Bug 54258 – Make class Mutex work for Pthreads implementation with recursive mutexes

• Component: Web Template Framework
• Status: RESOLVED INVALID

[Daniel Bates]

Some Pthreads implementations may not support non-recursive mutexes (i.e. mutex type PTHREAD_MUTEX_NORMAL). For systems with such an implementation we should make the class Mutex mimic the behavior it has when compiled on systems that have a Pthreads implementation that support  mutex type PTHREAD_MUTEX_NORMAL.

[Daniel Bates]

Created attachment 82074 [details]
Patch

[Daniel Bates]

Created attachment 82077 [details]
Patch

[WebKit Review Bot]

Attachment 82077 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'Source/JavaScriptCore/ChangeLog', u'Source..." exit_code: 1

Source/JavaScriptCore/wtf/ThreadingPthreads.cpp:389:  Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.  [readability/comparison_to_zero] [5]
Total errors found: 1 in 4 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Daniel Bates]

(In reply to comment #3)
> Attachment 82077 [details] did not pass style-queue:
> [...]
> Source/JavaScriptCore/wtf/ThreadingPthreads.cpp:389:  Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.  [readability/comparison_to_zero] [5]
> Total errors found: 1 in 4 files

On success pthread_cond_timedwait() explicitly returns 0. I thought to preserve this explicit mention of 0. I can change code to read "return !pthread_cond_timedwait(...)".

[Alexey Proskuryakov]

Comment on attachment 82077 [details]
Patch

I think that this has race conditions in updating m_recursionCount. In unlock(), you decrease the count after unlocking, but another thread can already take this mutex, and be increasing the count.

pthread_cond_wait() temporarily releases the mutex - should the count be decreased, too?

Following tryLock() was just too difficult for me.

Can you name the platform this is needed for? Can a proper mutex be added to its pthreads implementation instead?

[Daniel Bates]

A proper mutex was added to the pthreads implementation for the platform.