[PATCH AUTOSEL 5.16 003/109] drm/amd/display: Add signal type check when verify stream backends same

[PATCH AUTOSEL 5.16 003/109] drm/amd/display: Add signal type check when verify stream backends same

[Sasha Levin]

From: Dale Zhao <dale.zhao@amd.com>

[ Upstream commit 047db281c026de5971cedb5bb486aa29bd16a39d ]

[Why]
For allow eDP hot-plug feature, the stream signal may change to VIRTUAL
when plug-out and back to eDP when plug-in. OS will still setPathMode
with same timing for each plugging, but eDP gets no stream update as we
don't check signal type changing back as keeping it VIRTUAL. It's also
unsafe for future cases that stream signal is switched with same timing.

[How]
Check stream signal type change include previous HDMI signal case.

Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: Dale Zhao <dale.zhao@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/core/dc_resource.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
index 6b066ceab412..3aa2040d2475 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
@@ -1640,6 +1640,9 @@ static bool are_stream_backends_same(
 	if (is_timing_changed(stream_a, stream_b))
 		return false;
 
+	if (stream_a->signal != stream_b->signal)
+		return false;
+
 	if (stream_a->dpms_off != stream_b->dpms_off)
 		return false;
 
-- 
2.34.1

[PATCH AUTOSEL 5.16 007/109] drm/amd/amdgpu/amdgpu_cs: fix refcount leak of a dma_fence obj

[Sasha Levin]

From: Xin Xiong <xiongx18@fudan.edu.cn>

[ Upstream commit dfced44f122c500004a48ecc8db516bb6a295a1b ]

This issue takes place in an error path in
amdgpu_cs_fence_to_handle_ioctl(). When `info->in.what` falls into
default case, the function simply returns -EINVAL, forgetting to
decrement the reference count of a dma_fence obj, which is bumped
earlier by amdgpu_cs_get_fence(). This may result in reference count
leaks.

Fix it by decreasing the refcount of specific object before returning
the error code.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Xin Xiong <xiongx18@fudan.edu.cn>
Signed-off-by: Xin Tan <tanxin.ctf@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 0311d799a010..894869789041 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -1510,6 +1510,7 @@ int amdgpu_cs_fence_to_handle_ioctl(struct drm_device *dev, void *data,
 		return 0;
 
 	default:
+		dma_fence_put(fence);
 		return -EINVAL;
 	}
 }
-- 
2.34.1

[PATCH AUTOSEL 5.16 008/109] drm/amd/display: Fix memory leak

[Sasha Levin]

From: Yongzhi Liu <lyz_cs@pku.edu.cn>

[ Upstream commit 5d5c6dba2b43e28845d7d7ed32a36802329a5f52 ]

[why]
Resource release is needed on the error handling path
to prevent memory leak.

[how]
Fix this by adding kfree on the error handling path.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Yongzhi Liu <lyz_cs@pku.edu.cn>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../amd/display/amdgpu_dm/amdgpu_dm_debugfs.c | 80 ++++++++++++++-----
 1 file changed, 60 insertions(+), 20 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
index f4e829ec8e10..ab58bcb11677 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
@@ -227,8 +227,10 @@ static ssize_t dp_link_settings_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -389,8 +391,10 @@ static ssize_t dp_phy_settings_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user((*(rd_buf + result)), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -1317,8 +1321,10 @@ static ssize_t dp_dsc_clock_en_read(struct file *f, char __user *buf,
 				break;
 	}
 
-	if (!pipe_ctx)
+	if (!pipe_ctx) {
+		kfree(rd_buf);
 		return -ENXIO;
+	}
 
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
@@ -1334,8 +1340,10 @@ static ssize_t dp_dsc_clock_en_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -1504,8 +1512,10 @@ static ssize_t dp_dsc_slice_width_read(struct file *f, char __user *buf,
 				break;
 	}
 
-	if (!pipe_ctx)
+	if (!pipe_ctx) {
+		kfree(rd_buf);
 		return -ENXIO;
+	}
 
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
@@ -1521,8 +1531,10 @@ static ssize_t dp_dsc_slice_width_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -1689,8 +1701,10 @@ static ssize_t dp_dsc_slice_height_read(struct file *f, char __user *buf,
 				break;
 	}
 
-	if (!pipe_ctx)
+	if (!pipe_ctx) {
+		kfree(rd_buf);
 		return -ENXIO;
+	}
 
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
@@ -1706,8 +1720,10 @@ static ssize_t dp_dsc_slice_height_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -1870,8 +1886,10 @@ static ssize_t dp_dsc_bits_per_pixel_read(struct file *f, char __user *buf,
 				break;
 	}
 
-	if (!pipe_ctx)
+	if (!pipe_ctx) {
+		kfree(rd_buf);
 		return -ENXIO;
+	}
 
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
@@ -1887,8 +1905,10 @@ static ssize_t dp_dsc_bits_per_pixel_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -2046,8 +2066,10 @@ static ssize_t dp_dsc_pic_width_read(struct file *f, char __user *buf,
 				break;
 	}
 
-	if (!pipe_ctx)
+	if (!pipe_ctx) {
+		kfree(rd_buf);
 		return -ENXIO;
+	}
 
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
@@ -2063,8 +2085,10 @@ static ssize_t dp_dsc_pic_width_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -2103,8 +2127,10 @@ static ssize_t dp_dsc_pic_height_read(struct file *f, char __user *buf,
 				break;
 	}
 
-	if (!pipe_ctx)
+	if (!pipe_ctx) {
+		kfree(rd_buf);
 		return -ENXIO;
+	}
 
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
@@ -2120,8 +2146,10 @@ static ssize_t dp_dsc_pic_height_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -2175,8 +2203,10 @@ static ssize_t dp_dsc_chunk_size_read(struct file *f, char __user *buf,
 				break;
 	}
 
-	if (!pipe_ctx)
+	if (!pipe_ctx) {
+		kfree(rd_buf);
 		return -ENXIO;
+	}
 
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
@@ -2192,8 +2222,10 @@ static ssize_t dp_dsc_chunk_size_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -2247,8 +2279,10 @@ static ssize_t dp_dsc_slice_bpg_offset_read(struct file *f, char __user *buf,
 				break;
 	}
 
-	if (!pipe_ctx)
+	if (!pipe_ctx) {
+		kfree(rd_buf);
 		return -ENXIO;
+	}
 
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
@@ -2264,8 +2298,10 @@ static ssize_t dp_dsc_slice_bpg_offset_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -3255,8 +3291,10 @@ static ssize_t dcc_en_bits_read(
 	dc->hwss.get_dcc_en_bits(dc, dcc_en_bits);
 
 	rd_buf = kcalloc(rd_buf_size, sizeof(char), GFP_KERNEL);
-	if (!rd_buf)
+	if (!rd_buf) {
+		kfree(dcc_en_bits);
 		return -ENOMEM;
+	}
 
 	for (i = 0; i < num_pipes; i++)
 		offset += snprintf(rd_buf + offset, rd_buf_size - offset,
@@ -3269,8 +3307,10 @@ static ssize_t dcc_en_bits_read(
 		if (*pos >= rd_buf_size)
 			break;
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 		buf += 1;
 		size -= 1;
 		*pos += 1;
-- 
2.34.1

[PATCH AUTOSEL 5.16 009/109] drm/amd/display: Use PSR version selected during set_psr_caps

[Sasha Levin]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[ Upstream commit b80ddeb29d9df449f875f0b6f5de08d7537c02b8 ]

[Why]
If the DPCD caps specifies a PSR version newer than PSR_VERSION_1 then
we fallback to using PSR_VERSION_1 in amdgpu_dm_set_psr_caps.

This gets overriden with the raw DPCD value in amdgpu_dm_link_setup_psr,
which can result in DMCUB hanging if we pass in an unsupported PSR
version number.

[How]
Fix the hang by using link->psr_settings.psr_version directly during
amdgpu_dm_link_setup_psr.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c
index c022e56f9459..90962fb91916 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c
@@ -74,10 +74,8 @@ bool amdgpu_dm_link_setup_psr(struct dc_stream_state *stream)
 
 	link = stream->link;
 
-	psr_config.psr_version = link->dpcd_caps.psr_caps.psr_version;
-
-	if (psr_config.psr_version > 0) {
-		psr_config.psr_exit_link_training_required = 0x1;
+	if (link->psr_settings.psr_version != DC_PSR_VERSION_UNSUPPORTED) {
+		psr_config.psr_version = link->psr_settings.psr_version;
 		psr_config.psr_frame_capture_indication_req = 0;
 		psr_config.psr_rfb_setup_time = 0x37;
 		psr_config.psr_sdp_transmit_line_num_deadline = 0x20;
-- 
2.34.1

[PATCH AUTOSEL 5.16 015/109] drm/amdkfd: Don't take process mutex for svm ioctls

[Sasha Levin]

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

[ Upstream commit ac7c48c0cce00d03b3c95fddcccb0a45257e33e3 ]

SVM ioctls take proper svms->lock to handle race conditions, don't need
take process mutex to serialize ioctls. This also fixes circular locking
warning:

WARNING: possible circular locking dependency detected

  Possible unsafe locking scenario:

        CPU0                    CPU1
        ----                    ----
   lock((work_completion)(&svms->deferred_list_work));
                                lock(&process->mutex);
                     lock((work_completion)(&svms->deferred_list_work));
   lock(&process->mutex);

   *** DEADLOCK ***

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_chardev.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index 24ebd61395d8..3aaf10c778d7 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
@@ -1840,13 +1840,9 @@ static int kfd_ioctl_svm(struct file *filep, struct kfd_process *p, void *data)
 	if (!args->start_addr || !args->size)
 		return -EINVAL;
 
-	mutex_lock(&p->mutex);
-
 	r = svm_ioctl(p, args->op, args->start_addr, args->size, args->nattr,
 		      args->attrs);
 
-	mutex_unlock(&p->mutex);
-
 	return r;
 }
 #else
-- 
2.34.1

[PATCH AUTOSEL 5.16 016/109] drm/amdkfd: Ensure mm remain valid in svm deferred_list work

[Sasha Levin]

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

[ Upstream commit 367c9b0f1b8750a704070e7ae85234d591290434 ]

svm_deferred_list work should continue to handle deferred_range_list
which maybe split to child range to avoid child range leak, and remove
ranges mmu interval notifier to avoid mm mm_count leak. So taking mm
reference when adding range to deferred list, to ensure mm is valid in
the scheduled deferred_list_work, and drop the mm referrence after range
is handled.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reported-by: Ruili Ji <ruili.ji@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 | 62 ++++++++++++++++------------
 1 file changed, 36 insertions(+), 26 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index c0b8f4ff80b8..ea1c5aaf659a 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -1926,10 +1926,9 @@ svm_range_update_notifier_and_interval_tree(struct mm_struct *mm,
 }
 
 static void
-svm_range_handle_list_op(struct svm_range_list *svms, struct svm_range *prange)
+svm_range_handle_list_op(struct svm_range_list *svms, struct svm_range *prange,
+			 struct mm_struct *mm)
 {
-	struct mm_struct *mm = prange->work_item.mm;
-
 	switch (prange->work_item.op) {
 	case SVM_OP_NULL:
 		pr_debug("NULL OP 0x%p prange 0x%p [0x%lx 0x%lx]\n",
@@ -2013,34 +2012,41 @@ static void svm_range_deferred_list_work(struct work_struct *work)
 	pr_debug("enter svms 0x%p\n", svms);
 
