| Message ID | 20210324152219.1456012-1-vkuznets@redhat.com (mailing list archive) |
|---|---|
| State | Mainlined, archived |
| Headers | show |
| Series | ACPI: processor: Fix CPU0 wakeup in acpi_idle_play_dead() | expand |
On Wed, Mar 24, 2021 at 4:23 PM Vitaly Kuznetsov <vkuznets@redhat.com> wrote: > > Commit 496121c02127 ("ACPI: processor: idle: Allow probing on platforms > with one ACPI C-state") broke CPU0 hotplug on certain systems, e.g. > I'm observing the following on AWS Nitro (e.g r5b.xlarge but other > instance types are affected as well): > > # echo 0 > /sys/devices/system/cpu/cpu0/online > # echo 1 > /sys/devices/system/cpu/cpu0/online > <10 seconds delay> > -bash: echo: write error: Input/output error > > In fact, the above mentioned commit only revealed the problem and did > not introduce it. On x86, to wakeup CPU an NMI is being used and > hlt_play_dead()/mwait_play_dead() loops are prepared to handle it: > > /* > * If NMI wants to wake up CPU0, start CPU0. > */ > if (wakeup_cpu0()) > start_cpu0(); > > cpuidle_play_dead() -> acpi_idle_play_dead() (which is now being called on > systems where it wasn't called before the above mentioned commit) serves > the same purpose but it doesn't have a path for CPU0. What happens now on > wakeup is: > - NMI is sent to CPU0 > - wakeup_cpu0_nmi() works as expected > - we get back to while (1) loop in acpi_idle_play_dead() > - safe_halt() puts CPU0 to sleep again. > > The straightforward/minimal fix is add the special handling for CPU0 on x86 > and that's what the patch is doing. > > Fixes: 496121c02127 ("ACPI: processor: idle: Allow probing on platforms with one ACPI C-state") > Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com> This looks reasonable to me, but I'll give others some time to respond. > --- > arch/x86/include/asm/smp.h | 1 + > arch/x86/kernel/smpboot.c | 2 +- > drivers/acpi/processor_idle.c | 7 +++++++ > 3 files changed, 9 insertions(+), 1 deletion(-) > > diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h > index c0538f82c9a2..57ef2094af93 100644 > --- a/arch/x86/include/asm/smp.h > +++ b/arch/x86/include/asm/smp.h > @@ -132,6 +132,7 @@ void native_play_dead(void); > void play_dead_common(void); > void wbinvd_on_cpu(int cpu); > int wbinvd_on_all_cpus(void); > +bool wakeup_cpu0(void); > > void native_smp_send_reschedule(int cpu); > void native_send_call_func_ipi(const struct cpumask *mask); > diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c > index 02813a7f3a7c..f877150a91da 100644 > --- a/arch/x86/kernel/smpboot.c > +++ b/arch/x86/kernel/smpboot.c > @@ -1659,7 +1659,7 @@ void play_dead_common(void) > local_irq_disable(); > } > > -static bool wakeup_cpu0(void) > +bool wakeup_cpu0(void) > { > if (smp_processor_id() == 0 && enable_start_cpu0) > return true; > diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c > index d93e400940a3..eb1101b16c08 100644 > --- a/drivers/acpi/processor_idle.c > +++ b/drivers/acpi/processor_idle.c > @@ -29,6 +29,7 @@ > */ > #ifdef CONFIG_X86 > #include <asm/apic.h> > +#include <asm/cpu.h> > #endif > > #define _COMPONENT ACPI_PROCESSOR_COMPONENT > @@ -541,6 +542,12 @@ static int acpi_idle_play_dead(struct cpuidle_device *dev, int index) > wait_for_freeze(); > } else > return -ENODEV; > + > +#ifdef CONFIG_X86 > + /* If NMI wants to wake up CPU0, start CPU0. */ > + if (wakeup_cpu0()) > + start_cpu0(); > +#endif > } > > /* Never reached */ > -- > 2.30.2 >
On Wed, Mar 24, 2021 at 4:37 PM Rafael J. Wysocki <rafael@kernel.org> wrote: > > On Wed, Mar 24, 2021 at 4:23 PM Vitaly Kuznetsov <vkuznets@redhat.com> wrote: > > > > Commit 496121c02127 ("ACPI: processor: idle: Allow probing on platforms > > with one ACPI C-state") broke CPU0 hotplug on certain systems, e.g. > > I'm observing the following on AWS Nitro (e.g r5b.xlarge but other > > instance types are affected as well): > > > > # echo 0 > /sys/devices/system/cpu/cpu0/online > > # echo 1 > /sys/devices/system/cpu/cpu0/online > > <10 seconds delay> > > -bash: echo: write error: Input/output error > > > > In fact, the above mentioned commit only revealed the problem and did > > not introduce it. On x86, to wakeup CPU an NMI is being used and > > hlt_play_dead()/mwait_play_dead() loops are prepared to handle it: > > > > /* > > * If NMI wants to wake up CPU0, start CPU0. > > */ > > if (wakeup_cpu0()) > > start_cpu0(); > > > > cpuidle_play_dead() -> acpi_idle_play_dead() (which is now being called on > > systems where it wasn't called before the above mentioned commit) serves > > the same purpose but it doesn't have a path for CPU0. What happens now on > > wakeup is: > > - NMI is sent to CPU0 > > - wakeup_cpu0_nmi() works as expected > > - we get back to while (1) loop in acpi_idle_play_dead() > > - safe_halt() puts CPU0 to sleep again. > > > > The straightforward/minimal fix is add the special handling for CPU0 on x86 > > and that's what the patch is doing. > > > > Fixes: 496121c02127 ("ACPI: processor: idle: Allow probing on platforms with one ACPI C-state") > > Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com> > > This looks reasonable to me, but I'll give others some time to respond. No comments, so applied as 5.12-rc material, thanks! > > --- > > arch/x86/include/asm/smp.h | 1 + > > arch/x86/kernel/smpboot.c | 2 +- > > drivers/acpi/processor_idle.c | 7 +++++++ > > 3 files changed, 9 insertions(+), 1 deletion(-) > > > > diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h > > index c0538f82c9a2..57ef2094af93 100644 > > --- a/arch/x86/include/asm/smp.h > > +++ b/arch/x86/include/asm/smp.h > > @@ -132,6 +132,7 @@ void native_play_dead(void); > > void play_dead_common(void); > > void wbinvd_on_cpu(int cpu); > > int wbinvd_on_all_cpus(void); > > +bool wakeup_cpu0(void); > > > > void native_smp_send_reschedule(int cpu); > > void native_send_call_func_ipi(const struct cpumask *mask); > > diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c > > index 02813a7f3a7c..f877150a91da 100644 > > --- a/arch/x86/kernel/smpboot.c > > +++ b/arch/x86/kernel/smpboot.c > > @@ -1659,7 +1659,7 @@ void play_dead_common(void) > > local_irq_disable(); > > } > > > > -static bool wakeup_cpu0(void) > > +bool wakeup_cpu0(void) > > { > > if (smp_processor_id() == 0 && enable_start_cpu0) > > return true; > > diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c > > index d93e400940a3..eb1101b16c08 100644 > > --- a/drivers/acpi/processor_idle.c > > +++ b/drivers/acpi/processor_idle.c > > @@ -29,6 +29,7 @@ > > */ > > #ifdef CONFIG_X86 > > #include <asm/apic.h> > > +#include <asm/cpu.h> > > #endif > > > > #define _COMPONENT ACPI_PROCESSOR_COMPONENT > > @@ -541,6 +542,12 @@ static int acpi_idle_play_dead(struct cpuidle_device *dev, int index) > > wait_for_freeze(); > > } else > > return -ENODEV; > > + > > +#ifdef CONFIG_X86 > > + /* If NMI wants to wake up CPU0, start CPU0. */ > > + if (wakeup_cpu0()) > > + start_cpu0(); > > +#endif > > } > > > > /* Never reached */ > > -- > > 2.30.2 > >
diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h index c0538f82c9a2..57ef2094af93 100644 --- a/arch/x86/include/asm/smp.h +++ b/arch/x86/include/asm/smp.h @@ -132,6 +132,7 @@ void native_play_dead(void); void play_dead_common(void); void wbinvd_on_cpu(int cpu); int wbinvd_on_all_cpus(void); +bool wakeup_cpu0(void); void native_smp_send_reschedule(int cpu); void native_send_call_func_ipi(const struct cpumask *mask); diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 02813a7f3a7c..f877150a91da 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1659,7 +1659,7 @@ void play_dead_common(void) local_irq_disable(); } -static bool wakeup_cpu0(void) +bool wakeup_cpu0(void) { if (smp_processor_id() == 0 && enable_start_cpu0) return true; diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c index d93e400940a3..eb1101b16c08 100644 --- a/drivers/acpi/processor_idle.c +++ b/drivers/acpi/processor_idle.c @@ -29,6 +29,7 @@ */ #ifdef CONFIG_X86 #include <asm/apic.h> +#include <asm/cpu.h> #endif #define _COMPONENT ACPI_PROCESSOR_COMPONENT @@ -541,6 +542,12 @@ static int acpi_idle_play_dead(struct cpuidle_device *dev, int index) wait_for_freeze(); } else return -ENODEV; + +#ifdef CONFIG_X86 + /* If NMI wants to wake up CPU0, start CPU0. */ + if (wakeup_cpu0()) + start_cpu0(); +#endif } /* Never reached */
Commit 496121c02127 ("ACPI: processor: idle: Allow probing on platforms with one ACPI C-state") broke CPU0 hotplug on certain systems, e.g. I'm observing the following on AWS Nitro (e.g r5b.xlarge but other instance types are affected as well): # echo 0 > /sys/devices/system/cpu/cpu0/online # echo 1 > /sys/devices/system/cpu/cpu0/online <10 seconds delay> -bash: echo: write error: Input/output error In fact, the above mentioned commit only revealed the problem and did not introduce it. On x86, to wakeup CPU an NMI is being used and hlt_play_dead()/mwait_play_dead() loops are prepared to handle it: /* * If NMI wants to wake up CPU0, start CPU0. */ if (wakeup_cpu0()) start_cpu0(); cpuidle_play_dead() -> acpi_idle_play_dead() (which is now being called on systems where it wasn't called before the above mentioned commit) serves the same purpose but it doesn't have a path for CPU0. What happens now on wakeup is: - NMI is sent to CPU0 - wakeup_cpu0_nmi() works as expected - we get back to while (1) loop in acpi_idle_play_dead() - safe_halt() puts CPU0 to sleep again. The straightforward/minimal fix is add the special handling for CPU0 on x86 and that's what the patch is doing. Fixes: 496121c02127 ("ACPI: processor: idle: Allow probing on platforms with one ACPI C-state") Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com> --- arch/x86/include/asm/smp.h | 1 + arch/x86/kernel/smpboot.c | 2 +- drivers/acpi/processor_idle.c | 7 +++++++ 3 files changed, 9 insertions(+), 1 deletion(-)