Re: [Qemu-devel] [PATCH] 9p: Convert use of atoi to qemu_strtol to allow error checking

[Greg Kurz]

On Mon, 12 Mar 2018 13:08:29 +0000
Daniel P. Berrangé <address@hidden> wrote:

> On Mon, Mar 12, 2018 at 02:02:29PM +0100, Greg Kurz wrote:
> > On Mon, 12 Mar 2018 07:12:52 -0500
> > Eric Blake <address@hidden> wrote:
> >   
> > > On 03/11/2018 03:12 PM, Nia Alarie wrote:  
> > > > Signed-off-by: Nia Alarie <address@hidden>
> > > > ---
> > > >   hw/9pfs/9p.c | 12 ++++++++++--
> > > >   1 file changed, 10 insertions(+), 2 deletions(-)
> > > >     
> > >   
> > > >       } else if (perm & P9_STAT_MODE_LINK) {
> > > > -        int32_t ofid = atoi(extension.data);
> > > > -        V9fsFidState *ofidp = get_fid(pdu, ofid);
> > > > +        long ofid;
> > > > +        V9fsFidState *ofidp;
> > > > +
> > > > +        if (qemu_strtol(extension.data, NULL, 10, &ofid) ||
> > > > +            ofid > INT32_MAX || ofid < INT32_MIN) {    
> > > 
> > > Dan has a pending patch that will add qemu_strtoi, which might be a 
> > > nicer fit for this situation:
> > > 
> > > https://lists.gnu.org/archive/html/qemu-devel/2018-02/msg00952.html
> > > 
> > > int32_t is not necessarily int, but all platforms that compile qemu have 
> > > 'int32_t' and 'int' both at 32 bits, so it's simpler to change to 'int 
> > > ofid' and use Dan's function than it is to parse to long and then do 
> > > bounds checking.  Except that Dan still needs to post an updated version 
> > > of his patch...
> > >   
> > 
> > I wasn't aware of Dan's patch but I agree it would result in a nicer
> > change for 9p. This being said, tomorrow is soft freeze... is there
> > a chance Dan's patch reaches master anytime soon ?  
> 
> I just sent an update of my series. If it gets acked by Eric, I'd be able
> to send a pull request today.
> 
> Regards,
> Daniel

Great !

Nia,

Could you please respin your patch on top of Daniel's series, using
qemu_strtoui() ?

Thanks,

--
Greg