Page MenuHomePhabricator
Feed Advanced Search

Today

craig.topper added a reviewer for D71532: [InstCombine] Teach removeBitcastsFromLoadStoreOnMinMax not to change the size of a store.: ABataev.
Sun, Dec 15, 4:15 PM · Restricted Project
craig.topper created D71532: [InstCombine] Teach removeBitcastsFromLoadStoreOnMinMax not to change the size of a store..
Sun, Dec 15, 4:15 PM · Restricted Project

Yesterday

craig.topper committed rG1dc0c8af5e7d: [LegalizeTypes] Teach BitcastToInt_ATOMIC_SWAP to only create FP16_TO_FP when… (authored by craig.topper).
[LegalizeTypes] Teach BitcastToInt_ATOMIC_SWAP to only create FP16_TO_FP when…
Sat, Dec 14, 3:12 PM
craig.topper committed rG95ce8f94986c: [LegalizeTypes] In PromoteFloatOp_SETCC, don't both querying for transforming… (authored by craig.topper).
[LegalizeTypes] In PromoteFloatOp_SETCC, don't both querying for transforming…
Sat, Dec 14, 3:12 PM
craig.topper added inline comments to D71442: [X86] Add calculation for elements in structures in getting uniform base for the Gather/Scatter intrinsic..
Sat, Dec 14, 1:40 PM · Restricted Project

Fri, Dec 13

craig.topper added a comment to D62731: Add support for options -frounding-math, -ftrapping-math, -ffp-model=, and -ffp-exception-behavior=, : Specify floating point behavior.

It's good that people are looking at achieving better modeling for the x86 backend, but we need to have a plan that doesn't require heroic effort just to get basic correctness.

Do you mean in the backend? If so, I don't think that's possible. The backends just don't have any sort of feature that could be used to get conservatively correct behavior for cheap the way intrinsics give it to us in the middle end. Once you go into instruction selection things get very low level in a hurry.

I'm looking for simple ways to modeling X86 intrinsics, but haven't find better one than modeling it one by one.

I would suggest that we need a function/call attribute roughly on the level of readonly / readnone, maybe readfponly, that says that a function has no side-effects and no dependencies on anything *except* the FP state.

Do you mean mark it at the declaration of intrinsics? Is it reasonable to mark except on dependent intrinsics?

Basic queries like Instruction::mayReadMemory() that are supposed to be used generically in code-motion transforms would then return true for calls marked that way only if they're FP-constrained functions.

Middle end or back end? I think in middle end you may need to change all releated passes to get such information to prevent optimization. And in back end, I think we can simply chain intrinsics marked except with other FP nodes like what common code doing.

Fri, Dec 13, 6:37 PM · Restricted Project, Restricted Project
craig.topper updated subscribers of D71467: [FPEnv] Generate constrained FP comparisons from clang.

The bug with __builtin_isless should be a really easy fix; the builtin just needs to be flagged as having custom type-checking, and then we need to make sure we do appropriate promotions on the arguments (but we probably do).

Fri, Dec 13, 3:09 PM · Restricted Project, Restricted Project
craig.topper accepted D71205: [NFC] Add a SDValue overload for SelectionDAG::getMemBasePlusOffset().

LGTM

Fri, Dec 13, 9:38 AM · Restricted Project
craig.topper committed rG5c80a4f454ac: [LegalizeTypes] Remove unnecessary if before calling ReplaceValueWith on the… (authored by craig.topper).
[LegalizeTypes] Remove unnecessary if before calling ReplaceValueWith on the…
Fri, Dec 13, 12:16 AM
craig.topper added inline comments to D71442: [X86] Add calculation for elements in structures in getting uniform base for the Gather/Scatter intrinsic..
Fri, Dec 13, 12:05 AM · Restricted Project

Thu, Dec 12

craig.topper added inline comments to D71442: [X86] Add calculation for elements in structures in getting uniform base for the Gather/Scatter intrinsic..
Thu, Dec 12, 10:19 PM · Restricted Project
craig.topper added inline comments to D71442: [X86] Add calculation for elements in structures in getting uniform base for the Gather/Scatter intrinsic..
Thu, Dec 12, 10:19 PM · Restricted Project
craig.topper added inline comments to D71442: [X86] Add calculation for elements in structures in getting uniform base for the Gather/Scatter intrinsic..
Thu, Dec 12, 7:37 PM · Restricted Project

Wed, Dec 11

craig.topper added a comment to D71373: [ConstantFolding] Disable constant folding for bitcasts to x86_fp80.

I was looking into trying to see if we could store this in APFloat. And I just noticed that we mark pseudonan, pseudoinfinity, and unnformals all as fcnan, and copy their original significand from the apint. I believe since the integer bit was 0, this makes apfloat encode them as a pseudonan. Then I think arithmetic operations with one of these as an operand will propagate the pseudonan. Should we have forced the integer bit of the significand when we created it to make it a true nan?

Wed, Dec 11, 4:37 PM · Restricted Project
craig.topper added inline comments to D69275: Add constrained int->FP intrinsics.
Wed, Dec 11, 3:42 PM · Restricted Project
craig.topper committed rG4b452952fed7: [LegalizeTypes] In SoftenFloatRes_FP_EXTEND, move the check for input already… (authored by craig.topper).
[LegalizeTypes] In SoftenFloatRes_FP_EXTEND, move the check for input already…
Wed, Dec 11, 12:56 PM
craig.topper created D71373: [ConstantFolding] Disable constant folding for bitcasts to x86_fp80.
Wed, Dec 11, 12:09 PM · Restricted Project
craig.topper added inline comments to D71360: Honor -fuse-init-array when os is not specified on x86.
Wed, Dec 11, 11:48 AM · Restricted Project
craig.topper added a comment to D71360: Honor -fuse-init-array when os is not specified on x86.

Is it possible to add a test for this?

Wed, Dec 11, 8:52 AM · Restricted Project
craig.topper added a comment to D71145: [InstCombine] Allow to limit the max number of iterations.

In my experience extra iterations are often caused by bad worklist management when updating nodes sometimes. For example https://reviews.llvm.org/D50990 was supposed to fix one case of that, but I never got around to finding the original test where I observed the issue. If that patch still applies does it have any effect on any of the extra iterations in the table?

Wed, Dec 11, 8:03 AM · Restricted Project
craig.topper committed rG3adc819b7a07: [X86] Erase dead LEA instruction after converting it to MOV in FixupLEAPass… (authored by craig.topper).
[X86] Erase dead LEA instruction after converting it to MOV in FixupLEAPass…
Wed, Dec 11, 7:53 AM

Tue, Dec 10

craig.topper committed rGd4345636e678: [LegalizeTypes] Remove manual worklist management from SoftenFloatRes_FP_EXTEND. (authored by craig.topper).
[LegalizeTypes] Remove manual worklist management from SoftenFloatRes_FP_EXTEND.
Tue, Dec 10, 10:34 PM
craig.topper added a comment to D71312: [InstCombine] X / (select C, X, -X) -> select C ? 1 : -1.

Seems like we should maybe teach InstCombiner::FoldOpIntoSelect to handle cases where one of the select operands becomes a constant if we fold. But we might only call FoldOpIntoSelect when one of the operands of the binop is a constant today.

Tue, Dec 10, 6:00 PM · Restricted Project
craig.topper added inline comments to D71312: [InstCombine] X / (select C, X, -X) -> select C ? 1 : -1.
Tue, Dec 10, 5:42 PM · Restricted Project
craig.topper committed rG935d41e4bd63: [X86] Split v64i1 arguments into 2 v32i1s that will be promoted to v32i8 under… (authored by craig.topper).
[X86] Split v64i1 arguments into 2 v32i1s that will be promoted to v32i8 under…
Tue, Dec 10, 5:33 PM
craig.topper added a comment to D71312: [InstCombine] X / (select C, X, -X) -> select C ? 1 : -1.

Seems to be what gcc really implements https://godbolt.org/z/ZBooz9

Tue, Dec 10, 5:05 PM · Restricted Project
craig.topper added a comment to D71312: [InstCombine] X / (select C, X, -X) -> select C ? 1 : -1.

Isn't it sufficient to just check that the true/false values of the select are x and -x. The condition itself doesn't matter. x / (select c, x, -x) -> select c ? 1 : -1

Tue, Dec 10, 4:46 PM · Restricted Project
craig.topper added a comment to D71312: [InstCombine] X / (select C, X, -X) -> select C ? 1 : -1.

Isn't abs(INT_MIN) undefined?

expensive div is removed:

_Z4fooai: # @_Z4fooai
  mov eax, edi
  mov ecx, edi
  neg ecx
  cmovl ecx, edi
  cdq
  idiv ecx
  ret
_Z5fooari: # @_Z5fooari
  mov eax, edi
  sar eax, 31
  or eax, 1
  ret

GCC knows this trick too.

Tue, Dec 10, 4:36 PM · Restricted Project
craig.topper committed rG88dacbd43625: [X86] Go back to considering v64i1 as a legal type under min-legal-vector… (authored by craig.topper).
[X86] Go back to considering v64i1 as a legal type under min-legal-vector…
Tue, Dec 10, 3:10 PM
craig.topper added a reverting change for rG3e1aee2ba717: [X86] Don't consider v64i1 as a legal type unless v64i8 is also a legal type.: rG88dacbd43625: [X86] Go back to considering v64i1 as a legal type under min-legal-vector….
Tue, Dec 10, 3:10 PM
craig.topper accepted D70007: [Intrinsic] Add fixed point division intrinsics..

I don't really like the code In SelectionDAGBuilder, but I don't have a better solution either. So LGTM.

Tue, Dec 10, 2:09 PM · Restricted Project
craig.topper accepted D70582: [FPEnv][X86] Constrained FCmp intrinsics enabling on X86.

signaling->signalling

Not to be too nitpicky, but I've been using the US spelling (with one 'l') in the common code parts of the patch :-) Should we be consistent here?

I found IEEE754 uses signaling, I think it's better to follow it as well as common code.

Tue, Dec 10, 1:15 PM · Restricted Project
craig.topper added inline comments to D69275: Add constrained int->FP intrinsics.
Tue, Dec 10, 12:56 PM · Restricted Project
craig.topper updated the diff for D71130: [X86] Supplement to D69275 including promotion support.

Rebase on latest D69275

Tue, Dec 10, 12:46 PM · Restricted Project
craig.topper added inline comments to D69275: Add constrained int->FP intrinsics.
Tue, Dec 10, 11:51 AM · Restricted Project
craig.topper added inline comments to D69275: Add constrained int->FP intrinsics.
Tue, Dec 10, 11:32 AM · Restricted Project
craig.topper added inline comments to D69275: Add constrained int->FP intrinsics.
Tue, Dec 10, 10:46 AM · Restricted Project
craig.topper added inline comments to D70582: [FPEnv][X86] Constrained FCmp intrinsics enabling on X86.
Tue, Dec 10, 12:43 AM · Restricted Project
craig.topper added inline comments to D70582: [FPEnv][X86] Constrained FCmp intrinsics enabling on X86.
Tue, Dec 10, 12:25 AM · Restricted Project

Mon, Dec 9

craig.topper added inline comments to D70582: [FPEnv][X86] Constrained FCmp intrinsics enabling on X86.
Mon, Dec 9, 11:39 PM · Restricted Project
craig.topper added inline comments to D70582: [FPEnv][X86] Constrained FCmp intrinsics enabling on X86.
Mon, Dec 9, 3:34 PM · Restricted Project
craig.topper added inline comments to D71205: [NFC] Add a SDValue overload for SelectionDAG::getMemBasePlusOffset().
Mon, Dec 9, 2:10 PM · Restricted Project
craig.topper committed rG505aa2410db6: [Attr] Move ParsedTargetAttr out of the TargetAttr class (authored by craig.topper).
[Attr] Move ParsedTargetAttr out of the TargetAttr class
Mon, Dec 9, 12:44 PM
craig.topper closed D71159: [Attr] Move ParsedTargetAttr out of the TargetAttr class.
Mon, Dec 9, 12:44 PM · Restricted Project
craig.topper added a comment to D71218: [FPEnv] Remove unnecessary rounding mode argument for constrained intrinsics.

I've also been wondering if we should drop rounding from ceil, floor, trunc, and round. Looks like lround/llround don't have rounding.

Mon, Dec 9, 11:29 AM · Restricted Project
craig.topper accepted D71184: [X86] Add supports for strict float operation fpextend/fpround/fsqrt.

Forgot click Accept I guess. Please fix those two comments before committing

Mon, Dec 9, 10:43 AM · Restricted Project
craig.topper added a comment to D71184: [X86] Add supports for strict float operation fpextend/fpround/fsqrt.

LGTM with those two comments addressed.

Mon, Dec 9, 10:16 AM · Restricted Project

Sun, Dec 8

craig.topper added a comment to D71178: [x86] Don't assume sign-extension of arguments smaller than 32-bits..

Shouldn’t we instead stop the sext and zext attributes from being added by the frontend?

Sun, Dec 8, 5:33 PM · Restricted Project

Fri, Dec 6

craig.topper added inline comments to D71159: [Attr] Move ParsedTargetAttr out of the TargetAttr class.
Fri, Dec 6, 6:32 PM · Restricted Project
craig.topper created D71159: [Attr] Move ParsedTargetAttr out of the TargetAttr class.
Fri, Dec 6, 5:21 PM · Restricted Project
craig.topper reopened D68627: [Sema][X86] Consider target attribute into the checks in validateOutputSize and validateInputSize..
Fri, Dec 6, 5:20 PM · Restricted Project
craig.topper committed rGe1578fd2b79f: [Sema][X86] Consider target attribute into the checks in validateOutputSize and… (authored by craig.topper).
[Sema][X86] Consider target attribute into the checks in validateOutputSize and…
Fri, Dec 6, 3:31 PM
craig.topper closed D68627: [Sema][X86] Consider target attribute into the checks in validateOutputSize and validateInputSize..
Fri, Dec 6, 3:31 PM · Restricted Project
craig.topper committed rG28b573d2497e: [TargetLowering] Fix another potential FPE in expandFP_TO_UINT (authored by craig.topper).
[TargetLowering] Fix another potential FPE in expandFP_TO_UINT
Fri, Dec 6, 2:16 PM
craig.topper closed D67105: [TargetLowering] Fix another potential FPE in expandFP_TO_UINT.
Fri, Dec 6, 2:16 PM · Restricted Project
craig.topper added a comment to D71046: Support Intel "l" suffixes for x86_64 R8-R15 registers..

I contacted our documentation people yesterday to point out this difference between Intel and AMD documentation. They have agreed to fix this in the next release of the SDM.

Fri, Dec 6, 12:43 PM · Restricted Project
craig.topper updated the diff for D67105: [TargetLowering] Fix another potential FPE in expandFP_TO_UINT.

Fix bug in X86 code

Fri, Dec 6, 11:56 AM · Restricted Project
craig.topper abandoned D68255: [X86] Remove AVX/AVX512 check from validateOperandSize, just always accept 512.
Fri, Dec 6, 10:05 AM
craig.topper created D71130: [X86] Supplement to D69275 including promotion support.
Fri, Dec 6, 10:03 AM · Restricted Project
craig.topper accepted D70650: [MC] Rewrite tablegen for printInstrAlias to comiple faster, NFC.

LGTM

Fri, Dec 6, 9:35 AM · Restricted Project
craig.topper updated the diff for D69275: Add constrained int->FP intrinsics.

Restore @kpn 's last diff.

Fri, Dec 6, 9:35 AM · Restricted Project
craig.topper added a comment to D69275: Add constrained int->FP intrinsics.
In D69275#1772541, @kpn wrote:

I'm still working on this ticket daily! I'm trying to merge the two vector unrolling functions like Ulrich suggested. But I ran into problems that lead me to think we may have a serious issue lurking that we'll need to fix. That's what I've been working on: trying to understand the issue and see if it needs further investigation.

If you are in a hurry then you could have sent me an email and I would have uploaded the diffs I've got without further investigation.

I'm leaving attached the comments on my work that I've been adding but haven't submitted until now.

Fri, Dec 6, 8:48 AM · Restricted Project
craig.topper updated the diff for D69275: Add constrained int->FP intrinsics.

Change the signature of SelectionDAGLegalize::ExpandLegalINT_TO_FP to allow it to update the Results vector directly.

Fri, Dec 6, 1:13 AM · Restricted Project
craig.topper updated the diff for D69275: Add constrained int->FP intrinsics.

Upload with context

Fri, Dec 6, 12:36 AM · Restricted Project
craig.topper updated the diff for D69275: Add constrained int->FP intrinsics.

-Improve some of the X86 code.
-Add Promote support. Use it for i8/i16 on X86.
-Remove changes to UnrollVectorOp which seemed to be unexercised
-Some cleanup in ExpandLegalINT_TO_FP
-Drop the changes to getNode. We can't fold NOP conversions here and asserts were recently added in another patch.

