September 2013 - ghc-tickets

[GHC] #8060: Undefined symbols when using Template Haskell linked from another object with unexposed modules
by GHC 08 Jan '16

08 Jan '16

#8060: Undefined symbols when using Template Haskell linked from another object with unexposed modules -----------------------------------+--------------------------------------- Reporter: tvynr | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Linux Architecture: | Type of failure: Compile-time crash Unknown/Multiple | Test Case: Difficulty: Unknown | Blocking: Blocked By: | Related Tickets: | -----------------------------------+--------------------------------------- I have constructed two Cabal projects named "first" and "second". The "first" project defines three files: Logging.hs (which provides some basic logging functionality) and FirstExample.hs and FirstDependent.hs (the latter of which imports the former). In the cabal file, only Logging and FirstDependent are exposed modules. FirstDependent uses a Template Haskell function "loggingFunctions" which defines logging functions for that module. Using cabal or cabal-dev to build this package works quite nicely. The second project depends on the first project and contains a single file, Second.hs. This file imports Logging but neither of the other two files. It also uses the loggingFunctions routine to generate local logging functions. The presence of the "$(loggingFunctions)" splice in Second.hs makes the problem visible, but the real problem is in the first project. Upon compilation of the "second" project with cabal or cabal- dev, the following error is generated: {{{ cabal-dev -s ../cabal-dev-sandbox --with-cabal-install=cabal configure Resolving dependencies... Configuring second-0.1... cabal-dev -s ../cabal-dev-sandbox --with-cabal-install=cabal build Building second-0.1... Preprocessing library second-0.1... [1 of 1] Compiling Second ( lib-src/Second.hs, dist/build/Second.o ) Loading package ghc-prim ... linking ... done. Loading package integer-gmp ... linking ... done. Loading package base ... linking ... done. Loading package array-0.4.0.1 ... linking ... done. Loading package deepseq-1.3.0.1 ... linking ... done. Loading package containers-0.5.0.0 ... linking ... done. Loading package filepath-1.3.0.1 ... linking ... done. Loading package old-locale-1.0.0.5 ... linking ... done. Loading package time-1.4.0.1 ... linking ... done. Loading package bytestring-0.10.0.2 ... linking ... done. Loading package unix-2.6.0.1 ... linking ... done. Loading package directory-1.2.0.1 ... linking ... done. Loading package transformers-0.3.0.0 ... linking ... done. Loading package mtl-2.1.2 ... linking ... done. Loading package text-0.11.3.1 ... linking ... done. Loading package parsec-3.1.3 ... linking ... done. Loading package network-2.4.1.2 ... linking ... done. Loading package process-1.1.0.2 ... linking ... done. Loading package hslogger-1.2.1 ... linking ... done. Loading package pretty-1.1.1.0 ... linking ... done. Loading package template-haskell ... linking ... done. Loading package first-0.2 ... linking ... ghc: /home/some- directories/example/first/../cabal-dev- sandbox//lib/first-0.2/ghc-7.6.1.20121207/HSfirst-0.2.o: unknown symbol `firstzm0zi2_FirstExample_zdfShowErrorSpec3_closure' ghc: unable to load package `first-0.2' make[1]: *** [build] Error 1 }}} Upon further investigation, it would seem that the .o file constructed to represent the "first" project has not been completely constructed: {{{ $ objdump -t cabal-dev- sandbox/lib/first-0.2/ghc-7.6.1.20121207/HSfirst-0.2.o | grep '\*UND\*' | grep first 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfOrdErrorSpeczuzdczg_info 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfShowErrorSpec3_closure 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfOrdErrorSpeczuzdczl_info 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfEqErrorSpeczuzdczeze_info 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfOrdErrorSpeczuzdczlze_info 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfOrdErrorSpeczuzdccompare_info 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfOrdErrorSpeczuzdczgze_info }}} (While the above dump is for ghc 7.6.1, this has also been tested on ghc 7.6.3 as well as on both i386- and amd64-based OS installs.) If the "$(loggingFunctions)" splice is removed from Second.hs, then the project builds (presumably because the logic used by non-Template Haskell projects for linking differs from that of Template Haskell projects). Fortunately, it seems that exposing every module of the "first" project resolves this issue. But please note that the FirstExample.hs and FirstDependent.hs files are completely unrelated to the Second.hs file; the latter does not depend on the former in any way. It seems strange that I can't use Template Haskell across projects without exposing all of my project modules. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8060> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #8182: Parser.y.pp needs special treatment with -fcmm-sink
by GHC 01 Jan '16

