summary |
shortlog | log |
commit |
commitdiff |
tree
first ⋅ prev ⋅ next
Raspbian Automatic git importer [Thu, 22 Dec 2016 06:09:51 +0000 (06:09 +0000)]
Merge ghc (7.6.3-21) import into refs/heads/workingbranch
Debian Haskell Group [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
ppc64el
Gbp-Pq: Name ppc64el.patch
Debian Haskell Group [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
arm64
Gbp-Pq: Name arm64.patch
Debian Haskell Group [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
bit-big-endian
commit
a4b1a43542b11d09dd3b603d82c5a0e99da67d74
Author: Austin Seipp <austin@well-typed.com>
Date: Fri Nov 1 22:17:01 2013 -0500
Fix loop on 64bit Big-Endian platforms (#8134)
This is a fun one.
In the RTS, `cas` expects a pointer to StgWord which will translate to
unsigned long (8 bytes under LP64.) But we had previously declared
token_locked as *StgBool* - which evaluates to 'int' (4 bytes under
LP64.) That means we fail to provide enough storage for the cas
primitive, causing it to corrupt memory on a 64bit platform.
Hilariously, this somehow did not affect little-endian platforms (ARM,
x86, etc) before. That's because to clear our lock token, we would say:
token_locked = 0;
But because token_locked is 32bits technically, this only writes to
half of the 64bit quantity. On a Big-Endian machine, this won't do
anything. That is, token_locked starts as 0:
/ token_locked
|
v
0x00000000
and the first cas modifies the memory to:
/ valid / corrupted
| |
v v
0x00000000 0x00000001
We then clear token_locked, but this doesn't change the corrupted 4
bytes of memory. And then we try to lock the token again, spinning until
it is released - clearly a deadlock.
Related: Windows (amd64) doesn't follow LP64, but LLP64, where both
int and long are 4 bytes, so this shouldn't change anything on these
platforms.
Thanks to Reid Barton for helping the diagnosis. Also, thanks to Jens
Peterson who confirmed this also fixes building GHC on Fedora/ppc64 and
Fedora/s390x.
Authored-by: Gustavo Luiz Duarte <gustavold@linux.vnet.ibm.com>
Signed-off-by: Austin Seipp <austin@well-typed.com>
Gbp-Pq: Name 64-bit-big-endian
Geoffrey Mainland [Wed, 12 Jun 2013 12:31:49 +0000 (13:31 +0100)]
Avoid generating empty llvm.used definitions.
Avoid generating empty llvm.used definitions.
LLVM 3.3rc3 complains when the llvm.used global is an empty array, so don't
define llvm.used at all when it would be empty.
Gbp-Pq: Name llvm-3.3-compat
Colin Watson [Mon, 8 Apr 2013 23:02:27 +0000 (00:02 +0100)]
Handle sign bit when generating veneer for ARM Thumb branch relocation
Gbp-Pq: Name Handle-sign-bit-when-generating-veneer-for-ARM-Thumb.patch
Debian Haskell Group [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
no-missing-haddock-file-warning
Gbp-Pq: Name no-missing-haddock-file-warning
Debian Haskell Group [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
ARM-VFPv3D16
Gbp-Pq: Name ARM-VFPv3D16
Debian Haskell Group [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
use-debian-gen_contents_index
===================================================================
Gbp-Pq: Name use-debian-gen_contents_index
Debian Haskell Group [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
haddock-hardcode-ghc-paths
===================================================================
Gbp-Pq: Name haddock-hardcode-ghc-paths
Debian Haskell Group [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
system-libffi
This patch could be replaced by a configure call if
http://hackage.haskell.org/trac/ghc/ticket/5743 were fixed.
Gbp-Pq: Name system-libffi
Debian Haskell Group [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
cabal-bug-1087
Preprocessed files would be written to dist/build instead of
dist/build/my-test-suite/my-test-suite-tmp, causing them not to be
found during compilation.
Fixes #1087.
Author: Johan Tibell <johan.tibell@gmail.com>
Origin: https://github.com/haskell/cabal/commit/
dd691fa6791d67981388ec044b28343879d8c2b1
Gbp-Pq: Name cabal-bug-1087.patch
Debian Haskell Group [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
ghc-ignore-ANN-annotation
Gbp-Pq: Name ghc-ignore-ANN-annotation.patch
Joachim Breitner [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
ghc (7.6.3-21) unstable; urgency=medium
* Fix watch file.
* Let ghc-haddock break on ghc-doc (<< 7.6.3-20~). If I used piuparts
correctly, this finally closes: #781649.
