Page MenuHomePhabricator

MatzeB (Matthias Braun)
User

Projects

User does not belong to any projects.

User Details

User Since
Aug 13 2013, 3:10 PM (464 w, 14 h)

Not actively working on LLVM at the moment

Recent Activity

Tue, Jun 28

Herald added a project to D107695: [llvm] [lit] Support forcing lexical test order: Restricted Project.

Hey, I don't know if this has anything to do with this change.

Tue, Jun 28, 3:45 PM · Restricted Project, Restricted Project
MatzeB committed rGa145a32544a8: build_symbolizer: Set LLVM_DEFAULT_TARGET_TRIPLE in llvm builds (authored by MatzeB).
build_symbolizer: Set LLVM_DEFAULT_TARGET_TRIPLE in llvm builds
Tue, Jun 28, 2:50 PM · Restricted Project, Restricted Project
MatzeB committed rG2030e6496aae: Avoid ar_to_bc.sh script in symbolizer build; add -ignore-non-bitcode flag to… (authored by MatzeB).
Avoid ar_to_bc.sh script in symbolizer build; add -ignore-non-bitcode flag to…
Tue, Jun 28, 2:50 PM · Restricted Project, Restricted Project, Restricted Project
MatzeB closed D128731: build_symbolizer: Set LLVM_DEFAULT_TARGET_TRIPLE in llvm builds.
Tue, Jun 28, 2:50 PM · Restricted Project, Restricted Project
MatzeB closed D128729: Avoid ar_to_bc.sh script in symbolizer build; add -ignore-non-bitcode flag to llvm-link.
Tue, Jun 28, 2:50 PM · Restricted Project, Restricted Project, Restricted Project
MatzeB added a comment to D128729: Avoid ar_to_bc.sh script in symbolizer build; add -ignore-non-bitcode flag to llvm-link.

Judging from reaction from folks to older diffs, I have a feeling the script may also be running within apple internal builds...

Tue, Jun 28, 10:14 AM · Restricted Project, Restricted Project, Restricted Project
MatzeB updated subscribers of D128729: Avoid ar_to_bc.sh script in symbolizer build; add -ignore-non-bitcode flag to llvm-link.

It's not clear from description that it's done to fix those scripts.

Tue, Jun 28, 10:11 AM · Restricted Project, Restricted Project, Restricted Project
MatzeB updated the summary of D128729: Avoid ar_to_bc.sh script in symbolizer build; add -ignore-non-bitcode flag to llvm-link.
Tue, Jun 28, 10:10 AM · Restricted Project, Restricted Project, Restricted Project
MatzeB updated the diff for D128729: Avoid ar_to_bc.sh script in symbolizer build; add -ignore-non-bitcode flag to llvm-link.

Added test for new llvm-link option.

Tue, Jun 28, 9:44 AM · Restricted Project, Restricted Project, Restricted Project
MatzeB added a comment to D128729: Avoid ar_to_bc.sh script in symbolizer build; add -ignore-non-bitcode flag to llvm-link.

Do we need a test for llvm-link?

Tue, Jun 28, 9:32 AM · Restricted Project, Restricted Project, Restricted Project
MatzeB updated the diff for D128729: Avoid ar_to_bc.sh script in symbolizer build; add -ignore-non-bitcode flag to llvm-link.
Tue, Jun 28, 7:16 AM · Restricted Project, Restricted Project, Restricted Project
MatzeB added a comment to D128731: build_symbolizer: Set LLVM_DEFAULT_TARGET_TRIPLE in llvm builds.

For background: We had failing builds because the host clang was configured for a default triple of x86_64-redhat-linux-gnu but the symbolizer build using x86_64-unknown-linux-gnu. I suspect this started failing in combination with the new LLVM_ENABLE_PER_TARGET_RUNTIME_DIR default in LLVM expecting libcxx headers and libraries in a subdirectory matching the target triple.

Tue, Jun 28, 7:13 AM · Restricted Project, Restricted Project
MatzeB requested review of D128731: build_symbolizer: Set LLVM_DEFAULT_TARGET_TRIPLE in llvm builds.
Tue, Jun 28, 7:10 AM · Restricted Project, Restricted Project
MatzeB updated the diff for D128729: Avoid ar_to_bc.sh script in symbolizer build; add -ignore-non-bitcode flag to llvm-link.
Tue, Jun 28, 7:04 AM · Restricted Project, Restricted Project, Restricted Project
MatzeB added inline comments to D126728: ar_to_bc.sh: Ignore non-bitcode files in archives.
Tue, Jun 28, 7:02 AM · Restricted Project, Restricted Project
MatzeB requested review of D128729: Avoid ar_to_bc.sh script in symbolizer build; add -ignore-non-bitcode flag to llvm-link.
Tue, Jun 28, 7:02 AM · Restricted Project, Restricted Project, Restricted Project

