Re: [Xen-devel] [PATCH v5 3/8] xen: add basic hypervisor filesystem support

[Julien Grall <julien@xxxxxxx>]

Hi,

On 19/02/2020 15:49, Jan Beulich wrote:
> On 19.02.2020 09:11, Juergen Gross wrote:
> > +int hypfs_write_bool(struct hypfs_entry_leaf *leaf,
> > +                     XEN_GUEST_HANDLE_PARAM(void) uaddr, unsigned long ulen)
> > +{
> > +    union {
> > +        char buf[8];
> > +        uint8_t u8;
> > +        uint16_t u16;
> > +        uint32_t u32;
> > +        uint64_t u64;
> > +    } u;
> > +
> > +    ASSERT(leaf->e.type == XEN_HYPFS_TYPE_UINT && leaf->e.size <= 8);
> > +
> > +    if ( ulen != leaf->e.size )
> > +        return -EDOM;
> > +
> > +    if ( copy_from_guest(u.buf, uaddr, ulen) )
> > +        return -EFAULT;
> > +
> > +    switch ( leaf->e.size )
> > +    {
> > +    case 1:
> > +        *(uint8_t *)leaf->write_ptr = !!u.u8;
> > +        break;
> > +    case 2:
> > +        *(uint16_t *)leaf->write_ptr = !!u.u16;
> > +        break;
> > +    case 4:
> > +        *(uint32_t *)leaf->write_ptr = !!u.u32;
> > +        break;
> > +    case 8:
> > +        *(uint64_t *)leaf->write_ptr = !!u.u64;
> > +        break;
> > +    }
> Looking at this again, is there really a need for uint64_t support
> here? I.e. can't you cap at unsigned long (or even unsigned int),
> and perhaps additionally avoid use of fixed width types here
> altogether (some trickery may be needed for 32-bit's
> sizeof(long) == sizeof(int))?
On Arm, we want to have the interface exactly the same for 64-bit and 
32-bit. If you plan to not support 8-byte here, this would effectively 
mean that 8-byte could not be used on Arm.
However, I can't find any use of hypfs_write_bool, so I am not entirely 
sure why we would want to cater size other than 1. Juerge, do you mind 
expanding your thoughts?
Cheers,

--
Julien Grall

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel