[PATCH V4 0/3] riscv: kexec: Fxiup crash_save percpu and machine_kexec_mask_interrupts

[PATCH V4 0/3] riscv: kexec: Fxiup crash_save percpu and machine_kexec_mask_interrupts

[guoren]

From: Guo Ren <guoren@linux.alibaba.com>

Current riscv kexec can't crash_save percpu states and disable
interrupts properly. The patch series fix them, make kexec work correct.

Also cleanup arm64, x86 declarations in asm/smp.h.

Changlogs:

v4:
 - Add cpu_ops[cpu]->cpu_stop() in ipi_cpu_crash_stop
 - Wording optimization in comments

V3:
https://lore.kernel.org/linux-riscv/20220819025444.2121315-1-guoren@kernel.org/
 - Fixup compile problem with !SMP, which reported by lkp@intel.com
 - Cleanup declarations in asm/smp.h
 - Add reviewed-by

V2:
https://lore.kernel.org/linux-riscv/20220817161258.748836-1-guoren@kernel.org/ 
 - Add Fixes tags
 - Remove extern from bool smp_crash_stop_failed(void)

V1:
https://lore.kernel.org/linux-riscv/20220816012701.561435-1-guoren@kernel.org/

Guo Ren (3):
  riscv: kexec: Fixup irq controller broken in kexec crash path
  riscv: kexec: Fixup crash_smp_send_stop without multi cores
  arch: crash: Remove duplicate declaration in smp.h

 arch/arm64/include/asm/smp.h      |  1 -
 arch/riscv/include/asm/smp.h      |  3 +
 arch/riscv/kernel/machine_kexec.c | 46 +++++++++++----
 arch/riscv/kernel/smp.c           | 97 ++++++++++++++++++++++++++++++-
 arch/x86/include/asm/crash.h      |  1 -
 5 files changed, 133 insertions(+), 15 deletions(-)

-- 
2.36.1

[PATCH V4 1/3] riscv: kexec: Fixup irq controller broken in kexec crash path

[guoren]

From: Guo Ren <guoren@linux.alibaba.com>

If a crash happens on cpu3 and all interrupts are binding on cpu0, the
bad irq routing will cause a crash kernel which can't receive any irq.
Because crash kernel won't clean up all harts' PLIC enable bits in
enable registers. This patch is similar to 9141a003a491 ("ARM: 7316/1:
kexec: EOI active and mask all interrupts in kexec crash path") and
78fd584cdec0 ("arm64: kdump: implement machine_crash_shutdown()"), and
PowerPC also has the same mechanism.

Fixes: fba8a8674f68 ("RISC-V: Add kexec support")
Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@kernel.org>
Reviewed-by: Xianting Tian <xianting.tian@linux.alibaba.com>
Cc: Nick Kossifidis <mick@ics.forth.gr>
Cc: Palmer Dabbelt <palmer@rivosinc.com>
---
 arch/riscv/kernel/machine_kexec.c | 35 +++++++++++++++++++++++++++++++
 1 file changed, 35 insertions(+)

diff --git a/arch/riscv/kernel/machine_kexec.c b/arch/riscv/kernel/machine_kexec.c
index ee79e6839b86..db41c676e5a2 100644
--- a/arch/riscv/kernel/machine_kexec.c
+++ b/arch/riscv/kernel/machine_kexec.c
@@ -15,6 +15,8 @@
 #include <linux/compiler.h>	/* For unreachable() */
 #include <linux/cpu.h>		/* For cpu_down() */
 #include <linux/reboot.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
 
 /*
  * kexec_image_info - Print received image details
@@ -154,6 +156,37 @@ void crash_smp_send_stop(void)
 	cpus_stopped = 1;
 }
 
+static void machine_kexec_mask_interrupts(void)
+{
+	unsigned int i;
+	struct irq_desc *desc;
+
+	for_each_irq_desc(i, desc) {
+		struct irq_chip *chip;
+		int ret;
+
+		chip = irq_desc_get_chip(desc);
+		if (!chip)
+			continue;
+
+		/*
+		 * First try to remove the active state. If this
+		 * fails, try to EOI the interrupt.
+		 */
+		ret = irq_set_irqchip_state(i, IRQCHIP_STATE_ACTIVE, false);
+
+		if (ret && irqd_irq_inprogress(&desc->irq_data) &&
+		    chip->irq_eoi)
+			chip->irq_eoi(&desc->irq_data);
+
+		if (chip->irq_mask)
+			chip->irq_mask(&desc->irq_data);
+
+		if (chip->irq_disable && !irqd_irq_disabled(&desc->irq_data))
+			chip->irq_disable(&desc->irq_data);
+	}
+}
+
 /*
  * machine_crash_shutdown - Prepare to kexec after a kernel crash
  *
@@ -169,6 +202,8 @@ machine_crash_shutdown(struct pt_regs *regs)
 	crash_smp_send_stop();
 
 	crash_save_cpu(regs, smp_processor_id());
+	machine_kexec_mask_interrupts();
+
 	pr_info("Starting crashdump kernel...\n");
 }
 
-- 
2.36.1

[PATCH V4 2/3] riscv: kexec: Fixup crash_smp_send_stop without multi cores

[guoren]

From: Guo Ren <guoren@linux.alibaba.com>

Current crash_smp_send_stop is the same as the generic one in
kernel/panic and misses crash_save_cpu in percpu. This patch is inspired
by 78fd584cdec0 ("arm64: kdump: implement machine_crash_shutdown()")
and adds the same mechanism for riscv.

Fixes: ad943893d5f1 ("RISC-V: Fixup schedule out issue in machine_crash_shutdown()")
Reviewed-by: Xianting Tian <xianting.tian@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@kernel.org>
Cc: Nick Kossifidis <mick@ics.forth.gr>
---
 arch/riscv/include/asm/smp.h      |  3 +
 arch/riscv/kernel/machine_kexec.c | 21 ++-----
 arch/riscv/kernel/smp.c           | 97 ++++++++++++++++++++++++++++++-
 3 files changed, 103 insertions(+), 18 deletions(-)

diff --git a/arch/riscv/include/asm/smp.h b/arch/riscv/include/asm/smp.h
index d3443be7eedc..3831b638ecab 100644
--- a/arch/riscv/include/asm/smp.h
+++ b/arch/riscv/include/asm/smp.h
@@ -50,6 +50,9 @@ void riscv_set_ipi_ops(const struct riscv_ipi_ops *ops);
 /* Clear IPI for current CPU */
 void riscv_clear_ipi(void);
 
+/* Check other CPUs stop or not */
+bool smp_crash_stop_failed(void);
+
 /* Secondary hart entry */
 asmlinkage void smp_callin(void);
 
diff --git a/arch/riscv/kernel/machine_kexec.c b/arch/riscv/kernel/machine_kexec.c
index db41c676e5a2..2d139b724bc8 100644
--- a/arch/riscv/kernel/machine_kexec.c
+++ b/arch/riscv/kernel/machine_kexec.c
@@ -140,22 +140,6 @@ void machine_shutdown(void)
 #endif
 }
 
-/* Override the weak function in kernel/panic.c */
-void crash_smp_send_stop(void)
-{
-	static int cpus_stopped;
-
-	/*
-	 * This function can be called twice in panic path, but obviously
-	 * we execute this only once.
-	 */
-	if (cpus_stopped)
-		return;
-
-	smp_send_stop();
-	cpus_stopped = 1;
-}
-
 static void machine_kexec_mask_interrupts(void)
 {
 	unsigned int i;
@@ -230,6 +214,11 @@ machine_kexec(struct kimage *image)
 	void *control_code_buffer = page_address(image->control_code_page);
 	riscv_kexec_method kexec_method = NULL;
 
+#ifdef CONFIG_SMP
+	WARN(smp_crash_stop_failed(),
+		"Some CPUs may be stale, kdump will be unreliable.\n");
+#endif
+
 	if (image->type != KEXEC_TYPE_CRASH)
 		kexec_method = control_code_buffer;
 	else
diff --git a/arch/riscv/kernel/smp.c b/arch/riscv/kernel/smp.c
index 760a64518c58..8c3b59f1f9b8 100644
--- a/arch/riscv/kernel/smp.c
+++ b/arch/riscv/kernel/smp.c
@@ -12,6 +12,7 @@
 #include <linux/clockchips.h>
 #include <linux/interrupt.h>
 #include <linux/module.h>
+#include <linux/kexec.h>
 #include <linux/profile.h>
 #include <linux/smp.h>
 #include <linux/sched.h>
@@ -22,11 +23,13 @@
 #include <asm/sbi.h>
 #include <asm/tlbflush.h>
 #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_MAX
@@ -71,6 +74,32 @@ static void ipi_stop(void)
 		wait_for_interrupt();
 }
 
+#ifdef CONFIG_KEXEC_CORE
+static atomic_t waiting_for_crash_ipi = ATOMIC_INIT(0);
+
+static inline void ipi_cpu_crash_stop(unsigned int cpu, struct pt_regs *regs)
+{
+	crash_save_cpu(regs, cpu);
+
+	atomic_dec(&waiting_for_crash_ipi);
+
+	local_irq_disable();
+
+#ifdef CONFIG_HOTPLUG_CPU
+	if (cpu_has_hotplug(cpu))
+		cpu_ops[cpu]->cpu_stop();
+#endif
+
+	for(;;)
+		wait_for_interrupt();
+}
+#else
+static inline void ipi_cpu_crash_stop(unsigned int cpu, struct pt_regs *regs)
+{
+	unreachable();
+}
+#endif
+
 static const struct riscv_ipi_ops *ipi_ops __ro_after_init;
 
 void riscv_set_ipi_ops(const struct riscv_ipi_ops *ops)
@@ -124,8 +153,9 @@ void arch_irq_work_raise(void)
 
 void handle_IPI(struct pt_regs *regs)
 {
-	unsigned long *pending_ipis = &ipi_data[smp_processor_id()].bits;
-	unsigned long *stats = ipi_data[smp_processor_id()].stats;
+	unsigned int cpu = smp_processor_id();
+	unsigned long *pending_ipis = &ipi_data[cpu].bits;
+	unsigned long *stats = ipi_data[cpu].stats;
 
 	riscv_clear_ipi();
 
@@ -154,6 +184,10 @@ void handle_IPI(struct pt_regs *regs)
 			ipi_stop();
 		}
 
+		if (ops & (1 << IPI_CPU_CRASH_STOP)) {
+			ipi_cpu_crash_stop(cpu, get_irq_regs());
+		}
+
 		if (ops & (1 << IPI_IRQ_WORK)) {
 			stats[IPI_IRQ_WORK]++;
 			irq_work_run();
@@ -176,6 +210,7 @@ static const char * const ipi_names[] = {
 	[IPI_RESCHEDULE]	= "Rescheduling interrupts",
 	[IPI_CALL_FUNC]		= "Function call interrupts",
 	[IPI_CPU_STOP]		= "CPU stop interrupts",
+	[IPI_CPU_CRASH_STOP]	= "CPU stop (for crash dump) interrupts",
 	[IPI_IRQ_WORK]		= "IRQ work interrupts",
 	[IPI_TIMER]		= "Timer broadcast interrupts",
 };
@@ -235,6 +270,64 @@ void smp_send_stop(void)
 			   cpumask_pr_args(cpu_online_mask));
 }
 
+#ifdef CONFIG_KEXEC_CORE
+/*
+ * The number of CPUs online, not counting this CPU (which may not be
+ * fully online and so not counted in num_online_cpus()).
+ */
+static inline unsigned int num_other_online_cpus(void)
+{
+	unsigned int this_cpu_online = cpu_online(smp_processor_id());
+
+	return num_online_cpus() - this_cpu_online;
+}
+
+void crash_smp_send_stop(void)
+{
+	static int cpus_stopped;
+	cpumask_t mask;
+	unsigned long timeout;
+
+	/*
+	 * This function can be called twice in panic path, but obviously
+	 * we execute this only once.
+	 */
+	if (cpus_stopped)
+		return;
+
+	cpus_stopped = 1;
+
+	/*
+	 * If this cpu is the only one alive at this point in time, online or
+	 * not, there are no stop messages to be sent around, so just back out.
+	 */
+	if (num_other_online_cpus() == 0)
+		return;
+
+	cpumask_copy(&mask, cpu_online_mask);
+	cpumask_clear_cpu(smp_processor_id(), &mask);
+
+	atomic_set(&waiting_for_crash_ipi, num_other_online_cpus());
+
+	pr_crit("SMP: stopping secondary CPUs\n");
+	send_ipi_mask(&mask, IPI_CPU_CRASH_STOP);
+
+	/* Wait up to one second for other CPUs to stop */
+	timeout = USEC_PER_SEC;
+	while ((atomic_read(&waiting_for_crash_ipi) > 0) && timeout--)
+		udelay(1);
+
+	if (atomic_read(&waiting_for_crash_ipi) > 0)
+		pr_warn("SMP: failed to stop secondary CPUs %*pbl\n",
+			cpumask_pr_args(&mask));
+}
+
+bool smp_crash_stop_failed(void)
+{
+	return (atomic_read(&waiting_for_crash_ipi) > 0);
+}
+#endif
+
 void smp_send_reschedule(int cpu)
 {
 	send_ipi_single(cpu, IPI_RESCHEDULE);
-- 
2.36.1

[PATCH V4 3/3] arch: crash: Remove duplicate declaration in smp.h

[guoren]

From: Guo Ren <guoren@linux.alibaba.com>

Remove crash_smp_send_stop declarations in arm64, x86 asm/smp.h which
has been done in include/linux/smp.h.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@kernel.org>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
---
 arch/arm64/include/asm/smp.h | 1 -
 arch/x86/include/asm/crash.h | 1 -
 2 files changed, 2 deletions(-)

diff --git a/arch/arm64/include/asm/smp.h b/arch/arm64/include/asm/smp.h
index fc55f5a57a06..a108ac93fd8f 100644
--- a/arch/arm64/include/asm/smp.h
+++ b/arch/arm64/include/asm/smp.h
@@ -141,7 +141,6 @@ static inline void cpu_panic_kernel(void)
  */
 bool cpus_are_stuck_in_kernel(void);
 
-extern void crash_smp_send_stop(void);
 extern bool smp_crash_stop_failed(void);
 extern void panic_smp_self_stop(void);
 
diff --git a/arch/x86/include/asm/crash.h b/arch/x86/include/asm/crash.h
index 8b6bd63530dc..6a9be4907c82 100644
--- a/arch/x86/include/asm/crash.h
+++ b/arch/x86/include/asm/crash.h
@@ -7,6 +7,5 @@ struct kimage;
 int crash_load_segments(struct kimage *image);
 int crash_setup_memmap_entries(struct kimage *image,
 		struct boot_params *params);
-void crash_smp_send_stop(void);
 
 #endif /* _ASM_X86_CRASH_H */
-- 
2.36.1

Re: [PATCH V4 3/3] arch: crash: Remove duplicate declaration in smp.h

[Heiko Stuebner]

Am Mittwoch, 21. September 2022, 05:31:34 CEST schrieb guoren@kernel.org:
> From: Guo Ren <guoren@linux.alibaba.com>
> 
> Remove crash_smp_send_stop declarations in arm64, x86 asm/smp.h which
> has been done in include/linux/smp.h.

