summaryrefslogtreecommitdiff
path: root/compiler
Commit message (Collapse)AuthorAgeFilesLines
* CmmToLlvm: rename LCGConfig -> LlvmCgConfigdoyougnu2021-12-149-107/+107
| | | | | | | | | | | | | | | | | | | | CmmToLlvm: renamce lcgPlatform -> llvmCgPlatform CmmToLlvm: rename lcgContext -> llvmCgContext CmmToLlvm: rename lcgFillUndefWithGarbage CmmToLlvm: rename lcgSplitSections CmmToLlvm: lcgBmiVersion -> llvmCgBmiVersion CmmToLlvm: lcgLlvmVersion -> llvmCgLlvmVersion CmmToLlvm: lcgDoWarn -> llvmCgDoWarn CmmToLlvm: lcgLlvmConfig -> llvmCgLlvmConfig CmmToLlvm: llvmCgPlatformMisc --> llvmCgLlvmTarget
* CmmToLlvm.Config: strictify LlvmConfig fielddoyougnu2021-12-141-2/+3
|
* LLVM.CodeGen: use fast-string literalsdoyougnu2021-12-141-34/+145
| | | | | | | That is remove factorization of common strings and string building code for the LLVM code gen ops. Replace these with string literals to obey the FastString rewrite rule in GHC.Data.FastString and compute the string length at compile time
* SysTools.Tasks Llvm.Types: remove redundant importdoyougnu2021-12-142-3/+2
| | | | | | | | Llvm.Types: remove redundant import SysTools.Tasks: remove redundant import - namely CmmToLlvm.Base
* CmmToLlvm: Remove DynFlags, add LlvmCgConfigdoyougnu2021-12-1413-226/+251
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | CodeOutput: LCGConfig, add handshake initLCGConfig Add two modules: GHC.CmmToLlvm.Config -- to hold the Llvm code gen config GHC.Driver.Config.CmmToLlvm -- for initialization, other utils CmmToLlvm: remove HasDynFlags, add LlvmConfig CmmToLlvm: add lcgContext to LCGConfig CmmToLlvm.Base: DynFlags --> LCGConfig Llvm: absorb LlvmOpts into LCGConfig CmmToLlvm.Ppr: swap DynFlags --> LCGConfig CmmToLlvm.CodeGen: swap DynFlags --> LCGConfig CmmToLlvm.CodeGen: swap DynFlags --> LCGConfig CmmToLlvm.Data: swap LlvmOpts --> LCGConfig CmmToLlvm: swap DynFlags --> LCGConfig CmmToLlvm: move LlvmVersion to CmmToLlvm.Config Additionally: - refactor Config and initConfig to hold LlvmVersion - push IO needed to get LlvmVersion to boundary between Cmm and LLvm code generation - remove redundant imports, this is much cleaner! CmmToLlvm.Config: store platformMisc_llvmTarget instead of all of platformMisc
* ghc-bin: Add --merge-objs modeBen Gamari2021-12-144-27/+38
| | | | | | | | | | This adds a new mode, `--merge-objs`, which can be used to produce merged GHCi library objects. As future work we will rip out the object-merging logic in Hadrian and Cabal and instead use this mode. Closes #20712.
* compiler: Use withFile instead of bracketBen Gamari2021-12-141-2/+1
| | | | A minor refactoring noticed by hlint.
* compiler: Drop `Maybe ModLocation` from T_MergeForeignBen Gamari2021-12-143-8/+8
| | | | This field was entirely unused.
