- ghc-commits - Haskell.org

[Git][ghc/ghc][wip/spj-try-opt-coercion] Wibble flag naming
by Simon Peyton Jones (＠simonpj) 25 May '26

25 May '26

Simon Peyton Jones pushed to branch wip/spj-try-opt-coercion at Glasgow Haskell Compiler / GHC Commits: 70f8849d by Simon Peyton Jones at 2026-05-25T15:24:06+01:00 Wibble flag naming - - - - - 1 changed file: - compiler/GHC/Driver/Session.hs Changes: ===================================== compiler/GHC/Driver/Session.hs ===================================== @@ -2503,8 +2503,8 @@ fFlagsDeps = [ flagSpec "cmm-sink" Opt_CmmSink, flagSpec "cmm-static-pred" Opt_CmmStaticPred, flagSpec "opt-coercion" Opt_OptCoercion, - flagSpec "opt-relco-simpleopt" Opt_OptReflCoSimpleOpt, - flagSpec "opt-relco-simplifier" Opt_OptReflCoSimplifier, + flagSpec "opt-reflco-simpleopt" Opt_OptReflCoSimpleOpt, + flagSpec "opt-reflco-simplifier" Opt_OptReflCoSimplifier, flagSpec "cse" Opt_CSE, flagSpec "stg-cse" Opt_StgCSE, flagSpec "stg-lift-lams" Opt_StgLiftLams, View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/70f8849dd2587ccd800b4d9472720e1… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/70f8849dd2587ccd800b4d9472720e1… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/fix-hpc-bytecode-modname] ghci: fix module name string lifetime in hs_hpc_module invocation
by Cheng Shao (＠TerrorJack) 25 May '26

25 May '26

Cheng Shao pushed to branch wip/fix-hpc-bytecode-modname at Glasgow Haskell Compiler / GHC Commits: 704164be by Cheng Shao at 2026-05-25T13:59:10+00:00 ghci: fix module name string lifetime in hs_hpc_module invocation This patch makes hpcAddModule pass a properly malloced module name string to hs_hpc_module, instead of using useAsCString which causes use-after-free of module name string. Fixes #27297. Co-authored-by: Codex <codex(a)openai.com> - - - - - 4 changed files: - + libraries/ghc-boot/GHC/Data/ShortByteString.hs - libraries/ghc-boot/ghc-boot.cabal.in - libraries/ghci/GHCi/Coverage.hs - libraries/ghci/GHCi/Run.hs Changes: ===================================== libraries/ghc-boot/GHC/Data/ShortByteString.hs ===================================== @@ -0,0 +1,17 @@ +module GHC.Data.ShortByteString + ( newCStringFromSBS + ) where + +import Prelude + +import qualified Data.ByteString.Short as SBS +import Foreign +import Foreign.C + +newCStringFromSBS :: SBS.ShortByteString -> IO CString +newCStringFromSBS sbs = + SBS.useAsCStringLen sbs $ \(src, len) -> do + dst <- mallocBytes (len + 1) + copyBytes dst src len + pokeByteOff dst len (0 :: Word8) + pure dst ===================================== libraries/ghc-boot/ghc-boot.cabal.in ===================================== @@ -51,6 +51,7 @@ Library exposed-modules: GHC.BaseDir + GHC.Data.ShortByteString GHC.Data.ShortText GHC.Data.SizedSeq GHC.Data.SmallArray ===================================== libraries/ghci/GHCi/Coverage.hs ===================================== @@ -9,9 +9,9 @@ import Prelude -- See note [Why do we import Prelude here?] import Control.Exception import Data.ByteString.Short (ShortByteString) -import qualified Data.ByteString.Short as SBS import Data.Word import Foreign +import GHC.Data.ShortByteString import GHC.Foreign (CString) import GHC.Utils.Encoding.UTF8 (utf8DecodeShortByteString) import GHCi.ObjLink (lookupSymbol) @@ -31,17 +31,19 @@ hpcAddModule :: -- ^ Name of the ticks array found in the c-stub. IO () hpcAddModule modlName ticks hash tickboxes = do - SBS.useAsCString modlName $ \modlNameLiteral -> do - -- we need to find the reference to the ticks array. - lookupSymbol tickboxes >>= \ case - Nothing -> do - -- the symbol is not found, this is a bug! - throwIO $ ErrorCall $ "hpcAddModule: failed to find symbol " <> utf8DecodeShortByteString tickboxes - Just tickBoxRef -> do - -- Calling 'hs_hpc_module' multiple times is safe, it will add the module only once. - hpc_register_module modlNameLiteral (fromIntegral ticks) (fromIntegral hash) (castPtr tickBoxRef) - -- calling 'hpc_startup' multiple times is safe, it will only be initialised once. - hpc_startup + -- we need to find the reference to the ticks array. + lookupSymbol tickboxes >>= \ case + Nothing -> do + -- the symbol is not found, this is a bug! + throwIO $ ErrorCall $ "hpcAddModule: failed to find symbol " <> utf8DecodeShortByteString tickboxes + Just tickBoxRef -> do + -- hs_hpc_module stores the module name pointer in the RTS hash table + -- until exitHpc, so pass a malloced C string. + modlNameLiteral <- newCStringFromSBS modlName + -- Calling 'hs_hpc_module' multiple times is safe, it will add the module only once. + hpc_register_module modlNameLiteral (fromIntegral ticks) (fromIntegral hash) (castPtr tickBoxRef) + -- calling 'hpc_startup' multiple times is safe, it will only be initialised once. + hpc_startup foreign import ccall unsafe "hs_hpc_module" hpc_register_module :: CString -> Word32 -> Word32 -> Ptr Word64 -> IO () ===================================== libraries/ghci/GHCi/Run.hs ===================================== @@ -37,6 +37,9 @@ import Control.Monad import Data.ByteString (ByteString) import qualified Data.ByteString.Short.Internal as BS import qualified Data.ByteString.Unsafe as B +#if defined(PROFILING) +import GHC.Data.ShortByteString +#endif import GHC.Exts import qualified GHC.Exts.Heap as Heap import GHC.Stack @@ -447,13 +450,6 @@ mkCostCentres mod ccs = do c_srcspan <- newCStringFromSBS srcspan toRemotePtr <$> c_mkCostCentre c_name c_module c_srcspan - newCStringFromSBS sbs = do - let len = BS.length sbs - buf <- mallocBytes $ len + 1 - BS.copyToPtr sbs 0 buf (fromIntegral len) - pokeByteOff buf len (0 :: Word8) - pure buf - foreign import ccall unsafe "mkCostCentre" c_mkCostCentre :: Ptr CChar -> Ptr CChar -> Ptr CChar -> IO (Ptr CostCentre) #else View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/704164be09e80185ce78727a4569f8c… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/704164be09e80185ce78727a4569f8c… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/fix-hpc-bytecode-modname
by Cheng Shao (＠TerrorJack) 25 May '26

25 May '26

Cheng Shao pushed new branch wip/fix-hpc-bytecode-modname at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/fix-hpc-bytecode-modname You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/mangoiv/ghc-9.12-bp] linker: add --optimistic-linking flag
by Magnus (＠MangoIV) 25 May '26

