Page MenuHomePhabricator

scott.linder (Scott Linder)
User

Projects

User does not belong to any projects.

User Details

User Since
Jan 15 2018, 8:31 AM (91 w, 1 h)

Recent Activity

Mon, Oct 7

scott.linder added a comment to D53768: Add VerboseOutputStream to CompilerInstance.

Ping

Mon, Oct 7, 9:34 AM

Tue, Oct 1

scott.linder added a comment to D53768: Add VerboseOutputStream to CompilerInstance.

Ping

Tue, Oct 1, 2:02 PM

Fri, Sep 20

scott.linder updated the diff for D53768: Add VerboseOutputStream to CompilerInstance.

Same patch, this time with a working test.

Fri, Sep 20, 12:22 PM
scott.linder updated the diff for D53768: Add VerboseOutputStream to CompilerInstance.

After reading again I think I understand the ask now. Is this closer to what you had in mind?

Fri, Sep 20, 11:20 AM
scott.linder added inline comments to D53768: Add VerboseOutputStream to CompilerInstance.
Fri, Sep 20, 7:35 AM

Mon, Sep 16

scott.linder updated the diff for D53768: Add VerboseOutputStream to CompilerInstance.

Actually rebase this time

Mon, Sep 16, 2:56 PM
scott.linder updated the diff for D53768: Add VerboseOutputStream to CompilerInstance.

Rebase and ping

Mon, Sep 16, 12:20 PM

Sep 10 2019

scott.linder committed rL371529: Request commit access for scott.linder.
Request commit access for scott.linder
Sep 10 2019, 8:51 AM

Aug 28 2019

scott.linder committed rG04f6f2542186: [AMDGPU] Fix bug when calculating user_spgr_count for Code Object V3 assembler (authored by scott.linder).
[AMDGPU] Fix bug when calculating user_spgr_count for Code Object V3 assembler
Aug 28 2019, 12:38 PM
scott.linder committed rL370250: [AMDGPU] Fix bug when calculating user_spgr_count for Code Object V3 assembler.
[AMDGPU] Fix bug when calculating user_spgr_count for Code Object V3 assembler
Aug 28 2019, 12:37 PM
scott.linder closed D66900: [AMDGPU] Fix bug when calculating user_spgr_count for Code Object V3 assembler.
Aug 28 2019, 12:37 PM · Restricted Project
scott.linder created D66900: [AMDGPU] Fix bug when calculating user_spgr_count for Code Object V3 assembler.
Aug 28 2019, 11:33 AM · Restricted Project

Jul 3 2019

scott.linder accepted D63886: [AMDGPU] Added a new metadata for multi grid sync implicit argument..
Jul 3 2019, 10:04 AM · Restricted Project

Jul 1 2019

scott.linder added a comment to D63886: [AMDGPU] Added a new metadata for multi grid sync implicit argument..

We should update the v2 and v3 metadata portions of docs/AMDGPUUsage.rst, see https://llvm.org/docs/AMDGPUUsage.html#code-object-metadata

Jul 1 2019, 1:11 PM · Restricted Project

Jun 26 2019

scott.linder committed rGc95046501dc2: Fix leaks in LLVMCreateDisasmCPUFeatures (authored by scott.linder).
Fix leaks in LLVMCreateDisasmCPUFeatures
Jun 26 2019, 9:14 AM
scott.linder committed rL364444: Fix leaks in LLVMCreateDisasmCPUFeatures.
Fix leaks in LLVMCreateDisasmCPUFeatures
Jun 26 2019, 9:13 AM
scott.linder closed D63795: Fix leaks in LLVMCreateDisasmCPUFeatures.
Jun 26 2019, 9:13 AM · Restricted Project

Jun 25 2019

scott.linder created D63795: Fix leaks in LLVMCreateDisasmCPUFeatures.
Jun 25 2019, 3:09 PM · Restricted Project

Jun 14 2019

scott.linder accepted D63356: UpdateTestChecks: Consider .section as end of function for AMDGPU.
Jun 14 2019, 1:34 PM
scott.linder added a comment to D63337: [AMDGPU] Don't constrain callees with inlinehint from inlining on MaxBB check to prevent link failure.

The undefined references in the linker are OK. The user has to provide an external definition, or link against one. I don't think there is anything to do about this in LLVM.

Jun 14 2019, 9:12 AM · Restricted Project

May 8 2019

scott.linder accepted D60682: [AMDGPU] Fixed +DumpCode.
May 8 2019, 11:38 AM · Restricted Project

May 3 2019

scott.linder added a comment to D61097: [Sema] Emit warning for visibility attribute on internal-linkage declaration.

Richard Smith on cfe-dev pointed out some cases where this patch is incorrect, stemming from trying to calculate the linkage too early; the warning will either have to work without the use of isExternallyVisible or will have to be emitted later. This was reverted in r359858 for now.

May 3 2019, 1:17 PM · Restricted Project, Restricted Project

May 2 2019

scott.linder committed rGdaa3c5b1325e: [Sema] Emit warning for visibility attribute on internal-linkage declaration (authored by scott.linder).
[Sema] Emit warning for visibility attribute on internal-linkage declaration
May 2 2019, 12:02 PM
scott.linder committed rL359814: [Sema] Emit warning for visibility attribute on internal-linkage declaration.
[Sema] Emit warning for visibility attribute on internal-linkage declaration
May 2 2019, 12:02 PM
scott.linder committed rC359814: [Sema] Emit warning for visibility attribute on internal-linkage declaration.
[Sema] Emit warning for visibility attribute on internal-linkage declaration
May 2 2019, 12:02 PM
scott.linder closed D61097: [Sema] Emit warning for visibility attribute on internal-linkage declaration.
May 2 2019, 12:02 PM · Restricted Project, Restricted Project

May 1 2019

scott.linder committed rGc86769bfbf1b: Bump DIAG_SIZE_SEMA, as we've hit it. (authored by scott.linder).
Bump DIAG_SIZE_SEMA, as we've hit it.
May 1 2019, 9:45 AM
scott.linder committed rL359702: Bump DIAG_SIZE_SEMA, as we've hit it..
Bump DIAG_SIZE_SEMA, as we've hit it.
May 1 2019, 9:43 AM
scott.linder committed rC359702: Bump DIAG_SIZE_SEMA, as we've hit it..
Bump DIAG_SIZE_SEMA, as we've hit it.
May 1 2019, 9:43 AM
scott.linder updated the diff for D61097: [Sema] Emit warning for visibility attribute on internal-linkage declaration.

Use distinct sema diagnostic and do not return early.

May 1 2019, 9:32 AM · Restricted Project, Restricted Project

Apr 29 2019

scott.linder created D61274: [Sema][AST] Explicit visibility for OpenCL/CUDA kernels/variables.
Apr 29 2019, 12:18 PM · Restricted Project

Apr 26 2019

scott.linder added a comment to D60967: Move setTargetAttributes after setGVProperties in SetFunctionAttributes.

@rjmccall I'm not sure if this is the right place to continue discussing this, but I don't have a patch I am happy with and I would rather not post something half-baked.

Apr 26 2019, 5:16 PM · Restricted Project
scott.linder updated the diff for D61097: [Sema] Emit warning for visibility attribute on internal-linkage declaration.

Clean up use of auto

Apr 26 2019, 8:04 AM · Restricted Project, Restricted Project
scott.linder added inline comments to D61097: [Sema] Emit warning for visibility attribute on internal-linkage declaration.
Apr 26 2019, 8:04 AM · Restricted Project, Restricted Project

Apr 24 2019

scott.linder created D61097: [Sema] Emit warning for visibility attribute on internal-linkage declaration.
Apr 24 2019, 3:00 PM · Restricted Project, Restricted Project
scott.linder added a comment to D60967: Move setTargetAttributes after setGVProperties in SetFunctionAttributes.

@rjmccall Would you expect similar conflicts in explicit visibility to result in diagnostics? For example, marking a static variable with an explicit visibility attribute doesn't warn, instead the explicit visibility attribute is silently ignored. GCC 7.3 complains with warning: ‘__visibility__’ attribute ignored [-Wattributes]

Apr 24 2019, 12:49 PM · Restricted Project

Apr 23 2019

scott.linder committed rGfb59fef7dcd0: Move setTargetAttributes after setGVProperties in SetFunctionAttributes (authored by scott.linder).
Move setTargetAttributes after setGVProperties in SetFunctionAttributes
Apr 23 2019, 2:50 PM
scott.linder committed rL359039: Move setTargetAttributes after setGVProperties in SetFunctionAttributes.
Move setTargetAttributes after setGVProperties in SetFunctionAttributes
Apr 23 2019, 2:50 PM
scott.linder committed rC359039: Move setTargetAttributes after setGVProperties in SetFunctionAttributes.
Move setTargetAttributes after setGVProperties in SetFunctionAttributes
Apr 23 2019, 2:50 PM
scott.linder closed D60967: Move setTargetAttributes after setGVProperties in SetFunctionAttributes.
Apr 23 2019, 2:50 PM · Restricted Project
scott.linder added a comment to D60967: Move setTargetAttributes after setGVProperties in SetFunctionAttributes.

I suspect that other OpenCL and CUDA implementations don't care at all about symbol visibility for device-side code generation, and giving kernel functions default visibility seems like the right thing to do for the (relatively few) things at the AST level that are sensitive to that, like template visibility. Would you mind reaching out to other implementors about that?

This patch seems fine to me regardless.

Apr 23 2019, 2:38 PM · Restricted Project
scott.linder added a comment to D60967: Move setTargetAttributes after setGVProperties in SetFunctionAttributes.

Shouldn't it be an error if the user tries to give it hidden visibility?

We effectively consider the user explicitly specifying that a symbol is e.g. a kernel to also carry with it visibility information. We don't want to require the user to redundantly specify that a kernel is not hidden, when it is never meaningful for it to be hidden.

I understand, but if the user explicitly gives it hidden visibility, you should still diagnose that.

Also, shouldn't you just handle this by treating the kernel attribute as a source of explicit visibility at the Sema/AST level?

Apr 23 2019, 2:03 PM · Restricted Project
scott.linder added a comment to D60967: Move setTargetAttributes after setGVProperties in SetFunctionAttributes.

Shouldn't it be an error if the user tries to give it hidden visibility?

Apr 23 2019, 12:47 PM · Restricted Project
scott.linder added a comment to D60967: Move setTargetAttributes after setGVProperties in SetFunctionAttributes.

It seems reasonable to me for target hooks to run after global hooks, but can I ask why AMDGPU specifically relies on this?

Apr 23 2019, 12:06 PM · Restricted Project
scott.linder committed rG3eed961973d6: [AMDGPU] Fix hidden argument metadata duplication for V3 (authored by scott.linder).
[AMDGPU] Fix hidden argument metadata duplication for V3
Apr 23 2019, 7:32 AM
scott.linder closed D60915: [AMDGPU] Fix hidden argument metadata duplication for V3.

r358992

Apr 23 2019, 7:31 AM · Restricted Project
scott.linder committed rL358992: [AMDGPU] Fix hidden argument metadata duplication for V3.
[AMDGPU] Fix hidden argument metadata duplication for V3
Apr 23 2019, 7:29 AM

Apr 22 2019

scott.linder created D60967: Move setTargetAttributes after setGVProperties in SetFunctionAttributes.
Apr 22 2019, 9:06 AM · Restricted Project

Apr 19 2019

scott.linder created D60915: [AMDGPU] Fix hidden argument metadata duplication for V3.
Apr 19 2019, 1:00 PM · Restricted Project

Apr 17 2019

scott.linder added a comment to D60682: [AMDGPU] Fixed +DumpCode.

