| Message ID | 20250615153751.1707298-5-tasos@tasossah.com |
|---|---|
| State | New |
| Headers | show |
| Series | Resolve various memory-related issues | expand |
diff --git a/src/speedy.c b/src/speedy.c index 2c85dc4..7883b86 100644 --- a/src/speedy.c +++ b/src/speedy.c @@ -1617,7 +1617,7 @@ static void quarter_blit_vertical_packed422_scanline_mmxext( uint8_t *output, ui one += 8; three += 8; } - width = width & 0x7; + width = width & 0x3; /* Handle last few pixels. */ for( i = width * 2; i; --i ) {
There are three loops in quarter_blit_vertical_packed422_scanline_mmxext(). The first one processes the scanline in chunks of 16. The width is then masked to process the remaining pixels in chunks of 4. It is then masked again to process any remaining pixels. The last mask is set incorrectly, which results in the last loop running even if all pixels have been processed by the previous one. Resolve it by setting the correct mask (divisor - 1) for the last loop. Signed-off-by: Tasos Sahanidis <tasos@tasossah.com> --- src/speedy.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)