[openib-general] [PATCH] diags/saquery: fix node_desc.description as string usages

Michael S. Tsirkin mst at mellanox.co.il
Tue Oct 31 07:06:50 PST 2006


Quoting r. Sasha Khapyorsky <sashak at voltaire.com>:
> Subject: Re: [PATCH] diags/saquery: fix node_desc.description as string usages
> 
> On 15:46 Tue 31 Oct     , Michael S. Tsirkin wrote:
> > Quoting r. Sasha Khapyorsky <sashak at voltaire.com>:
> > > > > Look at this for example:
> > > > > http://sourceware.org/ml/bug-glibc/2005-02/msg00123.html
> > > > 
> > > > Hmm, yea. Do you understand the answer there?
> > > > Does not make sense to me ...
> > > 
> > > I less care about the answer, but more about valgrind output (btw
> > > cannot see this with glibc-2.5) and similar "corner case" issues.
> > 
> > Well, I understand what you are saying, but whether a work-around is worth it
> > depends on whether the issue is still relevant in distros in use today
> 
> as well as with VC++/unknown DDK version?

Again, should be fine according to VC++ docs.
Whether there's a bug in this function, or some other function, is
an open question but I don't really see why make an exception here.

> 
> > - the
> > message you quote is from 2005.  After all, you never know whether some other
> > piece of code triggers a compiler bug in some rare case on some outdated distro.
> 
> Right, but this one is "known".
> 
> > Can you run valgrind on the simple test and check?
> 
> Yes, it was fine with glibc-2.5.
> 
> > In any case, I think we need a commment so we can clean this up
> > sometime in the future.
> 
> IMO it is better to just truncate original buffer by putting '\0' at the
> end and don't think too much about "possible impacts", but feel free to
> submit the patch.

I would just use the %.*s precision and not worry - its probably fixed in all
distros already.

-- 
MST




More information about the general mailing list