llvm-toolchain-4.0.git
8 years agoMerge version 1:4.0.1-1+rpi1 and 1:4.0.1-8 to produce 1:4.0.1-8+rpi1 archive/raspbian/1%4.0.1-8+rpi1 raspbian/1%4.0.1-8+rpi1
Raspbian forward pporter [Thu, 26 Oct 2017 17:30:05 +0000 (18:30 +0100)]
Merge version 1:4.0.1-1+rpi1 and 1:4.0.1-8 to produce 1:4.0.1-8+rpi1

8 years agoMerge llvm-toolchain-4.0 (1:4.0.1-8) import into refs/heads/workingbranch
Sylvestre Ledru [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
Merge llvm-toolchain-4.0 (1:4.0.1-8) import into refs/heads/workingbranch

8 years agorL306379
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
rL306379

commit 7df06519765b14e1b08d7034c82c45a0a653eb25
Author: Chandler Carruth <chandlerc@gmail.com>
Date:   Tue Jun 27 08:32:03 2017 +0000

    [SROA] Fix PR32902 by more carefully propagating !nonnull metadata.

    This is based heavily on the work done ni D34285. I mostly wanted to do
    test cleanup for the author to save them some time, but I had a really
    hard time understanding why it was so hard to write better test cases
    for these issues.

    The problem is that because SROA does a second rewrite of the loads and
    because we *don't* propagate !nonnull for non-pointer loads, we first
    introduced invalid !nonnull metadata and then stripped it back off just
    in time to avoid most ways of this PR manifesting. Moving to the more
    careful utility only fixes this by changing the predicate to look at the
    new load's type rather than the target type. However, that *does* fix
    the bug, and the utility is much nicer including adding range metadata
    to model the nonnull property after a conversion to an integer.

    However, we have bigger problems because we don't actually propagate
    *range* metadata, and the utility to do this extracted from instcombine
    isn't really in good shape to do this currently. It *only* handles the
    case of copying range metadata from an integer load to a pointer load.
    It doesn't even handle the trivial cases of propagating from one integer
    load to another when they are the same width! This utility will need to
    be beefed up prior to using in this location to get the metadata to
    fully survive.

    And even then, we need to go and teach things to turn the range metadata
    into an assume the way we do with nonnull so that when we *promote* an
    integer we don't lose the information.

    All of this will require a new test case that looks kind-of like
    `preserve-nonnull.ll` does here but focuses on range metadata. It will
    also likely require more testing because it needs to correctly handle
    changes to the integer width, especially as SROA actively tries to
    change the integer width!

    Last but not least, I'm a little worried about hooking the range
    metadata up here because the instcombine logic for converting from
    a range metadata *to* a nonnull metadata node seems broken in the face
    of non-zero address spaces where null is not mapped to the integer `0`.
    So that probably needs to get fixed with test cases both in SROA and in
    instcombine to cover it.

    But this *does* extract the core PR fix from D34285 of preventing the
    !nonnull metadata from being propagated in a broken state just long
    enough to feed into promotion and crash value tracking.

    On D34285 there is some discussion of zero-extend handling because it
    isn't necessary. First, the new load size covers all of the non-undef
    (ie, possibly initialized) bits. This may even extend past the original
    alloca if loading those bits could produce valid data. The only way its
    valid for us to zero-extend an integer load in SROA is if the original
    code had a zero extend or those bits were undef. And we get to assume
    things like undef *never* satifies nonnull, so non undef bits can
    participate here. No need to special case the zero-extend handling, it
    just falls out correctly.

    The original credit goes to Ariel Ben-Yehuda! I'm mostly landing this to
    save a few rounds of trivial edits fixing style issues and test case
    formulation.

    Differental Revision: D34285

    git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@306379 91177308-0d34-0410-b5e6-96231b3b80d8

Gbp-Pq: Name rL306379.diff

8 years agorL306358
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
rL306358

commit 156cc49e505986a1659adaa3a0b5a070372377c8
Author: Chandler Carruth <chandlerc@gmail.com>
Date:   Tue Jun 27 03:08:45 2017 +0000

    [SROA] Further test cleanup and add a test for the actual propagation of
    the nonnull attribute distinct from rewriting it into an assume.

    git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@306358 91177308-0d34-0410-b5e6-96231b3b80d8

Gbp-Pq: Name rL306358.diff

8 years agorL306353
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
rL306353

commit 477bd758b48cb96477d3dd4cf3b36bf2706e8c1d
Author: Chandler Carruth <chandlerc@gmail.com>
Date:   Tue Jun 27 02:23:15 2017 +0000

    [SROA] Clean up a test case a bit prior to adding more testing for
    nonnull as part of fixing PR32902.

    git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@306353 91177308-0d34-0410-b5e6-96231b3b80d8

Gbp-Pq: Name rL306353.diff

8 years agorL306267
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
rL306267

commit 5a057dc8edbb63887f8c611dd8ddf1b76997f07c
Author: Chandler Carruth <chandlerc@gmail.com>
Date:   Mon Jun 26 03:31:31 2017 +0000

    [InstCombine] Factor the logic for propagating !nonnull and !range
    metadata out of InstCombine and into helpers.

    NFC, this just exposes the logic used by InstCombine when propagating
    metadata from one load instruction to another. The plan is to use this
    in SROA to address PR32902.

    If anyone has better ideas about how to factor this or name variables,
    I'm all ears, but this seemed like a pretty good start and lets us make
    progress on the PR.

    This is based on a patch by Ariel Ben-Yehuda (D34285).

    git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@306267 91177308-0d34-0410-b5e6-96231b3b80d8

Gbp-Pq: Name rL306267.diff

8 years agorL298540
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
rL298540

commit 2b66aeed20b61d42cd2782b94d5419c6faca49a1
Author: Luqman Aden <me@luqman.ca>
Date:   Wed Mar 22 19:16:39 2017 +0000

    Preserve nonnull metadata on Loads through SROA & mem2reg.

    Summary:
    https://llvm.org/bugs/show_bug.cgi?id=31142 :

    SROA was dropping the nonnull metadata on loads from allocas that got optimized out. This patch simply preserves nonnull metadata on loads through SROA and mem2reg.

    Reviewers: chandlerc, efriedma

    Reviewed By: efriedma

    Subscribers: hfinkel, spatel, efriedma, arielb1, davide, llvm-commits

    Differential Revision: https://reviews.llvm.org/D27114

    git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@298540 91177308-0d34-0410-b5e6-96231b3b80d8

Gbp-Pq: Name rL298540.diff

8 years agorL305193-backport
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
rL305193-backport

commit 2b622a393ce80c6157d32a50bf67d6b830729469
Author: Than McIntosh <thanm@google.com>
Date:   Mon Jun 12 14:56:02 2017 +0000

    StackColoring: smarter check for slot overlap

    Summary:
    The old check for slot overlap treated 2 slots `S` and `T` as
    overlapping if there existed a CFG node in which both of the slots could
    possibly be active. That is overly conservative and caused stack blowups
    in Rust programs. Instead, check whether there is a single CFG node in
    which both of the slots are possibly active *together*.

    Fixes PR32488.

    Patch by Ariel Ben-Yehuda <ariel.byd@gmail.com>

    Reviewers: thanm, nagisa, llvm-commits, efriedma, rnk

    Reviewed By: thanm

    Subscribers: dotdash

    Differential Revision: https://reviews.llvm.org/D31583

    git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@305193 91177308-0d34-0410-b5e6-96231b3b80d8

Gbp-Pq: Name rL305193-backport.diff

8 years agoarmhf-bitfield
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
armhf-bitfield

===================================================================

Gbp-Pq: Name armhf-bitfield.diff

8 years agopr81066
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
pr81066

# DP; Fix PR sanitizer/81066, taken from the gcc-7-branch.

libsanitizer/

2017-07-17  Jakub Jelinek  <jakub@redhat.com>

Backported from mainline
2017-07-14  Jakub Jelinek  <jakub@redhat.com>

PR sanitizer/81066
* sanitizer_common/sanitizer_linux.h: Cherry-pick upstream r307969.
* sanitizer_common/sanitizer_linux.cc: Likewise.
* sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cc: Likewise.
* tsan/tsan_platform_linux.cc: Likewise.

Gbp-Pq: Name pr81066.diff

8 years agoftfbs-gcc
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
ftfbs-gcc

===================================================================

Gbp-Pq: Name ftfbs-gcc.diff

8 years agoadd_symbols_versioning
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
add_symbols_versioning

This patch adds a very simple linker script to version the lib's symbols
and thus trying to avoid crashes if an application loads two different
LLVM versions (as long as they do not share data between them).

Note that we deliberately *don't* make LLVM_3.9 depend on LLVM_3.8:
they're incompatible and the whole point of this patch is
to tell the linker that.

Author: Rebecca N. Palmer <rebecca_palmer@zoho.com>
Author: Lisandro Damían Nicanor Pérez Meyer <lisandro@debian.org>
Bug-Debian: https://bugs.debian.org/848368

Gbp-Pq: Name add_symbols_versioning.patch

8 years agopthread-link
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
pthread-link

1;4601;0c===================================================================

Gbp-Pq: Name pthread-link.diff

8 years agosilent-amdgpu-test-failing
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
silent-amdgpu-test-failing

===================================================================

Gbp-Pq: Name silent-amdgpu-test-failing.diff

8 years agosilent-test-failing-codeverage
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
silent-test-failing-codeverage

===================================================================

Gbp-Pq: Name silent-test-failing-codeverage.diff

8 years agolldb-missing-install
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
lldb-missing-install

===================================================================

Gbp-Pq: Name lldb-missing-install.diff

8 years agolldb-addversion-suffix-to-llvm-server-exec
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
lldb-addversion-suffix-to-llvm-server-exec

lldb-server exec users always /usr/bin/lldb-server. Server is required
for any debugging with lldb which makes it unusable unless default version
package has been installed. Small changes to code and debian/rules allows
a workaround for lldb-server start up.

To use this one needs to add cmake defination during configure. eg
-DDEBIAN_VERSION_SUFFIX=-$(LLVM_VERSION)

Better implementation would be to use /usr/share/llvm-$(VERSION)/bin but
that change seems to require a big change to the path handling code
which could then break something else.

This probably should have upstream bug but I couldn't find any existing report.

Gbp-Pq: Name lldb-addversion-suffix-to-llvm-server-exec.patch

8 years agoclang-fix-cmpxchg8-detection-on-i386
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
clang-fix-cmpxchg8-detection-on-i386

libcxx atomic tests for old i386 fail with wrong Atomic inline width.
cmpxchg8b instruction is required for 8 byte atomics that clang was
assuming.

Too bad _GCC_ATOMIC_LLONG_LOCK_FREE 2 isn't supported even with this change
because llvm doesn't support unaligned atomic compare and exchange operation.
Fallback calls to libatomic.so should handle long long lock free but clang
can't tell program if libatomic is always lock free.

Related bug: https://llvm.org/bugs/show_bug.cgi?id=19355

Gbp-Pq: Name clang-fix-cmpxchg8-detection-on-i386.patch

8 years agofix-scan-view-path
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
fix-scan-view-path

===================================================================

Gbp-Pq: Name fix-scan-view-path.diff

8 years agoclang-tidy-run-bin
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
clang-tidy-run-bin

===================================================================

Gbp-Pq: Name clang-tidy-run-bin.diff

8 years agodisable-llvm-symbolizer-test
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
disable-llvm-symbolizer-test

Gbp-Pq: Name disable-llvm-symbolizer-test.diff

8 years agosilent-gold-utils
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
silent-gold-utils

Gbp-Pq: Name silent-gold-utils.diff

8 years agodisable-source-interleave
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
disable-source-interleave

===================================================================

Gbp-Pq: Name disable-source-interleave.diff

8 years agolldb-link-atomic-cmake
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
lldb-link-atomic-cmake

Gbp-Pq: Name lldb-link-atomic-cmake.patch

8 years agofix-cmake-config-prefix
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
fix-cmake-config-prefix

## Description: add some description
## Origin/Author: add some origin or author
## Bug: bug URL

Gbp-Pq: Name fix-cmake-config-prefix.diff

8 years agomips-fpxx-enable
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
mips-fpxx-enable

Gbp-Pq: Name mips-fpxx-enable.diff

8 years agoscan-view-fix-path
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
scan-view-fix-path

Gbp-Pq: Name scan-view-fix-path.diff

8 years agoinstall-scan-build-py
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
install-scan-build-py

Gbp-Pq: Name install-scan-build-py.diff

8 years agoclang-analyzer-force-version
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
clang-analyzer-force-version

# Force the version of clang in the analyzer
# This was causing the static analyzer to fail silently if the clang & clang++ are
# not installed

