[GHC] #11108: Weak references related crash

GHC

18 Nov 2015 18 Nov '15

6:07 a.m.

#11108: Weak references related crash ----------------------------------------+---------------------------------- Reporter: Saulzar | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Runtime System | Version: 7.10.2 Keywords: | Operating System: Linux Architecture: Unknown/Multiple | Type of failure: Runtime crash Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: ----------------------------------------+---------------------------------- I've been getting a Segmentation fault relating to Weak reference use as part of a FRP library (reflex). Attached is a more minimal example which crashes. With debugging on the original problem gave me these problems (on different runs): Most frequently just Segmentation fault (with no stack trace) Occasionally this: {{{ bench: internal error: ASSERTION FAILED: file rts/sm/GCAux.c, line 44 (GHC version 7.10.2 for x86_64_unknown_linux) Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug Aborted (core dumped) }}} and more rarely this: {{{ bench: internal error: WEAK object entered! (GHC version 7.10.2 for x86_64_unknown_linux) Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug Aborted (core dumped) }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

Show replies by date

GHC

18 Nov 18 Nov

6:08 a.m.

GHC

6:38 a.m.

GHC

12:11 p.m.

GHC

25 Nov 25 Nov

6:15 p.m.

GHC

6:21 p.m.

GHC

6:32 p.m.

GHC

29 Feb 29 Feb

7:59 a.m.

GHC

24 Mar 24 Mar

3:43 p.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: jme Type: bug | Status: new Priority: high | Milestone: 8.2.1 Component: Runtime System | Version: 7.10.2 Resolution: | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746 | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by jme): * owner: => jme * related: => #11746 Comment: I'll give this one a shot. It looks like the weak pointer is ending up in an older generation than its key. The key gets GCed during a minor collection, but the weak pointer is not "tidied" until the next major collection. At that point, its pointer to the key is no longer valid. This appears to be the same issue as #11746. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:7 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

29 Mar 29 Mar

6:29 a.m.

GHC

2:28 p.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: jme Type: bug | Status: new Priority: high | Milestone: 8.2.1 Component: Runtime System | Version: 7.10.2 Resolution: | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746 | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by jme): In case anyone needs a workaround for this bug, try running with `+RTS -G1 -RTS` (so the garbage collector only uses one generation). -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:9 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

10 Apr 10 Apr

8:07 p.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: jme Type: bug | Status: new Priority: high | Milestone: 8.2.1 Component: Runtime System | Version: 7.10.2 Resolution: | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746 | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by akio): * cc: akio (added) Comment: Oh... it sounds like I messed things up when I implemented per-generation weak pointer lists. Please let me know if there is anything I can help with. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:10 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

11 Apr 11 Apr

12:44 a.m.

...

Please let me know if there is anything I can help with. Thanks! The fix should be relatively straightforward, but if I do have any questions, I'll be sure to ask.

-- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:11 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

23 Apr 23 Apr

11:47 a.m.

GHC

3:28 p.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: jme Type: bug | Status: new Priority: high | Milestone: 8.2.1 Component: Runtime System | Version: 7.10.2 Resolution: | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746,#11972 | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by ryantrinkle): Hi jme/akio, Is there any way I can help get this resolved? It's a pretty serious issue for code that uses Weak pointers heavily, and -G1 seems like it may create a performance issue in production applications. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:13 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

3:38 p.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: jme Type: bug | Status: new Priority: high | Milestone: 8.2.1 Component: Runtime System | Version: 7.10.2 Resolution: | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746,#11972 | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by ezyang): If it's a regression from per-generation weak pointer lists, we can just revert that commit, right? -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:14 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

6:12 p.m.

...

If it's a regression from per-generation weak pointer lists, we can just revert that commit, right? Yes, that should be fine. I'm swamped at the moment, so if someone could take care of this, that would be great.

-- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:15 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

25 Apr 25 Apr

5:40 a.m.

GHC

9 May 9 May

8:26 p.m.

GHC

10 May 10 May

8:37 a.m.

...

If it's a regression from per-generation weak pointer lists, we can just revert that commit, right?

