[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH v6 1/3] memory: drop guest writes to read-only ram device reg
From: |
Paolo Bonzini |
Subject: |
Re: [PATCH v6 1/3] memory: drop guest writes to read-only ram device regions |
Date: |
Thu, 28 May 2020 07:10:46 +0200 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.6.0 |
On 28/05/20 06:35, Yan Zhao wrote:
> On Tue, May 26, 2020 at 10:26:35AM +0100, Peter Maydell wrote:
>> On Mon, 25 May 2020 at 11:20, Paolo Bonzini <pbonzini@redhat.com> wrote:
>>> Not all of them, only those that need to return MEMTX_ERROR. I would
>>> like some guidance from Peter as to whether (or when) reads from ROMs
>>> should return MEMTX_ERROR. This way, we can use that information to
>>> device what the read-only ram-device regions should do.
>>
>> In general I think writes to ROMs (and indeed reads from ROMs) should
>> not return MEMTX_ERROR. I think that in real hardware you could have
>> a ROM that behaved either way; so our default behaviour should probably
>> be to do what we've always done and not report a MEMTX_ERROR. (If we
>> needed to I suppose we should implement a MEMTX_ERROR-reporting ROM,
>> but to be honest there aren't really many real ROMs in systems these
>> days: it's more often flash, whose response to writes is defined
>> by the spec and is I think to ignore writes which aren't the
>> magic "shift to program-the-flash-mode" sequence.)
>>
> then should I just drop the writes to read-only ram-device regions and
> vfio regions without returning MEMTX_ERROR?
> do you think it's good?
I am not really sure, I have to think more about it. I think read-only
RAMD regions are slightly different because the guest can expect "magic"
behavior from RAMD regions (e.g. registers that trigger I/O on writes)
that are simply not there for ROM. So I'm still inclined to queue your
v6 patch series.
Thanks,
Paolo