Mon, Jun 27

MatzeB added inline comments to D126728: ar_to_bc.sh: Ignore non-bitcode files in archives.
Mon, Jun 27, 9:09 PM · Restricted Project, Restricted Project
MatzeB committed rTa2f875798692: [CMake][SingleSource] Fix compile with profile generation (authored by aidengrossman).
[CMake][SingleSource] Fix compile with profile generation
Mon, Jun 27, 4:41 PM · Restricted Project
MatzeB closed D128225: [CMake][SingleSource] Fix compile with profile generation.
Mon, Jun 27, 4:41 PM · Restricted Project
MatzeB added inline comments to D127516: RegisterCoalescer: Fix verifier error when merging copy of undef.
Mon, Jun 27, 11:31 AM · Restricted Project, Restricted Project
MatzeB added inline comments to D127516: RegisterCoalescer: Fix verifier error when merging copy of undef.
Mon, Jun 27, 11:12 AM · Restricted Project, Restricted Project
MatzeB added a comment to D127516: RegisterCoalescer: Fix verifier error when merging copy of undef.

For background: My understanding of IMPLICIT_DEF is that it was a bit of a workaround/hack necessary because for a given VReg on a control flow join you can have different "undef"/"not undef" merge. Unfortunately being out of SSA form we can't just attach an undef flag to the respective PHI operand as there is no PHI anymore. I always felt a bit uneasy with IMPLICIT_DEF getting used in other situations where they are not at the end of a block just for joining with a different definition/live value in the next block...

Mon, Jun 27, 11:11 AM · Restricted Project, Restricted Project
MatzeB added a comment to D127516: RegisterCoalescer: Fix verifier error when merging copy of undef.

I am not sure I understand this change... Wasn't the intention of the existing that code that when you have something like:

%2 = COPY undef %1
  USE %2

that you merge the regs to remove the copy and add an undef to the use operands? =>

USE undef %1
Mon, Jun 27, 11:02 AM · Restricted Project, Restricted Project
MatzeB added a comment to D128583: CodeGen: Remove AliasAnalysis from regalloc.

Yes, this could potentially regress rematerializing instructions from target load intrinsics. We would have to pass AA through getTgtMemIntrinsic to allow targets to set the flag themselves.

Mon, Jun 27, 10:49 AM · Restricted Project, Restricted Project
MatzeB added a comment to D128583: CodeGen: Remove AliasAnalysis from regalloc.

Maybe we should leave this diff open for a day or two to wait for extra feedback?

Mon, Jun 27, 10:39 AM · Restricted Project, Restricted Project
MatzeB accepted D128583: CodeGen: Remove AliasAnalysis from regalloc.

I always found it very unfortunate that we keep IR references around in MIR to do alias analysis queries. This appears to remove a lot of those users (all but the ones in the schedule graph construction?), so I highly welcome the change!

Mon, Jun 27, 10:38 AM · Restricted Project, Restricted Project

Thu, Jun 23

MatzeB added a comment to D128437: [fastregalloc] Fix bug when there is no register class..

Are you sure your bot touble actually comes from GlobalIsel? While GIsel is unfinished and likely broken on x86 it also shouldn‘t be enabled anywhere except a few dedicated globalisel tests…

Thu, Jun 23, 8:35 AM · Restricted Project, Restricted Project
MatzeB requested changes to D128437: [fastregalloc] Fix bug when there is no register class..

When there is a bug producing invalid IR then we better crash than hide it.

Thu, Jun 23, 8:11 AM · Restricted Project, Restricted Project

Mon, Jun 20

MatzeB accepted D128225: [CMake][SingleSource] Fix compile with profile generation.

LGTM

Mon, Jun 20, 2:49 PM · Restricted Project
MatzeB added a comment to D128225: [CMake][SingleSource] Fix compile with profile generation.

Could you instead exclude the test in the CMakeLists.txt file? Given that this is source from the gcc torture suite that we may want to update again in the future, it's probably best if we can avoid custom changes in the source.

Mon, Jun 20, 2:22 PM · Restricted Project
MatzeB accepted D128175: Corrected some typos found in MachineSink.cpp.

LGTM

Mon, Jun 20, 8:22 AM · Restricted Project, Restricted Project
MatzeB accepted D125079: [fastregalloc] Enhance the heuristics for liveout in self loop..

LGTM

Mon, Jun 20, 8:22 AM · Restricted Project, Restricted Project
MatzeB accepted D128176: Update usage comments in Printable.h. NFC..

LGTM

Mon, Jun 20, 8:20 AM · Restricted Project, Restricted Project

Thu, Jun 16

MatzeB added inline comments to D126754: [MachineSink] Clear kill flags on operands outside loop.
Thu, Jun 16, 2:11 PM · Restricted Project, Restricted Project
MatzeB accepted D126754: [MachineSink] Clear kill flags on operands outside loop.

LGTM

Thu, Jun 16, 2:09 PM · Restricted Project, Restricted Project
MatzeB added a comment to D127103: CodeGen: Split out MachineVerifier's liveness tracking.

I didn't go into any detailed review here, but some top of the head thoughs:

  • We already have LiveIntervals which was supposed to be good enough for everyone. I agree that there is no efficient API for querying every vreg live at a certain program point. But I'd feel more comfortable adding a convenience or caching layer for that on top of LiveIntervals...
  • We already have a 2nd system LiveVariables (which we failed to actually deprecate enough to remove).
  • This is adding a 3rd one, so I think it ought to have some better justification than "simplicity over speed".
Thu, Jun 16, 2:07 PM · Restricted Project, Restricted Project

Wed, Jun 15

MatzeB added a comment to D127848: [CodeGen] Fix the bug of machine sink.

Yes that's what I wrote. It seems there is an exception that if for all operands you get readReg() == false (that includes undef) then it is indeed allowed to have a VReg without a definition.

Wed, Jun 15, 9:03 AM · Restricted Project, Restricted Project
MatzeB added a comment to D127848: [CodeGen] Fix the bug of machine sink.

I see this code in the MachineVerifier:

Wed, Jun 15, 8:57 AM · Restricted Project, Restricted Project
MatzeB added inline comments to D127848: [CodeGen] Fix the bug of machine sink.
Wed, Jun 15, 8:51 AM · Restricted Project, Restricted Project
MatzeB added a comment to D127848: [CodeGen] Fix the bug of machine sink.

I'm also not sure right now if vregs without definitions are even legal MIR while we are still in MachineSSA... Checking that now

Wed, Jun 15, 8:47 AM · Restricted Project, Restricted Project
MatzeB added a comment to D127848: [CodeGen] Fix the bug of machine sink.

Please choose a better title!

Wed, Jun 15, 8:46 AM · Restricted Project, Restricted Project

Fri, Jun 10

MatzeB added a comment to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

Your commit message seems to use the original summary.

Fri, Jun 10, 8:35 AM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB added a comment to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

@JakeEgan I did a quick look for the code in that stacktrace but nothing jumped out on me. It‘s with inlining effects obscuring it. But I think that crash must be investigated on an AIX machine, there‘s no indication that this directly related to my change which works on all other systems.

Fri, Jun 10, 8:33 AM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB added inline comments to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
Fri, Jun 10, 8:29 AM · Restricted Project, Restricted Project, Restricted Project, Restricted Project

Mon, Jun 6

MatzeB accepted D126048: [SplitKit] Handle early clobber + tied to def correctly.

LGTM, thanks

Mon, Jun 6, 6:21 PM · Restricted Project, Restricted Project
MatzeB accepted D127109: llvm-reduce: Don't assert on functions which don't track liveness.

oh looks like we also mark the function parameters as live-in causing a bunch of test changes. Anyway I guess this change better reflects the reality of what registers are in the live-in lists. LGTM

Mon, Jun 6, 4:07 PM · Restricted Project, Restricted Project
MatzeB added a comment to D127109: llvm-reduce: Don't assert on functions which don't track liveness.

The background for this property is that we only really compute the livein lists at the end of register allocation (in the VirtRegRewriter or FastRegAlloc). Before that point using the liveins is pretty much always wrong because they are not complete/correct.

Mon, Jun 6, 12:07 PM · Restricted Project, Restricted Project
MatzeB added inline comments to D126754: [MachineSink] Clear kill flags on operands outside loop.
Mon, Jun 6, 9:46 AM · Restricted Project, Restricted Project

Jun 2 2022

MatzeB added a comment to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

Hi, this caused arm-float-abi-lto.c to fail on AIX. The failure went away for a few builds, then came back. Could you take a look?

https://lab.llvm.org/buildbot/#/builders/214/builds/1625/steps/6/logs/FAIL__Clang__arm-float-abi-lto_c

