diff mbox series

[v11,10/11] RISC-V: Add supported for ordered booting method using HSM

Message ID 20200318011144.91532-11-atish.patra@wdc.com (mailing list archive)
State New, archived
Headers show
Series Add support for SBI v0.2 and CPU hotplug | expand

Commit Message

Atish Patra March 18, 2020, 1:11 a.m. UTC
Currently, all harts have to jump Linux in RISC-V. This complicates the
multi-stage boot process as every transient stage also has to ensure all
harts enter to that stage and jump to Linux afterwards. It also obstructs
a clean Kexec implementation.

SBI HSM extension provides alternate solutions where only a single hart
need to boot and enter Linux. The booting hart can bring up secondary
harts one by one afterwards.

Add SBI HSM based cpu_ops that implements an ordered booting method in
RISC-V. This change is also backward compatible with older firmware not
implementing HSM extension. If a latest kernel is used with older
firmware, it will continue to use the default spinning booting method.

Signed-off-by: Atish Patra <atish.patra@wdc.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
---
 arch/riscv/kernel/Makefile      |  3 ++
 arch/riscv/kernel/cpu_ops.c     | 10 +++-
 arch/riscv/kernel/cpu_ops_sbi.c | 81 +++++++++++++++++++++++++++++++++
 arch/riscv/kernel/head.S        | 26 +++++++++++
 arch/riscv/kernel/smpboot.c     |  2 +-
 arch/riscv/kernel/traps.c       |  2 +-
 6 files changed, 121 insertions(+), 3 deletions(-)
 create mode 100644 arch/riscv/kernel/cpu_ops_sbi.c

Comments

Bin Meng April 5, 2020, 2:37 p.m. UTC | #1
On Wed, Mar 18, 2020 at 9:12 AM Atish Patra <atish.patra@wdc.com> wrote:
>
> Currently, all harts have to jump Linux in RISC-V. This complicates the
> multi-stage boot process as every transient stage also has to ensure all
> harts enter to that stage and jump to Linux afterwards. It also obstructs
> a clean Kexec implementation.
>
> SBI HSM extension provides alternate solutions where only a single hart
> need to boot and enter Linux. The booting hart can bring up secondary
> harts one by one afterwards.
>
> Add SBI HSM based cpu_ops that implements an ordered booting method in
> RISC-V. This change is also backward compatible with older firmware not
> implementing HSM extension. If a latest kernel is used with older
> firmware, it will continue to use the default spinning booting method.
>
> Signed-off-by: Atish Patra <atish.patra@wdc.com>
> Reviewed-by: Anup Patel <anup@brainfault.org>
> ---
>  arch/riscv/kernel/Makefile      |  3 ++
>  arch/riscv/kernel/cpu_ops.c     | 10 +++-
>  arch/riscv/kernel/cpu_ops_sbi.c | 81 +++++++++++++++++++++++++++++++++
>  arch/riscv/kernel/head.S        | 26 +++++++++++
>  arch/riscv/kernel/smpboot.c     |  2 +-
>  arch/riscv/kernel/traps.c       |  2 +-
>  6 files changed, 121 insertions(+), 3 deletions(-)
>  create mode 100644 arch/riscv/kernel/cpu_ops_sbi.c
>
> diff --git a/arch/riscv/kernel/Makefile b/arch/riscv/kernel/Makefile
> index f81a6ff88005..a0be34b96846 100644
> --- a/arch/riscv/kernel/Makefile
> +++ b/arch/riscv/kernel/Makefile
> @@ -44,5 +44,8 @@ obj-$(CONFIG_PERF_EVENTS)     += perf_event.o
>  obj-$(CONFIG_PERF_EVENTS)      += perf_callchain.o
>  obj-$(CONFIG_HAVE_PERF_REGS)   += perf_regs.o
>  obj-$(CONFIG_RISCV_SBI)                += sbi.o
> +ifeq ($(CONFIG_RISCV_SBI), y)
> +obj-$(CONFIG_SMP) += cpu_ops_sbi.o
> +endif
>
>  clean:
> diff --git a/arch/riscv/kernel/cpu_ops.c b/arch/riscv/kernel/cpu_ops.c
> index 62705908eee5..c4c33bf02369 100644
> --- a/arch/riscv/kernel/cpu_ops.c
> +++ b/arch/riscv/kernel/cpu_ops.c
> @@ -18,6 +18,7 @@ const struct cpu_operations *cpu_ops[NR_CPUS] __ro_after_init;
>  void *__cpu_up_stack_pointer[NR_CPUS];
>  void *__cpu_up_task_pointer[NR_CPUS];
>
> +extern const struct cpu_operations cpu_ops_sbi;
>  extern const struct cpu_operations cpu_ops_spinwait;
>
>  void cpu_update_secondary_bootdata(unsigned int cpuid,
> @@ -34,5 +35,12 @@ void cpu_update_secondary_bootdata(unsigned int cpuid,
>
>  void __init cpu_set_ops(int cpuid)
>  {
> -       cpu_ops[cpuid] = &cpu_ops_spinwait;
> +#if IS_ENABLED(CONFIG_RISCV_SBI)
> +       if (sbi_probe_extension(SBI_EXT_HSM) > 0) {
> +               if (!cpuid)
> +                       pr_info("SBI v0.2 HSM extension detected\n");
> +               cpu_ops[cpuid] = &cpu_ops_sbi;
> +       } else
> +#endif
> +               cpu_ops[cpuid] = &cpu_ops_spinwait;
>  }
> diff --git a/arch/riscv/kernel/cpu_ops_sbi.c b/arch/riscv/kernel/cpu_ops_sbi.c
> new file mode 100644
> index 000000000000..66f3cded91f5
> --- /dev/null
> +++ b/arch/riscv/kernel/cpu_ops_sbi.c
> @@ -0,0 +1,81 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * HSM extension and cpu_ops implementation.
> + *
> + * Copyright (c) 2020 Western Digital Corporation or its affiliates.
> + */
> +
> +#include <linux/init.h>
> +#include <linux/mm.h>
> +#include <asm/cpu_ops.h>
> +#include <asm/sbi.h>
> +#include <asm/smp.h>
> +
> +extern char secondary_start_sbi[];
> +const struct cpu_operations cpu_ops_sbi;
> +
> +static int sbi_hsm_hart_start(unsigned long hartid, unsigned long saddr,
> +                             unsigned long priv)
> +{
> +       struct sbiret ret;
> +
> +       ret = sbi_ecall(SBI_EXT_HSM, SBI_EXT_HSM_HART_START,
> +                       hartid, saddr, priv, 0, 0, 0);
> +       if (ret.error)
> +               return sbi_err_map_linux_errno(ret.error);
> +       else
> +               return 0;
> +}
> +
> +#ifdef CONFIG_HOTPLUG_CPU
> +static int sbi_hsm_hart_stop(void)
> +{
> +       struct sbiret ret;
> +
> +       ret = sbi_ecall(SBI_EXT_HSM, SBI_EXT_HSM_HART_STOP, 0, 0, 0, 0, 0, 0);
> +
> +       if (ret.error)
> +               return sbi_err_map_linux_errno(ret.error);
> +       else
> +               return 0;
> +}
> +
> +static int sbi_hsm_hart_get_status(unsigned long hartid)
> +{
> +       struct sbiret ret;
> +
> +       ret = sbi_ecall(SBI_EXT_HSM, SBI_EXT_HSM_HART_STATUS,
> +                       hartid, 0, 0, 0, 0, 0);
> +       if (ret.error)
> +               return sbi_err_map_linux_errno(ret.error);
> +       else
> +               return ret.value;
> +}
> +#endif
> +
> +static int sbi_cpu_start(unsigned int cpuid, struct task_struct *tidle)
> +{
> +       int rc;
> +       unsigned long boot_addr = __pa_symbol(secondary_start_sbi);
> +       int hartid = cpuid_to_hartid_map(cpuid);
> +
> +       cpu_update_secondary_bootdata(cpuid, tidle);
> +       rc = sbi_hsm_hart_start(hartid, boot_addr, 0);
> +
> +       return rc;
> +}
> +
> +static int sbi_cpu_prepare(unsigned int cpuid)
> +{
> +       if (!cpu_ops_sbi.cpu_start) {
> +               pr_err("cpu start method not defined for CPU [%d]\n", cpuid);
> +               return -ENODEV;
> +       }
> +       return 0;
> +}
> +
> +const struct cpu_operations cpu_ops_sbi = {
> +       .name           = "sbi",
> +       .cpu_prepare    = sbi_cpu_prepare,
> +       .cpu_start      = sbi_cpu_start,
> +};
> diff --git a/arch/riscv/kernel/head.S b/arch/riscv/kernel/head.S
> index 173507395a6b..e5115d5e0b3a 100644
> --- a/arch/riscv/kernel/head.S
> +++ b/arch/riscv/kernel/head.S
> @@ -99,11 +99,37 @@ relocate:
>         ret
>  #endif /* CONFIG_MMU */
>  #ifdef CONFIG_SMP
> +       .global secondary_start_sbi
> +secondary_start_sbi:
> +       /* Mask all interrupts */
> +       csrw CSR_IE, zero
> +       csrw CSR_IP, zero
> +
> +       /* Load the global pointer */
> +       .option push
> +       .option norelax
> +               la gp, __global_pointer$

Is this line's indentation level intended?

> +       .option pop
> +
> +       /*
> +        * Disable FPU to detect illegal usage of
> +        * floating point in kernel space
> +        */
> +       li t0, SR_FS
> +       csrc CSR_STATUS, t0
> +
>         /* Set trap vector to spin forever to help debug */
>         la a3, .Lsecondary_park
>         csrw CSR_TVEC, a3
>
>         slli a3, a0, LGREG
> +       la a4, __cpu_up_stack_pointer
> +       la a5, __cpu_up_task_pointer
> +       add a4, a3, a4
> +       add a5, a3, a5
> +       REG_L sp, (a4)
> +       REG_L tp, (a5)
> +
>         .global secondary_start_common
>  secondary_start_common:
>
> diff --git a/arch/riscv/kernel/smpboot.c b/arch/riscv/kernel/smpboot.c
> index e89396a2a1af..4e9922790f6e 100644
> --- a/arch/riscv/kernel/smpboot.c
> +++ b/arch/riscv/kernel/smpboot.c
> @@ -143,7 +143,7 @@ void __init smp_cpus_done(unsigned int max_cpus)
>  /*
>   * C entry point for a secondary processor.
>   */
> -asmlinkage __visible void __init smp_callin(void)
> +asmlinkage __visible void smp_callin(void)
>  {
>         struct mm_struct *mm = &init_mm;
>
> diff --git a/arch/riscv/kernel/traps.c b/arch/riscv/kernel/traps.c
> index ffb3d94bf0cc..8e13ad45ccaa 100644
> --- a/arch/riscv/kernel/traps.c
> +++ b/arch/riscv/kernel/traps.c
> @@ -147,7 +147,7 @@ int is_valid_bugaddr(unsigned long pc)
>  }
>  #endif /* CONFIG_GENERIC_BUG */
>
> -void __init trap_init(void)
> +void trap_init(void)
>  {
>         /*
>          * Set sup0 scratch register to 0, indicating to exception vector

Should the above 2 be moved to patch [06/11] RISC-V: Move relocate and
few other functions out of __init ?

Other than that,
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>

Regards,
Bin
diff mbox series

Patch

diff --git a/arch/riscv/kernel/Makefile b/arch/riscv/kernel/Makefile
index f81a6ff88005..a0be34b96846 100644
--- a/arch/riscv/kernel/Makefile
+++ b/arch/riscv/kernel/Makefile
@@ -44,5 +44,8 @@  obj-$(CONFIG_PERF_EVENTS)	+= perf_event.o
 obj-$(CONFIG_PERF_EVENTS)	+= perf_callchain.o
 obj-$(CONFIG_HAVE_PERF_REGS)	+= perf_regs.o
 obj-$(CONFIG_RISCV_SBI)		+= sbi.o
+ifeq ($(CONFIG_RISCV_SBI), y)
+obj-$(CONFIG_SMP) += cpu_ops_sbi.o
+endif
 
 clean:
diff --git a/arch/riscv/kernel/cpu_ops.c b/arch/riscv/kernel/cpu_ops.c
index 62705908eee5..c4c33bf02369 100644
--- a/arch/riscv/kernel/cpu_ops.c
+++ b/arch/riscv/kernel/cpu_ops.c
@@ -18,6 +18,7 @@  const struct cpu_operations *cpu_ops[NR_CPUS] __ro_after_init;
 void *__cpu_up_stack_pointer[NR_CPUS];
 void *__cpu_up_task_pointer[NR_CPUS];
 
+extern const struct cpu_operations cpu_ops_sbi;
 extern const struct cpu_operations cpu_ops_spinwait;
 
 void cpu_update_secondary_bootdata(unsigned int cpuid,
@@ -34,5 +35,12 @@  void cpu_update_secondary_bootdata(unsigned int cpuid,
 
 void __init cpu_set_ops(int cpuid)
 {
-	cpu_ops[cpuid] = &cpu_ops_spinwait;
+#if IS_ENABLED(CONFIG_RISCV_SBI)
+	if (sbi_probe_extension(SBI_EXT_HSM) > 0) {
+		if (!cpuid)
+			pr_info("SBI v0.2 HSM extension detected\n");
+		cpu_ops[cpuid] = &cpu_ops_sbi;
+	} else
+#endif
+		cpu_ops[cpuid] = &cpu_ops_spinwait;
 }
diff --git a/arch/riscv/kernel/cpu_ops_sbi.c b/arch/riscv/kernel/cpu_ops_sbi.c
new file mode 100644
index 000000000000..66f3cded91f5
--- /dev/null
+++ b/arch/riscv/kernel/cpu_ops_sbi.c
@@ -0,0 +1,81 @@ 
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * HSM extension and cpu_ops implementation.
+ *
+ * Copyright (c) 2020 Western Digital Corporation or its affiliates.
+ */
+
+#include <linux/init.h>
+#include <linux/mm.h>
+#include <asm/cpu_ops.h>
+#include <asm/sbi.h>
+#include <asm/smp.h>
+
+extern char secondary_start_sbi[];
+const struct cpu_operations cpu_ops_sbi;
+
+static int sbi_hsm_hart_start(unsigned long hartid, unsigned long saddr,
+			      unsigned long priv)
+{
+	struct sbiret ret;
+
+	ret = sbi_ecall(SBI_EXT_HSM, SBI_EXT_HSM_HART_START,
+			hartid, saddr, priv, 0, 0, 0);
+	if (ret.error)
+		return sbi_err_map_linux_errno(ret.error);
+	else
+		return 0;
+}
+
+#ifdef CONFIG_HOTPLUG_CPU
+static int sbi_hsm_hart_stop(void)
+{
+	struct sbiret ret;
+
+	ret = sbi_ecall(SBI_EXT_HSM, SBI_EXT_HSM_HART_STOP, 0, 0, 0, 0, 0, 0);
+
+	if (ret.error)
+		return sbi_err_map_linux_errno(ret.error);
+	else
+		return 0;
+}
+
+static int sbi_hsm_hart_get_status(unsigned long hartid)
+{
+	struct sbiret ret;
+
+	ret = sbi_ecall(SBI_EXT_HSM, SBI_EXT_HSM_HART_STATUS,
+			hartid, 0, 0, 0, 0, 0);
+	if (ret.error)
+		return sbi_err_map_linux_errno(ret.error);
+	else
+		return ret.value;
+}
+#endif
+
+static int sbi_cpu_start(unsigned int cpuid, struct task_struct *tidle)
+{
+	int rc;
+	unsigned long boot_addr = __pa_symbol(secondary_start_sbi);
+	int hartid = cpuid_to_hartid_map(cpuid);
+
+	cpu_update_secondary_bootdata(cpuid, tidle);
+	rc = sbi_hsm_hart_start(hartid, boot_addr, 0);
+
+	return rc;
+}
+
+static int sbi_cpu_prepare(unsigned int cpuid)
+{
+	if (!cpu_ops_sbi.cpu_start) {
+		pr_err("cpu start method not defined for CPU [%d]\n", cpuid);
+		return -ENODEV;
+	}
+	return 0;
+}
+
+const struct cpu_operations cpu_ops_sbi = {
+	.name		= "sbi",
+	.cpu_prepare	= sbi_cpu_prepare,
+	.cpu_start	= sbi_cpu_start,
+};
diff --git a/arch/riscv/kernel/head.S b/arch/riscv/kernel/head.S
index 173507395a6b..e5115d5e0b3a 100644
--- a/arch/riscv/kernel/head.S
+++ b/arch/riscv/kernel/head.S
@@ -99,11 +99,37 @@  relocate:
 	ret
 #endif /* CONFIG_MMU */
 #ifdef CONFIG_SMP
+	.global secondary_start_sbi
+secondary_start_sbi:
+	/* Mask all interrupts */
+	csrw CSR_IE, zero
+	csrw CSR_IP, zero
+
+	/* Load the global pointer */
+	.option push
+	.option norelax
+		la gp, __global_pointer$
+	.option pop
+
+	/*
+	 * Disable FPU to detect illegal usage of
+	 * floating point in kernel space
+	 */
+	li t0, SR_FS
+	csrc CSR_STATUS, t0
+
 	/* Set trap vector to spin forever to help debug */
 	la a3, .Lsecondary_park
 	csrw CSR_TVEC, a3
 
 	slli a3, a0, LGREG
+	la a4, __cpu_up_stack_pointer
+	la a5, __cpu_up_task_pointer
+	add a4, a3, a4
+	add a5, a3, a5
+	REG_L sp, (a4)
+	REG_L tp, (a5)
+
 	.global secondary_start_common
 secondary_start_common:
 
diff --git a/arch/riscv/kernel/smpboot.c b/arch/riscv/kernel/smpboot.c
index e89396a2a1af..4e9922790f6e 100644
--- a/arch/riscv/kernel/smpboot.c
+++ b/arch/riscv/kernel/smpboot.c
@@ -143,7 +143,7 @@  void __init smp_cpus_done(unsigned int max_cpus)
 /*
  * C entry point for a secondary processor.
  */
-asmlinkage __visible void __init smp_callin(void)
+asmlinkage __visible void smp_callin(void)
 {
 	struct mm_struct *mm = &init_mm;
 
diff --git a/arch/riscv/kernel/traps.c b/arch/riscv/kernel/traps.c
index ffb3d94bf0cc..8e13ad45ccaa 100644
--- a/arch/riscv/kernel/traps.c
+++ b/arch/riscv/kernel/traps.c
@@ -147,7 +147,7 @@  int is_valid_bugaddr(unsigned long pc)
 }
 #endif /* CONFIG_GENERIC_BUG */
 
-void __init trap_init(void)
+void trap_init(void)
 {
 	/*
 	 * Set sup0 scratch register to 0, indicating to exception vector