[Git][ghc/ghc][wip/int64-everywhere] 159 commits: hadrian: add flag to skip rebuilding dependency information #17636

Sylvain Henry gitlab at gitlab.haskell.org
Mon Jul 20 14:20:39 UTC 2020



Sylvain Henry pushed to branch wip/int64-everywhere at Glasgow Haskell Compiler / GHC


Commits:
3033e0e4 by Adam Sandberg Ericsson at 2020-07-08T20:36:49-04:00
hadrian: add flag to skip rebuilding dependency information #17636

- - - - -
b7de4b96 by Stefan Schulze Frielinghaus at 2020-07-09T09:49:22-04:00
Fix GHCi :print on big-endian platforms

On big-endian platforms executing

  import GHC.Exts
  data Foo = Foo Float# deriving Show
  foo = Foo 42.0#
  foo
  :print foo

results in an arithmetic overflow exception which is caused by function
index where moveBytes equals
  word_size - (r + item_size_b) * 8
Here we have a mixture of units. Both, word_size and item_size_b have
unit bytes whereas r has unit bits.  On 64-bit platforms moveBytes
equals then
  8 - (0 + 4) * 8
which results in a negative and therefore invalid second parameter for a
shiftL operation.

In order to make things more clear the expression
  (word .&. (mask `shiftL` moveBytes)) `shiftR` moveBytes
is equivalent to
  (word `shiftR` moveBytes) .&. mask
On big-endian platforms the shift must be a left shift instead of a
right shift. For symmetry reasons not a mask is used but two shifts in
order to zero out bits. Thus the fixed version equals
  case endian of
    BigEndian    -> (word `shiftL` moveBits) `shiftR` zeroOutBits `shiftL` zeroOutBits
    LittleEndian -> (word `shiftR` moveBits) `shiftL` zeroOutBits `shiftR` zeroOutBits

Fixes #16548 and #14455

- - - - -
3656dff8 by Sylvain Henry at 2020-07-09T09:50:01-04:00
LLVM: fix MO_S_Mul2 support (#18434)

The value indicating if the carry is useful wasn't taken into account.

- - - - -
d9f09506 by Simon Peyton Jones at 2020-07-10T10:33:44-04:00
Define multiShotIO and use it in mkSplitUniqueSupply

This patch is part of the ongoing eta-expansion saga;
see #18238.

It implements a neat trick (suggested by Sebastian Graf)
that allows the programmer to disable the default one-shot behaviour
of IO (the "state hack").  The trick is to use a new multiShotIO
function; see Note [multiShotIO].  For now, multiShotIO is defined
here in Unique.Supply; but it should ultimately be moved to the IO
library.

The change is necessary to get good code for GHC's unique supply;
see Note [Optimising the unique supply].

However it makes no difference to GHC as-is.  Rather, it makes
a difference when a subsequent commit

   Improve eta-expansion using ArityType

lands.

- - - - -
bce695cc by Simon Peyton Jones at 2020-07-10T10:33:44-04:00
Make arityType deal with join points

As Note [Eta-expansion and join points] describes,
this patch makes arityType deal correctly with join points.
What was there before was not wrong, but yielded lower
arities than it could.

Fixes #18328

In base GHC this makes no difference to nofib.

        Program           Size    Allocs   Runtime   Elapsed  TotalMem
--------------------------------------------------------------------------------
         n-body          -0.1%     -0.1%     -1.2%     -1.1%      0.0%
--------------------------------------------------------------------------------
            Min          -0.1%     -0.1%    -55.0%    -56.5%      0.0%
            Max          -0.0%      0.0%    +16.1%    +13.4%      0.0%
 Geometric Mean          -0.0%     -0.0%    -30.1%    -31.0%     -0.0%

But it starts to make real difference when we land the change to the
way mkDupableAlts handles StrictArg, in fixing #13253 and friends.
I think this is because we then get more non-inlined join points.

- - - - -
2b7c71cb by Simon Peyton Jones at 2020-07-11T12:17:02-04:00
Improve eta-expansion using ArityType

As #18355 shows, we were failing to preserve one-shot info when
eta-expanding.  It's rather easy to fix, by using ArityType more,
rather than just Arity.

This patch is important to suport the one-shot monad trick;
see #18202.  But the extra tracking of one-shot-ness requires
the patch

   Define multiShotIO and use it in mkSplitUniqueSupply

If that patch is missing, ths patch makes things worse in
GHC.Types.Uniq.Supply.  With it, however, we see these improvements

    T3064     compiler bytes allocated -2.2%
    T3294     compiler bytes allocated -1.3%
    T12707    compiler bytes allocated -1.3%
    T13056    compiler bytes allocated -2.2%

Metric Decrease:
    T3064
    T3294
    T12707
    T13056

- - - - -
de139cc4 by Artem Pelenitsyn at 2020-07-12T02:53:20-04:00
add reproducer for #15630

- - - - -
c4de6a7a by Andreas Klebinger at 2020-07-12T02:53:55-04:00
Give Uniq[D]FM a phantom type for its key.