01 Jan '16

#8182: Parser.y.pp needs special treatment with -fcmm-sink ------------------------------------+------------------------------------- Reporter: thoughtpolice | Owner: thoughtpolice Type: bug | Status: new Priority: highest | Milestone: 7.8.1 Component: Build System | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- This bug is really more of a reminder, but it's critical for the 7.8.1 release. Right now, `Parser.y.pp` uses a very weird hack to pass `-fcmm-sink` to the compiler. Here's the relevant comment: {{{ -- CPP tricks because we want the directives in the output of the -- first CPP pass. -- -- Clang note, 6/17/2013 by aseipp: It is *extremely* important (for -- some reason) that there be a line of whitespace between the two -- definitions here, and the subsequent use of __IF_GHC_77__ - this -- seems to be a bug in clang or something, where having the line of -- whitespace will make the preprocessor correctly format the rendered -- lines in the 'two step' CPP pass. No, this is not a joke. #define __IF_GHC_77__ #if __GLASGOW_HASKELL__ >= 707 #define __ENDIF__ #endif __IF_GHC_77__ -- Required on x86 to avoid the register allocator running out of -- stack slots when compiling this module with -fPIC -dynamic. {-# OPTIONS_GHC -fcmm-sink #-} __ENDIF__ }}} This is because the parser is first preprocessed before being run with GHC (which again preprocesses,) so we want the resulting `#ifdef` in the final `.hs` file. Things to note: * We really shouldn't be doing this, it's amazingly fragile. I think the correct thing to do is to ensure the build system correctly passes `-fcmm- sink` during the stage[2,3] build. * `./configure.ac` needs to check to see if the bootstrapping compiler is `ghc >= 7.7` (or `7.8`) and if it is, ''also'' pass `-fcmm-sink` during the stage1 build. * This hack needs to be removed once we can rely on >= 7.8 for bootstrap. Probably something like the hypothetical 7.12-7.14 timeframe. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8182> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #8347: Add a Strict LANGUAGE pragma
by GHC 16 Dec '15

16 Dec '15

#8347: Add a Strict LANGUAGE pragma ------------------------------------+------------------------------------- Reporter: tibbe | Owner: Type: feature request | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- A `Strict` `LANGUAGE` pragma would allow us to experiment with writing in a strict sub-language. Since this pragma works on a per module basis, it needs to be modular. I propose the following semantics: * Patterns are strict at the top-level, as if they had a top-level bang, unless an explicit ~ is used to make the pattern lazy. * Functions evaluate their parameters to WHNF. * Data type fields are strict (as if they had a bang-pattern). -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8347> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 14

[GHC] #8294: T7478 fails on Mac OS X with "unexpected bindingNone" from ld
by GHC 15 Dec '15

15 Dec '15

#8294: T7478 fails on Mac OS X with "unexpected bindingNone" from ld -----------------------------------+--------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: GHC API | Version: 7.7 Keywords: | Operating System: MacOS X Architecture: | Type of failure: Compile-time crash Unknown/Multiple | Test Case: T7478 Difficulty: Unknown | Blocking: Blocked By: | Related Tickets: | -----------------------------------+--------------------------------------- (Tangentially, the test script was swallowing up relevant output; you need to get rid of the output hook to see any debugging output from GHC proper). Linker step fails: {{{ *** Linker: /usr/bin/gcc -m64 -fno-stack-protector -DTABLES_NEXT_TO_CODE -m64 -v -o A -Wl,-no_compact_unwind C.o -L/Users/ezyang/Dev/ghc-init/libraries/base /dist-install/build -L/Users/ezyang/Dev/ghc-init/libraries/integer-gmp /dist-install/build -L/Users/ezyang/Dev/ghc-init/libraries/ghc-prim/dist- install/build -L/Users/ezyang/Dev/ghc-init/rts/dist/build /var/folders/l8/1lth36c11yg8dv_3kz779_th0000gn/T/ghc26290_0/ghc26290_10.o -lHSbase-4.7.0.0-ghc7.7.20130913 -lHSinteger-gmp-0.5.1.0-ghc7.7.20130913 -lHSghc-prim-0.3.1.0-ghc7.7.20130913 -lHSrts-ghc7.7.20130913 -lffi -liconv -lm -ldl -Wl,-u,_ghczmprim_GHCziTypes_Izh_static_info -Wl,-u,_ghczmprim_GHCziTypes_Czh_static_info -Wl,-u,_ghczmprim_GHCziTypes_Fzh_static_info -Wl,-u,_ghczmprim_GHCziTypes_Dzh_static_info -Wl,-u,_base_GHCziPtr_Ptr_static_info -Wl,-u,_ghczmprim_GHCziTypes_Wzh_static_info -Wl,-u,_base_GHCziInt_I8zh_static_info -Wl,-u,_base_GHCziInt_I16zh_static_info -Wl,-u,_base_GHCziInt_I32zh_static_info -Wl,-u,_base_GHCziInt_I64zh_static_info -Wl,-u,_base_GHCziWord_W8zh_static_info -Wl,-u,_base_GHCziWord_W16zh_static_info -Wl,-u,_base_GHCziWord_W32zh_static_info -Wl,-u,_base_GHCziWord_W64zh_static_info -Wl,-u,_base_GHCziStable_StablePtr_static_info -Wl,-u,_ghczmprim_GHCziTypes_Izh_con_info -Wl,-u,_ghczmprim_GHCziTypes_Czh_con_info -Wl,-u,_ghczmprim_GHCziTypes_Fzh_con_info -Wl,-u,_ghczmprim_GHCziTypes_Dzh_con_info -Wl,-u,_base_GHCziPtr_Ptr_con_info -Wl,-u,_base_GHCziPtr_FunPtr_con_info -Wl,-u,_base_GHCziStable_StablePtr_con_info -Wl,-u,_ghczmprim_GHCziTypes_False_closure -Wl,-u,_ghczmprim_GHCziTypes_True_closure -Wl,-u,_base_GHCziPack_unpackCString_closure -Wl,-u,_base_GHCziIOziException_stackOverflow_closure -Wl,-u,_base_GHCziIOziException_heapOverflow_closure -Wl,-u,_base_ControlziExceptionziBase_nonTermination_closure -Wl,-u,_base_GHCziIOziException_blockedIndefinitelyOnMVar_closure -Wl,-u,_base_GHCziIOziException_blockedIndefinitelyOnSTM_closure -Wl,-u,_base_ControlziExceptionziBase_nestedAtomically_closure -Wl,-u,_base_GHCziWeak_runFinalizzerBatch_closure -Wl,-u,_base_GHCziTopHandler_flushStdHandles_closure -Wl,-u,_base_GHCziTopHandler_runIO_closure -Wl,-u,_base_GHCziTopHandler_runNonIO_closure -Wl,-u,_base_GHCziConcziIO_ensureIOManagerIsRunning_closure -Wl,-u,_base_GHCziConcziIO_ioManagerCapabilitiesChanged_closure -Wl,-u,_base_GHCziConcziSync_runSparks_closure -Wl,-u,_base_GHCziConcziSignal_runHandlers_closure -Wl,-search_paths_first Using built-in specs. Target: i686-apple-darwin11 Configured with: /private/var/tmp/llvmgcc42/llvmgcc42-2336.11~182/src/configure --disable- checking --enable-werror --prefix=/Applications/Xcode.app/Contents/Developer/usr/llvm-gcc-4.2 --mandir=/share/man --enable-languages=c,objc,c++,obj-c++ --program- prefix=llvm- --program-transform-name=/^[cg][^.-]*$/s/$/-4.2/ --with- slibdir=/usr/lib --build=i686-apple-darwin11 --enable- llvm=/private/var/tmp/llvmgcc42/llvmgcc42-2336.11~182/dst- llvmCore/Developer/usr/local --program-prefix=i686-apple-darwin11- --host=x86_64-apple-darwin11 --target=i686-apple-darwin11 --with-gxx- include-dir=/usr/include/c++/4.2.1 Thread model: posix gcc version 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.11.00) /usr/llvm-gcc-4.2/bin/../libexec/gcc/i686-apple-darwin11/4.2.1/collect2 -dynamic -arch x86_64 -macosx_version_min 10.8.3 -weak_reference_mismatches non-weak -o A -lcrt1.10.6.o -L/Users/ezyang/Dev /ghc-init/libraries/base/dist-install/build -L/Users/ezyang/Dev/ghc- init/libraries/integer-gmp/dist-install/build -L/Users/ezyang/Dev/ghc- init/libraries/ghc-prim/dist-install/build -L/Users/ezyang/Dev/ghc- init/rts/dist/build -L/usr/llvm-gcc-4.2/bin/../lib/gcc/i686-apple- darwin11/4.2.1/x86_64 -L/Applications/Xcode.app/Contents/Developer/usr /llvm-gcc-4.2/lib/gcc/i686-apple-darwin11/4.2.1/x86_64 -L/usr/llvm- gcc-4.2/bin/../lib/gcc/i686-apple-darwin11/4.2.1 -L/usr/llvm- gcc-4.2/bin/../lib/gcc -L/Applications/Xcode.app/Contents/Developer/usr /llvm-gcc-4.2/lib/gcc/i686-apple-darwin11/4.2.1 -L/usr/llvm- gcc-4.2/bin/../lib/gcc/i686-apple-darwin11/4.2.1/../../.. -L/Applications/Xcode.app/Contents/Developer/usr/llvm-gcc-4.2/lib/gcc/i686 -apple-darwin11/4.2.1/../../.. -no_compact_unwind C.o /var/folders/l8/1lth36c11yg8dv_3kz779_th0000gn/T/ghc26290_0/ghc26290_10.o -lHSbase-4.7.0.0-ghc7.7.20130913 -lHSinteger-gmp-0.5.1.0-ghc7.7.20130913 -lHSghc-prim-0.3.1.0-ghc7.7.20130913 -lHSrts-ghc7.7.20130913 -lffi -liconv -lm -ldl -u _ghczmprim_GHCziTypes_Izh_static_info -u _ghczmprim_GHCziTypes_Czh_static_info -u _ghczmprim_GHCziTypes_Fzh_static_info -u _ghczmprim_GHCziTypes_Dzh_static_info -u _base_GHCziPtr_Ptr_static_info -u _ghczmprim_GHCziTypes_Wzh_static_info -u _base_GHCziInt_I8zh_static_info -u _base_GHCziInt_I16zh_static_info -u _base_GHCziInt_I32zh_static_info -u _base_GHCziInt_I64zh_static_info -u _base_GHCziWord_W8zh_static_info -u _base_GHCziWord_W16zh_static_info -u _base_GHCziWord_W32zh_static_info -u _base_GHCziWord_W64zh_static_info -u _base_GHCziStable_StablePtr_static_info -u _ghczmprim_GHCziTypes_Izh_con_info -u _ghczmprim_GHCziTypes_Czh_con_info -u _ghczmprim_GHCziTypes_Fzh_con_info -u _ghczmprim_GHCziTypes_Dzh_con_info -u _base_GHCziPtr_Ptr_con_info -u _base_GHCziPtr_FunPtr_con_info -u _base_GHCziStable_StablePtr_con_info -u _ghczmprim_GHCziTypes_False_closure -u _ghczmprim_GHCziTypes_True_closure -u _base_GHCziPack_unpackCString_closure -u _base_GHCziIOziException_stackOverflow_closure -u _base_GHCziIOziException_heapOverflow_closure -u _base_ControlziExceptionziBase_nonTermination_closure -u _base_GHCziIOziException_blockedIndefinitelyOnMVar_closure -u _base_GHCziIOziException_blockedIndefinitelyOnSTM_closure -u _base_ControlziExceptionziBase_nestedAtomically_closure -u _base_GHCziWeak_runFinalizzerBatch_closure -u _base_GHCziTopHandler_flushStdHandles_closure -u _base_GHCziTopHandler_runIO_closure -u _base_GHCziTopHandler_runNonIO_closure -u _base_GHCziConcziIO_ensureIOManagerIsRunning_closure -u _base_GHCziConcziIO_ioManagerCapabilitiesChanged_closure -u _base_GHCziConcziSync_runSparks_closure -u _base_GHCziConcziSignal_runHandlers_closure -search_paths_first -lSystem -lgcc -lSystem ld: warning: PIE disabled. Absolute addressing (perhaps -mdynamic-no-pic) not allowed in code signed PIE, but used in _Main_main_info from C.o. To fix this warning, don't compile with -mdynamic-no-pic or link with -Wl,-no_pie final section layout: __TEXT/__text addr=0x100000D30, size=0x00000218, fileOffset=0x00000D30, type=1 __TEXT/__stubs addr=0x100000F48, size=0x0000001E, fileOffset=0x00000F48, type=27 __TEXT/__stub_helper addr=0x100000F68, size=0x00000042, fileOffset=0x00000F68, type=31 __TEXT/__eh_frame addr=0x100000FB0, size=0x00000048, fileOffset=0x00000FB0, type=18 __DATA/__program_vars addr=0x100001000, size=0x00000028, fileOffset=0x00001000, type=29 __DATA/__got addr=0x100001028, size=0x00000020, fileOffset=0x00001028, type=28 __DATA/__nl_symbol_ptr addr=0x100001048, size=0x00000010, fileOffset=0x00001048, type=28 __DATA/__la_symbol_ptr addr=0x100001058, size=0x00000028, fileOffset=0x00001058, type=26 __DATA/__data addr=0x100001080, size=0x00000040, fileOffset=0x00001080, type=0 __DATA/__const addr=0x1000010C0, size=0x00000018, fileOffset=0x000010C0, type=0 __DATA/__common addr=0x1000010D8, size=0x00000020, fileOffset=0x00000000, type=24 ld: unexpected bindingNone in '_Main_main_info' from C.o for architecture x86_64 collect2: ld returned 1 exit status T7478: T7478: phase `Linker' failed (exitcode = 1) }}} Unfortunately, I don't know what a bindingNone is... -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8294> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #8245: ghc-pkg list prints packages in non-alphabetical order
by GHC 07 Dec '15

07 Dec '15

#8245: ghc-pkg list prints packages in non-alphabetical order ------------------------------------+------------------------------------- Reporter: nh2 | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- Instead, it prints the ones with the smallest version first. That does not make much sense. I will attach a patch. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8245> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #8098: Faulty Word64 arithmetic if optimized
by GHC 02 Dec '15

02 Dec '15

#8098: Faulty Word64 arithmetic if optimized ----------------------------+--------------------------------- Reporter: achp | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.4.2 Keywords: | Operating System: Windows Architecture: x86 | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ----------------------------+--------------------------------- The following program produces different results depending on optimization: Checked.hs: {{{ module Checked ((.-.)) where (.-.) :: (Bounded a, Integral a) => a -> a -> Maybe a x .-. y | x >= 0 && y < 0 = if z > x then Just z else Nothing | x >= 0 = if z <= x then Just z else Nothing | x < 0 && y > 0 = if z < x then Just z else Nothing | otherwise = Just z where z = x - y }}} Main.hs: {{{ module Main (main) where import Data.Maybe import Data.Word import Numeric import Checked main :: IO () main = print $ showHex (fromJust ((0xFFFFFFFF00000000 :: Word64) .-. 0x00000000FFFFFFFF)) "" }}} {{{ D:\dev\haskell\ch1>ghc --make Checked.hs Main.hs [1 of 2] Compiling Checked ( Checked.hs, Checked.o ) [2 of 2] Compiling Main ( Main.hs, Main.o ) Linking Main.exe ... D:\dev\haskell\ch1>Main.exe "fffffffe00000001" D:\dev\haskell\ch1>del *.hi *.o *.exe D:\dev\haskell\ch1>ghc --make -O2 Checked.hs Main.hs [1 of 2] Compiling Checked ( Checked.hs, Checked.o ) [2 of 2] Compiling Main ( Main.hs, Main.o ) Linking Main.exe ... D:\dev\haskell\ch1>Main.exe "fffe00000001" D:\dev\haskell\ch1> }}} The bug disappears if definition and usage of (.-.) are in the same module. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8098> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #8079: Redo T7919 with a cabal file
by GHC 22 Nov '15

22 Nov '15

#8079: Redo T7919 with a cabal file -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: task | Status: new Priority: normal | Milestone: Component: Test Suite | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: Compile-time Difficulty: Easy (less than 1 | crash hour) | Test Case: Blocked By: | Blocking: Related Tickets: | -------------------------------------+------------------------------------- Right now T7919 fails because the new dynamic-by-default GHC means that Template Haskell must be built with dynamic; of course, GHC isn't clever enough to figure this out. If we Cabalize the test, the example should start building again (and if they don't, well, that's a bug!) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8079> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2

[GHC] #8199: Get rid of HEAP_ALLOCED
by GHC 22 Nov '15

22 Nov '15

#8199: Get rid of HEAP_ALLOCED ----------------------------------------------+---------------------------- Reporter: ezyang | Owner: ezyang Type: feature request | Status: new Priority: normal | Milestone: 7.10.1 Component: Compiler | Version: 7.7 Keywords: | Operating System: Architecture: Unknown/Multiple | Unknown/Multiple Difficulty: Project (more than a week) | Type of failure: Blocked By: | None/Unknown Related Tickets: | Test Case: | Blocking: ----------------------------------------------+---------------------------- This bug is to track progress of removing HEAP_ALLOCED from GHC, promising faster GC, as long as we can keep the cost of indirections down. The relevant wiki article: http://ghc.haskell.org/trac/ghc/wiki/Commentary/Rts/Storage/HeapAlloced ; we are implementing method 2. The attached patchset is *almost* able to support GHC itself; a ghc-stage2 built this way can compile most files, but is much slower in some cases, and probably has some correctness bugs. Improvements possible: * Don't pay for a double indirection when -fPIC is turned on. Probably the easiest way to do this is to *not* bake in the indirections into compiled code when it is -fPIC'd, and instead scribble over the GOT. However, I don't know how to go backwards from a symbol to a GOT entry, so we might need some heinous assembly hacks to get this working. * Closure tables should have their indirections short-circuited by the initialization code. * We are paying an indirection when a GC occurs when the closure is not in R1. According to the wiki page, technically this is not needed, but I don't know how to eliminate references to the closure itself from stg_gc_fun. * Properly handle the Windows DLL case (e.g. SRTs); this should be simpler now that everything is an indirection. * Bikeshed hs_main API changes (because closures are not valid prior to RTS initialization, so you have to pass in an indirection instead) * Improve static indirection and stable pointer registration, avoiding binary bloat from `__attribute(constructor)__` stubs. * Need to implement a new megablock tracking structure so we can free/check for lost blocks -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8199> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 52

[GHC] #8364: equip GHC with an accurate internal model of floating point
by GHC 07 Oct '15

