[SingleAccretion]

I've been taking a look at some x86 codegen-related issues recently, and solving one of them requires supporting containment for sources of PUTARG_STK/FIELD_LIST nodes.

However, the code for determining how exactly should this be handled is not "trivial", so to support that change, I am first making this one, which introduces a new helper class, OperandDesc, and a method to create one, to help codegen determine what emitter methods should be called, in case there are multiple possibilities.

No diffs as expected.

@tannergooding curious to know of your opinion on this.

[ghost]

Tagging subscribers to this area: @JulieLeeMSFT
See info in area-owners.md if you want to be subscribed.

Issue Details

---

I've been taking a look at some x86 codegen-related issues recently, and solving one of them requires supporting containment for sources of PUTARG_STK/FIELD_LIST nodes.

However, the code for determining how exactly should this be handled is not "trivial", so to support that change, I am first making this one, which introduces a new helper class, OperandDesc, and a method to create one, to help codegen determine what emitter methods should be called, in case there are multiple possibilities.

@tannergooding curious to know of your opinion on this.

Author:	SingleAccretion
Assignees:	-
Labels:	area-CodeGen-coreclr
Milestone:	-

[SingleAccretion]

@dotnet/jit-contrib

[tannergooding]

@tannergooding curious to know of your opinion on this.

Talked with Single about this in Discord. I think simplifying some of the checks/logic here is general goodness. In particular I like how most of the checks around which form to pick are centralized. The individual paths then just need to worry about passing the args along to the right emit method.

It might be good to see if it can be extended to the areas outside of HWI as well, such as to inst_RV_TT.

[echesakov]

/azp run runtime-coreclr jitstress

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).