Re: [PATCH RFC V5 4/5] xen, libxc: Request page fault injection via libxc

[Razvan Cojocaru <rcojocaru@bitdefender.com>]

On 08/08/2014 05:48 PM, Jan Beulich wrote:
>> +    if ( seg.attr.fields.dpl != 3 ) /* Guest is not in user mode */
>> +        return 0;
>> +
>> +
> 
> Double blank line.Didn't I ask you to clean up your patches in this
> regard already?

Not me, but point taken. :)

>> +    if ( curr->arch.hvm_vcpu.guest_cr[3]
>> +         != d->arch.hvm_domain.fault_info.address_space )
> 
> Do you really mean to compare the full CR3 value here, rather than
> just bits 12...51? In which case the address_space field likely would
> better be a GPFN.

You're right, I'll compare the values shifted by PAGE_SHIFT.

>> +}
>> +
>> +static void vmx_inject_pf(void)
>> +{
>> +    struct vcpu *curr = current;
>> +    struct domain *d = curr->domain;
>> +    int errcode = PFEC_user_mode;
>> +    uint64_t virtual_address = d->arch.hvm_domain.fault_info.virtual_address;
>> +    uint32_t write_access = d->arch.hvm_domain.fault_info.write_access;
>> +
>> +    d->arch.hvm_domain.fault_info.address_space = 0;
>> +    d->arch.hvm_domain.fault_info.virtual_address = 0;
>> +    d->arch.hvm_domain.fault_info.write_access = 0;
> 
> Are these necessary? Because ...
> 
>> +    d->arch.hvm_domain.fault_info.valid = 0;
> 
> ... I would hope that this one is properly guarding all uses of the
> other fields.

No, they're not necessary anymore. Thank you for pointing that out.


Thanks,
Razvan Cojocaru