Mesa uses +DumpCode as well. Do we have a good solution for generation both ELF and textual output?

Apr 17 2019, 7:42 AM · Restricted Project

Apr 4 2019

scott.linder committed rGce2b61b2994b: Fix some MCTargetOptions Doxygen comments (NFC) (authored by scott.linder).
Fix some MCTargetOptions Doxygen comments (NFC)
Apr 4 2019, 1:38 PM
scott.linder committed rL357723: Fix some MCTargetOptions Doxygen comments (NFC).
Fix some MCTargetOptions Doxygen comments (NFC)
Apr 4 2019, 1:38 PM

Mar 29 2019

scott.linder committed rG0bc9f15ddd1e: [AMDGPU] Add an additional Code Object V3 assembler example (authored by scott.linder).
[AMDGPU] Add an additional Code Object V3 assembler example
Mar 29 2019, 10:49 AM
scott.linder committed rL357289: [AMDGPU] Add an additional Code Object V3 assembler example.
[AMDGPU] Add an additional Code Object V3 assembler example
Mar 29 2019, 10:48 AM
scott.linder closed D59949: [AMDGPU] Add an additional Code Object V3 assembler example.
Mar 29 2019, 10:48 AM · Restricted Project
scott.linder committed rGa377be6eb655: [AMDGPU] Switch default DWARF version to 5 (authored by scott.linder).
[AMDGPU] Switch default DWARF version to 5
Mar 29 2019, 10:45 AM
scott.linder committed rC357285: [AMDGPU] Switch default DWARF version to 5.
[AMDGPU] Switch default DWARF version to 5
Mar 29 2019, 10:45 AM
scott.linder committed rL357285: [AMDGPU] Switch default DWARF version to 5.
[AMDGPU] Switch default DWARF version to 5
Mar 29 2019, 10:45 AM
scott.linder closed D59008: [AMDGPU] Switch default dwarf version to 5.
Mar 29 2019, 10:45 AM · Restricted Project

Mar 28 2019

scott.linder updated the diff for D59008: [AMDGPU] Switch default dwarf version to 5.

Add a test to confirm split-dwarf is supported for the amdhsa OS in the driver.

Mar 28 2019, 11:19 AM · Restricted Project
scott.linder created D59949: [AMDGPU] Add an additional Code Object V3 assembler example.
Mar 28 2019, 10:44 AM · Restricted Project
scott.linder committed rGac20b745735e: [AMDGPU] Clarify Code Object V2/V3 differences in AMDGPUUsage (authored by scott.linder).
[AMDGPU] Clarify Code Object V2/V3 differences in AMDGPUUsage
Mar 28 2019, 8:09 AM
scott.linder committed rL357176: [AMDGPU] Clarify Code Object V2/V3 differences in AMDGPUUsage.
[AMDGPU] Clarify Code Object V2/V3 differences in AMDGPUUsage
Mar 28 2019, 8:09 AM
scott.linder closed D59782: [AMDGPU] Clarify Code Object V2/V3 differences in AMDGPUUsage.
Mar 28 2019, 8:09 AM · Restricted Project

Mar 25 2019

scott.linder created D59782: [AMDGPU] Clarify Code Object V2/V3 differences in AMDGPUUsage.
Mar 25 2019, 10:48 AM · Restricted Project
scott.linder added a comment to D59008: [AMDGPU] Switch default dwarf version to 5.

Ping

Mar 25 2019, 10:44 AM · Restricted Project

Mar 20 2019

scott.linder accepted D57028: [AMDGPU] Added MsgPack format PAL metadata.

LGTM, thanks! One small change to a comment.

Mar 20 2019, 7:57 AM · Restricted Project

Mar 19 2019

scott.linder added inline comments to D57028: [AMDGPU] Added MsgPack format PAL metadata.
Mar 19 2019, 8:25 AM · Restricted Project
scott.linder accepted D57027: [AMDGPU] Factored PAL metadata handling out into its own class.

LGTM, thanks! I think you can also drop "provides a method for use by LLPC to write metadata into LLVM IR" from the commit message.

Mar 19 2019, 7:37 AM · Restricted Project

Mar 18 2019

scott.linder added inline comments to D57027: [AMDGPU] Factored PAL metadata handling out into its own class.
Mar 18 2019, 4:14 PM · Restricted Project

Mar 5 2019

scott.linder created D59008: [AMDGPU] Switch default dwarf version to 5.
Mar 5 2019, 5:11 PM · Restricted Project

Mar 4 2019

scott.linder committed rGefec1396accb: [AMDGPU] Implement AMDGPUMCInstrAnalysis (authored by scott.linder).
[AMDGPU] Implement AMDGPUMCInstrAnalysis
Mar 4 2019, 7:01 PM
scott.linder committed rL355373: [AMDGPU] Implement AMDGPUMCInstrAnalysis.
[AMDGPU] Implement AMDGPUMCInstrAnalysis
Mar 4 2019, 7:01 PM
scott.linder closed D58400: [AMDGPU] Implement AMDGPUMCInstrAnalysis.
Mar 4 2019, 7:01 PM · Restricted Project

Feb 27 2019

scott.linder added a comment to D58400: [AMDGPU] Implement AMDGPUMCInstrAnalysis.

I'm not sure how the question is whether signed integers are supported. The instruction does treat the offset as signed

Feb 27 2019, 2:12 PM · Restricted Project
scott.linder updated the diff for D58400: [AMDGPU] Implement AMDGPUMCInstrAnalysis.

After digging a bit more into how we parse/print operands for sopp_br I think there are some more fundamental decisions to make beyond just "do we support signed integers", so I want to avoid changing anything in this patch. I will revisit how we treat them in the future and make any breaking changes all at once, rather than spreading them out.

Feb 27 2019, 12:56 PM · Restricted Project

Feb 22 2019

scott.linder accepted D58512: AMDGPU/GlobalISel: Insert waterfall loop for vector indexing.

LGTM then, but like I said I'm not very familiar with GlobalISel so more eyes might be good

Feb 22 2019, 9:31 AM
scott.linder added inline comments to D58512: AMDGPU/GlobalISel: Insert waterfall loop for vector indexing.
Feb 22 2019, 8:49 AM
scott.linder added a comment to D58400: [AMDGPU] Implement AMDGPUMCInstrAnalysis.

Does anyone have an opinion on returning negative branch targets (e.g. <keep_symbol+0xfffffffffffe0018>)? I don't know how this would ever come up in hardware anyway, or what the hardware would do, but it doesn't seem very helpful in the disassembly.

The hardware doesn't know about the symbol? Do you mean for overflow?

Feb 22 2019, 7:55 AM · Restricted Project

Feb 21 2019

scott.linder added a comment to D58400: [AMDGPU] Implement AMDGPUMCInstrAnalysis.

Does anyone have an opinion on returning negative branch targets (e.g. <keep_symbol+0xfffffffffffe0018>)? I don't know how this would ever come up in hardware anyway, or what the hardware would do, but it doesn't seem very helpful in the disassembly.

Feb 21 2019, 9:19 AM · Restricted Project

Feb 20 2019

scott.linder updated the diff for D58400: [AMDGPU] Implement AMDGPUMCInstrAnalysis.

Improved error message.

Feb 20 2019, 2:25 PM · Restricted Project
scott.linder added inline comments to D58400: [AMDGPU] Implement AMDGPUMCInstrAnalysis.
Feb 20 2019, 2:24 PM · Restricted Project
scott.linder updated the diff for D58400: [AMDGPU] Implement AMDGPUMCInstrAnalysis.