07 Oct '15

#8364: equip GHC with an accurate internal model of floating point ----------------------------------------------+---------------------------- Reporter: carter | Owner: Type: feature request | Status: new Priority: high | Milestone: 7.10.1 Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Architecture: Unknown/Multiple | Unknown/Multiple Difficulty: Project (more than a week) | Type of failure: Blocked By: | None/Unknown Related Tickets: | Test Case: | Blocking: ----------------------------------------------+---------------------------- currently there really isn't a systematic / portable story for doing compile time reduction / evaluation for floating point numbers. there are some folks in the larger haskell community who are working on high precision soft float implementations / modelling. Once that work becomes available, might be worth exploring having something like that for doing controlled high precision evaluation of intermediate expressions in ghc proper. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8364> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #8335: Create more specialized entries to GC
by GHC 07 Oct '15

07 Oct '15

#8335: Create more specialized entries to GC ------------------------------------+------------------------------------- Reporter: jstolarek | Owner: Type: task | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- Consider this Cmm code (taken from #8326): {{{ {offset cFd: // stack check if ((Sp + -16) < SpLim) goto cFr; else goto cFs; cFr: // not enough place on the stack, call GC R2 = R2; R1 = A.f_closure; call (stg_gc_fun)(R2, R1) args: 8, res: 0, upd: 8; cFs: // scrutinize (x ># 0#) _sEU::I64 = R2; _sEV::I64 = %MO_S_Gt_W64(R2, 0); if (_sEV::I64 != 1) goto cFg; else goto cFo; cFg: // False branch Hp = Hp + 16; if (Hp > HpLim) goto cFy; else goto cFx; cFy: // not enough heap, call GC HpAlloc = 16; I64[Sp - 16] = cFf; R1 = _sEV::I64; I64[Sp - 8] = _sEU::I64; Sp = Sp - 16; call stg_gc_unbx_r1(R1) returns to cFf, args: 8, res: 8, upd: 8; cFf: // re-do the False branch _sEU::I64 = I64[Sp + 8]; Sp = Sp + 16; _sEV::I64 = R1; goto cFg; cFx: // RHS of False branch I64[Hp - 8] = GHC.Types.I#_con_info; I64[Hp] = -_sEU::I64; R1 = Hp - 7; call (P64[Sp])(R1) args: 8, res: 0, upd: 8; cFo: // True branch Hp = Hp + 16; if (Hp > HpLim) goto cFv; else goto cFu; cFv: // not enough heap, call GC HpAlloc = 16; I64[Sp - 16] = cFn; R1 = _sEV::I64; I64[Sp - 8] = _sEU::I64; Sp = Sp - 16; call stg_gc_unbx_r1(R1) returns to cFn, args: 8, res: 8, upd: 8; cFn: // re-do the True branch _sEU::I64 = I64[Sp + 8]; Sp = Sp + 16; _sEV::I64 = R1; goto cFo; cFu: // RHS of True branch I64[Hp - 8] = GHC.Types.I#_con_info; I64[Hp] = _sEU::I64; R1 = Hp - 7; call (P64[Sp])(R1) args: 8, res: 0, upd: 8; } }}} We perform two different calls to garbage collection here. Calling GC after making initial stack check obviously can't use any stack, so we pass parameter to GC in regsiters: {{{ R2 = R2; R1 = A.f_closure; call (stg_gc_fun)(R2, R1) args: 8, res: 0, upd: 8; }}} But in the calls that are made later on inside the function we do put some parameters on the stack, thus increasing stack usage: {{{ cFv: // not enough heap, call GC HpAlloc = 16; I64[Sp - 16] = cFn; R1 = _sEV::I64; I64[Sp - 8] = _sEU::I64; Sp = Sp - 16; call stg_gc_unbx_r1(R1) returns to cFn, args: 8, res: 8, upd: 8; }}} If we had more specialized entries to GC that would allow us to pass parameters in registers (or in some fixed memory location associated with TCO?) then this function would not require any stack at all. Note that there is a separate issue here - `_sEV` variable is in fact dead and should be eliminated (see #8327). -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8335> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 2