[PATCH 1/2] habanalabs: add prefetch flag to the MAP operation

[PATCH 1/2] habanalabs: add prefetch flag to the MAP operation

[Oded Gabbay]

From: Ohad Sharabi <osharabi@habana.ai>

This patch let the user decide whether the translations done in the
page tables will be fetched directly to the STLB right after the map.

We want to let the user control whether to perform prefetch upon map
operation.

To do so a memory flag was added, to be used in the MAP ioctl, called
HL_MEM_PREFETCH and if set- the mappings will be fetched directly to
the STLB after map operation.

Signed-off-by: Ohad Sharabi <osharabi@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
---
 drivers/misc/habanalabs/common/memory.c | 11 ++++++-----
 include/uapi/misc/habanalabs.h          |  1 +
 2 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/drivers/misc/habanalabs/common/memory.c b/drivers/misc/habanalabs/common/memory.c
index 326c2179628f..6face45c57e3 100644
--- a/drivers/misc/habanalabs/common/memory.c
+++ b/drivers/misc/habanalabs/common/memory.c
@@ -1250,11 +1250,12 @@ static int map_device_va(struct hl_ctx *ctx, struct hl_mem_in *args,
 	if (rc)
 		goto map_err;
 
-	/* already prefetch the relevant translations to the cache */
-	rc = hl_mmu_prefetch_cache_range(hdev, *vm_type, ctx->asid, ret_vaddr,
-						phys_pg_pack->total_size);
-	if (rc)
-		goto map_err;
+	if (args->flags & HL_MEM_PREFETCH) {
+		rc = hl_mmu_prefetch_cache_range(hdev, *vm_type, ctx->asid, ret_vaddr,
+							phys_pg_pack->total_size);
+		if (rc)
+			goto map_err;
+	}
 
 	mutex_unlock(&ctx->mmu_lock);
 
diff --git a/include/uapi/misc/habanalabs.h b/include/uapi/misc/habanalabs.h
index ca2af5f98056..3576bf2b4841 100644
--- a/include/uapi/misc/habanalabs.h
+++ b/include/uapi/misc/habanalabs.h
@@ -1134,6 +1134,7 @@ union hl_wait_cs_args {
 #define HL_MEM_SHARED		0x2
 #define HL_MEM_USERPTR		0x4
 #define HL_MEM_FORCE_HINT	0x8
+#define HL_MEM_PREFETCH		0x40
 
 /**
  * structure hl_mem_in - structure that handle input args for memory IOCTL
-- 
2.25.1

[PATCH 2/2] habanalabs: use get_task_pid() to take PID

[Oded Gabbay]

From: Dani Liberman <dliberman@habana.ai>

find_get_pid() isn't good in case the user process was run inside
docker.

As a result, we didn't had the PID and we couldn't kill the user
process in case the device got stuck and we needed to reset the
device.

Signed-off-by: Dani Liberman <dliberman@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
---
 drivers/misc/habanalabs/common/habanalabs_drv.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/misc/habanalabs/common/habanalabs_drv.c b/drivers/misc/habanalabs/common/habanalabs_drv.c
index 28549d34d99c..b85ca1e66eb2 100644
--- a/drivers/misc/habanalabs/common/habanalabs_drv.c
+++ b/drivers/misc/habanalabs/common/habanalabs_drv.c
@@ -237,7 +237,7 @@ int hl_device_open_ctrl(struct inode *inode, struct file *filp)
 	hpriv->filp = filp;
 	nonseekable_open(inode, filp);
 
-	hpriv->taskpid = find_get_pid(current->pid);
+	hpriv->taskpid = get_task_pid(current, PIDTYPE_PID);
 
 	mutex_lock(&hdev->fpriv_ctrl_list_lock);
 
-- 
2.25.1