From: Michael Kelley <[email protected]>
Add standard interrupt handler annotations to
hyperv_vector_handler().
Signed-off-by: Michael Kelley <[email protected]>
---
Changes in v2:
* Fixed From: line
---
arch/x86/kernel/cpu/mshyperv.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
index 4488cf0..20f6849 100644
--- a/arch/x86/kernel/cpu/mshyperv.c
+++ b/arch/x86/kernel/cpu/mshyperv.c
@@ -41,7 +41,7 @@ static void (*hv_stimer0_handler)(void);
static void (*hv_kexec_handler)(void);
static void (*hv_crash_handler)(struct pt_regs *regs);
-void hyperv_vector_handler(struct pt_regs *regs)
+__visible void __irq_entry hyperv_vector_handler(struct pt_regs *regs)
{
struct pt_regs *old_regs = set_irq_regs(regs);
--
2.7.4
On Tue, Apr 03, 2018 at 01:59:08PM -0700, [email protected] wrote:
> From: Michael Kelley <[email protected]>
>
> Add standard interrupt handler annotations to
> hyperv_vector_handler().
>
> Signed-off-by: Michael Kelley <[email protected]>
> ---
> Changes in v2:
> * Fixed From: line
> ---
> arch/x86/kernel/cpu/mshyperv.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
> index 4488cf0..20f6849 100644
> --- a/arch/x86/kernel/cpu/mshyperv.c
> +++ b/arch/x86/kernel/cpu/mshyperv.c
> @@ -41,7 +41,7 @@ static void (*hv_stimer0_handler)(void);
> static void (*hv_kexec_handler)(void);
> static void (*hv_crash_handler)(struct pt_regs *regs);
>
> -void hyperv_vector_handler(struct pt_regs *regs)
> +__visible void __irq_entry hyperv_vector_handler(struct pt_regs *regs)
What bug does this solve? What is wrong with the existing markings?
What does __visible and __irq_entry give us that we don't already have
and we need?
Are you really using LTO that requires this marking to prevent the code
from being removed?
thanks,
greg k-h
> -----Original Message-----
> From: Greg KH <[email protected]>
> Sent: Wednesday, April 4, 2018 1:16 AM
> To: Michael Kelley (EOSG) <[email protected]>
> Cc: [email protected]; [email protected]; [email protected];
> [email protected]; [email protected]; [email protected];
> [email protected]; [email protected]; Stephen Hemminger
> <[email protected]>; KY Srinivasan <[email protected]>
> Subject: Re: [PATCH v2 char-misc 1/1] x86/hyperv: Add interrupt handler annotations
>
> On Tue, Apr 03, 2018 at 01:59:08PM -0700, [email protected] wrote:
> > From: Michael Kelley <[email protected]>
> >
> > Add standard interrupt handler annotations to
> > hyperv_vector_handler().
> >
> > Signed-off-by: Michael Kelley <[email protected]>
> > ---
> > Changes in v2:
> > * Fixed From: line
> > ---
> > arch/x86/kernel/cpu/mshyperv.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
> > index 4488cf0..20f6849 100644
> > --- a/arch/x86/kernel/cpu/mshyperv.c
> > +++ b/arch/x86/kernel/cpu/mshyperv.c
> > @@ -41,7 +41,7 @@ static void (*hv_stimer0_handler)(void);
> > static void (*hv_kexec_handler)(void);
> > static void (*hv_crash_handler)(struct pt_regs *regs);
> >
> > -void hyperv_vector_handler(struct pt_regs *regs)
> > +__visible void __irq_entry hyperv_vector_handler(struct pt_regs *regs)
>
> What bug does this solve? What is wrong with the existing markings?
> What does __visible and __irq_entry give us that we don't already have
> and we need?
>
> Are you really using LTO that requires this marking to prevent the code
> from being removed?
Thomas Gleixner commented on Vitaly Kuznetsov's Hyper-V reenlightenment patch
that the interrupt handler should have these annotations: see
https://lkml.org/lkml/2018/1/14/145
I put the same annotations on the interrupt handler for stimer0 Direct Mode,
So this change makes the hyperv_vector_handler() consistent with
hv_stimer0_vector_handler() in the same source file. It does not fix any
immediate bug -- it's for consistency and alignment with what is apparently
standard practice.
Not sure what LTO is ...
Michael
>
> thanks,
>
> greg k-h
On Wed, Apr 04, 2018 at 03:47:03PM +0000, Michael Kelley (EOSG) wrote:
> > -----Original Message-----
> > From: Greg KH <[email protected]>
> > Sent: Wednesday, April 4, 2018 1:16 AM
> > To: Michael Kelley (EOSG) <[email protected]>
> > Cc: [email protected]; [email protected]; [email protected];
> > [email protected]; [email protected]; [email protected];
> > [email protected]; [email protected]; Stephen Hemminger
> > <[email protected]>; KY Srinivasan <[email protected]>
> > Subject: Re: [PATCH v2 char-misc 1/1] x86/hyperv: Add interrupt handler annotations
> >
> > On Tue, Apr 03, 2018 at 01:59:08PM -0700, [email protected] wrote:
> > > From: Michael Kelley <[email protected]>
> > >
> > > Add standard interrupt handler annotations to
> > > hyperv_vector_handler().
> > >
> > > Signed-off-by: Michael Kelley <[email protected]>
> > > ---
> > > Changes in v2:
> > > * Fixed From: line
> > > ---
> > > arch/x86/kernel/cpu/mshyperv.c | 2 +-
> > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
> > > index 4488cf0..20f6849 100644
> > > --- a/arch/x86/kernel/cpu/mshyperv.c
> > > +++ b/arch/x86/kernel/cpu/mshyperv.c
> > > @@ -41,7 +41,7 @@ static void (*hv_stimer0_handler)(void);
> > > static void (*hv_kexec_handler)(void);
> > > static void (*hv_crash_handler)(struct pt_regs *regs);
> > >
> > > -void hyperv_vector_handler(struct pt_regs *regs)
> > > +__visible void __irq_entry hyperv_vector_handler(struct pt_regs *regs)
> >
> > What bug does this solve? What is wrong with the existing markings?
> > What does __visible and __irq_entry give us that we don't already have
> > and we need?
> >
> > Are you really using LTO that requires this marking to prevent the code
> > from being removed?
>
> Thomas Gleixner commented on Vitaly Kuznetsov's Hyper-V reenlightenment patch
> that the interrupt handler should have these annotations: see
> https://lkml.org/lkml/2018/1/14/145
Ok, then someone needs to put a "Suggested-by:" or "Requested-by:" or
something like that tag here, right?
> I put the same annotations on the interrupt handler for stimer0 Direct Mode,
> So this change makes the hyperv_vector_handler() consistent with
> hv_stimer0_vector_handler() in the same source file. It does not fix any
> immediate bug -- it's for consistency and alignment with what is apparently
> standard practice.
>
> Not sure what LTO is ...
That's what the __visable marking fixes! Please go at least _read_ the
definition of the marking you are adding to a function before doing it.
Otherwise this is just cargo-cult-coding :(
thanks,
greg k-h