Stream: git-wasmtime

Topic: wasmtime / PR #3585 x64: assert that temporary and destin...


view this post on Zulip Wasmtime GitHub notifications bot (Dec 07 2021 at 19:46):

abrown opened PR #3585 from assert-tmp-dsts to main:

@cfallin, here is the change we just discussed to ensure that renaming operates on the same numbers of temporary and destination registers. I see it failing on the address64.wast spec test, which would indicate to me that either the assertion is too strong or the memory64 codegen has some hidden issue.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 07 2021 at 19:46):

abrown requested cfallin for a review on PR #3585.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 07 2021 at 23:31):

fitzgen submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 07 2021 at 23:41):

fitzgen submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 19:46):

abrown updated PR #3585 from assert-tmp-dsts to main.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 19:59):

abrown updated PR #3585 from assert-tmp-dsts to main.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 20:34):

fitzgen submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 20:34):

fitzgen created PR review comment:

Here and elsewhere in this commit:

(let ((unused_iflags Reg (writable_reg_to_reg (temp_writable_reg ty))))

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 20:34):

fitzgen edited PR review comment.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 20:37):

cfallin submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 20:37):

cfallin created PR review comment:

I'm actually not sure about that; this value is a placeholder for the iflags, so the type is logically unconnected to the iadd. (The register isn't actually used so it won't e.g. cause any additional pressure in whatever register class, but it seems odd to use ty here regardless.)

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 21:12):

abrown has marked PR #3585 as ready for review.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 21:34):

fitzgen created PR review comment:

Ah yes, that makes sense.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 21:34):

fitzgen submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 21:41):

cfallin submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 21:43):

cfallin submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 22:10):

fitzgen submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2021 at 22:10):

fitzgen merged PR #3585.


Last updated: Dec 23 2024 at 12:05 UTC