GitLab

I think someone accidentally a word in a Note in TcGenDeriv.

Parallel GC does not scan large allocation area (-A)
effectively as it does not do work stealing from nursery
by default.

That leads to large imbalance when only one of threads
overflows allocation area: most of GC threads finish
quickly (as there is not much to collect) and sit idle
waiting while single GC thread finishes scan of single
allocation area for that thread.

The patch enables work stealing for (equivalent of -qb0)
allocation area of -A32M or higher.

Tested on a highlighting-kate package from Trac #9221

On 8-core machine the difference is around 5% faster
of wall-clock time. On 24-core VM the speedup is 20%.
Signed-off-by: Sergei Trofimovich <siarheit@google.com>

Test Plan: measured wall time and GC parallelism on highlighting-kate build

Reviewers: austin, bgamari, erikd, simonmar

Reviewed By: bgamari, simonmar

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2483

GHC Trac Issues: #9221

In Trac #9221 one of problems using high --jobs=<N>
is amount of mutator (or GC) threads we crate.

We use userspace spinning-and-yielding (see ACQUIRE_SPIN_LOCK)
to acess work stealing queues. In case of
N-worker-threads > N-CPUs fraction of time when
thread holding spin lock gets descheduled by kernel
increases. That causes other threads to waste CPU time
before giving up CPU.
Signed-off-by: Sergei Trofimovich <siarheit@google.com>

Test Plan:
ghc --make -j8 and -j80 have comparable sys time
on a 8-core system.

Reviewers: austin, gintas, bgamari, simonmar

Reviewed By: bgamari, simonmar

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2482

GHC Trac Issues: #9221

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

Summary:
Back when derived `Generic` instances used to generate auxiliary datatypes,
they would also generate instances for those datatypes. Nowadays, GHC generics
uses a `DataKinds`-based encoding that requires neither auxiliary datatypes
(corresponding to the `DerivTyCon` constructor of `DerivStuff`) nor instances
for them (the `DerivInst` constructor of `DerivStuff`). It appears that
`DerivTyCon` constructor was removed at some point, but `DerivInst` never was.

No `DerivInst` values are ever constructed, so we can safely remove it.

Test Plan: It builds

Reviewers: austin, hvr, bgamari

Reviewed By: bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2481

Summary: Fixes #12530.

Test Plan: make test TEST=12530

Reviewers: austin, bgamari, hvr, goldfire

Reviewed By: goldfire

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2472

GHC Trac Issues: #12530

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

This fixes Trac #12531:

   class Foo x where
     foo :: forall a . x a -> x a
     default foo :: forall b . x b -> x b
     foo x = go
       where go :: x b
             go = undefined

We want 'b' to scope over the code for 'foo', but we were
using 'a' instead.

The import of UniqFM in HscTypes was only needed in stage2

This one-line change fixes Trac #12531.  Hooray.

Simple, non-invasive; can merge to 8.0.2

harendra authored Aug 25, 2016 and Ben Gamari committed Aug 25, 2016

Remove the doc targets completely not partially when disabled.
Otherwise even though we are able to build the targets but we will see
some strange errors.

Reviewers: austin, bgamari

Reviewed By: bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2470

indicating that it is bottom. This should help making the "empty cases"
lint error give less false alarms.

as proposed in #12435, as it easily gets in the way of development when
it reports false positives.

Any variable with useful information (strict or used-once) will not be
included in lazy_fv (according to splitFVs). If we now also remove them
from the strictness signatures, their uses are not recorded anywhere –
and then probably considered absent.

this fixes #12368.

It also refactors dmdFix a bit, removes some redundancies (such as
passing around an strictness signature right next to an id, when that id
is guaranteed to have been annotated with that strictness signature).

Note that when fixed-point iteration does not terminate, we
conservatively delete their strictness signatures (set them to nopSig).
But this loses the information on how its strict free variables are
used!

Lazily used variables already escape via lazy_fvs. We ensure that in the
case of an aborted fixed-point iteration, also the strict variables are
put there (with a conservative demand of topDmd).

Differential Revision: https://phabricator.haskell.org/D2392

Ryan Scott authored Aug 23, 2016 and Ben Gamari committed Aug 23, 2016

This adds new constructors `UnboxedSumE`, `UnboxedSumT`, and
`UnboxedSumP` to represent unboxed sums in Template Haskell.

One thing you can't currently do is, e.g., `reify ''(#||#)`, since I
don't believe unboxed sum type/data constructors can be written in
prefix form.  I will look at fixing that as part of #12514.

Fixes #12478.

