2024-04-02 11:18:26

by Harshit Mogalapalli

Subject: [PATCH] drm/panthor: Fix IS_ERR() vs NULL bug in group_process_tiler_oom()

panthor_vm_get_heap_pool() returns ERR_PTR on failure, update the check
accordingly.

Fixes: de8548813824 ("drm/panthor: Add the scheduler logical block")
Signed-off-by: Harshit Mogalapalli <[email protected]>
---
This is based on static analysis with smatch, only compile tested.
---
drivers/gpu/drm/panthor/panthor_sched.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/panthor/panthor_sched.c b/drivers/gpu/drm/panthor/panthor_sched.c
index 5f7803b6fc48..d34f213795a3 100644
--- a/drivers/gpu/drm/panthor/panthor_sched.c
+++ b/drivers/gpu/drm/panthor/panthor_sched.c
@@ -1343,7 +1343,10 @@ static int group_process_tiler_oom(struct panthor_group *group, u32 cs_id)
if (unlikely(csg_id < 0))
return 0;

- if (!heaps || frag_end > vt_end || vt_end >= vt_start) {
+ if (IS_ERR(heaps))
+ return PTR_ERR(heaps);
+
+ if (frag_end > vt_end || vt_end >= vt_start) {
ret = -EINVAL;
} else {
/* We do the allocation without holding the scheduler lock to avoid
--
2.39.3

2024-04-02 11:24:34

by Harshit Mogalapalli

Subject: Re: [PATCH] drm/panthor: Fix IS_ERR() vs NULL bug in group_process_tiler_oom()

Hi,

On 02/04/24 16:47, Harshit Mogalapalli wrote:
> panthor_vm_get_heap_pool() returns ERR_PTR on failure, update the check
> accordingly.
>
> Fixes: de8548813824 ("drm/panthor: Add the scheduler logical block")
> Signed-off-by: Harshit Mogalapalli <[email protected]>
> ---
> This is based on static analysis with smatch, only compile tested.
> ---
> drivers/gpu/drm/panthor/panthor_sched.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/panthor/panthor_sched.c b/drivers/gpu/drm/panthor/panthor_sched.c
> index 5f7803b6fc48..d34f213795a3 100644
> --- a/drivers/gpu/drm/panthor/panthor_sched.c
> +++ b/drivers/gpu/drm/panthor/panthor_sched.c
> @@ -1343,7 +1343,10 @@ static int group_process_tiler_oom(struct panthor_group *group, u32 cs_id)
> if (unlikely(csg_id < 0))
> return 0;
>
> - if (!heaps || frag_end > vt_end || vt_end >= vt_start) {
> + if (IS_ERR(heaps))
> + return PTR_ERR(heaps);
> +

Please ignore this patch, doesn't look correct on looking at more closely.

I should have just done this:

diff --git a/drivers/gpu/drm/panthor/panthor_sched.c
b/drivers/gpu/drm/panthor/panthor_sched.c
index 5f7803b6fc48..617df2b980d0 100644
--- a/drivers/gpu/drm/panthor/panthor_sched.c
+++ b/drivers/gpu/drm/panthor/panthor_sched.c
@@ -1343,7 +1343,7 @@ static int group_process_tiler_oom(struct
panthor_group *group, u32 cs_id)
if (unlikely(csg_id < 0))
return 0;

- if (!heaps || frag_end > vt_end || vt_end >= vt_start) {
+ if (IS_ERR(heaps) || frag_end > vt_end || vt_end >= vt_start) {
ret = -EINVAL;
} else {
/* We do the allocation without holding the scheduler
lock to avoid


Thanks,
Harshit


> + if (frag_end > vt_end || vt_end >= vt_start) {
> ret = -EINVAL;
> } else {
> /* We do the allocation without holding the scheduler lock to avoid