Address feedback. Sign-extend all s_branch immediates and change the assembler syntax to represent these as true negative numbers.

Feb 20 2019, 1:54 PM · Restricted Project
scott.linder added inline comments to D58400: [AMDGPU] Implement AMDGPUMCInstrAnalysis.
Feb 20 2019, 8:01 AM · Restricted Project

Feb 19 2019

scott.linder created D58400: [AMDGPU] Implement AMDGPUMCInstrAnalysis.
Feb 19 2019, 11:39 AM · Restricted Project

Feb 12 2019

scott.linder committed rG80a1ee46d87b: [AMDGPU] Require at least protected visibility for certain symbols (authored by scott.linder).
[AMDGPU] Require at least protected visibility for certain symbols
Feb 12 2019, 10:31 AM
scott.linder committed rC353870: [AMDGPU] Require at least protected visibility for certain symbols.
[AMDGPU] Require at least protected visibility for certain symbols
Feb 12 2019, 10:31 AM
scott.linder committed rL353870: [AMDGPU] Require at least protected visibility for certain symbols.
[AMDGPU] Require at least protected visibility for certain symbols
Feb 12 2019, 10:31 AM
scott.linder closed D56871: [AMDGPU] Require at least protected visibility for certain symbols.
Feb 12 2019, 10:31 AM · Restricted Project

Feb 11 2019

scott.linder committed rG72a0f4e8db40: [IRReader] Expose getLazyIRModule (authored by scott.linder).
[IRReader] Expose getLazyIRModule
Feb 11 2019, 2:01 PM
scott.linder committed rL353755: [IRReader] Expose getLazyIRModule.
[IRReader] Expose getLazyIRModule
Feb 11 2019, 2:01 PM
scott.linder closed D56203: [IRReader] Expose getLazyIRModule.
Feb 11 2019, 2:01 PM · Restricted Project
scott.linder added a comment to D56203: [IRReader] Expose getLazyIRModule.

Ping

Feb 11 2019, 9:50 AM · Restricted Project
scott.linder added a comment to D56871: [AMDGPU] Require at least protected visibility for certain symbols.

Ping

Feb 11 2019, 9:48 AM · Restricted Project

Feb 8 2019

Herald added a project to D57028: [AMDGPU] Added MsgPack format PAL metadata: Restricted Project.
Feb 8 2019, 8:16 AM · Restricted Project

Feb 5 2019

scott.linder committed rGe2c584741469: [AMDGPU] Consider XOR in waterfall loop as a terminator (authored by scott.linder).
[AMDGPU] Consider XOR in waterfall loop as a terminator
Feb 5 2019, 11:51 AM
scott.linder committed rL353207: [AMDGPU] Consider XOR in waterfall loop as a terminator.
[AMDGPU] Consider XOR in waterfall loop as a terminator
Feb 5 2019, 11:50 AM
scott.linder closed D57703: [AMDGPU] Consider XOR in waterfall loop as a terminator.
Feb 5 2019, 11:50 AM · Restricted Project
scott.linder updated the diff for D57703: [AMDGPU] Consider XOR in waterfall loop as a terminator.

I think MIR is the best place to look; this now tests that the right pseudo is produced and that spills don't interleave the terminators.

Feb 5 2019, 7:35 AM · Restricted Project

Feb 4 2019

scott.linder updated the diff for D57703: [AMDGPU] Consider XOR in waterfall loop as a terminator.

I agree the test does not test much at all, it was just the minimum I could think of considering none of our existing tests notice the change. I'm not sure what pass you mean by 'isel', but 'stop-after=amdgpu-isel' is too early to see the SI_INDIRECT_SRC_* psuedo expanded.

Feb 4 2019, 12:47 PM · Restricted Project