Skip to contentSkip to conversation

habanalabs: use GFP_ATOMIC under spin lock

PATCH In review Wei Yongjun 3 participants 4 messages Feb 22, 2019

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 habanalabs: use GFP_ATOMIC under spin lock

A spin lock is taken here so we should use GFP_ATOMIC.

Fixes 0feaf86d4e69 ("habanalabs: add virtual memory and MMU modules")
---
 drivers/misc/habanalabs/memory.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/misc/habanalabs/memory.c +1 -1
index 9236e52852c6..9e3491dc3b55 100644
@@ -120,7 +120,7 @@ static int alloc_device_memory(struct hl_ctx *ctx, struct hl_mem_in *args,
 
 	spin_lock(&vm->idr_lock);
 	handle = idr_alloc(&vm->phys_pg_pack_handles, phys_pg_pack, 1, 0,
-				GFP_KERNEL);
+				GFP_ATOMIC);
 	spin_unlock(&vm->idr_lock);
 
 	if (handle < 0) {
#1 lore raw
Quoted 16 lines · Wei Yongjun

A spin lock is taken here so we should use GFP_ATOMIC.

Fixes 0feaf86d4e69 ("habanalabs: add virtual memory and MMU modules")
---
 drivers/misc/habanalabs/memory.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/misc/habanalabs/memory.c +0 -0
index 9236e52852c6..9e3491dc3b55 100644

drivers/misc/habanalabs/memory.c static int alloc_device_memory(struct hl_ctx *ctx, struct hl_mem_in *args, · line 126

Quoted 8 lines
drivers/misc/habanalabs/memory.c +1 -1
@@ -120,7 +120,7 @@ static int alloc_device_memory(struct hl_ctx *ctx, struct hl_mem_in *args,
 
 	spin_lock(&vm->idr_lock);
 	handle = idr_alloc(&vm->phys_pg_pack_handles, phys_pg_pack, 1, 0,
-				GFP_KERNEL);
+				GFP_ATOMIC);
 	spin_unlock(&vm->idr_lock);
 
 	if (handle < 0) {

Ah, nice catch!

Oded, any objection to me taking this patch?

greg k-h

Reviewed-by
#2 lore raw
Quoted 29 lines · Greg Kroah-Hartman
Quoted 16 lines · Wei Yongjun

A spin lock is taken here so we should use GFP_ATOMIC.

Fixes 0feaf86d4e69 ("habanalabs: add virtual memory and MMU modules")
---
 drivers/misc/habanalabs/memory.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/misc/habanalabs/memory.c +0 -0
index 9236e52852c6..9e3491dc3b55 100644

drivers/misc/habanalabs/memory.c static int alloc_device_memory(struct hl_ctx *ctx, struct hl_mem_in *args, · line 126

Quoted 8 lines
drivers/misc/habanalabs/memory.c +1 -1
@@ -120,7 +120,7 @@ static int alloc_device_memory(struct hl_ctx *ctx, struct hl_mem_in *args,
 
      spin_lock(&vm->idr_lock);
      handle = idr_alloc(&vm->phys_pg_pack_handles, phys_pg_pack, 1, 0,
-                             GFP_KERNEL);
+                             GFP_ATOMIC);
      spin_unlock(&vm->idr_lock);
 
      if (handle < 0) {

Ah, nice catch!

Quoted 3 lines

Oded, any objection to me taking this patch?

greg k-h

No, of course not.
This patch is:

#3 lore raw
Quoted 31 lines · Oded Gabbay
Quoted 29 lines · Greg Kroah-Hartman
Quoted 16 lines · Wei Yongjun

A spin lock is taken here so we should use GFP_ATOMIC.

Fixes 0feaf86d4e69 ("habanalabs: add virtual memory and MMU modules")
---
 drivers/misc/habanalabs/memory.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/misc/habanalabs/memory.c +0 -0
index 9236e52852c6..9e3491dc3b55 100644

drivers/misc/habanalabs/memory.c static int alloc_device_memory(struct hl_ctx *ctx, struct hl_mem_in *args, · line 126

Quoted 8 lines
drivers/misc/habanalabs/memory.c +1 -1
@@ -120,7 +120,7 @@ static int alloc_device_memory(struct hl_ctx *ctx, struct hl_mem_in *args,
 
      spin_lock(&vm->idr_lock);
      handle = idr_alloc(&vm->phys_pg_pack_handles, phys_pg_pack, 1, 0,
-                             GFP_KERNEL);
+                             GFP_ATOMIC);
      spin_unlock(&vm->idr_lock);
 
      if (handle < 0) {

Ah, nice catch!

Quoted 6 lines
Quoted 3 lines

Oded, any objection to me taking this patch?

greg k-h

No, of course not.
This patch is:

Great, will go queue it up now, thanks.

greg k-h