[PATCH v4 19/24] drm/amdkfd: add queue remapping

[James Zhu]

Add queue remapping to ensure that any waves executing the PC sampling
part of the trap handler are done before kfd_pc_sample_stop returns,
and that no new waves enter that part of the trap handler afterwards.
This avoids race conditions that could lead to use-after-free. Unmapping
and remapping the queues either waits for the waves to drain, or preempts
them with CWSR, which itself executes a trap and waits for previous traps
to finish.

Signed-off-by: James Zhu <James.Zhu at amd.com>
---
 drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c | 11 +++++++++++
 drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h |  5 +++++
 drivers/gpu/drm/amd/amdkfd/kfd_pc_sampling.c          |  4 +++-
 3 files changed, 19 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
index c0e71543389a..a3f57be63f4f 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -3155,6 +3155,17 @@ int debug_refresh_runlist(struct device_queue_manager *dqm)
 	return debug_map_and_unlock(dqm);
 }
 
+void remap_queue(struct device_queue_manager *dqm,
+				enum kfd_unmap_queues_filter filter,
+				uint32_t filter_param,
+				uint32_t grace_period)
+{
+	dqm_lock(dqm);
+	if (!dqm->dev->kfd->shared_resources.enable_mes)
+		execute_queues_cpsch(dqm, filter, filter_param, grace_period);
+	dqm_unlock(dqm);
+}
+
 #if defined(CONFIG_DEBUG_FS)
 
 static void seq_reg_dump(struct seq_file *m,
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
index cf7e182588f8..f8aae3747a36 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
@@ -303,6 +303,11 @@ int debug_lock_and_unmap(struct device_queue_manager *dqm);
 int debug_map_and_unlock(struct device_queue_manager *dqm);
 int debug_refresh_runlist(struct device_queue_manager *dqm);
 
+void remap_queue(struct device_queue_manager *dqm,
+				enum kfd_unmap_queues_filter filter,
+				uint32_t filter_param,
+				uint32_t grace_period);
+
 static inline unsigned int get_sh_mem_bases_32(struct kfd_process_device *pdd)
 {
 	return (pdd->lds_base >> 16) & 0xFF;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pc_sampling.c b/drivers/gpu/drm/amd/amdkfd/kfd_pc_sampling.c
index 53e44e68408e..df2f4bfd0cda 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_pc_sampling.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_pc_sampling.c
@@ -24,6 +24,7 @@
 #include "kfd_priv.h"
 #include "amdgpu_amdkfd.h"
 #include "kfd_pc_sampling.h"
+#include "kfd_device_queue_manager.h"
 
 struct supported_pc_sample_info {
 	uint32_t ip_version;
@@ -115,9 +116,10 @@ static int kfd_pc_sample_stop(struct kfd_process_device *pdd,
 
 	kfd_process_set_trap_pc_sampling_flag(&pdd->qpd,
 		pdd->dev->pcs_data.hosttrap_entry.base.pc_sample_info.method, false);
+	remap_queue(pdd->dev->dqm,
+		KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD);
 
 	if (pc_sampling_stop) {
-
 		mutex_lock(&pdd->dev->pcs_data.mutex);
 		pdd->dev->pcs_data.hosttrap_entry.base.target_simd = 0;
 		pdd->dev->pcs_data.hosttrap_entry.base.target_wave_slot = 0;
-- 
2.25.1