[PATCH 0/5] riscv: Make IPI_MAX visible and use it consistently

[PATCH 0/5] riscv: Make IPI_MAX visible and use it consistently

[Guo Ren]

From: "GUO Ren (XuanTie)" <guoren@kernel.org>

This series removes the implicit assumption that RISC-V supports exactly
eight IPI message types.

Currently, the SBI, CLINT and ACLINT SSWI IPI providers use
BITS_PER_BYTE when sizing the generic IPI mux, while IMSIC carries a
separate IMSIC_NR_IPI definition set to 8. These values happen to match
IPI_MAX today, but neither is the proper source of truth for the number
of RISC-V IPI message types.

Move enum ipi_message_type to asm/smp.h so IPI providers can use IPI_MAX
directly, then replace the BITS_PER_BYTE and IMSIC_NR_IPI uses with
IPI_MAX.

This also makes adding future RISC-V IPI message types independent of
the current eight-entry assumption.

GUO Ren (XuanTie) (5):
  riscv: smp: Move enum ipi_message_type to asm/smp.h
  riscv: sbi: Use IPI_MAX for SBI IPI muxing
  clocksource: clint: Use IPI_MAX for IPI muxing
  irqchip/aclint-sswi: Use IPI_MAX for IPI muxing
  irqchip/imsic: Use IPI_MAX instead of IMSIC_NR_IPI

 arch/riscv/include/asm/smp.h            | 12 ++++++++++++
 arch/riscv/kernel/sbi-ipi.c             |  4 ++--
 arch/riscv/kernel/smp.c                 | 12 ------------
 drivers/clocksource/timer-clint.c       |  4 ++--
 drivers/irqchip/irq-aclint-sswi.c       |  4 ++--
 drivers/irqchip/irq-riscv-imsic-early.c |  4 ++--
 drivers/irqchip/irq-riscv-imsic-state.h |  1 -
 7 files changed, 20 insertions(+), 21 deletions(-)

-- 
2.43.0

[PATCH 1/5] riscv: smp: Move enum ipi_message_type to asm/smp.h

[Guo Ren]

From: "GUO Ren (XuanTie)" <guoren@kernel.org>

The IPI message type enumeration (and therefore IPI_MAX) is currently
private to arch/riscv/kernel/smp.c.  Several IPI providers need to know
the exact number of IPIs that the architecture requires, so move the
enum into the public header.

This is a pure code movement with no functional change.

Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>
---
 arch/riscv/include/asm/smp.h | 12 ++++++++++++
 arch/riscv/kernel/smp.c      | 12 ------------
 2 files changed, 12 insertions(+), 12 deletions(-)

diff --git a/arch/riscv/include/asm/smp.h b/arch/riscv/include/asm/smp.h
index 0ecc67641b09..bed39fff1f8a 100644
--- a/arch/riscv/include/asm/smp.h
+++ b/arch/riscv/include/asm/smp.h
@@ -15,6 +15,18 @@
 struct seq_file;
 extern unsigned long boot_cpu_hartid;
 
+enum ipi_message_type {
+	IPI_RESCHEDULE,
+	IPI_CALL_FUNC,
+	IPI_CPU_STOP,
+	IPI_CPU_CRASH_STOP,
+	IPI_IRQ_WORK,
+	IPI_TIMER,
+	IPI_CPU_BACKTRACE,
+	IPI_KGDB_ROUNDUP,
+	IPI_MAX
+};
+
 #ifdef CONFIG_SMP
 
 #include <linux/jump_label.h>
diff --git a/arch/riscv/kernel/smp.c b/arch/riscv/kernel/smp.c
index fa66f9c97d74..8930b62b15e7 100644
--- a/arch/riscv/kernel/smp.c
+++ b/arch/riscv/kernel/smp.c
@@ -28,18 +28,6 @@
 #include <asm/cacheflush.h>
 #include <asm/cpu_ops.h>
 
-enum ipi_message_type {
-	IPI_RESCHEDULE,
-	IPI_CALL_FUNC,
-	IPI_CPU_STOP,
-	IPI_CPU_CRASH_STOP,
-	IPI_IRQ_WORK,
-	IPI_TIMER,
-	IPI_CPU_BACKTRACE,
-	IPI_KGDB_ROUNDUP,
-	IPI_MAX
-};
-
 static const char * const ipi_names[] = {
 	[IPI_RESCHEDULE]	= "Rescheduling interrupts",
 	[IPI_CALL_FUNC]		= "Function call interrupts",
-- 
2.43.0

[PATCH 2/5] riscv: sbi: Use IPI_MAX for SBI IPI muxing

[Guo Ren]

From: "GUO Ren (XuanTie)" <guoren@kernel.org>

Now that IPI_MAX is visible from <asm/smp.h>, replace the hard-coded
BITS_PER_BYTE with the proper symbolic constant in the SBI IPI
extension driver.

No functional change; IPI_MAX is still 8.

Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>
---
 arch/riscv/kernel/sbi-ipi.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/riscv/kernel/sbi-ipi.c b/arch/riscv/kernel/sbi-ipi.c
index 0cc5559c08d8..eeec178a9b95 100644
--- a/arch/riscv/kernel/sbi-ipi.c
+++ b/arch/riscv/kernel/sbi-ipi.c
@@ -57,7 +57,7 @@ void __init sbi_ipi_init(void)
 		return;
 	}
 
-	virq = ipi_mux_create(BITS_PER_BYTE, sbi_send_ipi);
+	virq = ipi_mux_create(IPI_MAX, sbi_send_ipi);
 	if (virq <= 0) {
 		pr_err("unable to create muxed IPIs\n");
 		irq_dispose_mapping(sbi_ipi_virq);
@@ -75,7 +75,7 @@ void __init sbi_ipi_init(void)
 			  "irqchip/sbi-ipi:starting",
 			  sbi_ipi_starting_cpu, NULL);
 
-	riscv_ipi_set_virq_range(virq, BITS_PER_BYTE);
+	riscv_ipi_set_virq_range(virq, IPI_MAX);
 	pr_info("providing IPIs using SBI IPI extension\n");
 
 	/*
-- 
2.43.0

[PATCH 3/5] clocksource: clint: Use IPI_MAX for IPI muxing

[Guo Ren]

From: "GUO Ren (XuanTie)" <guoren@kernel.org>

Replace the hard-coded BITS_PER_BYTE with IPI_MAX now that the
constant is exported from riscv asm/smp.h.

Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>
---
 drivers/clocksource/timer-clint.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/clocksource/timer-clint.c b/drivers/clocksource/timer-clint.c
index 0bdd9d7ec545..e56eee7e3781 100644
--- a/drivers/clocksource/timer-clint.c
+++ b/drivers/clocksource/timer-clint.c
@@ -243,7 +243,7 @@ static int __init clint_timer_init_dt(struct device_node *np)
 	}
 
 #ifdef CONFIG_SMP
-	rc = ipi_mux_create(BITS_PER_BYTE, clint_send_ipi);
+	rc = ipi_mux_create(IPI_MAX, clint_send_ipi);
 	if (rc <= 0) {
 		pr_err("unable to create muxed IPIs\n");
 		rc = (rc < 0) ? rc : -ENODEV;
@@ -251,7 +251,7 @@ static int __init clint_timer_init_dt(struct device_node *np)
 	}
 
 	irq_set_chained_handler(clint_ipi_irq, clint_ipi_interrupt);
-	riscv_ipi_set_virq_range(rc, BITS_PER_BYTE);
+	riscv_ipi_set_virq_range(rc, IPI_MAX);
 	clint_clear_ipi();
 #endif
 
-- 
2.43.0

[PATCH 4/5] irqchip/aclint-sswi: Use IPI_MAX for IPI muxing

[Guo Ren]

From: "GUO Ren (XuanTie)" <guoren@kernel.org>

Replace the hard-coded BITS_PER_BYTE with IPI_MAX now that the constant
is exported from riscv asm/smp.h.

Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>
---
 drivers/irqchip/irq-aclint-sswi.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/irqchip/irq-aclint-sswi.c b/drivers/irqchip/irq-aclint-sswi.c
index ca06efd86fa1..5e010fc401f7 100644
--- a/drivers/irqchip/irq-aclint-sswi.c
+++ b/drivers/irqchip/irq-aclint-sswi.c
@@ -138,7 +138,7 @@ static int __init aclint_sswi_probe(struct fwnode_handle *fwnode)
 	}
 
 	/* Register SSWI irq and handler */
-	virq = ipi_mux_create(BITS_PER_BYTE, aclint_sswi_ipi_send);
+	virq = ipi_mux_create(IPI_MAX, aclint_sswi_ipi_send);
 	if (virq <= 0) {
 		pr_err("unable to create muxed IPIs\n");
 		irq_dispose_mapping(sswi_ipi_virq);
@@ -152,7 +152,7 @@ static int __init aclint_sswi_probe(struct fwnode_handle *fwnode)
 			  aclint_sswi_starting_cpu,
 			  aclint_sswi_dying_cpu);
 
-	riscv_ipi_set_virq_range(virq, BITS_PER_BYTE);
+	riscv_ipi_set_virq_range(virq, IPI_MAX);
 
 	return 0;
 }
-- 
2.43.0

[PATCH 5/5] irqchip/imsic: Use IPI_MAX instead of IMSIC_NR_IPI

[Guo Ren]

From: "GUO Ren (XuanTie)" <guoren@kernel.org>

IMSIC was defining its own IMSIC_NR_IPI (= 8) which happened to match
the architecture's IPI_MAX. Now that IPI_MAX is exported from riscv
asm/smp.h, use the architecture constant and drop the private define.

