coby (coby)
User

Projects

User does not belong to any projects.

User Details

User Since
Sep 8 2016, 7:41 AM (49 w, 2 d)

Recent Activity

Yesterday

coby added a comment to D36793: [X86AsmParser] Refactoring, (almost) NFC..

IntelExpr::isValid looks rather strange because it validates possible multiplication only. But if we have Scale == 1 and don't have either BaseReg or Disp?

Sat, Aug 19, 3:08 PM

Fri, Aug 18

coby updated the diff for D36845: [X86] Allow xacquire/xrelease prefixes.

Why did you move data32 to the top instead of leaving next to data16?

no particular reason. back in place

Fri, Aug 18, 1:26 PM
coby added a reviewer for D36793: [X86AsmParser] Refactoring, (almost) NFC.: mcrosier.
Fri, Aug 18, 12:47 PM
coby added a reviewer for D36793: [X86AsmParser] Refactoring, (almost) NFC.: m_zuckerman.
Fri, Aug 18, 1:38 AM
coby added a comment to D35621: X86 Asm can't work properly with symbolic Scale.

Coby, Reid, is eveything OK? Could you get LGTM?

Fri, Aug 18, 1:37 AM

Thu, Aug 17

coby created D36845: [X86] Allow xacquire/xrelease prefixes.
Thu, Aug 17, 12:51 PM
coby added a reviewer for D36793: [X86AsmParser] Refactoring, (almost) NFC.: RKSimon.
Thu, Aug 17, 5:45 AM

Wed, Aug 16

coby updated the summary of D36793: [X86AsmParser] Refactoring, (almost) NFC..
Wed, Aug 16, 7:15 AM
coby created D36794: Fixups to FE tests affected by D36793.
Wed, Aug 16, 7:15 AM
coby created D36793: [X86AsmParser] Refactoring, (almost) NFC..
Wed, Aug 16, 7:10 AM

Mon, Aug 14

coby added inline comments to D35890: [Diagnostics][Inline-Asm]Allow targets to provide elaborations regarding failures resulted from (GNU-inline-asm) modifiers' apply.
Mon, Aug 14, 11:39 AM
coby added a comment to D35621: X86 Asm can't work properly with symbolic Scale.

few minors + add fixes in clang tests to adopt the added diagnostics.
nothing else has caught my eye

Mon, Aug 14, 4:13 AM

Sun, Aug 13

coby committed rL310789: [X86][AsmParser][AVX512] Error appropriately when K0 is tried as a write-mask.
[X86][AsmParser][AVX512] Error appropriately when K0 is tried as a write-mask
Sun, Aug 13, 5:06 AM
coby closed D36570: [X86][AsmParser][AVX512] Error appropriately when K0 is tried as a write-mask by committing rL310789: [X86][AsmParser][AVX512] Error appropriately when K0 is tried as a write-mask.
Sun, Aug 13, 5:06 AM

Sat, Aug 12

coby added a comment to D35621: X86 Asm can't work properly with symbolic Scale.

I applied all changes suggested by Coby.

Sat, Aug 12, 11:41 AM

Thu, Aug 10

coby added a comment to D35774: [x86][inline-asm]Extend support for memory reference expression.

lads?
anyone?

Thu, Aug 10, 6:01 AM
coby added inline comments to D35621: X86 Asm can't work properly with symbolic Scale.
Thu, Aug 10, 5:05 AM
coby added inline comments to D35621: X86 Asm can't work properly with symbolic Scale.
Thu, Aug 10, 3:51 AM
coby added inline comments to D36570: [X86][AsmParser][AVX512] Error appropriately when K0 is tried as a write-mask.
Thu, Aug 10, 3:23 AM
coby added a reviewer for D36570: [X86][AsmParser][AVX512] Error appropriately when K0 is tried as a write-mask: m_zuckerman.
Thu, Aug 10, 3:16 AM
coby created D36570: [X86][AsmParser][AVX512] Error appropriately when K0 is tried as a write-mask.
Thu, Aug 10, 1:49 AM
coby abandoned D35204: [LLVM][x86][Inline Asm] support for GCC style inline asm - Y<x> constraints.

superseded by D36369

Thu, Aug 10, 12:34 AM
coby commandeered D35204: [LLVM][x86][Inline Asm] support for GCC style inline asm - Y<x> constraints.
Thu, Aug 10, 12:34 AM
coby abandoned D35205: [Clang][x86][Inline Asm] support for GCC style inline asm - Y<x> constraints.