Jun 2 2022, 2:40 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project

Jun 1 2022

MatzeB committed rGfe0f2c1e9967: Fix compilation in gold-plugin triggered by D125847 (authored by MatzeB).
Fix compilation in gold-plugin triggered by D125847
Jun 1 2022, 6:52 PM · Restricted Project, Restricted Project
MatzeB committed rG850d53a197f9: LTO: Decide upfront whether to use opaque/non-opaque pointer types (authored by MatzeB).
LTO: Decide upfront whether to use opaque/non-opaque pointer types
Jun 1 2022, 6:08 PM · Restricted Project, Restricted Project, Restricted Project
MatzeB closed D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
Jun 1 2022, 6:07 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB retitled D125847: LTO: Add option to initialize with opaque/non-opaque pointer types from LTO: Decide upfront whether to use opaque/non-opaque pointer types to LTO: Add option to initialize with opaque/non-opaque pointer types.
Jun 1 2022, 4:50 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB added a comment to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

address review feedback. I assume this is accepted and good to push when the buildkit builds look reasonable.

(I changed myself to a blocking review to check the component I mostly care about. I promise I'll perform a quick check.)

Jun 1 2022, 4:48 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
Jun 1 2022, 4:47 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB added inline comments to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
Jun 1 2022, 4:45 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB updated the summary of D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
Jun 1 2022, 3:44 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB updated the summary of D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
Jun 1 2022, 3:43 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

address review feedback. I assume this is accepted and good to push when the buildkit builds look reasonable.

Jun 1 2022, 3:41 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB committed rG53753531bc19: TensorFlowCompile: Add object file to list of sources rather than LINK_LIBS (authored by MatzeB).
TensorFlowCompile: Add object file to list of sources rather than LINK_LIBS
Jun 1 2022, 9:05 AM · Restricted Project, Restricted Project
MatzeB closed D126736: TensorFlowCompile: Add object file to list of sources rather than LINK_LIBS.
Jun 1 2022, 9:05 AM · Restricted Project, Restricted Project

May 31 2022

MatzeB updated the summary of D126736: TensorFlowCompile: Add object file to list of sources rather than LINK_LIBS.
May 31 2022, 5:21 PM · Restricted Project, Restricted Project
MatzeB requested review of D126736: TensorFlowCompile: Add object file to list of sources rather than LINK_LIBS.
May 31 2022, 5:20 PM · Restricted Project, Restricted Project
MatzeB added a comment to D125602: [X86][AMX][fastalloc] Allocate tile register based on its shape..

There's a lot going on here.

  • Could you extract the ShouldAllocClass fixes for fastregalloc into a separate diff so we can get discuss them separately and get feedback from AMDGPU folks who introduced this and are the major other user of this AFAIK.
  • I am still wrapping my head around tile registers and tile register configs; specifically I wonder if the support for that really needs to be integrated into the generic register allocation code or whether there is a way to materialize the register configurations in a post-pass. For example do you know how the legacy x86 x87-FPU support works, where we use the register allocator to allocate pseudo FP register fp0-fp7 and then use a post-pass in X86FloatingPoint to insert the necessary stack management operations after the fact. I am not saying it's the same problem, but it is an example of an instance where we managed to have the regalloc allocate to some intermediate pseudo registers and adapt to the complications (in that case register stacks) in a target-specific pass so we wouldn't need to introduce the concept of register stacks to the generic code.
May 31 2022, 9:18 AM · Restricted Project, Restricted Project
MatzeB added inline comments to D126048: [SplitKit] Handle early clobber + tied to def correctly.
May 31 2022, 8:46 AM · Restricted Project, Restricted Project
MatzeB added inline comments to D126048: [SplitKit] Handle early clobber + tied to def correctly.
May 31 2022, 8:44 AM · Restricted Project, Restricted Project

May 26 2022

MatzeB committed rGe267df8ce8d8: Use cmake Python3_EXECUTABLE variable instead of hardcoding (authored by MatzeB).
Use cmake Python3_EXECUTABLE variable instead of hardcoding
May 26 2022, 2:49 PM · Restricted Project, Restricted Project
MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
May 26 2022, 12:04 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB added inline comments to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
May 26 2022, 12:03 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

Address review feedback.

May 26 2022, 12:03 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB added a comment to D124306: [UpdateTestChecks] Auto-generate stub bodies for unused prefixes.

llvm/docs/GettingStarted.rst only asks for python >= 3.6 at the moment.

May 26 2022, 11:51 AM · Restricted Project, Restricted Project
MatzeB added a comment to D124306: [UpdateTestChecks] Auto-generate stub bodies for unused prefixes.

This change is breaking the build for me: BooleanOptionalAction is only available in python 3.9!

May 26 2022, 11:50 AM · Restricted Project, Restricted Project
MatzeB added a comment to D124306: [UpdateTestChecks] Auto-generate stub bodies for unused prefixes.