Gbp-Pq: Name clang-analyzer-force-version.diff

8 years agosilent-MCJIIT-tests
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
silent-MCJIIT-tests

Gbp-Pq: Name silent-MCJIIT-tests.diff

8 years agohurd-EIEIO-undef
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
hurd-EIEIO-undef

===================================================================

Gbp-Pq: Name hurd-EIEIO-undef.diff

8 years agolldb-libname
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
lldb-libname

Gbp-Pq: Name lldb-libname.diff

8 years agolldb-soname
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
lldb-soname

Gbp-Pq: Name lldb-soname.diff

8 years agosoname
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
soname

Gbp-Pq: Name 0044-soname.diff

8 years agofix-llvm-config-obj-src-root
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
fix-llvm-config-obj-src-root

Gbp-Pq: Name fix-llvm-config-obj-src-root.patch

8 years agodisable-display-PASS-UNSUPPORTED-XFAIL
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
disable-display-PASS-UNSUPPORTED-XFAIL

===================================================================

Gbp-Pq: Name disable-display-PASS-UNSUPPORTED-XFAIL.diff

8 years agosilent-more-tests
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
silent-more-tests

# Comment the tests for the code coverage (fails otherwise)

Gbp-Pq: Name silent-more-tests.diff

8 years agodo-not-fail-on-unexpected-pass
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
do-not-fail-on-unexpected-pass

===================================================================

Gbp-Pq: Name do-not-fail-on-unexpected-pass.diff

8 years agofix-clang-path-and-build
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
fix-clang-path-and-build

Gbp-Pq: Name fix-clang-path-and-build.diff

8 years agoremoveduplicatedeclaration
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
removeduplicatedeclaration

Gbp-Pq: Name removeduplicatedeclaration.diff

8 years agopython-clangpath
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
python-clangpath

Gbp-Pq: Name python-clangpath.diff

8 years agoatomic_library_2
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
atomic_library_2

Gbp-Pq: Name atomic_library_2.diff

8 years agoatomic_library_1
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
atomic_library_1

Gbp-Pq: Name atomic_library_1.diff

8 years agosilent-gold-test
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
silent-gold-test

Gbp-Pq: Name silent-gold-test.diff

8 years agouse-deb-json
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
use-deb-json

Gbp-Pq: Name use-deb-json.diff

8 years agohurd-pathmax
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
hurd-pathmax

===================================================================

Gbp-Pq: Name hurd-pathmax.diff

8 years agounwind-chain-inclusion
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
unwind-chain-inclusion

# Without this patch, the first local include of unwind.h might, with the
# __has_include_next, try to include the one from the system.
# It might be /usr/include/clang/3.4/include/unwind.h
# Because of the #ifndef __CLANG_UNWIND_H, it might never include any declaration
# from the system.

Gbp-Pq: Name unwind-chain-inclusion.diff

8 years agoclang-format-version
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
clang-format-version

Gbp-Pq: Name clang-format-version.diff

8 years agodeclare_clear_cache
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
declare_clear_cache

Gbp-Pq: Name declare_clear_cache.diff

8 years agoscan-build-clang-path
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
scan-build-clang-path

Gbp-Pq: Name scan-build-clang-path.diff

8 years agoDebian-version-info-and-bugreport
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
Debian-version-info-and-bugreport

Gbp-Pq: Name 0003-Debian-version-info-and-bugreport.patch

8 years agoset-correct-float-abi
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
set-correct-float-abi

Gbp-Pq: Name 26-set-correct-float-abi.diff

8 years agofix_clang_stdint
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
fix_clang_stdint

===================================================================

Gbp-Pq: Name 27-fix_clang_stdint.diff

8 years agostrlcpy_strlcat_warning_removed
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
strlcpy_strlcat_warning_removed

Gbp-Pq: Name 23-strlcpy_strlcat_warning_removed.diff

8 years agoclang_debian_version
LLVM Packaging Team [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
clang_debian_version

===================================================================

Gbp-Pq: Name 19-clang_debian_version.patch

8 years agollvm-toolchain-4.0 (1:4.0.1-8) unstable; urgency=medium
Sylvestre Ledru [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
llvm-toolchain-4.0 (1:4.0.1-8) unstable; urgency=medium

  [ Ximin Luo ]
  * Backport some patches (originally from rust, and upstreamed) to fix two
    failing tests in rustc.

  [ Sylvestre Ledru ]
  * Try to fix the mipsel FTBFS (Closes: #877567)
    I am trying the first option from the bug:
    - gsplit-dward on 32 bits archs
    - -g everywhere
    Many thanks to Adrian Bunk for that

[dgit import unpatched llvm-toolchain-4.0 1:4.0.1-8]

8 years agoImport llvm-toolchain-4.0_4.0.1-8.debian.tar.xz
Sylvestre Ledru [Wed, 18 Oct 2017 14:37:25 +0000 (15:37 +0100)]
Import llvm-toolchain-4.0_4.0.1-8.debian.tar.xz

[dgit import tarball llvm-toolchain-4.0 1:4.0.1-8 llvm-toolchain-4.0_4.0.1-8.debian.tar.xz]

8 years agoClean up and rename dgit generated patch. archive/raspbian/1%4.0.1-1+rpi1 raspbian/1%4.0.1-1+rpi1
Peter Michael Green [Thu, 6 Jul 2017 21:38:50 +0000 (21:38 +0000)]
Clean up and rename dgit generated patch.

8 years agoCommit Debian 3.0 (quilt) metadata
Peter Michael Green [Thu, 6 Jul 2017 21:36:15 +0000 (21:36 +0000)]
Commit Debian 3.0 (quilt) metadata

[dgit (3.10) quilt-fixup]

8 years agoRemove old cmake-arm-version-hack.patch. dgit will regenerate it.
Peter Michael Green [Thu, 6 Jul 2017 21:25:21 +0000 (21:25 +0000)]
Remove old cmake-arm-version-hack.patch. dgit will regenerate it.

8 years agoMerge raspbian changes from llvm-toolchain-3.9 packaging.
Peter Michael Green [Thu, 6 Jul 2017 20:49:03 +0000 (21:49 +0100)]
Merge raspbian changes from llvm-toolchain-3.9 packaging.

8 years agoPsudeomerge to tie in llvm-toolchain-3.9 upstream history.
root [Thu, 6 Jul 2017 20:39:58 +0000 (21:39 +0100)]
Psudeomerge to tie in llvm-toolchain-3.9 upstream history.

8 years agoMerge version 1:3.9.1-9+rpi1 and 1:3.9.1-10 to produce 1:3.9.1-10+rpi1
Raspbian forward pporter [Wed, 28 Jun 2017 23:45:49 +0000 (00:45 +0100)]
Merge version 1:3.9.1-9+rpi1 and 1:3.9.1-10 to produce 1:3.9.1-10+rpi1

8 years agoMerge version 1:3.9.1-7+rpi1 and 1:3.9.1-9 to produce 1:3.9.1-9+rpi1
Raspbian forward pporter [Tue, 27 Jun 2017 19:16:05 +0000 (20:16 +0100)]
Merge version 1:3.9.1-7+rpi1 and 1:3.9.1-9 to produce 1:3.9.1-9+rpi1

8 years agoMerge llvm-toolchain-3.9 (1:3.9.1-10) import into refs/heads/workingbranch
Sylvestre Ledru [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
Merge llvm-toolchain-3.9 (1:3.9.1-10) import into refs/heads/workingbranch

8 years agoFix computeKnownBits for ARMISD::CMOV
Pirama Arumuga Nainar [Thu, 23 Mar 2017 16:47:47 +0000 (16:47 +0000)]
Fix computeKnownBits for ARMISD::CMOV

Summary:
The true and false operands for the CMOV are operands 0 and 1.
ARMISelLowering.cpp::computeKnownBits was looking at operands 1 and 2
instead.  This can cause CMOV instructions to be incorrectly folded into
BFI if value set by the CMOV is another CMOV, whose known bits are
computed incorrectly.

This patch fixes the issue and adds a test case.

Reviewers: kristof.beyls, jmolloy

Subscribers: llvm-commits, aemerson, srhines, rengolin

Differential Revision: https://reviews.llvm.org/D31265

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@298624 91177308-0d34-0410-b5e6-96231b3b80d8

Gbp-Pq: Name rust-0032-Fix-computeKnownBits-for-ARMISD-CMOV.patch

8 years agofix-R_AARCH64_MOVW_UABS_G3-relocation
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
fix-R_AARCH64_MOVW_UABS_G3-relocation

Gbp-Pq: Name fix-R_AARCH64_MOVW_UABS_G3-relocation.patch

8 years agoadd_symbols_versioning
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
add_symbols_versioning

This patch adds a very simple linker script to version the lib's symbols
and thus trying to avoid crashes if an application loads two different
LLVM versions (as long as they do not share data between them).