Test Plan: make test TEST=T12478_{1,2,3}

Reviewers: osa1, goldfire, austin, bgamari

Reviewed By: goldfire, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2448

GHC Trac Issues: #12478

mniip authored Aug 23, 2016 and Ben Gamari committed Aug 23, 2016

Replaced error-prone index manipulation on a pointer array with
a simple fold on the array elements.

Test Plan: Added a test case that triggers the bug

Reviewers: hvr, austin, bgamari

Reviewed By: bgamari

Subscribers: simonpj, thomie

Differential Revision: https://phabricator.haskell.org/D2439

GHC Trac Issues: #12458

  #12520 was already fixed in HEAD. Adding a test to make sure it stays fixed.

The changes in #12403 and #12513 subtly changed the behavior of Template
Haskell reification and splicing. While the old behavior was certainly buggy,
it's possible that there's code in the wild that depended on the old behavior
to work. To err on the side of caution, I'll postpone these changes to GHC
8.2.1 instead of having them merged into GHC 8.0.2.

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

Summary:
Previously, TH would implicitly remove the parentheses when splicing in
singleton unboxed tuple types (e.g., turning `(# Int #)` into `Int`). Luckily,
the fix is simply to delete some code.

Fixes #12513.

Test Plan: make test TEST=T12513

Reviewers: hvr, bgamari, austin, goldfire

Reviewed By: goldfire

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2462

GHC Trac Issues: #12513

Summary:
This patch fixes issue with abort in GHCi on OpenBSD current
as of Aug 12 2016. The OpenBSD is more and more strict about usage
of writable and executable memory. Programs/applications which
requires such functionality need to be linked with -z wxneeded linker
flag and need to be run from the file-system mounted with wxallowed
mount option. If either of those options in not met, then problematic
program/application usually fail on some mmap/mprotect call which fail.

Reviewers: bgamari, austin, hvr

Subscribers: thomie, erikd

Differential Revision: https://phabricator.haskell.org/D2454

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

Test Plan: validate

Reviewers: simonpj, austin, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2444

GHC Trac Issues: #12472

Summary:
Kept this seperate from the previous patch for clarity.
Comes with a test.
Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: validate

Reviewers: simonpj, austin, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2220

GHC Trac Issues: #12035

Summary:
This makes ghc --make's retypecheck behavior more in line
with ghc -c, which is able to tie the knot as we are typechecking.
Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: validate

Reviewers: simonpj, austin, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2213

GHC Trac Issues: #12035

Summary:
This also drops the parked fix from
efa7b3a4


(though I didn't revert the refactoring).
Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: validate

Reviewers: simonpj, austin, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2211

GHC Trac Issues: #10083

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

Test Plan: validate

Reviewers: simonpj, austin, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2246

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

Test Plan: validate

Reviewers: simonpj, austin, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2209

Summary:
initIfaceTc was originally used to make sure when we typecheck
an interface, it can find the TyThings for things it itself
defined.  However, in the case of retypecheckLoop, this wasn't
necessary because we ALREADY tied the knot through the HPT.

This commit removes initIfaceTc, instead relying on the HPT
to tie the knot.  genModDetails' caller needed to be modified
to tie the knot, but there are not that many call-sites of
typecheckIface so the change is quite reasonable.

We also introduce a new 'initIfaceLoad', which does
NOT set up 'if_rec_types'.  It's used when we're
typechecking old, up-to-date interfaces in, since we're
never going to update the type environment.

The full details are in Note [Knot-tying typecheckIface].
Displeasingly, we need a special case to handle DFuns in
the case of tcHiBootIface, see
Note [DFun knot-tying special case] for the gory details.

I also added another test which tickles a bug in a buggy
version of this patch (see "Why the seq?")
Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

Test Plan: validate

Reviewers: simonpj, austin, bgamari

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2349

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

This reverts commit e3e2e49a.

I'm reverting because it makes ghc-stage2 seg-fault on
64-bit Windows machines.  Even ghc-stage2 --version seg-faults.

Summary:
For symmetry with `gcastWith` in `Data.Type.Equality`.

Fixes #12493.

Test Plan: It compiles

Reviewers: bgamari, goldfire, hvr, austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D2456

GHC Trac Issues: #12493

This reverts commit 773e3aad.

Simpler code, and simpler to understand.
No change in behaviour.

defaultCallStacks generates evidence bindings for call stacks,
but wasn't setting the binding site correctly.  As a result
they were simply discarded in the case of pattern synonyms,
giving rise to Trac #12489.

The fix is easy; and I added an ASSERT to catch the error earlier.

This fixes Trac #12406.  A not-in-scope error shoudl be an error
even if you have -fdefer-typed-holes.