Page MenuHomePhabricator

asbirlea (Alina Sbirlea)
Google

Projects

User does not belong to any projects.

User Details

User Since
Feb 10 2016, 9:51 AM (257 w, 5 d)

Recent Activity

Thu, Jan 14

asbirlea added a comment to D94363: [AA] Pass AAResults through AAQueryInfo.

Do you have an idea if there is overhead of adding the AAR object to each query (each query info object)? (vs the setAAResults being done once for all the AAs available).

Thu, Jan 14, 4:50 PM · Restricted Project

Wed, Jan 13

asbirlea accepted D90094: [BasicAA] Handle recursive queries more efficiently (NFCI).

lgtm.
Runtime impact results look good.

Wed, Jan 13, 4:15 PM · Restricted Project
asbirlea accepted D94559: [NewPM] Only non-trivially loop unswitch at -O3 and for non-optsize functions.

Thank you for confirming @SjoerdMeijer.

Wed, Jan 13, 1:11 PM · Restricted Project

Tue, Jan 12

asbirlea added a comment to D94376: [MemCpyOpt] Enable MemorySSA by default.

Could you test clang as well, and update test?

Tue, Jan 12, 3:42 PM · Restricted Project
asbirlea added a comment to D94448: [NewPM] Run non-trivial loop unswitching under -O2/3/s/z.

I'm curious if the compile time impact is complemented by impact in run time. Looking into that for O3.

Tue, Jan 12, 3:00 PM · Restricted Project

Mon, Jan 11

asbirlea accepted D94448: [NewPM] Run non-trivial loop unswitching under -O2/3/s/z.

lgtm.

Mon, Jan 11, 2:28 PM · Restricted Project

Thu, Jan 7

asbirlea added a comment to D93743: [NewPM] Run ObjC ARC passes.

The changes LGTM, but I'd rather defer final review to ahatanak@.
It would be good to have some testing but it could be addressed in a separate patch, potentially by folks more familiar with the ObjCARC passes.

Thu, Jan 7, 3:26 PM · Restricted Project, Restricted Project
asbirlea added a reviewer for D93929: [NewPM][Hexagon] Fix HexagonVectorLoopCarriedReusePass position in pipeline: quic_aankit.
Thu, Jan 7, 3:17 PM · Restricted Project

Wed, Jan 6

asbirlea accepted D92486: Set option default for enabling memory ssa for new pass manager loop sink pass to true..

I believe for the NPM this should work fine.
@nikic @fhahn Any objections?

Wed, Jan 6, 6:30 PM · Restricted Project
asbirlea accepted D94176: [SplitEdge] Add new parameter to SplitEdge to name the newly created basic block.

lgtm.

Wed, Jan 6, 6:21 PM · Restricted Project
asbirlea added a comment to D93503: [GVN] Propagate llvm.access.group metadata of loads.

Looks reasonable to me as well, that the metadata info should hold at the same loop level.

Wed, Jan 6, 5:54 PM · Restricted Project
asbirlea committed rG63aeaf754a78: [DominatorTree] Add support for mixed pre/post CFG views. (authored by asbirlea).
[DominatorTree] Add support for mixed pre/post CFG views.
Wed, Jan 6, 2:54 PM
asbirlea closed D93371: [DominatorTree] Add support for mixed pre/post CFG views..
Wed, Jan 6, 2:53 PM · Restricted Project

Tue, Jan 5

asbirlea accepted D91936: [BasicAA] Fix BatchAA results for phi-phi assumptions.

Apologies for the delay, I spent some more time understanding how the assumptions changed in the new implementation.

Tue, Jan 5, 5:50 PM · Restricted Project
asbirlea accepted D94085: [simd_ops] Build tests with -fpic.
Tue, Jan 5, 10:13 AM · Restricted Project
asbirlea added a comment to D94085: [simd_ops] Build tests with -fpic.

I'll look into regenerating the bitcode files, but in the mean time this looks good.

Tue, Jan 5, 10:13 AM · Restricted Project

Wed, Dec 30

asbirlea added a comment to rGb980bed34b96: [MSSAUpdater] Skip renaming when inserting def in unreachable block..

Thank you for pushing the fix. Based on a quick look, it seems reasonable to not do the renaming for unreachable blocks. I'll look in more depth next week.

Wed, Dec 30, 10:36 AM

Wed, Dec 23

asbirlea accepted D93577: [NFC][LICM] Minor improvements to debug output.

Thanks!

