rovka (Diana Picus)
User

Projects

User does not belong to any projects.

User Details

User Since
Apr 21 2016, 3:27 AM (57 w, 2 d)

Recent Activity

Wed, May 24

rovka committed rL303747: Revert "[SCEV] Do not fold dominated SCEVUnknown into AddRecExpr start".
Revert "[SCEV] Do not fold dominated SCEVUnknown into AddRecExpr start"
Wed, May 24, 7:16 AM

Thu, May 18

rovka committed rL303338: Fixup r303324 - temporary disable stndup interceptor, due to r302781 being buggy.
Fixup r303324 - temporary disable stndup interceptor, due to r302781 being buggy
Thu, May 18, 2:25 AM

Wed, May 17

rovka committed rL303252: Reland r303247: [ARM] GlobalISel: Remove dead instruction selection code.
Reland r303247: [ARM] GlobalISel: Remove dead instruction selection code
Wed, May 17, 5:56 AM
rovka committed rL303249: Revert "[ARM] GlobalISel: Remove dead instruction selection code".
Revert "[ARM] GlobalISel: Remove dead instruction selection code"
Wed, May 17, 5:09 AM
rovka committed rL303247: [ARM] GlobalISel: Remove dead instruction selection code.
[ARM] GlobalISel: Remove dead instruction selection code
Wed, May 17, 4:52 AM
rovka committed rL303243: Fixup r303240: Use llvm::to_string instead of std::to_string.
Fixup r303240: Use llvm::to_string instead of std::to_string
Wed, May 17, 2:38 AM
rovka committed rL303240: [GlobalISel][TableGen] Fix handling of default operands.
[GlobalISel][TableGen] Fix handling of default operands
Wed, May 17, 2:11 AM
rovka closed D33031: [GlobalISel][TableGen] Fix handling of default operands by committing rL303240: [GlobalISel][TableGen] Fix handling of default operands.
Wed, May 17, 2:10 AM

Fri, May 12

rovka committed rL302900: Add -lm to plot2fig.
Add -lm to plot2fig
Fri, May 12, 7:42 AM

Thu, May 11

rovka updated the diff for D33031: [GlobalISel][TableGen] Fix handling of default operands.

Removed the first loop and extracted a helper for adding renderers for default ops. I think this makes the body of the loop easier to read.

Thu, May 11, 5:53 AM
rovka committed rL302782: [ARM][GlobalISel] Legalize narrow scalar ops by widening.
[ARM][GlobalISel] Legalize narrow scalar ops by widening
Thu, May 11, 2:59 AM
rovka committed rL302778: [ARM][GlobalISel] Support for G_ANYEXT.
[ARM][GlobalISel] Support for G_ANYEXT
Thu, May 11, 1:41 AM
rovka committed rL302776: Revert "PR22877: When constructing an array via a constructor with a default….
Revert "PR22877: When constructing an array via a constructor with a default…
Thu, May 11, 1:23 AM

Wed, May 10

rovka added inline comments to D33031: [GlobalISel][TableGen] Fix handling of default operands.
Wed, May 10, 3:29 PM
rovka added inline comments to D33031: [GlobalISel][TableGen] Fix handling of default operands.
Wed, May 10, 12:19 AM
rovka created D33031: [GlobalISel][TableGen] Fix handling of default operands.
Wed, May 10, 12:11 AM

Tue, May 9

rovka committed rL302524: Revert "[Dwarf] Disable reference verification for now (PR32972)".
Revert "[Dwarf] Disable reference verification for now (PR32972)"
Tue, May 9, 6:19 AM
rovka committed rL302511: [ARM GlobalISel] Remove hand-written G_FADD selection.
[ARM GlobalISel] Remove hand-written G_FADD selection
Tue, May 9, 1:46 AM

Tue, May 2

rovka committed rL301905: [ARM] GlobalISel: Use TableGen instruction selector.
[ARM] GlobalISel: Use TableGen instruction selector
Tue, May 2, 2:53 AM

Fri, Apr 28

rovka added a comment to D32491: [globalisel][tablegen] Compute available feature bits correctly..

I just checked and the existing ARM tests fail without this patch because the available features aren't reset when entering the functions. They work with this patch + enabling the TableGen selector for ARM, so we'll have tests then :)

Fri, Apr 28, 8:24 AM
rovka added a comment to D32491: [globalisel][tablegen] Compute available feature bits correctly..

I think this patch is already doing a lot of things, I'd prefer committing the support for immediates separately and opening a PR for the test (and you can attach the X86 test to it so we can discuss with Igor). I need to double check this, but I think if I enable this for ARM (which I intend to do after this goes in) the existing ARM tests will already exercise this (for S_DIV).

Fri, Apr 28, 7:55 AM
rovka committed rL301633: [ARM] GlobalISel: fixup r301632.
[ARM] GlobalISel: fixup r301632
Fri, Apr 28, 2:33 AM
rovka committed rL301632: [ARM] GlobalISel: Get rid of ARMInstructionSelector.h. NFC..
[ARM] GlobalISel: Get rid of ARMInstructionSelector.h. NFC.
Fri, Apr 28, 2:23 AM
rovka committed rL301627: [ARM] GlobalISel: Tighten test. NFC.
[ARM] GlobalISel: Tighten test. NFC
Fri, Apr 28, 1:03 AM

Apr 27 2017

rovka added a comment to D32491: [globalisel][tablegen] Compute available feature bits correctly..

I'm having trouble improving the test case since the function-level predicates don't show up in any importable rules yet. I'll see if I can uncover one

Apr 27 2017, 8:05 AM
rovka committed rL301531: [ARM] GlobalISel: Fix extended stack operands.
[ARM] GlobalISel: Fix extended stack operands
Apr 27 2017, 3:36 AM

Apr 26 2017

rovka added a comment to D32529: [GlobalISel] Make multi-step legalization work..

Thanks for splitting this out! The ARM part looks excellent :)
I think you could add some target-independent tests for this in unittests/CodeGen/GlobalISel/LegalizerInfoTest.cpp, to make sure it works to WidenScalar/MoreElements etc to a type that has a Libcall or Custom action. That way we won't have to depend on ARM for testing this functionality.

Apr 26 2017, 8:44 AM
rovka accepted D32491: [globalisel][tablegen] Compute available feature bits correctly..
Apr 26 2017, 7:58 AM
rovka added a comment to D32491: [globalisel][tablegen] Compute available feature bits correctly..

Thanks for all the explanations!
LGTM with that extra test, we can keep discussing the remaining points but I think it's ok to iterate in-tree. I'd really, really like to see the default change from 0 to 1 on RecomputePerFunction, it seems safer that way.

Apr 26 2017, 7:58 AM
rovka added a comment to D32491: [globalisel][tablegen] Compute available feature bits correctly..

So, we have a new GISelAccessor with a new InstructionSelector for each unique subtarget (module + function attributes, as hashed in getSubtargetImpl). Why do we need to have module features and function features in each InstructionSelector? Isn't there a different InstructionSelector for each kind of function, and don't we magically get the right one when processing each function? What am I missing? It seems to me that we could get away with a single set of availableFeatures per InstructionSelector.

Apr 26 2017, 2:40 AM

Apr 25 2017

rovka accepted D32438: [LNT] Document how to capture linux perf profiles..

I tried it and I can confirm that it works. Thanks!

Apr 25 2017, 9:58 AM
rovka added a comment to D32275: [globalisel][tablegen] Add several GINodeEquiv's for operators that do not require additional support..

I decided to have a look at code coverage differences before committing this. This patch made no difference to the code coverage. On closer inspection it turns out beginFunction() isn't being executed so no rules with rule predicates can match. I'll take a look at this, maybe something from the rule-predicates patch didn't make it into the final commit.

Apr 25 2017, 3:46 AM

Apr 24 2017

rovka committed rL301166: [ARM] GlobalISel: Legalize s8 and s16 G_(S|U)DIV.
[ARM] GlobalISel: Legalize s8 and s16 G_(S|U)DIV
Apr 24 2017, 2:25 AM
rovka committed rL301164: [ARM] GlobalISel: Support G_(S|U)DIV for s32.
[ARM] GlobalISel: Support G_(S|U)DIV for s32
Apr 24 2017, 1:33 AM
rovka committed rL301163: [GlobalISel] Legalize G_(S|U)DIV libcalls.
[GlobalISel] Legalize G_(S|U)DIV libcalls
Apr 24 2017, 12:35 AM

Apr 23 2017

rovka committed rL301162: [ARM] GlobalISel: Select G_CONSTANT with CImm operands.
[ARM] GlobalISel: Select G_CONSTANT with CImm operands
Apr 23 2017, 11:43 PM

Apr 21 2017

rovka committed rL300971: [ARM] GlobalISel: Add support for G_TRUNC.
[ARM] GlobalISel: Add support for G_TRUNC
Apr 21 2017, 6:29 AM
rovka committed rL300967: [ARM] GlobalISel: Make struct arguments fail elegantly.
[ARM] GlobalISel: Make struct arguments fail elegantly
Apr 21 2017, 5:06 AM
rovka committed rL300955: Revert r300889, r300906, r300935, r300939.
Revert r300889, r300906, r300935, r300939
Apr 21 2017, 1:34 AM

Apr 20 2017

rovka accepted D31709: [NFC] Refactor DiagnosticRenderer to use FullSourceLoc.

I don't see anything wrong with this, I think you can commit it in a couple of days if nobody comes up with a reason why the DiagnosticRenderer shouldn't use FullSourceLoc.

Apr 20 2017, 5:20 AM
rovka accepted D31418: [globalisel][tablegen] Import SelectionDAG's rule predicates and support the equivalent in GIRule..

LGTM with nits.

Apr 20 2017, 5:10 AM
rovka committed rL300827: [ARM] Rename HW div feature to HW div Thumb. NFCI..
[ARM] Rename HW div feature to HW div Thumb. NFCI.
Apr 20 2017, 2:51 AM
rovka closed D32160: [ARM] Rename HW div feature to HW div Thumb. NFCI. by committing rL300827: [ARM] Rename HW div feature to HW div Thumb. NFCI..
Apr 20 2017, 2:51 AM

Apr 19 2017

rovka committed rL300665: [ARM] GlobalISel: Add support for G_MUL.
[ARM] GlobalISel: Add support for G_MUL
Apr 19 2017, 12:42 AM

Apr 18 2017

rovka committed rL300546: [ARM] GlobalISel: Add support for G_SUB.
[ARM] GlobalISel: Add support for G_SUB
Apr 18 2017, 5:48 AM
rovka created D32160: [ARM] Rename HW div feature to HW div Thumb. NFCI..
Apr 18 2017, 3:30 AM
rovka committed rL300536: [ARM] Check for correct HW div when lowering divmod.
[ARM] Check for correct HW div when lowering divmod
Apr 18 2017, 1:45 AM
rovka closed D32005: [ARM] Check for correct HW div when lowering divmod by committing rL300536: [ARM] Check for correct HW div when lowering divmod.
Apr 18 2017, 1:45 AM

Apr 13 2017

rovka added a comment to D32005: [ARM] Check for correct HW div when lowering divmod.

Thanks for the review

Apr 13 2017, 8:59 AM
rovka added reviewers for D32005: [ARM] Check for correct HW div when lowering divmod: jmolloy, olista01, samparker, compnerd.
Apr 13 2017, 4:19 AM
rovka updated the summary of D32005: [ARM] Check for correct HW div when lowering divmod.
Apr 13 2017, 4:17 AM
rovka created D32005: [ARM] Check for correct HW div when lowering divmod.
Apr 13 2017, 4:14 AM
rovka committed rL300180: Revert "Cache size per class size in SizeClassAllocatorXLocalCache.".
Revert "Cache size per class size in SizeClassAllocatorXLocalCache."
Apr 13 2017, 12:51 AM

Apr 11 2017

rovka committed rL299931: GlobalISel: Allow legalizing G_FADD to a libcall.
GlobalISel: Allow legalizing G_FADD to a libcall
Apr 11 2017, 4:05 AM
rovka accepted D31135: [globalisel][tablegen] Add experimental support for OperandWithDefaultOps, PredicateOperand, and OptionalDefOperand.

LGTM, thanks.

Apr 11 2017, 3:57 AM
rovka committed rL299928: Revert "Turn some C-style vararg into variadic templates".
Revert "Turn some C-style vararg into variadic templates"
Apr 11 2017, 3:20 AM

Apr 10 2017

rovka committed rL299841: [ARM] GlobalISel: Support G_FPOW for float and double.
[ARM] GlobalISel: Support G_FPOW for float and double
Apr 10 2017, 2:40 AM

Apr 7 2017

rovka committed rL299761: [ARM] GlobalISel: Test hard float properly.
[ARM] GlobalISel: Test hard float properly
Apr 7 2017, 5:17 AM
rovka committed rL299756: [ARM] GlobalISel: Support frem for 64-bit values.
[ARM] GlobalISel: Support frem for 64-bit values
Apr 7 2017, 4:02 AM
rovka committed rL299753: [ARM] GlobalISel: Support frem for 32-bit values.
[ARM] GlobalISel: Support frem for 32-bit values
Apr 7 2017, 2:54 AM

Apr 3 2017

rovka added inline comments to D31135: [globalisel][tablegen] Add experimental support for OperandWithDefaultOps, PredicateOperand, and OptionalDefOperand.
Apr 3 2017, 9:28 AM
rovka accepted D31054: [globalisel][tablegen] Fix patterns involving multiple ComplexPatterns..

LGTM.

Apr 3 2017, 7:14 AM

Mar 31 2017

rovka added inline comments to D30539: [tablegen][globalisel] Add support for nested instruction matching..
Mar 31 2017, 8:16 AM

Mar 30 2017

rovka added a comment to D30299: [tablegen] Optionally format tablegen targets with clang-format.

Hi Daniel,

That's interesting.
I am a bit concerned about making GISel special in terms of how the output looks. I can see the argument but I am also wondering if we don't what to have a control on that. For instance, in SDISel, each entry in the array is on its own line and clang-format could break that.

Is that the case for many of the files that we generate? If not, maybe we can make it opt-out instead?
Alternatively, clang-format accepts some configuration options, is there any combination of those that would make the output of SDISel look ok when clang-formatted?

Mar 30 2017, 4:41 AM

Mar 23 2017

rovka accepted D30973: [GlobalISel][X86] Support G_STORE/G_LOAD operation.

Could you commit the unrelated changes separately (e.g. the space changes and the include reordering)?
Other than that, LGTM.

Mar 23 2017, 3:35 AM

Mar 20 2017

rovka committed rL298254: [GlobalISel] Use the correct calling conv for calls.
[GlobalISel] Use the correct calling conv for calls
Mar 20 2017, 7:52 AM
rovka closed D31039: [GlobalISel] Use the correct calling conv for calls by committing rL298254: [GlobalISel] Use the correct calling conv for calls.
Mar 20 2017, 7:52 AM

Mar 17 2017

rovka updated the diff for D31039: [GlobalISel] Use the correct calling conv for calls.

Thanks for reviewing!

Mar 17 2017, 3:06 AM

Mar 16 2017

rovka created D31039: [GlobalISel] Use the correct calling conv for calls.
Mar 16 2017, 10:16 AM
rovka accepted D30535: [tablegen][globalisel] Capture instructions into locals and related infrastructure for multiple instructions matches..

LGTM.

Mar 16 2017, 9:14 AM

Mar 14 2017

rovka added a comment to D30535: [tablegen][globalisel] Capture instructions into locals and related infrastructure for multiple instructions matches..

I think this looks ok overall, just a few nits.

Mar 14 2017, 6:20 AM

Mar 13 2017

rovka accepted D30531: [tablegen][globalisel] Trivial changes to reduce size of following patch. NFC..

LGTM.

Mar 13 2017, 10:51 AM
rovka committed rL297621: [ARM] GlobalISel: Support SP in regbankselect.
[ARM] GlobalISel: Support SP in regbankselect
Mar 13 2017, 7:40 AM

Mar 1 2017

rovka committed rL296634: clang-format r296631.
clang-format r296631
Mar 1 2017, 8:06 AM
rovka committed rL296631: [ARM] GlobalISel: Lower call params that need extensions.
[ARM] GlobalISel: Lower call params that need extensions
Mar 1 2017, 7:47 AM

Feb 28 2017

rovka committed rL296473: [ARM] GlobalISel: Lower i32 and fp call parameters on the stack.
[ARM] GlobalISel: Lower i32 and fp call parameters on the stack
Feb 28 2017, 6:29 AM
rovka committed rL296471: [ARM] GlobalISel: Select 32-bit G_CONSTANT.
[ARM] GlobalISel: Select 32-bit G_CONSTANT
Feb 28 2017, 5:17 AM
rovka committed rL296469: [ARM] GlobalISel: Add mapping for G_CONSTANT.
[ARM] GlobalISel: Add mapping for G_CONSTANT
Feb 28 2017, 4:25 AM
rovka committed rL296468: [ARM] GlobalISel: Legalize 32-bit constants.
[ARM] GlobalISel: Legalize 32-bit constants
Feb 28 2017, 3:45 AM
rovka committed rL296462: [ARM] GlobalISel: Select G_GEP.
[ARM] GlobalISel: Select G_GEP
Feb 28 2017, 2:26 AM
rovka committed rL296455: [ARM] GlobalISel: Add reg bank mapping for G_GEP.
[ARM] GlobalISel: Add reg bank mapping for G_GEP
Feb 28 2017, 1:47 AM
rovka committed rL296452: [ARM] GlobalISel: Legalize G_GEP with 32-bit offsets.
[ARM] GlobalISel: Legalize G_GEP with 32-bit offsets
Feb 28 2017, 1:14 AM