25 May '26

Magnus pushed to branch wip/mangoiv/ghc-9.12-bp at Glasgow Haskell Compiler / GHC Commits: 11ce0b56 by doyougnu at 2026-05-25T14:42:22+02:00 linker: add --optimistic-linking flag This patch adds: - the --optimistic-linking flag which binds unknown symbols in the runtime linker to 0xDEADBEEF instead of exiting with failure - The test T25240 which tests these flags using dead code in the FFI system. - closes #25240 This patch is part of the upstreaming haskell.nix patches project. (cherry picked from commit 25121dbc57724498776c0d46f810612f4ca81cc4) - - - - - 20 changed files: - docs/users_guide/runtime_control.rst - libraries/ghc-internal/src/GHC/Internal/RTS/Flags.hsc - rts/Linker.c - rts/RtsFlags.c - rts/include/rts/Flags.h - rts/linker/elf_got.c - + testsuite/tests/ghci/linking/T25240/Makefile - + testsuite/tests/ghci/linking/T25240/T25240.hs - + testsuite/tests/ghci/linking/T25240/T25240.stderr - + testsuite/tests/ghci/linking/T25240/T25240a.hs - + testsuite/tests/ghci/linking/T25240/all.T - testsuite/tests/ghci/should_run/T18064.stderr - testsuite/tests/interface-stability/base-exports.stdout - testsuite/tests/interface-stability/base-exports.stdout-javascript-unknown-ghcjs - testsuite/tests/interface-stability/base-exports.stdout-mingw32 - testsuite/tests/interface-stability/base-exports.stdout-ws-32 - testsuite/tests/rts/linker/T11223/T11223_link_order_a_b_2_fail.stderr - testsuite/tests/rts/linker/T11223/T11223_link_order_a_b_2_fail.stderr-ws-64-mingw32 - testsuite/tests/rts/linker/T11223/T11223_simple_duplicate_lib.stderr - testsuite/tests/rts/linker/T11223/T11223_simple_duplicate_lib.stderr-ws-64-mingw32 Changes: ===================================== docs/users_guide/runtime_control.rst ===================================== @@ -397,6 +397,11 @@ Miscellaneous RTS options heap larger than 1T. ``-xr`` is a no-op if GHC is configured with ``--disable-large-address-space`` or if the platform is 32-bit. +.. rts-flag:: --optimistic-linking + + If given, instruct the runtime linker to try to continue linking in the + presence of an unresolved symbol. + .. _rts-options-gc: RTS options to control the garbage collector ===================================== libraries/ghc-internal/src/GHC/Internal/RTS/Flags.hsc ===================================== @@ -162,6 +162,7 @@ data MiscFlags = MiscFlags , disableDelayedOsMemoryReturn :: Bool , internalCounters :: Bool , linkerAlwaysPic :: Bool + , linkerOptimistic :: Bool , linkerMemBase :: Word -- ^ address to ask the OS for memory for the linker, 0 ==> off , ioManager :: IoManagerFlag @@ -536,6 +537,8 @@ getMiscFlags = do (#{peek MISC_FLAGS, internalCounters} ptr :: IO CBool)) <*> (toBool <$> (#{peek MISC_FLAGS, linkerAlwaysPic} ptr :: IO CBool)) + <*> (toBool <$> + (#{peek MISC_FLAGS, linkerOptimistic} ptr :: IO CBool)) <*> #{peek MISC_FLAGS, linkerMemBase} ptr <*> (toEnum . fromIntegral <$> (#{peek MISC_FLAGS, ioManager} ptr :: IO Word32)) ===================================== rts/Linker.c ===================================== @@ -980,10 +980,20 @@ SymbolAddr* lookupSymbol( SymbolName* lbl ) // lookupDependentSymbol directly. SymbolAddr* r = lookupDependentSymbol(lbl, NULL, NULL); if (!r) { - errorBelch("^^ Could not load '%s', dependency unresolved. " - "See top entry above.\n", lbl); - IF_DEBUG(linker, printLoadedObjects()); - fflush(stderr); + if (!RtsFlags.MiscFlags.linkerOptimistic) { + errorBelch("^^ Could not load '%s', dependency unresolved. " + "See top entry above. You might consider using --optimistic-linking\n", + lbl); + IF_DEBUG(linker, printLoadedObjects()); + fflush(stderr); + } else { + // if --optimistic-linking is passed into the RTS we allow the linker + // to optimistically continue + errorBelch("^^ Could not load '%s', dependency unresolved, " + "optimistically continuing\n", + lbl); + r = (void*) 0xDEADBEEF; + } } if (!runPendingInitializers()) { ===================================== rts/RtsFlags.c ===================================== @@ -269,6 +269,7 @@ void initRtsFlagsDefaults(void) RtsFlags.MiscFlags.disableDelayedOsMemoryReturn = false; RtsFlags.MiscFlags.internalCounters = false; RtsFlags.MiscFlags.linkerAlwaysPic = DEFAULT_LINKER_ALWAYS_PIC; + RtsFlags.MiscFlags.linkerOptimistic = false; RtsFlags.MiscFlags.linkerMemBase = 0; RtsFlags.MiscFlags.ioManager = IO_MNGR_FLAG_AUTO; #if defined(THREADED_RTS) && defined(mingw32_HOST_OS) @@ -998,6 +999,11 @@ error = true; OPTION_UNSAFE; RtsFlags.MiscFlags.generate_dump_file = true; } + else if (strequal("optimistic-linking", + &rts_argv[arg][2])) { + OPTION_UNSAFE; + RtsFlags.MiscFlags.linkerOptimistic = true; + } else if (strequal("null-eventlog-writer", &rts_argv[arg][2])) { OPTION_UNSAFE; ===================================== rts/include/rts/Flags.h ===================================== @@ -267,6 +267,7 @@ typedef struct _MISC_FLAGS { there as well. */ bool internalCounters; /* See Note [Internal Counters Stats] */ bool linkerAlwaysPic; /* Assume the object code is always PIC */ + bool linkerOptimistic; /* Should the runtime linker optimistically continue */ StgWord linkerMemBase; /* address to ask the OS for memory * for the linker, NULL ==> off */ IO_MANAGER_FLAG ioManager; /* The I/O manager to use. */ ===================================== rts/linker/elf_got.c ===================================== @@ -97,9 +97,22 @@ fillGot(ObjectCode * oc) { if(0 == strncmp(symbol->name,"_GLOBAL_OFFSET_TABLE_",21)) { symbol->addr = oc->info->got_start; } else { - errorBelch("Failed to lookup symbol: %s\n", + errorBelch("Failed to lookup symbol: %s," + " you might consider using --optimistic-linking\n", symbol->name); - return EXIT_FAILURE; + + // if --optimistic-linking is passed into the + // RTS we allow the linker to optimistically + // continue + if (RtsFlags.MiscFlags.linkerOptimistic) { + errorBelch("Failed to lookup symbol: %s," + " optimistically continuing.\n", + symbol->name); + symbol->addr = (void*) 0xDEADBEEF; + } else { + return EXIT_FAILURE; + } + } } } else { ===================================== testsuite/tests/ghci/linking/T25240/Makefile ===================================== @@ -0,0 +1,7 @@ +TOP=../../../.. +include $(TOP)/mk/boilerplate.mk +include $(TOP)/mk/test.mk + +.PHONY: T25240 +T25240: + "$(TEST_HC)" $(TEST_HC_OPTS_INTERACTIVE) T25240a.hs T25240.hs +RTS --optimistic-linking -RTS ===================================== testsuite/tests/ghci/linking/T25240/T25240.hs ===================================== @@ -0,0 +1,17 @@ +{-# LANGUAGE TemplateHaskell #-} + +{- + +When the RTS linker loads the T25240a module to run the pure foo splice, it +tries to resolve the func symbol even if this function isn't required to run the +splice code, i.e., its dead code. This test checks that by passing the +--optimistic-linking flag the RTS linker continues to link even in the presence +of unknown symbols. + +-} + +module T25240 where + +import T25240a + +$(pure foo) ===================================== testsuite/tests/ghci/linking/T25240/T25240.stderr ===================================== @@ -0,0 +1,2 @@ +ghc: ^^ Could not load 'func', dependency unresolved, optimistically continuing + ===================================== testsuite/tests/ghci/linking/T25240/T25240a.hs ===================================== @@ -0,0 +1,13 @@ +{-# LANGUAGE TemplateHaskell #-} + + +module T25240a + ( foo, func + ) where + + +foo :: [a] +foo = [] + +foreign import ccall "func" + func :: Int -> Int ===================================== testsuite/tests/ghci/linking/T25240/all.T ===================================== @@ -0,0 +1,3 @@ +# skip on darwin because the leading underscores will make the test fail +test('T25240', [when(leading_underscore(),skip), req_rts_linker, extra_files(['T25240a.hs'])], + makefile_test, ['T25240']) ===================================== testsuite/tests/ghci/should_run/T18064.stderr ===================================== @@ -1,2 +1,2 @@ -<interactive>: ^^ Could not load 'blah', dependency unresolved. See top entry above. +<interactive>: ^^ Could not load 'blah', dependency unresolved. See top entry above. You might consider using --optimistic-linking ===================================== testsuite/tests/interface-stability/base-exports.stdout ===================================== @@ -9120,7 +9120,7 @@ module GHC.RTS.Flags where type IoSubSystem :: * data IoSubSystem = IoPOSIX | IoNative type MiscFlags :: * - data MiscFlags = MiscFlags {tickInterval :: RtsTime, installSignalHandlers :: GHC.Types.Bool, installSEHHandlers :: GHC.Types.Bool, generateCrashDumpFile :: GHC.Types.Bool, generateStackTrace :: GHC.Types.Bool, machineReadable :: GHC.Types.Bool, disableDelayedOsMemoryReturn :: GHC.Types.Bool, internalCounters :: GHC.Types.Bool, linkerAlwaysPic :: GHC.Types.Bool, linkerMemBase :: GHC.Types.Word, ioManager :: IoManagerFlag, numIoWorkerThreads :: GHC.Internal.Word.Word32} + data