[FFmpeg-devel] [PATCH] libswscale/tests/colorspace: Add av_free() to avoid potential memory leak

[FFmpeg-devel] [PATCH] libswscale/tests/colorspace: Add av_free() to avoid potential memory leak

[Jiasheng Jiang]

Add av_free() to avoid potential memory leak.

Fixes: 3a74415d59 ("cosmetics, expand all tabs to four spaces.")
Signed-off-by: Jiasheng Jiang <jiashengjiangcool@gmail.com>
---
 libswscale/tests/colorspace.c | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/libswscale/tests/colorspace.c b/libswscale/tests/colorspace.c
index d1eddc82bf..d0fd03a03b 100644
--- a/libswscale/tests/colorspace.c
+++ b/libswscale/tests/colorspace.c
@@ -41,9 +41,12 @@ int main(int argc, char **argv)
     uint8_t *dstBuffer = av_malloc(SIZE);
     int failedNum      = 0;
     int passedNum      = 0;
+    int ret;
 
-    if (!srcBuffer || !dstBuffer)
-        return -1;
+    if (!srcBuffer || !dstBuffer) {
+        ret = -1;
+        goto end;
+    }
 
     av_log(NULL, AV_LOG_INFO, "memory corruption test ...\n");
     ff_sws_rgb2rgb_init();
@@ -167,5 +170,10 @@ int main(int argc, char **argv)
     av_log(NULL, AV_LOG_INFO,
            "\n%d converters passed, %d converters randomly overwrote memory\n",
            passedNum, failedNum);
-    return failedNum;
+    ret = failedNum;
+
+err:
+    av_free(srcBuffer);
+    av_free(dstBuffer);
+    return ret;
 }
-- 
2.25.1

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-devel] [PATCH] libswscale/tests/colorspace: Add av_free() to avoid potential memory leak

[Nicolas George]

Jiasheng Jiang (HE12025-08-08):
> Add av_free() to avoid potential memory leak.

This is main(), return means exit, no leak possible.

This applies to most cases in tests, even not in main(): if something
fails, the test will most likely exit with failure almost immediately,
leaks are not an issue.

Also:

> Fixes: 3a74415d59 ("cosmetics, expand all tabs to four spaces.")

This does not seem right. I do not know what is adding these comments,
but its attribution logic seems a little simplistic. Just because¹ the
same lines were changes last in that patch does not mean that patch is
what introduced the issue. It is quite obvious in that case with a
cosmetic commit.

Regards,

-- 
  Nicolas George


1: Strange construction, but apparently idiomatic.
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".