drm/nouveau/secboot: remove VLA usage

PATCH Superseded Gustavo A. R. Silva 2 participants 3 messages Mar 13, 2018

Subsystems (3)

DRM drivers DRM misc nouveau

lore

Share & tools

lore

Gustavo A. R. Silva

author

2018-03-13 14:48 UTC #0 lore raw

PATCH 1/1 drm/nouveau/secboot: remove VLA usage

In preparation to enabling -Wvla, remove VLA. In this particular case directly use macro NVKM_MSGQUEUE_CMDLINE_SIZE instead of local variable cmdline_size. Also, remove cmdline_size as it is not actually useful anymore.

The use of stack Variable Length Arrays needs to be avoided, as they can be a vector for stack exhaustion, which can be both a runtime bug or a security flaw. Also, in general, as code evolves it is easy to lose track of how big a VLA can get. Thus, we can end up having runtime failures that are hard to debug.

Also, fixed as part of the directive to remove all VLAs from the kernel: https://lkml.org/lkml/2018/3/7/621

Signed-off-by Gustavo A. R. Silva <[email protected]>

---
 drivers/gpu/drm/nouveau/nvkm/subdev/secboot/ls_ucode_msgqueue.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

drivers/gpu/drm/nouveau/nvkm/subdev/secboot/ls_ucode_msgqueue.c +4 -4

index 6f10b09..2da147b 100644
@@ -80,12 +80,12 @@ acr_ls_msgqueue_post_run(struct nvkm_msgqueue *queue,
 			 struct nvkm_falcon *falcon, u32 addr_args)
 {
 	struct nvkm_device *device = falcon->owner->device;
-	u32 cmdline_size = NVKM_MSGQUEUE_CMDLINE_SIZE;
-	u8 buf[cmdline_size];
+	u8 buf[NVKM_MSGQUEUE_CMDLINE_SIZE];
 
-	memset(buf, 0, cmdline_size);
+	memset(buf, 0, NVKM_MSGQUEUE_CMDLINE_SIZE);
 	nvkm_msgqueue_write_cmdline(queue, buf);
-	nvkm_falcon_load_dmem(falcon, buf, addr_args, cmdline_size, 0);
+	nvkm_falcon_load_dmem(falcon, buf, addr_args,
+			      NVKM_MSGQUEUE_CMDLINE_SIZE, 0);
 	/* rearm the queue so it will wait for the init message */
 	nvkm_msgqueue_reinit(queue);

Signature · 1 line

2.7.4

David Laight

2018-03-13 16:09 UTC #1 lore raw

From: Gustavo A. R. Silva

Quoted 5 lines

Sent: 13 March 2018 14:48
In preparation to enabling -Wvla, remove VLA. In this particular case directly use macro NVKM_MSGQUEUE_CMDLINE_SIZE instead of local variable cmdline_size. Also, remove cmdline_size as it is not actually useful anymore.

...

Quoted 9 lines

drivers/gpu/drm/nouveau/nvkm/subdev/secboot/ls_ucode_msgqueue.c +0 -0

index 6f10b09..2da147b 100644
@@ -80,12 +80,12 @@ acr_ls_msgqueue_post_run(struct nvkm_msgqueue *queue,
 			 struct nvkm_falcon *falcon, u32 addr_args)
 {
 	struct nvkm_device *device = falcon->owner->device;

drivers/gpu/drm/nouveau/nvkm/subdev/secboot/ls_ucode_msgqueue.c acr_ls_msgqueue_post_run(struct nvkm_msgqueue *queue, · line 88

Quoted 10 lines

drivers/gpu/drm/nouveau/nvkm/subdev/secboot/ls_ucode_msgqueue.c +4 -4

-	u32 cmdline_size = NVKM_MSGQUEUE_CMDLINE_SIZE;
-	u8 buf[cmdline_size];
+	u8 buf[NVKM_MSGQUEUE_CMDLINE_SIZE];
 
-	memset(buf, 0, cmdline_size);
+	memset(buf, 0, NVKM_MSGQUEUE_CMDLINE_SIZE);
 	nvkm_msgqueue_write_cmdline(queue, buf);
-	nvkm_falcon_load_dmem(falcon, buf, addr_args, cmdline_size, 0);
+	nvkm_falcon_load_dmem(falcon, buf, addr_args,
+			      NVKM_MSGQUEUE_CMDLINE_SIZE, 0);

I think I'd use 'sizeof (buf)' in both those lines.

David

Gustavo A. R. Silva

author

2018-03-13 16:18 UTC #2 lore raw

Hi David,

Quoted 20 lines · David Laight

From: Gustavo A. R. Silva
Quoted 5 lines
Sent: 13 March 2018 14:48
In preparation to enabling -Wvla, remove VLA. In this particular case directly use macro NVKM_MSGQUEUE_CMDLINE_SIZE instead of local variable cmdline_size. Also, remove cmdline_size as it is not actually useful anymore.
...
Quoted 13 lines
drivers/gpu/drm/nouveau/nvkm/subdev/secboot/ls_ucode_msgqueue.c +1 -2
index 6f10b09..2da147b 100644
@@ -80,12 +80,12 @@ acr_ls_msgqueue_post_run(struct nvkm_msgqueue *queue,
  			 struct nvkm_falcon *falcon, u32 addr_args)
  {
  	struct nvkm_device *device = falcon->owner->device;
-	u32 cmdline_size = NVKM_MSGQUEUE_CMDLINE_SIZE;
-	u8 buf[cmdline_size];
+	u8 buf[NVKM_MSGQUEUE_CMDLINE_SIZE];
 

Quoted 9 lines

drivers/gpu/drm/nouveau/nvkm/subdev/secboot/ls_ucode_msgqueue.c acr_ls_msgqueue_post_run(struct nvkm_msgqueue *queue, · line 88
Quoted 6 lines
drivers/gpu/drm/nouveau/nvkm/subdev/secboot/ls_ucode_msgqueue.c +3 -2
-	memset(buf, 0, cmdline_size);
+	memset(buf, 0, NVKM_MSGQUEUE_CMDLINE_SIZE);
  	nvkm_msgqueue_write_cmdline(queue, buf);
-	nvkm_falcon_load_dmem(falcon, buf, addr_args, cmdline_size, 0);
+	nvkm_falcon_load_dmem(falcon, buf, addr_args,
+			      NVKM_MSGQUEUE_CMDLINE_SIZE, 0);
I think I'd use 'sizeof (buf)' in both those lines.

Yeah. I like it.

I'll send v2 with that change shortly.

Thanks
--
Gustavo