Page MenuHomePhabricator

thanm (Than McIntosh)
User

Projects

User does not belong to any projects.

User Details

User Since
May 6 2015, 11:54 AM (214 w, 3 d)

Recent Activity

Mar 14 2019

thanm committed rG9f96f1f17aa9: [SampleFDO] add suffix elision control for fcn names (authored by thanm).
[SampleFDO] add suffix elision control for fcn names
Mar 14 2019, 6:58 AM
thanm committed rL356146: [SampleFDO] add suffix elision control for fcn names.
[SampleFDO] add suffix elision control for fcn names
Mar 14 2019, 6:55 AM
thanm closed D58832: [SampleFDO] add suffix elision control for fcn names.
Mar 14 2019, 6:55 AM · Restricted Project
thanm added a comment to D58832: [SampleFDO] add suffix elision control for fcn names.

Thanks for reviewing.

Mar 14 2019, 6:16 AM · Restricted Project

Mar 11 2019

thanm updated the diff for D58832: [SampleFDO] add suffix elision control for fcn names.

Additional changes suggested by Wei.

Mar 11 2019, 1:27 PM · Restricted Project
thanm added inline comments to D58832: [SampleFDO] add suffix elision control for fcn names.
Mar 11 2019, 1:19 PM · Restricted Project

Mar 8 2019

thanm added a comment to D58832: [SampleFDO] add suffix elision control for fcn names.
In D58832#1422205, @wmi wrote:

We need function attribute only when we want to use different strategies for different functions. Do we need the function granularity control over the suffix elision policy?

Mar 8 2019, 12:50 PM · Restricted Project

Mar 6 2019

thanm updated the diff for D58832: [SampleFDO] add suffix elision control for fcn names.

Formatting and variable/function naming.

Mar 6 2019, 7:29 AM · Restricted Project
thanm added inline comments to D58832: [SampleFDO] add suffix elision control for fcn names.
Mar 6 2019, 7:28 AM · Restricted Project

Mar 1 2019

thanm added reviewers for D58832: [SampleFDO] add suffix elision control for fcn names: mtrofin, danielcdh, wmi, cherry.
Mar 1 2019, 9:06 AM · Restricted Project
thanm created D58832: [SampleFDO] add suffix elision control for fcn names.
Mar 1 2019, 9:06 AM · Restricted Project

Nov 30 2018

thanm committed rL347998: [CodeGen] Prefer static frame index for STATEPOINT liveness args.
[CodeGen] Prefer static frame index for STATEPOINT liveness args
Nov 30 2018, 8:25 AM
thanm closed D53889: [CodeGen] Prefer static frame index for STATEPOINT liveness args.
Nov 30 2018, 8:25 AM

Nov 26 2018

thanm committed rL347584: [CodeGen] Support custom format of stack maps.
[CodeGen] Support custom format of stack maps
Nov 26 2018, 10:47 AM
thanm closed D53892: [CodeGen] Support custom format of stack maps.
Nov 26 2018, 10:47 AM
thanm committed rL347569: [CodeGen] Take SPAdj into account for STATEPOINT liveness args.
[CodeGen] Take SPAdj into account for STATEPOINT liveness args
Nov 26 2018, 8:19 AM
thanm closed D53603: [CodeGen] Take SPAdj into account for STATEPOINT liveness args.
Nov 26 2018, 8:18 AM

Nov 16 2018

thanm committed rL347061: [CodeGen] Expose some data types and accessors from StackMaps.
[CodeGen] Expose some data types and accessors from StackMaps
Nov 16 2018, 8:51 AM
thanm closed D54224: [CodeGen] Expose some data types and accessors from StackMaps.
Nov 16 2018, 8:51 AM
thanm committed rL347050: [IRVerifier] Allow StructRet in statepoint.
[IRVerifier] Allow StructRet in statepoint
Nov 16 2018, 6:30 AM
thanm closed D53602: [IRVerifier] Allow StructRet in statepoint.
Nov 16 2018, 6:30 AM

Nov 7 2018

thanm committed rL346336: [X86] improve split-stack machine BB placement.
[X86] improve split-stack machine BB placement
Nov 7 2018, 9:44 AM
thanm closed D54123: [X86] improve split-stack machine BB placement.
Nov 7 2018, 9:44 AM
thanm updated the summary of D54123: [X86] improve split-stack machine BB placement.
Nov 7 2018, 9:10 AM
thanm updated the diff for D54123: [X86] improve split-stack machine BB placement.

Comment change only.

Nov 7 2018, 9:09 AM
thanm added a comment to D54123: [X86] improve split-stack machine BB placement.

