Opened 9 years ago

Closed 9 years ago

#6542 closed enhancement (fixed)

llvm-3.6.1

Reported by: Fernando de Oliveira Owned by: Fernando de Oliveira
Priority: normal Milestone: 7.8
Component: BOOK Version: SVN
Severity: normal Keywords:
Cc:

Description

http://llvm.org/releases/3.6.1/llvm-3.6.1.src.tar.xz

http://llvm.org/releases/3.6.1/llvm-3.6.1.src.tar.xz.sig

http://llvm.org/releases/3.6.1/cfe-3.6.1.src.tar.xz

http://llvm.org/releases/3.6.1/cfe-3.6.1.src.tar.xz.sig

http://llvm.org/releases/3.6.1/compiler-rt-3.6.1.src.tar.xz

http://llvm.org/releases/3.6.1/compiler-rt-3.6.1.src.tar.xz.sig

http://lists.cs.uiuc.edu/pipermail/llvm-announce/2015-May/000059.html

Tom Stellard thomas.stellard at amd.com
Tue May 26 09:54:51 CDT 2015

This release contains bug-fixes for the LLVM 3.6.0 release and is both
API and ABI compatible with 3.6.0.

Fixes in this release include:

- Large number of bug fixes for MIPS target.
- Bug fixes for R600 target including several fixes for Volcanic
   Islands.
- Several bug fixes for core libraries

Change History (3)

comment:1 by Fernando de Oliveira, 9 years ago

Owner: changed from blfs-book@… to Fernando de Oliveira
Status: newassigned

I'm having problem with the tests, and was going to write that they (problems) do not matter for our intended i686 and x86_64 architectures, when discovered that some of the CodeGen/AArch64 and DebugInfo/AArch64 do work in my x86_64 system.

First, make -k check-all:

Testing Time: 120.24s
********************
Failing Tests (35):

...
[same as further below]
...

  Expected Passes    : 19940
  Expected Failures  : 93
  Unsupported Tests  : 203
  Unexpected Failures: 35

Later, reduced just to the two sets with failures, CodeGen/AArch64 and DebugInfo/AArch64:

$ cd llvm-3.6.1.src
$ Release/bin/llvm-lit test/CodeGen/AArch64 test/DebugInfo/AArch64

Testing Time: 2.75s
********************
Failing Tests (35):
    LLVM :: CodeGen/AArch64/aarch64_f16_be.ll
    LLVM :: CodeGen/AArch64/and-mask-removal.ll
    LLVM :: CodeGen/AArch64/arm64-2012-06-06-FPToUI.ll
    LLVM :: CodeGen/AArch64/arm64-abi.ll
    LLVM :: CodeGen/AArch64/arm64-abi_align.ll
    LLVM :: CodeGen/AArch64/arm64-anyregcc-crash.ll
    LLVM :: CodeGen/AArch64/arm64-anyregcc.ll
    LLVM :: CodeGen/AArch64/arm64-big-endian-bitconverts.ll
    LLVM :: CodeGen/AArch64/arm64-big-endian-vector-caller.ll
    LLVM :: CodeGen/AArch64/arm64-elf-globals.ll
    LLVM :: CodeGen/AArch64/arm64-fast-isel-call.ll
    LLVM :: CodeGen/AArch64/arm64-fast-isel-fcmp.ll
    LLVM :: CodeGen/AArch64/arm64-fast-isel-indirectbr.ll
    LLVM :: CodeGen/AArch64/arm64-fast-isel-noconvert.ll
    LLVM :: CodeGen/AArch64/arm64-fastisel-gep-promote-before-add.ll
    LLVM :: CodeGen/AArch64/arm64-fp128.ll
    LLVM :: CodeGen/AArch64/arm64-illegal-float-ops.ll
    LLVM :: CodeGen/AArch64/arm64-neon-mul-div.ll
    LLVM :: CodeGen/AArch64/arm64-neon-vector-list-spill.ll
    LLVM :: CodeGen/AArch64/arm64-platform-reg.ll
    LLVM :: CodeGen/AArch64/arm64-spill.ll
    LLVM :: CodeGen/AArch64/arm64-stackmap.ll
    LLVM :: CodeGen/AArch64/arm64-stacksave.ll
    LLVM :: CodeGen/AArch64/arm64-vcvt_f.ll
    LLVM :: CodeGen/AArch64/arm64-vfloatintrinsics.ll
    LLVM :: CodeGen/AArch64/fast-isel-folding.ll
    LLVM :: CodeGen/AArch64/illegal-float-ops.ll
    LLVM :: CodeGen/AArch64/ldst-opt.ll
    LLVM :: CodeGen/AArch64/machine-copy-prop.ll
    LLVM :: CodeGen/AArch64/neon-fpround_f128.ll
    LLVM :: CodeGen/AArch64/ragreedy-csr.ll
    LLVM :: CodeGen/AArch64/regress-tail-livereg.ll
    LLVM :: CodeGen/AArch64/sincos-expansion.ll
    LLVM :: CodeGen/AArch64/sincospow-vector-expansion.ll
    LLVM :: DebugInfo/AArch64/cfi-eof-prologue.ll

  Expected Passes    : 467
  Expected Failures  : 1
  Unsupported Tests  : 4
  Unexpected Failures: 35

Will wait until tomorrow, for some comment.

In case of no reply to this comment, will just add -k and "... fail for unknown reasons", if there is no objection.

comment:2 by Douglas R. Reno, 9 years ago

I would say go ahead with the change, since these tests do not apply to our architecture. I would be more worried if it were something related to x86_64 or i686, as we are not generating any code for the ARM 64-bit architecture anywhere within the book (that I am aware of).

comment:3 by Fernando de Oliveira, 9 years ago

Resolution: fixed
Status: assignedclosed

Fixed at r16045.

Note: See TracTickets for help on using tickets.