Note that we deliberately *don't* make LLVM_3.9 depend on LLVM_3.8:
they're incompatible and the whole point of this patch is
to tell the linker that.

Author: Rebecca N. Palmer <rebecca_palmer@zoho.com>
Author: Lisandro Damían Nicanor Pérez Meyer <lisandro@debian.org>
Bug-Debian: https://bugs.debian.org/848368

Gbp-Pq: Name add_symbols_versioning.patch

8 years agoallow-opencl-pointer-to-bool
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
allow-opencl-pointer-to-bool

Used by e.g. Blender on mesa-opencl-icd

Author: Anastasia Stulova
Origin: upstream https://reviews.llvm.org/rL294313
Bug: https://bugs.llvm.org/show_bug.cgi?id=30217
Bug-Debian: https://bugs.debian.org/857623

Gbp-Pq: Name 857623-allow-opencl-pointer-to-bool.diff

8 years agoFix ESan test failure on Debian Sid bot
Qin Zhao [Fri, 7 Oct 2016 20:53:35 +0000 (20:53 +0000)]
Fix ESan test failure on Debian Sid bot

Summary: Increase early allocation buffer size.

Reviewers: bruening

Subscribers: kubabrecka

Differential Revision: https://reviews.llvm.org/D25380

git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@283598 91177308-0d34-0410-b5e6-96231b3b80d8

Gbp-Pq: Name esan-Fix-ESan-test-failure-on-Debian-Sid-bot2.diff

8 years agoFix ESan test failure on Debian Sid bot
Qin Zhao [Mon, 3 Oct 2016 20:03:10 +0000 (20:03 +0000)]
Fix ESan test failure on Debian Sid bot

Summary:
Handles early allocation from dlsym by allocating memory from a local
static buffer.

Reviewers: bruening

Subscribers: kubabrecka

Differential Revision: https://reviews.llvm.org/D25193

git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@283139 91177308-0d34-0410-b5e6-96231b3b80d8

Gbp-Pq: Name esan-Fix-ESan-test-failure-on-Debian-Sid-bot.diff

8 years agoamdgpu-regression
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
amdgpu-regression

===================================================================

Gbp-Pq: Name amdgpu-regression.diff

8 years agolldb-server-link
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
lldb-server-link

===================================================================

Gbp-Pq: Name lldb-server-link.diff

8 years agolldb-addversion-suffix-to-llvm-server-exec
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
lldb-addversion-suffix-to-llvm-server-exec

===================================================================

Gbp-Pq: Name lldb-addversion-suffix-to-llvm-server-exec.patch

8 years agoclang-arm-default-vfp3-on-armv7a
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
clang-arm-default-vfp3-on-armv7a

===================================================================

Gbp-Pq: Name clang-arm-default-vfp3-on-armv7a.patch

8 years agolit-force-lang
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
lit-force-lang

===================================================================

Gbp-Pq: Name lit-force-lang.diff

8 years agoFix reallocation logic when IsInDlsymAllocPool(ptr) is true.
Maxim Ostapenko [Mon, 26 Sep 2016 08:11:21 +0000 (08:11 +0000)]
Fix reallocation logic when IsInDlsymAllocPool(ptr) is true.

git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@282389 91177308-0d34-0410-b5e6-96231b3b80d8

Gbp-Pq: Name upstream-asan-msan-fix-reallocation-logic.diff

8 years agoPrevent initialization failure with newer (2.23+) glibc in use.
Maxim Ostapenko [Fri, 23 Sep 2016 07:40:55 +0000 (07:40 +0000)]
Prevent initialization failure with newer (2.23+) glibc in use.

This patch is pretty the same as http://reviews.llvm.org/D20235 that we used
for ASan. Using the same hack for MSan fixes its initialization with newer
Glibc in use.