Wed, Dec 23, 4:49 PM · Restricted Project

Mon, Dec 21

asbirlea added a comment to D93577: [NFC][LICM] Minor improvements to debug output.

This seems like a useful utility to add to BasicBlock itself if it doesn't exit? (e.g. getNonEmptyName or getNameOrAsOperand).

Mon, Dec 21, 6:57 PM · Restricted Project
asbirlea updated the diff for D93371: [DominatorTree] Add support for mixed pre/post CFG views..

Address comments.

Mon, Dec 21, 6:48 PM · Restricted Project

Dec 16 2020

asbirlea added a comment to rG47321c311bdb: [X86][SSE] combineReductionToHorizontal - add vXi8 ISD::MUL reduction handling….

Many thanks!

Dec 16 2020, 11:00 AM

Dec 15 2020

asbirlea requested review of D93371: [DominatorTree] Add support for mixed pre/post CFG views..
Dec 15 2020, 10:37 PM · Restricted Project
asbirlea added a comment to rG47321c311bdb: [X86][SSE] combineReductionToHorizontal - add vXi8 ISD::MUL reduction handling….

We've bisected a test failure to this revision.

Dec 15 2020, 10:00 PM

Dec 14 2020

asbirlea committed rG5a2d954671e9: [NFC] Remove stray comment. (authored by asbirlea).
[NFC] Remove stray comment.
Dec 14 2020, 11:20 AM

Dec 11 2020

asbirlea added reviewers for D88714: [NewPM][CGSCC] Handle newly added functions not directly referenced by existing nodes: modocache, jdoerfert.
Dec 11 2020, 3:14 PM · Restricted Project

Dec 10 2020

asbirlea accepted D93002: [NPM] Support -fmerge-functions.

Thank you for adding this support.

Dec 10 2020, 10:26 AM · Restricted Project, Restricted Project

Dec 9 2020

asbirlea committed rG1289835a96eb: [MemorySSA/docs] Extend MemorySSA documentation. (authored by asbirlea).
[MemorySSA/docs] Extend MemorySSA documentation.
Dec 9 2020, 6:01 PM

Dec 8 2020

asbirlea added a comment to D92796: Pin -loop-reduce to legacy PM.

Currently LSR is not tested with the NPM in any circumstance.
There are no tests with -passes=loop-reduce, and LSR is only used in the codegen pass manager, so even with -DEXPERIMENTAL_NEW_PASS_MANAGER=ON, it's never in a NPM context.

Dec 8 2020, 1:23 PM · Restricted Project
asbirlea accepted D92818: [NFC] Rename IsCodeGenPass to ShouldPinPassToLegacyPM.

Fix clang-tidy warning (lowercase method names; the file needs a lot of other functions names fixed, but might as well start with these)

Dec 8 2020, 1:19 PM · Restricted Project
asbirlea updated subscribers of D92796: Pin -loop-reduce to legacy PM.
Dec 8 2020, 1:09 PM · Restricted Project
asbirlea accepted D92796: Pin -loop-reduce to legacy PM.

Discussed this offline, and proposed this solution to make progress and resolve the failing tests.
This does need exploring further though as:

  1. the behavior of current LSR will no longer be tested in the new pass manager, due to LCSSA being run as a canonicalization pass.
  2. adding it to a loop pass manager in the npm will be problematic, because it doesn't preserve LCSSA.
Dec 8 2020, 1:09 PM · Restricted Project

Dec 7 2020

asbirlea accepted D92436: [Time-report] Add a flag -ftime-report={per-pass,per-pass-run} to control the pass timing aggregation.

lgtm.

Dec 7 2020, 2:02 PM · Restricted Project, Restricted Project
asbirlea added a comment to D92623: [inliner] Apply nomerge attribute to all call sites inside inlined function..

According to: https://clang.llvm.org/docs/AttributeReference.html#nomerge, the existance of the attribute should prevent call site merging, so the change makes sense.
However, in the test example then, if there was branching and two @f calls inside @bar, it would be correct to merge the @f call sites (no attribute preventing that inside) when processing @bar. But after inlining, that's no longer allowed if the attribute is distributed to all callsites inside the function.
This is a contradiction that needs clarification from folks more familiar with the Inliner.

Dec 7 2020, 1:57 PM · Restricted Project
asbirlea added reviewers for D92623: [inliner] Apply nomerge attribute to all call sites inside inlined function.: mtrofin, tejohnson, chandlerc.
Dec 7 2020, 1:55 PM · Restricted Project
asbirlea added a comment to D92488: [LICM] Add a maximum for the number of instructions to be hoisted.