nit: the commit message could reference the commit that brought in the
generic declarations, which was
	6f1f942cd5fb ("smp: kernel/panic.c - silence warnings")

other than that
Reviewed-by: Heiko Stuebner <heiko@sntech.de>


> Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
> Signed-off-by: Guo Ren <guoren@kernel.org>
> Acked-by: Catalin Marinas <catalin.marinas@arm.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> ---
>  arch/arm64/include/asm/smp.h | 1 -
>  arch/x86/include/asm/crash.h | 1 -
>  2 files changed, 2 deletions(-)
> 
> diff --git a/arch/arm64/include/asm/smp.h b/arch/arm64/include/asm/smp.h
> index fc55f5a57a06..a108ac93fd8f 100644
> --- a/arch/arm64/include/asm/smp.h
> +++ b/arch/arm64/include/asm/smp.h
> @@ -141,7 +141,6 @@ static inline void cpu_panic_kernel(void)
>   */
>  bool cpus_are_stuck_in_kernel(void);
>  
> -extern void crash_smp_send_stop(void);
>  extern bool smp_crash_stop_failed(void);
>  extern void panic_smp_self_stop(void);
>  
> diff --git a/arch/x86/include/asm/crash.h b/arch/x86/include/asm/crash.h
> index 8b6bd63530dc..6a9be4907c82 100644
> --- a/arch/x86/include/asm/crash.h
> +++ b/arch/x86/include/asm/crash.h
> @@ -7,6 +7,5 @@ struct kimage;
>  int crash_load_segments(struct kimage *image);
>  int crash_setup_memmap_entries(struct kimage *image,
>  		struct boot_params *params);
> -void crash_smp_send_stop(void);
>  
>  #endif /* _ASM_X86_CRASH_H */
> 

Re: [PATCH V4 3/3] arch: crash: Remove duplicate declaration in smp.h

[Guo Ren]

On Fri, Sep 23, 2022 at 7:00 PM Heiko Stuebner <heiko@sntech.de> wrote:
>
> Am Mittwoch, 21. September 2022, 05:31:34 CEST schrieb guoren@kernel.org:
> > From: Guo Ren <guoren@linux.alibaba.com>
> >
> > Remove crash_smp_send_stop declarations in arm64, x86 asm/smp.h which
> > has been done in include/linux/smp.h.
>
> nit: the commit message could reference the commit that brought in the
> generic declarations, which was
>         6f1f942cd5fb ("smp: kernel/panic.c - silence warnings")
Good advice, thx.

>
> other than that
> Reviewed-by: Heiko Stuebner <heiko@sntech.de>
>
>
> > Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
> > Signed-off-by: Guo Ren <guoren@kernel.org>
> > Acked-by: Catalin Marinas <catalin.marinas@arm.com>
> > Cc: Arnd Bergmann <arnd@arndb.de>
> > Cc: Thomas Gleixner <tglx@linutronix.de>
> > ---
> >  arch/arm64/include/asm/smp.h | 1 -
> >  arch/x86/include/asm/crash.h | 1 -
> >  2 files changed, 2 deletions(-)
> >
> > diff --git a/arch/arm64/include/asm/smp.h b/arch/arm64/include/asm/smp.h
> > index fc55f5a57a06..a108ac93fd8f 100644
> > --- a/arch/arm64/include/asm/smp.h
> > +++ b/arch/arm64/include/asm/smp.h
> > @@ -141,7 +141,6 @@ static inline void cpu_panic_kernel(void)
> >   */
> >  bool cpus_are_stuck_in_kernel(void);
> >
> > -extern void crash_smp_send_stop(void);
> >  extern bool smp_crash_stop_failed(void);
> >  extern void panic_smp_self_stop(void);
> >
> > diff --git a/arch/x86/include/asm/crash.h b/arch/x86/include/asm/crash.h
> > index 8b6bd63530dc..6a9be4907c82 100644
> > --- a/arch/x86/include/asm/crash.h
> > +++ b/arch/x86/include/asm/crash.h
> > @@ -7,6 +7,5 @@ struct kimage;
> >  int crash_load_segments(struct kimage *image);
> >  int crash_setup_memmap_entries(struct kimage *image,
> >               struct boot_params *params);
> > -void crash_smp_send_stop(void);
> >
> >  #endif /* _ASM_X86_CRASH_H */
> >
>
>
>
>


-- 
Best Regards
 Guo Ren