externref stack maps · wasmtime · Zulip Chat Archive

no intentions for the JIT to read the registry -- but it is needed for GC, and that means wasmtime-runtime needs a way to get it from inside a libcall, and hanging it off the VMContext is the easiest way to do that

fitzgen (he/him) (Jun 12 2020 at 18:44):

@Alex Crichton

Alex Crichton (Jun 12 2020 at 18:44):

ah ok

Alex Crichton (Jun 12 2020 at 18:44):

I'm personally less of a fan than others I think about making code future-proof

Alex Crichton (Jun 12 2020 at 18:45):

I would personally just go ahead and drop it from this PR if it's not needed yet, saving it for later when it's needed

Alex Crichton (Jun 12 2020 at 18:45):

I'm always afraid that the actual feature would look different than a pre-implementation b/c requirements change over time or something like that

Alex Crichton (Jun 12 2020 at 18:45):

minor point though

fitzgen (he/him) (Jun 12 2020 at 18:46):

when next == end, we need to do a libcall to insert the reference into the table, which will basically do:

gc to maybe free space
insert into the newly freed space, or allocate a new chunk and insert into that

fitzgen (he/him) (Jun 12 2020 at 18:46):

I'm p sure of the requirement in this case

Alex Crichton (Jun 12 2020 at 18:47):

oh sorry yeah next/end make sense

Alex Crichton (Jun 12 2020 at 18:47):

but the registry for GC far in the future is what I mean

Alex Crichton (Jun 12 2020 at 18:47):

next/end is like "the next PR needs this" which is fine

Alex Crichton (Jun 12 2020 at 18:47):

but afaik we don't even have plans yet for GC in the future

fitzgen (he/him) (Jun 12 2020 at 18:48):

when I say "gc" I mean the wasmtime_runtime::gc function that is in the PR now

fitzgen (he/him) (Jun 12 2020 at 18:48):

not the wasm GC proposal

Alex Crichton (Jun 12 2020 at 18:48):

oh wait it's actually used?

Alex Crichton (Jun 12 2020 at 18:48):

I think I missed where it was used?

fitzgen (he/him) (Jun 12 2020 at 18:49):

table.get of a reference type will have a fast path where it can bump into the table if it isn't full, but if next == end, then the current chunk is full, so it will call out to host code via a libcall, that will do gc(); infallible_insert_into_table() as described above

this isn't in the current PR, but coming in the next PR