Since per-generation weak pointer lists are in since GHC 7.8, reverting it would mean creating another regression. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:18 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

11 May 11 May

9:18 a.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: akio Type: bug | Status: new Priority: high | Milestone: 8.2.1 Component: Runtime System | Version: 7.10.2 Resolution: | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746,#11972 | Differential Rev(s): Phab:D2189 Wiki Page: | -------------------------------------+------------------------------------- Comment (by Ben Gamari ): In [changeset:"9363f04d0ff22f3d898af35bb5432c4287e6dc9a/ghc" 9363f04/ghc]: {{{ #!CommitTicketReference repository="ghc" revision="9363f04d0ff22f3d898af35bb5432c4287e6dc9a" Handle promotion failures when scavenging a WEAK (#11108) Previously, we ignored promotion failures when evacuating fields of a WEAK object. When a failure happens, this resulted in an WEAK object pointing to another object in a younger generation, causing crashes. I used the test case from #11746 to check that the fix is working. However I haven't managed to produce a test case that quickly reproduces the issue. Test Plan: ./validate Reviewers: austin, bgamari, simonmar Reviewed By: simonmar Subscribers: thomie Differential Revision: https://phabricator.haskell.org/D2189 GHC Trac Issues: #11108 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:19 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

9:40 a.m.

GHC

11:17 a.m.

GHC

12 May 12 May

11:30 a.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: akio Type: bug | Status: closed Priority: high | Milestone: 8.0.1 Component: Runtime System | Version: 7.10.2 Resolution: fixed | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746,#11972 | Differential Rev(s): Phab:D2189 Wiki Page: | -------------------------------------+------------------------------------- Comment (by Ben Gamari ): In [changeset:"7c0b595e55d31f9f89e6dede11981e942c5bb32f/ghc" 7c0b595e/ghc]: {{{ #!CommitTicketReference repository="ghc" revision="7c0b595e55d31f9f89e6dede11981e942c5bb32f" Fix comments about scavenging WEAK objects This is a follow-up of D2189. If fixes some comments, deletes a section in the User's Guide about the bug, and updates .mailmap as suggested on the WorkinConventions wiki page. Test Plan: It compiles. Reviewers: austin, simonmar, bgamari Reviewed By: bgamari Subscribers: thomie Differential Revision: https://phabricator.haskell.org/D2202 GHC Trac Issues: #11108 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:22 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

7:59 p.m.

...

Here's a patch to add this test: https://phabricator.haskell.org/D2196 Oh I somehow didn't notice this patch, thank you for catching this.

-- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:23 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

9:01 p.m.

GHC

18 May 18 May

3:59 p.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: akio Type: bug | Status: closed Priority: high | Milestone: 8.0.1 Component: Runtime System | Version: 7.10.2 Resolution: fixed | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746,#11972 | Differential Rev(s): Phab:D2189 Wiki Page: | -------------------------------------+------------------------------------- Comment (by Ben Gamari ): In [changeset:"ba3e1fd37dc5004c4307ed205f6701b16faceb59/ghc" ba3e1fd/ghc]: {{{ #!CommitTicketReference repository="ghc" revision="ba3e1fd37dc5004c4307ed205f6701b16faceb59" Add a test for #11108 Reviewers: austin, simonmar, bgamari Reviewed By: bgamari Subscribers: thomie Differential Revision: https://phabricator.haskell.org/D2196 GHC Trac Issues: #11108 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:25 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

6 p.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: Type: bug | Status: new Priority: high | Milestone: 8.0.1 Component: Runtime System | Version: 7.10.2 Resolution: | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746,#11972 | Differential Rev(s): Phab:D2189 Wiki Page: | -------------------------------------+------------------------------------- Changes (by thomie): * owner: akio => * status: closed => new * resolution: fixed => Comment: That test is failing on Travis. Set `DYNAMIC_GHC_PROGRAMS=NO` in `mk/build.mk` to reproduce. Maybe somebody could have a look? {{{ =====> T11108(normal) 3321 of 5160 [0, 0, 0] +++ /tmp/ghctest/hXSuXa/1/2/3/./rts/T11108/T11108.run.stdout.normalised 2016-05-18 21:08:22.862178964 +0000 @@ -0,0 +1,10 @@ +1001 +1002 +1003 +1004 +1005 +1006 +1007 +1008 +1009 +1010 *** unexpected failure for T11108(normal) }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:26 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

19 May 19 May

8:14 a.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: Type: bug | Status: closed Priority: high | Milestone: 8.0.1 Component: Runtime System | Version: 7.10.2 Resolution: fixed | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746,#11972 | Differential Rev(s): Phab:D2189 Wiki Page: | -------------------------------------+------------------------------------- Changes (by bgamari): * status: new => closed * resolution: => fixed Comment: I believe the issue was merely a missing expected output file. Sorry about that! -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:27 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