Hard reject. This patch will not be accepted. This is a major design change.

Hoisting operations out of loops is treated as a canonicalization transform in LLVM. If the register pressure is problematic, the MachineLICM pass is supposed to undo the transformation. If you have cases where that is not happening, please explore why.

If after doing so, you still feel this is the right approach, please raise the design change on llvm-dev.

Dec 7 2020, 1:36 PM · Restricted Project

Dec 1 2020

asbirlea added a comment to D68484: [PATCH 01/27] [noalias] LangRef: noalias intrinsics and ptr_provenance documentation..

A few comments inline.

Dec 1 2020, 7:40 AM · Restricted Project

Nov 30 2020

asbirlea accepted D89207: [MemCpyOpt] Port to MemorySSA.

Yes.

Nov 30 2020, 6:03 PM · Restricted Project

Nov 25 2020

asbirlea accepted D91525: [LoopLoadElim] Make sure all loops are in simplify form. PR48150.
Nov 25 2020, 8:30 AM · Restricted Project

Nov 24 2020

asbirlea added inline comments to D87293: [GVN] Fix undef incoming value for phi node when new loop exit block created..
Nov 24 2020, 6:06 PM · Restricted Project
asbirlea added a comment to D91804: [NewPM] Add pipeline EP callback after initial frontend cleanup.

EarlySimplificationEPCallback?

Nov 24 2020, 5:06 PM · Restricted Project
asbirlea accepted D90539: Make CallInst::updateProfWeight emit i32 weights instead of i64.

Unblocking the fixing of the tests under NPM. Additional changes can be done as follow ups.

Nov 24 2020, 4:36 PM · Restricted Project
asbirlea accepted D91936: [BasicAA] Fix BatchAA results for phi-phi assumptions.

This is great, thank you for this fix!

Nov 24 2020, 4:29 PM · Restricted Project
asbirlea accepted D91914: [BasicAA] Generalize recursive phi alias analysis.

sgtm too.

Nov 24 2020, 4:10 PM · Restricted Project
asbirlea accepted D91482: [BasicAA] Remove unnecessary known size requirement.

Ack.

Nov 24 2020, 3:05 PM · Restricted Project
asbirlea accepted D91649: [AA] Split up LocationSize::unknown().

lgtm.
Please wait in case fhahn@ has additional comments.

Nov 24 2020, 3:03 PM · Restricted Project
asbirlea added a reviewer for D91649: [AA] Split up LocationSize::unknown(): fhahn.
Nov 24 2020, 3:01 PM · Restricted Project

Nov 23 2020

asbirlea added a comment to D91649: [AA] Split up LocationSize::unknown().

One more comment, the others look good.

Nov 23 2020, 6:00 PM · Restricted Project
asbirlea added a comment to D91804: [NewPM] Add pipeline EP callback after initial frontend cleanup.

I don't think "ModuleOptimizer" in the name makes sense, the association can come from the "Early" part. Either keep the naming as is, or "EarlyAfterStart" if you want to define order against the other callbacks.
ychen@ so you have a better naming suggestion?

Nov 23 2020, 12:48 PM · Restricted Project
asbirlea accepted D91741: [NPM] Share pass building options with legacy PM.

Thank you for the cleanup!

Nov 23 2020, 12:42 PM · Restricted Project
asbirlea accepted D91685: [test] Pin tests using -dot-callgraph to legacy PM.
Nov 23 2020, 11:30 AM · Restricted Project
asbirlea accepted D89587: [CGSCC] Detect devirtualization in more cases.
Nov 23 2020, 11:17 AM · Restricted Project

Nov 20 2020

asbirlea added a comment to D89207: [MemCpyOpt] Port to MemorySSA.

Since this is using MSSA walker API with a MemoryLocation (which does not cache), could this become costly in a pathologically constructed example? If so, could we have a (reasonably large) upper bound on the number of getClobbering (Access, Location) calls to avoid such cases?

I've started implementing this (by returning a DoNothingMemorySSAWalker when a limit is hit), but ended up wondering if it really makes sense: I believe that the number of MSSA walker queries we do in this pass is roughly proportional to the number of memsets/memcpys in the IR. For very big functions we may end up performing many queries, but the overall runtime should still be linear (each individual query is limited by the walker limit). Having a limit would make sense to me if something here had quadratic runtime (like e.g. the AST limit in LICM), but I don't think that's the case here.