Fri, Dec 6, 12:32 AM · Restricted Project
craig.topper commandeered D69275: Add constrained int->FP intrinsics.

Commandeering to update the X86 code and some other fixes/cleanup.

Fri, Dec 6, 12:26 AM · Restricted Project

Thu, Dec 5

craig.topper added inline comments to D70582: [FPEnv][X86] Constrained FCmp intrinsics enabling on X86.
Thu, Dec 5, 7:03 PM · Restricted Project
craig.topper committed rG8267be29955e: [X86] Make X86TargetLowering::BuildFILD return a std::pair of SDValues so we… (authored by craig.topper).
[X86] Make X86TargetLowering::BuildFILD return a std::pair of SDValues so we…
Thu, Dec 5, 6:16 PM
craig.topper added inline comments to D70650: [MC] Rewrite tablegen for printInstrAlias to comiple faster, NFC.
Thu, Dec 5, 4:44 PM · Restricted Project
craig.topper accepted D68757: [X86] Add strict fp support for instructions fadd/fsub/fmul/fdiv.

LGTM

Thu, Dec 5, 3:55 PM · Restricted Project
craig.topper accepted D70913: [FPEnv][SelectionDAG] Relax chain requirements.

LGTM

Thu, Dec 5, 3:54 PM · Restricted Project
craig.topper accepted D71023: [x86] add cost model special-case for insert/extract from element 0.

Seems reasonable. LGTM

Thu, Dec 5, 3:54 PM · Restricted Project
craig.topper accepted D69281: [FPEnv] Constrained FCmp intrinsics.

LGTM

Thu, Dec 5, 3:07 PM · Restricted Project
craig.topper added a comment to D71046: Support Intel "l" suffixes for x86_64 R8-R15 registers..

I also found this where NASM indicated they wouldn't support it https://sourceforge.net/p/nasm/bugs/324/

Thu, Dec 5, 2:17 PM · Restricted Project
craig.topper added a comment to D71046: Support Intel "l" suffixes for x86_64 R8-R15 registers..

Do you have examples of other tools that accept this? I checked the GNU assembler and it didn't accept r8l

Thu, Dec 5, 11:11 AM · Restricted Project
craig.topper committed rGf688570d5c54: [X86] Remove ProcIntelGLM/ProcIntelGLP/ProcIntelTRM and replace them with a… (authored by craig.topper).
[X86] Remove ProcIntelGLM/ProcIntelGLP/ProcIntelTRM and replace them with a…
Thu, Dec 5, 11:02 AM
craig.topper closed D71048: [X86] Remove ProcIntelGLM/ProcIntelGLP/ProcIntelTRM and replace them with a single feature flag covers the two places they were used..
Thu, Dec 5, 11:02 AM · Restricted Project
craig.topper added a comment to D69281: [FPEnv] Constrained FCmp intrinsics.

For X86, I think the one opcode or two is a wash. So I think this is fine.

Thu, Dec 5, 9:55 AM · Restricted Project

Wed, Dec 4

craig.topper created D71048: [X86] Remove ProcIntelGLM/ProcIntelGLP/ProcIntelTRM and replace them with a single feature flag covers the two places they were used..
Wed, Dec 4, 10:56 PM · Restricted Project
craig.topper added a comment to D71046: Support Intel "l" suffixes for x86_64 R8-R15 registers..

Test case?

Wed, Dec 4, 10:35 PM · Restricted Project
craig.topper committed rG3d43c73f26f2: [X86] Remove override of shouldUseStrictFP_TO_INT for fp80. NFC (authored by craig.topper).
[X86] Remove override of shouldUseStrictFP_TO_INT for fp80. NFC
Wed, Dec 4, 6:06 PM
craig.topper accepted D70999: Fix the macro fusion table for X86 according to Intel optimization manual.

LGTM

Wed, Dec 4, 6:05 PM · Restricted Project
craig.topper updated the diff for D67105: [TargetLowering] Fix another potential FPE in expandFP_TO_UINT.

Change the X86 code to the behavior as well. Regenerate the test changes which no longer applied cleanly.

Wed, Dec 4, 3:45 PM · Restricted Project
craig.topper commandeered D67105: [TargetLowering] Fix another potential FPE in expandFP_TO_UINT.

I'm going to commandeer this patch and update the equivalent code in X86. I'll post a new version later this afternoon.

