Page MenuHomePhabricator

hvdijk (Harald van Dijk)
User

Projects

User does not belong to any projects.

User Details

User Since
Aug 13 2019, 4:58 AM (68 w, 16 h)

Recent Activity

Yesterday

hvdijk requested review of D92346: [X86] Add TLS_(base_)addrX32 for X32 mode.
Mon, Nov 30, 1:25 PM · Restricted Project
hvdijk added a comment to D91338: [X86] Zero-extend pointers to i64 for x86_64.

I'm OK with this patch. Can you please check if expensive check is happy with it?

Mon, Nov 30, 10:52 AM · Restricted Project
hvdijk committed rGcdac34bd47a3: [X86] Zero-extend pointers to i64 for x86_64 (authored by hvdijk).
[X86] Zero-extend pointers to i64 for x86_64
Mon, Nov 30, 10:51 AM
hvdijk closed D91338: [X86] Zero-extend pointers to i64 for x86_64.
Mon, Nov 30, 10:51 AM · Restricted Project

Sun, Nov 29

hvdijk updated the diff for D16474: Use PC-relative address for x32 TLS address.

Update to current LLVM.

Sun, Nov 29, 2:12 PM · Restricted Project
hvdijk commandeered D16474: Use PC-relative address for x32 TLS address.
Sun, Nov 29, 2:11 PM · Restricted Project
hvdijk added a comment to D16474: Use PC-relative address for x32 TLS address.

@hvdijk Are you in a position to commandeer this phab revision and complete @hjl.tools's work?

Sun, Nov 29, 1:06 PM · Restricted Project
hvdijk added a comment to D16474: Use PC-relative address for x32 TLS address.

LLVM ERROR: Cannot select: 0x29507a0: ch,glue = X86ISD::TLSADDR 0x28fc140, TargetGlobalTLSAddress:i32<i32* @x> 0 [TF=7]

Sun, Nov 29, 6:04 AM · Restricted Project
hvdijk added a comment to D91338: [X86] Zero-extend pointers to i64 for x86_64.

@pengfei With the confirmation that we do need this, could you take another look? Does this look okay now?

Sun, Nov 29, 6:00 AM · Restricted Project
hvdijk committed rGfba0b65f7271: [libc++] hash<long double>: adjust for x86-64 ILP32 (authored by hvdijk).
[libc++] hash<long double>: adjust for x86-64 ILP32
Sun, Nov 29, 5:53 AM
hvdijk closed D91349: [libc++] hash<long double>: adjust for x86-64 ILP32.
Sun, Nov 29, 5:52 AM · Restricted Project

Sat, Nov 28

hvdijk committed rG78a30c830b53: [X86] Add -verify-machineinstrs to pic.ll (authored by hvdijk).
[X86] Add -verify-machineinstrs to pic.ll
Sat, Nov 28, 9:55 AM
hvdijk added a comment to D91933: [X86] Do not allow FixupSetCC to relax constraints.

If possible, please commit the pic.ll fix as a separate followup.

Sat, Nov 28, 9:49 AM · Restricted Project
hvdijk committed rG47e2fafbf3d9: [X86] Do not allow FixupSetCC to relax constraints (authored by hvdijk).
[X86] Do not allow FixupSetCC to relax constraints
Sat, Nov 28, 9:47 AM
hvdijk closed D91933: [X86] Do not allow FixupSetCC to relax constraints.
Sat, Nov 28, 9:47 AM · Restricted Project
hvdijk committed rG47c902ba8479: [X86] Have indirect calls take 64-bit operands in 64-bit modes (authored by hvdijk).
[X86] Have indirect calls take 64-bit operands in 64-bit modes
Sat, Nov 28, 8:49 AM
hvdijk closed D91924: [X86] Have indirect calls take 64-bit operands in 64-bit modes.
Sat, Nov 28, 8:49 AM · Restricted Project

Fri, Nov 27

hvdijk updated the diff for D91924: [X86] Have indirect calls take 64-bit operands in 64-bit modes.

Changed auto to Register

Fri, Nov 27, 11:12 AM · Restricted Project

Thu, Nov 26

hvdijk added a comment to D52050: [Driver] Fix architecture triplets and search paths for Linux x32.

Yeah, @hvdijk has made multiple other improvements which should finally allow the backend to be usable.

We still disagree on the search paths for libraries and headers though if I remember correctly.

Thu, Nov 26, 3:22 PM · Restricted Project
hvdijk added a comment to D91924: [X86] Have indirect calls take 64-bit operands in 64-bit modes.

We already have fast-isel-call tests - could we add gnux32 test coverage there?

Thu, Nov 26, 2:58 PM · Restricted Project
hvdijk added inline comments to D91933: [X86] Do not allow FixupSetCC to relax constraints.
Thu, Nov 26, 2:21 PM · Restricted Project
hvdijk updated the diff for D91933: [X86] Do not allow FixupSetCC to relax constraints.

Add a comment about skipping the optimization when constraining fails.

Thu, Nov 26, 2:21 PM · Restricted Project
hvdijk added a comment to D91339: [X86] Include %rip for 32-bit RIP-relative relocs for x32.

What is the status of the expensive check failure? It has been present for five days now.

Thu, Nov 26, 1:24 AM · Restricted Project

Sun, Nov 22

hvdijk added inline comments to D91339: [X86] Include %rip for 32-bit RIP-relative relocs for x32.
Sun, Nov 22, 5:50 PM · Restricted Project
hvdijk updated the diff for D91933: [X86] Do not allow FixupSetCC to relax constraints.

Fix FileCheck syntax at the same time, using CHECK-X32-DAG rather than CHECK-DAG-X32 that @pengfei noticed in D91339 after it had been committed already.

Sun, Nov 22, 5:49 PM · Restricted Project
hvdijk added inline comments to D91339: [X86] Include %rip for 32-bit RIP-relative relocs for x32.
Sun, Nov 22, 5:03 PM · Restricted Project
hvdijk added a comment to D91339: [X86] Include %rip for 32-bit RIP-relative relocs for x32.

Thanks - please can you add -verify-machineinstrs to the pic.ll command lines as part of either of those patches?

Sun, Nov 22, 12:23 PM · Restricted Project
hvdijk updated the diff for D91933: [X86] Do not allow FixupSetCC to relax constraints.

As requested by @RKSimon, added -verify-machineinstrs to test/CodeGen/X86/pic.ll so that the problem is visible even when expensive checks are disabled.

Sun, Nov 22, 12:23 PM · Restricted Project
hvdijk added a comment to D91339: [X86] Include %rip for 32-bit RIP-relative relocs for x32.

I have now also opened https://reviews.llvm.org/D91933 hoping to fix the other pre-existing issue. With both https://reviews.llvm.org/D91924 and that, the test in here passes even when expensive checks are enabled.

Sun, Nov 22, 8:34 AM · Restricted Project
hvdijk requested review of D91933: [X86] Do not allow FixupSetCC to relax constraints.
Sun, Nov 22, 8:33 AM · Restricted Project
hvdijk added a comment to D91339: [X86] Include %rip for 32-bit RIP-relative relocs for x32.

@hvdijk On EXPENSIVE_CHECKS builds I'm seeing a build failure in pic.ll - please can you take a look?

Sun, Nov 22, 5:10 AM · Restricted Project
hvdijk added inline comments to D91924: [X86] Have indirect calls take 64-bit operands in 64-bit modes.
Sun, Nov 22, 1:58 AM · Restricted Project
hvdijk updated the diff for D91924: [X86] Have indirect calls take 64-bit operands in 64-bit modes.

Add MOV32rr to guarantee high bits will be zero.

Sun, Nov 22, 1:58 AM · Restricted Project

Sat, Nov 21

hvdijk updated the diff for D91924: [X86] Have indirect calls take 64-bit operands in 64-bit modes.
Sat, Nov 21, 5:02 PM · Restricted Project
hvdijk added inline comments to D91924: [X86] Have indirect calls take 64-bit operands in 64-bit modes.
Sat, Nov 21, 2:55 PM · Restricted Project
hvdijk updated the diff for D91924: [X86] Have indirect calls take 64-bit operands in 64-bit modes.

Forgot to run clang-format-diff before uploading, updated now.

Sat, Nov 21, 2:02 PM · Restricted Project
hvdijk requested review of D91924: [X86] Have indirect calls take 64-bit operands in 64-bit modes.
Sat, Nov 21, 1:58 PM · Restricted Project
hvdijk added a comment to D91913: Suppress non-conforming GNU paste extension in all standard-conforming modes.

Then the comment needs to be fixed too i would think?