Nov 20 2020, 2:41 PM · Restricted Project

Nov 19 2020

asbirlea accepted D91811: Port -lower-matrix-intrinsics-minimal to NPM.
Nov 19 2020, 3:08 PM · Restricted Project
asbirlea added inline comments to D91811: Port -lower-matrix-intrinsics-minimal to NPM.
Nov 19 2020, 2:00 PM · Restricted Project
asbirlea added a reviewer for D91767: [DomTree][NFC] Introduce function that finds common dom of multiple instructions: kuhar.
Nov 19 2020, 1:49 PM · Restricted Project
asbirlea added a comment to D91767: [DomTree][NFC] Introduce function that finds common dom of multiple instructions.

Algorithm lgtm.
+ @kuhar for additional review.

Nov 19 2020, 1:48 PM · Restricted Project
asbirlea added a comment to D91649: [AA] Split up LocationSize::unknown().
  • LocationSize::unknownMaybeNegative() seems somewhat ambiguous in isolation; unless I see the other name or have the context from this patch. I'm either keep LocationSize::unknown or make it even more verbose LocationSize::unknownPositiveOrNegative().

I'd like to not reuse the existing unknown() name, to force all users (including out of tree users) to make a decision one way or another. Not particularly attached to the names though. We could also drop the "unknown" prefix and make them LocationSize::nonNegative() and LocationSize::positiveOrNegative() for example.

Generally I'm not really fond of "negative" sizes as a concept, it just seemed like the most concise way to explain it. It would be more accurate to say that the size is always positive, but there may be a negative offset. So maybe LocationSize::nonNegative(), LocationSize::withNonNegativeOffset() and LocationSize::withAnyOffset(), where the first two are aliases but can be used according to desired semantics. Can't say I really like this either though.

Or maybe LocationSize::afterPointer() and LocationSize::beforeOrAfterPointer().

Nov 19 2020, 1:42 PM · Restricted Project

Nov 18 2020

asbirlea accepted D91743: [NPM] Remove -enable-npm-optnone flag.
Nov 18 2020, 2:46 PM · Restricted Project, Restricted Project

Nov 17 2020

asbirlea accepted D91476: [DCE] Always get TargetLibraryInfo.
Nov 17 2020, 8:23 PM · Restricted Project
asbirlea added a comment to D89587: [CGSCC] Detect devirtualization in more cases.

Could you clarify why a map and not a SmallVector?

I was expecting the indirect calls to be added by all CGSCC passes, e.g. lib/Transforms/IPO/Inliner.cpp:750. Adding the calls in updateCG... instead, introduces a single point of failure so less chances of having passes forget to add them; the one thing I'm not clear is if the updateCG... call may be added too late in a CGSCC pass; for the inliner, it seems fine (thank you for adding the tests) but I'm not clear if this will hold for other passes. I don't have a use case now, just making a note to verify this for coroutines.
The check that all accesses are still indirect on the return looks good.

It is definitely possible to have a call to updateCG... miss a devirtualization in a CGSCC pass if an indirect call is introduced and then promoted before the call to updateCG..., but the inliner looks ok. I think other CGSCC passes don't really devirtualize much. As long as we call updateCG... after passes that can devirtualize (inliner, function passes like instcombine, etc) we should be good. And we can always revisit if we find missed cases.

Nov 17 2020, 5:14 PM · Restricted Project
asbirlea accepted D91027: [BasicAA] Generalize base offset modulus handling.

This makes sense to me too. The additional comment was helpful with the gcd/difference condition.

Nov 17 2020, 5:10 PM · Restricted Project
asbirlea added inline comments to D91482: [BasicAA] Remove unnecessary known size requirement.
Nov 17 2020, 4:46 PM · Restricted Project
asbirlea added a comment to D91649: [AA] Split up LocationSize::unknown().
  • LocationSize::unknownMaybeNegative() seems somewhat ambiguous in isolation; unless I see the other name or have the context from this patch. I'm either keep LocationSize::unknown or make it even more verbose LocationSize::unknownPositiveOrNegative(). I think the most restrictive case ( LocationSize::unknownMaybeNegative() or renamed version) should be the default ~uint64_t(0), but don't feel strongly about it.
  • clang-format
Nov 17 2020, 4:38 PM · Restricted Project

Nov 16 2020

asbirlea added inline comments to D87531: [LoopNest] Extend `LPMUpdater` and adaptor to handle loop-nest passes.
Nov 16 2020, 5:39 PM · Restricted Project
asbirlea added a comment to D87045: [LoopNest] Handle loop-nest passes in LoopPassManager.

