Re: [Qemu-devel] [PATCH 1/6] qtest: add test framework

[Anthony Liguori <aliguori@us.ibm.com>]

On 01/16/2012 11:08 AM, Avi Kivity wrote:
> On 01/16/2012 06:59 PM, Stefan Hajnoczi wrote:
>>> +        }
>>> +        qtest_send_prefix(chr);
>>> +        qtest_send(chr, "OK 0x%04x\n", value);
>>
>> Endianness is a little weird here.  memory.c will byteswap if target
>> and device endianness differ.
>>
>> Imagine the case where we're on an x86 host, running a ppc guest,
>> reading from PCI configuration space (little-endian).  Since ppc
>> (target endian) is big-endian and the device is little-endian the
>> value read/written will be byteswapped.  However, our qtest runs on
>> the host and therefore we don't want that automatic swap (or we need
>> to neutralize it by performing another byteswap on top).
>>
>
> Good catch.  This is another example of how an access depends not only
> on the destination, but also on the source.  Here the source is not the
> cpu; it's qtest.

target-ppc never calls cpu_out.  It's prep_pci.c that calls it based on an MMIO 
operation and that's where the endianess conversion happens.

You wouldn't use outw() in a qtest harness when talking to a PCI device over PPC.

What we really want is to be able to send I/O directly to MemoryRegions but 
that's a bit different than what qtest is.

Regards,

Anthony Liguori

>