This fixes #17667 and should help to avoid such issues going forward.

The changes are mostly mechanical in nature. With two notable
exceptions.

* The register allocator.

  The register allocator references registers by distinct uniques.
  However they come from the types of VirtualReg, Reg or Unique in
  various places. As a result we sometimes cast the key type of the
  map and use functions which operate on the now typed map but take
  a raw Unique as actual key. The logic itself has not changed it
  just becomes obvious where we do so now.

* <Type>Env Modules.

As an example a ClassEnv is currently queried using the types `Class`,
`Name`, and `TyCon`. This is safe since for a distinct class value all
these expressions give the same unique.

    getUnique cls
    getUnique (classTyCon cls)
    getUnique (className cls)
    getUnique (tcName $ classTyCon cls)

This is for the most part contained within the modules defining the
interface. However it requires us to play dirty when we are given a
`Name` to lookup in a `UniqFM Class a` map. But again the logic did
not change and it's for the most part hidden behind the Env Module.

Some of these cases could be avoided by refactoring but this is left
for future work.

We also bump the haddock submodule as it uses UniqFM.

- - - - -
c2cfdfde by Aaron Allen at 2020-07-13T09:00:33-04:00
Warn about empty Char enumerations (#18402)

Currently the "Enumeration is empty" warning (-Wempty-enumerations)
only fires for numeric literals. This patch adds support for `Char`
literals so that enumerating an empty list of `Char`s will also
trigger the warning.

- - - - -
c3ac87ec by Stefan Schulze Frielinghaus at 2020-07-13T09:01:10-04:00
hadrian: build check-ppr dynamic if GHC is build dynamic

Fixes #18361

- - - - -
9ad072b4 by Simon Peyton Jones at 2020-07-13T14:52:49-04:00
Use dumpStyle when printing inlinings

This just makes debug-printing consistent,
and more informative.

- - - - -
e78c4efb by Simon Peyton Jones at 2020-07-13T14:52:49-04:00
Comments only

- - - - -
7ccb760b by Simon Peyton Jones at 2020-07-13T14:52:49-04:00
Reduce result discount in conSize

Ticket #18282 showed that the result discount given by conSize
was massively too large.  This patch reduces that discount to
a constant 10, which just balances the cost of the constructor
application itself.

Note [Constructor size and result discount] elaborates, as
does the ticket #18282.

Reducing result discount reduces inlining, which affects perf.  I
found that I could increase the unfoldingUseThrehold from 80 to 90 in
compensation; in combination with the result discount change I get
these overall nofib numbers:

        Program           Size    Allocs   Runtime   Elapsed  TotalMem
--------------------------------------------------------------------------------
          boyer          -0.2%     +5.4%     -3.2%     -3.4%      0.0%
       cichelli          -0.1%     +5.9%    -11.2%    -11.7%      0.0%
      compress2          -0.2%     +9.6%     -6.0%     -6.8%      0.0%
   cryptarithm2          -0.1%     -3.9%     -6.0%     -5.7%      0.0%
         gamteb          -0.2%     +2.6%    -13.8%    -14.4%      0.0%
         genfft          -0.1%     -1.6%    -29.5%    -29.9%      0.0%
             gg          -0.0%     -2.2%    -17.2%    -17.8%    -20.0%
           life          -0.1%     -2.2%    -62.3%    -63.4%      0.0%
           mate          +0.0%     +1.4%     -5.1%     -5.1%    -14.3%
         parser          -0.2%     -2.1%     +7.4%     +6.7%      0.0%
      primetest          -0.2%    -12.8%    -14.3%    -14.2%      0.0%
         puzzle          -0.2%     +2.1%    -10.0%    -10.4%      0.0%
            rsa          -0.2%    -11.7%     -3.7%     -3.8%      0.0%
         simple          -0.2%     +2.8%    -36.7%    -38.3%     -2.2%
   wheel-sieve2          -0.1%    -19.2%    -48.8%    -49.2%    -42.9%
--------------------------------------------------------------------------------
            Min          -0.4%    -19.2%    -62.3%    -63.4%    -42.9%
            Max          +0.3%     +9.6%     +7.4%    +11.0%    +16.7%
 Geometric Mean          -0.1%     -0.3%    -17.6%    -18.0%     -0.7%

I'm ok with these numbers, remembering that this change removes
an *exponential* increase in code size in some in-the-wild cases.

I investigated compress2.  The difference is entirely caused by this
function no longer inlining

WriteRoutines.$woutputCodes
  = \ (w :: [CodeEvent]) ->
      let result_s1Sr
            = case WriteRoutines.outputCodes_$s$woutput w 0# 0# 8# 9# of
                (# ww1, ww2 #) -> (ww1, ww2)
      in (# case result_s1Sr of (x, _) ->
              map @Int @Char WriteRoutines.outputCodes1 x
         , case result_s1Sr of { (_, y) -> y } #)

It was right on the cusp before, driven by the excessive result
discount.  Too bad!

Happily, the compiler/perf tests show a number of improvements:
    T12227     compiler bytes-alloc  -6.6%
    T12545     compiler bytes-alloc  -4.7%
    T13056     compiler bytes-alloc  -3.3%
    T15263     runtime  bytes-alloc -13.1%
    T17499     runtime  bytes-alloc -14.3%
    T3294      compiler bytes-alloc  -1.1%
    T5030      compiler bytes-alloc -11.7%
    T9872a     compiler bytes-alloc  -2.0%
    T9872b     compiler bytes-alloc  -1.2%
    T9872c     compiler bytes-alloc  -1.5%

Metric Decrease:
    T12227
    T12545
    T13056
    T15263
    T17499
    T3294
    T5030
    T9872a
    T9872b
    T9872c

- - - - -
7f0b671e by Ben Gamari at 2020-07-13T14:52:49-04:00
testsuite: Widen acceptance threshold on T5837

This test is positively tiny and consequently the bytes allocated
measurement will be relatively noisy. Consequently I have seen this
fail spuriously quite often.

- - - - -
118e1c3d by Alp Mestanogullari at 2020-07-14T21:30:52-04:00
compiler: re-engineer the treatment of rebindable if

Executing on the plan described in #17582, this patch changes the way if expressions
are handled in the compiler in the presence of rebindable syntax. We get rid of the
SyntaxExpr field of HsIf and instead, when rebindable syntax is on, we rewrite the HsIf
node to the appropriate sequence of applications of the local `ifThenElse` function.

In order to be able to report good error messages, with expressions as they were
written by the user (and not as desugared by the renamer), we make use of TTG
extensions to extend GhcRn expression ASTs with an `HsExpansion` construct, which
keeps track of a source (GhcPs) expression and the desugared (GhcRn) expression that
it gives rise to. This way, we can typecheck the latter while reporting the former in
error messages.

In order to discard the error context lines that arise from typechecking the desugared
expressions (because they talk about expressions that the user has not written), we
carefully give a special treatment to the nodes fabricated by this new renaming-time
transformation when typechecking them. See Note [Rebindable syntax and HsExpansion]
for more details. The note also includes a recipe to apply the same treatment to
other rebindable constructs.

Tests 'rebindable11' and 'rebindable12' have been added to make sure we report
identical error messages as before this patch under various circumstances.

We also now disable rebindable syntax when processing untyped TH quotes, as per
the discussion in #18102 and document the interaction of rebindable syntax and
Template Haskell, both in Note [Template Haskell quotes and Rebindable Syntax]
and in the user guide, adding a test to make sure that we do not regress in
that regard.

- - - - -
64c774b0 by Andreas Klebinger at 2020-07-14T21:31:27-04:00
Explain why keeping DynFlags in AnalEnv saves allocation.

- - - - -
254245d0 by Ben Gamari at 2020-07-14T21:32:03-04:00
docs/users-guide: Update default -funfolding-use-threshold value

This was changed in 3d2991f8 but I neglected to update the
documentation. Fixes #18419.

- - - - -
4c259f86 by Andreas Klebinger at 2020-07-14T21:32:41-04:00
Escape backslashes in json profiling reports properly.

I also took the liberty to do away the fixed buffer size for escaping.
Using a fixed size here can only lead to issues down the line.

Fixes #18438.

- - - - -
23797224 by Sergei Trofimovich at 2020-07-14T21:33:19-04:00
.gitlab: re-enable integer-simple substitute (BIGNUM_BACKEND)

Recently build system migrated from INTEGER_LIBRARY to BIGNUM_BACKEND.
But gitlab CI was never updated. Let's enable BIGNUM_BACKEND=native.

Bug: https://gitlab.haskell.org/ghc/ghc/-/issues/18437
Signed-off-by: Sergei Trofimovich <slyfox at gentoo.org>

- - - - -
e0db878a by Sergei Trofimovich at 2020-07-14T21:33:19-04:00
ghc-bignum: bring in sync .hs-boot files with module declarations

Before this change `BIGNUM_BACKEND=native` build was failing as:

```
libraries/ghc-bignum/src/GHC/Num/BigNat/Native.hs:708:16: error:
    * Variable not in scope: naturalFromBigNat# :: WordArray# -> t
    * Perhaps you meant one of these:
        `naturalFromBigNat' (imported from GHC.Num.Natural),
        `naturalToBigNat' (imported from GHC.Num.Natural)
    |
708 |           m' = naturalFromBigNat# m
    |
```

This happens because `.hs-boot` files are slightly out of date.
This change brings in data and function types in sync.