MiscFlags = MiscFlags {tickInterval :: RtsTime, installSignalHandlers :: GHC.Types.Bool, installSEHHandlers :: GHC.Types.Bool, generateCrashDumpFile :: GHC.Types.Bool, generateStackTrace :: GHC.Types.Bool, machineReadable :: GHC.Types.Bool, disableDelayedOsMemoryReturn :: GHC.Types.Bool, internalCounters :: GHC.Types.Bool, linkerAlwaysPic :: GHC.Types.Bool, linkerOptimistic :: GHC.Types.Bool, linkerMemBase :: GHC.Types.Word, ioManager :: IoManagerFlag, numIoWorkerThreads :: GHC.Internal.Word.Word32} type ParFlags :: * data ParFlags = ParFlags {nCapabilities :: GHC.Internal.Word.Word32, migrate :: GHC.Types.Bool, maxLocalSparks :: GHC.Internal.Word.Word32, parGcEnabled :: GHC.Types.Bool, parGcGen :: GHC.Internal.Word.Word32, parGcLoadBalancingEnabled :: GHC.Types.Bool, parGcLoadBalancingGen :: GHC.Internal.Word.Word32, parGcNoSyncWithIdle :: GHC.Internal.Word.Word32, parGcThreads :: GHC.Internal.Word.Word32, setAffinity :: GHC.Types.Bool} type ProfFlags :: * ===================================== testsuite/tests/interface-stability/base-exports.stdout-javascript-unknown-ghcjs ===================================== @@ -12162,7 +12162,7 @@ module GHC.RTS.Flags where type IoSubSystem :: * data IoSubSystem = IoPOSIX | IoNative type MiscFlags :: * - data MiscFlags = MiscFlags {tickInterval :: RtsTime, installSignalHandlers :: GHC.Types.Bool, installSEHHandlers :: GHC.Types.Bool, generateCrashDumpFile :: GHC.Types.Bool, generateStackTrace :: GHC.Types.Bool, machineReadable :: GHC.Types.Bool, disableDelayedOsMemoryReturn :: GHC.Types.Bool, internalCounters :: GHC.Types.Bool, linkerAlwaysPic :: GHC.Types.Bool, linkerMemBase :: GHC.Types.Word, ioManager :: IoManagerFlag, numIoWorkerThreads :: GHC.Internal.Word.Word32} + data MiscFlags = MiscFlags {tickInterval :: RtsTime, installSignalHandlers :: GHC.Types.Bool, installSEHHandlers :: GHC.Types.Bool, generateCrashDumpFile :: GHC.Types.Bool, generateStackTrace :: GHC.Types.Bool, machineReadable :: GHC.Types.Bool, disableDelayedOsMemoryReturn :: GHC.Types.Bool, internalCounters :: GHC.Types.Bool, linkerAlwaysPic :: GHC.Types.Bool, linkerOptimistic :: GHC.Types.Bool, linkerMemBase :: GHC.Types.Word, ioManager :: IoManagerFlag, numIoWorkerThreads :: GHC.Internal.Word.Word32} type ParFlags :: * data ParFlags = ParFlags {nCapabilities :: GHC.Internal.Word.Word32, migrate :: GHC.Types.Bool, maxLocalSparks :: GHC.Internal.Word.Word32, parGcEnabled :: GHC.Types.Bool, parGcGen :: GHC.Internal.Word.Word32, parGcLoadBalancingEnabled :: GHC.Types.Bool, parGcLoadBalancingGen :: GHC.Internal.Word.Word32, parGcNoSyncWithIdle :: GHC.Internal.Word.Word32, parGcThreads :: GHC.Internal.Word.Word32, setAffinity :: GHC.Types.Bool} type ProfFlags :: * ===================================== testsuite/tests/interface-stability/base-exports.stdout-mingw32 ===================================== @@ -9344,7 +9344,7 @@ module GHC.RTS.Flags where type IoSubSystem :: * data IoSubSystem = IoPOSIX | IoNative type MiscFlags :: * - data MiscFlags = MiscFlags {tickInterval :: RtsTime, installSignalHandlers :: GHC.Types.Bool, installSEHHandlers :: GHC.Types.Bool, generateCrashDumpFile :: GHC.Types.Bool, generateStackTrace :: GHC.Types.Bool, machineReadable :: GHC.Types.Bool, disableDelayedOsMemoryReturn :: GHC.Types.Bool, internalCounters :: GHC.Types.Bool, linkerAlwaysPic :: GHC.Types.Bool, linkerMemBase :: GHC.Types.Word, ioManager :: IoManagerFlag, numIoWorkerThreads :: GHC.Internal.Word.Word32} + data MiscFlags = MiscFlags {tickInterval :: RtsTime, installSignalHandlers :: GHC.Types.Bool, installSEHHandlers :: GHC.Types.Bool, generateCrashDumpFile :: GHC.Types.Bool, generateStackTrace :: GHC.Types.Bool, machineReadable :: GHC.Types.Bool, disableDelayedOsMemoryReturn :: GHC.Types.Bool, internalCounters :: GHC.Types.Bool, linkerAlwaysPic :: GHC.Types.Bool, linkerOptimistic :: GHC.Types.Bool, linkerMemBase :: GHC.Types.Word, ioManager :: IoManagerFlag, numIoWorkerThreads :: GHC.Internal.Word.Word32} type ParFlags :: * data ParFlags = ParFlags {nCapabilities :: GHC.Internal.Word.Word32, migrate :: GHC.Types.Bool, maxLocalSparks :: GHC.Internal.Word.Word32, parGcEnabled :: GHC.Types.Bool, parGcGen :: GHC.Internal.Word.Word32, parGcLoadBalancingEnabled :: GHC.Types.Bool, parGcLoadBalancingGen :: GHC.Internal.Word.Word32, parGcNoSyncWithIdle :: GHC.Internal.Word.Word32, parGcThreads :: GHC.Internal.Word.Word32, setAffinity :: GHC.Types.Bool} type ProfFlags :: * ===================================== testsuite/tests/interface-stability/base-exports.stdout-ws-32 ===================================== @@ -9120,7 +9120,7 @@ module GHC.RTS.Flags where type IoSubSystem :: * data IoSubSystem = IoPOSIX | IoNative type MiscFlags :: * - data MiscFlags = MiscFlags {tickInterval :: RtsTime, installSignalHandlers :: GHC.Types.Bool, installSEHHandlers :: GHC.Types.Bool, generateCrashDumpFile :: GHC.Types.Bool, generateStackTrace :: GHC.Types.Bool, machineReadable :: GHC.Types.Bool, disableDelayedOsMemoryReturn :: GHC.Types.Bool, internalCounters :: GHC.Types.Bool, linkerAlwaysPic :: GHC.Types.Bool, linkerMemBase :: GHC.Types.Word, ioManager :: IoManagerFlag, numIoWorkerThreads :: GHC.Internal.Word.Word32} + data MiscFlags = MiscFlags {tickInterval :: RtsTime, installSignalHandlers :: GHC.Types.Bool, installSEHHandlers :: GHC.Types.Bool, generateCrashDumpFile :: GHC.Types.Bool, generateStackTrace :: GHC.Types.Bool, machineReadable :: GHC.Types.Bool, disableDelayedOsMemoryReturn :: GHC.Types.Bool, internalCounters :: GHC.Types.Bool, linkerAlwaysPic :: GHC.Types.Bool, linkerOptimistic :: GHC.Types.Bool, linkerMemBase :: GHC.Types.Word, ioManager :: IoManagerFlag, numIoWorkerThreads :: GHC.Internal.Word.Word32} type ParFlags :: * data ParFlags = ParFlags {nCapabilities :: GHC.Internal.Word.Word32, migrate :: GHC.Types.Bool, maxLocalSparks :: GHC.Internal.Word.Word32, parGcEnabled :: GHC.Types.Bool, parGcGen :: GHC.Internal.Word.Word32, parGcLoadBalancingEnabled :: GHC.Types.Bool, parGcLoadBalancingGen :: GHC.Internal.Word.Word32, parGcNoSyncWithIdle :: GHC.Internal.Word.Word32, parGcThreads :: GHC.Internal.Word.Word32, setAffinity :: GHC.Types.Bool} type ProfFlags :: * ===================================== testsuite/tests/rts/linker/T11223/T11223_link_order_a_b_2_fail.stderr ===================================== @@ -9,7 +9,7 @@ This could be caused by: * Specifying the same object file twice on the GHCi command line * An incorrect `package.conf' entry, causing some object to be loaded twice. -ghc-stage2: ^^ Could not load 'c', dependency unresolved. See top entry above. +ghc-stage2: ^^ Could not load 'c', dependency unresolved. See top entry above. You might consider using --optimistic-linking GHC.ByteCode.Linker: can't find label ===================================== testsuite/tests/rts/linker/T11223/T11223_link_order_a_b_2_fail.stderr-ws-64-mingw32 ===================================== @@ -9,7 +9,7 @@ This could be caused by: * Specifying the same object file twice on the GHCi command line * An incorrect `package.conf' entry, causing some object to be loaded twice. -ghc-stage2.exe: ^^ Could not load 'c', dependency unresolved. See top entry above. +ghc-stage2.exe: ^^ Could not load 'c', dependency unresolved. See top entry above. You might consider using --optimistic-linking GHC.ByteCode.Linker: can't find label ===================================== testsuite/tests/rts/linker/T11223/T11223_simple_duplicate_lib.stderr ===================================== @@ -9,7 +9,7 @@ This could be caused by: * Specifying the same object file twice on the GHCi command line * An incorrect `package.conf' entry, causing some object to be loaded twice. -ghc-stage2: ^^ Could not load 'c', dependency unresolved. See top entry above. +ghc-stage2: ^^ Could not load 'c', dependency unresolved. See top entry above. You might consider using --optimistic-linking GHC.ByteCode.Linker: can't find label ===================================== testsuite/tests/rts/linker/T11223/T11223_simple_duplicate_lib.stderr-ws-64-mingw32 ===================================== @@ -9,7 +9,7 @@ This could be caused by: * Specifying the same object file twice on the GHCi command line * An incorrect `package.conf' entry, causing some object to be loaded twice. -ghc-stage2.exe: ^^ Could not load 'c', dependency unresolved. See top entry above. +ghc-stage2.exe: ^^ Could not load 'c', dependency unresolved. See top entry above. You might consider using --optimistic-linking GHC.ByteCode.Linker: can't find label View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/11ce0b56c671b0939625cb208a48e7c… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/11ce0b56c671b0939625cb208a48e7c… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/mangoiv/process-bump] 2 commits: libraries/process: bump submodule to v1.6.29.0
by Magnus (＠MangoIV) 25 May '26

25 May '26

Magnus pushed to branch wip/mangoiv/process-bump at Glasgow Haskell Compiler / GHC Commits: 2c79a63e by mangoiv at 2026-05-25T14:32:20+02:00 libraries/process: bump submodule to v1.6.29.0 This submodule bump resolves a segfault on macos 15. Fixes #27144 - - - - - d3569e83 by mangoiv at 2026-05-25T14:32:20+02:00 libraries/unix: in submodule, don't pick branch 2.7 The 2.7 branch is outdated and the module has been advanced far beyond it anyway, so remove that line. - - - - - 3 changed files: - .gitmodules - + changelog.d/bump-process - libraries/process Changes: ===================================== .gitmodules ===================================== @@ -82,7 +82,6 @@ path = libraries/unix url = https://gitlab.haskell.org/ghc/packages/unix.git ignore = untracked - branch = 2.7 [submodule "libraries/semaphore-compat"] path = libraries/semaphore-compat url = https://gitlab.haskell.org/ghc/semaphore-compat.git ===================================== changelog.d/bump-process ===================================== @@ -0,0 +1,8 @@ +section: packaging +issues: #27144 +mrs: !16096 +synopsis: + bump submodule to v1.6.29.0 +description: + This submodule bump resolves a segfault on macos 15 with + certain command line SDK versions. ===================================== libraries/process ===================================== @@ -1 +1 @@ -Subproject commit 72e5b7c75a17f543262674259b2ebf4a3bda390c +Subproject commit 92deb52c1781bf10ad390296dbc435abe103bfe4 View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/31576d9b05d3bec1244caac7dfa5a1… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/31576d9b05d3bec1244caac7dfa5a1… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/marge_bot_batch_merge_job] 2 commits: Ensure that SetOps.{minusList,unionListsOrd} can be specialized
by Marge Bot (＠marge-bot) 25 May '26

