Closed
Bug 637233
Opened 14 years ago
Closed 14 years ago
Android's VMPI atomic-integer operations return bad post-inc/dec values
Categories
(Tamarin Graveyard :: Virtual Machine, defect)
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: siwilkin, Unassigned)
References
Details
Attachments
(1 file, 1 obsolete file)
2.62 KB,
patch
|
edwsmith
:
superreview+
|
Details | Diff | Splinter Review |
Android's atomic-integer API returns pre-inc/dec values, whereas VMPI returns post-inc/dec values. This patch fixes VMPI's (trivial) conversion which is currently incorrect.
Reporter | ||
Comment 1•14 years ago
|
||
Attachment #515548 -
Flags: superreview?(edwsmith)
Attachment #515548 -
Flags: review?(kpalacz)
Reporter | ||
Comment 2•14 years ago
|
||
Attachment #515548 -
Attachment is obsolete: true
Attachment #515548 -
Flags: superreview?(edwsmith)
Attachment #515548 -
Flags: review?(kpalacz)
Reporter | ||
Updated•14 years ago
|
Attachment #515551 -
Flags: review?(kpalacz)
Updated•14 years ago
|
Attachment #515551 -
Flags: review?(fklockii)
Updated•14 years ago
|
Attachment #515551 -
Flags: review?(fklockii) → review+
Comment 3•14 years ago
|
||
changeset: 6009:b8166dce7f69
user: Simon Wilkinson <siwilkin>
summary: Bug 637233: Fixes return values for Android atomic-inc/dec-and-get (r=fklockii, r pending kpalacz)
http://hg.mozilla.org/tamarin-redux/rev/b8166dce7f69
Reporter | ||
Updated•14 years ago
|
Attachment #515551 -
Flags: superreview?(edwsmith)
Attachment #515551 -
Flags: review?(kpalacz)
Attachment #515551 -
Flags: review+
Comment 4•14 years ago
|
||
I think my comment on bug 629435 actually applies here, not there.. copied:
> I see that VMPI_atomicIncAndGet32WithBarrier() is being called in a loop to
> make sure the right number of increments happen, in ST_vmpi_threads.st. But
> the test doesn't check the exact correctness of the returned value (before or
> after value)
> each time.
>
> worth adding more tests?
Updated•14 years ago
|
Attachment #515551 -
Flags: superreview?(edwsmith) → superreview+
Reporter | ||
Comment 5•14 years ago
|
||
(In reply to comment #4)
> I think my comment on bug 629435 actually applies here, not there.. copied:
>
> > I see that VMPI_atomicIncAndGet32WithBarrier() is being called in a loop to
> > make sure the right number of increments happen, in ST_vmpi_threads.st. But
> > the test doesn't check the exact correctness of the returned value (before or
> > after value)
> > each time.
> >
> > worth adding more tests?
Seeing as though I messed it up so royally the first time around, it's probably worth adding a selftest.
Reporter | ||
Updated•14 years ago
|
Status: NEW → RESOLVED
Closed: 14 years ago
Resolution: --- → FIXED
You need to log in
before you can comment on or make changes to this bug.
Description
•