GitLab

Before this commit, GHC only supported LLVM 3.6. Now it only supports
LLVM 3.7 which was released in August 2015. LLVM version 3.6 and earlier
do not work on AArch64/Arm64, but 3.7 does.

Also:
* Add CC_Ghc constructor to LlvmCallConvention.
* Replace `maxSupportLlvmVersion`/`minSupportLlvmVersion` with
  a single `supportedLlvmVersion` variable.
* Get `supportedLlvmVersion` from version specified in configure.ac.
* Drop llvmVersion field from DynFlags (no longer needed because only
  one version is supported).

Test Plan: Validate on x86_64 and arm

Reviewers: bgamari, austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1320

GHC Trac Issues: #10953

Pushed by mistacke before it was ready.

This reverts commit 5dc3db74.

This fixes Trac #7883.

This adds proper support for,
  * `MO_AtomicRMW`
  * `MO_AtomicWrite`
  * `MO_CmpXChg`

Test Plan: Validate

Reviewers: rrnewton, austin

Subscribers: thomie

Differential Revision: https://phabricator.haskell.org/D1282

GHC Trac Issues: #7883

This includes:

- Adding new LlvmType called LMStructP that represents an unpacked
  struct (this is necessary since LLVM's instructions the
  llvm.sadd.with.overflow.* return an unpacked struct).

- Modifications to LlvmCodeGen.CodeGen to generate the LLVM
  instructions for the primops.

- Modifications to StgCmmPrim to actually use those three instructions
  if we use the LLVM backend (so far they were only used for NCG).

Test Plan: validate

Reviewers: austin, rwbarton, bgamari

Reviewed By: bgamari

Subscribers: thomie, bgamari

Differential Revision: https://phabricator.haskell.org/D991

GHC Trac Issues: #9430

Due to changes in LLVM 3.5 aliases now may only refer to definitions.
Previously to handle symbols defined outside of the current commpilation
unit GHC would emit both an `external` declaration, as well as an alias
pointing to it, e.g.,

    @stg_BCO_info = external global i8
    @stg_BCO_info$alias = alias private i8* @stg_BCO_info

Where references to `stg_BCO_info` will use the alias
`stg_BCO_info$alias`. This is not permitted under the new alias
behavior, resulting in errors resembling,

    Alias must point to a definition
    i8* @"stg_BCO_info$alias"

To fix this, we invert the naming relationship between aliases and
definitions. That is, now the symbol definition takes the name
`@stg_BCO_info$def` and references use the actual name, `@stg_BCO_info`.
This means the external symbols can be handled by simply emitting an
`external` declaration,

    @stg_BCO_info = external global i8

Whereas in the case of a forward declaration we emit,

    @stg_BCO_info = alias private i8* @stg_BCO_info$def

Reviewed By: austin

Differential Revision: https://phabricator.haskell.org/D155

In some cases, the layout of the LANGUAGE/OPTIONS_GHC lines has been
reorganized, while following the convention, to

- place `{-# LANGUAGE #-}` pragmas at the top of the source file, before
  any `{-# OPTIONS_GHC #-}`-lines.

- Moreover, if the list of language extensions fit into a single
  `{-# LANGUAGE ... -#}`-line (shorter than 80 characters), keep it on one
  line. Otherwise split into `{-# LANGUAGE ... -#}`-lines for each
  individual language extension. In both cases, try to keep the
  enumeration alphabetically ordered.
  (The latter layout is preferable as it's more diff-friendly)

While at it, this also replaces obsolete `{-# OPTIONS ... #-}` pragma
occurences by `{-# OPTIONS_GHC ... #-}` pragmas.

This combined patch reworks the LLVM backend in a number of ways:

1. Most prominently, we introduce a LlvmM monad carrying the contents of
   the old LlvmEnv around. This patch completely removes LlvmEnv and
   refactors towards standard library monad combinators wherever possible.

2. Support for streaming - we can now generate chunks of Llvm for Cmm as
   it comes in. This might improve our speed.

3. To allow streaming, we need a more flexible way to handle forward
   references. The solution (getGlobalPtr) unifies LlvmCodeGen.Data
   and getHsFunc as well.

4. Skip alloca-allocation for registers that are actually never written.
   LLVM will automatically eliminate these, but output is smaller and
   friendlier to human eyes this way.

5. We use LlvmM to collect references for llvm.used. This allows places
   other than cmmProcLlvmGens to generate entries.

I am not quite sure at what point it makes sense to look at arrays as
pointers, but I ran into at least one use case that strongly suggested
doing it this way (calculating the actual size of structures, to be exact).

Also give them a proper constructor - getGlobalVar and getGlobalValue
map directly to the accessors.

This patch reworks some parts of the LLVM pretty-printing code that were
still using Show and String. Now we should be using SDoc and Outputable
throughout. Note that many get*Name functions become pp*Name
here as a side-effect.

- MetaArgs is not needed, as variables are already meta data

- Same goes for MetaVal - its only reason for existing seems to be to
  support LLVM's strange pretty-printing for meta-data annotations, and
  I feel that is better to keep the data structure clean and handle it in
  the pretty-printing instead.

- Rename "MetaData" to "MetaAnnot". Meta-data is still meta-data when it
  is not associated with an expression or statement - for example compile
  unit data for debugging. I feel the old name was a bit misleading.

- Make the renamed MetaAnnot a proper data type instead of a type alias
  for a pair.

- Rename "MetaExpr" constructor to "MetaStruct". As the data is much more
  like a LLVM structure (not array, as it can contain values).

- Fix a warning

backend.

This patch adds support for LLVM vectors and vector operations to our internal
LLVM abstract syntax data types.

The regular structure type adds padding to conform to the platform ABI,
which causes problems with structures storing doubles under windows since
we don't conform to the platform ABI there. So we use packed structures
instead now that don't do any padding.

Patch from Erik de Castro Lopo <erikd@mega-nerd.com>.

These allow annotations of the code produced by the backend
which should bring some perforamnce gains. At the moment
the attributes aren't being used though.

This involved removing the old constant handling mechanism
which was fairly hard to use. Now being constant or not is
simply a property of a global variable instead of a separate
type.

We do this through a gnu as feature called subsections,
where you can put data/code into a numbered subsection
and those subsections will be joined together in descending
order by gas at compile time.

This was done as part of an honours thesis at UNSW, the paper describing the
work and results can be found at:

http://www.cse.unsw.edu.au/~pls/thesis/davidt-thesis.pdf

A Homepage for the backend can be found at:

http://hackage.haskell.org/trac/ghc/wiki/Commentary/Compiler/Backends/LLVM

Quick summary of performance is that for the 'nofib' benchmark suite, runtimes
are within 5% slower than the NCG and generally better than the C code
generator.  For some code though, such as the DPH projects benchmark, the LLVM
code generator outperforms the NCG and C code generator by about a 25%
reduction in run times.