[PATCH AUTOSEL 6.13 05/37] drm/amd/display: Update Cursor request mode to the beginning prefetch always

[PATCH AUTOSEL 6.13 05/37] drm/amd/display: Update Cursor request mode to the beginning prefetch always

[Sasha Levin]

From: Zhikai Zhai <zhikai.zhai@amd.com>

[ Upstream commit 4a4077b4b63a8404efd6d37fc2926f03fb25bace ]

[Why]
The double buffer cursor registers is updated by the cursor
vupdate event. There is a gap between vupdate and cursor data
fetch if cursor fetch data reletive to cursor position.
Cursor corruption will happen if we update the cursor surface
in this gap.

[How]
Modify the cursor request mode to the beginning prefetch always
and avoid wraparound calculation issues.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Zhikai Zhai <zhikai.zhai@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../amd/display/dc/hubp/dcn31/dcn31_hubp.c    |  2 +-
 .../amd/display/dc/hwss/dcn10/dcn10_hwseq.c   | 22 ++++++++-----------
 2 files changed, 10 insertions(+), 14 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn31/dcn31_hubp.c b/drivers/gpu/drm/amd/display/dc/hubp/dcn31/dcn31_hubp.c
index a65a0ddee6467..c671908ba7d06 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn31/dcn31_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn31/dcn31_hubp.c
@@ -44,7 +44,7 @@ void hubp31_set_unbounded_requesting(struct hubp *hubp, bool enable)
 	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
 
 	REG_UPDATE(DCHUBP_CNTL, HUBP_UNBOUNDED_REQ_MODE, enable);
-	REG_UPDATE(CURSOR_CONTROL, CURSOR_REQ_MODE, enable);
+	REG_UPDATE(CURSOR_CONTROL, CURSOR_REQ_MODE, 1);
 }
 
 void hubp31_soft_reset(struct hubp *hubp, bool reset)
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
index 44e405e9bc971..13f9e9b439f6a 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
@@ -1992,20 +1992,11 @@ static void delay_cursor_until_vupdate(struct dc *dc, struct pipe_ctx *pipe_ctx)
 	dc->hwss.get_position(&pipe_ctx, 1, &position);
 	vpos = position.vertical_count;
 
-	/* Avoid wraparound calculation issues */
-	vupdate_start += stream->timing.v_total;
-	vupdate_end += stream->timing.v_total;
-	vpos += stream->timing.v_total;
-
 	if (vpos <= vupdate_start) {
 		/* VPOS is in VACTIVE or back porch. */
 		lines_to_vupdate = vupdate_start - vpos;
-	} else if (vpos > vupdate_end) {
-		/* VPOS is in the front porch. */
-		return;
 	} else {
-		/* VPOS is in VUPDATE. */
-		lines_to_vupdate = 0;
+		lines_to_vupdate = stream->timing.v_total - vpos + vupdate_start;
 	}
 
 	/* Calculate time until VUPDATE in microseconds. */
@@ -2013,13 +2004,18 @@ static void delay_cursor_until_vupdate(struct dc *dc, struct pipe_ctx *pipe_ctx)
 		stream->timing.h_total * 10000u / stream->timing.pix_clk_100hz;
 	us_to_vupdate = lines_to_vupdate * us_per_line;
 
+	/* Stall out until the cursor update completes. */
+	if (vupdate_end < vupdate_start)
+		vupdate_end += stream->timing.v_total;
+
+	/* Position is in the range of vupdate start and end*/
+	if (lines_to_vupdate > stream->timing.v_total - vupdate_end + vupdate_start)
+		us_to_vupdate = 0;
+
 	/* 70 us is a conservative estimate of cursor update time*/
 	if (us_to_vupdate > 70)
 		return;
 
-	/* Stall out until the cursor update completes. */
-	if (vupdate_end < vupdate_start)
-		vupdate_end += stream->timing.v_total;
 	us_vupdate = (vupdate_end - vupdate_start + 1) * us_per_line;
 	udelay(us_to_vupdate + us_vupdate);
 }
-- 
2.39.5

[PATCH AUTOSEL 6.13 06/37] drm/amd/display: add workaround flag to link to force FFE preset

[Sasha Levin]

From: Brendan Tam <Brendan.Tam@amd.com>

[ Upstream commit 51d1b338541dea83fec8e6f95d3e46fa469a73a8 ]

[Why]
There have been instances of some monitors being unable to link train on
their reported link speed using their selected FFE preset. If a different
FFE preset is found that has a higher rate of success during link training
this workaround can be used to force its FFE preset.

[How]
A new link workaround flag is made called force_dp_ffe_preset. The flag is
checked in override_training_settings and will set lt_settings->ffe_preset
which is null if the flag is not set. The flag is then set in
override_lane_settings.

Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Brendan Tam <Brendan.Tam@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/dc.h                             | 2 ++
 .../gpu/drm/amd/display/dc/link/protocols/link_dp_training.c    | 2 ++
 2 files changed, 4 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index 08c5a315b3a67..1e1633b7771c5 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -1780,7 +1780,9 @@ struct dc_link {
 		bool dongle_mode_timing_override;
 		bool blank_stream_on_ocs_change;
 		bool read_dpcd204h_on_irq_hpd;
+		bool force_dp_ffe_preset;
 	} wa_flags;
+	union dc_dp_ffe_preset forced_dp_ffe_preset;
 	struct link_mst_stream_allocation_table mst_stream_alloc_table;
 
 	struct dc_link_status link_status;
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c
index 754c895e1bfbd..09024a8bac128 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c
@@ -736,6 +736,8 @@ void override_training_settings(
 		lt_settings->pre_emphasis = overrides->pre_emphasis;
 	if (overrides->post_cursor2 != NULL)
 		lt_settings->post_cursor2 = overrides->post_cursor2;
+	if (link->wa_flags.force_dp_ffe_preset && !dp_is_lttpr_present(link))
+		lt_settings->ffe_preset = &link->forced_dp_ffe_preset;
 	if (overrides->ffe_preset != NULL)
 		lt_settings->ffe_preset = overrides->ffe_preset;
 	/* Override HW lane settings with BIOS forced values if present */
-- 
2.39.5

[PATCH AUTOSEL 6.13 07/37] drm/amdgpu: Unlocked unmap only clear page table leaves

[Sasha Levin]

From: Philip Yang <Philip.Yang@amd.com>

[ Upstream commit 23b645231eeffdaf44021debac881d2f26824150 ]

SVM migration unmap pages from GPU and then update mapping to GPU to
recover page fault. Currently unmap clears the PDE entry for range
length >= huge page and free PTB bo, update mapping to alloc new PT bo.
There is race bug that the freed entry bo maybe still on the pt_free
list, reused when updating mapping and then freed, leave invalid PDE
entry and cause GPU page fault.

By setting the update to clear only one PDE entry or clear PTB, to
avoid unmap to free PTE bo. This fixes the race bug and improve the
unmap and map to GPU performance. Update mapping to huge page will
still free the PTB bo.

With this change, the vm->pt_freed list and work is not needed. Add
WARN_ON(unlocked) in amdgpu_vm_pt_free_dfs to catch if unmap to free the
PTB.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c    |  4 ---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h    |  4 ---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c | 43 +++++++----------------
 3 files changed, 13 insertions(+), 38 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index c9c48b782ec1b..48b2c0b3b315a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -2440,8 +2440,6 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	spin_lock_init(&vm->status_lock);
 	INIT_LIST_HEAD(&vm->freed);
 	INIT_LIST_HEAD(&vm->done);
-	INIT_LIST_HEAD(&vm->pt_freed);
-	INIT_WORK(&vm->pt_free_work, amdgpu_vm_pt_free_work);
 	INIT_KFIFO(vm->faults);
 
 	r = amdgpu_vm_init_entities(adev, vm);
@@ -2613,8 +2611,6 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 
 	amdgpu_amdkfd_gpuvm_destroy_cb(adev, vm);
 
-	flush_work(&vm->pt_free_work);
-
 	root = amdgpu_bo_ref(vm->root.bo);
 	amdgpu_bo_reserve(root, true);
 	amdgpu_vm_put_task_info(vm->task_info);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 5d119ac26c4fe..160889e5e64dd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -369,10 +369,6 @@ struct amdgpu_vm {
 	/* BOs which are invalidated, has been updated in the PTs */
 	struct list_head        done;
 
-	/* PT BOs scheduled to free and fill with zero if vm_resv is not hold */
-	struct list_head	pt_freed;
-	struct work_struct	pt_free_work;
-
 	/* contains the page directory */
 	struct amdgpu_vm_bo_base     root;
 	struct dma_fence	*last_update;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
index f78a0434a48fa..54ae0e9bc6d77 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
@@ -546,27 +546,6 @@ static void amdgpu_vm_pt_free(struct amdgpu_vm_bo_base *entry)
 	amdgpu_bo_unref(&entry->bo);
 }
 
-void amdgpu_vm_pt_free_work(struct work_struct *work)
-{
-	struct amdgpu_vm_bo_base *entry, *next;
-	struct amdgpu_vm *vm;
-	LIST_HEAD(pt_freed);
-
-	vm = container_of(work, struct amdgpu_vm, pt_free_work);
-
-	spin_lock(&vm->status_lock);
-	list_splice_init(&vm->pt_freed, &pt_freed);
-	spin_unlock(&vm->status_lock);
-
-	/* flush_work in amdgpu_vm_fini ensure vm->root.bo is valid. */
-	amdgpu_bo_reserve(vm->root.bo, true);
-
-	list_for_each_entry_safe(entry, next, &pt_freed, vm_status)
-		amdgpu_vm_pt_free(entry);
-
-	amdgpu_bo_unreserve(vm->root.bo);
-}
-
 /**
  * amdgpu_vm_pt_free_list - free PD/PT levels
  *
@@ -579,19 +558,15 @@ void amdgpu_vm_pt_free_list(struct amdgpu_device *adev,
 			    struct amdgpu_vm_update_params *params)
 {
 	struct amdgpu_vm_bo_base *entry, *next;
-	struct amdgpu_vm *vm = params->vm;
 	bool unlocked = params->unlocked;
 
 	if (list_empty(&params->tlb_flush_waitlist))
 		return;
 
-	if (unlocked) {
-		spin_lock(&vm->status_lock);
-		list_splice_init(&params->tlb_flush_waitlist, &vm->pt_freed);
-		spin_unlock(&vm->status_lock);
-		schedule_work(&vm->pt_free_work);
-		return;
-	}
+	/*
+	 * unlocked unmap clear page table leaves, warning to free the page entry.
+	 */
+	WARN_ON(unlocked);
 
 	list_for_each_entry_safe(entry, next, &params->tlb_flush_waitlist, vm_status)
 		amdgpu_vm_pt_free(entry);
@@ -899,7 +874,15 @@ int amdgpu_vm_ptes_update(struct amdgpu_vm_update_params *params,
 		incr = (uint64_t)AMDGPU_GPU_PAGE_SIZE << shift;
 		mask = amdgpu_vm_pt_entries_mask(adev, cursor.level);
 		pe_start = ((cursor.pfn >> shift) & mask) * 8;
-		entry_end = ((uint64_t)mask + 1) << shift;
+
+		if (cursor.level < AMDGPU_VM_PTB && params->unlocked)
+			/*
+			 * MMU notifier callback unlocked unmap huge page, leave is PDE entry,
+			 * only clear one entry. Next entry search again for PDE or PTE leave.
+			 */
+			entry_end = 1ULL << shift;
+		else
+			entry_end = ((uint64_t)mask + 1) << shift;
 		entry_end += cursor.pfn & ~(entry_end - 1);
 		entry_end = min(entry_end, end);
 
-- 
2.39.5

[PATCH AUTOSEL 6.13 15/37] drm/amd/display: Update FIXED_VS Link Rate Toggle Workaround Usage

[Sasha Levin]

From: Michael Strauss <michael.strauss@amd.com>

[ Upstream commit 7c6518c1c73199a230b5fc55ddfed3e5b9dc3290 ]

[WHY]
Previously the 128b/132b LTTPR support DPCD field was used to decide if
FIXED_VS training sequence required a rate toggle before initiating LT.

When running DP2.1 4.9.x.x compliance tests, emulated LTTPRs can report
no-128b/132b support which is then forwarded by the FIXED_VS retimer.
As a result this test exposes the rate toggle again, erroneously causing
failures as certain compliance sinks don't expect this behaviour.

[HOW]
Add new DPCD register defines/reads to read LTTPR IEEE OUI and device ID.

Decide whether to perform the rate toggle based on the LTTPR's IEEE OUI
which guarantees that we only perform the toggle on affected retimers.

Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Michael Strauss <michael.strauss@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/dc_dp_types.h         |  8 ++++++++
 .../display/dc/link/protocols/link_dp_capability.c   | 12 ++++++++++--
 .../protocols/link_dp_training_fixed_vs_pe_retimer.c |  3 ++-
 3 files changed, 20 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
index 8dd6eb044829a..aecaf06ba9990 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
@@ -1104,6 +1104,8 @@ struct dc_lttpr_caps {
 	union dp_main_link_channel_coding_lttpr_cap main_link_channel_coding;
 	union dp_128b_132b_supported_lttpr_link_rates supported_128b_132b_rates;
 	uint8_t aux_rd_interval[MAX_REPEATER_CNT - 1];
+	uint8_t lttpr_ieee_oui[3];
+	uint8_t lttpr_device_id[6];
 };
 
 struct dc_dongle_dfp_cap_ext {
@@ -1363,6 +1365,12 @@ struct dp_trace {
 #ifndef DP_BRANCH_VENDOR_SPECIFIC_START
 #define DP_BRANCH_VENDOR_SPECIFIC_START     0x50C
 #endif
+#ifndef DP_LTTPR_IEEE_OUI
+#define DP_LTTPR_IEEE_OUI 0xF003D
+#endif
+#ifndef DP_LTTPR_DEVICE_ID
+#define DP_LTTPR_DEVICE_ID 0xF0040
+#endif
 /** USB4 DPCD BW Allocation Registers Chapter 10.7 **/
 #ifndef DP_TUNNELING_CAPABILITIES
 #define DP_TUNNELING_CAPABILITIES			0xE000D /* 1.4a */
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
index 9dabaf682171d..d5d1f5ffd4fd8 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
@@ -1568,10 +1568,18 @@ enum dc_status dp_retrieve_lttpr_cap(struct dc_link *link)
 	/* Attempt to train in LTTPR transparent mode if repeater count exceeds 8. */
 	is_lttpr_present = dp_is_lttpr_present(link);
 
-	if (is_lttpr_present)
+	DC_LOG_DC("is_lttpr_present = %d\n", is_lttpr_present);
+
+	if (is_lttpr_present) {
 		CONN_DATA_DETECT(link, lttpr_dpcd_data, sizeof(lttpr_dpcd_data), "LTTPR Caps: ");
 
-	DC_LOG_DC("is_lttpr_present = %d\n", is_lttpr_present);
+		core_link_read_dpcd(link, DP_LTTPR_IEEE_OUI, link->dpcd_caps.lttpr_caps.lttpr_ieee_oui, sizeof(link->dpcd_caps.lttpr_caps.lttpr_ieee_oui));
+		CONN_DATA_DETECT(link, link->dpcd_caps.lttpr_caps.lttpr_ieee_oui, sizeof(link->dpcd_caps.lttpr_caps.lttpr_ieee_oui), "LTTPR IEEE OUI: ");
+
+		core_link_read_dpcd(link, DP_LTTPR_DEVICE_ID, link->dpcd_caps.lttpr_caps.lttpr_device_id, sizeof(link->dpcd_caps.lttpr_caps.lttpr_device_id));
+		CONN_DATA_DETECT(link, link->dpcd_caps.lttpr_caps.lttpr_device_id, sizeof(link->dpcd_caps.lttpr_caps.lttpr_device_id), "LTTPR Device ID: ");
+	}
+
 	return status;
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c
index ccf8096dde290..ce174ce5579c0 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c
@@ -270,7 +270,8 @@ enum link_training_result dp_perform_fixed_vs_pe_training_sequence(
 
 	rate = get_dpcd_link_rate(&lt_settings->link_settings);
 
-	if (!link->dpcd_caps.lttpr_caps.main_link_channel_coding.bits.DP_128b_132b_SUPPORTED) {
+	// Only perform toggle if FIXED_VS LTTPR reports no IEEE OUI
+	if (memcmp("\x0,\x0,\x0", &link->dpcd_caps.lttpr_caps.lttpr_ieee_oui[0], 3) == 0) {
 		/* Vendor specific: Toggle link rate */
 		toggle_rate = (rate == 0x6) ? 0xA : 0x6;
 
-- 
2.39.5

[PATCH AUTOSEL 6.13 16/37] drm/amd/display: stop DML2 from removing pipes based on planes

[Sasha Levin]

From: Mike Katsnelson <mike.katsnelson@amd.com>

[ Upstream commit 8adeff83a3b07fa6d0958ed51e1b38ba7469e448 ]

[Why]
Transitioning from low to high resolutions at high refresh rates caused grey corruption.
During the transition state, there is a period where plane size is based on low resultion
state and ODM slices are based on high resoultion state, causing the entire plane to be
contained in one ODM slice. DML2 would turn off the pipe for the ODM slice with no plane,
causing an underflow since the pixel rate for the higher resolution cannot be supported on
one pipe. This change stops DML2 from turning off pipes that are mapped to an ODM slice
with no plane. This is possible to do without negative consequences because pipes can now
take the minimum viewport and draw with zero recout size, removing the need to have the
pipe turned off.