Sat, Nov 21, 9:41 AM · Restricted Project
hvdijk requested review of D91913: Suppress non-conforming GNU paste extension in all standard-conforming modes.
Sat, Nov 21, 8:00 AM · Restricted Project
hvdijk added a comment to D91339: [X86] Include %rip for 32-bit RIP-relative relocs for x32.

LGTM - @hvdijk do you have commit access yet?

Sat, Nov 21, 7:17 AM · Restricted Project

Fri, Nov 20

hvdijk added a comment to D91349: [libc++] hash<long double>: adjust for x86-64 ILP32.

That looks correct to me with what I just read about x32 (I wasn't familiar with it). I can't imagine this is the only place that needs fixing for libc++ to work in x32?

Fri, Nov 20, 1:38 PM · Restricted Project

Sun, Nov 15

hvdijk added a comment to D91338: [X86] Zero-extend pointers to i64 for x86_64.

I never finished x32 work. You can use my branch as a reference when working on x32.

Sun, Nov 15, 12:13 AM · Restricted Project

Sat, Nov 14

hvdijk added a comment to D91338: [X86] Zero-extend pointers to i64 for x86_64.

Was there ever a follow-up on https://lkml.org/lkml/2018/12/10/1145 ?

Sat, Nov 14, 1:14 AM · Restricted Project
hvdijk added a comment to D91338: [X86] Zero-extend pointers to i64 for x86_64.

Hi Harald, thanks for thoroughly answering my questions. I still have doubts intention of this patch. Is there any bug related to it?

Sat, Nov 14, 12:49 AM · Restricted Project

Fri, Nov 13

hvdijk added inline comments to D91338: [X86] Zero-extend pointers to i64 for x86_64.
Fri, Nov 13, 7:57 AM · Restricted Project
hvdijk updated the diff for D91338: [X86] Zero-extend pointers to i64 for x86_64.

Remove no longer used [[REG:.*]] from test

Fri, Nov 13, 7:56 AM · Restricted Project
hvdijk retitled D91349: [libc++] hash<long double>: adjust for x86-64 ILP32 from [libc++] hash<long double>: adjust for ILP32 to [libc++] hash<long double>: adjust for x86-64 ILP32.
Fri, Nov 13, 4:22 AM · Restricted Project
hvdijk added inline comments to D91338: [X86] Zero-extend pointers to i64 for x86_64.
Fri, Nov 13, 4:12 AM · Restricted Project
hvdijk updated the diff for D91338: [X86] Zero-extend pointers to i64 for x86_64.

Remove no longer necessary FIXME comment

Fri, Nov 13, 4:11 AM · Restricted Project
hvdijk added inline comments to D91338: [X86] Zero-extend pointers to i64 for x86_64.
Fri, Nov 13, 1:20 AM · Restricted Project
hvdijk added a comment to D91339: [X86] Include %rip for 32-bit RIP-relative relocs for x32.

This is correct. Probably worth mentioning in the subject that this is for x32.

Fri, Nov 13, 12:39 AM · Restricted Project
hvdijk retitled D91339: [X86] Include %rip for 32-bit RIP-relative relocs for x32 from [X86] Include %rip for 32-bit PIC-relative relocs to [X86] Include %rip for 32-bit RIP-relative relocs for x32.
Fri, Nov 13, 12:32 AM · Restricted Project

Thu, Nov 12

hvdijk requested review of D91349: [libc++] hash<long double>: adjust for x86-64 ILP32.
Thu, Nov 12, 6:39 AM · Restricted Project
hvdijk requested review of D91339: [X86] Include %rip for 32-bit RIP-relative relocs for x32.
Thu, Nov 12, 5:17 AM · Restricted Project
hvdijk requested review of D91338: [X86] Zero-extend pointers to i64 for x86_64.
Thu, Nov 12, 5:15 AM · Restricted Project

Aug 27 2020

hvdijk requested review of D86695: [RFC] Call ParseLangArgs for all inputs.
Aug 27 2020, 4:09 AM · Restricted Project

Sep 25 2019

hvdijk added a comment to D66138: Let findCommutedOpIndices take const MachineInstr&.

Thanks! I do not have commit access, can somebody commit it for me?

Sep 25 2019, 3:56 AM · Restricted Project

Sep 24 2019

hvdijk added a comment to D66138: Let findCommutedOpIndices take const MachineInstr&.

ping

Sep 24 2019, 3:21 PM · Restricted Project

Aug 13 2019

hvdijk created D66138: Let findCommutedOpIndices take const MachineInstr&.
Aug 13 2019, 5:12 AM · Restricted Project