github-mirrors/FFmpeg
1
0
Fork 0
mirror of https://mirror.skon.top/https://github.com/FFmpeg/FFmpeg synced 2026-04-20 21:00:41 +08:00
Code Issues Packages Projects Releases Wiki Activity

swscale/ops_chain: check for exact linear mask match

Browse Source 
Makes this logic a lot simpler and less brittle. We can trivially adjust the
list of linear masks that are required, whenever it changes as a result of any
future modifications.

Signed-off-by: Niklas Haas <git@haasn.dev>
This commit is contained in:
Niklas Haas
2026-03-31 16:11:27 +02:00
parent e20a32d730
commit 9f0dded48d
1 changed files with 1 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
libswscale/ops_chain.c
View File

@@ -68,10 +68,6 @@ int ff_sws_op_chain_append(SwsOpChain *chain, SwsFuncPtr func,
* Likewise, if `ref->comps` has any components marked as unused, they must be
* marked as unused in `ops` as well.
*
* For SWS_OP_LINEAR, `ref->linear.mask` must be a strict superset of
* `op->linear.mask`, but may not contain any columns explicitly ignored by
* `op->comps.unused`.
*
* For unfiltered SWS_OP_READ/SWS_OP_WRITE, SWS_OP_SWAP_BYTES and
* SWS_OP_SWIZZLE, the exact type is not checked, just the size.
*
@@ -172,17 +168,8 @@ static int op_match(const SwsOp *op, const SwsOpEntry *entry)
av_assert1(entry->flexible);
break;
case SWS_OP_LINEAR:
/* All required elements must be present */
if (op->lin.mask & ~entry->linear_mask)
if (op->lin.mask != entry->linear_mask)
return 0;
/* To avoid operating on possibly undefined memory, filter out
* implementations that operate on more input components */
for (int i = 0; i < 4; i++) {
if ((entry->linear_mask & SWS_MASK_COL(i)) && op->comps.unused[i])
return 0;
}
/* Prioritize smaller implementations */
score += av_popcount(SWS_MASK_ALL ^ entry->linear_mask);
return score;
case SWS_OP_SCALE:
return av_cmp_q(op->scale.factor, entry->scale) ? 0 : score;