No objections.

Nov 16 2020, 5:34 PM · Restricted Project
asbirlea accepted D91383: [BasicAA] Make alias GEP positive offset handling symmetric.

As talked offline, this lgtm.

Nov 16 2020, 5:32 PM · Restricted Project
asbirlea added a comment to D89587: [CGSCC] Detect devirtualization in more cases.

Could you clarify why a map and not a SmallVector?

Nov 16 2020, 5:29 PM · Restricted Project

Nov 13 2020

asbirlea added inline comments to D91383: [BasicAA] Make alias GEP positive offset handling symmetric.
Nov 13 2020, 12:51 AM · Restricted Project

Nov 12 2020

asbirlea accepted D91383: [BasicAA] Make alias GEP positive offset handling symmetric.

I wouldn't be surprised if there are other asymmetries, thank you for digging further here.

Nov 12 2020, 5:06 PM · Restricted Project

Nov 11 2020

asbirlea added a comment to D89158: [NewPM] Provide method to run all pipeline callbacks, used for -O0.

AFAICT all issues were addressed.
@Meinersbur: are there any concerns left with this patch?

Nov 11 2020, 1:31 PM · Restricted Project, Restricted Project

Nov 10 2020

asbirlea accepted D91046: [CGSCC][Inliner] Handle new non-trivial edges in updateCGAndAnalysisManagerForPass.

This is great, thank you!

Nov 10 2020, 3:20 PM · Restricted Project

Nov 9 2020

asbirlea accepted D91095: [NewPM] Port -separate-const-offset-from-gep.
Nov 9 2020, 4:58 PM · Restricted Project

Nov 5 2020

asbirlea added a reviewer for D90870: [NFC][NewPM] Reuse PassBuilder callbacks with -O0: leonardchan.
Nov 5 2020, 1:07 PM · Restricted Project

Nov 4 2020

asbirlea accepted D90638: Introduce -dot-cfg-mssa option which creates dot-cfg style file with mssa comments included in source.

Resulting dot file for the test looks good.

Nov 4 2020, 6:40 PM · Restricted Project
asbirlea accepted D90784: [GVN] Fix MemorySSA update when replacing assume(false) with stores..

Thank you for the fix.

Nov 4 2020, 6:30 PM · Restricted Project
asbirlea accepted D89158: [NewPM] Provide method to run all pipeline callbacks, used for -O0.
Nov 4 2020, 1:53 PM · Restricted Project, Restricted Project
asbirlea accepted D90249: Add options to enable memoryssa for loopsink, expand loopsink testing and fix exposed bug in LICM.

lgtm.
Looking forward to having it enabled in a future patch.

Nov 4 2020, 11:39 AM · Restricted Project

Nov 3 2020

asbirlea accepted D90207: Port print-must-be-executed-contexts and print-mustexecute to NPM.

lgtm

Nov 3 2020, 6:40 PM · Restricted Project
asbirlea added reviewers for D90539: Make CallInst::updateProfWeight emit i32 weights instead of i64: wmi, mtrofin.
Nov 3 2020, 6:38 PM · Restricted Project
asbirlea accepted D88782: [MemorySSA] Use provided memory location even if instruction is call.

Thank you for detailed tests, this looks good!

Nov 3 2020, 6:33 PM · Restricted Project
asbirlea accepted D90545: [NewPM] Don't run before pass instrumentation on required passes.
Nov 3 2020, 6:25 PM · Restricted Project
asbirlea added a comment to D90249: Add options to enable memoryssa for loopsink, expand loopsink testing and fix exposed bug in LICM.

Thank you, just a couple of small comments below.

Nov 3 2020, 6:03 PM · Restricted Project

Nov 2 2020

asbirlea added a comment to D89207: [MemCpyOpt] Port to MemorySSA.

Thank you for working on this, the compile times look very promising and the tests results are great!

Nov 2 2020, 6:43 PM · Restricted Project
asbirlea requested review of D90660: [MemorySSA] Extend walker API taking in a MemoryLocation to take in a Stop access..
Nov 2 2020, 6:26 PM · Restricted Project
asbirlea added a comment to D90249: Add options to enable memoryssa for loopsink, expand loopsink testing and fix exposed bug in LICM.

