[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] Ping: [PATCH 0/3] x86emul: misc adjustments



>>> On 20.12.16 at 12:30, <andrew.cooper3@xxxxxxxxxx> wrote:
> On 20/12/2016 09:04, Jan Beulich wrote:
>>>>> On 12.12.16 at 10:38, <JBeulich@xxxxxxxx> wrote:
>>> These patches are grouped together merely because of contextual
>>> dependencies.
>>>
>>> 1: correct EFLAGS.TF handling
>>> 2: conditionally clear BNDn for branches
>>> 3: some REX related polishing
>>>
>>> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
>> Any word on these? I realize patch 1 will need to be re-based over
>> yesterday's XSA-204 patch, and as mentioned elsewhere may also
>> want to suppress setting of retire.sti when TF=1 and BTF=0, but
>> before sending a v2 of that I'd prefer knowing whether further
>> adjustments may be needed.
> 
> I'd like to get a better understanding of the interaction of shadows
> before passing review on patch 1.  I don't have any specific issues with
> it though.
> 
> The repeated reads of MSR_DEBUGCTL are a little concerning, but your
> proposed alternate for vendor_is() has given me an idea how to fix this
> once the MSR levelling work starts, so I am less concerned.

After re-basing there are no such repeated reads anymore afaict:
The MSR will be read once when TF is set (at the time where the
singlestep local variable get initialized), and then only again in
put_loop_count() (just like in v1). I'll send out v2 later.

Jan


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

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.