This reverts f48f5a9e

The prefixing does not work as comma
is stripped by $(addprefix) macro:

The following call
    $$(addprefix -optl-Wl, $$($1_$2_$3_ALL_LD_OPTS))

prefixes options with "-optl-Wl" not with "-optl-Wl,"

The simplest breakage can be seen by adding
    SRC_LD_OPTS += -O1
to mk/build.mk:

  <no location info>: error:
    Warning: Couldn't figure out linker information!
             Make sure you're using GNU ld, GNU gold
             or the built in OS X linker, etc.
  gcc: error: unrecognized command line option '-Wl-O1'

Another problem with original change is loss of ability
to pass options to gcc as a linker driver, for example:
    SRC_LD_OPTS += -flto

Signed-off-by: Sergei Trofimovich <siarheit@google.com>
(cherry picked from commit a6657bd0)

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

Currently uniques are 32-bits wide. 8 of these bits are for the unique
class, leaving only 24 for the unique number itself. This seems
dangerously small for a large project. Let's use the full range of the
native machine word.

We also add (now largely unnecessary) overflow check to ensure that the
unique number doesn't overflow.

Test Plan: Validate

Reviewers: simonmar, austin, niteria

Reviewed By: niteria

Subscribers: mpickering, thomie

Differential Revision: https://phabricator.haskell.org/D2844

GHC Trac Issues: #12944

(cherry picked from commit 0d213c18)

This reverts commit 6a5d13c4 due to
breakage on PPC AIX while compiling unordered-containers. See #13033.

There is two types of options passed directly to 'ld'
(and not to 'gcc' driver):
    - CONF_LD_LINKER_OPTS_STAGE$4
    - EXTRA_LD_OPTS

This changedoes two things:

- split 'EXTRA_LD_OPTS' into two variables:
    - EXTRA_LD_OPTS (accepts 'gcc' wrapper options)
    - EXTRA_LD_LINKER_OPTS (accepts raw 'ld' options)

- wraps all LD_LINKER options as '-Wl,' when passed
  to 'gcc' driver.

Fixes https://phabricator.haskell.org/D2776



Signed-off-by: Sergei Trofimovich <siarheit@google.com>
(cherry picked from commit 87c3b1d4)

This reverts commit 9d9eaeca due
to #12993.

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

GHC uses gcc, not ld, for linking. Consequently all flags to be
interpreted by ld need to be prefixed by -optl,-Wl on the GHC command
line.

Test Plan: Validate on OpenBSD

Reviewers: austin, rwbarton

Reviewed By: rwbarton

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2776

(cherry picked from commit f48f5a9e)

(cherry picked from commit 54706738cd452717c7ec1a91927c3941c8037c7b)

Before the change result of expression
    ArchSupportsSMP="$(if $(filter $(ARM_ISA),ARMv5 ARMv6),NO,YES)"
to evaluate to
    ArchSupportsSMP="YES"

After the change it's
    ArchSupportsSMP=YES

Thanks to orion for the fix!

Fixes Trac #12981
Signed-off-by: Sergei Trofimovich <siarheit@google.com>

(cherry picked from commit 52c5e553)

Summary:
We mentioned that there were "some programs" that failed to typecheck
due to #12784, but given how surprisingly common this issue has been, it'd
be prudent to at least give one example of the bug in the release notes.

Reviewers: simonpj, bgamari, austin, rwbarton

Reviewed By: rwbarton

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2786

GHC Trac Issues: #12784

(cherry picked from commit eec02ab7)

The test associated with this has given us too much trouble. It's not
worth the pain for a minor release.

This reverts commit fb0f4cf6.

Alexander Vershilov authored 8 years ago and Ben Gamari committed 8 years ago

When rts is forked it doesn't update toplevel handler, so UserInterrupt
exception is sent to Thread1 that doesn't exist in forked process.

We install toplevel handler when fork so signal will be delivered to the
new main thread.

Fixes #12903

Reviewers: simonmar, austin, erikd, bgamari

Reviewed By: bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2770

GHC Trac Issues: #12903

(cherry picked from commit 895a131f)

Simon Marlow authored 8 years ago and Ben Gamari committed 8 years ago

select() is limited to 1024 file descriptors.  This actually blew up
in a very hard-to-debug way in our production system when using the
hinotify package.

Test Plan:
libraries/tests pass, paricularly hGetBuf001 which exercises this
code.

Reviewers: niteria, erikd, austin, hvr, bgamari

Reviewed By: bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2785

GHC Trac Issues: #12912

(cherry picked from commit f46369b8)

Test Plan: Read it, commit it, merge it, ship it

Reviewers: goldfire, bgamari, austin, hvr, simonpj

Reviewed By: simonpj

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2774

GHC Trac Issues: #12907

(cherry picked from commit a452c6e5)

(cherry picked from commit be8a47f5)

Alan Zimmerman authored 8 years ago and Ben Gamari committed 8 years ago

(cherry picked from commit 36e36227)

When we had
  f :: ( _ ) => blah
we were failing to call growThetaTyVars, as we do in the
no-type-signature case, and that meant that we weren't generalising
over the right type variables.  I'm quite surprised this didn't cause
problems earlier.

Anyway Trac #12844 showed it up and this patch fixes it

(cherry picked from commit 1bfff60f)

Summary: Resolves #12881.