 	p = container_of(svms, struct kfd_process, svms);
-	/* Avoid mm is gone when inserting mmu notifier */
-	mm = get_task_mm(p->lead_thread);
-	if (!mm) {
-		pr_debug("svms 0x%p process mm gone\n", svms);
-		return;
-	}
-retry:
-	mmap_write_lock(mm);
-
-	/* Checking for the need to drain retry faults must be inside
-	 * mmap write lock to serialize with munmap notifiers.
-	 */
-	if (unlikely(atomic_read(&svms->drain_pagefaults))) {
-		mmap_write_unlock(mm);
-		svm_range_drain_retry_fault(svms);
-		goto retry;
-	}
 
 	spin_lock(&svms->deferred_list_lock);
 	while (!list_empty(&svms->deferred_range_list)) {
 		prange = list_first_entry(&svms->deferred_range_list,
 					  struct svm_range, deferred_list);
-		list_del_init(&prange->deferred_list);
 		spin_unlock(&svms->deferred_list_lock);
 
 		pr_debug("prange 0x%p [0x%lx 0x%lx] op %d\n", prange,
 			 prange->start, prange->last, prange->work_item.op);
 
+		mm = prange->work_item.mm;
+retry:
+		mmap_write_lock(mm);
+
+		/* Checking for the need to drain retry faults must be inside
+		 * mmap write lock to serialize with munmap notifiers.
+		 */
+		if (unlikely(atomic_read(&svms->drain_pagefaults))) {
+			mmap_write_unlock(mm);
+			svm_range_drain_retry_fault(svms);
+			goto retry;
+		}
+
+		/* Remove from deferred_list must be inside mmap write lock, for
+		 * two race cases:
+		 * 1. unmap_from_cpu may change work_item.op and add the range
+		 *    to deferred_list again, cause use after free bug.
+		 * 2. svm_range_list_lock_and_flush_work may hold mmap write
+		 *    lock and continue because deferred_list is empty, but
+		 *    deferred_list work is actually waiting for mmap lock.
+		 */
+		spin_lock(&svms->deferred_list_lock);
+		list_del_init(&prange->deferred_list);
+		spin_unlock(&svms->deferred_list_lock);
+
 		mutex_lock(&svms->lock);
 		mutex_lock(&prange->migrate_mutex);
 		while (!list_empty(&prange->child_list)) {
@@ -2051,19 +2057,20 @@ static void svm_range_deferred_list_work(struct work_struct *work)
 			pr_debug("child prange 0x%p op %d\n", pchild,
 				 pchild->work_item.op);
 			list_del_init(&pchild->child_list);
-			svm_range_handle_list_op(svms, pchild);
+			svm_range_handle_list_op(svms, pchild, mm);
 		}
 		mutex_unlock(&prange->migrate_mutex);
 
-		svm_range_handle_list_op(svms, prange);
+		svm_range_handle_list_op(svms, prange, mm);
 		mutex_unlock(&svms->lock);
+		mmap_write_unlock(mm);
+
+		/* Pairs with mmget in svm_range_add_list_work */
+		mmput(mm);
 
 		spin_lock(&svms->deferred_list_lock);
 	}
 	spin_unlock(&svms->deferred_list_lock);
-
-	mmap_write_unlock(mm);
-	mmput(mm);
 	pr_debug("exit svms 0x%p\n", svms);
 }
 
@@ -2081,6 +2088,9 @@ svm_range_add_list_work(struct svm_range_list *svms, struct svm_range *prange,
 			prange->work_item.op = op;
 	} else {
 		prange->work_item.op = op;
+
+		/* Pairs with mmput in deferred_list_work */
+		mmget(mm);
 		prange->work_item.mm = mm;
 		list_add_tail(&prange->deferred_list,
 			      &prange->svms->deferred_range_list);
-- 
2.34.1

[PATCH AUTOSEL 5.16 017/109] drm/amdkfd: svm range restore work deadlock when process exit

[Sasha Levin]

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

[ Upstream commit 6225bb3a88d22594aacea2485dc28ca12d596721 ]

kfd_process_notifier_release flush svm_range_restore_work
which calls svm_range_list_lock_and_flush_work to flush deferred_list
work, but if deferred_list work mmput release the last user, it will
call exit_mmap -> notifier_release, it is deadlock with below backtrace.