19 Jun 19 Jun

9:10 a.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: Type: bug | Status: new Priority: high | Milestone: 8.0.1 Component: Runtime System | Version: 7.10.2 Resolution: | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746,#11972 | Differential Rev(s): Phab:D2189 Wiki Page: | -------------------------------------+------------------------------------- Changes (by thomie): * status: closed => new * resolution: fixed => Comment: This test is failing for `WAY=ghci`, `WAY=hpc`. You can run `make TEST=T11108 slow` to test all ways at the same time. {{{ --- ./rts/T11108.run/T11108.stdout.normalised 2016-06-19 15:10:31.680324016 +0200 +++ ./rts/T11108.run/T11108.run.stdout.normalised 2016-06-19 15:10:31.680324016 +0200 @@ -1,10 +1,24 @@ 1001 -1002 -1003 -1004 -1005 -1006 -1007 -1008 -1009 -1010 +"died" +1001 +"died" +1001 +"died" +1001 +"died" +1001 +"died" +1001 +"died" +1001 +"died" +1001 +"died" +1001 +"died" +1001 +"died" +"died" +"died" +"died" +"died" *** unexpected failure for T11108(ghci) }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:28 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

20 Jun 20 Jun

10:36 a.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: Type: bug | Status: new Priority: high | Milestone: 8.0.1 Component: Runtime System | Version: 7.10.2 Resolution: | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746,#11972 | Differential Rev(s): Phab:D2189 Wiki Page: | -------------------------------------+------------------------------------- Comment (by Thomas Miedema ): In [changeset:"1d938aa3373f464f46dd3806a21aa85dda764ec6/ghc" 1d938aa/ghc]: {{{ #!CommitTicketReference repository="ghc" revision="1d938aa3373f464f46dd3806a21aa85dda764ec6" Testsuite: mark tests expect broken * CgStaticPointers, GcStaticPointers, ListStaticPointers, TcStaticPointers01, TcStaticPointers02: #12207 * T11535: #12210 * ffi017/ffi021: #12209 * T11108: #11108 * T9646: #9646 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:29 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

4 Sep 4 Sep

11:01 p.m.

GHC

5 Sep 5 Sep

4:59 p.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: Type: bug | Status: new Priority: high | Milestone: 8.0.1 Component: Runtime System | Version: 7.10.2 Resolution: | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746,#11972 | Differential Rev(s): Phab:D2189, Wiki Page: | Phab:D2512 -------------------------------------+------------------------------------- Comment (by Ben Gamari ): In [changeset:"a7a960e43c34e40e1656fa1505605f756a44bb71/ghc" a7a960e4/ghc]: {{{ #!CommitTicketReference repository="ghc" revision="a7a960e43c34e40e1656fa1505605f756a44bb71" Make the test for #11108 less fragile This change should close #11108 by fixing the test case. This commit fixes two issues: * Make sure that each weak pointer we allocate has a constructor as the key, not a thunk. A failure to do so meant these weak pointers died prematurely on the 'ghci' WAY. * Don't print anything in the finalizer, because they are not guaranteed to run. Test Plan: validate Reviewers: austin, simonmar, erikd, bgamari Reviewed By: erikd, bgamari Subscribers: thomie Differential Revision: https://phabricator.haskell.org/D2512 GHC Trac Issues: #11108 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:31 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

GHC

17 Oct 17 Oct

2:53 a.m.

#11108: Weak references related crash -------------------------------------+------------------------------------- Reporter: Saulzar | Owner: Type: bug | Status: closed Priority: high | Milestone: 8.0.1 Component: Runtime System | Version: 7.10.2 Resolution: fixed | Keywords: Operating System: Linux | Architecture: | Unknown/Multiple Type of failure: Runtime crash | Test Case: Blocked By: | Blocking: Related Tickets: #11746,#11972 | Differential Rev(s): Phab:D2189, Wiki Page: | Phab:D2512 -------------------------------------+------------------------------------- Changes (by akio): * status: new => closed * resolution: => fixed Comment: I believe this ticket can be closed now. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11108#comment:32 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

3139

Age (days ago)

3473

Last active (days ago)

List overview

Download

33 comments

1 participants

[GHC] #11108: Weak references related crash

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

GHC

tags

participants (1)