Re: [Xen-devel] [PATCH v2 17/19] x86/hvm: Avoid __hvm_copy() raising #PF behind the emulators back

[Tim Deegan <tim@xxxxxxx>]

At 16:32 +0000 on 28 Nov (1480350762), Andrew Cooper wrote:
> On 28/11/16 14:56, Tim Deegan wrote:
> > At 11:13 +0000 on 28 Nov (1480331614), Andrew Cooper wrote:
> >> Drop the call to hvm_inject_page_fault() in __hvm_copy(), and require 
> >> callers
> >> to inject the pagefault themselves.
> > This seems like it'd be easy to forget to DTRT with the fault,
> > especially in code being ported forward across this series.
> 
> Code ported across the series will have an API change to accommodate.
> 
> >
> > Would it be better to have hvm_copy &c take a callback function
> > instead of a pfinfo pointer?
> 
> I considered both of these options, but this option seemed cleaner at
> the time.
> 
> I am not fussed either way, but I don't see that a new function pointer
> would be any less easy to get wrong.

Righto.  As I said, Ack to the shadow bits anyway.

Tim.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel