On 04/07/2014 01:51 PM, Raghavendra K T wrote:
On 04/07/2014 10:08 PM, Waiman Long wrote:
On 04/07/2014 02:14 AM, Raghavendra K T wrote:
[...]
But I am seeing hang in overcommit cases. Gdb showed that many vcpus
are halted and there was no progress. Suspecting the problem /race with
halting, I removed the halt() part of kvm_hibernate(). I am yet to
take a closer look at the code on halt() related changes.
It seems like there may still be race conditions where the current code
is not handling correctly. I will look into that to see where the
problem is. BTW, what test do you use to produce the hang condition?
Running ebizzy on 2 of the vms simultaneously (for sometime in
repeated loop) could reproduce it.
Yes, I am able to reproduce the hang problem with ebizzy. BTW, could you
try to apply the attached patch file on top of the v8 patch series to
see if it can fix the hang problem?
What is the baseline for the performance improvement? Is it without the
unfair lock and PV qspinlock?
Baseline was 3.14-rc8 without any of the qspin patch series.
Does the baseline have PV ticketlock or without any PV support?
-Longman
diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h
index a35cd02..825c535 100644
--- a/arch/x86/include/asm/paravirt.h
+++ b/arch/x86/include/asm/paravirt.h
@@ -717,9 +717,10 @@ static __always_inline void __queue_kick_cpu(int cpu)
PVOP_VCALL1(pv_lock_ops.kick_cpu, cpu);
}
-static __always_inline void __queue_hibernate(enum pv_lock_stats type)
+static __always_inline void
+__queue_hibernate(enum pv_lock_stats type, s8 *state, s8 sval)
{
- PVOP_VCALL1(pv_lock_ops.hibernate, type);
+ PVOP_VCALL3(pv_lock_ops.hibernate, type, state, sval);
}
static __always_inline void __queue_lockstat(enum pv_lock_stats type)
diff --git a/arch/x86/include/asm/paravirt_types.h b/arch/x86/include/asm/paravirt_types.h
index a8564b9..0e204dd 100644
--- a/arch/x86/include/asm/paravirt_types.h
+++ b/arch/x86/include/asm/paravirt_types.h
@@ -346,7 +346,7 @@ enum pv_lock_stats {
struct pv_lock_ops {
#ifdef CONFIG_QUEUE_SPINLOCK
void (*kick_cpu)(int cpu);
- void (*hibernate)(enum pv_lock_stats type);
+ void (*hibernate)(enum pv_lock_stats type, s8 *state, s8 sval);
void (*lockstat)(enum pv_lock_stats type);
#else
struct paravirt_callee_save lock_spinning;
diff --git a/arch/x86/include/asm/pvqspinlock.h b/arch/x86/include/asm/pvqspinlock.h
index a632dcb..4b33769 100644
--- a/arch/x86/include/asm/pvqspinlock.h
+++ b/arch/x86/include/asm/pvqspinlock.h
@@ -134,7 +134,7 @@ static __always_inline void pv_head_spin_check(struct pv_qvars *pv, int *count,
*count = 0;
return;
}
- __queue_hibernate(PV_HALT_QHEAD);
+ __queue_hibernate(PV_HALT_QHEAD, &pv->cpustate, PV_CPU_HALTED);
__queue_lockstat((pv->cpustate == PV_CPU_KICKED)
? PV_WAKE_KICKED : PV_WAKE_SPURIOUS);
ACCESS_ONCE(pv->cpustate) = PV_CPU_ACTIVE;
@@ -169,7 +169,7 @@ static __always_inline void pv_queue_spin_check(struct pv_qvars *pv, int *count)
* pv_next_node_check pv_queue_spin_check
* ------------------ -------------------
* [1] qhead = true [3] cpustate = PV_CPU_HALTED
- * barrier() barrier()
+ * smp_mb() smp_mb()
* [2] if (cpustate [4] if (qhead)
* == PV_CPU_HALTED)
*
@@ -178,9 +178,10 @@ static __always_inline void pv_queue_spin_check(struct pv_qvars *pv, int *count)
* _QLOCK_LOCKED_SLOWPATH may or may not be set
* 3,4,1,2 - the CPU is halt and _QLOCK_LOCKED_SLOWPATH is set
*/
- barrier();
+ smp_mb();
if (!ACCESS_ONCE(pv->qhead)) {
- __queue_hibernate(PV_HALT_QNODE);
+ __queue_hibernate(PV_HALT_QNODE, &pv->cpustate,
+ PV_CPU_HALTED);
__queue_lockstat((pv->cpustate == PV_CPU_KICKED)
? PV_WAKE_KICKED : PV_WAKE_SPURIOUS);
} else {
@@ -208,7 +209,7 @@ pv_next_node_check(struct pv_qvars *pv, struct qspinlock *lock)
/*
* Make sure qhead flag is visible before checking the cpustate flag
*/
- barrier();
+ smp_mb();
if (ACCESS_ONCE(pv->cpustate) == PV_CPU_HALTED)
ACCESS_ONCE(((union arch_qspinlock *)lock)->lock)
= _QLOCK_LOCKED_SLOWPATH;
diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
index 7d97e58..ce4b74b 100644
--- a/arch/x86/kernel/kvm.c
+++ b/arch/x86/kernel/kvm.c
@@ -880,23 +880,29 @@ static void kvm_kick_cpu_stats(int cpu)
/*
* Halt the current CPU & release it back to the host
*/
-static void kvm_hibernate(enum pv_lock_stats type)
+static void kvm_hibernate(enum pv_lock_stats type, s8 *state, s8 sval)
{
unsigned long flags;
if (in_nmi())
return;
- kvm_halt_stats(type);
/*
* Make sure an interrupt handler can't upset things in a
* partially setup state.
*/
local_irq_save(flags);
+ /*
+ * Don't halt if the CPU state has been changed.
+ */
+ if (ACCESS_ONCE(*state) != sval)
+ goto out;
+ kvm_halt_stats(type);
if (arch_irqs_disabled_flags(flags))
halt();
else
safe_halt();
+out:
local_irq_restore(flags);
}
#endif /* !CONFIG_QUEUE_SPINLOCK */
diff --git a/arch/x86/xen/spinlock.c b/arch/x86/xen/spinlock.c
index 6bbe798..c597bbd 100644
--- a/arch/x86/xen/spinlock.c
+++ b/arch/x86/xen/spinlock.c
@@ -271,7 +271,7 @@ static void xen_kick_cpu(int cpu)
/*
* Halt the current CPU & release it back to the host
*/
-static void xen_hibernate(enum pv_lock_stats type)
+static void xen_hibernate(enum pv_lock_stats type, s8 *state, s8 sval)
{
int irq = __this_cpu_read(lock_kicker_irq);
unsigned long flags;
@@ -292,7 +292,11 @@ static void xen_hibernate(enum pv_lock_stats type)
/* Allow interrupts while blocked */
local_irq_restore(flags);
-
+ /*
+ * Don't halt if the CPU state has been changed.
+ */
+ if (ACCESS_ONCE(*state) != sval)
+ return;
/*
* If an interrupt happens here, it will leave the wakeup irq
* pending, which will cause xen_poll_irq() to return
