[Openchrome-devel] [Bug 93675] Segfault in xf86DestroyI2CDevRec

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Tue Feb 2 22:24:45 PST 2016


https://bugs.freedesktop.org/show_bug.cgi?id=93675

--- Comment #12 from Kyle Guinn <elyk03 at gmail.com> ---
(In reply to Kevin Brace from comment #11)
> Do you have ideas on how this bug can be fixed without having to patch xorg
> (like what your patch does)?

Yes, it should just be a matter of rearranging the code.  We have to do the
Create only after one of the probes succeeded and we know which bus to use. 
I'm attaching a patch of how it should look, but I have not yet tested it.


> Over at Bug 91966, I recently uploaded patches that will scan for the
> existence of VT1632A regardless whether or not it is actually hooked up.
> As you saw from the source code snippets I showed, currently VT1632A is
> scanned only when P4M800 Pro and related chpsets are used. 
> The patches I uploaded will let OpenChrome scan several I2C buses to figure
> out what is connected.
> This became necessary because the bug reporter's system has a DVI-I
> connector (a DVI connector with DVI and VGA coming out, but shares the I2C
> bus between the two).
> Unfortunately, he was having strange freezes with the second I2C bus
> (OpenChrome calls it I2C bus 2).
> I am wondering if the instability the bug reporter reported is related to
> what you called "Create without a matching Destroy."

Not likely, unless it's triggering the same segfault, but I wouldn't
characterize a segfault as "freezing".  I haven't read through that bug yet.

The Create/Destroy pair is just a calloc/free pair.  If you omit the Destroy,
then free isn't called and you have a memory leak.  By itself that shouldn't
cause instability.  My guess is it's a completely separate issue and
something's wrong with the scanning.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/openchrome-devel/attachments/20160203/51eb4d66/attachment.html>


More information about the Openchrome-devel mailing list