Move flush svm_range_restore_work to kfd_process_wq_release to avoid
deadlock. Then svm_range_restore_work take task->mm ref to avoid mm is
gone while validating and mapping ranges to GPU.

Workqueue: events svm_range_deferred_list_work [amdgpu]
Call Trace:
 wait_for_completion+0x94/0x100
 __flush_work+0x12a/0x1e0
 __cancel_work_timer+0x10e/0x190
 cancel_delayed_work_sync+0x13/0x20
 kfd_process_notifier_release+0x98/0x2a0 [amdgpu]
 __mmu_notifier_release+0x74/0x1f0
 exit_mmap+0x170/0x200
 mmput+0x5d/0x130
 svm_range_deferred_list_work+0x104/0x230 [amdgpu]
 process_one_work+0x220/0x3c0

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reported-by: Ruili Ji <ruili.ji@amd.com>
Tested-by: Ruili Ji <ruili.ji@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 |  1 -
 drivers/gpu/drm/amd/amdkfd/kfd_svm.c     | 15 +++++++++------
 2 files changed, 9 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
index b993011cfa64..990228711108 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -1150,7 +1150,6 @@ static void kfd_process_notifier_release(struct mmu_notifier *mn,
 
 	cancel_delayed_work_sync(&p->eviction_work);
 	cancel_delayed_work_sync(&p->restore_work);
-	cancel_delayed_work_sync(&p->svms.restore_work);
 
 	mutex_lock(&p->mutex);
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index ea1c5aaf659a..a1b0c6bda803 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -1589,13 +1589,14 @@ static void svm_range_restore_work(struct work_struct *work)
 
 	pr_debug("restore svm ranges\n");
 
-	/* kfd_process_notifier_release destroys this worker thread. So during
-	 * the lifetime of this thread, kfd_process and mm will be valid.
-	 */
 	p = container_of(svms, struct kfd_process, svms);
-	mm = p->mm;
-	if (!mm)
+
+	/* Keep mm reference when svm_range_validate_and_map ranges */
+	mm = get_task_mm(p->lead_thread);
+	if (!mm) {
+		pr_debug("svms 0x%p process mm gone\n", svms);
 		return;
+	}
 
 	svm_range_list_lock_and_flush_work(svms, mm);
 	mutex_lock(&svms->lock);
@@ -1649,6 +1650,7 @@ static void svm_range_restore_work(struct work_struct *work)
 out_reschedule:
 	mutex_unlock(&svms->lock);
 	mmap_write_unlock(mm);
+	mmput(mm);
 
 	/* If validation failed, reschedule another attempt */
 	if (evicted_ranges) {
@@ -2779,6 +2781,8 @@ void svm_range_list_fini(struct kfd_process *p)
 
 	pr_debug("pasid 0x%x svms 0x%p\n", p->pasid, &p->svms);
 
+	cancel_delayed_work_sync(&p->svms.restore_work);
+
 	/* Ensure list work is finished before process is destroyed */
 	flush_work(&p->svms.deferred_list_work);
 
@@ -2789,7 +2793,6 @@ void svm_range_list_fini(struct kfd_process *p)
 	atomic_inc(&p->svms.drain_pagefaults);
 	svm_range_drain_retry_fault(&p->svms);
 
-
 	list_for_each_entry_safe(prange, next, &p->svms.list, list) {
 		svm_range_unlink(prange);
 		svm_range_remove_notifier(prange);
-- 
2.34.1

[PATCH AUTOSEL 5.16 036/109] drm/amdgpu: Fix recursive locking warning

[Sasha Levin]

From: Rajneesh Bhardwaj <rajneesh.bhardwaj@amd.com>

[ Upstream commit 447c7997b62a5115ba4da846dcdee4fc12298a6a ]

Noticed the below warning while running a pytorch workload on vega10
GPUs. Change to trylock to avoid conflicts with already held reservation
locks.

[  +0.000003] WARNING: possible recursive locking detected
[  +0.000003] 5.13.0-kfd-rajneesh #1030 Not tainted
[  +0.000004] --------------------------------------------
[  +0.000002] python/4822 is trying to acquire lock:
[  +0.000004] ffff932cd9a259f8 (reservation_ww_class_mutex){+.+.}-{3:3},
at: amdgpu_bo_release_notify+0xc4/0x160 [amdgpu]
[  +0.000203]
              but task is already holding lock:
[  +0.000003] ffff932cbb7181f8 (reservation_ww_class_mutex){+.+.}-{3:3},
at: ttm_eu_reserve_buffers+0x270/0x470 [ttm]
[  +0.000017]
              other info that might help us debug this:
[  +0.000002]  Possible unsafe locking scenario:

[  +0.000003]        CPU0
[  +0.000002]        ----
[  +0.000002]   lock(reservation_ww_class_mutex);
[  +0.000004]   lock(reservation_ww_class_mutex);
[  +0.000003]
               *** DEADLOCK ***

[  +0.000002]  May be due to missing lock nesting notation

[  +0.000003] 7 locks held by python/4822:
[  +0.000003]  #0: ffff932c4ac028d0 (&process->mutex){+.+.}-{3:3}, at:
kfd_ioctl_map_memory_to_gpu+0x10b/0x320 [amdgpu]
[  +0.000232]  #1: ffff932c55e830a8 (&info->lock#2){+.+.}-{3:3}, at:
amdgpu_amdkfd_gpuvm_map_memory_to_gpu+0x64/0xf60 [amdgpu]
[  +0.000241]  #2: ffff932cc45b5e68 (&(*mem)->lock){+.+.}-{3:3}, at:
amdgpu_amdkfd_gpuvm_map_memory_to_gpu+0xdf/0xf60 [amdgpu]
[  +0.000236]  #3: ffffb2b35606fd28
(reservation_ww_class_acquire){+.+.}-{0:0}, at:
amdgpu_amdkfd_gpuvm_map_memory_to_gpu+0x232/0xf60 [amdgpu]
[  +0.000235]  #4: ffff932cbb7181f8
(reservation_ww_class_mutex){+.+.}-{3:3}, at:
ttm_eu_reserve_buffers+0x270/0x470 [ttm]
[  +0.000015]  #5: ffffffffc045f700 (*(sspp++)){....}-{0:0}, at:
drm_dev_enter+0x5/0xa0 [drm]
[  +0.000038]  #6: ffff932c52da7078 (&vm->eviction_lock){+.+.}-{3:3},
at: amdgpu_vm_bo_update_mapping+0xd5/0x4f0 [amdgpu]
[  +0.000195]
              stack backtrace:
[  +0.000003] CPU: 11 PID: 4822 Comm: python Not tainted
5.13.0-kfd-rajneesh #1030
[  +0.000005] Hardware name: GIGABYTE MZ01-CE0-00/MZ01-CE0-00, BIOS F02
08/29/2018
[  +0.000003] Call Trace:
[  +0.000003]  dump_stack+0x6d/0x89
[  +0.000010]  __lock_acquire+0xb93/0x1a90
[  +0.000009]  lock_acquire+0x25d/0x2d0
[  +0.000005]  ? amdgpu_bo_release_notify+0xc4/0x160 [amdgpu]
[  +0.000184]  ? lock_is_held_type+0xa2/0x110
[  +0.000006]  ? amdgpu_bo_release_notify+0xc4/0x160 [amdgpu]
[  +0.000184]  __ww_mutex_lock.constprop.17+0xca/0x1060
[  +0.000007]  ? amdgpu_bo_release_notify+0xc4/0x160 [amdgpu]
[  +0.000183]  ? lock_release+0x13f/0x270
[  +0.000005]  ? lock_is_held_type+0xa2/0x110
[  +0.000006]  ? amdgpu_bo_release_notify+0xc4/0x160 [amdgpu]
[  +0.000183]  amdgpu_bo_release_notify+0xc4/0x160 [amdgpu]
[  +0.000185]  ttm_bo_release+0x4c6/0x580 [ttm]
[  +0.000010]  amdgpu_bo_unref+0x1a/0x30 [amdgpu]
[  +0.000183]  amdgpu_vm_free_table+0x76/0xa0 [amdgpu]
[  +0.000189]  amdgpu_vm_free_pts+0xb8/0xf0 [amdgpu]
[  +0.000189]  amdgpu_vm_update_ptes+0x411/0x770 [amdgpu]
[  +0.000191]  amdgpu_vm_bo_update_mapping+0x324/0x4f0 [amdgpu]
[  +0.000191]  amdgpu_vm_bo_update+0x251/0x610 [amdgpu]
[  +0.000191]  update_gpuvm_pte+0xcc/0x290 [amdgpu]
[  +0.000229]  ? amdgpu_vm_bo_map+0xd7/0x130 [amdgpu]
[  +0.000190]  amdgpu_amdkfd_gpuvm_map_memory_to_gpu+0x912/0xf60
[amdgpu]
[  +0.000234]  kfd_ioctl_map_memory_to_gpu+0x182/0x320 [amdgpu]
[  +0.000218]  kfd_ioctl+0x2b9/0x600 [amdgpu]
[  +0.000216]  ? kfd_ioctl_unmap_memory_from_gpu+0x270/0x270 [amdgpu]
[  +0.000216]  ? lock_release+0x13f/0x270
[  +0.000006]  ? __fget_files+0x107/0x1e0
[  +0.000007]  __x64_sys_ioctl+0x8b/0xd0
[  +0.000007]  do_syscall_64+0x36/0x70
[  +0.000004]  entry_SYSCALL_64_after_hwframe+0x44/0xae
[  +0.000007] RIP: 0033:0x7fbff90a7317
[  +0.000004] Code: b3 66 90 48 8b 05 71 4b 2d 00 64 c7 00 26 00 00 00
48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f
05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 41 4b 2d 00 f7 d8 64 89 01 48
[  +0.000005] RSP: 002b:00007fbe301fe648 EFLAGS: 00000246 ORIG_RAX:
0000000000000010
[  +0.000006] RAX: ffffffffffffffda RBX: 00007fbcc402d820 RCX:
00007fbff90a7317
[  +0.000003] RDX: 00007fbe301fe690 RSI: 00000000c0184b18 RDI:
0000000000000004
[  +0.000003] RBP: 00007fbe301fe690 R08: 0000000000000000 R09:
00007fbcc402d880
[  +0.000003] R10: 0000000002001000 R11: 0000000000000246 R12:
00000000c0184b18
[  +0.000003] R13: 0000000000000004 R14: 00007fbf689593a0 R15:
00007fbcc402d820

Cc: Christian König <christian.koenig@amd.com>
Cc: Felix Kuehling <Felix.Kuehling@amd.com>
Cc: Alex Deucher <Alexander.Deucher@amd.com>

Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Rajneesh Bhardwaj <rajneesh.bhardwaj@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_object.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 4fcfc2313b8c..ea792b6a1c0f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -1286,7 +1286,8 @@ void amdgpu_bo_release_notify(struct ttm_buffer_object *bo)
 	    !(abo->flags & AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE))
 		return;
 
-	dma_resv_lock(bo->base.resv, NULL);
+	if (WARN_ON_ONCE(!dma_resv_trylock(bo->base.resv)))
+		return;
 
 	r = amdgpu_fill_buffer(abo, AMDGPU_POISON, bo->base.resv, &fence);
 	if (!WARN_ON(r)) {
-- 
2.34.1

[PATCH AUTOSEL 5.16 050/109] drm/amd/display: reset lane settings after each PHY repeater LT

[Sasha Levin]

From: Sung Joon Kim <sungkim@amd.com>

[ Upstream commit 3b853c316c9321e195414a6fb121d1c2d45b1e87 ]

[why]
In LTTPR non-transparent mode, we need
to reset the cached lane settings before performing
link training on the next PHY repeater. Otherwise,
the cached lane settings will be used for the next
clock recovery e.g. VS = MAX (3) which should not be
the case according to the DP specs. We expect to use
minimum lane settings on each clock recovery sequence.

[how]
Reset DPCD and HW lane settings on each repeater LT.
Set training pattern to 0 for the repeater that failed LT
at the proper place.

Reviewed-by: Meenakshikumar Somasundaram <Meenakshikumar.Somasundaram@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Jasdeep Dhillon <jdhillon@amd.com>
Signed-off-by: Sung Joon Kim <sungkim@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/core/dc_link_dp.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
index 135ea1c422f2..f46aa7f8c35d 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
@@ -2124,21 +2124,26 @@ static enum link_training_result dp_perform_8b_10b_link_training(
 				repeater_id--) {
 			status = perform_clock_recovery_sequence(link, lt_settings, repeater_id);
 
-			if (status != LINK_TRAINING_SUCCESS)
+			if (status != LINK_TRAINING_SUCCESS) {
+				repeater_training_done(link, repeater_id);
 				break;
+			}
 
 			status = perform_channel_equalization_sequence(link,
 					lt_settings,
 					repeater_id);
 
+			repeater_training_done(link, repeater_id);
+
 			if (status != LINK_TRAINING_SUCCESS)
 				break;
 
-			repeater_training_done(link, repeater_id);
+			for (lane = 0; lane < LANE_COUNT_DP_MAX; lane++) {
+				lt_settings->dpcd_lane_settings[lane].raw = 0;
+				lt_settings->hw_lane_settings[lane].VOLTAGE_SWING = 0;
+				lt_settings->hw_lane_settings[lane].PRE_EMPHASIS = 0;
+			}
 		}
-
-		for (lane = 0; lane < (uint8_t)lt_settings->link_settings.lane_count; lane++)
-			lt_settings->dpcd_lane_settings[lane].raw = 0;
 	}
 
 	if (status == LINK_TRAINING_SUCCESS) {
-- 
2.34.1

[PATCH AUTOSEL 5.16 062/109] drm/amdkfd: make CRAT table missing message informational only

[Sasha Levin]

From: Alex Deucher <alexander.deucher@amd.com>

[ Upstream commit 9dff13f9edf755a15f6507874185a3290c1ae8bb ]

The driver has a fallback so make the message informational
rather than a warning. The driver has a fallback if the
Component Resource Association Table (CRAT) is missing, so
make this informational now.

Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/1906
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_crat.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
index c33d689f29e8..e574aa32a111 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
@@ -1563,7 +1563,7 @@ int kfd_create_crat_image_acpi(void **crat_image, size_t *size)
 	/* Fetch the CRAT table from ACPI */
 	status = acpi_get_table(CRAT_SIGNATURE, 0, &crat_table);
 	if (status == AE_NOT_FOUND) {
-		pr_warn("CRAT table not found\n");
+		pr_info("CRAT table not found\n");
 		return -ENODATA;
 	} else if (ACPI_FAILURE(status)) {
 		const char *err = acpi_format_exception(status);
-- 
2.34.1