> - struct dc_state *dc_state,
> - struct dsc_mst_fairness_vars *vars)
> +int compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
> + struct dc_state *dc_state,
> + struct dsc_mst_fairness_vars *vars)
> {
> int i, j;
> struct dc_stream_state *stream;
> bool computed_streams[MAX_PIPES];
> struct amdgpu_dm_connector *aconnector;
> int link_vars_start_index = 0;
> + int ret = 0;
>
> for (i = 0; i < dc_state->stream_count; i++)
> computed_streams[i] = false;
> @@ -1118,17 +1123,19 @@ bool compute_mst_dsc_configs_for_state(struct
> drm_atomic_state *state,
> continue;
>
> if (dcn20_remove_stream_from_ctx(stream->ctx->dc,
> dc_state, stream) != DC_OK)
> - return false;
> + return -EINVAL;
>
> if (!is_dsc_need_re_compute(state, dc_state, stream->link))
> continue;
>
> mutex_lock(&aconnector->mst_mgr.lock);
> - if (!compute_mst_dsc_configs_for_link(state, dc_state,
> stream->link, vars,
> - &aconnector->mst_mgr,
> - &link_vars_start_index)) {
> +
> + ret = compute_mst_dsc_configs_for_link(state, dc_state,
> stream->link, vars,
> + &aconnector->mst_mgr,
> + &link_vars_start_index);
> + if (ret != 0) {
> mutex_unlock(&aconnector->mst_mgr.lock);
> - return false;
> + return ret;
> }
> mutex_unlock(&aconnector->mst_mgr.lock);
>
> @@ -1143,22 +1150,22 @@ bool compute_mst_dsc_configs_for_state(struct
> drm_atomic_state *state,
>
> if (stream->timing.flags.DSC == 1)
> if (dc_stream_add_dsc_to_resource(stream->ctx-
> >dc, dc_state, stream) != DC_OK)
> - return false;
> + return -EINVAL;
> }
>
> - return true;
> + return ret;
> }
>
> -static bool
> - pre_compute_mst_dsc_configs_for_state(struct drm_atomic_state
> *state,
> - struct dc_state *dc_state,
> - struct dsc_mst_fairness_vars
> *vars)
> +static int pre_compute_mst_dsc_configs_for_state(struct
> drm_atomic_state *state,
> + struct dc_state *dc_state,
> + struct dsc_mst_fairness_vars
> *vars)
> {
> int i, j;
> struct dc_stream_state *stream;
> bool computed_streams[MAX_PIPES];
> struct amdgpu_dm_connector *aconnector;
> int link_vars_start_index = 0;
> + int ret;
>
> for (i = 0; i < dc_state->stream_count; i++)
> computed_streams[i] = false;
> @@ -1184,13 +1191,12 @@ static bool
> continue;
>
> mutex_lock(&aconnector->mst_mgr.lock);
> - if (!compute_mst_dsc_configs_for_link(state, dc_state,
> stream->link, vars,
> - &aconnector->mst_mgr,
> - &link_vars_start_index)) {
> - mutex_unlock(&aconnector->mst_mgr.lock);
> - return false;
> - }
> + ret = compute_mst_dsc_configs_for_link(state, dc_state,
> stream->link, vars,
> + &aconnector->mst_mgr,
> + &link_vars_start_index);
> mutex_unlock(&aconnector->mst_mgr.lock);
> + if (ret != 0)
> + return ret;
>
> for (j = 0; j < dc_state->stream_count; j++) {
> if (dc_state->streams[j]->link == stream->link) @@ -
> 1198,7 +1204,7 @@ static bool
> }
> }
>
> - return true;
> + return ret;
> }
>
> static int find_crtc_index_in_state_by_stream(struct drm_atomic_state
> *state, @@ -1253,9 +1259,9 @@ static bool
> is_dsc_precompute_needed(struct drm_atomic_state *state)
> return ret;
> }
>
> -bool pre_validate_dsc(struct drm_atomic_state *state,
> - struct dm_atomic_state **dm_state_ptr,
> - struct dsc_mst_fairness_vars *vars)
> +int pre_validate_dsc(struct drm_atomic_state *state,
> + struct dm_atomic_state **dm_state_ptr,
> + struct dsc_mst_fairness_vars *vars)
> {
> int i;
> struct dm_atomic_state *dm_state;
> @@ -1264,11 +1270,12 @@ bool pre_validate_dsc(struct drm_atomic_state
> *state,
>
> if (!is_dsc_precompute_needed(state)) {
> DRM_INFO_ONCE("DSC precompute is not needed.\n");
> - return true;
> + return 0;
> }
> - if (dm_atomic_get_state(state, dm_state_ptr)) {
> + ret = dm_atomic_get_state(state, dm_state_ptr);
> + if (ret != 0) {
> DRM_INFO_ONCE("dm_atomic_get_state() failed\n");
> - return false;
> + return ret;
> }
> dm_state = *dm_state_ptr;
>
> @@ -1280,7 +1287,7 @@ bool pre_validate_dsc(struct drm_atomic_state
> *state,
>
> local_dc_state = kmemdup(dm_state->context, sizeof(struct
> dc_state), GFP_KERNEL);
> if (!local_dc_state)
> - return false;
> + return -ENOMEM;
>
> for (i = 0; i < local_dc_state->stream_count; i++) {
> struct dc_stream_state *stream = dm_state->context-
> >streams[i]; @@ -1316,9 +1323,9 @@ bool pre_validate_dsc(struct
> drm_atomic_state *state,
> if (ret != 0)
> goto clean_exit;
>
> - if (!pre_compute_mst_dsc_configs_for_state(state, local_dc_state,
> vars)) {
> + ret = pre_compute_mst_dsc_configs_for_state(state, local_dc_state,
> vars);
> + if (ret != 0) {
>
> DRM_INFO_ONCE("pre_compute_mst_dsc_configs_for_state()
> failed\n");
> - ret = -EINVAL;
> goto clean_exit;
> }
>
> @@ -1349,7 +1356,7 @@ bool pre_validate_dsc(struct drm_atomic_state
> *state,
>
> kfree(local_dc_state);
>
> - return (ret == 0);
> + return ret;
> }
>
> static unsigned int kbps_from_pbn(unsigned int pbn) diff --git
> a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h
> b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h
> index b92a7c5671aa2..97fd70df531bf 100644
> --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h
> +++
> b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h
> @@ -53,15 +53,15 @@ struct dsc_mst_fairness_vars {
> struct amdgpu_dm_connector *aconnector; };
>
> -bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
> - struct dc_state *dc_state,
> - struct dsc_mst_fairness_vars *vars);
> +int compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
> + struct dc_state *dc_state,
> + struct dsc_mst_fairness_vars *vars);
>
> bool needs_dsc_aux_workaround(struct dc_link *link);
>
> -bool pre_validate_dsc(struct drm_atomic_state *state,
> - struct dm_atomic_state **dm_state_ptr,
> - struct dsc_mst_fairness_vars *vars);
> +int pre_validate_dsc(struct drm_atomic_state *state,
> + struct dm_atomic_state **dm_state_ptr,
> + struct dsc_mst_fairness_vars *vars);
>
> enum dc_status dm_dp_mst_is_port_support_mode(
> struct amdgpu_dm_connector *aconnector,
> --
> 2.37.3
---
Regards,
Wayne Lin
