答复: [PATCH 1/1] Skip flatview_simplify() for cpu vendor zhaoxin

[FelixCui-oc]

hi Paolo,

>Felix, did you identify the regions whose simplification causes the bug?
>Is this RAM (for example due to the PAM registers) or something else?

yes, this bug is caused by write PAM register.The actual situation is that the

properties of some ranges are changed from RW to readonly.This situation 

will cause the old ranges to be unmapped.

Best regards

Felixcui-oc

---

发件人: Paolo Bonzini <pbonzini@redhat.com>
发送时间: 2020年10月21日 15:37:53
收件人: Alex Williamson
抄送: FelixCui-oc; Richard Henderson; Eduardo Habkost; qemu-devel@nongnu.org; RockCui-oc; Tony W Wang-oc; CobeChen-oc
主题: Re: [PATCH 1/1] Skip flatview_simplify() for cpu vendor zhaoxin

 

On 21/10/20 00:44, Alex Williamson wrote:
> Do we necessarily need a memory map ioctl for this or could it be the
> QEMU code that compares the old and new maps to trigger map and unmap
> ioctls?  For example (aiui) our race is that if we have contiguous
> memory regions A and B and flatview_simplify() tries to expand A and
> delete B we'll see a series of listener notifications deleting A and B
> and adding A'.  But the vfio QEMU code could parse the memory map to
> determine that old A + B is functionally equivalent to A' and do
> nothing.

I think the issue is a bit different, and in fact there are two sides of
the same issue.  Say you have A (large) and it is replaced by A'
(smaller) + B, then:

* the first part of A disappears for a moment before A' appears.  This
is something that QEMU can work around, by not doing anything

* the second part of A disappears for a moment before B appears.  This
is the root API issue and not something that QEMU can work around; and
in fact it is not even fixed by removing flatview_simplify.

Felix, did you identify the regions whose simplification causes the bug?
 Is this RAM (for example due to the PAM registers) or something else?

Paolo