Re: [PATCH] drivers/xen/xenbus: Replace deprecated strcpy in xenbus_transaction_end

[Jürgen Groß <jgross@xxxxxxxx>]

On 13.10.25 10:22, Jan Beulich wrote:
> On 13.10.2025 09:36, Jürgen Groß wrote:
> > On 13.10.25 08:59, Jan Beulich wrote:
> > > On 12.10.2025 21:55, Thorsten Blum wrote:
> > > > --- a/drivers/xen/xenbus/xenbus_xs.c
> > > > +++ b/drivers/xen/xenbus/xenbus_xs.c
> > > > @@ -546,16 +546,13 @@ int xenbus_transaction_start(struct xenbus_transaction *t)
> > > >    EXPORT_SYMBOL_GPL(xenbus_transaction_start);
> > > >    
> > > >    /* End a transaction.
> > > > - * If abandon is true, transaction is discarded instead of committed.
> > > > + * If abort is true, transaction is discarded instead of committed.
> > > >     */
> > > > -int xenbus_transaction_end(struct xenbus_transaction t, int abort)
> > > > +int xenbus_transaction_end(struct xenbus_transaction t, bool abort)
> > > >    {
> > > >         char abortstr[2];
> > > >    
> > > > -	if (abort)
> > > > -               strcpy(abortstr, "F");
> > > > -       else
> > > > -               strcpy(abortstr, "T");
> > > While at least in principle a compiler might be able to transform this into
> > > code not using any library function at all, ...
> > >
> > > > +       strscpy(abortstr, abort ? "F" : "T");
> > > ... the use of a n on-standard function (without equivalent compiler builtin)
> > > doesn't permit this. IOW why not simply switch to e.g.
> > >
> > >       char abortstr[2] = { [0] = abort ? 'F' : 'T' };
> > I would even go further and drop abortstr[] completely:
> >
> > diff --git a/drivers/xen/xenbus/xenbus_xs.c b/drivers/xen/xenbus/xenbus_xs.c
> > index 528682bf0c7f..c891af7165f5 100644
> > --- a/drivers/xen/xenbus/xenbus_xs.c
> > +++ b/drivers/xen/xenbus/xenbus_xs.c
> > @@ -550,14 +550,8 @@ EXPORT_SYMBOL_GPL(xenbus_transaction_start);
> >     */
> >    int xenbus_transaction_end(struct xenbus_transaction t, int abort)
> >    {
> > -       char abortstr[2];
> > -
> > -       if (abort)
> > -               strcpy(abortstr, "F");
> > -       else
> > -               strcpy(abortstr, "T");
> > -
> > -       return xs_error(xs_single(t, XS_TRANSACTION_END, abortstr, NULL));
> > +       return xs_error(xs_single(t, XS_TRANSACTION_END, abort ? "F" : "T",
> > +                       NULL));
> >    }
> >    EXPORT_SYMBOL_GPL(xenbus_transaction_end);
> Hmm, which xs_single() indeed takes a const char *, it then casts away const-
> ness before handing to xs_talkv().
Yes, the cast is needed as xs_talkv() can handle reads and writes. No problem in
this case, as the string is only read by xs_talkv() (write type operation).


Juergen

Attachment: OpenPGP_0xB0DE9DD628BF132F.asc
Description: OpenPGP public key

Attachment: OpenPGP_signature.asc
Description: OpenPGP digital signature