Re: [Xen-devel] RFC: PVH set vcpu info context in vmcs....

[Mukesh Rathor <mukesh.rathor@xxxxxxxxxx>]

On Wed, 14 Aug 2013 10:12:18 +0100
"Jan Beulich" <JBeulich@xxxxxxxx> wrote:

> >>> On 14.08.13 at 04:12, Mukesh Rathor <mukesh.rathor@xxxxxxxxxx>
> >>> wrote:
> > On Tue, 13 Aug 2013 11:56:36 +0100
> > "Jan Beulich" <JBeulich@xxxxxxxx> wrote:
> > 
...
> >> >     if ( v->vcpu_id == 0 )
> >> >         return 0;
> >> 
> >> Bogus/pointless.
> > 
> > No, we don't have anything for the boot vcpu. It's totally coming up
> > on the flat address space. For non boot, the vcpu is coming up on
> > the kernel GDT. Recall it's a PV guest (coming up in an HVM
> > container).
> 
> No, that's the wrong perspective. You either should never get here
> for vCPU 0, or you should refuse this for all already initialized
> vCPU-s.

Ok, i'll move the check to caller. FWIW, the vcpu->is_initialised is
inapplicable here because this is relevant for non-boot vcpus only.

>> 
>>     if ( ctxtp->user_regs.cs == 0 || (ctxtp->user_regs.cs & 3) == 3 )
>>         return -EINVAL;  

>Perhaps better check for any non-zero RPL, as the RPL needs
>to match of the (enforced) CS hidden fields?

Well, we now load the descriptor provided by the guest in the GDT
so that's kinda un needed now.. thats why i thought we don't need to
enforce RPL or check VGCF_in_kernel. But, that would work too...

thanks,
mukesh

Final(hopefully) version:

/*
 * Set vmcs fields in support of vcpu_op -> VCPUOP_initialise hcall used
 * to initialise a secondary vcpu prior to boot. (boot vcpu 0 context is
 * set by the tools via do_domctl -> vcpu_initialise).  Called from
 * arch_set_info_guest() which sets the (PVH relevant) non-vmcs fields.
 *
 * In case of linux:
 *     The call comes from cpu_initialize_context().  
 *
 * PVH 32bitfixme: this function needs to be modified for 32bit guest.
 */
int vmx_pvh_set_vcpu_info(struct vcpu *v, struct vcpu_guest_context *ctxtp)
{
    int rc;

    if ( !(ctxtp->flags & VGCF_in_kernel) )
        return -EINVAL;

    if ( ctxtp->ldt_base || ctxtp->ldt_ents ||
         ctxtp->user_regs.ss || ctxtp->user_regs.es || ctxtp->user_regs.ds ||
         ctxtp->user_regs.fs || ctxtp->user_regs.gs )
        return -EINVAL;

    if ( ctxtp->user_regs.cs == 0 || (ctxtp->user_regs.cs & 7) )
        return -EINVAL;

    vmx_vmcs_enter(v);
    __vmwrite(GUEST_GDTR_BASE, ctxtp->gdt.pvh.addr);
    __vmwrite(GUEST_GDTR_LIMIT, ctxtp->gdt.pvh.limit);

    /* IA-32e: ss/es/ds are ignored. */
    if ( (rc = hvm_load_segment_selector(v, x86_seg_cs, ctxtp->user_regs.cs)) )
        goto out;

    __vmwrite(GUEST_FS_BASE, ctxtp->fs_base);
    __vmwrite(GUEST_GS_BASE, ctxtp->gs_base_kernel);

    if ( (rc = vmx_add_guest_msr(MSR_SHADOW_GS_BASE)) )
        goto out;
    vmx_write_guest_msr(MSR_SHADOW_GS_BASE, ctxtp->gs_base_user);

 out:
    vmx_vmcs_exit(v);
    return rc;
}


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