This change is breaking the build for me: BooleanOptionalAction is only available in python 3.9!

May 26 2022, 11:50 AM · Restricted Project, Restricted Project

May 25 2022

MatzeB added a comment to D126048: [SplitKit] Handle early clobber + tied to def correctly.

This seems to make sense to me (though I am not super familiar with the code here). Some ideas for clearer code below:

May 25 2022, 5:01 PM · Restricted Project, Restricted Project
MatzeB accepted D124631: [TwoAddressInstructionPass] Special processing of STATEPOINT instruction..

Well I don't have too much experience with STATEPOINTs, but the change makes sense to me so far.

May 25 2022, 4:07 PM · Restricted Project, Restricted Project

May 24 2022

MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

Enable LTO by default and fix a whole bunch of LTO tests because of it.

May 24 2022, 6:12 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB accepted D126334: Move GCC-compatible pod-packing change to v15/old behavior available at v14 and below.

LGTM, thanks!

May 24 2022, 3:53 PM · Restricted Project, Restricted Project

May 23 2022

MatzeB added a comment to D123995: [MachineSink] replace MachineLoop with MachineCycle.

LGTM but let's wait for @sameerds for accept.

May 23 2022, 9:19 AM · Restricted Project, Restricted Project

May 19 2022

MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
May 19 2022, 4:01 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
May 19 2022, 4:01 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB added a comment to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

Changed the approach a bit; summary (and tests) are updated.

May 19 2022, 3:38 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB updated subscribers of D126010: Make sure the AsmPrinter doesn't emit any zero-sized symbols to `.debug_aranges`..
May 19 2022, 3:34 PM · Restricted Project, Restricted Project
MatzeB added reviewers for D126010: Make sure the AsmPrinter doesn't emit any zero-sized symbols to `.debug_aranges`.: aprantl, rnk, probinson.
May 19 2022, 3:22 PM · Restricted Project, Restricted Project
MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
May 19 2022, 3:21 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB added a comment to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

While trying to implement things I noticed that -Xclang options are only meant for the LLVM codegeneration part (aka CC1) of clang, they are not meant for the linker/LTO (which would be -Xlinker or -Wl) and consequently there is no easy way in the code to query -Xclang options and pass them on to the linker (nor would it be a good idea I think).

May 19 2022, 11:27 AM · Restricted Project, Restricted Project, Restricted Project, Restricted Project

May 18 2022

MatzeB committed rG8d03c49f498c: Extend switch condition in optimizeSwitchPhiConst when free (authored by MatzeB).
Extend switch condition in optimizeSwitchPhiConst when free
May 18 2022, 4:25 PM · Restricted Project, Restricted Project
MatzeB closed D124897: Extend switch condition in optimizeSwitchPhiConst when free.
May 18 2022, 4:25 PM · Restricted Project, Restricted Project
MatzeB updated the diff for D124897: Extend switch condition in optimizeSwitchPhiConst when free.
May 18 2022, 3:38 PM · Restricted Project, Restricted Project
MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
May 18 2022, 3:17 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
May 18 2022, 2:48 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB accepted D124044: llvm-reduce: Add pass to reduce MIR instruction flags.

As far as FrameSetup/FrameDestroy are concerned, it's a bug if the targets break on them.

May 18 2022, 2:02 PM · Restricted Project, Restricted Project
MatzeB added a comment to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

+1 to adding something to lto::Config.

May 18 2022, 1:50 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

update tests

May 18 2022, 1:49 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB added a comment to D125926: Remove duplicate fields in RAGreedy.

Therefore, if reserved registers information is changed between last time freezeReservedRegs is called and RAGreedy, it's not picked up by RCI.

May 18 2022, 1:40 PM · Restricted Project, Restricted Project
MatzeB accepted D125926: Remove duplicate fields in RAGreedy.

Nice catch! LGTM.

May 18 2022, 1:34 PM · Restricted Project, Restricted Project
MatzeB updated the summary of D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
May 18 2022, 10:14 AM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
May 18 2022, 10:08 AM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB updated the diff for D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.
May 18 2022, 10:07 AM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
MatzeB added a comment to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

Alternatively we could just drop support for LTO with typed pointers :)

May 18 2022, 10:07 AM · Restricted Project, Restricted Project, Restricted Project, Restricted Project

May 17 2022

MatzeB added a comment to D125847: LTO: Add option to initialize with opaque/non-opaque pointer types.

I am not completely sure however how to deal with users specifying -Xclang -no-opaque-pointers explicitly. With this change they would end up using opaque-pointers anyway in LTO modes...

May 17 2022, 8:49 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project