Feed All Stories

Today

hans committed rL322711: Creating release candidate rc1 from release_600 branch.
Creating release candidate rc1 from release_600 branch
Wed, Jan 17, 10:06 AM
tkrasnukha committed rL322718: [ARC] Add missing condition codes..
[ARC] Add missing condition codes.
Wed, Jan 17, 10:05 AM
hans committed rL322717: Creating release candidate rc1 from release_600 branch.
Creating release candidate rc1 from release_600 branch
Wed, Jan 17, 10:05 AM
hans committed rL322716: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 10:05 AM
hans committed rL322715: Creating release candidate rc1 from release_600 branch.
Creating release candidate rc1 from release_600 branch
Wed, Jan 17, 10:05 AM
hans committed rL322714: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 10:04 AM
cfang added inline comments to D39912: AMDGPU/SI: Implement d16 support for image intrinsics.
Wed, Jan 17, 10:04 AM
hans committed rL322713: Creating release candidate rc1 from release_600 branch.
Creating release candidate rc1 from release_600 branch
Wed, Jan 17, 10:04 AM
clayborg accepted D42182: Add LLDB_LOG_ERROR (?).
Wed, Jan 17, 10:04 AM
hans committed rL322712: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 10:03 AM
rnk accepted D41813: [builtins] Enable CRT_HAS_128BIT for MSVC.

Thanks, looks good. I will apply this.

Wed, Jan 17, 10:03 AM
hans committed rL322697: Creating release candidate rc1 from release_600 branch.
Creating release candidate rc1 from release_600 branch
Wed, Jan 17, 10:03 AM
hans committed rL322710: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 10:02 AM
hans committed rL322708: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 10:01 AM
craig.topper added a comment to D41895: [X86] Another attempt at support prefer-vector-width function attribute.

Ping

Wed, Jan 17, 10:01 AM
hans committed rL322707: Creating release candidate rc1 from release_600 branch.
Creating release candidate rc1 from release_600 branch
Wed, Jan 17, 10:01 AM
craig.topper added a reviewer for D41895: [X86] Another attempt at support prefer-vector-width function attribute: spatel.
Wed, Jan 17, 10:00 AM
hans committed rL322706: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 10:00 AM
hans committed rL322705: Creating release candidate rc1 from release_600 branch.
Creating release candidate rc1 from release_600 branch
Wed, Jan 17, 10:00 AM
hans committed rL322704: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 10:00 AM
tatyana-krasnukha closed D42172: [ARC] Add missing condition codes..
Wed, Jan 17, 10:00 AM
hans committed rL322703: Creating release candidate rc1 from release_600 branch.
Creating release candidate rc1 from release_600 branch
Wed, Jan 17, 10:00 AM
clayborg requested changes to D42145: [lldb] Use vFlash commands when writing to target's flash memory regions.

Very nice overall. See inlined comments. Big issues are:

  • GDBRemoteCommunication::WriteEscapedBinary() is not needed as StreamGDBRemote::PutEscapedBytes() does this already
  • Remove the batch start and end functions in process if possible and have ProcessGDBRemote::DoWriteMemory() just "do the right thing"
  • Can we actually cache the results of the qXfer:memory-map for the entire process lifetime?
  • Remove the new ProcessGDBRemote::GetMemoryMapRegion() and fold into GDBRemoteCommunicationClient::GetMemoryRegionInfo() as needed
Wed, Jan 17, 10:00 AM
sbc100 accepted D42075: [WebAssembly] Symbol changes #1, LLVM.

LGTM, I'll try to land this one now.

Wed, Jan 17, 10:00 AM
hans committed rL322702: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 9:56 AM
hans committed rL322701: Creating release candidate rc1 from release_600 branch.
Creating release candidate rc1 from release_600 branch
Wed, Jan 17, 9:56 AM
hans committed rL322700: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 9:54 AM
hans committed rL322698: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 9:54 AM
hans committed rL322696: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 9:54 AM
thegameg updated subscribers of D39386: [Power9] Allow gpr callee saved spills in prologue to vector registers rather than stack.
Wed, Jan 17, 9:54 AM
NoQ added inline comments to D42131: [analyzer] Helper method for checking whether two symbolic values are equal.
Wed, Jan 17, 9:54 AM
hans committed rL322694: Creating release directory for release_600..
Creating release directory for release_600.
Wed, Jan 17, 9:52 AM
rnk added a comment to D42171: X86CallFrameOptimization: Bail on win64cc calls.

Here is the MI in question:

Wed, Jan 17, 9:52 AM
hans committed rL322693: Merging r322003:.
Merging r322003:
Wed, Jan 17, 9:50 AM
spatel added a comment to D42088: [x86] shrink 'and' immediate values by setting the high bits (PR35907).

Here's an example of the SimplifyDemandedBits failure that avoids the preliminary div transform in the existing test:

Wed, Jan 17, 9:50 AM
ilya-biryukov updated the diff for D41991: [clangd] Always use preamble (even stale) for code completion.
  • s/buffer for MainFile/MainFile/
Wed, Jan 17, 9:47 AM
bogner added inline comments to D42159: [GlobalISel] Making MachineCSE runnable in the middle of the GlobalISel.
Wed, Jan 17, 9:44 AM
NoQ accepted D42155: [analyzer] introduce getSVal(Stmt *) helper on ExplodedNode, make sure the helper is used consistently.
Wed, Jan 17, 9:43 AM
NoQ added a comment to D42155: [analyzer] introduce getSVal(Stmt *) helper on ExplodedNode, make sure the helper is used consistently.

Looks pretty much awesome!

Wed, Jan 17, 9:43 AM
juliehockett committed rCTE322692: [clang-tidy] Fixing Fuchsia overloaded operator warning message.
[clang-tidy] Fixing Fuchsia overloaded operator warning message
Wed, Jan 17, 9:43 AM
juliehockett committed rL322692: [clang-tidy] Fixing Fuchsia overloaded operator warning message.
[clang-tidy] Fixing Fuchsia overloaded operator warning message
Wed, Jan 17, 9:43 AM
juliehockett closed D42120: [clang-tidy] Fixing Fuchsia overloaded operator warning message.
Wed, Jan 17, 9:43 AM · Restricted Project
RKSimon accepted D42153: [X86] When legalizing (v64i1 select i8, v64i1, v64i1) make sure not to introduce bitcasts to i64 in 32-bit mode.

LGTM

Wed, Jan 17, 9:41 AM
nlopes added a comment to D41944: [LLVM][IR][LIT] support of 'no-overflow' flag for sdiv\udiv instructions.

Before accepting this patch, we really need to see benchmark results. I'm not going to change clang to start emitting non-UB divs if the perf is going to be horrible. We need data.
Otherwise I don't see the need for this poison version of division. Could you elaborate if your plan is to expose this somehow to the application developer?

I'm sorry if this questions have been properly answered in the past. If so, could you please link them here?

In general the proposed feature allows compiler to start speculating div without worrying too much of div-by-zero etc. so for example you can do instruction hoisting or vectorizing predicated sdiv.
We are currently focused on vectorizing predicated div instruction and our implementation shows around 20-30% improvements on several tests of coremark-pro and denbench.

Wed, Jan 17, 9:40 AM
petecoup accepted D42172: [ARC] Add missing condition codes..

Hi Tatyana,

Wed, Jan 17, 9:40 AM
ioeric committed rL322691: [Tooling] Don't deduplicate tool results in the All-TUs executor..
[Tooling] Don't deduplicate tool results in the All-TUs executor.
Wed, Jan 17, 9:38 AM
ioeric committed rC322691: [Tooling] Don't deduplicate tool results in the All-TUs executor..
[Tooling] Don't deduplicate tool results in the All-TUs executor.
Wed, Jan 17, 9:38 AM
ioeric closed D42111: [Tooling] Don't deduplicate tool results in the All-TUs executor..
Wed, Jan 17, 9:38 AM
aemerson added a reviewer for D42175: [AArch64][GlobalISel] Add isel support for global values in the large code model: ab.
Wed, Jan 17, 9:38 AM
benhamilton committed rC322690: [Format] Improve ObjC header guessing heuristic.
[Format] Improve ObjC header guessing heuristic
Wed, Jan 17, 9:34 AM
benhamilton committed rL322690: [Format] Improve ObjC header guessing heuristic.
[Format] Improve ObjC header guessing heuristic
Wed, Jan 17, 9:34 AM
benhamilton closed D42135: [Format] Improve ObjC header guessing heuristic.
Wed, Jan 17, 9:34 AM
benhamilton updated the diff for D42135: [Format] Improve ObjC header guessing heuristic.

Use unordered_set, consolidate at-token checks

Wed, Jan 17, 9:32 AM
dp added inline comments to D42186: [AMDGPU][MC] Enabled disassembler for image atomic operations.
Wed, Jan 17, 9:31 AM
phosek committed rCRT322689: Revert "[libFuzzer] Support using libc++".
Revert "[libFuzzer] Support using libc++"
Wed, Jan 17, 9:26 AM
phosek committed rL322689: Revert "[libFuzzer] Support using libc++".
Revert "[libFuzzer] Support using libc++"
Wed, Jan 17, 9:26 AM
arsenm added inline comments to D42186: [AMDGPU][MC] Enabled disassembler for image atomic operations.
Wed, Jan 17, 9:25 AM
ioeric added a comment to D41946: [clangd] Add support for different file URI schemas..

Thanks for the comments! I've refactored the code according to your suggestions. PTAL

Wed, Jan 17, 9:25 AM
ilya-biryukov added inline comments to D42174: [clangd] Refactored threading in ClangdServer.
Wed, Jan 17, 9:25 AM
magabari added a comment to D41944: [LLVM][IR][LIT] support of 'no-overflow' flag for sdiv\udiv instructions.

Before accepting this patch, we really need to see benchmark results. I'm not going to change clang to start emitting non-UB divs if the perf is going to be horrible. We need data.
Otherwise I don't see the need for this poison version of division. Could you elaborate if your plan is to expose this somehow to the application developer?

I'm sorry if this questions have been properly answered in the past. If so, could you please link them here?

Wed, Jan 17, 9:25 AM
dp added a comment to D42186: [AMDGPU][MC] Enabled disassembler for image atomic operations.

Sorry, I used Size as if it were ValueType :-)

Wed, Jan 17, 9:24 AM
ioeric updated the diff for D41946: [clangd] Add support for different file URI schemas..
  • Addressed review comments.
Wed, Jan 17, 9:23 AM
Herald added a reviewer for D41946: [clangd] Add support for different file URI schemas.: jkorous-apple.
Wed, Jan 17, 9:23 AM
juliehockett updated the diff for D42116: [clang-tidy] Adding Fuchsia checker for trailing returns.

Updating matchers & fixing comments

Wed, Jan 17, 9:22 AM · Restricted Project
dp added inline comments to D42186: [AMDGPU][MC] Enabled disassembler for image atomic operations.
Wed, Jan 17, 9:21 AM
ncw added a comment to D42030: [WebAssembly] Define __heap_base global.

In fact, I wonder now as I re-do the calculations if I shouldn't in fact use 16KiB page size (reducing wastage to 10% for 129KiB allocations, at the expense of having a 16KiB mmap bitmap.)

Wed, Jan 17, 9:20 AM
sdesmalen added a comment to D41695: [Metadata] Extend 'count' field of DISubrange to take a metadata node.

Thanks @rnk for confirming and everyone for your input!

Wed, Jan 17, 9:20 AM
ABataev added a comment to D42168: [OpenMP] Generate unique name for offloading entries.
  1. Please, mark the record as packed.

Where and how do I need to do this? I found that PackedAttr is checked on FieldDecls but I'm not sure if we go through that code path and how I could add it...

Wed, Jan 17, 9:20 AM
jonpa committed rL322688: [SystemZ] Handle BRCTH branches correctly in SystemZLongBranch.cpp..
[SystemZ] Handle BRCTH branches correctly in SystemZLongBranch.cpp.
Wed, Jan 17, 9:17 AM
ncw added inline comments to D42030: [WebAssembly] Define __heap_base global.
Wed, Jan 17, 9:13 AM
hokein added inline comments to D41852: [clang-tidy] Don't generate fix for argument constructed from std::initializer_list..
Wed, Jan 17, 9:13 AM
jolesiak accepted D42135: [Format] Improve ObjC header guessing heuristic.
Wed, Jan 17, 9:10 AM
arsenm added inline comments to D42186: [AMDGPU][MC] Enabled disassembler for image atomic operations.
Wed, Jan 17, 9:08 AM
krytarowski added a comment to D42064: Add new NetBSD interceptors: getgrouplist(3) & getgroupmembership(3).

Tests from this revision require merged another patch in D42068

Wed, Jan 17, 9:08 AM · Restricted Project
krytarowski updated the diff for D42068: Add new interceptors for pwcache(3)-style functions.
  • correct typo
Wed, Jan 17, 9:07 AM · Restricted Project
ilya-biryukov requested changes to D41535: Add -vfsoverlay option for Clang-Tidy.

Sorry for the delay

Wed, Jan 17, 9:07 AM · Restricted Project
krytarowski updated the diff for D42068: Add new interceptors for pwcache(3)-style functions.
  • add tests
  • correct interceptors
Wed, Jan 17, 9:03 AM · Restricted Project
sammccall added a comment to D42173: [clangd] Simplify code handling compile commands.

This looks better overall to me.

Wed, Jan 17, 9:02 AM
syzaara added a comment to D39386: [Power9] Allow gpr callee saved spills in prologue to vector registers rather than stack.

@hfinkel Can you please finish reviewing this.

Wed, Jan 17, 9:02 AM
davide accepted D42182: Add LLDB_LOG_ERROR (?).

Forgot to click accept.

Wed, Jan 17, 9:01 AM
bollu added a comment to D42142: Adding Live-range reordering for Polly.

@Meinersbur I agree. However, I don't think that this is the right place for the code either. It is more a DependenceInfo thing, and maybe it makes sense to teach DependenceInfo to generate this information for us?

Wed, Jan 17, 9:00 AM
dp created D42186: [AMDGPU][MC] Enabled disassembler for image atomic operations.
Wed, Jan 17, 9:00 AM
sbc100 added inline comments to D42030: [WebAssembly] Define __heap_base global.
Wed, Jan 17, 8:59 AM
davide added a comment to D42182: Add LLDB_LOG_ERROR (?).

I think this is fine, but I'll defer to Zachary.

Wed, Jan 17, 8:56 AM
nlopes added a comment to D41944: [LLVM][IR][LIT] support of 'no-overflow' flag for sdiv\udiv instructions.

Before accepting this patch, we really need to see benchmark results. I'm not going to change clang to start emitting non-UB divs if the perf is going to be horrible. We need data.
Otherwise I don't see the need for this poison version of division. Could you elaborate if your plan is to expose this somehow to the application developer?

Wed, Jan 17, 8:53 AM
Hahnfeld added a comment to D42168: [OpenMP] Generate unique name for offloading entries.
  1. Please, mark the record as packed.
Wed, Jan 17, 8:52 AM
bkramer requested changes to D42185: [ASTMatcher] Add isScoped matcher for enumDecl..

dump_ast_matchers.py should be working again now.

Wed, Jan 17, 8:52 AM
d0k committed rL322687: Update dump_ast_matchers for many recent changes..
Update dump_ast_matchers for many recent changes.
Wed, Jan 17, 8:51 AM
d0k committed rC322687: Update dump_ast_matchers for many recent changes..
Update dump_ast_matchers for many recent changes.
Wed, Jan 17, 8:51 AM
sammccall accepted D41991: [clangd] Always use preamble (even stale) for code completion.

Thanks, LG!

Wed, Jan 17, 8:50 AM
ilya-biryukov updated the diff for D41991: [clangd] Always use preamble (even stale) for code completion.
  • Fixed a typo
Wed, Jan 17, 8:45 AM
seurer added a comment to D42157: [clang-cl] Let /FA output use intel assembly..

This patch fails on powerpc64

Wed, Jan 17, 8:44 AM
ilya-biryukov added inline comments to D41991: [clangd] Always use preamble (even stale) for code completion.
Wed, Jan 17, 8:44 AM
ilya-biryukov updated the diff for D41991: [clangd] Always use preamble (even stale) for code completion.
  • Use a better comment suggest by Sam
Wed, Jan 17, 8:44 AM
probinson added a comment to D41761: Introduce llvm.nospeculateload intrinsic.

It would be awesome to have static analysis rules to help identify *where* to put these intrinsics. Is somebody working on that? Or did I miss it?

Wed, Jan 17, 8:41 AM
ilya-biryukov updated the diff for D41991: [clangd] Always use preamble (even stale) for code completion.
  • Rewrote the comment for prepareCompilerInstance
  • Call CanReuse() and discard its results. We have clients that rely on having the stat() calls for files from preamble
Wed, Jan 17, 8:38 AM
probinson added a comment to D42082: Add DWARF for discriminated unions.

The DWARF spec says: "If the variant part has a discriminant, the discriminant is represented by a separate debugging information entry which is a child of the variant part entry. This entry has the form of a structure data member entry. The variant part entry will have a DW_AT_descr attribute whose value is a reference to the member entry for the discriminant."

Wed, Jan 17, 8:37 AM · debug-info
hans committed rL322686: Merging r321751, r321806, and r321878:.
Merging r321751, r321806, and r321878:
Wed, Jan 17, 8:35 AM
ilya-biryukov accepted D41852: [clang-tidy] Don't generate fix for argument constructed from std::initializer_list..

LGTM modulo a change to the 'initializer_list.h'. Do we really need it for this patch?

Wed, Jan 17, 8:34 AM
aaron.ballman requested changes to D42185: [ASTMatcher] Add isScoped matcher for enumDecl..

This needs to be added to the dynamic matcher registry and also should not be commit until the documentation is generated for it.

Wed, Jan 17, 8:32 AM