Bug: https://gitlab.haskell.org/ghc/ghc/-/issues/18437
Signed-off-by: Sergei Trofimovich <slyfox at gentoo.org>

- - - - -
c9f65c36 by Stefan Schulze Frielinghaus at 2020-07-14T21:33:57-04:00
rts/Disassembler.c: Use FMT_HexWord for printing values in hex format

- - - - -
58ae62eb by Matthias Andreas Benkard at 2020-07-14T21:34:35-04:00
macOS: Load frameworks without stating them first.

macOS Big Sur makes the following change to how frameworks are shipped
with the OS:

> New in macOS Big Sur 11 beta, the system ships with a built-in
> dynamic linker cache of all system-provided libraries. As part of
> this change, copies of dynamic libraries are no longer present on
> the filesystem. Code that attempts to check for dynamic library
> presence by looking for a file at a path or enumerating a directory
> will fail. Instead, check for library presence by attempting to
> dlopen() the path, which will correctly check for the library in the
> cache. (62986286)

https://developer.apple.com/documentation/macos-release-notes/macos-big-sur-11-beta-release-notes/

Therefore, the previous method of checking whether a library exists
before attempting to load it makes GHC.Runtime.Linker.loadFramework
fail to find frameworks installed at /System/Library/Frameworks.

GHC.Runtime.Linker.loadFramework now opportunistically loads the
framework libraries without checking for their existence first,
failing only if all attempts to load a given framework from any of the
various possible locations fail.

- - - - -
cdc4a6b0 by Matthias Andreas Benkard at 2020-07-14T21:34:35-04:00
loadFramework: Output the errors collected in all loading attempts.

With the recent change away from first finding and then loading a
framework, loadFramework had no way of communicating the real reason
why loadDLL failed if it was any reason other than the framework
missing from the file system.  It now collects all loading attempt
errors into a list and concatenates them into a string to return to
the caller.

- - - - -
51dbfa52 by Ben Gamari at 2020-07-15T04:05:34-04:00
StgToCmm: Use CmmRegOff smart constructor

Previously we would generate expressions of the form
`CmmRegOff BaseReg 0`. This should do no harm (and really should be
handled by the NCG anyways) but it's better to just generate a plain
`CmmReg`.

- - - - -
ae11bdfd by Ben Gamari at 2020-07-15T04:06:08-04:00
testsuite: Add regression test for #17744

Test due to @monoidal.

- - - - -
0e3c277a by Ben Gamari at 2020-07-15T16:41:01-04:00
Bump Cabal submodule

Updates a variety of tests as Cabal is now more strict about Cabal file
form.

- - - - -
ceed994a by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Drop Windows Vista support, require Windows 7

- - - - -
00a23bfd by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Update Windows FileSystem wrapper utilities.

- - - - -
459e1c5f by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Use SlimReaderLocks and ConditonalVariables provided by the OS instead of emulated ones

- - - - -
763088fc by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Small linker comment and ifdef cleanups

- - - - -
1a228ff9 by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Flush event logs eagerly.

- - - - -
e9e04dda by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Refactor Buffer structures to be able to track async operations

- - - - -
356dc3fe by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Implement new Console API

- - - - -
90e69f77 by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Add IOPort synchronization primitive

- - - - -
71245fcc by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Add new io-manager cmdline options

- - - - -
d548a3b3 by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Init Windows console Codepage to UTF-8.

- - - - -
58ef6366 by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Add unsafeSplat to GHC.Event.Array

- - - - -
d660725e by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Add size and iterate to GHC.Event.IntTable.

- - - - -
050da6dd by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Switch Testsuite to test winio by default

- - - - -
4bf542bf by Tamar Christina at 2020-07-15T16:41:01-04:00
winio: Multiple refactorings and support changes.

- - - - -
4489af6b by Tamar Christina at 2020-07-15T16:41:02-04:00
winio: core threaded I/O manager

- - - - -
64d8f2fe by Tamar Christina at 2020-07-15T16:41:02-04:00
winio: core non-threaded I/O manager

- - - - -
8da15a09 by Tamar Christina at 2020-07-15T16:41:02-04:00
winio: Fix a scheduler bug with the threaded-runtime.

- - - - -
84ea3d14 by Tamar Christina at 2020-07-15T16:41:02-04:00
winio: Relaxing some constraints in io-manager.

- - - - -
ccf0d107 by Tamar Christina at 2020-07-15T16:41:02-04:00
winio: Fix issues with non-threaded I/O manager after split.

- - - - -
b492fe6e by Tamar Christina at 2020-07-15T16:41:02-04:00
winio: Remove some barf statements that are a bit strict.

- - - - -
01423fd2 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Expand comments describing non-threaded loop

- - - - -
4b69004f by Tamar Christina at 2020-07-15T16:41:02-04:00
winio: fix FileSize unstat-able handles

- - - - -
9b384270 by Tamar Christina at 2020-07-15T16:41:02-04:00
winio: Implement new tempfile routines for winio

- - - - -
f1e0be82 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Fix input truncation when reading from handle.

This was caused by not upholding the read buffer invariant
that bufR == bufL == 0 for empty read buffers.

- - - - -
e176b625 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Fix output truncation for writes larger than buffer size

- - - - -
a831ce0e by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Rewrite bufWrite.

I think it's far easier to follow the code now.
It's also correct now as I had still missed a spot
where we didn't update the offset.

- - - - -
6aefdf62 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Fix offset set by bufReadEmpty.

bufReadEmpty returns the bytes read *including* content that
was already buffered,
But for calculating the offset we only care about the number
of bytes read into the new buffer.

- - - - -
750ebaee by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Clean up code surrounding IOPort primitives.

According to phyx these should only be read and written once per
object. Not neccesarily in that order.

To strengthen that guarantee the primitives will now throw an
exception if we violate this invariant.

As a consequence we can eliminate some code from their primops.
In particular code dealing with multiple queued readers/writers
now simply checks the invariant and throws an exception if it
was violated. That is in contrast to mvars which will do things
like wake up all readers, queue multi writers etc.

- - - - -
ffd31db9 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Fix multi threaded threadDelay and a few other small changes.

Multithreaded threadDelay suffered from a race condition
based on the ioManagerStatus. Since the status isn't needed
for WIO I removed it completely.

This resulted in a light refactoring, as consequence we will always
wake up the IO manager using interruptSystemManager, which uses
`postQueuedCompletionStatus` internally.

I also added a few comments which hopefully makes the code easier to
dive into for the next person diving in.

- - - - -
6ec26df2 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
wionio: Make IO subsystem check a no-op on non-windows platforms.

- - - - -
29bcd936 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Set handle offset when opening files in Append mode.

Otherwise we would truncate the file.

- - - - -
55c29700 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Remove debug event log trace

- - - - -
9acb9f40 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Fix sqrt and openFile009 test cases

- - - - -
57017cb7 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Allow hp2ps to build with -DDEBUG

- - - - -
b8cd9995 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Update output of T9681 since we now actually run it.

- - - - -
10af5b14 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: A few more improvements to the IOPort primitives.

- - - - -
39afc4a7 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Fix expected tempfiles output.

Tempfiles now works properly on windows, as such we can
delete the win32 specific output.

- - - - -
99db46e0 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Assign thread labels to IOManager threads.

- - - - -
be6af732 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Properly check for the tso of an incall to be zero.

- - - - -
e2c6dac7 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Mark FD instances as unsupported under WINIO.

- - - - -
fd02ceed by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Fix threadDelay maxBound invocations.

Instead of letting the ns timer overflow now clamp it at
(maxBound :: Word64) ns. That still gives a few hundred
years.

- - - - -
bc79f9f1 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Add comments/cleanup an import in base

- - - - -
1d197f4b by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Mark outstanding_service_requests volatile.

As far as I know C(99) gives no guarantees for code like

    bool condition;

    ...

    while(condition)
        sleep();

that condition will be updated if it's changed by another thread.
So we are explicit here and mark it as volatile, this will force
a reload from memory on each iteration.

- - - - -
dc438186 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Make last_event a local variable

- - - - -
2fc957c5 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Add comment about thread safety of processCompletion.

- - - - -
4c026b6c by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: nonthreaded: Create io processing threads in main thread.

We now set a flag in the IO thread. The scheduler when looking for work
will check the flag and create/queue threads accordingly.

We used to create these in the IO thread. This improved performance
but caused frequent segfaults. Thread creation/allocation is only safe to
do if nothing currently accesses the storeagemanager. However without
locks in the non-threaded runtime this can't be guaranteed.

This shouldn't change performance all too much.

In the past we had:
* IO: Create/Queue thread.
* Scheduler: Runs a few times. Eventually picks up IO processing thread.

Now it's:
* IO: Set flag to queue thread.
* Scheduler: Pick up flag, if set create/queue thread. Eventually picks up IO processing thread.

- - - - -
f47c7208 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Add an exported isHeapAlloced function to the RTS

- - - - -
cc5d7bb1 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Queue IO processing threads at the front of the queue.

This will unblock the IO thread sooner hopefully leading to higher
throughput in some situations.

- - - - -
e7630115 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: ThreadDelay001: Use higher resolution timer.

- - - - -
451b5f96 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Update T9681 output, disable T4808 on windows.

T4808 tests functionality of the FD interface which won't be supported
under WINIO.

T9681 just has it's expected output tweaked.

- - - - -
dd06f930 by Andreas Klebinger at 2020-07-15T16:41:02-04:00
winio: Wake io manager once per registerTimeout.

Which is implicitly done in editTimeouts, so need to wake it
up twice.

- - - - -
e87d0bf9 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Update placeholder comment with actual function name.

- - - - -
fc9025db by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Always lock win32 event queue

- - - - -
c24c9a1f by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Display thread labels when tracing scheduler events.

