* [PATCH] rv: Normalize opid IRQ state to a boolean
@ 2026-10-20 3:51 Karl Mehltretter
2026-11-11 10:07 ` Gabriele Monaco
0 siblings, 1 reply; 2+ messages in thread
From: Karl Mehltretter @ 2026-10-20 3:51 UTC (permalink / raw)
To: Steven Rostedt, Gabriele Monaco
Cc: Karl Mehltretter, Masami Hiramatsu, Mathieu Desnoyers, Nam Cao,
linux-trace-kernel, linux-kernel
The opid hybrid-automaton guards compare irq_off with 1. However,
irqs_disabled() only guarantees zero or nonzero. On 32-bit Arm,
arch_irqs_disabled_flags() returns the CPSR I bit, 0x80. Consequently,
valid events with interrupts disabled fail the guard and spuriously
trigger the reactor.
Normalize the value before returning it. Add a KUnit regression test
that handles a need-resched event with local interrupts disabled.
Without the normalization, the test triggers a reaction on Arm.
Fixes: 2b406fdb33387 ("rv: Convert the opid monitor to a hybrid automaton")
Assisted-by: LLM
Signed-off-by: Karl Mehltretter <kmehltretter@gmail.com>
---
Tested with an ARMv7 GCC 16 W=1 build and targeted SMP1 QEMU run:
rv_test_opid and the RV KUnit suite pass.
kernel/trace/rv/monitors/opid/opid.c | 2 +-
kernel/trace/rv/monitors/opid/opid_kunit.c | 6 ++++++
2 files changed, 7 insertions(+), 1 deletion(-)
diff --git a/kernel/trace/rv/monitors/opid/opid.c b/kernel/trace/rv/monitors/opid/opid.c
index 9ae619f176fa1..a14532cc41778 100644
--- a/kernel/trace/rv/monitors/opid/opid.c
+++ b/kernel/trace/rv/monitors/opid/opid.c
@@ -20,7 +20,7 @@
static u64 ha_get_env(struct ha_monitor *ha_mon, enum envs_opid env, u64 time_ns)
{
if (env == irq_off_opid)
- return irqs_disabled();
+ return !!irqs_disabled();
else if (env == preempt_off_opid) {
if (IS_ENABLED(CONFIG_PREEMPTION))
return (preempt_count() & PREEMPT_MASK) > 0;
diff --git a/kernel/trace/rv/monitors/opid/opid_kunit.c b/kernel/trace/rv/monitors/opid/opid_kunit.c
index 3cb087a74241e..8aa1fb9c62344 100644
--- a/kernel/trace/rv/monitors/opid/opid_kunit.c
+++ b/kernel/trace/rv/monitors/opid/opid_kunit.c
@@ -1,5 +1,6 @@
// SPDX-License-Identifier: GPL-2.0
#include <linux/kernel.h>
+#include <linux/irqflags.h>
#include <linux/rv.h>
#include <rv/kunit.h>
#include <trace/events/sched.h>
@@ -17,6 +18,11 @@ static void rv_test_opid(struct kunit *test)
guard(migrate)();
KUNIT_EXPECT_TRUE(test, preemptible());
+ /* Need resched with interrupts disabled */
+ scoped_guard(irq)
+ rv_opid_ops.handle_sched_need_resched(NULL, NULL, 0, TIF_NEED_RESCHED);
+ RV_KUNIT_EXPECT_NO_REACTION(test, ctx);
+
/* Wakeup with preemption and interrupts enabled */
RV_KUNIT_EXPECT_REACTION_HERE(test, ctx)
rv_opid_ops.handle_sched_waking(NULL, NULL);
base-commit: c03c1a14507dfaf8b856c2608f9ac216021fc222
--
2.39.5 (Apple Git-154)
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] rv: Normalize opid IRQ state to a boolean
2026-10-20 3:51 [PATCH] rv: Normalize opid IRQ state to a boolean Karl Mehltretter
@ 2026-11-11 10:07 ` Gabriele Monaco
0 siblings, 0 replies; 2+ messages in thread
From: Gabriele Monaco @ 2026-11-11 10:07 UTC (permalink / raw)
To: Karl Mehltretter, Steven Rostedt, Gabriele Monaco
Cc: Masami Hiramatsu, Mathieu Desnoyers, Nam Cao, linux-trace-kernel,
linux-kernel
On Tue Oct 20, 2026 at 5:51 AM CEST, Karl Mehltretter wrote:
> The opid hybrid-automaton guards compare irq_off with 1. However,
> irqs_disabled() only guarantees zero or nonzero. On 32-bit Arm,
> arch_irqs_disabled_flags() returns the CPSR I bit, 0x80. Consequently,
> valid events with interrupts disabled fail the guard and spuriously
> trigger the reactor.