[dgit import unpatched ghc 7.6.3-21]
Joachim Breitner [Wed, 1 Apr 2015 11:35:10 +0000 (11:35 +0000)]
Import ghc_7.6.3-21.debian.tar.xz
[dgit import tarball ghc 7.6.3-21 ghc_7.6.3-21.debian.tar.xz]
Joachim Breitner [Sun, 21 Apr 2013 17:12:24 +0000 (17:12 +0000)]
Import ghc_7.6.3.orig.tar.bz2
[dgit import orig ghc_7.6.3.orig.tar.bz2]
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
no-useless-timer
commit
a73d0a21c875ff539c2240e431eaf87173a92bd3
Author: Ian Lynagh <igloo@earth.li>
Date: Sun Jan 15 19:21:38 2012 +0000
Fix a #define
I don't think it was causing any problems, but
TimeToUS(x+y)
would have evaluated to
x + (y / 1000)
MERGED from commit
54121fffb5ee069d9b7a5628104ff1114ea87182
On the 7.4 branch, this was causing a wrong conversion in
getDelayTarget.
Signed-off-by: Paolo Capriotti <p.capriotti@gmail.com>
Gbp-Pq: Name no-useless-timer
Erik de Castro Lopo [Sat, 16 Jun 2012 15:48:05 +0000 (10:48 -0500)]
PPC: Handle right shift of > 31 bits. Fix #5900.
Gbp-Pq: Name fix-PPC-right-shift-bug
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
memcpy-ffi
commit
40c1106c338e209f07023d165f32bff0f75e2e54
Author: Paolo Capriotti <p.capriotti@gmail.com>
Date: Wed May 2 15:24:46 2012 +0100
Description:
Cast memory primops in the C backend (#5976)
To prevent conflicts with GCC builtins, generate identical code for
calls to mem primos and FFI calls.
Based on a patch by Joachim Breitner.
Bug: http://hackage.haskell.org/trac/ghc/ticket/5967
Origin: http://hackage.haskell.org/trac/ghc/changeset/
40c1106c338e209f07023d165f32bff0f75e2e54
Gbp-Pq: Name memcpy-ffi.patch
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
no-missing-haddock-file-warning
Gbp-Pq: Name no-missing-haddock-file-warning
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
hurd-is-ELF
Gbp-Pq: Name hurd-is-ELF
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
ARM-VFPv3D16
Gbp-Pq: Name ARM-VFPv3D16
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
armhf_llvm_abi
Gbp-Pq: Name armhf_llvm_abi
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
use-llvm-3.0
===================================================================
Gbp-Pq: Name use-llvm-3.0
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
no_ghci_on_powerpc
Gbp-Pq: Name no_ghci_on_powerpc
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
haddock-no-library
===================================================================
Gbp-Pq: Name haddock-no-library
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
use-debian-gen_contents_index
===================================================================
Gbp-Pq: Name use-debian-gen_contents_index
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
haddock-hardcode-ghc-paths
===================================================================
Gbp-Pq: Name haddock-hardcode-ghc-paths
Debian Haskell Group [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
system-libffi
This patch could be replaced by a configure call if
http://hackage.haskell.org/trac/ghc/ticket/5743 were fixed.
Gbp-Pq: Name system-libffi
Karel Gardas [Tue, 14 Feb 2012 07:03:07 +0000 (08:03 +0100)]
fix ARM StgCRun to not save and restore r11/fp register twice
Gbp-Pq: Name fix-ARM-StgCRun-to-not-save-and-restore-r11-fp-regis
Karel Gardas [Tue, 14 Feb 2012 07:01:47 +0000 (08:01 +0100)]
fix ARM's StgCRun clobbered register list for both ARM and Thumb modes
Gbp-Pq: Name fix-ARM-s-StgCRun-clobbered-register-list-for-both-A
Joachim Breitner [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
ghc (7.4.1-4) unstable; urgency=low
[ Erik de Castro Lopo ]
* Add debian/patches/fix-PPC-right-shift-bug which fixes upstream GHC bug:
http://hackage.haskell.org/trac/ghc/ticket/6156 (Closes: #677591)
* Refresh other patches.
* debian/contol: Add myself to uploaders and set DM-Upload-Allowed to yes.
[ Joachim Breitner ]
* debian/patches/no-useless-timer: Backported from GHC 7.4.2, (Closes:
#677096)
* Make sure GHC is using ld.bfd, as it passes arguments not understood by
the gold linker (Closes: #673081)
* Use a saner priority for the runhaskell alternative. (Closes: #676970).
TODO for later: Consider removing the alternatives code altogether.
[dgit import unpatched ghc 7.4.1-4]
Joachim Breitner [Sat, 16 Jun 2012 10:48:18 +0000 (10:48 +0000)]
Import ghc_7.4.1-4.debian.tar.gz
[dgit import tarball ghc 7.4.1-4 ghc_7.4.1-4.debian.tar.gz]
Joachim Breitner [Fri, 3 Feb 2012 15:27:31 +0000 (15:27 +0000)]
Import ghc_7.4.1.orig.tar.bz2
[dgit import orig ghc_7.4.1.orig.tar.bz2]