September 2016 - ghc-tickets

[GHC] #12582: HSOC Eventlog live profiling
by GHC 12 Jun '17

12 Jun '17

#12582: HSOC Eventlog live profiling -------------------------------------+------------------------------------- Reporter: NCrashed | Owner: NCrashed Type: feature | Status: new request | Priority: normal | Milestone: Component: Runtime | Version: 8.0.1 System | Keywords: eventlog | Operating System: Unknown/Multiple profile hsoc | Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Hi, it is ticket for results of my Haskell Summer of Code 2016 project [1]. The goal is provide feature for redirection of events to external profilers without an application termination. Proposed changes: * Changing eventlog file descriptor in runtime. {{{#!c /* * Set custom file stream for global event log sink. * * The function overwrites previous event log file pointer. Previouss * sink is closed only if closePrev flag is on. * * Writing to the sink is protected by global mutex. * * The function puts header to the new sink only when emitHeader flag * is on. User might not want the header if it is switching to * already existed eventlog handle that was switched away recently. */ void rts_setEventLogSink(FILE *sink, StgBool closePrev, StgBool emitHeader); /* * Get current file stream that is used for global event log sink. * * You shouldn't do anything with the pointer until * rts_setEventLogSink(otherFileOrNull, false) is called. After that * you can do anything with the file stream. */ FILE* rts_getEventLogSink(void); }}} * Special flag `-lm` for RTS to store eventlog in memory chunks. User space functions to retrieve the chunks from memory. {{{#!c /* * If RTS started with '-lm' flag then eventlog is stored in memory buffer. * * The function allows to pop chunks of the buffer. Return value of 0 means * that there is no any filled chunk of data. * * If the function returns nonzero value the parameter contains full chunk * of eventlog data with size of the returned value. Caller must free the * buffer, the buffer isn't referenced anywhere anymore. * * If nobody calls the function with '-lm' flag then the memory is kinda * to be exhausted. */ StgWord64 rts_getEventLogChunk(StgInt8** ptr); }}} * Feature for dynamic resize of eventlog buffers. {{{#!c /* * Reallocate inner buffers to match the new size. The size should be not * too small to contain at least one event. * * If RTS started with '-lm' the chunks of memory buffer is also resized. */ void rts_resizeEventLog(StgWord64 size); /* * Return current size of eventlog buffers. */ StgWord64 rts_getEventLogBuffersSize(void); }}} * Bindings to new features in `Debug.Trace` in `base` package. {{{#!hs -- | The 'setEventLogCFile' function changes current sink of the eventlog, if eventlog -- profiling is available and enabled at runtime. -- -- The second parameter defines whether old sink should be finalized and closed or not. -- Preserving it could be helpful for temporal redirection of eventlog data into not -- standard sink and then restoring to the default file sink. -- -- The third parameter defines whether new header section should be emitted to the new -- sink. Emitting header to already started eventlog streams will corrupt the structure -- of eventlog format. -- -- The function is more low-level than 'setEventLogHandle' but doesn't recreate underlying -- file descriptor and is intended to use with 'getEventLogCFile' to save and restore -- current sink of the eventlog. -- -- @since 4.10.0.0 setEventLogCFile :: Ptr CFile -> Bool -> Bool -> IO () -- | The 'getEventLogCFile' function returns current sink of the eventlog, if eventlog -- profiling is available and enabled at runtime. -- -- The function is intented to be used with 'setEventLogCFile' to save and restore -- current sink of the eventlog. -- -- @since 4.10.0.0 getEventLogCFile :: IO (Ptr CFile) -- | Setting size of internal eventlog buffers. The size should be not -- too small to contain at least one event. -- -- If RTS started with '-lm' the chunks of memory buffer is also resized. -- -- The larger the buffers the lesser overhead from event logging, but -- larger delays between data dumps. -- -- See also: 'getEventLogChunk', 'getEventLogBufferSize' setEventLogBufferSize :: Word -> IO () -- | Getting size of internal eventlog buffers. -- -- See also: 'setEventLogBufferSize', 'getEventLogChunk' getEventLogBufferSize :: IO Word -- | Get next portion of the eventlog data. -- -- If RTS started with '-lm' flag then eventlog is stored in memory buffer. -- -- The function allows to pop chunks out of the buffer. Return value of Nothing -- means that there is no any filled chunk of data. -- -- If the function returns nonzero value the parameter contains full chunk -- of eventlog data with size of the returned value. Caller must free the -- buffer with 'free' from 'Foreign.Marshal.Alloc', the buffer isn't referenced -- anywhere anymore. -- -- If nobody calls the function with '-lm' flag on then the memory is kinda -- to be exhausted. -- -- If '-lm' flag is off, the function returns always 'Nothing'. -- -- See also: 'setEventLogBufferSize' getEventLogChunk :: IO (Maybe CStringLen) }}} References: 1. [http://ncrashed.github.io/blog/posts/2016-06-12-hsoc-acceptance.html HSOC project description] 2. [http://ncrashed.github.io/blog/posts/2016-06-22-hsoc-rts.html Design choices for implementation] -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12582> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 6

[GHC] #11501: Building nofib/fibon returns permission denied
by GHC 08 Jun '17

08 Jun '17

#11501: Building nofib/fibon returns permission denied -------------------------------------+------------------------------------- Reporter: rem | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: NoFib | Version: 7.10.3 benchmark suite | Keywords: | Operating System: Linux Architecture: x86_64 | Type of failure: None/Unknown (amd64) | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Building nofib/fibon returns permission denied: {{{ == make boot - --no-print-directory; in /h/ywang30/nofib/fibon/Hackage/Bzlib ------------------------------------------------------------------------ // Codec/Compression/BZip/Stream.hsc make[3]: execvp: //: Permission denied make[3]: *** [Codec/Compression/BZip/Stream.hs] Error 127 }}} Steps to reproduce: Edit nofib/mk/boilerplate.mk:35 to NoFibSubDirs = fibon make clean && make boot in nofib/ -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11501> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 33

[GHC] #11317: Test prog003 fails with segfault on Windows (GHCi)
by GHC 07 Jun '17

07 Jun '17

#11317: Test prog003 fails with segfault on Windows (GHCi) --------------------------------------+------------------------------- Reporter: rdragon | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: GHCi | Version: 7.11 Keywords: | Operating System: Windows Architecture: x86_64 (amd64) | Type of failure: GHCi crash Test Case: | Blocked By: Blocking: | Related Tickets: #11234 Differential Rev(s): | Wiki Page: --------------------------------------+------------------------------- On my `x86_64` Windows machine test `prog003` fails with a segfault: {{{ cd ./ghci/prog003 && ghciWayFlags=-static HC="/home/Rik/ghc/inplace/bin /ghc-stage2.exe" HC_OPTS="-dcore-lint -dcmm-lint -dno-debug-output -no- user-package-db -rtsopts -fno-warn-tabs -fno-warn-missed-specialisations -fno-ghci-history " "/home/Rik/ghc/inplace/bin/ghc-stage2.exe" -dcore-lint -dcmm-lint -dno-debug-output -no-user-package-db -rtsopts -fno-warn-tabs -fno-warn-missed-specialisations -fno-ghci-history --interactive -v0 -ignore-dot-ghci +RTS -I0.1 -RTS <prog003.script > prog003.run.stdout 2> prog003.run.stderr >> prog003.run.stdout 2>> prog003.run.stderr Wrong exit code (expected 0 , actual 1 ) Stdout: Run 1 a :: Int -> Int 168 Run 2 (A.a,B.b,C.c,D.d) :: (Float -> Float, Float -> Float, Float -> Float, Float -> Float) 28.0 Run 3 (A.a,B.b,C.c,D.d) :: (Float -> Float, Float -> Float, Float -> Float, Float -> Float) 28.0 Run 4 (A.a,B.b,C.c,D.d) :: (Float -> Float, Float -> Float, Float -> Float, Float -> Float) 28.0 Run 5 (A.a,B.b,C.c,D.d) :: (Float -> Float, Float -> Float, Float -> Float, Float -> Float) 28.0 Run 6 (A.a,B.b,C.c,D.d) :: (Float -> Float, Float -> Float, Float -> Float, Float -> Float) 28.0 Run 7 (A.a,B.b,C.c,D.d) :: (Float -> Float, Float -> Float, Float -> Float, Float -> Float) 28.0 Run 8 Segmentation fault/access violation in generated code Stderr: *** unexpected failure for prog003(ghci) }}} (I do not think it is an access violation error since when I remove [https://github.com/ghc/ghc/blob/c2fab84239299176a8e72aa26ae894019d677bd9/te… line 65] and everything below line 70, the same stdout is generated.) This looks related to ticket #11234. The GHC version used was [http://git.haskell.org/ghc.git/commit/df6cb57b32d94b7f6f7c9a86207adfeee9712… df6cb57b32d94b7f6f7c9a86207adfeee9712ed6], although I already noticed the problem with a version a few days older. If someone could point me to information about how to install GDB suitable for debugging GHC on my Windows machine I will try to post the stack trace. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11317> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 15

