RE: [PATCH v5 4/5] common/domain: add a domain context record for shared_info...

["Durrant, Paul" <pdurrant@xxxxxxxxxxxx>]

> -----Original Message-----
> From: Jan Beulich <jbeulich@xxxxxxxx>
> Sent: 22 May 2020 15:34
> To: Paul Durrant <paul@xxxxxxx>
> Cc: xen-devel@xxxxxxxxxxxxxxxxxxxx; Durrant, Paul <pdurrant@xxxxxxxxxxxx>; 
> Ian Jackson
> <ian.jackson@xxxxxxxxxxxxx>; Wei Liu <wl@xxxxxxx>; Andrew Cooper 
> <andrew.cooper3@xxxxxxxxxx>; George
> Dunlap <george.dunlap@xxxxxxxxxx>; Julien Grall <julien@xxxxxxx>; Stefano 
> Stabellini
> <sstabellini@xxxxxxxxxx>
> Subject: RE: [EXTERNAL] [PATCH v5 4/5] common/domain: add a domain context 
> record for shared_info...
> 
> CAUTION: This email originated from outside of the organization. Do not click 
> links or open
> attachments unless you can confirm the sender and know the content is safe.
> 
> 
> 
> On 21.05.2020 18:19, Paul Durrant wrote:
> > @@ -1649,6 +1650,70 @@ int continue_hypercall_on_cpu(
> >      return 0;
> >  }
> >
> > +static int save_shared_info(const struct domain *d, struct domain_context 
> > *c,
> > +                            bool dry_run)
> > +{
> > +    struct domain_shared_info_context ctxt = {
> > +#ifdef CONFIG_COMPAT
> > +        .flags = has_32bit_shinfo(d) ? DOMAIN_SAVE_32BIT_SHINFO : 0,
> > +#endif
> > +        .buffer_size = sizeof(shared_info_t),
> 
> But this size varies between native and compat.
> 
> > +static int load_shared_info(struct domain *d, struct domain_context *c)
> > +{
> > +    struct domain_shared_info_context ctxt;
> > +    size_t hdr_size = offsetof(typeof(ctxt), buffer);
> > +    unsigned int i;
> > +    int rc;
> > +
> > +    rc = DOMAIN_LOAD_BEGIN(SHARED_INFO, c, &i);
> > +    if ( rc )
> > +        return rc;
> > +
> > +    if ( i ) /* expect only a single instance */
> > +        return -ENXIO;
> > +
> > +    rc = domain_load_data(c, &ctxt, hdr_size);
> > +    if ( rc )
> > +        return rc;
> > +
> > +    if ( ctxt.buffer_size != sizeof(shared_info_t) )
> > +        return -EINVAL;
> 
> While on the save side things could be left as they are (yet
> I'd prefer a change), this should be flexible enough to allow
> at least the smaller compat size as well in the compat case.

Ok, I guess we can optimize the buffer size down if only the compat version is 
needed. Seems like slightly pointless complexity though.

> I wonder whether any smaller sizes might be acceptable, once
> again with the rest getting zero-padded.
> 

If the need arises to zero extend an older shared_info variant then that can be 
done in future.

> > +    if ( ctxt.flags & DOMAIN_SAVE_32BIT_SHINFO )
> > +#ifdef CONFIG_COMPAT
> > +        has_32bit_shinfo(d) = true;
> > +#else
> > +        return -EINVAL;
> > +#endif
> 
> Am I mis-remembering or was a check lost of the remaining
> flags being zero? If I am, one needs adding in any case, imo.
> 

It wasn't flags before, but you're quite right; they should be zero-checked.

  Paul

> Jan