Wed, Dec 4, 1:04 PM · Restricted Project
craig.topper committed rGf730ac719db3: [X86] Add missing break to the end of the last case in a switch. NFC (authored by craig.topper).
[X86] Add missing break to the end of the last case in a switch. NFC
Wed, Dec 4, 12:35 PM
craig.topper added inline comments to D70999: Fix the macro fusion table for X86 according to Intel optimization manual.
Wed, Dec 4, 7:49 AM · Restricted Project
craig.topper requested changes to D70999: Fix the macro fusion table for X86 according to Intel optimization manual.

I wanted this to be the full change from the other patch. Including moving this all to X86BaseInfo.h

Wed, Dec 4, 7:39 AM · Restricted Project
craig.topper added a comment to D70157: Align branches within 32-Byte boundary.

Can you please put the macro fusion changes in a separate phabricator review. I’ll review it in the morning US time and if it all looks good we can get that part committed while the other comments are being addressed.

Wed, Dec 4, 12:44 AM · Restricted Project, Restricted Project

Tue, Dec 3

craig.topper added inline comments to D70157: Align branches within 32-Byte boundary.
Tue, Dec 3, 10:24 PM · Restricted Project, Restricted Project
craig.topper accepted D70938: [X86] Model DAZ and FTZ.

LGTM

Tue, Dec 3, 3:14 PM · Restricted Project
craig.topper accepted D70881: [X86] Model MXCSR for all AVX512 instructions.

LGTM

Tue, Dec 3, 2:55 PM · Restricted Project
craig.topper committed rGf586fd44e4ba: [FPEnv] [PowerPC] Lowering ppc_fp128 StrictFP Nodes to libcalls (authored by craig.topper).
[FPEnv] [PowerPC] Lowering ppc_fp128 StrictFP Nodes to libcalls
Tue, Dec 3, 2:18 PM
craig.topper closed D70867: [FPEnv] [PowerPC] Alternative lowering ppc_fp128 StrictFP Nodes to libcalls.
Tue, Dec 3, 2:17 PM · Restricted Project
craig.topper committed rG5ebbabc1af36: [InstCombine] Revert aafde063aaf09285c701c80cd4b543c2beb523e8 and… (authored by craig.topper).
[InstCombine] Revert aafde063aaf09285c701c80cd4b543c2beb523e8 and…
Tue, Dec 3, 2:08 PM
craig.topper added a reverting change for rG6749dc344667: [InstCombine] Don't transform bitcasts between x86_mmx and v1i64 into…: rG5ebbabc1af36: [InstCombine] Revert aafde063aaf09285c701c80cd4b543c2beb523e8 and….
Tue, Dec 3, 2:08 PM
craig.topper added a reverting change for rGaafde063aaf0: [InstCombine] Turn (extractelement <1 x i64/double> (bitcast (x86_mmx))) into a…: rG5ebbabc1af36: [InstCombine] Revert aafde063aaf09285c701c80cd4b543c2beb523e8 and….
Tue, Dec 3, 2:08 PM
craig.topper updated the diff for D70867: [FPEnv] [PowerPC] Alternative lowering ppc_fp128 StrictFP Nodes to libcalls.

-Bypass the node in STRICT_FP_EXTEND if the expanded VT matches in the put VT.
-Add a bunch of asserts to getNode to ensure we don't create bad STRICT_FP_EXTEND/STRICT_FP_ROUND.

Tue, Dec 3, 11:38 AM · Restricted Project
craig.topper added inline comments to D70867: [FPEnv] [PowerPC] Alternative lowering ppc_fp128 StrictFP Nodes to libcalls.
Tue, Dec 3, 11:10 AM · Restricted Project

Mon, Dec 2

craig.topper committed rG039664db87d2: [LegalizeDAG] Return true from ExpandNode for some nodes that don't have expand… (authored by craig.topper).
[LegalizeDAG] Return true from ExpandNode for some nodes that don't have expand…
Mon, Dec 2, 11:48 PM
craig.topper committed rGf92000187e14: [LegalizeDAG] When expanding vector SRA/SRL/SHL add the new BUILD_VECTOR to the… (authored by craig.topper).
[LegalizeDAG] When expanding vector SRA/SRL/SHL add the new BUILD_VECTOR to the…
Mon, Dec 2, 11:11 PM