This keeps the number of multiplexed IPIs in sync with the rest of the
RISC-V IPI infrastructure.

Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>
---
 drivers/irqchip/irq-riscv-imsic-early.c | 4 ++--
 drivers/irqchip/irq-riscv-imsic-state.h | 1 -
 2 files changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/irqchip/irq-riscv-imsic-early.c b/drivers/irqchip/irq-riscv-imsic-early.c
index 12efd241ce88..823f5f2ecb3d 100644
--- a/drivers/irqchip/irq-riscv-imsic-early.c
+++ b/drivers/irqchip/irq-riscv-imsic-early.c
@@ -67,12 +67,12 @@ static int __init imsic_ipi_domain_init(void)
 		return 0;
 
 	/* Create IMSIC IPI multiplexing */
-	virq = ipi_mux_create(IMSIC_NR_IPI, imsic_ipi_send);
+	virq = ipi_mux_create(IPI_MAX, imsic_ipi_send);
 	if (virq <= 0)
 		return virq < 0 ? virq : -ENOMEM;
 
 	/* Set vIRQ range */
-	riscv_ipi_set_virq_range(virq, IMSIC_NR_IPI);
+	riscv_ipi_set_virq_range(virq, IPI_MAX);
 
 	/* Announce that IMSIC is providing IPIs */
 	pr_info("%pfwP: providing IPIs using interrupt %d\n", imsic->fwnode, IMSIC_IPI_ID);
diff --git a/drivers/irqchip/irq-riscv-imsic-state.h b/drivers/irqchip/irq-riscv-imsic-state.h
index c42ee180b305..878cc192ccec 100644
--- a/drivers/irqchip/irq-riscv-imsic-state.h
+++ b/drivers/irqchip/irq-riscv-imsic-state.h
@@ -13,7 +13,6 @@
 #include <linux/timer.h>
 
 #define IMSIC_IPI_ID				1
-#define IMSIC_NR_IPI				8
 
 struct imsic_vector {
 	/* Fixed details of the vector */
-- 
2.43.0

Re: [PATCH 4/5] irqchip/aclint-sswi: Use IPI_MAX for IPI muxing

[Radu Rendec]

On Sun, 2026-08-16 at 07:00 +0000, Guo Ren wrote:
> From: "GUO Ren (XuanTie)" <guoren@kernel.org>
> 
> Replace the hard-coded BITS_PER_BYTE with IPI_MAX now that the constant
> is exported from riscv asm/smp.h.
> 
> Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>
> ---
>  drivers/irqchip/irq-aclint-sswi.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/irqchip/irq-aclint-sswi.c b/drivers/irqchip/irq-aclint-sswi.c
> index ca06efd86fa1..5e010fc401f7 100644
> --- a/drivers/irqchip/irq-aclint-sswi.c
> +++ b/drivers/irqchip/irq-aclint-sswi.c
> @@ -138,7 +138,7 @@ static int __init aclint_sswi_probe(struct fwnode_handle *fwnode)
>  	}
>  
>  	/* Register SSWI irq and handler */
> -	virq = ipi_mux_create(BITS_PER_BYTE, aclint_sswi_ipi_send);
> +	virq = ipi_mux_create(IPI_MAX, aclint_sswi_ipi_send);
>  	if (virq <= 0) {
>  		pr_err("unable to create muxed IPIs\n");
>  		irq_dispose_mapping(sswi_ipi_virq);
> @@ -152,7 +152,7 @@ static int __init aclint_sswi_probe(struct fwnode_handle *fwnode)
>  			  aclint_sswi_starting_cpu,
>  			  aclint_sswi_dying_cpu);
>  
> -	riscv_ipi_set_virq_range(virq, BITS_PER_BYTE);
> +	riscv_ipi_set_virq_range(virq, IPI_MAX);
>  
>  	return 0;
>  }

Reviewed-by: Radu Rendec <radu@rendec.net>

Re: [PATCH 5/5] irqchip/imsic: Use IPI_MAX instead of IMSIC_NR_IPI

[Radu Rendec]

