Still buffer overflows with rDNS tags

[ocococococ]

Using ASAN options, I tried to fix buffer overflows with rDNS tags.

Navigating through the original source code is not easy and even if I did not find where the actual overflow happens, I think the proposed workaround would/could be acceptable. Adding 4 more bytes than normally necessary when allocating memory storing rDNS name seems to bypass the problem. So maybe someone would be interested in testing this in another context or even finding a better solution.

Here comes a potential patch to solve the problem. reverseDNS.patch

Here comes a sample .sh script that adds many rDNS tags to a .m4a file converted from a .flac file. try.sh.zip

Here comes the ASAN report without the patch. asan.txt

[github-actions[bot]]

Thanks for filing an issue! Please note that this project is only passively maintained, so your best bet for getting an issue resolved is through a pull request that is easy to verify! Please read this for more information.

[miraclx]

Cross-referencing prior context: https://github.com/wez/atomicparsley/issues/44

[wez]

Thanks! Could you include the ASAN overflows you see without this patch in place, and turn it into a PR so that it is reviewable? I'm troubled by the comments about +4 padding and there's a comment in the picture above that has a hardcoded 16 that doesn't appear in the before changes; is that really 12 + 4 bytes of padding? I think getting to the bottom of that magic #4 value would be wise before assuming that this is fixed, as it feels a bit funky.

[ocococococ]

ASAN report uploaded above. PR #63 created.

By the way, much better, no more +4 magic number in proposed patch.