[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Xen-devel] [PATCH v15 17/21] x86/VPMU: Handle PMU interrupts for PV guests
- To: Jan Beulich <JBeulich@xxxxxxxx>
- From: Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx>
- Date: Wed, 03 Dec 2014 15:29:02 -0500
- Cc: kevin.tian@xxxxxxxxx, keir@xxxxxxx, suravee.suthikulpanit@xxxxxxx, andrew.cooper3@xxxxxxxxxx, tim@xxxxxxx, dietmar.hahn@xxxxxxxxxxxxxx, xen-devel@xxxxxxxxxxxxx, Aravind.Gopalakrishnan@xxxxxxx, jun.nakajima@xxxxxxxxx, dgdegra@xxxxxxxxxxxxx
- Delivery-date: Wed, 03 Dec 2014 20:27:52 +0000
- List-id: Xen developer discussion <xen-devel.lists.xen.org>
On 11/27/2014 03:59 AM, Jan Beulich wrote:
On 26.11.14 at 15:39, <boris.ostrovsky@xxxxxxxxxx> wrote:
On 11/25/2014 09:28 AM, Jan Beulich wrote:
+ else
+ {
+ struct segment_register seg;
+
+ hvm_get_segment_register(sampled, x86_seg_cs, &seg);
+ r->cs = seg.sel;
+ hvm_get_segment_register(sampled, x86_seg_ss, &seg);
+ r->ss = seg.sel;
+ if ( seg.attr.fields.dpl != 0 )
+ *flags |= PMU_SAMPLE_USER;
Is that how hardware treats it (CPL != 0 meaning user, rather
than CPL == 3)?
You mean how *software* (e.g. Linux kernel) treats it? If yes, then for
32-bit user_mode() checks for (CS == 3) and for 64-bit it's !!(CS & 3).
No, I meant hardware. There CPL qualified PMU aspects, and it was
those I had in mind to use as reference here.
Maybe you should surface CPL instead of a
boolean flag?
Yes, I think it may be better. Let the caller sort out how to interpret it.
-boris
Am I not already doing it by passing SS and CS to the guest?
No, neither SS.RPL nor CS.RPL formally represent CPL.
Jan
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel
|
