cdsspec-compiler.git
12 years agomodel: use proper size_t printf format
Brian Norris [Sun, 7 Oct 2012 22:02:16 +0000 (15:02 -0700)]
model: use proper size_t printf format

12 years agoschedule: add is_enabled() function
Brian Norris [Sun, 7 Oct 2012 21:57:42 +0000 (14:57 -0700)]
schedule: add is_enabled() function

The release sequence functionality needs to check whether a Thread is
currently enabled (rather than finding this information in a particular
Node).

Also, rename the (private) 'is_enabled' field to avoid a name conflict.

12 years agoMerge branch 'master' of ssh://demsky.eecs.uci.edu/home/git/model-checker
Brian Demsky [Sat, 6 Oct 2012 01:50:24 +0000 (18:50 -0700)]
Merge branch 'master' of ssh://demsky.eecs.uci.edu/home/git/model-checker

12 years agocreate enumeration for enabled information...switch from bools to the enumeration
Brian Demsky [Sat, 6 Oct 2012 00:15:49 +0000 (17:15 -0700)]
create enumeration for enabled information...switch from bools to the enumeration

12 years agocyclegraph: fix indentation
Brian Norris [Sat, 6 Oct 2012 00:07:49 +0000 (17:07 -0700)]
cyclegraph: fix indentation

12 years agocyclegraph: flag cycles for reflexive edges
Brian Norris [Sat, 6 Oct 2012 00:03:01 +0000 (17:03 -0700)]
cyclegraph: flag cycles for reflexive edges

The cyclegraph shouldn't fail (i.e., ASSERT()) when reflexive edges are
added; instead, they should be declared as cycles.

12 years ago1) more comments
Brian Demsky [Fri, 5 Oct 2012 02:44:14 +0000 (19:44 -0700)]
1) more comments
2) rename is_synchronizing with
3) realize that reordering is only necessary if we create a synchronization...not needed to break one...

12 years agoMerge branch 'master' of ssh://demsky.eecs.uci.edu/home/git/model-checker
Brian Demsky [Fri, 5 Oct 2012 00:37:38 +0000 (17:37 -0700)]
Merge branch 'master' of ssh://demsky.eecs.uci.edu/home/git/model-checker

12 years agochanges
Brian Demsky [Fri, 5 Oct 2012 00:37:12 +0000 (17:37 -0700)]
changes

12 years agonodestack: destroy Nodes properly in ~NodeStack()
Brian Norris [Thu, 4 Oct 2012 21:59:44 +0000 (14:59 -0700)]
nodestack: destroy Nodes properly in ~NodeStack()

12 years agonodestack: use initializer list
Brian Norris [Thu, 4 Oct 2012 21:57:37 +0000 (14:57 -0700)]
nodestack: use initializer list

12 years agomodel: remove unnecessary boolean variable
Brian Norris [Thu, 4 Oct 2012 19:11:05 +0000 (12:11 -0700)]
model: remove unnecessary boolean variable

12 years agomodel: assign "pending" return values in release_seq_heads()
Brian Norris [Thu, 4 Oct 2012 19:42:13 +0000 (12:42 -0700)]
model: assign "pending" return values in release_seq_heads()

Now that the lazy pending release sequences list is improved, return
more information from release_seq_heads().

Note that this change also affects the behavior of the last double loop;
now, even after discovering write(s) from one thread that might break the
release sequence, I continue exploring other threads. This has some pros
and cons:

Con:
* may cause more work in exploring many threads without gaining any
  extra information

Pro:
* may help discover a write that *certainly* breaks the sequence,
  whereas previously, we might have exited with an uncertain status
* builds a more thorough, complete set of writes that might break the
  release sequence

The last point is the main intention, since we will require a complete
set for end-of-execution fixups. The others balance out to a net benefit
(and possibly a bugfix).

12 years agomodel: update pending release sequence list type
Brian Norris [Thu, 4 Oct 2012 19:15:14 +0000 (12:15 -0700)]
model: update pending release sequence list type

In working toward proper resolution of pending release sequences, I will
change the pending release sequence list again. This time, a single list
entry will have the ability to hold all important info regarding a
single release sequence. For now, I only utilize it to store the
'acquire' operation, so there should be no real change in behavior yet.

12 years agomodel: add release_seq structure definition
Brian Norris [Thu, 4 Oct 2012 18:15:39 +0000 (11:15 -0700)]
model: add release_seq structure definition

12 years agofix bug...can't mo_check_promises until we're done resolving promises for a write...
Brian Demsky [Thu, 4 Oct 2012 08:32:45 +0000 (01:32 -0700)]
fix bug...can't mo_check_promises until we're done resolving promises for a write back in time

12 years agolocal commit... bug that prunes too many executions
Brian Demsky [Thu, 4 Oct 2012 08:21:11 +0000 (01:21 -0700)]
local commit...  bug that prunes too many executions

12 years agoMerge branch 'master' of /home/git/model-checker
Brian Demsky [Thu, 4 Oct 2012 00:28:26 +0000 (17:28 -0700)]
Merge branch 'master' of /home/git/model-checker

12 years agofix low hanging fruit when profiling...
Brian Demsky [Thu, 4 Oct 2012 00:26:43 +0000 (17:26 -0700)]
fix low hanging fruit when profiling...

dump typo was turning off optimized version of checkpointing...option 2 is by far fastest
was spending most of the time freeing an oversized hashtable

12 years agofixup more id vs. thread_id_t
Brian Norris [Thu, 4 Oct 2012 00:12:59 +0000 (17:12 -0700)]
fixup more id vs. thread_id_t

12 years agoaction: return synchronization status for ModelAction::read_from()
Brian Norris [Tue, 25 Sep 2012 23:29:14 +0000 (16:29 -0700)]
action: return synchronization status for ModelAction::read_from()

The ModelChecker may need to know if any synchronization was performed in
read_from().

Note that the return status is not used yet.

12 years agoaction: remove clockvector flag from print() method
Brian Norris [Wed, 3 Oct 2012 23:29:59 +0000 (16:29 -0700)]
action: remove clockvector flag from print() method

We don't actually need to tell ModelAction::print() when to print the
associated clock vector. The introduction of the 'print_cv' boolean flag
was to fix a separate bug that has been fixed. Now the clock vector can
be printed unconditionally, when it exists.

12 years agomodel: rename release_seq_head() -> release_seq_heads()
Brian Norris [Wed, 3 Oct 2012 23:07:54 +0000 (16:07 -0700)]
model: rename release_seq_head() -> release_seq_heads()

It may return multiple heads, so make it plural.

12 years agomodel: do not assume THREAD_FINISH is always the last action
Brian Norris [Wed, 3 Oct 2012 22:36:26 +0000 (15:36 -0700)]
model: do not assume THREAD_FINISH is always the last action

It's possible that there will be special fixup ModelActions added, so
just use the Thread::is_complete() method to check for a completed
Thread.

12 years agofixup usage of int vs. thread_id_t
Brian Norris [Wed, 3 Oct 2012 22:26:52 +0000 (15:26 -0700)]
fixup usage of int vs. thread_id_t

12 years agomodel: add const qualifier to get_thread()
Brian Norris [Wed, 3 Oct 2012 22:08:28 +0000 (15:08 -0700)]
model: add const qualifier to get_thread()

12 years agomodel: rearrange conditionals, fixup take_step()
Brian Norris [Wed, 3 Oct 2012 21:50:30 +0000 (14:50 -0700)]
model: rearrange conditionals, fixup take_step()

12 years agoMerge remote-tracking branch 'origin/master'
Brian Norris [Wed, 3 Oct 2012 20:54:22 +0000 (13:54 -0700)]
Merge remote-tracking branch 'origin/master'

12 years agomodeltypes: move small typedefs to own header
Brian Norris [Wed, 3 Oct 2012 20:17:47 +0000 (13:17 -0700)]
modeltypes: move small typedefs to own header

To prevent some unnecessary inter-header dependencies, we can move some
simple typedefs to a modeltypes.h header.

