xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Qing He <qing.he@intel.com>
To: Tim Deegan <Tim.Deegan@citrix.com>
Cc: "xen-devel@lists.xensource.com" <xen-devel@lists.xensource.com>
Subject: Re: [PATCH 11/17] vmx: nest: L2 tsc
Date: Fri, 21 May 2010 00:07:18 +0800	[thread overview]
Message-ID: <20100520160718.GE21538@qhe2-db> (raw)
In-Reply-To: <20100520114716.GR4164@whitby.uk.xensource.com>

On Thu, 2010-05-20 at 19:47 +0800, Tim Deegan wrote:
> At 10:41 +0100 on 22 Apr (1271932883), Qing He wrote:
> > L2 TSC needs special handling, either rdtsc exiting is
> > turned on or off
> 
> Looks OK to me, but I'm a bit behind on some of the recent changes to
> TSC handling. 

I'm also behind TSC changes, I'll try to understand more about those
changes, maybe after version 2.

Thanks,
Qing
> 
> Tim.
> 
> > Signed-off-by: Qing He <qing.he@intel.com>
> > 
> > ---
> >  arch/x86/hvm/vmx/nest.c        |   31 +++++++++++++++++++++++++++++++
> >  arch/x86/hvm/vmx/vmx.c         |    4 ++++
> >  include/asm-x86/hvm/vmx/nest.h |    2 ++
> >  3 files changed, 37 insertions(+)
> > 
> > diff -r 2f9ba6dbbe62 -r 2332586ff957 xen/arch/x86/hvm/vmx/nest.c
> > --- a/xen/arch/x86/hvm/vmx/nest.c	Thu Apr 22 22:30:09 2010 +0800
> > +++ b/xen/arch/x86/hvm/vmx/nest.c	Thu Apr 22 22:30:09 2010 +0800
> > @@ -533,6 +533,18 @@
> >   * Nested VMX context switch
> >   */
> >  
> > +u64 vmx_nest_get_tsc_offset(struct vcpu *v)
> > +{
> > +    u64 offset = 0;
> > +    struct vmx_nest_struct *nest = &v->arch.hvm_vmx.nest;
> > +
> > +    if ( __get_vvmcs(nest->vvmcs, CPU_BASED_VM_EXEC_CONTROL) &
> > +         CPU_BASED_USE_TSC_OFFSETING )
> > +        offset = __get_vvmcs(nest->vvmcs, TSC_OFFSET);
> > +
> > +    return offset;
> > +}
> > +
> >  static unsigned long vmcs_gstate_field[] = {
> >      /* 16 BITS */
> >      GUEST_ES_SELECTOR,
> > @@ -715,6 +727,8 @@
> >      hvm_set_cr4(__get_vvmcs(nest->vvmcs, GUEST_CR4));
> >      hvm_set_cr3(__get_vvmcs(nest->vvmcs, GUEST_CR3));
> >  
> > +    hvm_funcs.set_tsc_offset(v, v->arch.hvm_vcpu.cache_tsc_offset);
> > +
> >      vvmcs_to_shadow(nest->vvmcs, VM_ENTRY_INTR_INFO);
> >      vvmcs_to_shadow(nest->vvmcs, VM_ENTRY_EXCEPTION_ERROR_CODE);
> >      vvmcs_to_shadow(nest->vvmcs, VM_ENTRY_INSTRUCTION_LEN);
> > @@ -837,6 +851,8 @@
> >      hvm_set_cr4(__get_vvmcs(nest->vvmcs, HOST_CR4));
> >      hvm_set_cr3(__get_vvmcs(nest->vvmcs, HOST_CR3));
> >  
> > +    hvm_funcs.set_tsc_offset(v, v->arch.hvm_vcpu.cache_tsc_offset);
> > +
> >      __set_vvmcs(nest->vvmcs, VM_ENTRY_INTR_INFO, 0);
> >  }
> >  
> > @@ -1116,6 +1132,21 @@
> >  
> >          if ( control_bit_for_reason[i].bit & ctrl )
> >              nest->vmexit_pending = 1;
> > +        else if ( exit_reason == EXIT_REASON_RDTSC )
> > +        {
> > +            uint64_t tsc;
> > +
> > +            /*
> > +             * rdtsc can't be handled normally in the L0 handler
> > +             * if L1 doesn't want it
> > +             */
> > +            tsc = hvm_get_guest_tsc(v);
> > +            tsc += __get_vvmcs(nest->vvmcs, TSC_OFFSET);
> > +            regs->eax = (uint32_t)tsc;
> > +            regs->edx = (uint32_t)(tsc >> 32);
> > +
> > +            bypass_l0 = 1;
> > +        }
> >  
> >          break;
> >      }
> > diff -r 2f9ba6dbbe62 -r 2332586ff957 xen/arch/x86/hvm/vmx/vmx.c
> > --- a/xen/arch/x86/hvm/vmx/vmx.c	Thu Apr 22 22:30:09 2010 +0800
> > +++ b/xen/arch/x86/hvm/vmx/vmx.c	Thu Apr 22 22:30:09 2010 +0800
> > @@ -974,6 +974,10 @@
> >  static void vmx_set_tsc_offset(struct vcpu *v, u64 offset)
> >  {
> >      vmx_vmcs_enter(v);
> > +
> > +    if ( v->arch.hvm_vcpu.in_nesting )
> > +        offset += vmx_nest_get_tsc_offset(v);
> > +
> >      __vmwrite(TSC_OFFSET, offset);
> >  #if defined (__i386__)
> >      __vmwrite(TSC_OFFSET_HIGH, offset >> 32);
> > diff -r 2f9ba6dbbe62 -r 2332586ff957 xen/include/asm-x86/hvm/vmx/nest.h
> > --- a/xen/include/asm-x86/hvm/vmx/nest.h	Thu Apr 22 22:30:09 2010 +0800
> > +++ b/xen/include/asm-x86/hvm/vmx/nest.h	Thu Apr 22 22:30:09 2010 +0800
> > @@ -69,6 +69,8 @@
> >                                              unsigned long value);
> >  void vmx_nest_update_exception_bitmap(struct vcpu *v, unsigned long value);
> >  
> > +u64 vmx_nest_get_tsc_offset(struct vcpu *v);
> > +
> >  void vmx_nest_idtv_handling(void);
> >  
> >  int vmx_nest_l2_vmexit_handler(struct cpu_user_regs *regs,
> > 
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.xensource.com
> > http://lists.xensource.com/xen-devel
> 
> -- 
> Tim Deegan <Tim.Deegan@citrix.com>
> Principal Software Engineer, XenServer Engineering
> Citrix Systems UK Ltd.  (Company #02937203, SL9 0BG)

