Am 09.10.2013 um 01:31 schrieb Paul Mackerras <paulus@xxxxxxxxx>:
> On Tue, Oct 08, 2013 at 04:25:31PM +0200, Alexander Graf wrote:
>>
>> On 08.10.2013, at 16:12, Cédric Le Goater <clg@xxxxxxxxxx> wrote:
>>
>>> MMIO emulation reads the last instruction executed by the guest
>>> and then emulates. If the guest is running in Little Endian mode,
>>> the instruction needs to be byte-swapped before being emulated.
>>>
>>> This patch stores the last instruction in the endian order of the
>>> host, primarily doing a byte-swap if needed. The common code
>>> which fetches 'last_inst' uses a helper routine kvmppc_need_byteswap().
>>> and the exit paths for the Book3S PV and HR guests use their own
>>> version in assembly.
>>>
>>> Finally, the meaning of the 'is_bigendian' argument of the
>>> routines kvmppc_handle_load() of kvmppc_handle_store() is
>>> slightly changed to represent an eventual reverse operation. This
>>> is used in conjunction with kvmppc_is_bigendian() to determine if
>>> the instruction being emulated should be byte-swapped.
>>>
>>> Signed-off-by: Cédric Le Goater <clg@xxxxxxxxxx>
>>> ---
>>>
>>> Changes in v2:
>>>
>>> - replaced rldicl. by andi. to test the MSR_LE bit in the guest
>>> exit paths. (Paul Mackerras)
>>>
>>> - moved the byte swapping logic to kvmppc_handle_load() and
>>> kvmppc_handle_load() by changing the is_bigendian parameter
>>> meaning. (Paul Mackerras)
>>>
>>> arch/powerpc/include/asm/kvm_book3s.h | 9 ++++++++-
>>> arch/powerpc/include/asm/kvm_ppc.h | 10 +++++-----
>>> arch/powerpc/kvm/book3s_hv_rmhandlers.S | 11 +++++++++++
>>> arch/powerpc/kvm/book3s_segment.S | 10 ++++++++++
>>> arch/powerpc/kvm/emulate.c | 1 -
>>> arch/powerpc/kvm/powerpc.c | 16 ++++++++++++----
>>> 6 files changed, 46 insertions(+), 11 deletions(-)
>>>
>>> diff --git a/arch/powerpc/include/asm/kvm_book3s.h b/arch/powerpc/include/asm/kvm_book3s.h
>>> index 4ee6c66..f043e62 100644
>>> --- a/arch/powerpc/include/asm/kvm_book3s.h
>>> +++ b/arch/powerpc/include/asm/kvm_book3s.h
>>> @@ -289,7 +289,14 @@ static inline int kvmppc_ld32(struct kvm_vcpu *vcpu, ulong *eaddr,
>>> static inline int kvmppc_ld_inst(struct kvm_vcpu *vcpu, ulong *eaddr,
>>> u32 *inst)
>>> {
>>> - return kvmppc_ld32(vcpu, eaddr, inst, false);
>>> + int ret;
>>> +
>>> + ret = kvmppc_ld32(vcpu, eaddr, inst, false);
>>> +
>>> + if (kvmppc_need_byteswap(vcpu))
>>> + *inst = swab32(*inst);
>>
>> This logic wants to live in kvmppc_ld32(), no? Every 32bit access is going to need a byteswap, regardless of whether it's an instruction or not.
>
> True, until we get to POWER8 with its split little-endian support,
> where instructions and data can have different endianness...
How exactly does that work?
Alex
>
> Paul.
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