* Ghci environment: Do not remove shadowed idsJoachim Breitner2021-12-142-43/+25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Names defined earier but shadowed need to be kept around, e.g. for type signatures: ``` ghci> data T = T ghci> let t = T ghci> data T = T ghci> :t t t :: Ghci1.T ``` and indeed they can be used: ``` ghci> let t2 = Ghci1.T :: Ghci1.T ghci> :t t2 t2 :: Ghci1.T ``` However, previously this did not happen for ids (non-types), although they are still around under the qualified name internally: ``` ghci> let t = "other t" ghci> t' <interactive>:8:1: error: • Variable not in scope: t' • Perhaps you meant one of these: ‘Ghci2.t’ (imported from Ghci2), ‘t’ (line 7), ‘t2’ (line 5) ghci> Ghci2.t <interactive>:9:1: error: • GHC internal error: ‘Ghci2.t’ is not in scope during type checking, but it passed the renamer tcl_env of environment: [] • In the expression: Ghci2.t In an equation for ‘it’: it = Ghci2.t ``` This fixes the problem by simply removing the code that tries to remove shadowed ids from the environment. Now you can refer to shadowed ids using `Ghci2.t`, just like you can do for data and type constructors. This simplifies the code, makes terms and types more similar, and also fixes #20455. Now all names ever defined in GHCi are in `ic_tythings`, which is printed by `:show bindings`. But for that commands, it seems to be more ergonomic to only list those bindings that are not shadowed. Or, even if it is not more ergonomic, it’s the current behavour. So let's restore that by filtering in `icInScopeTTs`. Of course a single `TyThing` can be associated with many names. We keep it it in the bindings if _any_ of its names are still visible unqualifiedly. It's a judgement call. This commit also turns a rather old comment into a test files. The comment is is rather stale and things are better explained elsewhere. Fixes #925. Two test cases are regressing: T14052(ghci) ghc/alloc 2749444288.0 12192109912.0 +343.4% BAD T14052Type(ghci) ghc/alloc 7365784616.0 10767078344.0 +46.2% BAD This is not unexpected; the `ic_tythings list grows` a lot more if we don’t remove shadowed Ids. I tried to alleviate it a bit with earlier MRs, but couldn’t make up for it completely. Metric Increase: T14052 T14052Type
* GHC.Utils.Misc.only: Add doc string.Andreas Klebinger2021-12-141-0/+4
| | | | | | | This function expects a singleton list as argument but only checks this in debug builds. I've added a docstring saying so. Fixes #20797
* Add PromotedInfixT/PromotedUInfixT to THJakob Brünker2021-12-112-40/+65
| | | | | | | | | | | | Previously, it was not possible to refer to a data constructor using InfixT with a dynamically bound name (i.e. a name with NameFlavour `NameS` or `NameQ`) if a type constructor of the same name exists. This commit adds promoted counterparts to InfixT and UInfixT, analogously to how PromotedT is the promoted counterpart to ConT. Closes #20773
* Use file hash cache when hashing object file dependenciesMatthew Pickering2021-12-091-1/+2
| | | | | | | | | | | | | | | | | | | This fixes the immediate problem that we hash the same file multiple different times which causes quite a noticeably performance regression. In the future we can probably do better than this by storing the implementation hash in the interface file rather than dependending on hashing the object file. Related to #20604 which notes some inefficiencies with the current recompilation logic. Closes #20790 ------------------------- Metric Decrease: T14052Type -------------------------
* Remove `optLevel` from `DynFlags` (closes #20500)Gergo ERDI2021-12-091-17/+28
|
* Add `Opt_CoreConstantFolding` to turn on constant folding (#20500)Gergo ERDI2021-12-094-3/+8
| | | | | Previously, `-O1` and `-O2`, by way of their effect on the compilation pipeline, they implicitly turned on constant folding
* Use same optimization pipeline regardless of `optLevel` (#20500)Gergo ERDI2021-12-091-13/+9
|
* package imports: Take into account package visibility when renamingMatthew Pickering2021-12-097-19/+36
| | | | | | | | | | | | In 806e49ae the package imports refactoring code was modified to rename package imports. There was a small oversight which meant the code didn't account for module visibility. This patch fixes that oversight. In general the "lookupPackageName" function is unsafe to use as it doesn't account for package visiblity/thinning/renaming etc, there is just one use in the compiler which would be good to audit. Fixes #20779
* Perf: remove an indirection when fetching the unique maskSylvain Henry2021-12-071-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Slight decrease but still noticeable on CI: Baseline Test Metric value New value Change ----------------------------------------------------------------------------- ManyAlternatives(normal) ghc/alloc 747607676.0 747458936.0 -0.0% ManyConstructors(normal) ghc/alloc 4003722296.0 4003530032.0 -0.0% MultiLayerModules(normal) ghc/alloc 3064539560.0 3063984552.0 -0.0% MultiLayerModulesRecomp(normal) ghc/alloc 894700016.0 894700624.0 +0.0% PmSeriesG(normal) ghc/alloc 48410952.0 48262496.0 -0.3% PmSeriesS(normal) ghc/alloc 61561848.0 61415768.0 -0.2% PmSeriesT(normal) ghc/alloc 90975784.0 90829360.0 -0.2% PmSeriesV(normal) ghc/alloc 60405424.0 60259008.0 -0.2% T10421(normal) ghc/alloc 113275928.0 113137168.0 -0.1% T10421a(normal) ghc/alloc 79195676.0 79050112.0 -0.2% T10547(normal) ghc/alloc 28720176.0 28710008.0 -0.0% T10858(normal) ghc/alloc 180992412.0 180857400.0 -0.1% T11195(normal) ghc/alloc 283452220.0 283293832.0 -0.1% T11276(normal) ghc/alloc 137882128.0 137745840.0 -0.1% T11303b(normal) ghc/alloc 44453956.0 44309184.0 -0.3% T11374(normal) ghc/alloc 248118668.0 247979880.0 -0.1% T11545(normal) ghc/alloc 971994728.0 971852696.0 -0.0% T11822(normal) ghc/alloc 131544864.0 131399024.0 -0.1% T12150(optasm) ghc/alloc 79336468.0 79191888.0 -0.2% T12227(normal) ghc/alloc 495064180.0 494943040.0 -0.0% T12234(optasm) ghc/alloc 57198468.0 57053568.0 -0.3% T12425(optasm) ghc/alloc 90928696.0 90793440.0 -0.1% T12545(normal) ghc/alloc 1695417772.0 1695275744.0 -0.0% T12707(normal) ghc/alloc 956258984.0 956138864.0 -0.0% T13035(normal) ghc/alloc 102279484.0 102132616.0 -0.1% T13056(optasm) ghc/alloc 367196556.0 367066408.0 -0.0% T13253(normal) ghc/alloc 334365844.0 334255264.0 -0.0% T13253-spj(normal) ghc/alloc 125474884.0 125328672.0 -0.1% T13379(normal) ghc/alloc 359185604.0 359036960.0 -0.0% T13701(normal) ghc/alloc 2403026480.0 2402677464.0 -0.0% T13719(normal) ghc/alloc 4192234752.0 4192039448.0 -0.0% T14052(ghci) ghc/alloc 2745868552.0 2747706176.0 +0.1% T14052Type(ghci) ghc/alloc 7335937964.0 7336283280.0 +0.0% T14683(normal) ghc/alloc 2992557736.0 2992436872.0 -0.0% T14697(normal) ghc/alloc 363391248.0 363222920.0 -0.0% T15164(normal) ghc/alloc 1292578008.0 1292434240.0 -0.0% T15304(normal) ghc/alloc 1279603472.0 1279465944.0 -0.0% T15630(normal) ghc/alloc 161707776.0 161602632.0 -0.1% T16190(normal) ghc/alloc 276904644.0 276555264.0 -0.1% T16577(normal) ghc/alloc 7573033016.0 7572982752.0 -0.0% T16875(normal) ghc/alloc 34937980.0 34796592.0 -0.4% T17096(normal) ghc/alloc 287436348.0 287299368.0 -0.0% T17516(normal) ghc/alloc 1714727484.0 1714617664.0 -0.0% T17836(normal) ghc/alloc 1091095748.0 1090958168.0 -0.0% T17836b(normal) ghc/alloc 52467912.0 52321296.0 -0.3% T17977(normal) ghc/alloc 44971660.0 44826480.0 -0.3% T17977b(normal) ghc/alloc 40941128.0 40793160.0 -0.4% T18140(normal) ghc/alloc 82363124.0 82213056.0 -0.2% T18223(normal) ghc/alloc 1168448128.0 1168333624.0 -0.0% T18282(normal) ghc/alloc 131577844.0 131440400.0 -0.1% T18304(normal) ghc/alloc 86988664.0 86844432.0 -0.2% T18478(normal) ghc/alloc 742992400.0 742871136.0 -0.0% T18698a(normal) ghc/alloc 337654412.0 337526792.0 -0.0% T18698b(normal) ghc/alloc 398840772.0 398716472.0 -0.0% T18923(normal) ghc/alloc 68964992.0 68818768.0 -0.2% T1969(normal) ghc/alloc 764285884.0 764156168.0 -0.0% T19695(normal) ghc/alloc 1395577984.0 1395552552.0 -0.0% T20049(normal) ghc/alloc 89159032.0 89012952.0 -0.2% T3064(normal) ghc/alloc 191194856.0 191051816.0 -0.1% T3294(normal) ghc/alloc 1604762016.0 1604656488.0 -0.0% T4801(normal) ghc/alloc 296829368.0 296687824.0 -0.0% T5030(normal) ghc/alloc 364720540.0 364580152.0 -0.0% T5321FD(normal) ghc/alloc 271090004.0 270950824.0 -0.1% T5321Fun(normal) ghc/alloc 301244320.0 301102960.0 -0.0% T5631(normal) ghc/alloc 576154548.0 576022904.0 -0.0% T5642(normal) ghc/alloc 471105876.0 470967552.0 -0.0% T5837(normal) ghc/alloc 36328620.0 36186720.0 -0.4% T6048(optasm) ghc/alloc 103125988.0 102981024.0 -0.1% T783(normal) ghc/alloc 386945556.0 386795984.0 -0.0% T9020(optasm) ghc/alloc 247835012.0 247696704.0 -0.1% T9198(normal) ghc/alloc 47556208.0 47413784.0 -0.3% T9233(normal) ghc/alloc 682210596.0 682069960.0 -0.0% T9630(normal) ghc/alloc 1429689648.0 1429581168.0 -0.0% T9675(optasm) ghc/alloc 431092812.0 430943192.0 -0.0% T9872a(normal) ghc/alloc 1705052592.0 1705042064.0 -0.0% T9872b(normal) ghc/alloc 2180406760.0 2180395784.0 -0.0% T9872c(normal) ghc/alloc 1760508464.0 1760497936.0 -0.0% T9872d(normal) ghc/alloc 501517968.0 501309464.0 -0.0% T9961(normal) ghc/alloc 354037204.0 353891576.0 -0.0% TcPlugin_RewritePerf(normal) ghc/alloc 2381708520.0 2381550824.0 -0.0% WWRec(normal) ghc/alloc 589553520.0 589407216.0 -0.0% hard_hole_fits(normal) ghc/alloc 492122188.0 492470648.0 +0.1% hie002(normal) ghc/alloc 9336434800.0 9336443496.0 +0.0% parsing001(normal) ghc/alloc 537680944.0 537659824.0 -0.0% geo. mean -0.1%
* generalize GHC.Cmm.Dataflow to work over any node typeNorman Ramsey2021-12-071-31/+46
| | | | | | See #20725. The commit includes source-code changes and a test case.
* More permissive parsing of higher-rank type IPssheaf2021-12-071-1/+1
| | | | | | | | | | | | | | | | | The parser now accepts implicit parameters with higher-rank types, such as `foo :: (?ip :: forall a. a -> a) => ...` Before this patch, we instead insisted on parentheses like so: `foo :: (?ip :: (forall a. a -> a)) => ...` The rest of the logic surrounding implicit parameters is unchanged; in particular, even with ImpredicativeTypes, this idiom is not likely to be very useful. Fixes #20654
* compiler: Eliminate accidental loop in GHC.SysTools.BaseDirBen Gamari2021-12-071-2/+2
| | | | | | | | As noted in #20757, `GHC.SysTools.BaseDir.findToolDir` previously contained an loop, which would be triggered in the case that the search failed. Closes #20757.
* Revert "Data.List specialization to []"Matthew Pickering2021-12-031-4/+1
| | | | | | | | | | This reverts commit bddecda1a4c96da21e3f5211743ce5e4c78793a2. This implements the first step in the plan formulated in #20025 to improve the communication and migration strategy for the proposed changes to Data.List. Requires changing the haddock submodule to update the test output.
* Require all dirty_MUT_VAR callers to do explicit stg_MUT_VAR_CLEAN_info ↵nineonine2021-12-022-5/+15
| | | | comparison (#20088)
* CmmToC: Cast possibly-signed results as unsignedBen Gamari2021-12-021-10/+40
| | | | | | | C11 rule 6.3.1.1 dictates that all small integers used in expressions be implicitly converted to `signed int`. However, Cmm semantics require that the width of the operands be preserved with zero-extension semantics. For this reason we must recast sub-word arithmetic results as unsigned.
* CmmToC: Always cast arguments as unsignedBen Gamari2021-12-021-4/+28
| | | | | | | As noted in Note [When in doubt, cast arguments as unsigned], we must ensure that arguments have the correct signedness since some operations (e.g. `%`) have different semantics depending upon signedness.
* CmmToC: Zero-extend sub-word size resultsBen Gamari2021-12-021-6/+41
| | | | | As noted in Note [Zero-extending sub-word signed results] we must explicitly zero-extend the results of sub-word-sized signed operations.
* CmmToC: Fix width of shift operationsBen Gamari2021-12-021-3/+15
| | | | | Under C's implicit widening rules, the result of an operation like (a >> b) where a::Word8 and b::Word will have type Word, yet we want Word.
* nativeGen/aarch64: Fix handling of subword valuesBen Gamari2021-12-023-83/+228
| | | | | | | | | | | | | | | | Here we rework the handling of sub-word operations in the AArch64 backend, fixing a number of bugs and inconsistencies. In short, we now impose the invariant that all subword values are represented in registers in zero-extended form. Signed arithmetic operations are then responsible for sign-extending as necessary. Possible future work: * Use `CMP`s extended register form to avoid burning an instruction in sign-extending the second operand. * Track sign-extension state of registers to elide redundant sign extensions in blocks with frequent sub-word signed arithmetic.
* cmm/opt: Fold away shifts larger than shiftee widthBen Gamari2021-12-021-2/+12
| | | | | This is necessary for lint-correctness since we no longer allow such shifts in Cmm.
* nativeGen/aarch64: Don't rely on register width to determine amodeBen Gamari2021-12-021-12/+16
| | | | | We might be loading, e.g., a 16- or 8-bit value, in which case the register width is not reflective of the loaded element size.
* cmm: Disallow shifts larger than shifteeBen Gamari2021-12-022-0/+19
| | | | | | | | | | | | Previously primops.txt.pp stipulated that the word-size shift primops were only defined for shift offsets in [0, word_size). However, there was no further guidance for the definition of Cmm's sub-word size shift MachOps. Here we fix this by explicitly disallowing (checked in many cases by CmmLint) shift operations where the shift offset is larger than the shiftee. This is consistent with LLVM's shift operations, avoiding the miscompilation noted in #20637.
* ncg/aarch64: Don't sign extend loadsBen Gamari2021-12-021-2/+2
| | | | | Previously we would emit the sign-extending LDS[HB] instructions for sub-word loads. However, this is wrong, as noted in #20638.
* cmm: narrow when folding signed quotientsBen Gamari2021-12-021-2/+2
| | | | | | | | | | | | | | | | | | Previously the constant-folding behavior for MO_S_Quot and MO_S_Rem failed to narrow its arguments, meaning that a program like: %zx64(%quot(%lobits8(0x00e1::bits16), 3::bits8)) would be miscompiled. Specifically, this program should reduce as %lobits8(0x00e1::bits16) == -31 %quot(%lobits8(0x00e1::bits16), 3::bits8) == -10 %zx64(%quot(%lobits8(0x00e1::bits16), 3::bits8)) == 246 However, with this bug the `%lobits8(0x00e1::bits16)` would instead be treated as `+31`, resulting in the incorrect result of `75`. (cherry picked from commit 94e197e3dbb9a48991eb90a03b51ea13d39ba4cc)
* nativeGen/x86: Don't encode large shift offsetsBen Gamari2021-12-021-1/+10
| | | | | | | | | | | Handle the case of a shift larger than the width of the shifted value. This is necessary since x86 applies a mask of 0x1f to the shift amount, meaning that, e.g., `shr 47, $eax` will actually shift by 47 & 0x1f == 15. See #20626. (cherry picked from commit 31370f1afe1e2f071b3569fb5ed4a115096127ca)
* Dump non-module specific info to file #20316Carrie Xu2021-12-013-16/+15
| | | | | | | | | - Change the dumpPrefix to FilePath, and default to non-module - Add dot to seperate dump-file-prefix and suffix - Modify user guide to introduce how dump files are named - This commit does not affect Ghci dump file naming. See also #17500
* rts: Refactor SRT representation selectionBen Gamari2021-12-012-9/+31
| | | | | | The goal here is to make the SRT selection logic a bit clearer and allow configurations which we currently don't support (e.g. using a full word in the info table even when TNTC is used).