25 May '26

Marge Bot pushed to branch wip/marge_bot_batch_merge_job at Glasgow Haskell Compiler / GHC Commits: 156f09b8 by Simon Jakobi at 2026-05-25T08:30:50-04:00 Ensure that SetOps.{minusList,unionListsOrd} can be specialized ...by marking them INLINABLE. Haddock allocates 0.1–0.3% less as a result. This also removes some redundant constraints on unionListsOrd. - - - - - 6a76e7aa by Cheng Shao at 2026-05-25T08:30:50-04:00 wasm: ensure post-linker output is synchronous ESM This patch fixes wasm backend's post-linker output script to ensure it's synchronous ESM and doesn't use top-level await, which doesn't work in ServiceWorkers. Fixes #27257. - - - - - 3 changed files: - + changelog.d/wasm-fix-serviceworker - compiler/GHC/Data/List/SetOps.hs - utils/jsffi/prelude.mjs Changes: ===================================== changelog.d/wasm-fix-serviceworker ===================================== @@ -0,0 +1,4 @@ +section: wasm-backend +synopsis: Ensure post-linker output is synchronous ESM and fix loading in ServiceWorker +issues: #27257 +mrs: !16093 ===================================== compiler/GHC/Data/List/SetOps.hs ===================================== @@ -60,12 +60,13 @@ getNth xs n = assertPpr (xs `lengthExceeds` n) (ppr n $$ ppr xs) $ -- -- Uses a set internally to record duplicates. This makes it slightly slower for -- very small lists but avoids quadratic behaviour for large lists. -unionListsOrd :: (HasDebugCallStack, Outputable a, Ord a) => [a] -> [a] -> [a] +unionListsOrd :: Ord a => [a] -> [a] -> [a] unionListsOrd xs ys - -- Since both arguments don't have internal duplicates we can just take all of xs - -- and every element of ys that's not already in xs. + -- Since both arguments don't have internal duplicates we can just take all of ys + -- and every element of xs that's not already in ys. = let set_ys = S.fromList ys in (filter (\e -> not $ S.member e set_ys) xs) ++ ys +{-# INLINABLE unionListsOrd #-} -- Ensure the function can be specialized. -- | Assumes that the arguments contain no duplicates unionLists :: (HasDebugCallStack, Outputable a, Eq a) => [a] -> [a] -> [a] @@ -108,6 +109,7 @@ minusList xs [y] = filter (/= y) xs minusList xs ys = filter (`S.notMember` yss) xs where yss = S.fromList ys +{-# INLINABLE minusList #-} -- Ensure the function can be specialized. {- ************************************************************************ ===================================== utils/jsffi/prelude.mjs ===================================== @@ -39,19 +39,13 @@ export class JSValManager { // To benchmark different setImmediate() implementations in the // browser, use https://github.com/jphpsf/setImmediate-shim-demo as a // starting point. -export const setImmediate = await (async () => { - // node, bun, or other scripts might have set this up in the browser +export const setImmediate = (() => { + // node, deno, bun, or other scripts might have set this up in the + // browser if (globalThis.setImmediate) { return globalThis.setImmediate; } - // deno - if (globalThis.Deno) { - try { - return (await import("node:timers")).setImmediate; - } catch {} - } - // https://developer.mozilla.org/en-US/docs/Web/API/Scheduler/postTask if (globalThis.scheduler) { return (cb, ...args) => scheduler.postTask(() => cb(...args)); View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/4d8d8e379bb06b8b0416d8e4422908… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/4d8d8e379bb06b8b0416d8e4422908… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/mangoiv/ghc-9.12-bp] AArch64: use ASR not LSR for MO_U_Shr at W8/W16
by Magnus (＠MangoIV) 25 May '26

25 May '26

Magnus pushed to branch wip/mangoiv/ghc-9.12-bp at Glasgow Haskell Compiler / GHC Commits: 82c521c6 by Ian Duncan at 2026-05-25T13:39:24+02:00 AArch64: use ASR not LSR for MO_U_Shr at W8/W16 The unsigned right shift (MO_U_Shr) for sub-word widths (W8, W16) with a variable shift amount was emitting ASR (arithmetic/signed shift right) after zero-extending with UXTB/UXTH. This should be LSR (logical/unsigned shift right). After zero-extension the upper bits happen to be 0 so ASR produces the same result, but it is semantically wrong and would break if the zero-extension were ever optimized away. Includes assembly output test (grep for lsr) and runtime test verifying unsigned right shift of Word8 and Word16 values. (cherry picked from commit 50188615342098345fc2822ea223ab23791bbf49) - - - - - 12 changed files: - + changelog.d/T26979 - compiler/GHC/CmmToAsm/AArch64/CodeGen.hs - + testsuite/tests/codeGen/should_gen_asm/aarch64-shl-subword.asm - + testsuite/tests/codeGen/should_gen_asm/aarch64-shl-subword.hs - + testsuite/tests/codeGen/should_gen_asm/aarch64-ushr-subword.asm - + testsuite/tests/codeGen/should_gen_asm/aarch64-ushr-subword.hs - testsuite/tests/codeGen/should_gen_asm/all.T - + testsuite/tests/codeGen/should_run/aarch64-subword-ops.hs - + testsuite/tests/codeGen/should_run/aarch64-subword-ops.stdout - + testsuite/tests/codeGen/should_run/aarch64-ushr-subword-run.hs - + testsuite/tests/codeGen/should_run/aarch64-ushr-subword-run.stdout - testsuite/tests/codeGen/should_run/all.T Changes: ===================================== changelog.d/T26979 ===================================== @@ -0,0 +1,11 @@ +section: compiler +issues: #26979 +mrs: !15620 +synopsis: + On AArch64, use an arithmetic instead of a logical right shift when + sign extending at 8/16 bit word size. +description: + The unsigned right shift ``MO_U_Shr`` for sub-word widths (``W8``, ``W16``) + with a variable shift amount now correctly emits ``LSR`` instead of ``ASR``. + ``ASR`` was semantically wrong, and would break if the zero-extension were + ever optimized away. ===================================== compiler/GHC/CmmToAsm/AArch64/CodeGen.hs ===================================== @@ -525,6 +525,15 @@ opRegWidth w = pprPanic "opRegWidth" (text "Unsupported width" <+> ppr w) -- sub-word-size value always contains the zero-extended form of that value -- in between operations. -- +-- IMPORTANT: this invariant only holds within a single expression tree as +-- generated by the NCG (via truncateReg after each sub-word operation). It +-- does NOT hold at function entry points or across basic block boundaries, +-- because the GHC calling convention does not guarantee that callers +-- zero-extend sub-word arguments. Therefore, any operation that is sensitive +-- to the upper bits of its input (e.g. unsigned right shift, unsigned +-- division) must explicitly zero- or sign-extend its operands rather than +-- assuming they are already extended. +-- -- For instance, consider the program, -- -- test(bits64 buffer) @@ -543,7 +552,7 @@ opRegWidth w = pprPanic "opRegWidth" (text "Unsupported width" <+> ppr w) -- Next we compute `c`: The `%not` requires no extension of its operands, but -- we must still truncate the result back down to 8-bits. Finally the `%shrl` -- requires no extension and no truncate since we can assume that --- `c` is zero-extended. +-- `c` is zero-extended (it was produced by a truncateReg in the same block). -- -- TODO: -- Don't use Width in Operands @@ -877,17 +886,30 @@ getRegister' config plat expr CmmMachOp (MO_U_Quot w) [x, y] | w == W8 -> do (reg_x, _format_x, code_x) <- getSomeReg x (reg_y, _format_y, code_y) <- getSomeReg y - return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` annExpr expr (UXTB (OpReg w reg_x) (OpReg w reg_x)) `snocOL` - (UXTB (OpReg w reg_y) (OpReg w reg_y)) `snocOL` - (UDIV (OpReg w dst) (OpReg w reg_x) (OpReg w reg_y))) + tmp_x <- getNewRegNat (intFormat w) + tmp_y <- getNewRegNat (intFormat w) + return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` annExpr expr (UXTB (OpReg w tmp_x) (OpReg w reg_x)) `snocOL` + (UXTB (OpReg w tmp_y) (OpReg w reg_y)) `snocOL` + (UDIV (OpReg w dst) (OpReg w tmp_x) (OpReg w tmp_y))) CmmMachOp (MO_U_Quot w) [x, y] | w == W16 -> do (reg_x, _format_x, code_x) <- getSomeReg x (reg_y, _format_y, code_y) <- getSomeReg y - return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` annExpr expr (UXTH (OpReg w reg_x) (OpReg w reg_x)) `snocOL` - (UXTH (OpReg w reg_y) (OpReg w reg_y)) `snocOL` - (UDIV (OpReg w dst) (OpReg w reg_x) (OpReg w reg_y))) + tmp_x <- getNewRegNat (intFormat w) + tmp_y <- getNewRegNat (intFormat w) + return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` annExpr expr (UXTH (OpReg w tmp_x) (OpReg w reg_x)) `snocOL` + (UXTH (OpReg w tmp_y) (OpReg w reg_y)) `snocOL` + (UDIV (OpReg w dst) (OpReg w tmp_x) (OpReg w tmp_y))) -- 2. Shifts. x << n, x >> n. + -- Sub-word left shifts by a constant: use UBFM (UBFIZ alias) to shift + -- and mask in a single instruction. See Note [Signed arithmetic on AArch64]. + CmmMachOp (MO_Shl w) [x, (CmmLit (CmmInt n _))] | w == W8, 0 <= n, n < 8 -> do + (reg_x, _format_x, code_x) <- getSomeReg x + return $ Any (intFormat w) (\dst -> code_x `snocOL` annExpr expr (UBFM (OpReg w dst) (OpReg w reg_x) (OpImm (ImmInteger ((32 - n) `mod` 32))) (OpImm (ImmInteger (7 - n))))) + CmmMachOp (MO_Shl w) [x, (CmmLit (CmmInt n _))] | w == W16, 0 <= n, n < 16 -> do + (reg_x, _format_x, code_x) <- getSomeReg x + return $ Any (intFormat w) (\dst -> code_x `snocOL` annExpr expr (UBFM (OpReg w dst) (OpReg w reg_x) (OpImm (ImmInteger ((32 - n) `mod` 32))) (OpImm (ImmInteger (15 - n))))) + CmmMachOp (MO_Shl w) [x, (CmmLit (CmmInt n _))] | w == W32 || w == W64 , 0 <= n, n < fromIntegral (widthInBits w) -> do @@ -901,8 +923,11 @@ getRegister' config plat expr CmmMachOp (MO_S_Shr w) [x, y] | w == W8 -> do (reg_x, _format_x, code_x) <- getSomeReg x (reg_y, _format_y, code_y) <- getSomeReg y - return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` annExpr expr (SXTB (OpReg w reg_x) (OpReg w reg_x)) `snocOL` - (ASR (OpReg w dst) (OpReg w reg_x) (OpReg w reg_y)) `snocOL` + -- Use a temporary register to avoid sign-extending reg_x in-place, + -- as other operations may use reg_x. + tmp <- getNewRegNat (intFormat w) + return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` annExpr expr (SXTB (OpReg w tmp) (OpReg w reg_x)) `snocOL` + (ASR (OpReg w dst) (OpReg w tmp) (OpReg w reg_y)) `snocOL` (UXTB (OpReg w dst) (OpReg w dst))) -- See Note [Signed arithmetic on AArch64] CmmMachOp (MO_S_Shr w) [x, (CmmLit (CmmInt n _))] | w == W16, 0 <= n, n < 16 -> do @@ -912,8 +937,11 @@ getRegister' config plat expr CmmMachOp (MO_S_Shr w) [x, y] | w == W16 -> do (reg_x, _format_x, code_x) <- getSomeReg x (reg_y, _format_y, code_y) <- getSomeReg y - return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` annExpr expr (SXTH (OpReg w reg_x) (OpReg w reg_x)) `snocOL` - (ASR (OpReg w dst) (OpReg w reg_x) (OpReg w reg_y)) `snocOL` + -- Use a temporary register to avoid sign-extending reg_x in-place, + -- as other operations may use reg_x. + tmp <- getNewRegNat (intFormat w) + return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` annExpr expr (SXTH (OpReg w tmp) (OpReg w reg_x)) `snocOL` + (ASR (OpReg w dst) (OpReg w tmp) (OpReg w reg_y)) `snocOL` (UXTH (OpReg w dst) (OpReg w dst))) -- See Note [Signed arithmetic on AArch64] CmmMachOp (MO_S_Shr w) [x, (CmmLit (CmmInt n _))] @@ -928,8 +956,8 @@ getRegister' config plat expr CmmMachOp (MO_U_Shr w) [x, y] | w == W8 -> do (reg_x, _format_x, code_x) <- getSomeReg x (reg_y, _format_y, code_y) <- getSomeReg y - return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` annExpr expr (UXTB (OpReg w reg_x) (OpReg w reg_x)) `snocOL` - (ASR (OpReg w dst) (OpReg w reg_x) (OpReg w reg_y))) + tmp <- getNewRegNat (intFormat w) + return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` UXTB (OpReg w tmp) (OpReg w reg_x) `snocOL` annExpr expr (LSR (OpReg w dst) (OpReg w tmp) (OpReg w reg_y))) CmmMachOp (MO_U_Shr w) [x, (CmmLit (CmmInt n _))] | w == W16, 0 <= n, n < 16 -> do (reg_x, _format_x, code_x) <- getSomeReg x @@ -937,8 +965,8 @@ getRegister' config plat expr CmmMachOp (MO_U_Shr w) [x, y] | w == W16 -> do (reg_x, _format_x, code_x) <- getSomeReg x (reg_y, _format_y, code_y) <- getSomeReg y - return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` annExpr expr (UXTH (OpReg w reg_x) (OpReg w reg_x)) - `snocOL` (ASR (OpReg w dst) (OpReg w reg_x) (OpReg w reg_y))) + tmp <- getNewRegNat (intFormat w) + return $ Any (intFormat w) (\dst -> code_x `appOL` code_y `snocOL` UXTH (OpReg w tmp) (OpReg w reg_x) `snocOL` annExpr expr (LSR (OpReg w dst) (OpReg w tmp) (OpReg w reg_y))) CmmMachOp (MO_U_Shr w) [x, (CmmLit (CmmInt n _))] | w == W32 || w == W64 ===================================== testsuite/tests/codeGen/should_gen_asm/aarch64-shl-subword.asm ===================================== @@ -0,0 +1 @@ +ubfm ===================================== testsuite/tests/codeGen/should_gen_asm/aarch64-shl-subword.hs ===================================== @@ -0,0 +1,8 @@ +{-# LANGUAGE MagicHash #-} +module ShlSubWord (shlW8) where + +import GHC.Exts +import GHC.Word + +shlW8 :: Word8 -> Word8 +shlW8 (W8# w) = W8# (uncheckedShiftLWord8# w 4#) ===================================== testsuite/tests/codeGen/should_gen_asm/aarch64-ushr-subword.asm ===================================== @@ -0,0 +1 @@ +lsr ===================================== testsuite/tests/codeGen/should_gen_asm/aarch64-ushr-subword.hs ===================================== @@ -0,0 +1,9 @@ +{-# LANGUAGE MagicHash #-} +module UShrSubWord (ushrW8) where + +import GHC.Exts +import GHC.Word + +ushrW8 :: Word8 -> Int -> Word8 +ushrW8 x n = x `shiftR` n + where shiftR (W8# w) (I# i) = W8# (wordToWord8# (word8ToWord# w `uncheckedShiftRL#` i)) ===================================== testsuite/tests/codeGen/should_gen_asm/all.T ===================================== @@ -12,3 +12,11 @@ test('bytearray-memcpy-unroll', is_amd64_codegen, compile_grep_asm, ['hs', True, test('T18137', [when(opsys('darwin'), skip), only_ways(llvm_ways)], compile_grep_asm, ['hs', False, '-fllvm -split-sections']) test('T24941', [only_ways(['optasm'])], compile, ['-fregs-graph']) +is_aarch64_codegen = [ + unless(arch('aarch64'), skip), + when(unregisterised(), skip), +] + +# AArch64-specific tests +test('aarch64-ushr-subword', is_aarch64_codegen, compile_grep_asm, ['hs', True, '-O']) +test('aarch64-shl-subword', is_aarch64_codegen, compile_grep_asm, ['hs', True, '-O']) ===================================== testsuite/tests/codeGen/should_run/aarch64-subword-ops.hs ===================================== @@ -0,0 +1,115 @@ +{-# LANGUAGE MagicHash #-} +module Main where + +import GHC.Exts +import GHC.Word +import GHC.Int + +-- Uses sub-word primops directly so that the NCG sees MO_Shl W8, +-- MO_U_Shr W8, MO_S_Shr W8 etc. (the Bits class widens to Word#/Int#). + +-- NOINLINE to prevent constant folding. + +-- MO_U_Shr W8/W16 variable shift +{-# NOINLINE ushrW8 #-} +ushrW8 :: Word8 -> Int -> Word8 +ushrW8 (W8# w) (I# i) = W8# (uncheckedShiftRLWord8# w i) + +{-# NOINLINE ushrW16 #-} +ushrW16 :: Word16 -> Int -> Word16 +ushrW16 (W16# w) (I# i) = W16# (uncheckedShiftRLWord16# w i) + +-- MO_S_Shr W8/W16 variable shift +{-# NOINLINE sshrI8 #-} +sshrI8 :: Int8 -> Int -> Int8 +sshrI8 (I8# x) (I# i) = I8# (uncheckedShiftRAInt8# x i) + +{-# NOINLINE sshrI16 #-} +sshrI16 :: Int16 -> Int -> Int16 +sshrI16 (I16# x) (I# i) = I16# (uncheckedShiftRAInt16# x i) + +-- MO_Shl W8/W16 variable shift +{-# NOINLINE shlW8 #-} +shlW8 :: Word8 -> Int -> Word8 +shlW8 (W8# w) (I# i) = W8# (uncheckedShiftLWord8# w i) + +{-# NOINLINE shlW16 #-} +shlW16 :: Word16 -> Int -> Word16 +shlW16 (W16# w) (I# i) = W16# (uncheckedShiftLWord16# w i) + +-- quot exercising MO_U_Quot W8/W16 +{-# NOINLINE quotW8 #-} +quotW8 :: Word8 -> Word8 -> Word8 +quotW8 (W8# x) (W8# y) = W8# (quotWord8# x y) + +{-# NOINLINE quotW16 #-} +quotW16 :: Word16 -> Word16 -> Word16 +quotW16 (W16# x) (W16# y) = W16# (quotWord16# x y) + +-- Register clobbering: use a value both in a shift/quot and afterward. +-- If the sign/zero extension clobbers the source register, the second +-- use sees the wrong value. + +{-# NOINLINE sshrAndAdd8 #-} +sshrAndAdd8 :: Int8 -> Int -> Int8 +sshrAndAdd8 a n = sshrI8 a n + a + +{-# NOINLINE sshrAndAdd16 #-} +sshrAndAdd16 :: Int16 -> Int -> Int16 +sshrAndAdd16 a n = sshrI16 a n + a + +{-# NOINLINE quotAndAdd8 #-} +quotAndAdd8 :: Word8 -> Word8 -> Word8 +quotAndAdd8 a b = quotW8 a b + a + b + +{-# NOINLINE quotAndAdd16 #-} +quotAndAdd16 :: Word16 -> Word16 -> Word16 +quotAndAdd16 a b = quotW16 a b + a + b + +main :: IO () +main = do + putStrLn "-- MO_U_Shr variable shift" + print (ushrW8 0x80 1) -- 64 + print (ushrW8 0xFF 4) -- 15 + print (ushrW8 0x42 0) -- 66 + print (ushrW16 0x8000 1) -- 16384 + print (ushrW16 0xFFFF 8) -- 255 + print (ushrW16 0x1234 0) -- 4660 + + putStrLn "-- MO_S_Shr variable shift" + print (sshrI8 (-1) 1) -- -1 + print (sshrI8 (-128) 1) -- -64 + print (sshrI8 127 1) -- 63 + print (sshrI8 0x42 3) -- 8 + print (sshrI16 (-1) 1) -- -1 + print (sshrI16 (-32768) 1) -- -16384 + print (sshrI16 32767 8) -- 127 + + putStrLn "-- MO_Shl variable shift" + print (shlW8 0x01 0) -- 1 + print (shlW8 0x01 4) -- 16 + print (shlW8 0xFF 1) -- 254 + print (shlW8 0x42 3) -- 16 + print (shlW16 0x0001 0) -- 1 + print (shlW16 0x0001 8) -- 256 + print (shlW16 0xFFFF 1) -- 65534 + print (shlW16 0x1234 4) -- 9024 + + putStrLn "-- MO_U_Quot" + print (quotW8 255 10) -- 25 + print (quotW8 200 7) -- 28 + print (quotW8 1 1) -- 1 + print (quotW16 65535 256) -- 255 + print (quotW16 1000 3) -- 333 + + putStrLn "-- register clobbering: shift + reuse" + print (sshrAndAdd8 (-128) 1) -- 64 (wraps: -64 + -128 = -192 = 64 as Int8) + print (sshrAndAdd8 0x42 1) -- 99 + print (sshrAndAdd16 (-32768) 1) -- 16384 (wraps) + print (sshrAndAdd16 0x1234 4) -- 4951 + + putStrLn "-- register clobbering: quot + reuse" + print (quotAndAdd8 200 7) -- 235 + print (quotAndAdd8 255 10) -- 34 (wraps: 290 mod 256) + print (quotAndAdd16 1000 3) -- 1336 + print (quotAndAdd16 65535 256) -- 510 (wraps: 66046 mod 65536) ===================================== testsuite/tests/codeGen/should_run/aarch64-subword-ops.stdout ===================================== @@ -0,0 +1,40 @@ +-- MO_U_Shr variable shift +64 +15 +66 +16384 +255 +4660 +-- MO_S_Shr variable shift +-1 +-64 +63 +8 +-1 +-16384 +127 +-- MO_Shl variable shift +1 +16 +254 +16 +1 +256 +65534 +9024 +-- MO_U_Quot +25 +28 +1 +255 +333 +-- register clobbering: shift + reuse +64 +99 +16384 +4951 +-- register clobbering: quot + reuse +235 +34 +1336 +510 ===================================== testsuite/tests/codeGen/should_run/aarch64-ushr-subword-run.hs ===================================== @@ -0,0 +1,9 @@ +import Data.Bits (shiftR) +import Data.Word (Word8, Word16) + +main :: IO () +main = do + print (shiftR (0x80 :: Word8) 1) + print (shiftR (0xFF :: Word8) 4) + print (shiftR (0x8000 :: Word16) 1) + print (shiftR (0xFFFF :: Word16) 8) ===================================== testsuite/tests/codeGen/should_run/aarch64-ushr-subword-run.stdout ===================================== @@ -0,0 +1,4 @@ +64 +15 +16384 +255 ===================================== testsuite/tests/codeGen/should_run/all.T ===================================== @@ -275,3 +275,6 @@ test('T27072d', [req_c, only_ways(['dyn']), when(not opsys('darwin'), skip), # Skipped on Darwin (Apple linker doesn't support --wrap). test('T27072w', [req_c, js_skip, when(opsys('darwin'), skip)], compile_and_run, ['T27072w_c.c -no-hs-main -optl-Wl,--wrap=hs_spt_remove']) +# AArch64-specific runtime tests +test('aarch64-ushr-subword-run', [unless(arch('aarch64'), skip)], compile_and_run, ['-O']) +test('aarch64-subword-ops', [unless(arch('aarch64'), skip)], compile_and_run, ['-O']) View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/82c521c62b1842672e6f411750df4c6… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/82c521c62b1842672e6f411750df4c6… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/spj-try-opt-coercion] Wibbles
by Simon Peyton Jones (＠simonpj) 25 May '26

25 May '26

Simon Peyton Jones pushed to branch wip/spj-try-opt-coercion at Glasgow Haskell Compiler / GHC Commits: 7574a932 by Simon Peyton Jones at 2026-05-25T13:20:20+01:00 Wibbles - - - - - 1 changed file: - compiler/GHC/Core/SimpleOpt.hs Changes: ===================================== compiler/GHC/Core/SimpleOpt.hs ===================================== @@ -35,7 +35,7 @@ import GHC.Core.Type hiding ( substTy, extendTvSubst, extendCvSubst, extendTvSub , isInScope, substTyVarBndr, cloneTyVarBndr ) import GHC.Core.Predicate( isCoVarType ) import GHC.Core.Coercion hiding ( substCo, substCoVarBndr ) -import GHC.Core.Coercion.Opt( optCoRefl, optTransCo ) +import GHC.Core.Coercion.Opt( optCoRefl ) import GHC.Types.Literal import GHC.Types.Id @@ -308,13 +308,13 @@ type SimpleClo = (SimpleOptEnv, InExpr) data SimpleContItem = ApplyToArg SimpleClo - | CastIt OutCoercion OutType + | CastIt OutCoercion -- The OutType is the corecionRKind of the coercion -- Used to make reflexivity checking more efficient instance Outputable SimpleContItem where ppr (ApplyToArg (_, arg)) = text "ARG" <+> ppr arg - ppr (CastIt co _) = text "CAST" <+> ppr co + ppr (CastIt co) = text "CAST" <+> ppr co data SimpleOptEnv = SOE { soe_opts :: {-# UNPACK #-} !SimpleOpts @@ -488,7 +488,7 @@ simple_app env e0@(Lam {}) as0@(_:_) where (env', b') = subst_opt_bndr env b -- See Note [Eliminate casts in function position] - do_beta env e@(Lam b _) as@(CastIt out_co _ : rest) + do_beta env e@(Lam b _) as@(CastIt out_co : rest) | isNonCoVarId b -- Optimise the inner lambda to make it an 'OutExpr', which makes it -- possible to call 'pushCoercionIntoLambda' with the 'OutCoercion' 'co'. @@ -563,18 +563,10 @@ add_cast env co1 as = as | otherwise = case as of - CastIt co2 _ : rest --- ToDo: get rid of the type field in CastIt? --- | ty2 `eqTypeIgnoringMultiplicity` coercionLKind opt_co1 --- -> rest --- ToDo: do we want to call optTransCo here? --- | otherwise -> CastIt (optTransCo in_scope opt_co1 co2) tyR : rest - | otherwise -> CastIt (mkTransCo opt_co1 co2) tyR : rest - _ -> CastIt opt_co1 tyR : as + CastIt co2 : rest -> CastIt (mkTransCo opt_co1 co2) : rest + _ -> CastIt opt_co1 : as where - in_scope = soeInScope env opt_co1 = simple_opt_co env co1 - tyR = coercionRKind opt_co1 rebuild_app :: HasDebugCallStack => SimpleOptEnv -> OutExpr -> [SimpleContItem] -> OutExpr @@ -583,7 +575,7 @@ rebuild_app env fun args = foldl mk_app fun args in_scope = soeInScope env mk_app out_fun = \case ApplyToArg arg -> App out_fun (simple_opt_clo in_scope arg) - CastIt co _ -> mk_cast out_fun co + CastIt co -> mk_cast out_fun co mk_cast :: CoreExpr -> CoercionR -> CoreExpr -- Does a full reflexivity check, unlike GHC.Core.Utils.mkCast, View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/7574a9326107b5a3df9ab1deb8e5c4c… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/7574a9326107b5a3df9ab1deb8e5c4c… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc][wip/mangoiv/process-bump] 2 commits: libraries/process: bump submodule to v1.6.29.0
by Magnus (＠MangoIV) 25 May '26

25 May '26

Magnus pushed to branch wip/mangoiv/process-bump at Glasgow Haskell Compiler / GHC Commits: 842285f5 by mangoiv at 2026-05-25T14:12:52+02:00 libraries/process: bump submodule to v1.6.29.0 This submodule bump resolves a segfault on macos 15. Fixes #27144 - - - - - 31576d9b by mangoiv at 2026-05-25T14:12:52+02:00 libraries/unix: in submodule, don't pick branch 2.7 The 2.7 branch is outdated and the module has been advanced far beyond it anyway, so remove that line. - - - - - 3 changed files: - .gitmodules - + changelog.d/bump-process - libraries/process Changes: ===================================== .gitmodules ===================================== @@ -82,7 +82,6 @@ path = libraries/unix url = https://gitlab.haskell.org/ghc/packages/unix.git ignore = untracked - branch = 2.7 [submodule "libraries/semaphore-compat"] path = libraries/semaphore-compat url = https://gitlab.haskell.org/ghc/semaphore-compat.git ===================================== changelog.d/bump-process ===================================== @@ -0,0 +1,8 @@ +section: libraries +issues: #27144 +mrs: !16096 +synopsis: + bump submodule to v1.6.29.0 +description: + This submodule bump resolves a segfault on macos 15 with + certain command line SDK versions. ===================================== libraries/process ===================================== @@ -1 +1 @@ -Subproject commit 72e5b7c75a17f543262674259b2ebf4a3bda390c +Subproject commit 92deb52c1781bf10ad390296dbc435abe103bfe4 View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/4d9ea2cd2ed566318d03d84b15755e… -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/4d9ea2cd2ed566318d03d84b15755e… You're receiving this email because of your account on gitlab.haskell.org.

1 0

[Git][ghc/ghc] Pushed new branch wip/sjakobi/multi-caret
by Simon Jakobi (＠sjakobi2) 25 May '26

25 May '26

Simon Jakobi pushed new branch wip/sjakobi/multi-caret at Glasgow Haskell Compiler / GHC -- View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/tree/wip/sjakobi/multi-caret You're receiving this email because of your account on gitlab.haskell.org.

1 0