On Sun, 2026-08-16 at 07:00 +0000, Guo Ren wrote:
> From: "GUO Ren (XuanTie)" <guoren@kernel.org>
> 
> IMSIC was defining its own IMSIC_NR_IPI (= 8) which happened to match
> the architecture's IPI_MAX. Now that IPI_MAX is exported from riscv
> asm/smp.h, use the architecture constant and drop the private define.
> 
> This keeps the number of multiplexed IPIs in sync with the rest of the
> RISC-V IPI infrastructure.
> 
> Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>
> ---
>  drivers/irqchip/irq-riscv-imsic-early.c | 4 ++--
>  drivers/irqchip/irq-riscv-imsic-state.h | 1 -
>  2 files changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/irqchip/irq-riscv-imsic-early.c b/drivers/irqchip/irq-riscv-imsic-early.c
> index 12efd241ce88..823f5f2ecb3d 100644
> --- a/drivers/irqchip/irq-riscv-imsic-early.c
> +++ b/drivers/irqchip/irq-riscv-imsic-early.c
> @@ -67,12 +67,12 @@ static int __init imsic_ipi_domain_init(void)
>  		return 0;
>  
>  	/* Create IMSIC IPI multiplexing */
> -	virq = ipi_mux_create(IMSIC_NR_IPI, imsic_ipi_send);
> +	virq = ipi_mux_create(IPI_MAX, imsic_ipi_send);
>  	if (virq <= 0)
>  		return virq < 0 ? virq : -ENOMEM;
>  
>  	/* Set vIRQ range */
> -	riscv_ipi_set_virq_range(virq, IMSIC_NR_IPI);
> +	riscv_ipi_set_virq_range(virq, IPI_MAX);
>  
>  	/* Announce that IMSIC is providing IPIs */
>  	pr_info("%pfwP: providing IPIs using interrupt %d\n", imsic->fwnode, IMSIC_IPI_ID);
> diff --git a/drivers/irqchip/irq-riscv-imsic-state.h b/drivers/irqchip/irq-riscv-imsic-state.h
> index c42ee180b305..878cc192ccec 100644
> --- a/drivers/irqchip/irq-riscv-imsic-state.h
> +++ b/drivers/irqchip/irq-riscv-imsic-state.h
> @@ -13,7 +13,6 @@
>  #include <linux/timer.h>
>  
>  #define IMSIC_IPI_ID				1
> -#define IMSIC_NR_IPI				8
>  
>  struct imsic_vector {
>  	/* Fixed details of the vector */

Reviewed-by: Radu Rendec <radu@rendec.net>

Re: [PATCH 5/5] irqchip/imsic: Use IPI_MAX instead of IMSIC_NR_IPI

[Guo Ren]

On Sun, Aug 16, 2026 at 11:28 PM Radu Rendec <radu@rendec.net> wrote:
>
> On Sun, 2026-08-16 at 07:00 +0000, Guo Ren wrote:
> > From: "GUO Ren (XuanTie)" <guoren@kernel.org>
> >
> > IMSIC was defining its own IMSIC_NR_IPI (= 8) which happened to match
> > the architecture's IPI_MAX. Now that IPI_MAX is exported from riscv
> > asm/smp.h, use the architecture constant and drop the private define.
> >
> > This keeps the number of multiplexed IPIs in sync with the rest of the
> > RISC-V IPI infrastructure.
> >
> > Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>
> > ---
> >  drivers/irqchip/irq-riscv-imsic-early.c | 4 ++--
> >  drivers/irqchip/irq-riscv-imsic-state.h | 1 -
> >  2 files changed, 2 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/irqchip/irq-riscv-imsic-early.c b/drivers/irqchip/irq-riscv-imsic-early.c
> > index 12efd241ce88..823f5f2ecb3d 100644
> > --- a/drivers/irqchip/irq-riscv-imsic-early.c
> > +++ b/drivers/irqchip/irq-riscv-imsic-early.c
> > @@ -67,12 +67,12 @@ static int __init imsic_ipi_domain_init(void)
> >               return 0;
> >
> >       /* Create IMSIC IPI multiplexing */
> > -     virq = ipi_mux_create(IMSIC_NR_IPI, imsic_ipi_send);
> > +     virq = ipi_mux_create(IPI_MAX, imsic_ipi_send);
> >       if (virq <= 0)
> >               return virq < 0 ? virq : -ENOMEM;
> >
> >       /* Set vIRQ range */
> > -     riscv_ipi_set_virq_range(virq, IMSIC_NR_IPI);
> > +     riscv_ipi_set_virq_range(virq, IPI_MAX);
> >
> >       /* Announce that IMSIC is providing IPIs */
> >       pr_info("%pfwP: providing IPIs using interrupt %d\n", imsic->fwnode, IMSIC_IPI_ID);
> > diff --git a/drivers/irqchip/irq-riscv-imsic-state.h b/drivers/irqchip/irq-riscv-imsic-state.h
> > index c42ee180b305..878cc192ccec 100644
> > --- a/drivers/irqchip/irq-riscv-imsic-state.h
> > +++ b/drivers/irqchip/irq-riscv-imsic-state.h
> > @@ -13,7 +13,6 @@
> >  #include <linux/timer.h>
> >
> >  #define IMSIC_IPI_ID                         1
> > -#define IMSIC_NR_IPI                         8
> >
> >  struct imsic_vector {
> >       /* Fixed details of the vector */
>
> Reviewed-by: Radu Rendec <radu@rendec.net>

Thanks for the review, Radu. However, this patch depends on [1]. Would
you mind also reviewing [1]?

