Drivers often need to first disable an interrupt, carry out some
action, and then reenable the interrupt. Introduce support for the
"guard" notation for this so that the following is possible:
...
scoped_cond_guard(mutex_intr, return -EINTR, &data->sysfs_mutex) {
guard(disable_irq)(&client->irq);
error = elan_acquire_baseline(data);
if (error)
return error;
}
...
Signed-off-by: Dmitry Torokhov <[email protected]>
---
include/linux/interrupt.h | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h
index 5c9bdd3ffccc..3a36e64119c8 100644
--- a/include/linux/interrupt.h
+++ b/include/linux/interrupt.h
@@ -5,6 +5,7 @@
#include <linux/kernel.h>
#include <linux/bitops.h>
+#include <linux/cleanup.h>
#include <linux/cpumask.h>
#include <linux/irqreturn.h>
#include <linux/irqnr.h>
@@ -235,6 +236,9 @@ extern void enable_percpu_irq(unsigned int irq, unsigned int type);
extern bool irq_percpu_is_enabled(unsigned int irq);
extern void irq_wake_thread(unsigned int irq, void *dev_id);
+DEFINE_LOCK_GUARD_1(disable_irq, int,
+ disable_irq(*_T->lock), enable_irq(*_T->lock))
+
extern void disable_nmi_nosync(unsigned int irq);
extern void disable_percpu_nmi(unsigned int irq);
extern void enable_nmi(unsigned int irq);
--
2.45.1.288.g0e0cd299f1-goog
--
Dmitry
On Thu, May 30 2024 at 11:07, Dmitry Torokhov wrote:
> Drivers often need to first disable an interrupt, carry out some
> action, and then reenable the interrupt. Introduce support for the
> "guard" notation for this so that the following is possible:
>
> ...
>
> scoped_cond_guard(mutex_intr, return -EINTR, &data->sysfs_mutex) {
> guard(disable_irq)(&client->irq);
>
> error = elan_acquire_baseline(data);
> if (error)
> return error;
> }
>
> ...
>
> Signed-off-by: Dmitry Torokhov <[email protected]>
I assume you want to carry this through the input tree, so you can
modify the drivers.
If so:
Reviewed-by: Thomas Gleixner <[email protected]>
On Tue, Jun 04, 2024 at 08:51:06PM +0200, Thomas Gleixner wrote:
> On Thu, May 30 2024 at 11:07, Dmitry Torokhov wrote:
> > Drivers often need to first disable an interrupt, carry out some
> > action, and then reenable the interrupt. Introduce support for the
> > "guard" notation for this so that the following is possible:
> >
> > ...
> >
> > scoped_cond_guard(mutex_intr, return -EINTR, &data->sysfs_mutex) {
> > guard(disable_irq)(&client->irq);
> >
> > error = elan_acquire_baseline(data);
> > if (error)
> > return error;
> > }
> >
> > ...
> >
> > Signed-off-by: Dmitry Torokhov <[email protected]>
>
> I assume you want to carry this through the input tree, so you can
> modify the drivers.
>
> If so:
>
> Reviewed-by: Thomas Gleixner <[email protected]>
Thanks Thomas. I'll stick this patch into an immutable branch off v6.9
in case you (or someone else) wants to pull it in and start using it.
Thanks.
--
Dmitry
On Tue, Jun 04, 2024 at 01:55:34PM -0700, Dmitry Torokhov wrote:
> On Tue, Jun 04, 2024 at 08:51:06PM +0200, Thomas Gleixner wrote:
> > On Thu, May 30 2024 at 11:07, Dmitry Torokhov wrote:
> > > Drivers often need to first disable an interrupt, carry out some
> > > action, and then reenable the interrupt. Introduce support for the
> > > "guard" notation for this so that the following is possible:
> > >
> > > ...
> > >
> > > scoped_cond_guard(mutex_intr, return -EINTR, &data->sysfs_mutex) {
> > > guard(disable_irq)(&client->irq);
> > >
> > > error = elan_acquire_baseline(data);
> > > if (error)
> > > return error;
> > > }
> > >
> > > ...
> > >
> > > Signed-off-by: Dmitry Torokhov <[email protected]>
> >
> > I assume you want to carry this through the input tree, so you can
> > modify the drivers.
> >
> > If so:
> >
> > Reviewed-by: Thomas Gleixner <[email protected]>
>
> Thanks Thomas. I'll stick this patch into an immutable branch off v6.9
> in case you (or someone else) wants to pull it in and start using it.
I made:
git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input.git ib/6.9-disable-irq-guard
Thanks.
--
Dmitry