| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
ModLocation is the data type which tells you the locations of all the
build products which can affect recompilation. It is now computed in one
place and not modified through the pipeline. Important locations will
now just consult ModLocation rather than construct the dynamic object
path incorrectly.
* Add paths for dynamic object and dynamic interface files to
ModLocation.
* Always use the paths from mod location when looking for where to find
any interface or object file.
* Always use the paths in a ModLocation when deciding where to write an
interface and object file.
* Remove `dynamicOutputFile` and `dynamicOutputHi` functions which
*calculated* (incorrectly) the location of `dyn_o` and `dyn_hi` files.
* Don't set `outputFile_` and so-on in `enableCodeGenWhen`, `-o` and
hence `outputFile_` should not affect the location of object files in
`--make` mode. It is now sufficient to just update the ModLocation with
the temporary paths.
* In `hscGenBackendPipeline` don't recompute the `ModLocation` to
account for `-dynamic-too`, the paths are now accurate from the start
of the run.
* Rename `getLocation` to `mkOneShotModLocation`, as that's the only
place it's used. Increase the locality of the definition by moving it
close to the use-site.
* Load the dynamic interface from ml_dyn_hi_file rather than attempting
to reconstruct it in load_dynamic_too.
* Add a variety of tests to check how -o -dyno etc interact with each
other.
Some other clean-ups
* DeIOify mkHomeModLocation and friends, they are all pure functions.
* Move FinderOpts into GHC.Driver.Config.Finder, next to initFinderOpts.
* Be more precise about whether we mean outputFile or outputFile_: there
were many places where outputFile was used but the result shouldn't have
been affected by `-dyno` (for example the filename of the resulting
executable). In these places dynamicNow would never be set but it's
still more precise to not allow for this possibility.
* Typo fixes suffices -> suffixes in the appropiate places.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's important that when -finfo-table-map is enabled that we generate
IPE entries just for those info tables which are actually used. To this
end, the info tables which are used are collected just before code
generation starts and entries only created for those tables.
Not accounted for in this scheme was the dead code elimination in the
native code generator. When compiling GHC this optimisation removed an
info table which had an IPE entry which resulting in the following kind
of linker error:
```
/home/matt/ghc-with-debug/_build/stage1/lib/../lib/x86_64-linux-ghc-9.3.20210928/libHSCabal-3.5.0.0-ghc9.3.20210928.so: error: undefined reference to '.Lc5sS_info'
/home/matt/ghc-with-debug/_build/stage1/lib/../lib/x86_64-linux-ghc-9.3.20210928/libHSCabal-3.5.0.0-ghc9.3.20210928.so: error: undefined reference to '.Lc5sH_info'
/home/matt/ghc-with-debug/_build/stage1/lib/../lib/x86_64-linux-ghc-9.3.20210928/libHSCabal-3.5.0.0-ghc9.3.20210928.so: error: undefined reference to '.Lc5sm_info'
collect2: error: ld returned 1 exit status
`cc' failed in phase `Linker'. (Exit code: 1)
Development.Shake.cmd, system command failed
```
Unfortunately, by the time this optimisation happens the structure of
the CmmInfoTable has been lost, we only have the generated code for the
info table to play with so we can no longer just collect all the used
info tables and generate the IPE map.
This leaves us with two options:
1. Return a list of the names of the discarded info tables and then
remove them from the map. This is awkward because we need to do code
generation for the map as well.
2. Just disable this small code size optimisation when -finfo-table-map
is enabled. The option produces very big object files anyway.
Option 2 is much easier to implement and means we don't have to thread
information around awkwardly. It's at the cost of slightly larger object
files (as dead code is not eliminated).
Disabling this optimisation allows an IPE build of GHC to complete
successfully.
Fixes #20428
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Use DiagOpts for diagnostic options instead of directly querying
DynFlags (#17957).
Surprising performance improvements on CI:
T4801(normal) ghc/alloc 313236344.0 306515216.0 -2.1% GOOD
T9961(normal) ghc/alloc 384502736.0 380584384.0 -1.0% GOOD
ManyAlternatives(normal) ghc/alloc 797356128.0 786644928.0 -1.3%
ManyConstructors(normal) ghc/alloc 4389732432.0 4317740880.0 -1.6%
T783(normal) ghc/alloc 408142680.0 402812176.0 -1.3%
Metric Decrease:
T4801
T9961
T783
ManyAlternatives
ManyConstructors
Bump haddock submodule
|
| |
|
|
|
|
|
|
|
|
| |
This is an attempt at reducing the number of dependencies of the Parser
(as reported by CountParserDeps). Modules in GHC.Parser.* don't import
GHC.Driver.Session directly anymore.
Sadly some GHC.Driver.* modules are still transitively imported and the
number of dependencies didn't decrease. But it's a step in the right
direction.
|
|
|
Introduce LogFlags as a independent subset of DynFlags used for logging.
As a consequence in many places we don't have to pass both Logger and
DynFlags anymore.
The main reason for this refactoring is that I want to refactor the
systools interfaces: for now many systools functions use DynFlags both
to use the Logger and to fetch their parameters (e.g. ldInputs for the
linker). I'm interested in refactoring the way they fetch their
parameters (i.e. use dedicated XxxOpts data types instead of DynFlags)
for #19877. But if I did this refactoring before refactoring the Logger,
we would have duplicate parameters (e.g. ldInputs from DynFlags and
linkerInputs from LinkerOpts). Hence this patch first.
Some flags don't really belong to LogFlags because they are subsystem
specific (e.g. most DumpFlags). For example -ddump-asm should better be
passed in NCGConfig somehow. This patch doesn't fix this tight coupling:
the dump flags are part of the UI but they are passed all the way down
for example to infer the file name for the dumps.
Because LogFlags are a subset of the DynFlags, we must update the former
when the latter changes (not so often). As a consequence we now use
accessors to read/write DynFlags in HscEnv instead of using `hsc_dflags`
directly.
In the process I've also made some subsystems less dependent on DynFlags:
- CmmToAsm: by passing some missing flags via NCGConfig (see new fields
in GHC.CmmToAsm.Config)
- Core.Opt.*:
- by passing -dinline-check value into UnfoldingOpts
- by fixing some Core passes interfaces (e.g. CallArity, FloatIn)
that took DynFlags argument for no good reason.
- as a side-effect GHC.Core.Opt.Pipeline.doCorePass is much less
convoluted.
|