12 years agomodel: don't include schedule.h
Brian Norris [Wed, 3 Oct 2012 20:16:49 +0000 (13:16 -0700)]
model: don't include schedule.h

12 years agonodestack: move has_priority() out of header
Brian Norris [Wed, 3 Oct 2012 20:50:51 +0000 (13:50 -0700)]
nodestack: move has_priority() out of header

12 years agomodel: move get_thread() implementations out of header
Brian Norris [Wed, 3 Oct 2012 20:37:02 +0000 (13:37 -0700)]
model: move get_thread() implementations out of header

12 years agomodel: move get_current_thread() implementation out of header
Brian Norris [Wed, 3 Oct 2012 20:16:00 +0000 (13:16 -0700)]
model: move get_current_thread() implementation out of header

12 years agoMerge branch 'master' of /home/git/model-checker
Brian Demsky [Wed, 3 Oct 2012 19:38:27 +0000 (12:38 -0700)]
Merge branch 'master' of /home/git/model-checker

12 years agoa number of fixes to add missing mo_graph edges to speed up detection of infeasible
Brian Demsky [Wed, 3 Oct 2012 19:37:53 +0000 (12:37 -0700)]
a number of fixes to add missing mo_graph edges to speed up detection of infeasible
traces

12 years agomodel: update switch_to_master() comment
Brian Norris [Wed, 3 Oct 2012 19:37:22 +0000 (12:37 -0700)]
model: update switch_to_master() comment

12 years agoMakefile: eliminate malloc.c warning
Brian Norris [Wed, 3 Oct 2012 19:13:03 +0000 (12:13 -0700)]
Makefile: eliminate malloc.c warning

malloc.c: In function ‘sys_trim’:
malloc.c:4289:20: warning: unused variable ‘newsize’ [-Wunused-variable]
malloc.c: In function ‘destroy_mspace’:
malloc.c:5452:13: warning: unused variable ‘base’ [-Wunused-variable]

12 years agomymemory: fix indentation for ModelAlloc
Brian Norris [Wed, 3 Oct 2012 19:06:20 +0000 (12:06 -0700)]
mymemory: fix indentation for ModelAlloc

12 years agorename 'MyAlloc' to 'ModelAlloc'
Brian Norris [Wed, 3 Oct 2012 18:50:35 +0000 (11:50 -0700)]
rename 'MyAlloc' to 'ModelAlloc'

12 years agomymemory: reword comments
Brian Norris [Wed, 3 Oct 2012 18:37:26 +0000 (11:37 -0700)]
mymemory: reword comments

12 years agosnapshot: rename 'SnapShot' -> 'Snapshot'
Brian Norris [Wed, 3 Oct 2012 18:29:01 +0000 (11:29 -0700)]
snapshot: rename 'SnapShot' -> 'Snapshot'

This capitalization always messes me up

12 years agocommon: remove excess semicolon
Brian Norris [Wed, 3 Oct 2012 18:26:17 +0000 (11:26 -0700)]
common: remove excess semicolon

12 years agomymemory: define SNAPSHOTALLOC appropriately
Brian Norris [Wed, 3 Oct 2012 18:15:05 +0000 (11:15 -0700)]
mymemory: define SNAPSHOTALLOC appropriately

SNAPSHOTALLOC can be filled in with meaningful allocators, rather than
using the global default allocators.

12 years agoreplace malloc/calloc/free with snapshot_{malloc/calloc/free}
Brian Norris [Wed, 3 Oct 2012 18:08:36 +0000 (11:08 -0700)]
replace malloc/calloc/free with snapshot_{malloc/calloc/free}

We need to explicitly declare when we are requesting snapshotting
memory, and we need to avoid using malloc, etc., directly. For now, this
has no functional change, as the 'snapshot_*' functions just call the
stdlib functions, but soon we may switch allocators to use a private
heap.

12 years agomymemory: make snapshot_{malloc,calloc,free} externally linkable
Brian Norris [Wed, 3 Oct 2012 18:05:36 +0000 (11:05 -0700)]
mymemory: make snapshot_{malloc,calloc,free} externally linkable

As we begin to use these functions, they can't be simply declared
statically in the header; they need to be linkable in hashtable.h, for
instance.