- - - - -
06542b03 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Refactor non-threaded runner thread and scheduler interface.

Only use a single communication point (registerAlertableWait) to inform
the C side aobut both timeouts to use as well as outstanding requests.

Also queue a haskell processing thread after each return from alertable
waits. This way there is no risk of us missing a timer event.

- - - - -
256299b1 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Remove outstanding_requests from runner.

We used a variable to keep track of situations where we got
entries from the IO port, but all of them had already been
canceled. While we can avoid some work that way this case
seems quite rare.

So we give up on tracking this and instead always assume at
least one of the returned entries is valid.

If that's not the case no harm is done, we just perform some
additional work. But it makes the runner easier to reason about.

In particular we don't need to care if another thread modifies
oustanding_requests after we return from waiting on the IO Port.

- - - - -
3ebd8ad9 by Tamar Christina at 2020-07-15T16:41:03-04:00
winio: Various fixes related to rebase and testdriver

- - - - -
6be6bcba by Tamar Christina at 2020-07-15T16:41:03-04:00
winio: Fix rebase artifacts

- - - - -
2c649dc3 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Rename unsafeSplat to unsafeCopyFromBuffer

- - - - -
a18b73f3 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Remove unused size/iterate operations from IntTable

- - - - -
16bab48e by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Detect running IO Backend via peeking at RtsConfig

- - - - -
8b8405a0 by Tamar Christina at 2020-07-15T16:41:03-04:00
winio: update temp path so GCC etc can handle it.

Also fix PIPE support, clean up error casting, fix memory leaks

- - - - -
2092bc54 by Ben Gamari at 2020-07-15T16:41:03-04:00
winio: Minor comments/renamings

- - - - -
a5b5b6c0 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Checking if an error code indicates completion is now a function.

- - - - -
362176fd by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Small refactor in withOverlappedEx

- - - - -
32e20597 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: A few comments and commented out dbxIO

- - - - -
a4bfc1d9 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Don't drop buffer offset in byteView/cwcharView

- - - - -
b3ad2a54 by Tamar Christina at 2020-07-15T16:41:03-04:00
winio: revert BHandle changes.

- - - - -
3dcd87e2 by Ben Gamari at 2020-07-15T16:41:03-04:00
winio: Fix imports

- - - - -
5a371890 by Tamar Christina at 2020-07-15T16:41:03-04:00
winio: update ghc-cabal to handle new Cabal submodule bump

- - - - -
d07ebe0d by Ben Gamari at 2020-07-15T16:41:03-04:00
winio: Only compile sources on Windows

- - - - -
dcb42393 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Actually return Nothing on EOF for non-blocking read

- - - - -
895a3beb by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Deduplicate logic in encodeMultiByte[Raw]IO.

- - - - -
e06e6734 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Deduplicate openFile logic

- - - - -
b59430c0 by Tamar Christina at 2020-07-15T16:41:03-04:00
winio: fix -werror issue in encoding file

- - - - -
f8d39a51 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Don't mention windows specific functions when building on Linux.

- - - - -
6a533d2a by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: add a note about file locking in the RTS.

- - - - -
cf37ce34 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Add version to @since annotation

- - - - -
0fafa2eb by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Rename GHC.Conc.IOCP -> GHC.Conc.WinIO

- - - - -
1854fc23 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Expand GHC.Conc.POSIX description

It now explains users may not use these functions when
using the old IO manager.

- - - - -
fcc7ba41 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Fix potential spaceleak in __createUUIDTempFileErrNo

- - - - -
6b3fd9fa by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Remove redundant -Wno-missing-signatures pragmas

- - - - -
916fc861 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Make it explicit that we only create one IO manager

- - - - -
f260a721 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Note why we don't use blocking waits.

- - - - -
aa0a4bbf by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Remove commented out pragma

- - - - -
d679b544 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Remove redundant buffer write in Handle/Text.hs:bufReadEmpty

- - - - -
d3f94368 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Rename SmartHandles to StdHandles

- - - - -
bd6b8ec1 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: add comment stating failure behaviour for getUniqueFileInfo.

- - - - -
12846b85 by Andreas Klebinger at 2020-07-15T16:41:03-04:00
winio: Update IOPort haddocks.

- - - - -
9f39fb14 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Add a note cross reference

- - - - -
62dd5a73 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Name Haskell/OS I/O Manager explicitly in Note

- - - - -
fa807828 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Expand BlockedOnIOCompletion description.

- - - - -
f0880a1d by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Remove historical todos

- - - - -
8e58e714 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Update note, remove debugging pragma.

- - - - -
aa4d84d5 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: flushCharReadBuffer shouldn't need to adjust offsets.

- - - - -
e580893a by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Remove obsolete comment about cond. variables

- - - - -
d54e9d79 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: fix initial linux validate build

- - - - -
3cd4de46 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Fix ThreadDelay001 CPP

- - - - -
c88b1b9f by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Fix openFile009 merge conflict leftover

- - - - -
849e8889 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Accept T9681 output.

GHC now reports String instead of [Char].

- - - - -
e7701818 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Fix cabal006 after upgrading cabal submodule

Demand cabal 2.0 syntax instead of >= 1.20 as required by newer cabal versions.

- - - - -
a44f0373 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Fix stderr output for ghci/linking/dyn tests.

We used to filter rtsopts, i opted to instead just accept the warning of it having no effect.
This works both for -rtsopts, as well as -with-rtsopts which winio adds.

- - - - -
515d9896 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Adjust T15261b stdout for --io-manager flag.

- - - - -
949aaacc by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Adjust T5435_dyn_asm stderr

The warning about rtsopts having no consequences is expected.
So accept new stderr.

- - - - -
7d424e1e by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Also accept T7037 stderr

- - - - -
1f009768 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: fix cabal04 by filtering rts args

- - - - -
981a9f2e by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: fix cabal01 by accepting expected stderr

- - - - -
b7b0464e by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: fix safePkg01 by accepting expected stderr

- - - - -
32734b29 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: fix T5435_dyn_gcc by accepting expected stderr

- - - - -
acc5cebf by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: fix tempfiles test on linux

- - - - -
c577b789 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Accept accepted stderr for T3807

- - - - -
c108c527 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Accept accepted stderr for linker_unload

- - - - -
2b0b9a08 by Andreas Klebinger at 2020-07-15T16:41:04-04:00
winio: Accept accepted stderr for linker_unload_multiple_objs

- - - - -
67afb03c by Tamar Christina at 2020-07-15T16:41:04-04:00
winio: clarify wording on conditional variables.

- - - - -
3bd41572 by Tamar Christina at 2020-07-15T16:41:04-04:00
winio: clarify comment on cooked mode.

- - - - -
ded58a03 by Tamar Christina at 2020-07-15T16:41:04-04:00
winio: update lockfile signature and remove mistaken symbol in rts.

- - - - -
2143c492 by Ben Gamari at 2020-07-15T16:41:04-04:00
testsuite: Add winio and winio_threaded ways

Reverts many of the testsuite changes

- - - - -
c0979cc5 by Ben Gamari at 2020-07-16T10:56:54-04:00
Merge remote-tracking branch 'origin/wip/winio'

- - - - -
750a1595 by Ben Gamari at 2020-07-18T07:26:41-04:00
rts: Add --copying-gc flag to reverse effect of --nonmoving-gc

Fixes #18281.

- - - - -
6ba6a881 by Hécate at 2020-07-18T07:26:42-04:00
Implement `fullCompilerVersion`

Follow-up of https://gitlab.haskell.org/ghc/ghc/-/issues/18403

This MR adds `fullCompilerVersion`, a function that shares the same
backend as the `--numeric-version` GHC flag, exposing a full,
three-digit version datatype.

- - - - -
e6cf27df by Hécate at 2020-07-18T07:26:43-04:00
Add a Lint hadrian rule and an .hlint.yaml file in base/

- - - - -
bcb177dd by Simon Peyton Jones at 2020-07-18T07:26:43-04:00
Allow multiple case branches to have a higher rank type

As #18412 points out, it should be OK for multiple case alternatives
to have a higher rank type, provided they are all the same.

This patch implements that change.  It sweeps away
GHC.Tc.Gen.Match.tauifyMultipleBranches, and friends, replacing it
with an enhanced version of fillInferResult.

The basic change to fillInferResult is to permit the case in which
another case alternative has already filled in the result; and in
that case simply unify.  It's very simple actually.

See the new Note [fillInferResult] in TcMType

Other refactoring:

- Move all the InferResult code to one place, in GHC.Tc.Utils.TcMType
  (previously some of it was in Unify)

- Move tcInstType and friends from TcMType to Instantiate, where it
  more properly belongs.  (TCMType was getting very long.)

- - - - -
e5525a51 by Simon Peyton Jones at 2020-07-18T07:26:43-04:00
Improve typechecking of NPlusK patterns

This patch (due to Richard Eisenberg) improves
documentation of the wrapper returned by tcSubMult
(see Note [Wrapper returned from tcSubMult] in
 GHC.Tc.Utils.Unify).

And, more substantially, it cleans up the multiplicity
handling in the typechecking of NPlusKPat

- - - - -
12f90352 by Krzysztof Gogolewski at 2020-07-18T07:26:45-04:00
Remove {-# CORE #-} pragma (part of #18048)

This pragma has no effect since 2011.
It was introduced for External Core, which no longer exists.

Updates haddock submodule.

- - - - -
e504c913 by Simon Peyton Jones at 2020-07-18T07:26:45-04:00
Refactor the simplification of join binders

This MR (for #18449) refactors the Simplifier's treatment
of join-point binders.

