Re: [Xen-devel] [PATCH for-4.6] xenstored: log tdb message via xenstored's logging mechanisms

[Wei Liu <wei.liu2@xxxxxxxxxx>]

On Thu, Jan 08, 2015 at 01:10:38PM +0000, Ian Campbell wrote:
> On Thu, 2015-01-08 at 13:04 +0000, Ian Campbell wrote:
> > On Thu, 2015-01-08 at 12:57 +0000, Wei Liu wrote:
> > > On Thu, Jan 08, 2015 at 12:33:29PM +0000, Ian Campbell wrote:
> > > > ping.
> > > > 
> > > > On Mon, 2014-12-15 at 13:18 +0000, Ian Campbell wrote:
> > > > > TDB provides us with a callback for this purpose. Use it in both
> > > > > xenstored and xs_tdb_dump.
> > > > > 
> > > > > While at it make the existing log() macro tollerate memory failures.
> > > > > 
> > > > > Signed-off-by: Ian Campbell <ian.campbell@xxxxxxxxxx>
> > > > > ---
> > > > >  tools/xenstore/xenstored_core.c |   39 
> > > > > +++++++++++++++++++++++++++++++++------
> > > > >  tools/xenstore/xs_tdb_dump.c    |   12 +++++++++++-
> > > > >  2 files changed, 44 insertions(+), 7 deletions(-)
> > > > > 
> > > > > diff --git a/tools/xenstore/xenstored_core.c 
> > > > > b/tools/xenstore/xenstored_core.c
> > > > > index 4eaff57..3fd9a20 100644
> > > > > --- a/tools/xenstore/xenstored_core.c
> > > > > +++ b/tools/xenstore/xenstored_core.c
> > > > > @@ -89,9 +89,14 @@ static void check_store(void);
> > > > >  #define log(...)                                                     
> > > > > \
> > > > >       do {                                                            
> > > > > \
> > > > >               char *s = talloc_asprintf(NULL, __VA_ARGS__);           
> > > > > \
> > > > > -             trace("%s\n", s);                                       
> > > > > \
> > > > > -             syslog(LOG_ERR, "%s",  s);                              
> > > > > \
> > > > > -             talloc_free(s);                                         
> > > > > \
> > > > > +             if (s) {                                                
> > > > > \
> > > > > +                     trace("%s\n", s);                               
> > > > > \
> > > > > +                     syslog(LOG_ERR, "%s",  s);                      
> > > > > \
> > > > > +                     talloc_free(s);                                 
> > > > > \
> > > > > +             } else {                                                
> > > > > \
> > > > > +                     trace("talloc failure during logging\n");       
> > > > > \
> > > > > +                     syslog(LOG_ERR, "talloc failure during 
> > > > > logging\n"); \
> > > > > +             }                                                       
> > > > > \
> > > 
> > > talloc_free can tolerate NULL pointer.
> > 
> > Good point, I'll fix.
> 
> Hrm, it tolerates but it does return failure (-1) instead of success in
> that case.
> 
> Literally nowhere in our code base actually checks the return value from
> talloc_free, but nonetheless given that I think it would prefer to keep
> the talloc_free where it is and therefore only pass a valid pointer to
> it.
> 

Fair enough.

Wei.

> Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel