[GHC] #8114: GHC panic when building `thyme`
#8114: GHC panic when building `thyme` -----------------------------------+--------------------------------------- Reporter: Ptharien's | Owner: Flame | Status: new Type: bug | Milestone: Priority: normal | Version: 7.6.3 Component: Compiler | Operating System: MacOS X Keywords: | Type of failure: Compile-time crash Architecture: x86_64 (amd64) | Test Case: Difficulty: Unknown | Blocking: Blocked By: | Related Tickets: | -----------------------------------+--------------------------------------- While attempting to build `thyme` from Hackage with core, stg, and cmm linting and -O2 enabled, GHC crashed with an internal panic. See attached terminal log. {{{ rrdhcp-10-32-234-247:thyme-0.2.4.1 Alexander$ ghc --version The Glorious Glasgow Haskell Compilation System, version 7.6.3 rrdhcp-10-32-234-247:thyme-0.2.4.1 Alexander$ cabal --version cabal-install version 1.16.0.2 using version 1.16.0 of the Cabal library rrdhcp-10-32-234-247:thyme-0.2.4.1 Alexander$ which ghc /usr/local/bin/ghc rrdhcp-10-32-234-247:thyme-0.2.4.1 Alexander$ which cabal /Users/Alexander/Library/Haskell/bin/cabal }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8114 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#8114: GHC panic when building `thyme` ---------------------------------------+---------------------------------- Reporter: Ptharien's Flame | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Resolution: | Keywords: Operating System: MacOS X | Architecture: x86_64 (amd64) Type of failure: Compile-time crash | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: ---------------------------------------+---------------------------------- Comment (by monoidal): As far as I can tell, the origin of the STG Lint error is Data.Text.Array in the text package. Here's the crucial part of that module. To reproduce, save this code as Array.hs and compile with `ghc -dstg-lint Array`. This gives a Lint error in 7.6.3 and HEAD. {{{ {-# LANGUAGE MagicHash, RecordWildCards, UnboxedTuples #-} module Array where import GHC.Base (ByteArray#, MutableByteArray#, unsafeCoerce#) import GHC.ST (ST(..), runST) data Array = Array { aBA :: ByteArray# } data MArray s = MArray { maBA :: MutableByteArray# s } unsafeFreeze :: MArray s -> ST s Array unsafeFreeze MArray{..} = ST $ \s# -> (# s#, Array (unsafeCoerce# maBA) #) }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8114#comment:1 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#8114: GHC panic when building `thyme` ---------------------------------------+---------------------------------- Reporter: Ptharien's Flame | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Resolution: | Keywords: Operating System: MacOS X | Architecture: x86_64 (amd64) Type of failure: Compile-time crash | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: ---------------------------------------+---------------------------------- Comment (by carter): the lint error message (using the minimal example that monoidal provides), on ghc 7.6.3 on OS X is the following {{{ carter repoScratcher/testtyme ยป ghc -dstg-lint Array [1 of 1] Compiling Array ( Array.hs, Array.o ) stgEqType: unequal ghc-prim:GHC.Prim.ByteArray#{(w) tc 3f} ghc-prim:GHC.Prim.MutableByteArray#{(w) tc 31p} s{tv h} [tv] ghc: panic! (the 'impossible' happened) (GHC version 7.6.3 for x86_64-apple-darwin): *** Stg Lint ErrMsgs: in Stg2Stg *** <no location info>: Warning: [RHS of sat_sgA{v} [lid] :: main:Array.Array{tc rfc}] In a RHS constructor application, con type doesn't match arg types: Constructor type: ghc-prim:GHC.Prim.ByteArray#{(w) tc 3f} -> main:Array.Array{tc rfc} Arg types: ghc-prim:GHC.Prim.MutableByteArray#{(w) tc 31p} s{tv h} [tv] *** Offending Program *** main:Array.unsafeFreeze{v rff} [gid] :: forall s{tv afh} [tv]. main:Array.MArray{tc rf9} s{tv afh} [tv] -> base:GHC.ST.ST{tc r4k} s{tv afh} [tv] main:Array.Array{tc rfc} [GblId, Arity=1, Unf=OtherCon []] = [] \r srt:SRT:[(01Z, base:GHC.Base.${v 01Z} [gid])] [ds{v sgf} [lid]] case ds{v sgf} [lid] :: Alg main:Array.MArray{tc rf9} of (wild{v sgD} [lid] [Occ=Dead] :: main:Array.MArray{tc rf9} s{tv h} [tv]) { -- lvs: [ds{v sgf} [lid]]; rhs lvs: []; srt:SRT:[(01Z, base:GHC.Base.${v 01Z} [gid])] main:Array.MArray{d rfa} (ds1{v sgn} [lid] :: ghc- prim:GHC.Prim.MutableByteArray#{(w) tc 31p} s{tv h} [tv]) -> let { sat_sgB{v} [lid] :: base:GHC.ST.STRep{tc r4i} s{tv h} [tv] main:Array.Array{tc rfc} [LclId] = [ds1{v sgn} [lid]] \r srt:SRT:[] [s#{v sgm} [lid]] let { sat_sgA{v} [lid] :: main:Array.Array{tc rfc} [LclId] = NO_CCS main:Array.Array{d rfd}! [ds1{v sgn} [lid]]; } in ghc-prim:GHC.Prim.(#,#){(w) d 84} [s#{v sgm} [lid] sat_sgA{v} [lid]]; } in let { sat_sgC{v} [lid] :: base:GHC.ST.STRep{tc r4i} s{tv h} [tv] main:Array.Array{tc rfc} -> base:GHC.ST.ST{tc r4k} s{tv h} [tv] main:Array.Array{tc rfc} [LclId] = [] \r srt:SRT:[] [tpl{v sgj} [lid]] tpl{v sgj} [lid]; } in base:GHC.Base.${v 01Z} [gid] sat_sgC{v} [lid] sat_sgB{v} [lid]; }; main:Array.aBA{v rfe} [gid[[RecSel]]] :: main:Array.Array{tc rfc} -> ghc-prim:GHC.Prim.ByteArray#{(w) tc 3f} [GblId[[RecSel]], Arity=1, Caf=NoCafRefs, Unf=OtherCon []] = [] \r srt:SRT:[] [ds{v sgr} [lid]] case ds{v sgr} [lid] :: Alg main:Array.Array{tc rfc} of (wild{v sgE} [lid] [Occ=Dead] :: main:Array.Array{tc rfc}) { -- lvs: [ds{v sgr} [lid]]; rhs lvs: []; srt:SRT:[] main:Array.Array{d rfd} (ds1{v sgu} [lid] :: ghc- prim:GHC.Prim.ByteArray#{(w) tc 3f}) -> ds1{v sgu} [lid]; }; main:Array.maBA{v rfb} [gid[[RecSel]]] :: forall s{tv afg} [tv]. main:Array.MArray{tc rf9} s{tv afg} [tv] -> ghc-prim:GHC.Prim.MutableByteArray#{(w) tc 31p} s{tv afg} [tv] [GblId[[RecSel]], Arity=1, Caf=NoCafRefs, Unf=OtherCon []] = [] \r srt:SRT:[] [ds{v sgw} [lid]] case ds{v sgw} [lid] :: Alg main:Array.MArray{tc rf9} of (wild{v sgF} [lid] [Occ=Dead] :: main:Array.MArray{tc rf9} s{tv d} [tv]) { -- lvs: [ds{v sgw} [lid]]; rhs lvs: []; srt:SRT:[] main:Array.MArray{d rfa} (ds1{v sgz} [lid] :: ghc- prim:GHC.Prim.MutableByteArray#{(w) tc 31p} s{tv d} [tv]) -> ds1{v sgz} [lid]; }; main:Array.MArray{v rfr} [gid[DataCon]] :: forall s{tv afg} [tv]. ghc-prim:GHC.Prim.MutableByteArray#{(w) tc 31p} s{tv afg} [tv] -> main:Array.MArray{tc rf9} s{tv afg} [tv] [GblId[DataCon], Arity=1, Caf=NoCafRefs, Str=DmdType Tm, Unf=OtherCon []] = [] \r srt:SRT:[] [eta_B1{v} [lid]] main:Array.MArray{d rfa} [eta_B1{v} [lid]]; main:Array.Array{v rfn} [gid[DataCon]] :: ghc-prim:GHC.Prim.ByteArray#{(w) tc 3f} -> main:Array.Array{tc rfc} [GblId[DataCon], Arity=1, Caf=NoCafRefs, Str=DmdType Tm, Unf=OtherCon []] = [] \r srt:SRT:[] [eta_B1{v} [lid]] main:Array.Array{d rfd} [eta_B1{v} [lid]]; *** End of Offense *** Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8114#comment:2 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#8114: GHC panic when building `thyme` ---------------------------------------+---------------------------------- Reporter: Ptharien's Flame | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Resolution: | Keywords: Operating System: MacOS X | Architecture: x86_64 (amd64) Type of failure: Compile-time crash | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: ---------------------------------------+---------------------------------- Comment (by carter): the meat of the error message is {{{ stgEqType: unequal ghc-prim:GHC.Prim.ByteArray#{(w) tc 3f} ghc-prim:GHC.Prim.MutableByteArray#{(w) tc 31p} s{tv h} [tv] ghc: panic! (the 'impossible' happened) (GHC version 7.6.3 for x86_64-apple-darwin): *** Stg Lint ErrMsgs: in Stg2Stg *** <no location info>: Warning: [RHS of sat_sgA{v} [lid] :: main:Array.Array{tc rfc}] In a RHS constructor application, con type doesn't match arg types: Constructor type: ghc-prim:GHC.Prim.ByteArray#{(w) tc 3f} -> main:Array.Array{tc rfc} Arg types: ghc-prim:GHC.Prim.MutableByteArray#{(w) tc 31p} s{tv h} [tv] }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8114#comment:3 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#8114: GHC panic when building `thyme` ---------------------------------------+---------------------------------- Reporter: Ptharien's Flame | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Resolution: | Keywords: Operating System: MacOS X | Architecture: x86_64 (amd64) Type of failure: Compile-time crash | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: ---------------------------------------+---------------------------------- Comment (by monoidal): See also Simon's comment at #5345. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8114#comment:4 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#8114: STG lint panic -------------------------------------+------------------------------------- Reporter: Ptharien's | Owner: Flame | Status: new Type: bug | Milestone: Priority: normal | Version: 7.6.3 Component: Compiler | Keywords: Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: Compile- | Related Tickets: #5345 time crash | Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Changes (by thomie): * os: MacOS X => Unknown/Multiple * architecture: x86_64 (amd64) => Unknown/Multiple * related: => #5345 Comment: Confirmed on Linux with ghc-7.9.20141115 using example from comment:1. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8114#comment:5 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#8114: STG lint panic -------------------------------------+------------------------------------- Reporter: Ptharien's Flame | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Compile-time | Unknown/Multiple crash | Test Case: Blocked By: | Blocking: Related Tickets: #5345 | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by bgamari): Apart from the STG lint issue (which I really can't comment on), it's not clear to me why `text` is using `unsafeCoerce#` here instead of `unfreezeByteArray#`, which seems to pre-date `Data.Text.Array`. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8114#comment:6 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#8114: STG lint panic -------------------------------------+------------------------------------- Reporter: Ptharien's Flame | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Compile-time | Unknown/Multiple crash | Test Case: Blocked By: | Blocking: Related Tickets: #5345 | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by bgamari): Indeed reworking the code in question to use `unsafeFreezeByteArray#` resolves the STG lint violation. I've opened a [[https://github.com/bos/text/pull/144|pull request]] against `text` fixing this. Given that this issue is only observed with `unsafeCoerce#` and STG linting is already a bit sketchy, I'm going to close this as won't fix. Thanks for your report! -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8114#comment:7 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#8114: STG lint panic -------------------------------------+------------------------------------- Reporter: Ptharien's Flame | Owner: Type: bug | Status: closed Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Resolution: wontfix | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Compile-time | Unknown/Multiple crash | Test Case: Blocked By: | Blocking: Related Tickets: #5345 | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by bgamari): * status: new => closed * resolution: => wontfix -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8114#comment:8 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
participants (1)
-
GHC