Differential Revision: https://reviews.llvm.org/D24736

git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@282232 91177308-0d34-0410-b5e6-96231b3b80d8

Gbp-Pq: Name upstream-msan-prevent-initialization-failure.diff

8 years agoclang-fix-cmpxchg8-detection-on-i386
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
clang-fix-cmpxchg8-detection-on-i386

libcxx atomic tests for old i386 fail with wrong Atomic inline width.
cmpxchg8b instruction is required for 8 byte atomics that clang was
assuming.

Too bad _GCC_ATOMIC_LLONG_LOCK_FREE 2 isn't supported even with this change
because llvm doesn't support unaligned atomic compare and exchange operation.
Fallback calls to libatomic.so should handle long long lock free but clang
can't tell program if libatomic is always lock free.

Related bug: https://llvm.org/bugs/show_bug.cgi?id=19355

Gbp-Pq: Name clang-fix-cmpxchg8-detection-on-i386.patch

8 years agofix-scan-view-path
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
fix-scan-view-path

===================================================================

Gbp-Pq: Name fix-scan-view-path.diff

8 years agodrop-wrong-hack-arm64
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
drop-wrong-hack-arm64

Author: Gianfranco Costamagna <locutusofborg@debian.org>

Gbp-Pq: Name drop-wrong-hack-arm64.patch

8 years agoclang-tidy-run-bin
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
clang-tidy-run-bin

===================================================================

Gbp-Pq: Name clang-tidy-run-bin.diff

8 years agokfreebsd-support
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
kfreebsd-support

===================================================================

Gbp-Pq: Name kfreebsd-support.diff

8 years agodisable-llvm-symbolizer-test
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
disable-llvm-symbolizer-test

===================================================================

Gbp-Pq: Name disable-llvm-symbolizer-test.diff

8 years agolldb-link-atomic-cmake
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
lldb-link-atomic-cmake

Gbp-Pq: Name lldb-link-atomic-cmake.patch

8 years agofix-cmake-config-prefix
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
fix-cmake-config-prefix

## Description: add some description
## Origin/Author: add some origin or author
## Bug: bug URL

Gbp-Pq: Name fix-cmake-config-prefix.diff

8 years agodisable-test-failing-sparc-i386
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
disable-test-failing-sparc-i386

===================================================================

Gbp-Pq: Name disable-test-failing-sparc-i386.diff

8 years agomips-fpxx-enable
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
mips-fpxx-enable

===================================================================

Gbp-Pq: Name mips-fpxx-enable.diff

8 years agoscan-view-fix-path
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
scan-view-fix-path

Gbp-Pq: Name scan-view-fix-path.diff

8 years agoinstall-scan-build-py
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
install-scan-build-py

===================================================================

Gbp-Pq: Name install-scan-build-py.diff

8 years agoclang-analyzer-force-version
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
clang-analyzer-force-version

# Force the version of clang in the analyzer
# This was causing the static analyzer to fail silently if the clang & clang++ are
# not installed

Gbp-Pq: Name clang-analyzer-force-version.diff

8 years agosilent-MCJIIT-tests
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
silent-MCJIIT-tests

===================================================================

Gbp-Pq: Name silent-MCJIIT-tests.diff

8 years agohurd-EIEIO-undef
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
hurd-EIEIO-undef

===================================================================

Gbp-Pq: Name hurd-EIEIO-undef.diff

8 years agolldb-libname
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
lldb-libname

Gbp-Pq: Name lldb-libname.diff

8 years agolldb-soname
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
lldb-soname

Gbp-Pq: Name lldb-soname.diff

8 years agosoname
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
soname

===================================================================

Gbp-Pq: Name 0044-soname.diff

8 years agocompiler-rt-path
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
compiler-rt-path

===================================================================

Gbp-Pq: Name compiler-rt-path.diff

8 years agofix-llvm-config-obj-src-root
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
fix-llvm-config-obj-src-root

===================================================================

Gbp-Pq: Name fix-llvm-config-obj-src-root.patch

8 years agodisable-display-PASS-UNSUPPORTED-XFAIL
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
disable-display-PASS-UNSUPPORTED-XFAIL

===================================================================

Gbp-Pq: Name disable-display-PASS-UNSUPPORTED-XFAIL.diff

8 years agosilent-more-tests
LLVM Packaging Team [Sun, 18 Jun 2017 17:12:15 +0000 (18:12 +0100)]
silent-more-tests

# Comment the tests for the code coverage (fails otherwise)

Gbp-Pq: Name silent-more-tests.diff