[1]: https://lore.kernel.org/linux-riscv/20260816070049.2097442-2-guoren@kernel.org/

-- 
Best Regards
 Guo Ren

Re: [PATCH 1/5] riscv: smp: Move enum ipi_message_type to asm/smp.h

[Radu Rendec]

On Sun, 2026-08-16 at 07:00 +0000, Guo Ren wrote:
> From: "GUO Ren (XuanTie)" <guoren@kernel.org>
> 
> The IPI message type enumeration (and therefore IPI_MAX) is currently
> private to arch/riscv/kernel/smp.c.  Several IPI providers need to know
> the exact number of IPIs that the architecture requires, so move the
> enum into the public header.
> 
> This is a pure code movement with no functional change.
> 
> Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>
> ---
>  arch/riscv/include/asm/smp.h | 12 ++++++++++++
>  arch/riscv/kernel/smp.c      | 12 ------------
>  2 files changed, 12 insertions(+), 12 deletions(-)
> 
> diff --git a/arch/riscv/include/asm/smp.h b/arch/riscv/include/asm/smp.h
> index 0ecc67641b09..bed39fff1f8a 100644
> --- a/arch/riscv/include/asm/smp.h
> +++ b/arch/riscv/include/asm/smp.h
> @@ -15,6 +15,18 @@
>  struct seq_file;
>  extern unsigned long boot_cpu_hartid;
>  
> +enum ipi_message_type {
> +	IPI_RESCHEDULE,
> +	IPI_CALL_FUNC,
> +	IPI_CPU_STOP,
> +	IPI_CPU_CRASH_STOP,
> +	IPI_IRQ_WORK,
> +	IPI_TIMER,
> +	IPI_CPU_BACKTRACE,
> +	IPI_KGDB_ROUNDUP,
> +	IPI_MAX
> +};
> +
>  #ifdef CONFIG_SMP
>  
>  #include <linux/jump_label.h>
> diff --git a/arch/riscv/kernel/smp.c b/arch/riscv/kernel/smp.c
> index fa66f9c97d74..8930b62b15e7 100644
> --- a/arch/riscv/kernel/smp.c
> +++ b/arch/riscv/kernel/smp.c
> @@ -28,18 +28,6 @@
>  #include <asm/cacheflush.h>
>  #include <asm/cpu_ops.h>
>  
> -enum ipi_message_type {
> -	IPI_RESCHEDULE,
> -	IPI_CALL_FUNC,
> -	IPI_CPU_STOP,
> -	IPI_CPU_CRASH_STOP,
> -	IPI_IRQ_WORK,
> -	IPI_TIMER,
> -	IPI_CPU_BACKTRACE,
> -	IPI_KGDB_ROUNDUP,
> -	IPI_MAX
> -};
> -
>  static const char * const ipi_names[] = {
>  	[IPI_RESCHEDULE]	= "Rescheduling interrupts",
>  	[IPI_CALL_FUNC]		= "Function call interrupts",

Reviewed-by: Radu Rendec <radu@rendec.net>

Re: [PATCH 5/5] irqchip/imsic: Use IPI_MAX instead of IMSIC_NR_IPI

[Radu Rendec]

On Mon, 2026-08-17 at 21:05 +0800, Guo Ren wrote:
> On Sun, Aug 16, 2026 at 11:28 PM Radu Rendec <radu@rendec.net> wrote:
> > 
> > On Sun, 2026-08-16 at 07:00 +0000, Guo Ren wrote:
> > > From: "GUO Ren (XuanTie)" <guoren@kernel.org>
> > > 
> > > IMSIC was defining its own IMSIC_NR_IPI (= 8) which happened to match
> > > the architecture's IPI_MAX. Now that IPI_MAX is exported from riscv
> > > asm/smp.h, use the architecture constant and drop the private define.
> > > 
> > > This keeps the number of multiplexed IPIs in sync with the rest of the
> > > RISC-V IPI infrastructure.
> > > 
> > > Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>
> > > ---
> > >  drivers/irqchip/irq-riscv-imsic-early.c | 4 ++--
> > >  drivers/irqchip/irq-riscv-imsic-state.h | 1 -
> > >  2 files changed, 2 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/drivers/irqchip/irq-riscv-imsic-early.c b/drivers/irqchip/irq-riscv-imsic-early.c
> > > index 12efd241ce88..823f5f2ecb3d 100644
> > > --- a/drivers/irqchip/irq-riscv-imsic-early.c
> > > +++ b/drivers/irqchip/irq-riscv-imsic-early.c
> > > @@ -67,12 +67,12 @@ static int __init imsic_ipi_domain_init(void)
> > >               return 0;
> > > 
> > >       /* Create IMSIC IPI multiplexing */
> > > -     virq = ipi_mux_create(IMSIC_NR_IPI, imsic_ipi_send);
> > > +     virq = ipi_mux_create(IPI_MAX, imsic_ipi_send);
> > >       if (virq <= 0)
> > >               return virq < 0 ? virq : -ENOMEM;
> > > 
> > >       /* Set vIRQ range */
> > > -     riscv_ipi_set_virq_range(virq, IMSIC_NR_IPI);
> > > +     riscv_ipi_set_virq_range(virq, IPI_MAX);
> > > 
> > >       /* Announce that IMSIC is providing IPIs */
> > >       pr_info("%pfwP: providing IPIs using interrupt %d\n", imsic->fwnode, IMSIC_IPI_ID);
> > > diff --git a/drivers/irqchip/irq-riscv-imsic-state.h b/drivers/irqchip/irq-riscv-imsic-state.h
> > > index c42ee180b305..878cc192ccec 100644
> > > --- a/drivers/irqchip/irq-riscv-imsic-state.h
> > > +++ b/drivers/irqchip/irq-riscv-imsic-state.h
> > > @@ -13,7 +13,6 @@
> > >  #include <linux/timer.h>
> > > 
> > >  #define IMSIC_IPI_ID                         1
> > > -#define IMSIC_NR_IPI                         8
> > > 
> > >  struct imsic_vector {
> > >       /* Fixed details of the vector */
> > 
> > Reviewed-by: Radu Rendec <radu@rendec.net>
> 
> Thanks for the review, Radu. However, this patch depends on [1]. Would
> you mind also reviewing [1]?
> 
> [1]: https://lore.kernel.org/linux-riscv/20260816070049.2097442-2-guoren@kernel.org/

You're welcome! Sure, that makes sense, I reviewed [1] as well.

-- 
Best regards,
Radu

Re: [PATCH 1/5] riscv: smp: Move enum ipi_message_type to asm/smp.h

[Anup Patel]

On Sun, Aug 16, 2026 at 12:31 PM Guo Ren <guoren@kernel.org> wrote:
>
> From: "GUO Ren (XuanTie)" <guoren@kernel.org>
>
> The IPI message type enumeration (and therefore IPI_MAX) is currently
> private to arch/riscv/kernel/smp.c.  Several IPI providers need to know
> the exact number of IPIs that the architecture requires, so move the
> enum into the public header.
>
> This is a pure code movement with no functional change.
>
> Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>

LGTM.

Reviewed-by: Anup Patel <anup@brainfault.org>

Regards,
Anup

> ---
>  arch/riscv/include/asm/smp.h | 12 ++++++++++++
>  arch/riscv/kernel/smp.c      | 12 ------------
>  2 files changed, 12 insertions(+), 12 deletions(-)
>
> diff --git a/arch/riscv/include/asm/smp.h b/arch/riscv/include/asm/smp.h
> index 0ecc67641b09..bed39fff1f8a 100644
> --- a/arch/riscv/include/asm/smp.h
> +++ b/arch/riscv/include/asm/smp.h
> @@ -15,6 +15,18 @@
>  struct seq_file;
>  extern unsigned long boot_cpu_hartid;
>
> +enum ipi_message_type {
> +       IPI_RESCHEDULE,
> +       IPI_CALL_FUNC,
> +       IPI_CPU_STOP,
> +       IPI_CPU_CRASH_STOP,
> +       IPI_IRQ_WORK,
> +       IPI_TIMER,
> +       IPI_CPU_BACKTRACE,
> +       IPI_KGDB_ROUNDUP,
> +       IPI_MAX
> +};
> +
>  #ifdef CONFIG_SMP
>
>  #include <linux/jump_label.h>
> diff --git a/arch/riscv/kernel/smp.c b/arch/riscv/kernel/smp.c
> index fa66f9c97d74..8930b62b15e7 100644
> --- a/arch/riscv/kernel/smp.c
> +++ b/arch/riscv/kernel/smp.c
> @@ -28,18 +28,6 @@
>  #include <asm/cacheflush.h>
>  #include <asm/cpu_ops.h>
>
> -enum ipi_message_type {
> -       IPI_RESCHEDULE,
> -       IPI_CALL_FUNC,
> -       IPI_CPU_STOP,
> -       IPI_CPU_CRASH_STOP,
> -       IPI_IRQ_WORK,
> -       IPI_TIMER,
> -       IPI_CPU_BACKTRACE,
> -       IPI_KGDB_ROUNDUP,
> -       IPI_MAX
> -};
> -
>  static const char * const ipi_names[] = {
>         [IPI_RESCHEDULE]        = "Rescheduling interrupts",
>         [IPI_CALL_FUNC]         = "Function call interrupts",
> --
> 2.43.0
>