[GHC] #8703: Use guard pages rather than heap checks
by GHC 07 Jun '17

07 Jun '17

#8703: Use guard pages rather than heap checks ------------------------------------+------------------------------------- Reporter: schyler | Owner: simonmar Type: feature request | Status: new Priority: normal | Milestone: Component: Runtime System | Version: Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- By mmap'ing memory using MAP_ANONYMOUS (or /dev/zero) as PROT_NONE (Windows has an equivalent functionality with VirtualAlloc), it's possible to create a page that will fault when read or written to. It may be possible to remove the heap checks in the runtime, speeding up allocations, and instead expand the heap when a fault is raised. This is likely cheaper for large pages full of many small allocations than the cost of the heap checking branch. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8703> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #10600: -fwarn-incomplete-patterns doesn't work with -fno-code
by GHC 05 Jun '17

05 Jun '17

#10600: -fwarn-incomplete-patterns doesn't work with -fno-code -------------------------------------+------------------------------------- Reporter: akio | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Incorrect Unknown/Multiple | warning at compile-time Test Case: | Blocked By: Blocking: | Related Tickets: Differential Revisions: | -------------------------------------+------------------------------------- `-fwarn-incomplete-patterns` doensn't seem to generate any warnings when `-fno-code` is specified. To reproduce, save this module as `foo.hs`. {{{#!hs module Foo where foo True = 4 }}} and compile with `-fwarn-incomplete-patterns`, with and without `-fno- code`. {{{ % ghc -fwarn-incomplete-patterns -fforce-recomp foo.hs [1 of 1] Compiling Foo ( foo.hs, foo.o ) foo.hs:2:1: Warning: Pattern match(es) are non-exhaustive In an equation for ‘foo’: Patterns not matched: False % ghc -fwarn-incomplete-patterns -fforce-recomp -fno-code foo.hs [1 of 1] Compiling Foo ( foo.hs, nothing ) % }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10600> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 26

[GHC] #11032: Missing result type handling for State# s in foreign import prim.
by GHC 02 Jun '17

02 Jun '17

#11032: Missing result type handling for State# s in foreign import prim. -------------------------------------+------------------------------------- Reporter: ekmett | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.2 (FFI) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: GHC rejects Unknown/Multiple | valid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- I've been using `foreign import prim` for a while now to craft custom primops on an as-needed basis. Attempting to write {{{#!hs {-# LANGUAGE MagicHash #-} {-# LANGUAGE UnboxedTuples #-} {-# LANGUAGE UnliftedFFITypes #-} {-# LANGUAGE GHCForeignImportPrim #-} {-# LANGUAGE ForeignFunctionInterface #-} foreign import prim "pinThreadzh" pinThread# :: State# s -> (# State# s, Int#, Int# #) foreign import prim "unpinThreadzh" unpinThread# :: State# s -> State# s }}} threw me for a loop when it didn't work. The former works fine. The latter complains that `State# s` isn't a valid result type! {{{ src/Concurrent/Internal/Thread.hs:23:1: Unacceptable result type in foreign declaration: ‘State# s’ cannot be marshalled in a foreign call When checking declaration: foreign import prim safe "static unpinThreadzh" unpinThread# :: State# s -> State# s }}} Builtin primitives work this way all the time, so I'm guessing we just missed it in a case statement somewhere inside the compiler when dealing with `foreign import prim` support. I can likely hack around it by giving back something like `(# State# s #)` for now but that is a rather unsatisfying solution. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11032> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4

[GHC] #12499: Support multiple library import libs
by GHC 02 Jun '17

02 Jun '17

#12499: Support multiple library import libs -------------------------------------+------------------------------------- Reporter: Phyx- | Owner: Type: feature | Status: new request | Priority: normal | Milestone: 8.2.1 Component: Runtime | Version: 8.0.1 System (Linker) | Keywords: | Operating System: Windows Architecture: | Type of failure: GHC rejects Unknown/Multiple | valid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Our implementation of import libraries only support single library import libs. These are by far the most common ones. However it is possible to make multi library import libraries. E.g. one import library pointing to multiple dlls. `binutils` correct supports these and with #5987 GHC will be relying on them to work. It makes sense that we are able to load them in GHCi as well. This will be required for a fully working dynamic GHC on Windows. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12499> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5

[GHC] #12237: Constraint resolution vs. type family resolution vs. TypeErrors
by GHC 01 Jun '17

01 Jun '17

#12237: Constraint resolution vs. type family resolution vs. TypeErrors -------------------------------------+------------------------------------- Reporter: cactus | Owner: Type: feature | Status: new request | Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Unknown/Multiple CustomTypeErrors | Architecture: | Type of failure: Other Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- Via http://stackoverflow.com/q/37769351/477476, given the following setup: {{{#!hs {-# LANGUAGE DataKinds, TypeFamilies, TypeOperators #-} {-# LANGUAGE FlexibleContexts, MultiParamTypeClasses, UndecidableInstances #-} import Data.Proxy import GHC.TypeLits type family TEq a b where TEq a a = a TEq a b = TypeError (Text "TEq error") type family F a where F () = () F (a, b) = TEq (F a) (F b) F a = TypeError (Text "Unsupported type: " :<>: ShowType a) }}} and the following usages of `F` and `TEq`: {{{#!hs class (repr ~ F a) => C repr a foo :: (C (F a) a) => proxy a -> () foo _ = () main :: IO () main = print $ foo (Proxy :: Proxy (Int, ())) }}} This results in {{{ * No instance for (C (TEq (TypeError ...) ()) (Int, ())) arising from a use of `foo' * In the second argument of `($)', namely `foo (Proxy :: Proxy (Int, ()))' In the expression: print $ foo (Proxy :: Proxy (Int, ())) In an equation for `main': main = print $ foo (Proxy :: Proxy (Int, ())) }}} but it would be preferable to bail out earlier when `F Int` is resolved to `TypeError ...` instead of propagating that all the way to the `C` constraint. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12237> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #11834: GHC master, not compiling on Archlinux
by GHC 29 May '17

29 May '17

#11834: GHC master, not compiling on Archlinux -------------------------------------+------------------------------------- Reporter: nitrix | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: (Linking) | Keywords: | Operating System: Linux Architecture: x86_64 | Type of failure: Building GHC (amd64) | failed Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- /usr/bin/ld: -r and -pie may not be used together http://lpaste.net/159726 It appears on Archlinux, with GCC version 5.3.0, there's an implicit -pie flag passed to `ld`. I looked into compiler/main/DriverPipeline.hs at joinObjectFiles, it is unconditionally doing "-Wl,-r", thus the error above. One suggested fix on IRC is to modify getCompilerInfo to know if it supports -no-pie, then add it to the flags that are used. I tried and got somewhere lost in compiler/main/DynFlags.hs. (Completely new to GHC) Hopefully someone can figure it out, it seems important for development and I really want to get involved :) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11834> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 12

[GHC] #12648: Stack overflow when using monad-unlift
by GHC 27 May '17

27 May '17

#12648: Stack overflow when using monad-unlift -------------------------------------+------------------------------------- Reporter: nh2 | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: Compile-time Unknown/Multiple | crash Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- I found that with the following simple program I could make runghc/ghci crash with `*** Exception: stack overflow`: {{{ #!/usr/bin/env stack -- stack --resolver lts-7.1 --install-ghc runghc --package monad- unlift-0.2.0 {-# LANGUAGE FlexibleContexts #-} import Control.Monad.Trans.Unlift f :: (MonadBaseUnlift m IO) => m a f = do _ <- askUnliftBase return () }}} I'm not super sure if that's OK. Probably not; even if it does trigger an undecideable case, shouldn't ghc tell me about it instead of just crashing? The `monad-unlift` code used is here: https://github.com/fpco/monad- unlift/blob/adb7869adccc66247cbb1a6175277802e9c098e1/monad- unlift/Control/Monad/Trans/Unlift.hs -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12648> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3