diff options
author | Simon Marlow <marlowsd@gmail.com> | 2017-06-03 20:26:13 +0100 |
---|---|---|
committer | Simon Marlow <marlowsd@gmail.com> | 2017-06-08 08:38:09 +0100 |
commit | 598472908ebb08f6811b892f285490554c290ae3 (patch) | |
tree | 84079aceefe1a7eacda507f104c0f0d4d8c12417 /rts/HeapStackCheck.cmm | |
parent | bca56bd040de64315564cdac4b7e943fc8a75ea0 (diff) | |
download | haskell-598472908ebb08f6811b892f285490554c290ae3.tar.gz |
Fix a lost-wakeup bug in BLACKHOLE handling (#13751)
Summary:
The problem occurred when
* Threads A & B evaluate the same thunk
* Thread A context-switches, so the thunk gets blackholed
* Thread C enters the blackhole, creates a BLOCKING_QUEUE attached to
the blackhole and thread A's `tso->bq` queue
* Thread B updates the blackhole with a value, overwriting the BLOCKING_QUEUE
* We GC, replacing A's update frame with stg_enter_checkbh
* Throw an exception in A, which ignores the stg_enter_checkbh frame
Now we have C blocked on A's tso->bq queue, but we forgot to check the
queue because the stg_enter_checkbh frame has been thrown away by the
exception.
The solution and alternative designs are discussed in Note [upd-black-hole].
This also exposed a bug in the interpreter, whereby we were sometimes
context-switching without calling `threadPaused()`. I've fixed this
and added some Notes.
Test Plan:
* `cd testsuite/tests/concurrent && make slow`
* validate
Reviewers: niteria, bgamari, austin, erikd
Reviewed By: erikd
Subscribers: rwbarton, thomie
GHC Trac Issues: #13751
Differential Revision: https://phabricator.haskell.org/D3630
Diffstat (limited to 'rts/HeapStackCheck.cmm')
-rw-r--r-- | rts/HeapStackCheck.cmm | 24 |
1 files changed, 0 insertions, 24 deletions
diff --git a/rts/HeapStackCheck.cmm b/rts/HeapStackCheck.cmm index 001e5f6299..85fb1cbef6 100644 --- a/rts/HeapStackCheck.cmm +++ b/rts/HeapStackCheck.cmm @@ -230,30 +230,6 @@ stg_gc_prim_p_ll } /* ----------------------------------------------------------------------------- - stg_enter_checkbh is just like stg_enter, except that we also call - checkBlockingQueues(). The point of this is that the GC can - replace an stg_marked_upd_frame with an stg_enter_checkbh if it - finds that the BLACKHOLE has already been updated by another - thread. It would be unsafe to use stg_enter, because there might - be an orphaned BLOCKING_QUEUE now. - -------------------------------------------------------------------------- */ - -/* The stg_enter_checkbh frame has the same shape as an update frame: */ - -INFO_TABLE_RET ( stg_enter_checkbh, RET_SMALL, - UPDATE_FRAME_FIELDS(W_,P_,info_ptr,ccs,p2,updatee)) - return (P_ ret) -{ - foreign "C" checkBlockingQueues(MyCapability() "ptr", - CurrentTSO); - - // we need to return updatee now. Note that it might be a pointer - // to an indirection or a tagged value, we don't know which, so we - // need to ENTER() rather than return(). - ENTER(updatee); -} - -/* ----------------------------------------------------------------------------- Info tables for returning values of various types. These are used when we want to push a frame on the stack that will return a value to the frame underneath it. |