Re: [PATCH 2/5] riscv: sbi: Use IPI_MAX for SBI IPI muxing

[Anup Patel]

On Sun, Aug 16, 2026 at 12:31 PM Guo Ren <guoren@kernel.org> wrote:
>
> From: "GUO Ren (XuanTie)" <guoren@kernel.org>
>
> Now that IPI_MAX is visible from <asm/smp.h>, replace the hard-coded
> BITS_PER_BYTE with the proper symbolic constant in the SBI IPI
> extension driver.
>
> No functional change; IPI_MAX is still 8.
>
> Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>

LGTM.

Reviewed-by: Anup Patel <anup@brainfault.org>

Regards,
Anup

> ---
>  arch/riscv/kernel/sbi-ipi.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/arch/riscv/kernel/sbi-ipi.c b/arch/riscv/kernel/sbi-ipi.c
> index 0cc5559c08d8..eeec178a9b95 100644
> --- a/arch/riscv/kernel/sbi-ipi.c
> +++ b/arch/riscv/kernel/sbi-ipi.c
> @@ -57,7 +57,7 @@ void __init sbi_ipi_init(void)
>                 return;
>         }
>
> -       virq = ipi_mux_create(BITS_PER_BYTE, sbi_send_ipi);
> +       virq = ipi_mux_create(IPI_MAX, sbi_send_ipi);
>         if (virq <= 0) {
>                 pr_err("unable to create muxed IPIs\n");
>                 irq_dispose_mapping(sbi_ipi_virq);
> @@ -75,7 +75,7 @@ void __init sbi_ipi_init(void)
>                           "irqchip/sbi-ipi:starting",
>                           sbi_ipi_starting_cpu, NULL);
>
> -       riscv_ipi_set_virq_range(virq, BITS_PER_BYTE);
> +       riscv_ipi_set_virq_range(virq, IPI_MAX);
>         pr_info("providing IPIs using SBI IPI extension\n");
>
>         /*
> --
> 2.43.0
>

Re: [PATCH 3/5] clocksource: clint: Use IPI_MAX for IPI muxing

[Anup Patel]

On Sun, Aug 16, 2026 at 12:31 PM Guo Ren <guoren@kernel.org> wrote:
>
> From: "GUO Ren (XuanTie)" <guoren@kernel.org>
>
> Replace the hard-coded BITS_PER_BYTE with IPI_MAX now that the
> constant is exported from riscv asm/smp.h.
>
> Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>

LGTM.

Reviewed-by: Anup Patel <anup@brainfault.org>

Regards,
Anup

> ---
>  drivers/clocksource/timer-clint.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/clocksource/timer-clint.c b/drivers/clocksource/timer-clint.c
> index 0bdd9d7ec545..e56eee7e3781 100644
> --- a/drivers/clocksource/timer-clint.c
> +++ b/drivers/clocksource/timer-clint.c
> @@ -243,7 +243,7 @@ static int __init clint_timer_init_dt(struct device_node *np)
>         }
>
>  #ifdef CONFIG_SMP
> -       rc = ipi_mux_create(BITS_PER_BYTE, clint_send_ipi);
> +       rc = ipi_mux_create(IPI_MAX, clint_send_ipi);
>         if (rc <= 0) {
>                 pr_err("unable to create muxed IPIs\n");
>                 rc = (rc < 0) ? rc : -ENODEV;
> @@ -251,7 +251,7 @@ static int __init clint_timer_init_dt(struct device_node *np)
>         }
>
>         irq_set_chained_handler(clint_ipi_irq, clint_ipi_interrupt);
> -       riscv_ipi_set_virq_range(rc, BITS_PER_BYTE);
> +       riscv_ipi_set_virq_range(rc, IPI_MAX);
>         clint_clear_ipi();
>  #endif
>
> --
> 2.43.0
>

Re: [PATCH 4/5] irqchip/aclint-sswi: Use IPI_MAX for IPI muxing

[Anup Patel]

On Sun, Aug 16, 2026 at 12:31 PM Guo Ren <guoren@kernel.org> wrote:
>
> From: "GUO Ren (XuanTie)" <guoren@kernel.org>
>
> Replace the hard-coded BITS_PER_BYTE with IPI_MAX now that the constant
> is exported from riscv asm/smp.h.
>
> Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>

LGTM.

Reviewed-by: Anup Patel <anup@brainfault.org>

Regards,
Anup

> ---
>  drivers/irqchip/irq-aclint-sswi.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/irqchip/irq-aclint-sswi.c b/drivers/irqchip/irq-aclint-sswi.c
> index ca06efd86fa1..5e010fc401f7 100644
> --- a/drivers/irqchip/irq-aclint-sswi.c
> +++ b/drivers/irqchip/irq-aclint-sswi.c
> @@ -138,7 +138,7 @@ static int __init aclint_sswi_probe(struct fwnode_handle *fwnode)
>         }
>
>         /* Register SSWI irq and handler */
> -       virq = ipi_mux_create(BITS_PER_BYTE, aclint_sswi_ipi_send);
> +       virq = ipi_mux_create(IPI_MAX, aclint_sswi_ipi_send);
>         if (virq <= 0) {
>                 pr_err("unable to create muxed IPIs\n");
>                 irq_dispose_mapping(sswi_ipi_virq);
> @@ -152,7 +152,7 @@ static int __init aclint_sswi_probe(struct fwnode_handle *fwnode)
>                           aclint_sswi_starting_cpu,
>                           aclint_sswi_dying_cpu);
>
> -       riscv_ipi_set_virq_range(virq, BITS_PER_BYTE);
> +       riscv_ipi_set_virq_range(virq, IPI_MAX);
>
>         return 0;
>  }
> --
> 2.43.0
>

Re: [PATCH 5/5] irqchip/imsic: Use IPI_MAX instead of IMSIC_NR_IPI

[Anup Patel]

On Sun, Aug 16, 2026 at 12:31 PM Guo Ren <guoren@kernel.org> wrote:
>
> From: "GUO Ren (XuanTie)" <guoren@kernel.org>
>
> IMSIC was defining its own IMSIC_NR_IPI (= 8) which happened to match
> the architecture's IPI_MAX. Now that IPI_MAX is exported from riscv
> asm/smp.h, use the architecture constant and drop the private define.
>
> This keeps the number of multiplexed IPIs in sync with the rest of the
> RISC-V IPI infrastructure.
>
> Signed-off-by: GUO Ren (XuanTie) <guoren@kernel.org>

LGTM.

Reviewed-by: Anup Patel <anup@brainfault.org>

Regards,
Anup

> ---
>  drivers/irqchip/irq-riscv-imsic-early.c | 4 ++--
>  drivers/irqchip/irq-riscv-imsic-state.h | 1 -
>  2 files changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/irqchip/irq-riscv-imsic-early.c b/drivers/irqchip/irq-riscv-imsic-early.c
> index 12efd241ce88..823f5f2ecb3d 100644
> --- a/drivers/irqchip/irq-riscv-imsic-early.c
> +++ b/drivers/irqchip/irq-riscv-imsic-early.c
> @@ -67,12 +67,12 @@ static int __init imsic_ipi_domain_init(void)
>                 return 0;
>
>         /* Create IMSIC IPI multiplexing */
> -       virq = ipi_mux_create(IMSIC_NR_IPI, imsic_ipi_send);
> +       virq = ipi_mux_create(IPI_MAX, imsic_ipi_send);
>         if (virq <= 0)
>                 return virq < 0 ? virq : -ENOMEM;
>
>         /* Set vIRQ range */
> -       riscv_ipi_set_virq_range(virq, IMSIC_NR_IPI);
> +       riscv_ipi_set_virq_range(virq, IPI_MAX);
>
>         /* Announce that IMSIC is providing IPIs */
>         pr_info("%pfwP: providing IPIs using interrupt %d\n", imsic->fwnode, IMSIC_IPI_ID);
> diff --git a/drivers/irqchip/irq-riscv-imsic-state.h b/drivers/irqchip/irq-riscv-imsic-state.h
> index c42ee180b305..878cc192ccec 100644
> --- a/drivers/irqchip/irq-riscv-imsic-state.h
> +++ b/drivers/irqchip/irq-riscv-imsic-state.h
> @@ -13,7 +13,6 @@
>  #include <linux/timer.h>
>
>  #define IMSIC_IPI_ID                           1
> -#define IMSIC_NR_IPI                           8
>
>  struct imsic_vector {
>         /* Fixed details of the vector */
> --
> 2.43.0
>