superseded by D36371

Thu, Aug 10, 12:33 AM
coby commandeered D35205: [Clang][x86][Inline Asm] support for GCC style inline asm - Y<x> constraints.
Thu, Aug 10, 12:33 AM

Wed, Aug 9

coby closed D36229: [X86][Asm] Allow negative immediate to appear before bracketed expression.

commited as rL310528

Wed, Aug 9, 11:10 PM
coby committed rL310529: [X86][Asm] Allow negative immediate to appear before bracketed expression.
[X86][Asm] Allow negative immediate to appear before bracketed expression
Wed, Aug 9, 2:53 PM
coby closed D36230: [X86][Asm] Allow negative immediate to appear before bracketed expression by committing rL310529: [X86][Asm] Allow negative immediate to appear before bracketed expression.
Wed, Aug 9, 2:53 PM
coby committed rL310528: [X86][Asm] Allow negative immediate to appear before bracketed expression.
[X86][Asm] Allow negative immediate to appear before bracketed expression
Wed, Aug 9, 2:50 PM
coby added inline comments to D36229: [X86][Asm] Allow negative immediate to appear before bracketed expression.
Wed, Aug 9, 2:20 PM
coby added a reviewer for D35621: X86 Asm can't work properly with symbolic Scale: rnk.

adding Reid.

Wed, Aug 9, 10:47 AM
coby added inline comments to D35621: X86 Asm can't work properly with symbolic Scale.
Wed, Aug 9, 10:46 AM
coby added inline comments to D35621: X86 Asm can't work properly with symbolic Scale.
Wed, Aug 9, 10:39 AM
coby committed rL310485: [X86][Asm]Allow far jmp/call to be picked when using explicit FWORD size….
[X86][Asm]Allow far jmp/call to be picked when using explicit FWORD size…
Wed, Aug 9, 8:35 AM
coby closed D35846: [X86][Asm]Allow far jmp/call to be picked when using explicit FWORD size specifier by committing rL310485: [X86][Asm]Allow far jmp/call to be picked when using explicit FWORD size….
Wed, Aug 9, 8:35 AM
coby committed rL310472: [X86][Ms-InlineAsm] Extend MS Dot operator to accept "this" + struct/class….
[X86][Ms-InlineAsm] Extend MS Dot operator to accept "this" + struct/class…
Wed, Aug 9, 6:32 AM
coby closed D36450: [X86][Ms-InlineAsm] Extend MS Dot operator to accept "this" + struct/class pointers aliases by committing rL310472: [X86][Ms-InlineAsm] Extend MS Dot operator to accept "this" + struct/class….
Wed, Aug 9, 6:32 AM
coby committed rL310467: [AsmParser][AVX512]Enhance OpMask/Zero/Merge syntax check rubostness.
[AsmParser][AVX512]Enhance OpMask/Zero/Merge syntax check rubostness
Wed, Aug 9, 5:33 AM
coby closed D35785: [AsmParser][AVX512]Enhance OpMask/Zero/Merge syntax check rubostness by committing rL310467: [AsmParser][AVX512]Enhance OpMask/Zero/Merge syntax check rubostness.
Wed, Aug 9, 5:33 AM
coby added inline comments to D36450: [X86][Ms-InlineAsm] Extend MS Dot operator to accept "this" + struct/class pointers aliases.
Wed, Aug 9, 4:50 AM

Tue, Aug 8

coby updated the diff for D35785: [AsmParser][AVX512]Enhance OpMask/Zero/Merge syntax check rubostness.

Added intel-dialect test as well

Tue, Aug 8, 3:49 AM
coby added a comment to D35621: X86 Asm can't work properly with symbolic Scale.

A few minors - does anyone have any more comments?

Tue, Aug 8, 3:02 AM
coby added a reviewer for D35621: X86 Asm can't work properly with symbolic Scale: coby.
Tue, Aug 8, 2:56 AM
coby created D36450: [X86][Ms-InlineAsm] Extend MS Dot operator to accept "this" + struct/class pointers aliases.
Tue, Aug 8, 12:39 AM
coby added reviewers for D35890: [Diagnostics][Inline-Asm]Allow targets to provide elaborations regarding failures resulted from (GNU-inline-asm) modifiers' apply: rnk, m_zuckerman.
Tue, Aug 8, 12:30 AM
coby added a reviewer for D36371: [Clang][x86][Inline Asm] support for GCC style inline asm - Y<x> constraints: m_zuckerman.
Tue, Aug 8, 12:30 AM
coby added a reviewer for D35774: [x86][inline-asm]Extend support for memory reference expression: m_zuckerman.
Tue, Aug 8, 12:29 AM

Mon, Aug 7

coby updated the diff for D35846: [X86][Asm]Allow far jmp/call to be picked when using explicit FWORD size specifier.

