* [FFmpeg-devel] [PATCH] avutil/cuda_check: add error handling to ff_cuda_check
@ 2025-12-11 2:51 Chris Boyd via ffmpeg-devel
2025-12-11 17:52 ` [FFmpeg-devel] " Timo Rothenpieler via ffmpeg-devel
0 siblings, 1 reply; 2+ messages in thread
From: Chris Boyd via ffmpeg-devel @ 2025-12-11 2:51 UTC (permalink / raw)
To: 'ffmpeg-devel@ffmpeg.org'; +Cc: Chris Boyd
Add variable initialization and return value checks when calling
cuGetErrorName() and cuGetErrorString(). Although the documentation for
these methods seems to imply that the provided char* will be set to NULL
on failure, we have seen crashes when logging these strings with
uninitialized memory on machines that have Nvidia drivers installed but
no longer have an Nvidia GPU.
Signed-off-by: Christopher Boyd <cboyd@valvesoftware.com>
---
libavutil/cuda_check.h | 14 ++++++++++----
1 file changed, 10 insertions(+), 4 deletions(-)
diff --git a/libavutil/cuda_check.h b/libavutil/cuda_check.h
index f5a9234eaf..763009ce92 100644
--- a/libavutil/cuda_check.h
+++ b/libavutil/cuda_check.h
@@ -33,16 +33,22 @@ static inline int ff_cuda_check(void *avctx,
void *cuGetErrorName_fn, void *cuGetErrorString_fn,
CUresult err, const char *func)
{
- const char *err_name;
- const char *err_string;
+ const char *err_name = NULL;
+ const char *err_string = NULL;
+ CUresult get_err;
av_log(avctx, AV_LOG_TRACE, "Calling %s\n", func);
if (err == CUDA_SUCCESS)
return 0;
- ((cuda_check_GetErrorName *)cuGetErrorName_fn)(err, &err_name);
- ((cuda_check_GetErrorString *)cuGetErrorString_fn)(err, &err_string);
+ get_err = ((cuda_check_GetErrorName *)cuGetErrorName_fn)(err, &err_name);
+ if (get_err != CUDA_SUCCESS)
+ err_name = NULL;
+
+ get_err = ((cuda_check_GetErrorString *)cuGetErrorString_fn)(err, &err_string);
+ if (get_err != CUDA_SUCCESS)
+ err_string = NULL;
av_log(avctx, AV_LOG_ERROR, "%s failed", func);
if (err_name && err_string)
--
2.43.0
_______________________________________________
ffmpeg-devel mailing list -- ffmpeg-devel@ffmpeg.org
To unsubscribe send an email to ffmpeg-devel-leave@ffmpeg.org
^ permalink raw reply [flat|nested] 2+ messages in thread* [FFmpeg-devel] Re: [PATCH] avutil/cuda_check: add error handling to ff_cuda_check
2025-12-11 2:51 [FFmpeg-devel] [PATCH] avutil/cuda_check: add error handling to ff_cuda_check Chris Boyd via ffmpeg-devel
@ 2025-12-11 17:52 ` Timo Rothenpieler via ffmpeg-devel
0 siblings, 0 replies; 2+ messages in thread
From: Timo Rothenpieler via ffmpeg-devel @ 2025-12-11 17:52 UTC (permalink / raw)
To: ffmpeg-devel; +Cc: Timo Rothenpieler
[-- Attachment #1.1.1.1: Type: text/plain, Size: 2033 bytes --]
On 11.12.2025 03:51, Chris Boyd via ffmpeg-devel wrote:
> Add variable initialization and return value checks when calling
> cuGetErrorName() and cuGetErrorString(). Although the documentation for
> these methods seems to imply that the provided char* will be set to NULL
> on failure, we have seen crashes when logging these strings with
> uninitialized memory on machines that have Nvidia drivers installed but
> no longer have an Nvidia GPU.
>
> Signed-off-by: Christopher Boyd <cboyd@valvesoftware.com>
> ---
> libavutil/cuda_check.h | 14 ++++++++++----
> 1 file changed, 10 insertions(+), 4 deletions(-)
>
> diff --git a/libavutil/cuda_check.h b/libavutil/cuda_check.h
> index f5a9234eaf..763009ce92 100644
> --- a/libavutil/cuda_check.h
> +++ b/libavutil/cuda_check.h
> @@ -33,16 +33,22 @@ static inline int ff_cuda_check(void *avctx,
> void *cuGetErrorName_fn, void *cuGetErrorString_fn,
> CUresult err, const char *func)
> {
> - const char *err_name;
> - const char *err_string;
> + const char *err_name = NULL;
> + const char *err_string = NULL;
> + CUresult get_err;
>
> av_log(avctx, AV_LOG_TRACE, "Calling %s\n", func);
>
> if (err == CUDA_SUCCESS)
> return 0;
>
> - ((cuda_check_GetErrorName *)cuGetErrorName_fn)(err, &err_name);
> - ((cuda_check_GetErrorString *)cuGetErrorString_fn)(err, &err_string);
> + get_err = ((cuda_check_GetErrorName *)cuGetErrorName_fn)(err, &err_name);
> + if (get_err != CUDA_SUCCESS)
> + err_name = NULL;
> +
> + get_err = ((cuda_check_GetErrorString *)cuGetErrorString_fn)(err, &err_string);
> + if (get_err != CUDA_SUCCESS)
> + err_string = NULL;
>
> av_log(avctx, AV_LOG_ERROR, "%s failed", func);
> if (err_name && err_string)
LGTM
fyi: You can also submit patches to
https://code.ffmpeg.org/FFmpeg/FFmpeg as PR now, and development is
rapidly shifting to that.
[-- Attachment #1.1.1.2: OpenPGP public key --]
[-- Type: application/pgp-keys, Size: 3203 bytes --]
[-- Attachment #1.2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 840 bytes --]
[-- Attachment #2: Type: text/plain, Size: 163 bytes --]
_______________________________________________
ffmpeg-devel mailing list -- ffmpeg-devel@ffmpeg.org
To unsubscribe send an email to ffmpeg-devel-leave@ffmpeg.org
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2025-12-11 17:53 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2025-12-11 2:51 [FFmpeg-devel] [PATCH] avutil/cuda_check: add error handling to ff_cuda_check Chris Boyd via ffmpeg-devel
2025-12-11 17:52 ` [FFmpeg-devel] " Timo Rothenpieler via ffmpeg-devel
Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
This inbox may be cloned and mirrored by anyone:
git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git
# If you have public-inbox 1.1+ installed, you may
# initialize and index your mirror using the following commands:
public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
ffmpegdev@gitmailbox.com
public-inbox-index ffmpegdev
Example config snippet for mirrors.
AGPL code for this site: git clone https://public-inbox.org/public-inbox.git