Mailing List Archive: [PATCH] x86/vmx: reorder code in vmx_deliver_posted

[PATCH] x86/vmx: reorder code in vmx_deliver_posted_intr

Jul 30, 2020, 7:03 AM

Post #1 of 4 (11 views)

Remove the unneeded else branch, which allows to reduce the
indentation of a larger block of code, while making the flow of the
function more obvious.

No functional change intended.

Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
xen/arch/x86/hvm/vmx/vmx.c | 55 ++++++++++++++++++--------------------
1 file changed, 26 insertions(+), 29 deletions(-)

diff --git a/xen/arch/x86/hvm/vmx/vmx.c b/xen/arch/x86/hvm/vmx/vmx.c
index eb54aadfba..7773dcae1b 100644
--- a/xen/arch/x86/hvm/vmx/vmx.c
+++ b/xen/arch/x86/hvm/vmx/vmx.c
@@ -2003,6 +2003,8 @@ static void __vmx_deliver_posted_interrupt(struct vcpu *v)

static void vmx_deliver_posted_intr(struct vcpu *v, u8 vector)
{
+ struct pi_desc old, new, prev;
+
if ( pi_test_and_set_pir(vector, &v->arch.hvm.vmx.pi_desc) )
return;

@@ -2014,41 +2016,36 @@ static void vmx_deliver_posted_intr(struct vcpu *v, u8 vector)
* VMEntry as it used to be.
*/
pi_set_on(&v->arch.hvm.vmx.pi_desc);
+ vcpu_kick(v);
+ return;
}
- else
- {
- struct pi_desc old, new, prev;

- prev.control = v->arch.hvm.vmx.pi_desc.control;
+ prev.control = v->arch.hvm.vmx.pi_desc.control;

- do {
- /*
- * Currently, we don't support urgent interrupt, all
- * interrupts are recognized as non-urgent interrupt,
- * Besides that, if 'ON' is already set, no need to
- * sent posted-interrupts notification event as well,
- * according to hardware behavior.
- */
- if ( pi_test_sn(&prev) || pi_test_on(&prev) )
- {
- vcpu_kick(v);
- return;
- }
-
- old.control = v->arch.hvm.vmx.pi_desc.control &
- ~((1 << POSTED_INTR_ON) | (1 << POSTED_INTR_SN));
- new.control = v->arch.hvm.vmx.pi_desc.control |
- (1 << POSTED_INTR_ON);
+ do {
+ /*
+ * Currently, we don't support urgent interrupt, all
+ * interrupts are recognized as non-urgent interrupt,
+ * Besides that, if 'ON' is already set, no need to
+ * sent posted-interrupts notification event as well,
+ * according to hardware behavior.
+ */
+ if ( pi_test_sn(&prev) || pi_test_on(&prev) )
+ {
+ vcpu_kick(v);
+ return;
+ }

- prev.control = cmpxchg(&v->arch.hvm.vmx.pi_desc.control,
- old.control, new.control);
- } while ( prev.control != old.control );
+ old.control = v->arch.hvm.vmx.pi_desc.control &
+ ~((1 << POSTED_INTR_ON) | (1 << POSTED_INTR_SN));
+ new.control = v->arch.hvm.vmx.pi_desc.control |
+ (1 << POSTED_INTR_ON);

- __vmx_deliver_posted_interrupt(v);
- return;
- }
+ prev.control = cmpxchg(&v->arch.hvm.vmx.pi_desc.control,
+ old.control, new.control);
+ } while ( prev.control != old.control );

- vcpu_kick(v);
+ __vmx_deliver_posted_interrupt(v);
}

static void vmx_sync_pir_to_irr(struct vcpu *v)
--
2.28.0

Re: [PATCH] x86/vmx: reorder code in vmx_deliver_posted_intr [ In reply to ]

jbeulich at suse

Jul 31, 2020, 6:05 AM

Post #2 of 4 (11 views)

Permalink

On 30.07.2020 16:03, Roger Pau Monne wrote:
> Remove the unneeded else branch, which allows to reduce the
> indentation of a larger block of code, while making the flow of the
> function more obvious.
>
> No functional change intended.
>
> Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>

Reviewed-by: Jan Beulich <jbeulich@suse.com>

One minor request (could likely be taken care of while
committing):

> @@ -2014,41 +2016,36 @@ static void vmx_deliver_posted_intr(struct vcpu *v, u8 vector)
> * VMEntry as it used to be.
> */
> pi_set_on(&v->arch.hvm.vmx.pi_desc);
> + vcpu_kick(v);
> + return;
> }
> - else
> - {
> - struct pi_desc old, new, prev;
>
> - prev.control = v->arch.hvm.vmx.pi_desc.control;
> + prev.control = v->arch.hvm.vmx.pi_desc.control;
>
> - do {
> - /*
> - * Currently, we don't support urgent interrupt, all
> - * interrupts are recognized as non-urgent interrupt,
> - * Besides that, if 'ON' is already set, no need to
> - * sent posted-interrupts notification event as well,
> - * according to hardware behavior.
> - */
> - if ( pi_test_sn(&prev) || pi_test_on(&prev) )
> - {
> - vcpu_kick(v);
> - return;
> - }
> -
> - old.control = v->arch.hvm.vmx.pi_desc.control &
> - ~((1 << POSTED_INTR_ON) | (1 << POSTED_INTR_SN));
> - new.control = v->arch.hvm.vmx.pi_desc.control |
> - (1 << POSTED_INTR_ON);
> + do {
> + /*
> + * Currently, we don't support urgent interrupt, all
> + * interrupts are recognized as non-urgent interrupt,
> + * Besides that, if 'ON' is already set, no need to
> + * sent posted-interrupts notification event as well,
> + * according to hardware behavior.
> + */

Would be nice to s/sent/send/ here as you move it (maybe also
remove the plural from "posted-interrupts") and - if possible -
re-flow for the now increased space on the right side.

Jan

Re: [PATCH] x86/vmx: reorder code in vmx_deliver_posted_intr [ In reply to ]

roger.pau at citrix

Jul 31, 2020, 6:56 AM

Post #3 of 4 (11 views)

Permalink

On Fri, Jul 31, 2020 at 03:05:52PM +0200, Jan Beulich wrote:
> On 30.07.2020 16:03, Roger Pau Monne wrote:
> > Remove the unneeded else branch, which allows to reduce the
> > indentation of a larger block of code, while making the flow of the
> > function more obvious.
> >
> > No functional change intended.
> >
> > Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
>
> Reviewed-by: Jan Beulich <jbeulich@suse.com>
>
> One minor request (could likely be taken care of while
> committing):
>
> > @@ -2014,41 +2016,36 @@ static void vmx_deliver_posted_intr(struct vcpu *v, u8 vector)
> > * VMEntry as it used to be.
> > */
> > pi_set_on(&v->arch.hvm.vmx.pi_desc);
> > + vcpu_kick(v);
> > + return;
> > }
> > - else
> > - {
> > - struct pi_desc old, new, prev;
> >
> > - prev.control = v->arch.hvm.vmx.pi_desc.control;
> > + prev.control = v->arch.hvm.vmx.pi_desc.control;
> >
> > - do {
> > - /*
> > - * Currently, we don't support urgent interrupt, all
> > - * interrupts are recognized as non-urgent interrupt,
> > - * Besides that, if 'ON' is already set, no need to
> > - * sent posted-interrupts notification event as well,
> > - * according to hardware behavior.
> > - */
> > - if ( pi_test_sn(&prev) || pi_test_on(&prev) )
> > - {
> > - vcpu_kick(v);
> > - return;
> > - }
> > -
> > - old.control = v->arch.hvm.vmx.pi_desc.control &
> > - ~((1 << POSTED_INTR_ON) | (1 << POSTED_INTR_SN));
> > - new.control = v->arch.hvm.vmx.pi_desc.control |
> > - (1 << POSTED_INTR_ON);
> > + do {
> > + /*
> > + * Currently, we don't support urgent interrupt, all
> > + * interrupts are recognized as non-urgent interrupt,
> > + * Besides that, if 'ON' is already set, no need to
> > + * sent posted-interrupts notification event as well,
> > + * according to hardware behavior.
> > + */
>
> Would be nice to s/sent/send/ here as you move it (maybe also
> remove the plural from "posted-interrupts") and - if possible -
> re-flow for the now increased space on the right side.