So should we test for it as well as the intel syntax?

Added some missing tests (intel & gnu)

Mon, Aug 7, 7:36 AM
coby added a comment to D35846: [X86][Asm]Allow far jmp/call to be picked when using explicit FWORD size specifier.

Is this consistent with GNU assembler behavior? Or does GNU assembler accept both versions?

Mon, Aug 7, 5:38 AM

Sun, Aug 6

coby added a comment to D35774: [x86][inline-asm]Extend support for memory reference expression.

pi|\|g

Sun, Aug 6, 11:38 AM
coby updated the summary of D36369: [LLVM][x86][Inline Asm] support for GCC style inline asm - Y<x> constraints.
Sun, Aug 6, 2:33 AM
coby updated the summary of D36371: [Clang][x86][Inline Asm] support for GCC style inline asm - Y<x> constraints.
Sun, Aug 6, 2:33 AM
coby created D36371: [Clang][x86][Inline Asm] support for GCC style inline asm - Y<x> constraints.
Sun, Aug 6, 2:32 AM
coby updated the summary of D36369: [LLVM][x86][Inline Asm] support for GCC style inline asm - Y<x> constraints.
Sun, Aug 6, 2:30 AM
coby created D36369: [LLVM][x86][Inline Asm] support for GCC style inline asm - Y<x> constraints.
Sun, Aug 6, 2:29 AM

Thu, Aug 3

coby added a reviewer for D35890: [Diagnostics][Inline-Asm]Allow targets to provide elaborations regarding failures resulted from (GNU-inline-asm) modifiers' apply: rengolin.
Thu, Aug 3, 11:23 AM

Wed, Aug 2

coby committed rL309841: [AsmParser][GAS-compatibility] Ignore an empty 'p2align' directive.
[AsmParser][GAS-compatibility] Ignore an empty 'p2align' directive
Wed, Aug 2, 10:37 AM
coby closed D36060: [AsmParser][GAS-compatibility] Ignore an empty 'p2align' directive by committing rL309841: [AsmParser][GAS-compatibility] Ignore an empty 'p2align' directive.
Wed, Aug 2, 10:36 AM
coby updated the summary of D36229: [X86][Asm] Allow negative immediate to appear before bracketed expression.
Wed, Aug 2, 10:29 AM
coby created D36230: [X86][Asm] Allow negative immediate to appear before bracketed expression.
Wed, Aug 2, 10:29 AM
coby added reviewers for D36229: [X86][Asm] Allow negative immediate to appear before bracketed expression: rnk, echristo.
Wed, Aug 2, 10:27 AM
coby created D36229: [X86][Asm] Allow negative immediate to appear before bracketed expression.
Wed, Aug 2, 10:27 AM

Tue, Aug 1

coby committed rL309672: [x86][inline-asm]Allow a pack of Control Debug to be properly picked.
[x86][inline-asm]Allow a pack of Control Debug to be properly picked
Tue, Aug 1, 3:51 AM
coby closed D36074: [x86][inline-asm]Allow a pack of Control Debug to be properly picked by committing rL309672: [x86][inline-asm]Allow a pack of Control Debug to be properly picked.
Tue, Aug 1, 3:51 AM
coby added inline comments to D35785: [AsmParser][AVX512]Enhance OpMask/Zero/Merge syntax check rubostness.
Tue, Aug 1, 12:00 AM

Mon, Jul 31

coby added inline comments to D35785: [AsmParser][AVX512]Enhance OpMask/Zero/Merge syntax check rubostness.
Mon, Jul 31, 1:17 PM
coby added a comment to D35785: [AsmParser][AVX512]Enhance OpMask/Zero/Merge syntax check rubostness.

ping

Mon, Jul 31, 11:41 AM
coby created D36074: [x86][inline-asm]Allow a pack of Control Debug to be properly picked.
Mon, Jul 31, 12:57 AM