Feb 24 2017

rovka committed rL296122: [ARM] GlobalISel: Select G_STORE.
[ARM] GlobalISel: Select G_STORE
Feb 24 2017, 6:13 AM
rovka committed rL296120: Minor test fix.
Minor test fix
Feb 24 2017, 5:39 AM
rovka committed rL296115: [ARM] GlobalISel: Add reg bank mappings for stores.
[ARM] GlobalISel: Add reg bank mappings for stores
Feb 24 2017, 5:19 AM
rovka committed rL296111: Fixup r296105 - only run tests on Mips.
Fixup r296105 - only run tests on Mips
Feb 24 2017, 4:59 AM
rovka committed rL296108: [ARM] GlobalISel: Legalize stores.
[ARM] GlobalISel: Legalize stores
Feb 24 2017, 3:40 AM
rovka committed rL296104: Revert "[ARM] GlobalISel: Legalize stores".
Revert "[ARM] GlobalISel: Legalize stores"
Feb 24 2017, 2:47 AM
rovka committed rL296103: [ARM] GlobalISel: Legalize stores.
[ARM] GlobalISel: Legalize stores
Feb 24 2017, 2:31 AM

Feb 23 2017

rovka accepted D29712: [globalisel] Decouple src pattern operands from dst pattern operands..

LGTM.

Feb 23 2017, 8:29 AM
rovka accepted D29711: [globalisel] Sort RuleMatchers by priority..

I don't see anything wrong with it, I think it's ok to iterate on the ordering and other tricky parts in tree.

Feb 23 2017, 8:13 AM
rovka committed rL295973: [ARM] GlobalISel: Lower call returns.
[ARM] GlobalISel: Lower call returns
Feb 23 2017, 6:30 AM
rovka committed rL295971: [ARM] GlobalISel: Lower call parameters in regs.
[ARM] GlobalISel: Lower call parameters in regs
Feb 23 2017, 5:37 AM

Feb 22 2017

rovka added a comment to D30047: [GlobalISel] Split LowLevelType to a fundamental type in libSupport..

Ok, let's go with the standalone method. I'd like to give it a more descriptive name though seeing as it's in the llvm namespace. How about getLLTForType()?

For the header name, I agree with not having LowLevelTypes.h in both libraries. Let's go with GlobalISel/Utils.h.

Feb 22 2017, 1:05 AM

Feb 21 2017

rovka added a comment to D30047: [GlobalISel] Split LowLevelType to a fundamental type in libSupport..

I've been trying to preserve the current style of LLT usage but having seen this I realize there's no need to do that.

I think we should make llvm::getLLT() a member of Type (see example usage below and https://reviews.llvm.org/differential/diff/88897/) but aside from that and a couple nits I think this seems like a reasonable way to fix the layering problem.

Feb 21 2017, 6:43 AM
rovka accepted D29953: [ARM] Classification Improvements to ARM Sched-Models. NFCI..

Thanks, Javed. LGTM.

Feb 21 2017, 4:23 AM
rovka committed rL295716: [ARM] GlobalISel: Lower calls to void() functions.
[ARM] GlobalISel: Lower calls to void() functions
Feb 21 2017, 3:45 AM

Feb 20 2017

rovka committed rL295667: Revert "[tsan] Provide external tags (object types) via debugging API".
Revert "[tsan] Provide external tags (object types) via debugging API"
Feb 20 2017, 7:38 AM
rovka committed rL295662: [ARM] GlobalISel: Don't select atomic loads.
[ARM] GlobalISel: Don't select atomic loads
Feb 20 2017, 6:57 AM

Feb 17 2017

rovka committed rL295446: [ARM] GlobalISel: Clean up some helpers.
[ARM] GlobalISel: Clean up some helpers
Feb 17 2017, 5:55 AM