Oh, sure, I should have realized myself. Feel free to adjust at commit
if you don't mind. I would also adjust 'non-urgent interrupts'.

Thanks, Roger.

RE: [PATCH] x86/vmx: reorder code in vmx_deliver_posted_intr [ In reply to ]

kevin.tian at intel

Aug 13, 2020, 10:49 PM

Post #4 of 4 (9 views)

Permalink

> From: Roger Pau Monne <roger.pau@citrix.com>
> Sent: Thursday, July 30, 2020 10:03 PM
>
> Remove the unneeded else branch, which allows to reduce the
> indentation of a larger block of code, while making the flow of the
> function more obvious.
>
> No functional change intended.
>
> Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>

Reviewed-by: Kevin Tian <kevin.tian@intel.com>

> ---
> xen/arch/x86/hvm/vmx/vmx.c | 55 ++++++++++++++++++--------------------
> 1 file changed, 26 insertions(+), 29 deletions(-)
>
> diff --git a/xen/arch/x86/hvm/vmx/vmx.c b/xen/arch/x86/hvm/vmx/vmx.c
> index eb54aadfba..7773dcae1b 100644
> --- a/xen/arch/x86/hvm/vmx/vmx.c
> +++ b/xen/arch/x86/hvm/vmx/vmx.c
> @@ -2003,6 +2003,8 @@ static void __vmx_deliver_posted_interrupt(struct
> vcpu *v)
>
> static void vmx_deliver_posted_intr(struct vcpu *v, u8 vector)
> {
> + struct pi_desc old, new, prev;
> +
> if ( pi_test_and_set_pir(vector, &v->arch.hvm.vmx.pi_desc) )
> return;
>
> @@ -2014,41 +2016,36 @@ static void vmx_deliver_posted_intr(struct vcpu
> *v, u8 vector)
> * VMEntry as it used to be.
> */
> pi_set_on(&v->arch.hvm.vmx.pi_desc);
> + vcpu_kick(v);
> + return;
> }
> - else
> - {
> - struct pi_desc old, new, prev;
>
> - prev.control = v->arch.hvm.vmx.pi_desc.control;
> + prev.control = v->arch.hvm.vmx.pi_desc.control;
>
> - do {
> - /*
> - * Currently, we don't support urgent interrupt, all
> - * interrupts are recognized as non-urgent interrupt,
> - * Besides that, if 'ON' is already set, no need to
> - * sent posted-interrupts notification event as well,
> - * according to hardware behavior.
> - */
> - if ( pi_test_sn(&prev) || pi_test_on(&prev) )
> - {
> - vcpu_kick(v);
> - return;
> - }
> -
> - old.control = v->arch.hvm.vmx.pi_desc.control &
> - ~((1 << POSTED_INTR_ON) | (1 << POSTED_INTR_SN));
> - new.control = v->arch.hvm.vmx.pi_desc.control |
> - (1 << POSTED_INTR_ON);
> + do {
> + /*
> + * Currently, we don't support urgent interrupt, all
> + * interrupts are recognized as non-urgent interrupt,
> + * Besides that, if 'ON' is already set, no need to
> + * sent posted-interrupts notification event as well,
> + * according to hardware behavior.
> + */
> + if ( pi_test_sn(&prev) || pi_test_on(&prev) )
> + {
> + vcpu_kick(v);
> + return;
> + }
>
> - prev.control = cmpxchg(&v->arch.hvm.vmx.pi_desc.control,
> - old.control, new.control);
> - } while ( prev.control != old.control );
> + old.control = v->arch.hvm.vmx.pi_desc.control &
> + ~((1 << POSTED_INTR_ON) | (1 << POSTED_INTR_SN));
> + new.control = v->arch.hvm.vmx.pi_desc.control |
> + (1 << POSTED_INTR_ON);
>
> - __vmx_deliver_posted_interrupt(v);
> - return;
> - }
> + prev.control = cmpxchg(&v->arch.hvm.vmx.pi_desc.control,
> + old.control, new.control);
> + } while ( prev.control != old.control );
>
> - vcpu_kick(v);
> + __vmx_deliver_posted_interrupt(v);
> }
>
> static void vmx_sync_pir_to_irr(struct vcpu *v)
> --
> 2.28.0