* Use Monoid in hptSomeThingsBelowUsSylvain Henry2021-11-291-18/+15
| | | | | | | | | | It seems to have a moderate but good impact on perf tests in CI. In particular: MultiLayerModules(normal) ghc/alloc 3125771138.7 3065532240.0 -1.9% So it's likely that huge projects will benefit from this.
* TTG: replace Void/NoExtCon with DataConCantHappenKrzysztof Gogolewski2021-11-2913-147/+134
| | | | | | | | There were two ways to indicate that a TTG constructor is unused in a phase: `NoExtCon` and `Void`. This unifies the code, and uses the name 'DataConCantHappen', following the discussion at MR 7041. Updates haddock submodule
* Make ambient MinGW support a proper settingsJohn Ericson2021-11-274-26/+39
| | | | | | | Get rid of `USE_INPLACE_MINGW_TOOLCHAIN` and use a settings file entry instead. The CPP setting was originally introduced in f065b6b012.
* Allow keywords which can be used as variables to be used with ↵Matthew Pickering2021-11-261-1/+1
| | | | | | | | | | | | | OverloadedDotSyntax There are quite a few keywords which are allowed to be used as variables. Such as "as", "dependency" etc. These weren't accepted by OverloadedDotSyntax. The fix is pretty simple, use the varid production rather than raw VARID. Fixes #20723
* Reflect type change in the haddock commentKai Prott2021-11-261-1/+1
|
* Fix plugin type to GHC.Plugins.PluginKai Prott2021-11-261-1/+1
|
* Simplify printQualificationKai Prott2021-11-261-7/+2
|
* Remove duplicate importKai Prott2021-11-261-1/+0
|
* Improve error message for mis-typed plugins #20671Kai Prott2021-11-262-15/+25
| | | | Previously, when a plugin could not be loaded because it was incorrectly typed, the error message only printed the expected but not the actual type. This commit augments the error message such that both types are printed and the corresponding module is printed as well.
* compiler: Mark GHC.Prelude as Haddock no-homeBen Gamari2021-11-251-0/+1
| | | | This significantly improves Haddock documentation generated by nix.
* Allow boring class declarations in hs-boot filessheaf2021-11-251-1/+9
| | | | | | | | | | | | | | | | | | | | | | | There are two different ways of declaring a class in an hs-boot file: - a full declaration, where everything is written as it is in the .hs file, - an abstract declaration, where class methods and superclasses are left out. However, a declaration with no methods and a trivial superclass, such as: class () => C a was erroneously considered to be an abstract declaration, because the superclass is trivial. This is remedied by a one line fix in GHC.Tc.TyCl.tcClassDecl1. This patch also further clarifies the documentation around class declarations in hs-boot files. Fixes #20661, #20588.
* Add `llvmOptLevel` to `DynFlags` (#20500)Gergo ERDI2021-11-252-5/+7
|
* Add specific optimization flag for Cmm control flow analysis (#20500)Gergo ERDI2021-11-253-1/+4
|
* Add specific optimization flag for fast PAP calls (#6084, #20500)Gergo ERDI2021-11-253-1/+4
|
* Use `simplify` in non-optimizing build pipeline (#20500)Gergo ERDI2021-11-251-4/+2
|
View Lorry Controller Status