[PATCH V3 0/2] Introduce PSCI-0.2 supports

[PATCH V3 0/2] Introduce PSCI-0.2 supports

[suravee.suthikulpanit]

From: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>

This patch series introduce PSCI-0.2 supports for:
  * PSCI_VERSION
  * CPU_ON
  * SYSTEM_OFF
  * SYSTEM_RESET

Then, it modifies the shutdown/restart logic to use PSCI-0.2 interface
by default when available.

Note: This probably do not need to target the 4.5 release.

Changes in V3:
        - Updated commit messages
 
Changes in V2:
    * 1/2:
        - Use "uint32_t" instead of "int" for psci_ver. (Julien, Ian)
        - In psci_init(), modify logic to first try PSCI-0.2, and fallback to 0.1
          when not supports (Julien, Ian).
        - Fix call_psci_system_off() and psci_system_reset()
        - Rewording of error messages (Julien, Ian)


Suravee Suthikulpanit (2):
  xen/arm: Initial support for PSCI-0.2 (cpu_on, system_off,
    system_reset)
  xen/arm: Use PSCI-0.2 for machine_halt/restart by default

 xen/arch/arm/arm64/smpboot.c |  2 +-
 xen/arch/arm/platform.c      |  2 +-
 xen/arch/arm/psci.c          | 80 +++++++++++++++++++++++++++++++++++++++-----
 xen/arch/arm/shutdown.c      | 25 ++++++++++----
 xen/include/asm-arm/psci.h   |  4 ++-
 5 files changed, 95 insertions(+), 18 deletions(-)

-- 
1.9.3

[PATCH V3 1/2] xen/arm: Initial support for PSCI-0.2

[suravee.suthikulpanit]

From: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>

This patch adds SMC calls to suport a subset of PSCI-0.2 functions
(PSCI_VERSION, CPU_ON, SYSTEM_OFF, SYSTEM_RESET).

By default, the psci_init() will use PSCI-0.2. Otherwise, it will
use PSCI-0.1 if PSCI-0.2 fails or un-supported.

To add support for PSCI_VERSION, this patch replaces the "bool_t psci_available"
variable with "int psci_ver", which contains the PSCI_VERSION as described in the
PSCI-0.2 spec.

Signed-off-by: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
---
 xen/arch/arm/arm64/smpboot.c |  2 +-
 xen/arch/arm/platform.c      |  2 +-
 xen/arch/arm/psci.c          | 80 +++++++++++++++++++++++++++++++++++++++-----
 xen/include/asm-arm/psci.h   |  4 ++-
 4 files changed, 76 insertions(+), 12 deletions(-)

diff --git a/xen/arch/arm/arm64/smpboot.c b/xen/arch/arm/arm64/smpboot.c
index 9146476..341cc77 100644
--- a/xen/arch/arm/arm64/smpboot.c
+++ b/xen/arch/arm/arm64/smpboot.c
@@ -54,7 +54,7 @@ static void __init smp_spin_table_init(int cpu, struct dt_device_node *dn)
 
 static int __init smp_psci_init(int cpu)
 {
-    if ( !psci_available )
+    if ( !psci_ver )
     {
         printk("CPU%d asks for PSCI, but DTB has no PSCI node\n", cpu);
         return -ENODEV;
diff --git a/xen/arch/arm/platform.c b/xen/arch/arm/platform.c
index 74c3328..cb4cda8 100644
--- a/xen/arch/arm/platform.c
+++ b/xen/arch/arm/platform.c
@@ -110,7 +110,7 @@ int __init platform_specific_mapping(struct domain *d)
 #ifdef CONFIG_ARM_32
 int __init platform_cpu_up(int cpu)
 {
-    if ( psci_available )
+    if ( psci_ver )
         return call_psci_cpu_on(cpu);
 
     if ( platform && platform->cpu_up )
diff --git a/xen/arch/arm/psci.c b/xen/arch/arm/psci.c
index b6360d5..604ff4c 100644
--- a/xen/arch/arm/psci.c
+++ b/xen/arch/arm/psci.c
@@ -23,7 +23,7 @@
 #include <xen/smp.h>
 #include <asm/psci.h>
 
-bool_t psci_available;
+uint32_t psci_ver;
 
 #ifdef CONFIG_ARM_32
 #define REG_PREFIX "r"
@@ -58,16 +58,23 @@ int call_psci_cpu_on(int cpu)
                                 cpu_logical_map(cpu), __pa(init_secondary), 0);
 }
 