[How]
In map_pipes_from_plane(), remove "check" that skips ODM slices that are not covered by
the plane. This prevents the pipes for those ODM slices from being freed.

Reviewed-by: Ovidiu Bunea <ovidiu.bunea@amd.com>
Signed-off-by: Mike Katsnelson <mike.katsnelson@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../display/dc/dml2/dml2_dc_resource_mgmt.c   | 26 -------------------
 1 file changed, 26 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c
index 1ed21c1b86a5b..a966abd407881 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c
@@ -532,26 +532,6 @@ static void calculate_odm_slices(const struct dc_stream_state *stream, unsigned
 	odm_slice_end_x[odm_factor - 1] = stream->src.width - 1;
 }
 
-static bool is_plane_in_odm_slice(const struct dc_plane_state *plane, unsigned int slice_index, unsigned int *odm_slice_end_x, unsigned int num_slices)
-{
-	unsigned int slice_start_x, slice_end_x;
-
-	if (slice_index == 0)
-		slice_start_x = 0;
-	else
-		slice_start_x = odm_slice_end_x[slice_index - 1] + 1;
-
-	slice_end_x = odm_slice_end_x[slice_index];
-
-	if (plane->clip_rect.x + plane->clip_rect.width < slice_start_x)
-		return false;
-
-	if (plane->clip_rect.x > slice_end_x)
-		return false;
-
-	return true;
-}
-
 static void add_odm_slice_to_odm_tree(struct dml2_context *ctx,
 		struct dc_state *state,
 		struct dc_pipe_mapping_scratch *scratch,
@@ -791,12 +771,6 @@ static void map_pipes_for_plane(struct dml2_context *ctx, struct dc_state *state
 	sort_pipes_for_splitting(&scratch->pipe_pool);
 
 	for (odm_slice_index = 0; odm_slice_index < scratch->odm_info.odm_factor; odm_slice_index++) {
-		// We build the tree for one ODM slice at a time.
-		// Each ODM slice shares a common OPP
-		if (!is_plane_in_odm_slice(plane, odm_slice_index, scratch->odm_info.odm_slice_end_x, scratch->odm_info.odm_factor)) {
-			continue;
-		}
-
 		// Now we have a list of all pipes to be used for this plane/stream, now setup the tree.
 		scratch->odm_info.next_higher_pipe_for_odm_slice[odm_slice_index] = add_plane_to_blend_tree(ctx, state,
 				plane,
-- 
2.39.5

[PATCH AUTOSEL 6.13 18/37] drm/amdkfd: clamp queue size to minimum

[Sasha Levin]

From: David Yat Sin <David.YatSin@amd.com>

[ Upstream commit e90711946b53590371ecce32e8fcc381a99d6333 ]

If queue size is less than minimum, clamp it to minimum to prevent
underflow when writing queue mqd.

Signed-off-by: David Yat Sin <David.YatSin@amd.com>
Reviewed-by: Jay Cornwall <jay.cornwall@amd.com>
Reviewed-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 10 ++++++++++
 include/uapi/linux/kfd_ioctl.h           |  2 ++
 2 files changed, 12 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index 065d878414591..33df35cab4679 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
@@ -212,6 +212,11 @@ static int set_queue_properties_from_user(struct queue_properties *q_properties,
 		return -EINVAL;
 	}
 
+	if (args->ring_size < KFD_MIN_QUEUE_RING_SIZE) {
+		args->ring_size = KFD_MIN_QUEUE_RING_SIZE;
+		pr_debug("Size lower. clamped to KFD_MIN_QUEUE_RING_SIZE");
+	}
+
 	if (!access_ok((const void __user *) args->read_pointer_address,
 			sizeof(uint32_t))) {
 		pr_err("Can't access read pointer\n");
@@ -461,6 +466,11 @@ static int kfd_ioctl_update_queue(struct file *filp, struct kfd_process *p,
 		return -EINVAL;
 	}
 
+	if (args->ring_size < KFD_MIN_QUEUE_RING_SIZE) {
+		args->ring_size = KFD_MIN_QUEUE_RING_SIZE;
+		pr_debug("Size lower. clamped to KFD_MIN_QUEUE_RING_SIZE");
+	}
+
 	properties.queue_address = args->ring_base_address;
 	properties.queue_size = args->ring_size;
 	properties.queue_percent = args->queue_percentage & 0xFF;
diff --git a/include/uapi/linux/kfd_ioctl.h b/include/uapi/linux/kfd_ioctl.h
index fa9f9846b88e4..b0160b09987c1 100644
--- a/include/uapi/linux/kfd_ioctl.h
+++ b/include/uapi/linux/kfd_ioctl.h
@@ -62,6 +62,8 @@ struct kfd_ioctl_get_version_args {
 #define KFD_MAX_QUEUE_PERCENTAGE	100
 #define KFD_MAX_QUEUE_PRIORITY		15
 
+#define KFD_MIN_QUEUE_RING_SIZE		1024
+
 struct kfd_ioctl_create_queue_args {
 	__u64 ring_base_address;	/* to KFD */
 	__u64 write_pointer_address;	/* from KFD */
-- 
2.39.5

[PATCH AUTOSEL 6.13 19/37] drm/amdkfd: Fix mode1 reset crash issue

[Sasha Levin]

From: Philip Yang <Philip.Yang@amd.com>

[ Upstream commit f0b4440cdc1807bb6ec3dce0d6de81170803569b ]

If HW scheduler hangs and mode1 reset is used to recover GPU, KFD signal
user space to abort the processes. After process abort exit, user queues
still use the GPU to access system memory before h/w is reset while KFD
cleanup worker free system memory and free VRAM.

There is use-after-free race bug that KFD allocate and reuse the freed
system memory, and user queue write to the same system memory to corrupt
the data structure and cause driver crash.

To fix this race, KFD cleanup worker terminate user queues, then flush
reset_domain wq to wait for any GPU ongoing reset complete, and then
free outstanding BOs.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdkfd/kfd_process.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
index edfe0b4788f44..211e268d10ab9 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -35,6 +35,7 @@
 #include <linux/pm_runtime.h>
 #include "amdgpu_amdkfd.h"
 #include "amdgpu.h"
+#include "amdgpu_reset.h"
 
 struct mm_struct;
 
@@ -1140,6 +1141,17 @@ static void kfd_process_remove_sysfs(struct kfd_process *p)
 	p->kobj = NULL;
 }
 
+/*
+ * If any GPU is ongoing reset, wait for reset complete.
+ */
+static void kfd_process_wait_gpu_reset_complete(struct kfd_process *p)
+{
+	int i;
+
+	for (i = 0; i < p->n_pdds; i++)
+		flush_workqueue(p->pdds[i]->dev->adev->reset_domain->wq);
+}
+
 /* No process locking is needed in this function, because the process
  * is not findable any more. We must assume that no other thread is
  * using it any more, otherwise we couldn't safely free the process
@@ -1154,6 +1166,11 @@ static void kfd_process_wq_release(struct work_struct *work)
 	kfd_process_dequeue_from_all_devices(p);
 	pqm_uninit(&p->pqm);
 
+	/*
+	 * If GPU in reset, user queues may still running, wait for reset complete.
+	 */
+	kfd_process_wait_gpu_reset_complete(p);
+
 	/* Signal the eviction fence after user mode queues are
 	 * destroyed. This allows any BOs to be freed without
 	 * triggering pointless evictions or waiting for fences.
-- 
2.39.5

[PATCH AUTOSEL 6.13 20/37] drm/amdkfd: Fix pqm_destroy_queue race with GPU reset

[Sasha Levin]

From: Philip Yang <Philip.Yang@amd.com>

[ Upstream commit 7919b4cad5545ed93778f11881ceee72e4dbed66 ]

If GPU in reset, destroy_queue return -EIO, pqm_destroy_queue should
delete the queue from process_queue_list and free the resource.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
index 1d538e874140c..6ca0efa2c3c97 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
@@ -529,7 +529,7 @@ int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid)
 			pr_err("Pasid 0x%x destroy queue %d failed, ret %d\n",
 				pqm->process->pasid,
 				pqn->q->properties.queue_id, retval);
-			if (retval != -ETIME)
+			if (retval != -ETIME && retval != -EIO)
 				goto err_destroy_queue;
 		}
 		kfd_procfs_del_queue(pqn->q);
-- 
2.39.5

[PATCH AUTOSEL 6.13 21/37] drm/amdkfd: debugfs hang_hws skip GPU with MES

[Sasha Levin]

From: Philip Yang <Philip.Yang@amd.com>

[ Upstream commit fe9d0061c413f8fb8c529b18b592b04170850ded ]

debugfs hang_hws is used by GPU reset test with HWS, for MES this crash
the kernel with NULL pointer access because dqm->packet_mgr is not setup
for MES path.

Skip GPU with MES for now, MES hang_hws debugfs interface will be
supported later.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Kent Russell <kent.russell@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdkfd/kfd_device.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index 35caa71f317dc..463771b7dd4a7 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -1519,6 +1519,11 @@ int kfd_debugfs_hang_hws(struct kfd_node *dev)
 		return -EINVAL;
 	}
 
+	if (dev->kfd->shared_resources.enable_mes) {
+		dev_err(dev->adev->dev, "Inducing MES hang is not supported\n");
+		return -EINVAL;
+	}
+
 	return dqm_debugfs_hang_hws(dev->dqm);
 }
 
-- 
2.39.5

[PATCH AUTOSEL 6.13 26/37] drm/amdgpu: handle amdgpu_cgs_create_device() errors in amd_powerplay_create()

[Sasha Levin]

From: Wentao Liang <vulab@iscas.ac.cn>

[ Upstream commit 1435e895d4fc967d64e9f5bf81e992ac32f5ac76 ]

Add error handling to propagate amdgpu_cgs_create_device() failures
to the caller. When amdgpu_cgs_create_device() fails, release hwmgr
and return -ENOMEM to prevent null pointer dereference.

[v1]->[v2]: Change error code from -EINVAL to -ENOMEM. Free hwmgr.

Signed-off-by: Wentao Liang <vulab@iscas.ac.cn>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/pm/powerplay/amd_powerplay.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/gpu/drm/amd/pm/powerplay/amd_powerplay.c b/drivers/gpu/drm/amd/pm/powerplay/amd_powerplay.c
index 26624a716fc60..f8434158a4022 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/amd_powerplay.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/amd_powerplay.c
@@ -51,6 +51,11 @@ static int amd_powerplay_create(struct amdgpu_device *adev)
 	hwmgr->adev = adev;
 	hwmgr->not_vf = !amdgpu_sriov_vf(adev);
 	hwmgr->device = amdgpu_cgs_create_device(adev);
+	if (!hwmgr->device) {
+		kfree(hwmgr);
+		return -ENOMEM;
+	}
+
 	mutex_init(&hwmgr->msg_lock);
 	hwmgr->chip_family = adev->family;
 	hwmgr->chip_id = adev->asic_type;
-- 
2.39.5

[PATCH AUTOSEL 6.13 27/37] drm/amd/display: Prevent VStartup Overflow

[Sasha Levin]

From: Ryan Seto <ryanseto@amd.com>

[ Upstream commit 29c1c20496a7a9bafe2bc2f833d69aa52e0f2c2d ]

[Why]
For some VR headsets with large blanks, it's possible
to overflow the OTG_VSTARTUP_PARAM:VSTARTUP_START
register. This can lead to incorrect DML calculations
and underflow downstream.

[How]
Min the calcualted max_vstartup_lines with the max
value of the register.

Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Ryan Seto <ryanseto@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c  | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
index ecfa3c898e09d..4285cb7fcbe71 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
@@ -12,6 +12,7 @@
 #define DML2_MAX_FMT_420_BUFFER_WIDTH 4096
 #define DML_MAX_NUM_OF_SLICES_PER_DSC 4
 #define ALLOW_SDPIF_RATE_LIMIT_PRE_CSTATE
+#define DML_MAX_VSTARTUP_START 1023
 
 const char *dml2_core_internal_bw_type_str(enum dml2_core_internal_bw_type bw_type)
 {
@@ -3648,6 +3649,7 @@ static unsigned int CalculateMaxVStartup(
 	dml2_printf("DML::%s: vblank_avail = %u\n", __func__, vblank_avail);
 	dml2_printf("DML::%s: max_vstartup_lines = %u\n", __func__, max_vstartup_lines);
 #endif
+	max_vstartup_lines = (unsigned int)math_min2(max_vstartup_lines, DML_MAX_VSTARTUP_START);
 	return max_vstartup_lines;
 }
 
-- 
2.39.5

[PATCH AUTOSEL 6.13 29/37] drm/amdgpu: Fix the race condition for draining retry fault

[Sasha Levin]

From: Emily Deng <Emily.Deng@amd.com>

[ Upstream commit f844732e3ad9c4b78df7436232949b8d2096d1a6 ]

Issue:
In the scenario where svm_range_restore_pages is called, but
svm->checkpoint_ts has not been set and the retry fault has not been
drained, svm_range_unmap_from_cpu is triggered and calls svm_range_free.
Meanwhile, svm_range_restore_pages continues execution and reaches
svm_range_from_addr. This results in a "failed to find prange..." error,
 causing the page recovery to fail.

How to fix:
Move the timestamp check code under the protection of svm->lock.

v2:
Make sure all right locks are released before go out.

v3:
Directly goto out_unlock_svms, and return -EAGAIN.

v4:
Refine code.

Signed-off-by: Emily Deng <Emily.Deng@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 31 +++++++++++++++-------------
 1 file changed, 17 insertions(+), 14 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index 1b04d54f9a69c..4eeea51c9a0eb 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -2992,19 +2992,6 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid,
 		goto out;
 	}
 
-	/* check if this page fault time stamp is before svms->checkpoint_ts */
-	if (svms->checkpoint_ts[gpuidx] != 0) {
-		if (amdgpu_ih_ts_after(ts,  svms->checkpoint_ts[gpuidx])) {
-			pr_debug("draining retry fault, drop fault 0x%llx\n", addr);
-			r = 0;
-			goto out;
-		} else
-			/* ts is after svms->checkpoint_ts now, reset svms->checkpoint_ts
-			 * to zero to avoid following ts wrap around give wrong comparing
-			 */
-			svms->checkpoint_ts[gpuidx] = 0;
-	}
-
 	if (!p->xnack_enabled) {
 		pr_debug("XNACK not enabled for pasid 0x%x\n", pasid);
 		r = -EFAULT;
@@ -3024,6 +3011,21 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid,
 	mmap_read_lock(mm);
 retry_write_locked:
 	mutex_lock(&svms->lock);
+
+	/* check if this page fault time stamp is before svms->checkpoint_ts */
+	if (svms->checkpoint_ts[gpuidx] != 0) {
+		if (amdgpu_ih_ts_after(ts,  svms->checkpoint_ts[gpuidx])) {
+			pr_debug("draining retry fault, drop fault 0x%llx\n", addr);
+			r = -EAGAIN;
+			goto out_unlock_svms;
+		} else {
+			/* ts is after svms->checkpoint_ts now, reset svms->checkpoint_ts
+			 * to zero to avoid following ts wrap around give wrong comparing
+			 */
+			svms->checkpoint_ts[gpuidx] = 0;
+		}
+	}
+
 	prange = svm_range_from_addr(svms, addr, NULL);
 	if (!prange) {
 		pr_debug("failed to find prange svms 0x%p address [0x%llx]\n",
@@ -3149,7 +3151,8 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid,
 	mutex_unlock(&svms->lock);
 	mmap_read_unlock(mm);
 
-	svm_range_count_fault(node, p, gpuidx);
+	if (r != -EAGAIN)
+		svm_range_count_fault(node, p, gpuidx);
 
 	mmput(mm);
 out:
-- 
2.39.5

[PATCH AUTOSEL 6.13 32/37] drm/amdgpu: grab an additional reference on the gang fence v2

[Sasha Levin]

From: Christian König <christian.koenig@amd.com>

[ Upstream commit 0d9a95099dcb05b5f4719c830d15bf4fdcad0dc2 ]

We keep the gang submission fence around in adev, make sure that it
stays alive.

v2: fix memory leak on retry

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 3780d50fd3ae8..61b3f676a5df0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -6707,18 +6707,26 @@ struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
 {
 	struct dma_fence *old = NULL;
 
+	dma_fence_get(gang);
 	do {
 		dma_fence_put(old);
 		old = amdgpu_device_get_gang(adev);
 		if (old == gang)
 			break;
 
-		if (!dma_fence_is_signaled(old))
+		if (!dma_fence_is_signaled(old)) {
+			dma_fence_put(gang);
 			return old;
+		}
 
 	} while (cmpxchg((struct dma_fence __force **)&adev->gang_submit,
 			 old, gang) != old);
 
+	/*
+	 * Drop it once for the exchanged reference in adev and once for the
+	 * thread local reference acquired in amdgpu_device_get_gang().
+	 */
+	dma_fence_put(old);
 	dma_fence_put(old);
 	return NULL;
 }
-- 
2.39.5