243e25112d
The XIVE interrupt controller is the new interrupt controller
found in POWER9. It supports advanced virtualization capabilities
among other things.
Currently we use a set of firmware calls that simulate the old
"XICS" interrupt controller but this is fairly inefficient.
This adds the framework for using XIVE along with a native
backend which OPAL for configuration. Later, a backend allowing
the use in a KVM or PowerVM guest will also be provided.
This disables some fast path for interrupts in KVM when XIVE is
enabled as these rely on the firmware emulation code which is no
longer available when the XIVE is used natively by Linux.
A latter patch will make KVM also directly exploit the XIVE, thus
recovering the lost performance (and more).
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
[mpe: Fixup pr_xxx("XIVE:"...), don't split pr_xxx() strings,
tweak Kconfig so XIVE_NATIVE selects XIVE and depends on POWERNV,
fix build errors when SMP=n, fold in fixes from Ben:
Don't call cpu_online() on an invalid CPU number
Fix irq target selection returning out of bounds cpu#
Extra sanity checks on cpu numbers
]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
302 lines
7.7 KiB
C
302 lines
7.7 KiB
C
/*
|
|
* SMP support for PowerNV machines.
|
|
*
|
|
* Copyright 2011 IBM Corp.
|
|
*
|
|
* This program is free software; you can redistribute it and/or
|
|
* modify it under the terms of the GNU General Public License
|
|
* as published by the Free Software Foundation; either version
|
|
* 2 of the License, or (at your option) any later version.
|
|
*/
|
|
|
|
#include <linux/kernel.h>
|
|
#include <linux/module.h>
|
|
#include <linux/sched.h>
|
|
#include <linux/sched/hotplug.h>
|
|
#include <linux/smp.h>
|
|
#include <linux/interrupt.h>
|
|
#include <linux/delay.h>
|
|
#include <linux/init.h>
|
|
#include <linux/spinlock.h>
|
|
#include <linux/cpu.h>
|
|
|
|
#include <asm/irq.h>
|
|
#include <asm/smp.h>
|
|
#include <asm/paca.h>
|
|
#include <asm/machdep.h>
|
|
#include <asm/cputable.h>
|
|
#include <asm/firmware.h>
|
|
#include <asm/vdso_datapage.h>
|
|
#include <asm/cputhreads.h>
|
|
#include <asm/xics.h>
|
|
#include <asm/xive.h>
|
|
#include <asm/opal.h>
|
|
#include <asm/runlatch.h>
|
|
#include <asm/code-patching.h>
|
|
#include <asm/dbell.h>
|
|
#include <asm/kvm_ppc.h>
|
|
#include <asm/ppc-opcode.h>
|
|
|
|
#include "powernv.h"
|
|
|
|
#ifdef DEBUG
|
|
#include <asm/udbg.h>
|
|
#define DBG(fmt...) udbg_printf(fmt)
|
|
#else
|
|
#define DBG(fmt...)
|
|
#endif
|
|
|
|
static void pnv_smp_setup_cpu(int cpu)
|
|
{
|
|
if (xive_enabled())
|
|
xive_smp_setup_cpu();
|
|
else if (cpu != boot_cpuid)
|
|
xics_setup_cpu();
|
|
|
|
#ifdef CONFIG_PPC_DOORBELL
|
|
if (cpu_has_feature(CPU_FTR_DBELL))
|
|
doorbell_setup_this_cpu();
|
|
#endif
|
|
}
|
|
|
|
static int pnv_smp_kick_cpu(int nr)
|
|
{
|
|
unsigned int pcpu = get_hard_smp_processor_id(nr);
|
|
unsigned long start_here =
|
|
__pa(ppc_function_entry(generic_secondary_smp_init));
|
|
long rc;
|
|
uint8_t status;
|
|
|
|
BUG_ON(nr < 0 || nr >= NR_CPUS);
|
|
|
|
/*
|
|
* If we already started or OPAL is not supported, we just
|
|
* kick the CPU via the PACA
|
|
*/
|
|
if (paca[nr].cpu_start || !firmware_has_feature(FW_FEATURE_OPAL))
|
|
goto kick;
|
|
|
|
/*
|
|
* At this point, the CPU can either be spinning on the way in
|
|
* from kexec or be inside OPAL waiting to be started for the
|
|
* first time. OPAL v3 allows us to query OPAL to know if it
|
|
* has the CPUs, so we do that
|
|
*/
|
|
rc = opal_query_cpu_status(pcpu, &status);
|
|
if (rc != OPAL_SUCCESS) {
|
|
pr_warn("OPAL Error %ld querying CPU %d state\n", rc, nr);
|
|
return -ENODEV;
|
|
}
|
|
|
|
/*
|
|
* Already started, just kick it, probably coming from
|
|
* kexec and spinning
|
|
*/
|
|
if (status == OPAL_THREAD_STARTED)
|
|
goto kick;
|
|
|
|
/*
|
|
* Available/inactive, let's kick it
|
|
*/
|
|
if (status == OPAL_THREAD_INACTIVE) {
|
|
pr_devel("OPAL: Starting CPU %d (HW 0x%x)...\n", nr, pcpu);
|
|
rc = opal_start_cpu(pcpu, start_here);
|
|
if (rc != OPAL_SUCCESS) {
|
|
pr_warn("OPAL Error %ld starting CPU %d\n", rc, nr);
|
|
return -ENODEV;
|
|
}
|
|
} else {
|
|
/*
|
|
* An unavailable CPU (or any other unknown status)
|
|
* shouldn't be started. It should also
|
|
* not be in the possible map but currently it can
|
|
* happen
|
|
*/
|
|
pr_devel("OPAL: CPU %d (HW 0x%x) is unavailable"
|
|
" (status %d)...\n", nr, pcpu, status);
|
|
return -ENODEV;
|
|
}
|
|
|
|
kick:
|
|
return smp_generic_kick_cpu(nr);
|
|
}
|
|
|
|
#ifdef CONFIG_HOTPLUG_CPU
|
|
|
|
static int pnv_smp_cpu_disable(void)
|
|
{
|
|
int cpu = smp_processor_id();
|
|
|
|
/* This is identical to pSeries... might consolidate by
|
|
* moving migrate_irqs_away to a ppc_md with default to
|
|
* the generic fixup_irqs. --BenH.
|
|
*/
|
|
set_cpu_online(cpu, false);
|
|
vdso_data->processorCount--;
|
|
if (cpu == boot_cpuid)
|
|
boot_cpuid = cpumask_any(cpu_online_mask);
|
|
if (xive_enabled())
|
|
xive_smp_disable_cpu();
|
|
else
|
|
xics_migrate_irqs_away();
|
|
return 0;
|
|
}
|
|
|
|
static void pnv_smp_cpu_kill_self(void)
|
|
{
|
|
unsigned int cpu;
|
|
unsigned long srr1, wmask;
|
|
u32 idle_states;
|
|
|
|
/* Standard hot unplug procedure */
|
|
local_irq_disable();
|
|
idle_task_exit();
|
|
current->active_mm = NULL; /* for sanity */
|
|
cpu = smp_processor_id();
|
|
DBG("CPU%d offline\n", cpu);
|
|
generic_set_cpu_dead(cpu);
|
|
smp_wmb();
|
|
|
|
wmask = SRR1_WAKEMASK;
|
|
if (cpu_has_feature(CPU_FTR_ARCH_207S))
|
|
wmask = SRR1_WAKEMASK_P8;
|
|
|
|
idle_states = pnv_get_supported_cpuidle_states();
|
|
|
|
/* We don't want to take decrementer interrupts while we are offline,
|
|
* so clear LPCR:PECE1. We keep PECE2 (and LPCR_PECE_HVEE on P9)
|
|
* enabled as to let IPIs in.
|
|
*/
|
|
mtspr(SPRN_LPCR, mfspr(SPRN_LPCR) & ~(u64)LPCR_PECE1);
|
|
|
|
/*
|
|
* Hard-disable interrupts, and then clear irq_happened flags
|
|
* that we can safely ignore while off-line, since they
|
|
* are for things for which we do no processing when off-line
|
|
* (or in the case of HMI, all the processing we need to do
|
|
* is done in lower-level real-mode code).
|
|
*/
|
|
hard_irq_disable();
|
|
local_paca->irq_happened &= ~(PACA_IRQ_DEC | PACA_IRQ_HMI);
|
|
|
|
while (!generic_check_cpu_restart(cpu)) {
|
|
/*
|
|
* Clear IPI flag, since we don't handle IPIs while
|
|
* offline, except for those when changing micro-threading
|
|
* mode, which are handled explicitly below, and those
|
|
* for coming online, which are handled via
|
|
* generic_check_cpu_restart() calls.
|
|
*/
|
|
kvmppc_set_host_ipi(cpu, 0);
|
|
|
|
ppc64_runlatch_off();
|
|
|
|
if (cpu_has_feature(CPU_FTR_ARCH_300)) {
|
|
srr1 = power9_idle_stop(pnv_deepest_stop_psscr_val,
|
|
pnv_deepest_stop_psscr_mask);
|
|
} else if (idle_states & OPAL_PM_WINKLE_ENABLED) {
|
|
srr1 = power7_winkle();
|
|
} else if ((idle_states & OPAL_PM_SLEEP_ENABLED) ||
|
|
(idle_states & OPAL_PM_SLEEP_ENABLED_ER1)) {
|
|
srr1 = power7_sleep();
|
|
} else {
|
|
srr1 = power7_nap(1);
|
|
}
|
|
|
|
ppc64_runlatch_on();
|
|
|
|
/*
|
|
* If the SRR1 value indicates that we woke up due to
|
|
* an external interrupt, then clear the interrupt.
|
|
* We clear the interrupt before checking for the
|
|
* reason, so as to avoid a race where we wake up for
|
|
* some other reason, find nothing and clear the interrupt
|
|
* just as some other cpu is sending us an interrupt.
|
|
* If we returned from power7_nap as a result of
|
|
* having finished executing in a KVM guest, then srr1
|
|
* contains 0.
|
|
*/
|
|
if (((srr1 & wmask) == SRR1_WAKEEE) ||
|
|
((srr1 & wmask) == SRR1_WAKEHVI) ||
|
|
(local_paca->irq_happened & PACA_IRQ_EE)) {
|
|
if (cpu_has_feature(CPU_FTR_ARCH_300)) {
|
|
if (xive_enabled())
|
|
xive_flush_interrupt();
|
|
else
|
|
icp_opal_flush_interrupt();
|
|
} else
|
|
icp_native_flush_interrupt();
|
|
} else if ((srr1 & wmask) == SRR1_WAKEHDBELL) {
|
|
unsigned long msg = PPC_DBELL_TYPE(PPC_DBELL_SERVER);
|
|
asm volatile(PPC_MSGCLR(%0) : : "r" (msg));
|
|
}
|
|
local_paca->irq_happened &= ~(PACA_IRQ_EE | PACA_IRQ_DBELL);
|
|
smp_mb();
|
|
|
|
if (cpu_core_split_required())
|
|
continue;
|
|
|
|
if (srr1 && !generic_check_cpu_restart(cpu))
|
|
DBG("CPU%d Unexpected exit while offline !\n", cpu);
|
|
}
|
|
|
|
/* Re-enable decrementer interrupts */
|
|
mtspr(SPRN_LPCR, mfspr(SPRN_LPCR) | LPCR_PECE1);
|
|
DBG("CPU%d coming online...\n", cpu);
|
|
}
|
|
|
|
#endif /* CONFIG_HOTPLUG_CPU */
|
|
|
|
static int pnv_cpu_bootable(unsigned int nr)
|
|
{
|
|
/*
|
|
* Starting with POWER8, the subcore logic relies on all threads of a
|
|
* core being booted so that they can participate in split mode
|
|
* switches. So on those machines we ignore the smt_enabled_at_boot
|
|
* setting (smt-enabled on the kernel command line).
|
|
*/
|
|
if (cpu_has_feature(CPU_FTR_ARCH_207S))
|
|
return 1;
|
|
|
|
return smp_generic_cpu_bootable(nr);
|
|
}
|
|
|
|
static int pnv_smp_prepare_cpu(int cpu)
|
|
{
|
|
if (xive_enabled())
|
|
return xive_smp_prepare_cpu(cpu);
|
|
return 0;
|
|
}
|
|
|
|
static void __init pnv_smp_probe(void)
|
|
{
|
|
if (xive_enabled())
|
|
xive_smp_probe();
|
|
else
|
|
xics_smp_probe();
|
|
}
|
|
|
|
static struct smp_ops_t pnv_smp_ops = {
|
|
.message_pass = smp_muxed_ipi_message_pass,
|
|
.cause_ipi = NULL, /* Filled at runtime by xi{cs,ve}_smp_probe() */
|
|
.probe = pnv_smp_probe,
|
|
.prepare_cpu = pnv_smp_prepare_cpu,
|
|
.kick_cpu = pnv_smp_kick_cpu,
|
|
.setup_cpu = pnv_smp_setup_cpu,
|
|
.cpu_bootable = pnv_cpu_bootable,
|
|
#ifdef CONFIG_HOTPLUG_CPU
|
|
.cpu_disable = pnv_smp_cpu_disable,
|
|
.cpu_die = generic_cpu_die,
|
|
#endif /* CONFIG_HOTPLUG_CPU */
|
|
};
|
|
|
|
/* This is called very early during platform setup_arch */
|
|
void __init pnv_smp_init(void)
|
|
{
|
|
smp_ops = &pnv_smp_ops;
|
|
|
|
#ifdef CONFIG_HOTPLUG_CPU
|
|
ppc_md.cpu_die = pnv_smp_cpu_kill_self;
|
|
#endif
|
|
}
|