Re: [PATCH v2 for 4.5] ioreq-server: handle the lack of a default emulator properly

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 29/09/14 13:20, Jan Beulich wrote:
>>>> On 29.09.14 at 12:59, <andrew.cooper3@citrix.com> wrote:
>> On 29/09/14 11:21, Paul Durrant wrote:
>>> I started porting QEMU over to use the new ioreq server API and hit a
>>> problem with PCI bus enumeration. Because, with my patches, QEMU only
>>> registers to handle config space accesses for the PCI device it implements
>>> all other attempts by the guest to access 0xcfc go nowhere and this was
>>> causing the vcpu to wedge up because nothing was completing the I/O.
>>>
>>> This patch introduces an I/O completion handler into the hypervisor for the
>>> case where no ioreq server matches a particular request. Read requests are
>>> completed with 0xf's in the data buffer, writes and all other I/O req types
>>> are ignored.
>>>
>>> Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
>>> Cc: Keir Fraser <keir@xen.org>
>>> Cc: Jan Beulich <jbeulich@suse.com>
>>> ---
>>> v2: - First non-RFC submission
>>>     - Removed warning on unemulated MMIO accesses
>>>
>>>  xen/arch/x86/hvm/hvm.c |   35 ++++++++++++++++++++++++++++++++---
>>>  1 file changed, 32 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c
>>> index 5c7e0a4..822ac37 100644
>>> --- a/xen/arch/x86/hvm/hvm.c
>>> +++ b/xen/arch/x86/hvm/hvm.c
>>> @@ -2386,8 +2386,7 @@ static struct hvm_ioreq_server 
>> *hvm_select_ioreq_server(struct domain *d,
>>>      if ( list_empty(&d->arch.hvm_domain.ioreq_server.list) )
>>>          return NULL;
>>>  
>>> -    if ( list_is_singular(&d->arch.hvm_domain.ioreq_server.list) ||
>>> -         (p->type != IOREQ_TYPE_COPY && p->type != IOREQ_TYPE_PIO) )
>>> +    if ( p->type != IOREQ_TYPE_COPY && p->type != IOREQ_TYPE_PIO )
>>>          return d->arch.hvm_domain.default_ioreq_server;
>>>  
>>>      cf8 = d->arch.hvm_domain.pci_cf8;
>>> @@ -2618,12 +2617,42 @@ bool_t hvm_send_assist_req_to_ioreq_server(struct 
>> hvm_ioreq_server *s,
>>>      return 0;
>>>  }
>>>  
>>> +static bool_t hvm_complete_assist_req(ioreq_t *p)
>>> +{
>>> +    switch (p->type)
>>> +    {
>>> +    case IOREQ_TYPE_COPY:
>>> +    case IOREQ_TYPE_PIO:
>>> +        if ( p->dir == IOREQ_READ )
>>> +        {
>>> +            if ( !p->data_is_ptr )
>>> +                p->data = ~0ul;
>>> +            else
>>> +            {
>>> +                int i, sign = p->df ? -1 : 1;
>>> +                uint32_t data = ~0;
>>> +
>>> +                for ( i = 0; i < p->count; i++ )
>>> +                    hvm_copy_to_guest_phys(p->data + sign * i * p->size, &data,
>>> +                                           p->size);
>> This is surely bogus for an `ins` which crosses a page boundary?
> Crossing page boundaries gets dealt with up the call stack in
> hvmemul_linear_to_phys(), namely the path exiting with
> X86EMUL_UNHANDLEABLE when done == 0.
>
> Jan
>

Paul also pointed this out in person, which indicates that
hvm_copy_to_guest_phys() is indeed correct in this case.

Therefore it is fine, but only because the caller guarentees that
"p->data + sign * i * p->size" does not cross a page boundary.


However, what I cant spot is any logic which copes with addr not being
aligned with bytes_per_rep.  This appears to be valid in x86, and would
constitute an individual repetition accessing two pages.

~Andrew