Thanks for the review. I've fixed up the header comment.

Nov 7 2018, 9:09 AM

Nov 5 2018

thanm added a reviewer for D54123: [X86] improve split-stack machine BB placement: javed.absar.
Nov 5 2018, 2:24 PM
thanm added a reviewer for D54123: [X86] improve split-stack machine BB placement: cherry.
Nov 5 2018, 2:21 PM
thanm created D54123: [X86] improve split-stack machine BB placement.
Nov 5 2018, 2:20 PM

Oct 25 2018

thanm added a comment to D53698: [X86] Fix llc invocation on MIR test case.

LGTM. I ran the new recipe in the debugger and it is indeed exercising the PR37130 fix code correctly (lifetime marker set is different, but that doesn't matter in this case).

Oct 25 2018, 7:04 AM

Jul 11 2018

thanm added a comment to D49203: [PM] Use getAnalysisIfAvailable / addUsedIfAvailable with StackProtector.

Thanks for looking into this-- I've definitely run into the "Unable to schedule pass" issue but hadn't been able to sort out the cause.

Jul 11 2018, 1:33 PM

Jun 26 2018

thanm committed rL335604: [X86,ARM] Retain split-stack prolog check for sibling calls.
[X86,ARM] Retain split-stack prolog check for sibling calls
Jun 26 2018, 7:16 AM
thanm closed D48444: [X86] Insure split-stack prolog check is retained for sibling calls.
Jun 26 2018, 7:16 AM

Jun 25 2018

thanm updated the diff for D48444: [X86] Insure split-stack prolog check is retained for sibling calls.

Update commit message to reflect that this is for both ARM and x86.

Jun 25 2018, 7:12 AM
thanm added a comment to D48444: [X86] Insure split-stack prolog check is retained for sibling calls.

If there are no other comments or concerns on this patch, I plan to submit it tomorrow. Thanks -NM.

Jun 25 2018, 7:10 AM

Jun 21 2018

thanm added a reviewer for D48444: [X86] Insure split-stack prolog check is retained for sibling calls: cherry.
Jun 21 2018, 10:27 AM
thanm created D48444: [X86] Insure split-stack prolog check is retained for sibling calls.
Jun 21 2018, 10:25 AM

May 29 2018

thanm committed rL333416: StackColoring: better handling of statically unreachable code.
StackColoring: better handling of statically unreachable code
May 29 2018, 6:56 AM
thanm closed D47372: StackColoring: better handling of statically unreachable code.
May 29 2018, 6:56 AM
thanm updated the diff for D47372: StackColoring: better handling of statically unreachable code.

Clean up unused meta data in testcase. Fix typo in commit msg.

May 29 2018, 6:55 AM
thanm added a comment to D47372: StackColoring: better handling of statically unreachable code.

Thanks for reviewing. I'll remove the cruft from the testcase and check it in.

May 29 2018, 6:17 AM

May 25 2018

thanm added a reviewer for D47372: StackColoring: better handling of statically unreachable code: sdardis.
May 25 2018, 6:43 AM
thanm created D47372: StackColoring: better handling of statically unreachable code.
May 25 2018, 6:40 AM
thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

I agree, given the nature of the change. I'll prepare another patch.

May 25 2018, 6:39 AM

May 22 2018

thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

Sorry I'm not sure what to do in that case. Locally it runs fine on my machine, and I've tested it on another system and it's ok there. You're probably safe enough to re-land the change without the test case.

May 22 2018, 7:32 AM

May 18 2018

thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

Not immediately, can you forward me the links to some of the failed builds?

May 18 2018, 2:01 PM
thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

Hi Simon,

May 18 2018, 11:15 AM
thanm committed rL332742: Revert changes from D46265..
Revert changes from D46265.
May 18 2018, 10:51 AM
thanm closed D47061: Revert recently added MIR test (causes buildbot failures)..
May 18 2018, 10:51 AM
thanm added a comment to D47061: Revert recently added MIR test (causes buildbot failures)..

I don't think that removing only the test (especially when there is only 1 test) is correct. You should probably either fix the test, or revert the entire commit and resubmit for review once you have fixed the test.

May 18 2018, 10:39 AM
thanm added a comment to D47061: Revert recently added MIR test (causes buildbot failures)..

Failing buildbot run is:

May 18 2018, 6:13 AM
thanm created D47061: Revert recently added MIR test (causes buildbot failures)..
May 18 2018, 6:11 AM
thanm committed rL332707: StackColoring: better handling of statically unreachable code.
StackColoring: better handling of statically unreachable code
May 18 2018, 5:29 AM
thanm closed D46265: StackColoring: better handling of statically unreachable code.
May 18 2018, 5:29 AM

May 16 2018

thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

Thanks! Review comments and suggestions much appreciated. I will try to submit this later in the week.

May 16 2018, 8:59 AM
thanm updated the diff for D46265: StackColoring: better handling of statically unreachable code.

Incorporate changes suggested by Simon:

  • remove option to disable statically unreachable block elim
  • modify MIR test to include "-start-before isel"
May 16 2018, 6:37 AM

May 14 2018

thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

I've tried the revised patch, but I am still seeing flakiness when running this on my darwin machine.

May 14 2018, 12:32 PM

May 11 2018

thanm updated the diff for D46265: StackColoring: better handling of statically unreachable code.

Revised test case. New hidden option to turn off statically
unreachable block elimination.

May 11 2018, 8:36 AM
thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

Sorry for the delay, but this change looks ok, but the supplied test case is flaky and crashes sometimes on my machine:

Digging in with LLDB gives differing results (somewhat expected). Can you run valgrind on the run of llc for this test case and report your findings?

May 11 2018, 8:13 AM

May 10 2018

thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

Friendly ping... let me know if I need to recruit additional reviews.

May 10 2018, 10:19 AM

May 3 2018

thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

Ready for review at this point.

May 3 2018, 11:12 AM
thanm updated the diff for D46265: StackColoring: better handling of statically unreachable code.

Add testcase.

May 3 2018, 11:08 AM

May 1 2018

thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

@thanm You can take the reproducer in the bug report, produce LLVM-IR from it (-S -emit-llvm), then run it through llc with "-stop-before stack-slot-coloring" to get MIR output. The test file would then be invoked with the same options as before except you'd add '-start-before stack-slot-coloring'.

Thank you! I will look into doing that.

May 1 2018, 1:21 PM

Apr 30 2018

thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

@thanm You can take the reproducer in the bug report, produce LLVM-IR from it (-S -emit-llvm), then run it through llc with "-stop-before stack-slot-coloring" to get MIR output. The test file would then be invoked with the same options as before except you'd add '-start-before stack-slot-coloring'.

Apr 30 2018, 7:24 AM
thanm added a comment to D46265: StackColoring: better handling of statically unreachable code.

I would welcome suggestions on how to write a regression test / lit test for this issue. The tricky part seems to be insuring that a statically unreachable BB survives all the way through the llc pipeline to make it to stack coloring.

Apr 30 2018, 7:07 AM
thanm added reviewers for D46265: StackColoring: better handling of statically unreachable code: inouehrs, nadav, hfinkel, Eugene.Zelenko.
Apr 30 2018, 7:05 AM
thanm created D46265: StackColoring: better handling of statically unreachable code.
Apr 30 2018, 7:01 AM

Mar 1 2018

thanm committed rL326433: [CodeGen] fix argument attribute in lowering statepoint/patchpoint.
[CodeGen] fix argument attribute in lowering statepoint/patchpoint
Mar 1 2018, 5:34 AM
thanm closed D43832: [CodeGen] fix argument attribute in lowering statepoint/patchpoint.
Mar 1 2018, 5:34 AM

Sep 28 2017

thanm added a comment to D35844: Correct dwarf unwind information in function epilogue.

Remark: there are several outstanding LLVM bugs for this problem:

Sep 28 2017, 2:56 PM

Sep 27 2017

thanm committed rL314335: [CodeGen] Emit necessary .note sections for -fsplit-stack.
[CodeGen] Emit necessary .note sections for -fsplit-stack
Sep 27 2017, 12:35 PM
thanm closed D38051: [CodeGen] Emit necessary .note sections for -fsplit-stack by committing rL314335: [CodeGen] Emit necessary .note sections for -fsplit-stack.
Sep 27 2017, 12:35 PM

Jun 12 2017

thanm committed rL305193: StackColoring: smarter check for slot overlap.
StackColoring: smarter check for slot overlap
Jun 12 2017, 7:56 AM
thanm closed D31583: StackColoring: smarter check for slot overlap by committing rL305193: StackColoring: smarter check for slot overlap.
Jun 12 2017, 7:56 AM

Jun 8 2017

thanm accepted D31583: StackColoring: smarter check for slot overlap.

FWIW I tried a self-build of clang with and without the change and looked at the -stats output to see what sort of improvement there is. Looks pretty modest (<1%) but it is positive overall.

Jun 8 2017, 5:32 AM

Jun 6 2017

thanm requested changes to D31583: StackColoring: smarter check for slot overlap.

I pulled your patch and did a bootstrap build with it, no issues. Overall this seems like an improvement over the initial patch. Please see inline comments.

Jun 6 2017, 5:28 AM

May 15 2017

thanm added a comment to D31583: StackColoring: smarter check for slot overlap.

Ready and just waiting for review.

May 15 2017, 8:12 AM

Apr 4 2017

thanm added a comment to D31583: StackColoring: smarter check for slot overlap.

Hi Ariel,

Apr 4 2017, 1:58 PM

Jun 1 2016

thanm committed rL271412: Better fix for PR27903..
Better fix for PR27903.
Jun 1 2016, 11:01 AM
thanm closed D20739: Better fix for PR27903. by committing rL271412: Better fix for PR27903..
Jun 1 2016, 11:01 AM
thanm updated the diff for D20739: Better fix for PR27903..

Incorporate code review suggestions from Wei.

Jun 1 2016, 9:41 AM
thanm added inline comments to D20739: Better fix for PR27903..
Jun 1 2016, 9:40 AM

May 27 2016

thanm retitled D20739: Better fix for PR27903. from to Better fix for PR27903..
May 27 2016, 11:06 AM
thanm committed rL271003: Disable lifetime-start-on-first-use analysis..
Disable lifetime-start-on-first-use analysis.
May 27 2016, 8:34 AM
thanm closed D20731: Disable lifetime-start-on-first-use analysis. by committing rL271003: Disable lifetime-start-on-first-use analysis..
May 27 2016, 8:34 AM
thanm updated the diff for D20731: Disable lifetime-start-on-first-use analysis..

Incorporate code review suggestions by Teresa.

May 27 2016, 8:31 AM
thanm added a comment to D20731: Disable lifetime-start-on-first-use analysis..

Thanks. Fixed the issues you cited, will post a revised patch.

May 27 2016, 8:31 AM
thanm retitled D20731: Disable lifetime-start-on-first-use analysis. from to Disable lifetime-start-on-first-use analysis..
May 27 2016, 8:11 AM

May 26 2016

thanm added a comment to D20453: Remove optnone/opt-bisect check from StackColoring.

LGTM

May 26 2016, 5:36 PM
thanm added a comment to D20453: Remove optnone/opt-bisect check from StackColoring.

Hi Andy,
I think I agree with probinson about the need for removing the markers in the opt-none case. I think the most expedient thing to do would be to make the skipFunction() check at the same point in the code where "DisableColoring" is checked (line 1008), this way the "removeAllMarkers()" helper will be called on the way out. This has the minor compile-time drawback that there is some analysis done to get to this point, but that seems like a minor concern for bisections purposes.
Thanks, Than

May 26 2016, 3:56 PM

May 24 2016

thanm closed D18827: Rework/enhance stack coloring data flow analysis..

Submitted in r270559.

May 24 2016, 9:08 AM
thanm retitled D20569: Debugging enhancements for BFI cfg dot graphs. from to Debugging enhancements for BFI cfg dot graphs..
May 24 2016, 9:06 AM
thanm committed rL270559: Rework/enhance stack coloring data flow analysis..
Rework/enhance stack coloring data flow analysis.
May 24 2016, 6:30 AM

May 22 2016

thanm added a comment to D18827: Rework/enhance stack coloring data flow analysis..

Another friendly ping... I'd like to get this submitted at some point.

May 22 2016, 12:49 PM

May 18 2016

thanm added a comment to D18827: Rework/enhance stack coloring data flow analysis..

Friendly ping...

May 18 2016, 5:50 PM

May 16 2016

thanm updated the diff for D18827: Rework/enhance stack coloring data flow analysis..

Incorporate fixes suggested by Wei.

May 16 2016, 11:46 AM
thanm added a comment to D18827: Rework/enhance stack coloring data flow analysis..

Will post a revised patch shortly.

May 16 2016, 11:46 AM
thanm updated the diff for D18827: Rework/enhance stack coloring data flow analysis..

Improvements to comments.

May 16 2016, 9:19 AM
thanm added a comment to D18827: Rework/enhance stack coloring data flow analysis..

Wei, you suggested adding an assert that triggers if an instruction is found otuside the the start/end lifetime that accesses memory.

May 16 2016, 9:16 AM

May 12 2016

thanm added a comment to D18827: Rework/enhance stack coloring data flow analysis..
In D18827#429001, @wmi wrote:

Thanks for the detailed implementation notes and spec2006 data. The spec2006 data is a overall number which looks good. I guess for individual file there is no regression anymore, right?

May 12 2016, 4:48 PM