xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Dario Faggioli <dario.faggioli@citrix.com>
To: "Wu, Feng" <feng.wu@intel.com>
Cc: "Tian, Kevin" <kevin.tian@intel.com>,
	"keir@xen.org" <keir@xen.org>,
	George Dunlap <george.dunlap@eu.citrix.com>,
	"andrew.cooper3@citrix.com" <andrew.cooper3@citrix.com>,
	xen-devel <xen-devel@lists.xen.org>,
	Jan Beulich <JBeulich@suse.com>,
	"Zhang, Yang Z" <yang.z.zhang@intel.com>
Subject: Re: Fwd: [v3 14/15] Update Posted-Interrupts Descriptor during vCPU scheduling
Date: Tue, 14 Jul 2015 18:02:37 +0200	[thread overview]
Message-ID: <1436889757.13522.185.camel@citrix.com> (raw)
In-Reply-To: <E959C4978C3B6342920538CF579893F00260F191@SHSMSX104.ccr.corp.intel.com>


[-- Attachment #1.1: Type: text/plain, Size: 8880 bytes --]

On Tue, 2015-07-14 at 14:08 +0000, Wu, Feng wrote:
> 
> > -----Original Message-----
> > From: Dario Faggioli [mailto:dario.faggioli@citrix.com]

> >  - do you need to perform an action upon context switch (on prev and/or
> >    next vcpu)? If yes, there's an arch specific path in there already;
> >  - do you need to perform an action when a vcpu wakes-up? If yes, we
> >    need an arch hook in vcpu_wake();
> >  - do you need to perform an action when a vcpu goes to sleep? If yes,
> >    we need an arch hook in vcpu_sleep_nosync();
> > 
> > I think this makes a more than fair solution. I happen to like it even
> > better than the centralized approach, actually! That is for personal
> > taste, but also because I think it may be useful for others too, in
> > future, to be able to execute arch specific code, e.g., upon wakes-up,
> > in which case we will be able to use the hook that we're introducing
> > here for PI.
> > 
> > Thanks and Regards,
> > Dario
> 
> Hi Dario,
> 
Hi,

> Thanks for the suggestion! I made a draft patch for this idea, 
>
Great!

> It may have
> some issues since It is just a draft version, kind of like prototype, I post
> it here just like to know whether it is meet your expectation, if it is I
> can continue with this direction and this may speed up the upstreaming
> process.
>
Yes, I think this is a good approach, and the proper way for this
feature to interact with the scheduler.

I appreciate it is a draft, so I'm not performing a thorough review, but
I'll try to at least give some comments, in the hope that it helps.

> diff --git a/xen/arch/x86/hvm/vmx/vmx.c b/xen/arch/x86/hvm/vmx/vmx.c
> index 6eebc1a..7e678c8 100644
> --- a/xen/arch/x86/hvm/vmx/vmx.c
> +++ b/xen/arch/x86/hvm/vmx/vmx.c
> @@ -740,6 +740,81 @@ static void vmx_ctxt_switch_from(struct vcpu *v)
>      vmx_save_guest_msrs(v);
>      vmx_restore_host_msrs();
>      vmx_save_dr(v);
> +
> +    if ( iommu_intpost )
> +    {
>
I'd put an helper together ( vmx_<something>_pi() ) and put the body of
this if in it.

Then, either just call it unconditionally from here and have, in the
helper, something like this:

 if ( !iommu_intpost )
   return;

Or just have this in here:

 if ( iommu_intpost )
  vmx_<something>_pi();

> +        struct pi_desc *pi_desc = &v->arch.hvm_vmx.pi_desc;
> +        struct pi_desc old, new;
> +        unsigned long flags;
> +
> +        if ( vcpu_runnable(v) || !test_bit(_VPF_blocked, &v->pause_flags) )
> +        {
>
Aha! So, AFAICT, this means we can deal with preemptions, sleeps and
blockings (as can be seen below) here in _ctxt_switch_from, i.e., we
don't have to call in this code from vcpu_sleep_nosync(), like we were,
when tying this to vcpu_runstate_change())... nice! :-D

> +            /*
> +             * The vCPU is preempted or sleeped. 
>
"has been preempted or went to sleep" ?

> We don't need to send
> +             * notification event to a non-running vcpu, the interrupt
> +             * information will be delivered to it before VM-ENTRY when
> +             * the vcpu is scheduled to run next time.
> +             */
> +            pi_set_sn(pi_desc);
> +
> +        }
> +        else if ( test_bit(_VPF_blocked, &v->pause_flags) )
> +        {
> +            /* The vCPU is blocked */
>
This comment does not add much, I'd kill it.

> +            ASSERT(v->arch.hvm_vmx.pi_block_cpu == -1);
> +
> +            /*
> +             * The vCPU is blocked on the block list. 
>
What about "The vCPU is blocking, we need to add it to one of the per
pCPU lists."

> Add the blocked
> +             * vCPU on the list of the v->arch.hvm_vmx.pi_block_cpu,
>
What you're doing seems more "Add the vCPU to the blocked list of
v->processor, which will be the target of the wake-up notification".

> +             * which is the destination of the wake-up notification event.
> +             */
> +            v->arch.hvm_vmx.pi_block_cpu = v->processor;
> +            spin_lock_irqsave(&per_cpu(pi_blocked_vcpu_lock,
> +                              v->arch.hvm_vmx.pi_block_cpu), flags);
> +            list_add_tail(&v->arch.hvm_vmx.pi_blocked_vcpu_list,
> +                          &per_cpu(pi_blocked_vcpu, v->arch.hvm_vmx.pi_block_cpu));
> +            spin_unlock_irqrestore(&per_cpu(pi_blocked_vcpu_lock,
> +                               v->arch.hvm_vmx.pi_block_cpu), flags);
> +
> +            do {
> +                old.control = new.control = pi_desc->control;
> +
> +                /*
> +                 * We should not block the vCPU if
> +                 * an interrupt was posted for it.
> +                 */
> +
> +                if ( old.on )
> +                {
> +                    /*
> +                     * The vCPU will be removed from the block list
> +                     * during its state transferring from RUNSTATE_blocked
> +                     * to RUNSTATE_runnable after the following tasklet
> +                     * is executed.
>
We can avoid referencing RUNSTATEs at all, can't we? Just say something
about the vCPU leaving the blocked vCPUs list on the wake-up path.

> +                     */
> +                    tasklet_schedule(&v->arch.hvm_vmx.pi_vcpu_wakeup_tasklet);
> +                    return;
> +                }
> +
> +                /*
> +                 * Change the 'NDST' field to v->arch.hvm_vmx.pi_block_cpu,
> +                 * so when external interrupts from assigned deivces happen,
> +                 * wakeup notifiction event will go to
> +                 * v->arch.hvm_vmx.pi_block_cpu, then in pi_wakeup_interrupt()
> +                 * we can find the vCPU in the right list to wake up.
> +                 */
> +                if ( x2apic_enabled )
> +                    new.ndst = cpu_physical_id(v->arch.hvm_vmx.pi_block_cpu);
> +                else
> +                    new.ndst = MASK_INSR(cpu_physical_id(
> +                                     v->arch.hvm_vmx.pi_block_cpu),
> +                                     PI_xAPIC_NDST_MASK);
> +                new.sn = 0;
> +                new.nv = pi_wakeup_vector;
> +            } while ( cmpxchg(&pi_desc->control, old.control, new.control)
> +                      != old.control );
> +        }
> +    }
ISTR, Jan had some comments on this code (variable names, etc.). It's
probably goes without saying that those still applies.

>  static void vmx_ctxt_switch_to(struct vcpu *v)
> @@ -764,6 +839,22 @@ static void vmx_ctxt_switch_to(struct vcpu *v)
> 
>      vmx_restore_guest_msrs(v);
>      vmx_restore_dr(v);
> +
> +    if ( iommu_intpost )
> +    {
>
You may consider having an helper for this too, for symmetry with the
above case, but this is less of an issue, IMO.

> +        struct pi_desc *pi_desc = &v->arch.hvm_vmx.pi_desc;
> +
> +        ASSERT( pi_desc->sn == 1 );
                  ^space

Above you wrote:

  ASSERT(v->arch.hvm_vmx.pi_block_cpu == -1);
         ^no space

Please, pick up one format (ideally, following suit from other
occurrences in the file, if any), and be consistent.

> +
> +        if ( x2apic_enabled )
> +            write_atomic(&pi_desc->ndst, cpu_physical_id(v->processor));
> +        else
> +            write_atomic(&pi_desc->ndst,
> +                         MASK_INSR(cpu_physical_id(v->processor),
> +                         PI_xAPIC_NDST_MASK));
> +
> +        pi_clear_sn(pi_desc);
> +    }
>  }

> +void arch_vcpu_wake(struct vcpu *v)
> +{
> +    if ( !iommu_intpost || (v->runstate.state != RUNSTATE_blocked) )
> +        return;
> +
> +    if ( likely(vcpu_runnable(v)) ||
> +         !test_bit(_VPF_blocked, &v->pause_flags) )
> +    {
Invert this and bail if true? Well, a matter of taste, I guess... but it
will save one level of indentation.

> +        struct pi_desc *pi_desc = &v->arch.hvm_vmx.pi_desc;
> +        unsigned long flags;
> +
> +        /*
> +         * blocked -> runnable/offline
> +         * If the state is transferred from RUNSTATE_blocked,
> +         * we should set 'NV' feild back to posted_intr_vector,
> +         * so the Posted-Interrupts can be delivered to the vCPU
> +         * by VT-d HW after it is scheduled to run.
> +         */
>
Again, make the comment describe things in a RUNSTATE independent way
(e.g., in terms of 'generic states', like "it's preempted", "it's
blocked", "it's runnable"; or in terms of flags; or both).

Thanks and Regards,
Dario
-- 
<<This happens because I choose it to happen!>> (Raistlin Majere)
-----------------------------------------------------------------
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)

[-- Attachment #1.2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

[-- Attachment #2: Type: text/plain, Size: 126 bytes --]

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

  parent reply	other threads:[~2015-07-14 16:02 UTC|newest]

Thread overview: 155+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-06-24  5:18 [v3 00/15] Add VT-d Posted-Interrupts support Feng Wu
2015-06-24  5:18 ` [v3 01/15] Vt-d Posted-intterrupt (PI) design Feng Wu
2015-06-24  6:15   ` Meng Xu
2015-06-24  6:19     ` Wu, Feng
2015-07-08  7:21   ` Tian, Kevin
2015-07-08  7:29     ` Wu, Feng
2015-06-24  5:18 ` [v3 02/15] Add helper macro for X86_FEATURE_CX16 feature detection Feng Wu
2015-06-24 17:31   ` Andrew Cooper
2015-07-08  7:23   ` Tian, Kevin
2015-06-24  5:18 ` [v3 03/15] Add cmpxchg16b support for x86-64 Feng Wu
2015-06-24 18:35   ` Andrew Cooper
2015-07-08  7:06     ` Wu, Feng
2015-07-08  8:12       ` Jan Beulich
2015-07-08  8:33         ` Wu, Feng
2015-07-08  8:43           ` Jan Beulich
2015-07-08  8:50             ` Wu, Feng
2015-07-08  8:50         ` Andrew Cooper
2015-07-10 12:57   ` Jan Beulich
2015-06-24  5:18 ` [v3 04/15] iommu: Add iommu_intpost to control VT-d Posted-Interrupts feature Feng Wu
2015-06-25  9:06   ` Andrew Cooper
2015-06-25  9:47     ` Wu, Feng
2015-06-25 10:16       ` Andrew Cooper
2015-06-25 12:47         ` Wu, Feng
2015-07-08  7:30   ` Tian, Kevin
2015-06-24  5:18 ` [v3 05/15] vt-d: VT-d Posted-Interrupts feature detection Feng Wu
2015-06-25 10:21   ` Andrew Cooper
2015-06-25 13:02     ` Wu, Feng
2015-07-08  7:32   ` Tian, Kevin
2015-07-08  8:00     ` Wu, Feng
2015-06-24  5:18 ` [v3 06/15] vmx: Extend struct pi_desc to support VT-d Posted-Interrupts Feng Wu
2015-06-29 15:04   ` Andrew Cooper
2015-07-08  7:48   ` Tian, Kevin
2015-07-10 13:08   ` Jan Beulich
2015-07-15  2:40     ` Wu, Feng
2015-07-15  8:20       ` Jan Beulich
2015-07-15  8:26         ` Wu, Feng
2015-07-15  8:36           ` Jan Beulich
2015-07-15  8:43             ` Wu, Feng
2015-07-15  9:28               ` Jan Beulich
2015-07-15  9:30                 ` Wu, Feng
2015-07-15  3:13     ` Wu, Feng
2015-06-24  5:18 ` [v3 07/15] vmx: Initialize VT-d Posted-Interrupts Descriptor Feng Wu
2015-06-29 15:32   ` Andrew Cooper
2015-06-30  1:46     ` Wu, Feng
2015-06-30  2:32     ` Dario Faggioli
2015-07-08  7:53   ` Tian, Kevin
2015-06-24  5:18 ` [v3 08/15] Suppress posting interrupts when 'SN' is set Feng Wu
2015-06-29 15:41   ` Andrew Cooper
2015-06-30  1:48     ` Wu, Feng
2015-07-08  9:06   ` Tian, Kevin
2015-07-08 10:11     ` Wu, Feng
2015-07-08 11:31       ` Tian, Kevin
2015-07-08 11:58         ` Wu, Feng
2015-07-10 13:20   ` Jan Beulich
2015-06-24  5:18 ` [v3 09/15] vt-d: Extend struct iremap_entry to support VT-d Posted-Interrupts Feng Wu
2015-06-29 16:04   ` Andrew Cooper
2015-06-30  1:52     ` Wu, Feng
2015-07-08  9:10   ` Tian, Kevin
2015-07-10 13:27   ` Jan Beulich
2015-06-24  5:18 ` [v3 10/15] vt-d: Add API to update IRTE when VT-d PI is used Feng Wu
2015-06-29 16:22   ` Andrew Cooper
2015-07-08  9:59   ` Tian, Kevin
2015-07-08 10:12     ` Wu, Feng
2015-07-10 14:01   ` Jan Beulich
2015-07-15  6:04     ` Wu, Feng
2015-07-15  8:24       ` Jan Beulich
2015-07-15  8:38         ` Wu, Feng
2015-07-15  8:46           ` Jan Beulich
2015-07-15  8:55             ` Wu, Feng
2015-07-15  9:32               ` Jan Beulich
2015-06-24  5:18 ` [v3 11/15] Update IRTE according to guest interrupt config changes Feng Wu
2015-06-29 16:46   ` Andrew Cooper
2015-07-08 10:22   ` Tian, Kevin
2015-07-08 10:31     ` Wu, Feng
2015-07-08 11:46       ` Tian, Kevin
2015-07-08 11:52         ` Wu, Feng
2015-07-08 11:54           ` Tian, Kevin
2015-07-10 14:23   ` Jan Beulich
2015-06-24  5:18 ` [v3 12/15] vmx: posted-interrupt handling when vCPU is blocked Feng Wu
2015-06-29 17:07   ` Andrew Cooper
2015-07-08 10:36     ` Wu, Feng
2015-07-08 10:48       ` Jan Beulich
     [not found]   ` <559181F9.6020106@citrix.com>
2015-06-30  2:51     ` Fwd: " Dario Faggioli
2015-06-30  2:59       ` Wu, Feng
2015-06-30  9:46         ` Dario Faggioli
2015-06-30 10:11   ` Andrew Cooper
2015-07-01 13:26     ` Dario Faggioli
2015-07-02  4:27       ` Wu, Feng
2015-07-02  8:30         ` Dario Faggioli
2015-07-02  8:58           ` Wu, Feng
2015-07-02 10:09             ` Dario Faggioli
2015-07-02 10:41               ` Wu, Feng
2015-07-02 10:30           ` Andrew Cooper
2015-07-02 10:56             ` Wu, Feng
2015-07-02 12:04             ` Dario Faggioli
2015-07-02 12:10               ` Wu, Feng
2015-07-02 12:16               ` Andrew Cooper
2015-07-02 12:38                 ` Dario Faggioli
2015-07-02 12:59                   ` Andrew Cooper
2015-07-03  1:33                     ` Wu, Feng
2015-07-02  4:25     ` Wu, Feng
2015-07-08 11:00   ` Tian, Kevin
2015-07-08 11:02     ` Wu, Feng
2015-07-08 12:46     ` Jan Beulich
2015-07-08 13:09       ` Andrew Cooper
2015-07-08 22:49         ` Tian, Kevin
2015-07-09  7:25           ` Jan Beulich
2015-07-10  6:21             ` Wu, Feng
2015-07-10  6:32               ` Jan Beulich
2015-07-10  7:29                 ` Wu, Feng
2015-07-10  8:49                   ` Jan Beulich
2015-07-10  8:57                     ` Wu, Feng
2015-07-08 22:31       ` Tian, Kevin
2015-06-24  5:18 ` [v3 13/15] vmx: Properly handle notification event when vCPU is running Feng Wu
2015-07-08 11:03   ` Tian, Kevin
2015-07-10 14:40   ` Jan Beulich
2015-06-24  5:18 ` [v3 14/15] Update Posted-Interrupts Descriptor during vCPU scheduling Feng Wu
     [not found]   ` <55918214.4030102@citrix.com>
2015-06-30  2:58     ` Fwd: " Dario Faggioli
2015-07-02  4:32       ` Wu, Feng
2015-07-02  4:34         ` Wu, Feng
2015-07-02  8:20         ` Dario Faggioli
2015-07-09  3:09           ` Wu, Feng
2015-07-09  8:18             ` Dario Faggioli
2015-07-09 11:19             ` George Dunlap
2015-07-09 11:29               ` George Dunlap
2015-07-09 11:38               ` Wu, Feng
2015-07-09 12:42                 ` Dario Faggioli
2015-07-10  0:07                   ` Wu, Feng
2015-07-10 12:40                     ` Dario Faggioli
2015-07-10 13:47                       ` Konrad Rzeszutek Wilk
2015-07-10 13:59                         ` Dario Faggioli
2015-07-09 12:53                 ` George Dunlap
2015-07-09 13:44                   ` Jan Beulich
2015-07-09 14:18                     ` Dario Faggioli
2015-07-09 14:27                       ` George Dunlap
2015-07-09 14:47                         ` Dario Faggioli
2015-07-10  5:59                         ` Wu, Feng
2015-07-10  6:22                           ` Jan Beulich
2015-07-10 11:05                             ` Dario Faggioli
2015-07-14  5:44                               ` Wu, Feng
2015-07-14 14:08                               ` Wu, Feng
2015-07-14 14:54                                 ` Jan Beulich
2015-07-14 15:20                                   ` Dario Faggioli
2015-07-14 16:41                                     ` George Dunlap
2015-07-14 16:02                                 ` Dario Faggioli [this message]
2015-07-15  0:54                                   ` Wu, Feng
2015-07-17  7:46                                   ` Wu, Feng
2015-07-17 10:13                                     ` Dario Faggioli
2015-07-17 22:57                                       ` Wu, Feng
2015-07-18 13:43                                         ` Dario Faggioli
2015-07-10  0:15                   ` Wu, Feng
2015-07-08 11:24   ` Tian, Kevin
2015-07-10 14:48   ` Jan Beulich
2015-06-24  5:18 ` [v3 15/15] Add a command line parameter for VT-d posted-interrupts Feng Wu
2015-07-08 11:25   ` Tian, Kevin

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=1436889757.13522.185.camel@citrix.com \
    --to=dario.faggioli@citrix.com \
    --cc=JBeulich@suse.com \
    --cc=andrew.cooper3@citrix.com \
    --cc=feng.wu@intel.com \
    --cc=george.dunlap@eu.citrix.com \
    --cc=keir@xen.org \
    --cc=kevin.tian@intel.com \
    --cc=xen-devel@lists.xen.org \
    --cc=yang.z.zhang@intel.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).