12 years agoaction: reword comments
Brian Norris [Wed, 3 Oct 2012 17:12:56 +0000 (10:12 -0700)]
action: reword comments

Why all the ...?

12 years agoextra file committed accidentally
Brian Demsky [Wed, 3 Oct 2012 10:04:21 +0000 (03:04 -0700)]
extra file committed accidentally

12 years agomissing change
Brian Demsky [Wed, 3 Oct 2012 10:03:02 +0000 (03:03 -0700)]
missing change

12 years agomspace_malloc will call into mmap if it runs out of memory... this does not play...
Brian Demsky [Wed, 3 Oct 2012 09:57:32 +0000 (02:57 -0700)]
mspace_malloc will call into mmap if it runs out of memory...  this does not play well with
snapshotting the mspace of course...

turn off the mmap option and give us a bigger heap...
add assertions to malloc/calloc/realloc if they return NULL (which is now possible with MMAP turned off)

12 years agoerror
Brian Demsky [Wed, 3 Oct 2012 08:28:12 +0000 (01:28 -0700)]
error

12 years agorandom memory leak fixes and memory access fixes
Brian Demsky [Wed, 3 Oct 2012 08:20:48 +0000 (01:20 -0700)]
random memory leak fixes and memory access fixes

12 years agomodel: debug print - pending release sequences
Brian Norris [Tue, 2 Oct 2012 01:15:12 +0000 (18:15 -0700)]
model: debug print - pending release sequences

12 years agotest: add a "pending release sequences" test
Brian Norris [Wed, 3 Oct 2012 01:52:43 +0000 (18:52 -0700)]
test: add a "pending release sequences" test

This helps test the end-of-execution "pending release sequence"
functionality, as well as other release-sequence-related code. It also
uncovered the regression that was just reverted prior to this.

12 years ago(bug) revert JOIN/LOCK simplification
Brian Norris [Wed, 3 Oct 2012 01:48:05 +0000 (18:48 -0700)]
(bug) revert JOIN/LOCK simplification

A test case throws an assertion when these changes are included. Maybe
I'll make a proper fix later, but for now, I mostly revert these.

Revert "model: remove obsolete ModelChecker::do_complete_join()"
This reverts commit cef10a2b49af5da16ffe59c5b9ddd210c668fbac.

Revert "model: unify JOIN- and LOCK-related sleep/wake code"
This reverts commit 620ae95ce4fed006d18a41b6ccfd949d7e77f677.

12 years agomodel: one release sequence may help resolve another
Brian Norris [Wed, 26 Sep 2012 00:00:53 +0000 (17:00 -0700)]
model: one release sequence may help resolve another

12 years agomodel: THREAD_FINISH triggers release sequence check
Brian Norris [Wed, 3 Oct 2012 01:09:40 +0000 (18:09 -0700)]
model: THREAD_FINISH triggers release sequence check

A THREAD_FINISH action may ensure that a Thread is no longer able to
break release sequences, so all pending release sequences should be
checked.

12 years agoMerge branch 'master' of ssh://demsky.eecs.uci.edu/home/git/model-checker
Brian Demsky [Wed, 3 Oct 2012 01:07:02 +0000 (18:07 -0700)]
Merge branch 'master' of ssh://demsky.eecs.uci.edu/home/git/model-checker

12 years agoa bug fix
Brian Demsky [Wed, 3 Oct 2012 01:06:46 +0000 (18:06 -0700)]
a bug fix

12 years agoassign sequence numbers after initial processing
Brian Norris [Wed, 3 Oct 2012 00:54:26 +0000 (17:54 -0700)]
assign sequence numbers after initial processing

When re-ordering the execution of JOIN and LOCK operations (e.g., when
they attempt to execute when not enabled), the sequence numbers became
out of order. Also, RMW operations would cause sequence numbers to jump
by 2 every time.

This fix arranges initialization such that a ModelAction is only
assigned a sequence number after it passes some initial checks to ensure
that it will be executed and utilized immediately. Otherwise, it will
either be discarded or held pending until it is enabled.

12 years agoaction: edit some spacing
Brian Norris [Wed, 3 Oct 2012 00:58:01 +0000 (17:58 -0700)]
action: edit some spacing

12 years agomodel: remove obsolete ModelChecker::do_complete_join()
Brian Norris [Wed, 3 Oct 2012 00:29:56 +0000 (17:29 -0700)]
model: remove obsolete ModelChecker::do_complete_join()

This function was folded into the one place it is used, now.

12 years agomodel: unify JOIN- and LOCK-related sleep/wake code
Brian Norris [Wed, 3 Oct 2012 00:28:02 +0000 (17:28 -0700)]
model: unify JOIN- and LOCK-related sleep/wake code

We might as well use similar code paths for JOIN and LOCK operations
that must wait on another thread to re-enable it. The side benefit here
is that the JOIN action is now placed at the correct sequence point in
the execution trace.

12 years agomodel: refactor mutex thread-blocking code
Brian Norris [Wed, 3 Oct 2012 00:14:46 +0000 (17:14 -0700)]
model: refactor mutex thread-blocking code

There are a few functions that are used in suboptimal ways.

* use the ModelChecker::get_thread(ModelAction *) version of overloaded
  'get_thread()' function

* the add/remove interfaces were used inconsistently previously; for
  "waking" we directly utilized Scheduler::add_thread, whereas the
  "sleeping" case used ModelChecker::remove_thread

* use the Scheduler::sleep and Scheduler::wake functions for sleep/wake
  instead of explicitly adding/removing

12 years agoaction: don't print clock vector when it is invalid
Brian Norris [Tue, 2 Oct 2012 19:33:21 +0000 (12:33 -0700)]
action: don't print clock vector when it is invalid

Due to clock vector snapshotting, a ModelAction may hold a pointer to
an inavlid ClockVector. This fix allows callers of ModelAction::print()
to prevent it from printing a corrupted ClockVector.

12 years agosmall hack to optimize our common case for snapshotting and set config option to...
Brian Demsky [Tue, 2 Oct 2012 21:54:59 +0000 (14:54 -0700)]
small hack to optimize our common case for snapshotting and set config option to turn this back on.

hack assumes that same pages will be dirty across different executions so just copies them and doesn't reprotect them

12 years agofix bug.... not quite perfect for linux locks, but runs to completion...
Brian Demsky [Tue, 2 Oct 2012 21:29:10 +0000 (14:29 -0700)]
fix bug....  not quite perfect for linux locks, but runs to completion...

Revert "model: revert broken bugfix"

This reverts commit ae7fcd2e5e499f72d9d1530bdc293f4fbc5f0644.

12 years agoclockvector: snapshot the whole object
Brian Norris [Tue, 2 Oct 2012 19:17:16 +0000 (12:17 -0700)]
clockvector: snapshot the whole object

12 years agomodel: revert broken bugfix
Brian Norris [Tue, 2 Oct 2012 18:16:12 +0000 (11:16 -0700)]
model: revert broken bugfix

This reverts parts of the following commit:

 commit 7ba211ec7d1cacbaa3ef2027f1a0f534888f1708

The bugfix broke our basic test case (userprog)

12 years agomodel: fix "earliest diverge" NULL pointer exception
Brian Norris [Tue, 2 Oct 2012 18:02:46 +0000 (11:02 -0700)]
model: fix "earliest diverge" NULL pointer exception

12 years agostart support for drawing execution diagrams
Brian Demsky [Tue, 2 Oct 2012 09:09:13 +0000 (02:09 -0700)]
start support for drawing execution diagrams

fix some signed/unsigned warnings

12 years agoFixed bug breaking our consolidation of future values...
Brian Demsky [Tue, 2 Oct 2012 07:29:10 +0000 (00:29 -0700)]
Fixed bug breaking our consolidation of future values...

12 years agoTwo change:
Brian Demsky [Tue, 2 Oct 2012 04:48:34 +0000 (21:48 -0700)]
Two change:
(1) Bug fix...  Check that rf is non-null before calling
r_modification

(2) Add a hook into assert mechanism for the debugger

12 years agorename MYFREE -> model_free
Brian Norris [Tue, 2 Oct 2012 01:18:07 +0000 (18:18 -0700)]
rename MYFREE -> model_free

12 years agorename MYCALLOC -> model_calloc
Brian Norris [Tue, 2 Oct 2012 01:17:26 +0000 (18:17 -0700)]
rename MYCALLOC -> model_calloc

12 years agorename MYMALLOC -> model_malloc
Brian Norris [Tue, 2 Oct 2012 01:16:24 +0000 (18:16 -0700)]
rename MYMALLOC -> model_malloc

12 years agomodel: refactor/reword initialize_curr_action()
Brian Norris [Tue, 2 Oct 2012 00:37:02 +0000 (17:37 -0700)]
model: refactor/reword initialize_curr_action()

12 years agoclockvector: snapshot our clock vectors
Brian Norris [Tue, 2 Oct 2012 00:30:01 +0000 (17:30 -0700)]
clockvector: snapshot our clock vectors

Clock vectors need to be snapshotted

12 years agomymemory: add placeholder "snapshot_{calloc,malloc,free}" functions
Brian Norris [Tue, 2 Oct 2012 00:11:52 +0000 (17:11 -0700)]
mymemory: add placeholder "snapshot_{calloc,malloc,free}" functions

Eventually, we may want a separate heap for the model-checker's
snapshotted data (vs. user's snapshotted data), so create allocator
stubs for this. This allows us to begin annotating our malloc()'s
properly.

12 years agomodel: always re-calculate clock vectors
Brian Norris [Tue, 2 Oct 2012 00:20:34 +0000 (17:20 -0700)]
model: always re-calculate clock vectors

Clock vectors are mutable, via synchronization. We will need to both
snapshot and recompute clock vectors as they are generated. Right now,
we have some incorrect assumptions in place. This bugfix prevents stale
clock vectors (from previous executions) from polluting subsequent
executions.

12 years agoanother bug fix
Brian Demsky [Mon, 1 Oct 2012 23:27:49 +0000 (16:27 -0700)]
another bug fix

12 years agobug fix...recompute promises of RMW actions at divergence points
Brian Demsky [Mon, 1 Oct 2012 22:17:11 +0000 (15:17 -0700)]
bug fix...recompute promises of RMW actions at divergence points

12 years agomodel: utilize bad_synchronization flag
Brian Norris [Mon, 1 Oct 2012 20:29:54 +0000 (13:29 -0700)]
model: utilize bad_synchronization flag

12 years agomodel: add "bad synchronization" flag
Brian Norris [Mon, 1 Oct 2012 20:29:00 +0000 (13:29 -0700)]
model: add "bad synchronization" flag

Can be used for aborting a trace when an incorrectly-ordered
synchronization occurs.

12 years agoaction: synchronize_with - return status for out-of-order synchronization
Brian Norris [Mon, 1 Oct 2012 20:20:56 +0000 (13:20 -0700)]
action: synchronize_with - return status for out-of-order synchronization

"synchronize_with()" now can return a boolean status, rather than just
using an ASSERT(). This allows synchronize_with() to be called with
actions that are against the program trace order, then return a status
to signify success/failure. This will allow, for instance, release
sequence calculations to simply abort this single execution, not the
whole model-checker.

12 years agomodel: remove debug print
Brian Norris [Mon, 1 Oct 2012 17:55:23 +0000 (10:55 -0700)]
model: remove debug print

12 years agomissing commit of mo_graph changes
Brian Demsky [Sat, 29 Sep 2012 03:58:27 +0000 (20:58 -0700)]
missing commit of mo_graph changes

12 years agofix bug from moving read_from check_recency...check_recency had the assumption that...
Brian Demsky [Fri, 28 Sep 2012 10:16:00 +0000 (03:16 -0700)]
fix bug from moving read_from check_recency...check_recency had the assumption that the read_from field was set...

12 years agonodestack: add debug prints
Brian Norris [Thu, 27 Sep 2012 19:29:05 +0000 (12:29 -0700)]
nodestack: add debug prints

12 years agomodel: push mo_graph cycle check into release_seq code
Brian Norris [Thu, 27 Sep 2012 17:53:59 +0000 (10:53 -0700)]
model: push mo_graph cycle check into release_seq code

Release sequences traveral should only be undertaken if the mo_graph has
no cycles. Instead of just making this check a hack within
resolve_promises(), make it part of every release sequence exploration.

12 years agomodel: add RMW violation debug print
Brian Norris [Thu, 27 Sep 2012 17:53:37 +0000 (10:53 -0700)]
model: add RMW violation debug print

12 years agodoc: add release sequence notes
Brian Norris [Thu, 27 Sep 2012 17:15:28 +0000 (10:15 -0700)]
doc: add release sequence notes

12 years agodocs: move to subdirectory
Brian Norris [Thu, 27 Sep 2012 16:58:10 +0000 (09:58 -0700)]
docs: move to subdirectory

12 years agomodel: avoid infinite loop in release_seq_head()
Brian Norris [Wed, 26 Sep 2012 02:23:51 +0000 (19:23 -0700)]
model: avoid infinite loop in release_seq_head()

To understand the problem I'm solving here, note that resolve_promises()
calls ModelAction::read_from() which calls ModelChecker::release_seq_head().

Now, release_seq_head() has the basic assumption that the mo_graph
doesn't have cycles (or specifically, in this case it's important that
sequences of RMW's don't form loops in the reads-from relation).
Unfortunately, resolve_promises() can create such a cycle (by assigning
the "reads-from" value) before checking if that would create a cycle.
This will trigger a check to release_seq_head, which gets stuck in an
infinite loop...

The solution, at least for this targeted portion of code, is to:
(1) First add the relevant RMW edge, if possible
(2) If (1) didn't create mo_graph cycles:
       Then assign reads-from, check release sequences, update
       synchronization
(3) Calculate other normal mo_graph edges

This way, (2) will simply be skipped if we have cycles, avoiding the
problem.

12 years agomodel: bugfix - iterator naming conflict
Brian Norris [Wed, 26 Sep 2012 18:16:17 +0000 (11:16 -0700)]
model: bugfix - iterator naming conflict

There's already another iterator named 'it' in resolve_release_sequences().
Although this iterator probably will mask the earlier iterator (giving correct
behavior), it's probably safer to rename the later one for clarity.

12 years agomodel: re-check release sequences after THREAD_JOIN
Brian Norris [Tue, 25 Sep 2012 23:56:02 +0000 (16:56 -0700)]
model: re-check release sequences after THREAD_JOIN

If a THREAD-specific operation (from process_thread_action()) added any
synchronization, we should re-check all release sequences.

12 years agomodel: add current action work_queue comment
Brian Norris [Tue, 25 Sep 2012 23:59:30 +0000 (16:59 -0700)]
model: add current action work_queue comment

12 years agomodel: mutex synchronization -> re-check release sequences
Brian Norris [Thu, 20 Sep 2012 22:26:37 +0000 (15:26 -0700)]
model: mutex synchronization -> re-check release sequences

All synchronization updates should generate a re-check of release sequences.
This commit corrects this for mutexes.

12 years agomodel: trivial - rename 'updated' -> 'update'
Brian Norris [Tue, 25 Sep 2012 23:50:59 +0000 (16:50 -0700)]
model: trivial - rename 'updated' -> 'update'

12 years agomodel: bugfix - release sequences - handle Thread completion
Brian Norris [Tue, 25 Sep 2012 23:39:07 +0000 (16:39 -0700)]
model: bugfix - release sequences - handle Thread completion

A completed Thread cannot generate any new writes that would break release
sequences.

12 years agomodel: fixup spelling (and formatting) in comments
Brian Norris [Tue, 25 Sep 2012 22:33:51 +0000 (15:33 -0700)]
model: fixup spelling (and formatting) in comments

relese -> release