Today

chandlerc added a comment to D33535: [PM] Teach the PGO instrumentation pasess to run GlobalDCE before instrumenting code..

Thanks, landing this to fix the immediate issue (with comments addressed).

Wed, May 24, 11:43 PM
chandlerc committed rL303843: [PM/Unswitch] Fix a bug in the domtree update logic for the new unswitch.
[PM/Unswitch] Fix a bug in the domtree update logic for the new unswitch
Wed, May 24, 11:33 PM
chandlerc closed D32740: [PM/Unswitch] Fix a bug in the domtree update logic for the new unswitch pass. by committing rL303843: [PM/Unswitch] Fix a bug in the domtree update logic for the new unswitch.
Wed, May 24, 11:33 PM
vitalybuka committed rL303842: [compiler-rt] Change default of allow_user_segv_handler to true.
[compiler-rt] Change default of allow_user_segv_handler to true
Wed, May 24, 11:29 PM
vitalybuka closed D32443: [asan] Change default of allow_user_segv_handler to true by committing rL303842: [compiler-rt] Change default of allow_user_segv_handler to true.
Wed, May 24, 11:29 PM
ctopper committed rL303841: [MVT] Fix the identation of the start of the MVT class. NFC.
[MVT] Fix the identation of the start of the MVT class. NFC
Wed, May 24, 11:15 PM
EricWF retitled D33538: [coroutines] Support "coroutines" feature in module map requires clause from [coroutines] Support "coroutines" feature to module map "requires" to [coroutines] Support "coroutines" feature in module map requires clause.
Wed, May 24, 11:11 PM
baloghadamsoftware added a comment to D32592: [Analyzer] Iterator Checker - Part 1: Minimal Checker for a Simple Test Case.

Any comments regarding the last changes?

Wed, May 24, 11:08 PM
EricWF updated the diff for D33538: [coroutines] Support "coroutines" feature in module map requires clause.
  • Alphabetize newly added switch case.
Wed, May 24, 11:07 PM
dblaikie accepted D33533: [IR] Add an iterator and range accessor for the PHI nodes of a basic block..
Wed, May 24, 11:07 PM
chandlerc added a comment to D32740: [PM/Unswitch] Fix a bug in the domtree update logic for the new unswitch pass..

Over the past week or so, I think we have done enough playing around with incremental dominators at this point (and i believe jakub has working prototype code) that i feel confident enough that we can make it work, and so i'm not worried about this patch really.
Maybe it's code we destroy in a few months, but it doesn't seem worth holding up progress over.
It obviously could turn out i'm completely wrong (IE as we bring a design to upstream and start on a non-prototype, something becomes intractable), but we can still deal with it then.

Wed, May 24, 11:06 PM
baloghadamsoftware abandoned D32350: [Analyzer] Exception checker for misuse: uncaught/noncompliant throws.

Superseded by D33537

Wed, May 24, 11:06 PM
baloghadamsoftware added a comment to D33537: [clang-tidy] Exception Escape Checker.

Supersedes D32350

Wed, May 24, 11:06 PM
EricWF created D33538: [coroutines] Support "coroutines" feature in module map requires clause.
Wed, May 24, 11:06 PM
baloghadamsoftware created D33537: [clang-tidy] Exception Escape Checker.
Wed, May 24, 11:05 PM
davidxl accepted D33535: [PM] Teach the PGO instrumentation pasess to run GlobalDCE before instrumenting code..

lgtm

Wed, May 24, 11:00 PM
davide added inline comments to D33535: [PM] Teach the PGO instrumentation pasess to run GlobalDCE before instrumenting code..
Wed, May 24, 10:53 PM
davide accepted D33535: [PM] Teach the PGO instrumentation pasess to run GlobalDCE before instrumenting code..

I can see the bloat as we inline with a ridicolously high threshold before instrumenting without DCE'ing, so the patch seems a decent idea.
Also, DCE is basically free compile-time wise. If you want to improve I'd do it on the DCE side instead of the instrumentation side, but that needs a larger discussion which definitely doesn't fit in the margin of this page.

Wed, May 24, 10:52 PM
ctopper committed rL303840: [SelectionDAG] Fix off by one in a compare in getOperationAction..
[SelectionDAG] Fix off by one in a compare in getOperationAction.
Wed, May 24, 10:38 PM
grosser committed rL303839: Drop newline in docs builder to see if Polly docs are updated.
Drop newline in docs builder to see if Polly docs are updated
Wed, May 24, 10:38 PM
chandlerc added a reviewer for D33535: [PM] Teach the PGO instrumentation pasess to run GlobalDCE before instrumenting code.: davidxl.

Add David to the reviewer set in case he can look sooner. I'd love to land this and unblock testing of the new PM + PGO.

Wed, May 24, 10:36 PM
atanasyan updated the diff for D31528: [ELF][MIPS] Multi-GOT implementation.

Rebased against the trunk.

Wed, May 24, 10:36 PM · lld
singam-sanjay added a comment to D33414: [Polly] [PPCGCodeGeneration] do not allow LLVM intrinsics to be tracked as Subtree Values..

I believe this patch covers only a special case of a more general problem. The pointer to the intrinsic is added in

static int findReferencesInBlock(struct SubtreeReferences &References,
                                 const ScopStmt *Stmt, const BasicBlock *BB) {
  for (const Instruction &Inst : *BB)
    for (Value *SrcVal : Inst.operands()) {
      auto *Scope = References.LI.getLoopFor(BB);
      if (canSynthesize(SrcVal, References.S, &References.SE, Scope)) {
        References.SCEVs.insert(References.SE.getSCEVAtScope(SrcVal, Scope));

because it is used as an operand of a call instruction.

Wed, May 24, 10:36 PM
EricWF committed rL303838: Remove <experimental/coroutine> from the module map for now. It doesn't work….
Remove <experimental/coroutine> from the module map for now. It doesn't work…
Wed, May 24, 10:30 PM
EricWF created D33536: [coroutines] Bump __cpp_coroutines version.
Wed, May 24, 10:24 PM
chandlerc added a comment to D33535: [PM] Teach the PGO instrumentation pasess to run GlobalDCE before instrumenting code..

Forgot to mention, I also added a test to verify that GlobalDCE won't remove available_externally functions that are actually in use in the module so it won't impair subsequent optimization passes.

Wed, May 24, 10:15 PM
chandlerc created D33535: [PM] Teach the PGO instrumentation pasess to run GlobalDCE before instrumenting code..
Wed, May 24, 10:15 PM
EricWF committed rL303837: Disable the coroutines tests until Clang bumps __cpp_coroutines to reflect….
Disable the coroutines tests until Clang bumps __cpp_coroutines to reflect…
Wed, May 24, 10:12 PM
EricWF committed rL303836: Add <experimental/coroutine>.
Add <experimental/coroutine>
Wed, May 24, 9:37 PM
dberlin added a comment to D32740: [PM/Unswitch] Fix a bug in the domtree update logic for the new unswitch pass..

At this point, I think we have done enough playing around with incremental dominators at this point that i feel confident enough that we can make it work, and so i'm not worried about this patch really.
Maybe it's code we destroy in a few months, but it doesn't seem worth holding up progress over.
It obviously could turn out i'm completely wrong, but we can still deal with it then.

Wed, May 24, 9:12 PM
EricWF committed rL303835: Fix broken links on C++1z status page.
Fix broken links on C++1z status page
Wed, May 24, 9:09 PM
george.burgess.iv updated the diff for D32332: Add support for transparent overloadable functions in clang.

Fix the aforementioned issue; PTAL.

Wed, May 24, 9:07 PM
EricWF created D33534: [coroutines] Diagnose when promise types fail to declare either return_void or return_value..
Wed, May 24, 8:47 PM
chandlerc added inline comments to D33533: [IR] Add an iterator and range accessor for the PHI nodes of a basic block..
Wed, May 24, 8:12 PM
dblaikie added inline comments to D33533: [IR] Add an iterator and range accessor for the PHI nodes of a basic block..
Wed, May 24, 8:08 PM
chandlerc committed rL303834: [LegacyPM] Make the 'addLoop' method accept a loop to add rather than.
[LegacyPM] Make the 'addLoop' method accept a loop to add rather than
Wed, May 24, 8:01 PM
chandlerc closed D33528: [LegacyPM] Make the 'addLoop' method accept a loop to add rather than having it internally allocate the loop. by committing rL303834: [LegacyPM] Make the 'addLoop' method accept a loop to add rather than.
Wed, May 24, 8:01 PM
chandlerc added a comment to D33528: [LegacyPM] Make the 'addLoop' method accept a loop to add rather than having it internally allocate the loop..

This is fine assuming you audited all the callers (a quick grep shows just these two).

Wed, May 24, 7:56 PM
chandlerc created D33533: [IR] Add an iterator and range accessor for the PHI nodes of a basic block..
Wed, May 24, 7:55 PM
marshall committed rL303833: Add non-parallel version of for_each_n (+tests) from the Parallelism TS.
Add non-parallel version of for_each_n (+tests) from the Parallelism TS
Wed, May 24, 7:30 PM