-int __init psci_init(void)
+void call_psci_system_off(void)
+{
+    if ( psci_ver > XEN_PSCI_V_0_1 )
+        __invoke_psci_fn_smc(PSCI_0_2_FN_SYSTEM_OFF, 0, 0, 0);
+}
+
+void call_psci_system_reset(void)
+{
+    if ( psci_ver > XEN_PSCI_V_0_1 )
+        __invoke_psci_fn_smc(PSCI_0_2_FN_SYSTEM_RESET, 0, 0, 0);
+}
+
+int __init psci_is_smc_method(const struct dt_device_node *psci)
 {
-    const struct dt_device_node *psci;
     int ret;
     const char *prop_str;
 
-    psci = dt_find_compatible_node(NULL, NULL, "arm,psci");
-    if ( !psci )
-        return -ENODEV;
-
     ret = dt_property_read_string(psci, "method", &prop_str);
     if ( ret )
     {
@@ -85,19 +92,74 @@ int __init psci_init(void)
         return -EINVAL;
     }
 
+    return 0;
+}
+
+int __init psci_init_0_1(void)
+{
+    int ret;
+    const struct dt_device_node *psci;
+
+    psci = dt_find_compatible_node(NULL, NULL, "arm,psci");
+    if ( !psci )
+        return -EOPNOTSUPP;
+
+    ret = psci_is_smc_method(psci);
+    if ( ret )
+        return -EINVAL;
+
     if ( !dt_property_read_u32(psci, "cpu_on", &psci_cpu_on_nr) )
     {
         printk("/psci node is missing the \"cpu_on\" property\n");
         return -ENOENT;
     }
 
-    psci_available = 1;
+    psci_ver = XEN_PSCI_V_0_1;
+
+    printk(XENLOG_INFO "Using PSCI-0.1 for SMP bringup\n");
+
+    return 0;
+}
+
+int __init psci_init_0_2(void)
+{
+    int ret;
+    const struct dt_device_node *psci;
+
+    psci = dt_find_compatible_node(NULL, NULL, "arm,psci-0.2");
+    if ( !psci )
+	return -EOPNOTSUPP;
+
+    ret = psci_is_smc_method(psci);
+    if ( ret )
+        return -EINVAL;
+
+    psci_ver = __invoke_psci_fn_smc(PSCI_0_2_FN_PSCI_VERSION, 0, 0, 0);
+
+    if ( psci_ver != XEN_PSCI_V_0_2 )
+    {
+        printk("Error: PSCI version %#x is not supported.\n", psci_ver);
+        return -EOPNOTSUPP;
+    }
+
+    psci_cpu_on_nr = PSCI_0_2_FN_CPU_ON;
 
-    printk(XENLOG_INFO "Using PSCI for SMP bringup\n");
+    printk(XENLOG_INFO "Using PSCI-0.2 for SMP bringup\n");
 
     return 0;
 }
 
+int __init psci_init(void)
+{
+    int ret;
+
+    ret = psci_init_0_2();
+    if ( ret )
+        ret = psci_init_0_1();
+
+    return ret;
+}
+
 /*
  * Local variables:
  * mode: C
diff --git a/xen/include/asm-arm/psci.h b/xen/include/asm-arm/psci.h
index 9777c03..5d17ee3 100644
--- a/xen/include/asm-arm/psci.h
+++ b/xen/include/asm-arm/psci.h
@@ -13,10 +13,12 @@
 #define PSCI_DISABLED               -8
 
 /* availability of PSCI on the host for SMP bringup */
-extern bool_t psci_available;
+extern uint32_t psci_ver;
 
 int psci_init(void);
 int call_psci_cpu_on(int cpu);
+void call_psci_system_off(void);
+void call_psci_system_reset(void);
 
 /* functions to handle guest PSCI requests */
 int32_t do_psci_cpu_on(uint32_t vcpuid, register_t entry_point);
-- 
1.9.3

[PATCH V3 2/2] xen/arm: Use PSCI-0.2 for machine_halt/restart by default

[suravee.suthikulpanit]

From: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>

"machine_halt()" and "machine_restart()" are modified to use PSCI interface
by default if PSCI-0.2 is supported. The "raw_machine_reset()" is also removed
since this is unnecessary.

Also, machine_halt() is modified to add 10ms delay to make sure all auxiliary CPUs
are offlined before calling the shutdown.

For non-PSCI, platform_poweroff() and platform_reset() are used instead.

Signed-off-by: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
---
 xen/arch/arm/shutdown.c | 25 +++++++++++++++++++------
 1 file changed, 19 insertions(+), 6 deletions(-)

diff --git a/xen/arch/arm/shutdown.c b/xen/arch/arm/shutdown.c
index adc0529..4988b03 100644
--- a/xen/arch/arm/shutdown.c
+++ b/xen/arch/arm/shutdown.c
@@ -5,11 +5,7 @@
 #include <xen/lib.h>
 #include <xen/smp.h>
 #include <asm/platform.h>
-
-static void raw_machine_reset(void)
-{
-    platform_reset();
-}
+#include <asm/psci.h>
 
 static void noreturn halt_this_cpu(void *arg)
 {
@@ -18,10 +14,23 @@ static void noreturn halt_this_cpu(void *arg)
 
 void machine_halt(void)
 {
+    int timeout = 10;
+
     watchdog_disable();
     console_start_sync();
     local_irq_enable();
     smp_call_function(halt_this_cpu, NULL, 0);
+    local_irq_disable();
+
+    /* Wait at most another 10ms for all other CPUs to go offline. */
+    while ( (num_online_cpus() > 1) && (timeout-- > 0) )
+        mdelay(1);
+
+    /* This is mainly for PSCI-0.2, which does not return if success. */
+    call_psci_system_off();
+
+    /* Alternative halt procedure */
+    platform_poweroff();
     halt_this_cpu(NULL);
 }
 
@@ -39,9 +48,13 @@ void machine_restart(unsigned int delay_millisecs)
     while ( (num_online_cpus() > 1) && (timeout-- > 0) )
         mdelay(1);
 
+    /* This is mainly for PSCI-0.2, which does not return if success. */
+    call_psci_system_reset();
+
+    /* Alternative reset procedure */
     while ( 1 )
     {
-        raw_machine_reset();
+        platform_reset();
         mdelay(100);
     }
 }
-- 
1.9.3

Re: [PATCH V3 0/2] Introduce PSCI-0.2 supports

[Ian Campbell]

On Fri, 2014-10-03 at 09:31 -0500, suravee.suthikulpanit@amd.com wrote:
> From: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>

Both patches: 

Acked-by: Ian Campbell <ian.campbell@citrix.com>

> 
> This patch series introduce PSCI-0.2 supports for:
>   * PSCI_VERSION
>   * CPU_ON
>   * SYSTEM_OFF
>   * SYSTEM_RESET
> 
> Then, it modifies the shutdown/restart logic to use PSCI-0.2 interface
> by default when available.
> 
> Note: This probably do not need to target the 4.5 release.

The patches are far less scary than I feared they would be and they are
harmless on any platform which doesn't advertise pscsi-0.2 in the device
tree.

I think we currently don't support any pscsi-0.2 aware platforms so
there is no possibility of regressing those, but I fully expect such
systems to become available in the lifetime of 4.5. Testing wise we know
that you have tested on at least one real system, and it's all pretty
simple in any case.

I think we should consider applying for 4.5.

Konrad? What are your thoughts?

Ian.

Re: [PATCH V3 0/2] Introduce PSCI-0.2 supports

[Konrad Rzeszutek Wilk]

On Wed, Oct 08, 2014 at 12:59:46PM +0100, Ian Campbell wrote:
> On Fri, 2014-10-03 at 09:31 -0500, suravee.suthikulpanit@amd.com wrote:
> > From: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
> 
> Both patches: 
> 
> Acked-by: Ian Campbell <ian.campbell@citrix.com>
> 
> > 
> > This patch series introduce PSCI-0.2 supports for:
> >   * PSCI_VERSION
> >   * CPU_ON
> >   * SYSTEM_OFF
> >   * SYSTEM_RESET
> > 
> > Then, it modifies the shutdown/restart logic to use PSCI-0.2 interface
> > by default when available.
> > 
> > Note: This probably do not need to target the 4.5 release.
> 
> The patches are far less scary than I feared they would be and they are
> harmless on any platform which doesn't advertise pscsi-0.2 in the device
> tree.
> 
> I think we currently don't support any pscsi-0.2 aware platforms so
> there is no possibility of regressing those, but I fully expect such
> systems to become available in the lifetime of 4.5. Testing wise we know
> that you have tested on at least one real system, and it's all pretty
> simple in any case.
> 
> I think we should consider applying for 4.5.
> 
> Konrad? What are your thoughts?

Release-Acked-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>

And thank you for enumerating the cons and pros.
> 
> Ian.
> 
> 

Re: [PATCH V3 0/2] Introduce PSCI-0.2 supports

[Ian Campbell]

On Wed, 2014-10-08 at 08:47 -0400, Konrad Rzeszutek Wilk wrote:
> On Wed, Oct 08, 2014 at 12:59:46PM +0100, Ian Campbell wrote:
> > On Fri, 2014-10-03 at 09:31 -0500, suravee.suthikulpanit@amd.com wrote:
> > > From: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
> > 
> > Both patches: 
> > 
> > Acked-by: Ian Campbell <ian.campbell@citrix.com>
[...]
> Release-Acked-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>

Thanks, applied.

Ian.