Re: [PATCH v1 0/7] Implement support for external IPT monitoring

[Roger Pau Monné <roger.pau@xxxxxxxxxx>]

On Wed, Jun 17, 2020 at 10:20:20PM +0200, Michał Leszczyński wrote:
> ----- 17 cze 2020 o 18:19, Andrew Cooper andrew.cooper3@xxxxxxxxxx napisał(a):
> 
> > On 17/06/2020 04:02, Tamas K Lengyel wrote:
> >> On Tue, Jun 16, 2020 at 2:17 PM Andrew Cooper <andrew.cooper3@xxxxxxxxxx> 
> >> wrote:
> >>> On 16/06/2020 19:47, Michał Leszczyński wrote:
> >>>> ----- 16 cze 2020 o 20:17, Andrew Cooper andrew.cooper3@xxxxxxxxxx 
> >>>> napisał(a):
> >>>>
> >>>>> Are there any restrictions on EPT being enabled in the first place?  I'm
> >>>>> not aware of any, and in principle we could use this functionality for
> >>>>> PV guests as well (using the CPL filter).  Therefore, I think it would
> >>>>> be helpful to not tie the functionality to HVM guests, even if that is
> >>>>> the only option enabled to start with.
> >>>> I think at the moment it's not required to have EPT. This patch series 
> >>>> doesn't
> >>>> use any translation feature flags, so the output address is always a 
> >>>> machine
> >>>> physical address, regardless of context. I will check if it could be 
> >>>> easily
> >>>> used with PV.
> >>> If its trivial to add PV support then please do.  If its not, then don't
> >>> feel obliged, but please do at least consider how PV support might look
> >>> in the eventual feature.
> >>>
> >>> (Generally speaking, considering "how would I make this work in other
> >>> modes where it is possible" leads to a better design.)
> >>>
> >>>>> The buffer mapping and creation logic is fairly problematic.  Instead of
> >>>>> fighting with another opencoded example, take a look at the IOREQ
> >>>>> server's use of "acquire resource" which is a mapping interface which
> >>>>> supports allocating memory on behalf of the guest, outside of the guest
> >>>>> memory, for use by control tools.
> >>>>>
> 
> 
> One thing that remains unclear to me is the "acquire resource" part. Could 
> you give some more details on that?
> 
> Assuming that buffers are allocated right from the domain creation, what 
> mechanism (instead of xc_map_foreign_range) should I use to map the IPT 
> buffers into Dom0?

Take a look at demu's demu_initialize function [0] (and it's usage of
xenforeignmemory_map_resource), you likely need something similar for
the trace buffers, introducing a new XENMEM_resource_trace_data kind
of resource (naming subject to change), and use the id field in
xen_mem_acquire_resource to signal which vCPU buffer you want to
map.

That's usable by both PV and HVM guests.

Roger.

[0] 
http://xenbits.xen.org/gitweb/?p=people/pauldu/demu.git;a=blob;f=demu.c;h=f785b394d0cf141dffa05bdddecf338214358aea;hb=refs/heads/master#l453