Sun, Jul 30

coby created D36060: [AsmParser][GAS-compatibility] Ignore an empty 'p2align' directive.
Sun, Jul 30, 7:35 AM
coby committed rL309510: [x86][inline-asm][ms-compat] legalize the use of "jc/jz short <op>".
[x86][inline-asm][ms-compat] legalize the use of "jc/jz short <op>"
Sun, Jul 30, 4:14 AM
coby closed D35893: [x86][inline-asm][ms-compat] legalize the use of "jc/jz short <op>" by committing rL309510: [x86][inline-asm][ms-compat] legalize the use of "jc/jz short <op>".
Sun, Jul 30, 4:14 AM
coby committed rL309509: [x86][inline-asm][ms-compat] legalize the use of "jc/jz short <op>".
[x86][inline-asm][ms-compat] legalize the use of "jc/jz short <op>"
Sun, Jul 30, 4:13 AM
coby closed D35892: [x86][inline-asm][ms-compat] legalize the use of "jc/jz short <op>" by committing rL309509: [x86][inline-asm][ms-compat] legalize the use of "jc/jz short <op>".
Sun, Jul 30, 4:13 AM
coby added a comment to D35890: [Diagnostics][Inline-Asm]Allow targets to provide elaborations regarding failures resulted from (GNU-inline-asm) modifiers' apply.

anyone?

Sun, Jul 30, 4:00 AM
coby added a reviewer for D35785: [AsmParser][AVX512]Enhance OpMask/Zero/Merge syntax check rubostness: m_zuckerman.
Sun, Jul 30, 4:00 AM
coby committed rL309508: [x86][inline-asm]Allow a pack of Control Regs to be properly picked.
[x86][inline-asm]Allow a pack of Control Regs to be properly picked
Sun, Jul 30, 3:21 AM
coby closed D35903: [x86][inline-asm]Allow a pack of Control Regs to be properly picked by committing rL309508: [x86][inline-asm]Allow a pack of Control Regs to be properly picked.
Sun, Jul 30, 3:21 AM
coby added a comment to D35774: [x86][inline-asm]Extend support for memory reference expression.
In D35774#822227, @rnk wrote:
In D35774#819706, @coby wrote:

I would really prefer to see fewer rewrites and more pass through of inline assembly to MC
As do I. What are you suggesting? Keep rewrites to the absolute necessary minimum?
Pepole do some funky stuff on MS-Inline-Asm, a large portion of it should be 'legalized' in order for the Assembler to be able to digest.
The other way around is to allow this nonsense in the Assembler as well.
Both options are bad, but I believe the latter is a bit more.

If it's of any comfort - have this 'new' rewrite is to be allowed - AOK_Imm/AOK_ImmPrefix/AOK_SizeDirective can be implemented as a special case of it (AOK_DotOperator should be cancelled).
Net effect will remain the same - but it'll be (possibly) less of a mess of code.

Hm, rewriting everything in terms of a general AOK_IntelExpr does sound better, even if it is kind of the opposite direction from what I was thinking. :) What I'm imagining is that any time there is a symbolic reference to C/C++ enums, globals, or variables, we basically fully parse that operand and decompose it into the usual parts of an X86 memory operand: segment, base, index, scale, displacement, etc, check that those parts make sense, and emit a completely rewritten operand. Right now the parsing code in MC has a hard time deciding what's legal because it doesn't know whether certain C/C++ variable references refer to stack locals or global variables. A local variable will require using a base register, but a global will not, but on x86_64, a global will require %rip-relative addressing, which can't use base+index registers.

Doing this exposes a lot of x86 details that aren't really supposed to be in the supposedly target neutral asmparser interface. That battle may already be lost, though.

Sun, Jul 30, 3:07 AM

Sat, Jul 29

coby added a reviewer for D35785: [AsmParser][AVX512]Enhance OpMask/Zero/Merge syntax check rubostness: igorb.
Sat, Jul 29, 12:11 PM

Wed, Jul 26

coby created D35903: [x86][inline-asm]Allow a pack of Control Regs to be properly picked.
Wed, Jul 26, 10:19 AM
coby updated the summary of D35892: [x86][inline-asm][ms-compat] legalize the use of "jc/jz short <op>".
Wed, Jul 26, 8:21 AM
coby created D35893: [x86][inline-asm][ms-compat] legalize the use of "jc/jz short <op>".
Wed, Jul 26, 8:21 AM
coby created D35892: [x86][inline-asm][ms-compat] legalize the use of "jc/jz short <op>".
Wed, Jul 26, 8:18 AM
coby updated the diff for D35890: [Diagnostics][Inline-Asm]Allow targets to provide elaborations regarding failures resulted from (GNU-inline-asm) modifiers' apply.