  reply	other threads:[~2010-05-20 16:07 UTC|newest]

Thread overview: 51+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-04-22  9:41 [PATCH 00/17][RFC] Nested virtualization for VMX Qing He
2010-04-22  9:41 ` [PATCH 01/17] vmx: nest: fix CR4.VME in update_guest_cr Qing He
2010-05-20  9:26   ` Tim Deegan
2010-05-20  9:36     ` Qing He
2010-04-22  9:41 ` [PATCH 02/17] vmx: nest: rename host_vmcs Qing He
2010-04-22  9:41 ` [PATCH 03/17] vmx: nest: wrapper for control update Qing He
2010-05-20  9:34   ` Tim Deegan
2010-05-20  9:46     ` Qing He
2010-05-20 12:57       ` Keir Fraser
2010-04-22  9:41 ` [PATCH 04/17] vmx: nest: domain and vcpu flags Qing He
2010-05-20  9:37   ` Tim Deegan
2010-05-20  9:51     ` Christoph Egger
2010-05-20  9:54     ` Qing He
2010-05-20 10:55       ` Tim Deegan
2010-05-20 12:53         ` Qing He
2010-05-20 14:06           ` Christoph Egger
2010-04-22  9:41 ` [PATCH 05/17] vmx: nest: nested control structure Qing He
2010-04-22  9:41 ` [PATCH 06/17] vmx: nest: virtual vmcs layout Qing He
2010-04-22  9:41 ` [PATCH 07/17] vmx: nest: handling VMX instruction exits Qing He
2010-05-20 10:53   ` Tim Deegan
2010-05-20 13:28     ` Qing He
2010-04-22  9:41 ` [PATCH 08/17] vmx: nest: L1 <-> L2 context switch Qing He
2010-05-20 11:11   ` Tim Deegan
2010-05-20 13:49     ` Qing He
2010-05-21  9:19       ` Tim Deegan
2010-05-21 10:31         ` Qing He
2010-05-25 15:27           ` Tim Deegan
2010-04-22  9:41 ` [PATCH 09/17] vmx: nest: interrupt Qing He
2010-05-20 11:21   ` Tim Deegan
2010-05-20 15:55     ` Qing He
2010-04-22  9:41 ` [PATCH 10/17] vmx: nest: VMExit handler in L2 Qing He
2010-05-20 11:44   ` Tim Deegan
2010-05-20 16:06     ` Qing He
2010-05-21  8:42       ` Tim Deegan
2010-05-21 10:35         ` Qing He
2010-05-25 15:34           ` Tim Deegan
2010-04-22  9:41 ` [PATCH 11/17] vmx: nest: L2 tsc Qing He
2010-05-20 11:47   ` Tim Deegan
2010-05-20 16:07     ` Qing He [this message]
2010-04-22  9:41 ` [PATCH 12/17] vmx: nest: CR0.TS and #NM Qing He
2010-04-22  9:41 ` [PATCH 13/17] vmx: nest: capability reporting MSRs Qing He
2010-05-20 11:52   ` Tim Deegan
2010-04-22  9:41 ` [PATCH 14/17] vmx: nest: enable virtual VMX Qing He
2010-04-22  9:41 ` [PATCH 15/17] vmx: nest: virtual ept for nested Qing He
2010-05-20 12:21   ` Tim Deegan
2010-05-21 10:24     ` Qing He
2010-05-25 16:02       ` Tim Deegan
2010-04-22  9:41 ` [PATCH 16/17] vmx: nest: hvmtrace " Qing He
2010-04-22  9:41 ` [PATCH 17/17] tools: nest: allow enabling nesting Qing He
2010-04-22 10:15 ` [PATCH 00/17][RFC] Nested virtualization for VMX Christoph Egger
2010-04-23 10:10   ` He, Qing

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20100520160718.GE21538@qhe2-db \
    --to=qing.he@intel.com \
    --cc=Tim.Deegan@citrix.com \
    --cc=xen-devel@lists.xensource.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).