Test Plan: Read it, commit it, merge it, ship it

Reviewers: hvr, simonpj, austin, bgamari

Reviewed By: simonpj

Subscribers: simonpj, thomie

Differential Revision: https://phabricator.haskell.org/D2760

GHC Trac Issues: #12881

(cherry picked from commit abd4a4c1)

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

Strangely my previous attempts at resolving this all seemed to end in
perplexing segmentation faults in the GHC testsuite (including some
rather recent attempts). Somehow this attempt miraculously works.

However, there was one wrinkle that I still need to work out fully: we
need to consider Lits as non-trivial in cpeArg. Failure to do this means
that we would transform something like,

    $trModule = TrModule "HelloWorld"#

into

    $trModule = case "HelloWorld"# of x { __DEFAULT -> TrModule x }

Which then fails the consistentStgInfo check in CoreToStg for reasons
that I am still trying to work out.

Mark T12757 as fixed

Reviewers: simonmar, simonpj, austin

Reviewed By: simonpj

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2666

GHC Trac Issues: #11158

(cherry picked from commit 967dd5c9)

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

Test Plan: Validate, expected to fail

Reviewers: austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2665

GHC Trac Issues: #12757

(cherry picked from commit b5460dd6)

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

Test Plan: `validate --slow`

Reviewers: austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2536

GHC Trac Issues: #12019

(cherry picked from commit d1b4fec1)

This is a stopgap fix for GHC 8.0 bug #12485: in particular,
it relaxes need for -package-db flags to be given in
dependency order.  The trade-off is that we are a lot more
unsafe when there are packages with duplicate 'id's in
the database stack: the new code will not do an ABI compatibility
check: if two packages have the same 'id', they are assumed to
be ABI compatible.  If this is not true, GHC may build
segfaulting executables.

Missing test updates, but I'm putting it up so people can
take a look.

In GHC 8.2, we'll record ABIs for all dependencies, allowing
GHC to make better decisions about shadowing.

Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: validate

Reviewers: austin, niteria, bgamari, simonmar

Reviewed By: simonmar

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2613

GHC Trac Issues: #12485

Facundo Domínguez authored 8 years ago and Ben Gamari committed 8 years ago

It turned out that finalizers were run too early and information
resulting from simplifying constraints was not available.

This patch runs finalizers after a first call to simplifyTop, and
then calls simplifyTop a second time to deal with constraints
that could result from running the finalizers.

Fixes T12777

Test Plan: ./validate

Reviewers: goldfire, simonpj, bgamari, austin

Reviewed By: simonpj

Subscribers: mpickering, mboes, thomie

Differential Revision: https://phabricator.haskell.org/D2659

GHC Trac Issues: #12777

(cherry picked from commit 231a3ae1)

This patch makes the RTS linker skip 64-bit symbol table entries.  See
https://mail.haskell.org/pipermail/ghc-devs/2016-November/013210.html

Test Plan: validate

Reviewers: austin, erikd, simonmar, bgamari

Reviewed By: bgamari

Subscribers: osa1, thomie

Differential Revision: https://phabricator.haskell.org/D2697

GHC Trac Issues: #12827

(cherry picked from commit 1b336d90)

Peter Trommler authored 8 years ago and Ben Gamari committed 8 years ago

On architectures with weak memory consistency a write barrier
is needed before the write to the pointer array.

Fixes #12469

Test Plan: rebuilt Stackage nightly twice on powerpc64le

Reviewers: hvr, rrnewton, erikd, austin, simonmar, bgamari

Reviewed By: erikd, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2525

GHC Trac Issues: #12469

(cherry picked from commit 2cb8cc26)

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

It seems like GCC versions prior to 4.8 exit with code 0 when faced with
an unrecognized flag. Silly compilers.

Test Plan: Validate

Reviewers: hvr, austin, ggreif

Reviewed By: ggreif

Subscribers: thomie, erikd

Differential Revision: https://phabricator.haskell.org/D2707

GHC Trac Issues: #12759

(cherry picked from commit 011af2bf)

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

As reported in #12812, the runtime system fails to build when linked
with gold due to a missing dependency on libpthread.

Additionally, rts/package.conf.in uses the WORD_SIZE_IN_BITS macro
defined by MachDeps.h, which it does not #include. Fix this.

Test Plan: Validate with gold linker

Reviewers: hsyl20, austin, erikd, simonmar

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2695

GHC Trac Issues: #12816

(cherry picked from commit 2cfbee89)

Ben Gamari authored 8 years ago and Ben Gamari committed 8 years ago

Otherwise we end up looking in the wrong place for dynamic libraries on
Windows. This addresses a regression introduced by D2611. See #12479.

Test Plan: validate across platforms

Reviewers: austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2640

GHC Trac Issues: #12479

(cherry picked from commit 75e69edb498511cf2b28dec4e14e6d11232f3e32)

Instead of pulling a token and looking for `'('` or `')'`,
just look for the character itself. This prevents us from
lexing every single item twice, once to see if it's a
left parenthesis and once to actually parse it.

Partially fixes #12665

Test Plan: Validate

Reviewers: austin, bgamari, hvr

Reviewed By: bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2623

GHC Trac Issues: #12665