[PATCH] D151326: [Assignment Tracking] Ignore stores to a negative offset from an alloca

Orlando Cazalet-Hyams via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Wed May 24 06:41:18 PDT 2023

Orlando added a comment.

In D151326#4368227 <https://reviews.llvm.org/D151326#4368227>, @jryans wrote:

> The `llvm.org/PR123`-style link in the patch summary doesn't seem to work. Perhaps for GitHub-only issues you need to use a direct link like https://github.com/llvm/llvm-project/issues/62838?

Aha, I didn't set the link target properly. Fixed, thanks for pointing that out.

> The patch looks good to me overall, thanks for working on this. I am wondering though... What would happen with a positive-but-still-out-of-bounds offset here?

Good question. Those assignments are tracked, for better or for worse. e.g. we get a DBG_VALUE describing that out of bounds offset covering a fragment of the variable which is also out of bounds. I can see the argument for not tracking them. I don't //think// actively tracking them causes any issues though. Possibly we might see a location list rather than a single location if one of these out-of-bounds locations had been DSE'd. I think it is worth looking into further. I'll add it to my list as a low priority task (due to the apparent "harmless" nature of it), if you're happy with that?



More information about the llvm-commits mailing list