>
> Normalize the value before returning it. Add a KUnit regression test
> that handles a need-resched event with local interrupts disabled.
> Without the normalization, the test triggers a reaction on Arm.
>
> Fixes: 2b406fdb33387 ("rv: Convert the opid monitor to a hybrid automaton")
> Assisted-by: LLM
> Signed-off-by: Karl Mehltretter <kmehltretter@gmail.com>
> ---
Looks reasonable, thanks!
Acked-by: Gabriele Monaco <gmonaco@redhat.com>
> Tested with an ARMv7 GCC 16 W=1 build and targeted SMP1 QEMU run:
> rv_test_opid and the RV KUnit suite pass.
>
> kernel/trace/rv/monitors/opid/opid.c | 2 +-
> kernel/trace/rv/monitors/opid/opid_kunit.c | 6 ++++++
> 2 files changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/trace/rv/monitors/opid/opid.c b/kernel/trace/rv/monitors/opid/opid.c
> index 9ae619f176fa1..a14532cc41778 100644
> --- a/kernel/trace/rv/monitors/opid/opid.c
> +++ b/kernel/trace/rv/monitors/opid/opid.c
> @@ -20,7 +20,7 @@
> static u64 ha_get_env(struct ha_monitor *ha_mon, enum envs_opid env, u64 time_ns)
> {
> if (env == irq_off_opid)
> - return irqs_disabled();
> + return !!irqs_disabled();
> else if (env == preempt_off_opid) {
> if (IS_ENABLED(CONFIG_PREEMPTION))
> return (preempt_count() & PREEMPT_MASK) > 0;
> diff --git a/kernel/trace/rv/monitors/opid/opid_kunit.c b/kernel/trace/rv/monitors/opid/opid_kunit.c
> index 3cb087a74241e..8aa1fb9c62344 100644
> --- a/kernel/trace/rv/monitors/opid/opid_kunit.c
> +++ b/kernel/trace/rv/monitors/opid/opid_kunit.c
> @@ -1,5 +1,6 @@
> // SPDX-License-Identifier: GPL-2.0
> #include <linux/kernel.h>
> +#include <linux/irqflags.h>
> #include <linux/rv.h>
> #include <rv/kunit.h>
> #include <trace/events/sched.h>
> @@ -17,6 +18,11 @@ static void rv_test_opid(struct kunit *test)
> guard(migrate)();
> KUNIT_EXPECT_TRUE(test, preemptible());
>
> + /* Need resched with interrupts disabled */
> + scoped_guard(irq)
> + rv_opid_ops.handle_sched_need_resched(NULL, NULL, 0, TIF_NEED_RESCHED);
> + RV_KUNIT_EXPECT_NO_REACTION(test, ctx);
> +
> /* Wakeup with preemption and interrupts enabled */
> RV_KUNIT_EXPECT_REACTION_HERE(test, ctx)
> rv_opid_ops.handle_sched_waking(NULL, NULL);
>
> base-commit: c03c1a14507dfaf8b856c2608f9ac216021fc222
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-11-11 10:07 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-10-20 3:51 [PATCH] rv: Normalize opid IRQ state to a boolean Karl Mehltretter
2026-11-11 10:07 ` Gabriele Monaco
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®