The current behavior is:

  • for the NPM, always use MSSA for LoopSink, which is a Function pass.
  • for the LPM, LoopSink is a Loop pass and it will use MSSA conditioned on the EnableMSSALoopDependency flag which is set to true, hence this patch is switching loop sink to use MSSA.
Nov 2 2020, 6:09 PM · Restricted Project
asbirlea committed rGf514b32a8993: [LICM] Add assert of AST/MSSA exclusiveness. (authored by asbirlea).
[LICM] Add assert of AST/MSSA exclusiveness.
Nov 2 2020, 6:05 PM
asbirlea accepted D90638: Introduce -dot-cfg-mssa option which creates dot-cfg style file with mssa comments included in source.
Nov 2 2020, 5:24 PM · Restricted Project
asbirlea added a reviewer for D90638: Introduce -dot-cfg-mssa option which creates dot-cfg style file with mssa comments included in source: dblaikie.
Nov 2 2020, 4:52 PM · Restricted Project
asbirlea added a comment to D90094: [BasicAA] Handle recursive queries more efficiently (NFCI).

This looks like a nice approach to clean up the recursion, thank you for working on this!

Nov 2 2020, 2:14 PM · Restricted Project

Oct 30 2020

asbirlea added a comment to D90249: Add options to enable memoryssa for loopsink, expand loopsink testing and fix exposed bug in LICM.

I guess my question was more regarding consistency. Why not use MSSA in the LPM as well and drop the AST usage entirely? Or provide a flag that can be used to switch between MSSA and AST version (see LICM or DSE for similar flags using MSSA or alternative analyses), just in case there are users of the AST version.

Oct 30 2020, 3:04 PM · Restricted Project
asbirlea accepted D90482: [NFC intended] Refactor SinkAndHoistLICMFlags to allow others to construct without exposing internals.

Thank you!

Oct 30 2020, 2:59 PM · Restricted Project

Oct 29 2020

asbirlea added a comment to D90249: Add options to enable memoryssa for loopsink, expand loopsink testing and fix exposed bug in LICM.
  • Could you separate the SinkHoistFlag refactoring into an NFC?
  • Could you clang-format and fix lints?
  • LICM changes look good. AFAICT the MSSA changes in LoopSink are accurate, but it may need more testing than the existing unit tests.
  • Why do you want the legacy pass manager to use the AST and the new pass manager MSSA?
Oct 29 2020, 6:37 PM · Restricted Project
asbirlea requested changes to D89991: [AliasAnalysis] Fix a bug in alias analysis interaction with CSE w/ memorySSA.

I checked in the changes adding AAQI as authored by you.
The change clearing the cache is not right, as discussed on llvm-dev, so I think this patch is not obsolete.

Oct 29 2020, 6:10 PM · Restricted Project
asbirlea committed rG9d93b150c916: [AA] Pass query info. (authored by asbirlea).
[AA] Pass query info.
Oct 29 2020, 6:08 PM
asbirlea accepted D90437: [NFC] Clean up PassBuilder.

Neat!

Oct 29 2020, 6:00 PM · Restricted Project, Restricted Project
asbirlea accepted D90012: [PruneEH] Pin tests to legacy PM.

lgtm

Oct 29 2020, 5:39 PM · Restricted Project
asbirlea accepted D89650: [DSE] Remove noop stores after killing stores for a MemoryDef..
Oct 29 2020, 5:37 PM · Restricted Project
asbirlea accepted D90371: [DSE] Improve partial overlap detection.

This LGTM.
I'd like to understand better what you mean by not calling partialOverwrite. If the functionality in that call would be included in the isOverwrite call, note that there are call-paths that would do more work than they do now, and I thought there was a previous refactoring done to avoid this.
I may be mis-understanding though, so feel free to send the additional changes and we can discuss on those.

Oct 29 2020, 4:37 PM · Restricted Project

Oct 28 2020

asbirlea added a reviewer for D85603: IR: Add convergence control operand bundle and intrinsics: resistor.
Oct 28 2020, 2:32 PM · Restricted Project
asbirlea accepted D90098: [BasicAA] Don't pass through AA metadata (NFCI).

This is an interesting point. IIUC, your argument is that the entry point into AA is always through AAResults and can rely on the invocation result for TBAA or ScopedAA from there, while the recursive invocations into these same analyses cannot yield stricter results. I cannot think of a case when this is not true.

Oct 28 2020, 1:28 PM · Restricted Project