Specifically, it puts together, into
     GHC.Core.Opt.Simplify.Env.adjustJoinPointType
two currently-separate ways in which we adjust the type of
a join point. As the comment says:

-- (adjustJoinPointType mult new_res_ty join_id) does two things:
--
--   1. Set the return type of the join_id to new_res_ty
--      See Note [Return type for join points]
--
--   2. Adjust the multiplicity of arrows in join_id's type, as
--      directed by 'mult'. See Note [Scaling join point arguments]

I think this actually fixes a latent bug, by ensuring that the
seIdSubst and seInScope have the right multiplicity on the type
of join points.

I did some tidying up while I was at it.  No more
setJoinResTy, or modifyJoinResTy: instead it's done locally in
Simplify.Env.adjustJoinPointType

- - - - -
49b265f0 by Chaitanya Koparkar at 2020-07-18T07:26:46-04:00
Fix minor typos in a Core.hs note

- - - - -
8d59aed6 by Stefan Schulze Frielinghaus at 2020-07-18T07:26:47-04:00
GHCi: Fix isLittleEndian

- - - - -
c26e81d1 by Ben Gamari at 2020-07-18T07:26:47-04:00
testsuite: Mark ghci tests as fragile under unreg compiler

In particular I have seen T16012 fail repeatedly under the
unregisterised compiler.

- - - - -
840bbfc3 by John Ericson at 2020-07-20T14:39:40+02:00
Make fixed-size `Int32#` and `Int64#`

The boxed Int64 uses Int64#, but Int32# still uses Int#. The 32-bit case
is less pressing to change because it is not a source of brittle
CPP---it is the same thing on all platforms.

We need Int64/Word64 constant folding to avoid the let/app restriction on Core, so
that is implemented now. 32-bit constant unfolding and 32-bit literals
are left as follow-up.

This is the bulk of #11953

- - - - -
d6574966 by John Ericson at 2020-07-20T14:39:40+02:00
Inline INT64 and WORD64 macros in primops.txt.pp

The definition is now unconditional so there is no reason for that CPP.

- - - - -
0298108e by Sylvain Henry at 2020-07-20T14:39:40+02:00
ghc-bignum: add support for Word64#/Int64# on 64-bit arch

- - - - -
f20f7005 by Sylvain Henry at 2020-07-20T16:19:03+02:00
Fix fingerprint Core generation

- - - - -
7889c584 by Sylvain Henry at 2020-07-20T16:19:23+02:00
Fix some tests

- - - - -


30 changed files:

- .gitlab-ci.yml
- .gitlab/ci.sh
- Makefile
- compiler/GHC/Builtin/Names.hs
- + compiler/GHC/Builtin/RebindableNames.hs
- compiler/GHC/Builtin/Types/Prim.hs
- compiler/GHC/Builtin/Utils.hs
- compiler/GHC/Builtin/primops.txt.pp
- compiler/GHC/Cmm/LayoutStack.hs
- compiler/GHC/Cmm/Parser.y
- compiler/GHC/Cmm/Sink.hs
- compiler/GHC/CmmToAsm.hs
- compiler/GHC/CmmToAsm/BlockLayout.hs
- compiler/GHC/CmmToAsm/Monad.hs
- compiler/GHC/CmmToAsm/Reg/Graph.hs
- compiler/GHC/CmmToAsm/Reg/Graph/Coalesce.hs
- compiler/GHC/CmmToAsm/Reg/Graph/Spill.hs
- compiler/GHC/CmmToAsm/Reg/Graph/SpillClean.hs
- compiler/GHC/CmmToAsm/Reg/Graph/SpillCost.hs
- compiler/GHC/CmmToAsm/Reg/Graph/Stats.hs
- compiler/GHC/CmmToAsm/Reg/Linear.hs
- compiler/GHC/CmmToAsm/Reg/Linear/Base.hs
- compiler/GHC/CmmToAsm/Reg/Linear/JoinToTargets.hs
- compiler/GHC/CmmToAsm/Reg/Linear/StackMap.hs
- compiler/GHC/CmmToAsm/Reg/Linear/Stats.hs
- compiler/GHC/CmmToAsm/Reg/Liveness.hs
- + compiler/GHC/CmmToAsm/Reg/Utils.hs
- compiler/GHC/CmmToAsm/X86/RegInfo.hs
- compiler/GHC/CmmToLlvm/Base.hs
- compiler/GHC/CmmToLlvm/CodeGen.hs


The diff was not included because it is too large.


View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/280b3843169580261f3816cf182d204476d20b0c...7889c584da61dd9da81e7df24d54fd8855e8ee68

-- 
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/280b3843169580261f3816cf182d204476d20b0c...7889c584da61dd9da81e7df24d54fd8855e8ee68
You're receiving this email because of your account on gitlab.haskell.org.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-commits/attachments/20200720/9f647cbe/attachment-0001.html>


More information about the ghc-commits mailing list