Re: [PATCH 2/3] hw/net: e1000e: Correct the initial value of VET register

[Jason Wang <jasowang@redhat.com>]

在 2021/7/2 下午2:12, Bin Meng 写道:
> On Fri, Jul 2, 2021 at 1:47 PM Jason Wang <jasowang@redhat.com> wrote:
>>
>> 在 2021/7/2 下午12:43, Bin Meng 写道:
>>> On Fri, Jul 2, 2021 at 11:29 AM Jason Wang <jasowang@redhat.com> wrote:
>>>> 在 2021/7/1 下午5:46, Bin Meng 写道:
>>>>> From: Christina Wang <christina.wang@windriver.com>
>>>>>
>>>>> The initial value of VLAN Ether Type (VET) register is 0x8100, as per
>>>>> the manual and real hardware.
>>>>>
>>>>> While Linux e1000e driver always writes VET register to 0x8100, it is
>>>>> not always the case for everyone. Drivers relying on the reset value
>>>>> of VET won't be able to transmit and receive VLAN frames in QEMU.
>>>>>
>>>>> Unlike e1000 in QEMU, e1000e uses a field 'vet' in "struct E1000Core"
>>>>> to cache the value of VET register, but the cache only gets updated
>>>>> when VET register is written. To always get a consistent VET value
>>>>> no matter VET is written or remains its reset value, drop the 'vet'
>>>>> field and use 'core->mac[VET]' directly.
>>>>>
>>>>> Reported-by: Markus Carlstedt <markus.carlstedt@windriver.com>
>>>>> Signed-off-by: Christina Wang <christina.wang@windriver.com>
>>>>> Signed-off-by: Bin Meng <bin.meng@windriver.com>
>>>>> Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
>>>>> ---
>>>>>
>>>>>     hw/net/e1000e_core.h |  2 --
>>>>>     hw/net/e1000e.c      |  6 ++----
>>>>>     hw/net/e1000e_core.c | 11 ++++++-----
>>>>>     3 files changed, 8 insertions(+), 11 deletions(-)
>>>>>
>>>>> diff --git a/hw/net/e1000e_core.h b/hw/net/e1000e_core.h
>>>>> index 4ddb4d2c39..07d722bc68 100644
>>>>> --- a/hw/net/e1000e_core.h
>>>>> +++ b/hw/net/e1000e_core.h
>>>>> @@ -105,8 +105,6 @@ struct E1000Core {
>>>>>         uint32_t itr_guest_value;
>>>>>         uint32_t eitr_guest_value[E1000E_MSIX_VEC_NUM];
>>>>>
>>>>> -    uint16_t vet;
>>>>> -
>>>>>         uint8_t permanent_mac[ETH_ALEN];
>>>>>
>>>>>         NICState *owner_nic;
>>>>> diff --git a/hw/net/e1000e.c b/hw/net/e1000e.c
>>>>> index a8a77eca95..1797e4a7cb 100644
>>>>> --- a/hw/net/e1000e.c
>>>>> +++ b/hw/net/e1000e.c
>>>>> @@ -602,8 +602,8 @@ static const VMStateDescription e1000e_vmstate_intr_timer = {
>>>>>
>>>>>     static const VMStateDescription e1000e_vmstate = {
>>>>>         .name = "e1000e",
>>>>> -    .version_id = 1,
>>>>> -    .minimum_version_id = 1,
>>>>> +    .version_id = 2,
>>>>> +    .minimum_version_id = 2,
>>>>>         .pre_save = e1000e_pre_save,
>>>>>         .post_load = e1000e_post_load,
>>>>>         .fields = (VMStateField[]) {
>>>>> @@ -645,8 +645,6 @@ static const VMStateDescription e1000e_vmstate = {
>>>>>             VMSTATE_UINT32_ARRAY(core.eitr_guest_value, E1000EState,
>>>>>                                  E1000E_MSIX_VEC_NUM),
>>>>>
>>>>> -        VMSTATE_UINT16(core.vet, E1000EState),
>>>> This is not the suggested way. We'd better not bump version in this case.
>>>>
>>>> How about update vet during post_load?
>>> But core.vet is removed in this patch. Not sure how to handle this?
>>
>> Keep using core.vet, sync core.vet with mac[VET] during post_load.
>>
> But keeping using core.vet in the e1000e_core.c will cause mismatch
> with mac[VET] as the commit message says.
>
> We can still keep the 'vet' field in the "struct E1000Core", and keep
> the "VMSTATE_UINT16(core.vet, E1000EState)" here, but it is useless in
> the new code.


The point is to unbreak migration to old version.

Thanks


>
> Regards,
> Bin
>