Added forgotten test

Wed, Jul 26, 7:43 AM
coby added a comment to D35890: [Diagnostics][Inline-Asm]Allow targets to provide elaborations regarding failures resulted from (GNU-inline-asm) modifiers' apply.

Any tests checking new diagnostics?

Wed, Jul 26, 7:36 AM
coby added a comment to D35890: [Diagnostics][Inline-Asm]Allow targets to provide elaborations regarding failures resulted from (GNU-inline-asm) modifiers' apply.

Please add whatever reviewers you see as fit :)
thx

Wed, Jul 26, 7:16 AM
coby created D35890: [Diagnostics][Inline-Asm]Allow targets to provide elaborations regarding failures resulted from (GNU-inline-asm) modifiers' apply.
Wed, Jul 26, 7:15 AM

Tue, Jul 25

coby created D35846: [X86][Asm]Allow far jmp/call to be picked when using explicit FWORD size specifier.
Tue, Jul 25, 10:15 AM
coby added inline comments to D33278: [LLVM][x86][Inline Asm] - Enum support for MS syntax.
Tue, Jul 25, 8:57 AM
coby updated the diff for D35774: [x86][inline-asm]Extend support for memory reference expression.

Addressed @rnk 's (most of) comments

Tue, Jul 25, 1:01 AM
coby added a comment to D35774: [x86][inline-asm]Extend support for memory reference expression.

I would really prefer to see fewer rewrites and more pass through of inline assembly to MC
As do I. What are you suggesting? Keep rewrites to the absolute necessary minimum?
Pepole do some funky stuff on MS-Inline-Asm, a large portion of it should be 'legalized' in order for the Assembler to be able to digest.
The other way around is to allow this nonsense in the Assembler as well.
Both options are bad, but I believe the latter is a bit more.

Tue, Jul 25, 12:59 AM

Mon, Jul 24

coby added a comment to D34739: [X86][InlineAsm][Ms Compatibility]Prefer variable name over a register when the two collides.
In D34739#819343, @rnk wrote:

I reverted this in rL308926 because it broke sanitizer-windows: http://lab.llvm.org:8011/builders/sanitizer-windows/builds/14535

With this patch, clang would reject this code:

namespace cl { }
void f() {
  __asm {
    mov al, cl
  }
}
Mon, Jul 24, 1:58 PM
coby created D35785: [AsmParser][AVX512]Enhance OpMask/Zero/Merge syntax check rubostness.
Mon, Jul 24, 12:16 AM
coby committed rL308867: [X86][InlineAsm][Ms Compatibility]Prefer variable name over a register when the….
[X86][InlineAsm][Ms Compatibility]Prefer variable name over a register when the…
Mon, Jul 24, 12:08 AM
coby closed D34740: [X86][InlineAsm][Ms Compatibility]Prefer variable name over a register when the two collides by committing rL308867: [X86][InlineAsm][Ms Compatibility]Prefer variable name over a register when the….
Mon, Jul 24, 12:08 AM
coby committed rL308866: [X86][InlineAsm][Ms Compatibility]Prefer variable name over a register when the….
[X86][InlineAsm][Ms Compatibility]Prefer variable name over a register when the…
Mon, Jul 24, 12:05 AM
coby closed D34739: [X86][InlineAsm][Ms Compatibility]Prefer variable name over a register when the two collides by committing rL308866: [X86][InlineAsm][Ms Compatibility]Prefer variable name over a register when the….
Mon, Jul 24, 12:05 AM

Sun, Jul 23

coby abandoned D26587: [AVX512][inline-asm] Fix AVX512 inline assembly instruction resolution when the size qualifier of a memory operand is not specified explicitly..

superseded by rL302179

Sun, Jul 23, 6:26 AM
coby updated the summary of D35774: [x86][inline-asm]Extend support for memory reference expression.
Sun, Jul 23, 6:09 AM
coby updated the summary of D35775: [x86][inline-asm]Extend support for memory reference expression.
Sun, Jul 23, 6:09 AM