[openib-general] [PATCHv2] osm: OSM crash TRIVIAL bug fix

[Hal Rosenstock]

Hi Yevgeny,

On Mon, 2006-08-21 at 11:27, Yevgeny Kliteynik wrote:
> Hi Hal.
> 
> My answers below. 
> I did check the code using the setup that discovered it.
> 
> This patch should make its way into both trunk and OFED 1.1 branch.

It's been on both since 8/14.

-- Hal

> Please let me know if there is anything else required for it.
> 
> Thanks,
> 
> Yevgeny
> 
> On Thu, 2006-08-17 at 09:35 -0400, Hal Rosenstock wrote:
> > Hi Yevgeny,
> > 
> > On Thu, 2006-08-17 at 09:28, Yevgeny Kliteynik wrote:
> > > Hi Hal.
> > > 
> > > > This line wrapped so there is something wrong with your mailer.
> > > 
> > > I'm using a different mailer now, so I hope that it's ok now.
> > 
> > Guess we'll see with your next patch with a long line...
> > 
> > > > > +  m->v = NULL; /* just make sure we do not point tofree'd madw */
> > > > 
> > > > Also, is this line really needed (and if so why) ? I know you did say 
> > > > "it cleans up old pointers to retrieved madw" but this shouldn't be 
> > > > accessed, right ?
> > > 
> > > You're right, it shouldn't be accessed.
> > 
> > Does the fix checked in work as is now ? Did you reverify ?
> 
> Yes, I did.
> 
> > > But generally, it's a good practice to assign a null to 
> > > any pointer that points to a freed memory, and should not
> > > be in use any more.
> > 
> > It's also good practice that when an issue is found in one place to look
> > for other occurences of the same issue.
> > 
> > I'm also not sure this is the general approach that OpenSM takes.
> 
> Right, I will try to clean these areas once I get to read them.
> 
> > -- Hal
> > 
> > > > Also, if this is added here, there are other places where the same
> > > > thing should be done ?
> > > 
> > > I just examined this area of code, so this is what I saw.
> >