[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH 1/2] CPUArchState: introduce per-cpu lock
From: |
liu ping fan |
Subject: |
Re: [Qemu-devel] [PATCH 1/2] CPUArchState: introduce per-cpu lock |
Date: |
Sun, 24 Jun 2012 22:05:31 +0800 |
On Fri, Jun 22, 2012 at 8:36 PM, Stefan Hajnoczi <address@hidden> wrote:
> On Thu, Jun 21, 2012 at 4:06 PM, Liu Ping Fan <address@hidden> wrote:
>> diff --git a/cpu-defs.h b/cpu-defs.h
>> index f49e950..7305822 100644
>> --- a/cpu-defs.h
>> +++ b/cpu-defs.h
>> @@ -30,6 +30,7 @@
>> #include "osdep.h"
>> #include "qemu-queue.h"
>> #include "targphys.h"
>> +#include "qemu-thread-posix.h"
>
> This breaks Windows, you need qemu-thread.h.
>
>>
>> #ifndef TARGET_LONG_BITS
>> #error TARGET_LONG_BITS must be defined before including this header
>> @@ -220,6 +221,7 @@ typedef struct CPUWatchpoint {
>> CPU_COMMON_THREAD \
>> struct QemuCond *halt_cond; \
>> int thread_kicked; \
>> + struct QemuMutex *cpu_lock; \
>
> It would be nicer to declare it QemuMutex cpu_lock (no pointer) so
> that you don't need to worry about malloc/free.
>
Yes :), I have considered about it, and not quite sure. But I figure
out the lock for per-device to break BQL will be like this for some
reason.
After all, have not decided yet.
>> struct qemu_work_item *queued_work_first, *queued_work_last; \
>> const char *cpu_model_str; \
>> struct KVMState *kvm_state; \
>> diff --git a/cpus.c b/cpus.c
>> index b182b3d..554f7bc 100644
>> --- a/cpus.c
>> +++ b/cpus.c
>> @@ -735,6 +735,7 @@ static void *qemu_kvm_cpu_thread_fn(void *arg)
>> env->thread_id = qemu_get_thread_id();
>> cpu_single_env = env;
>>
>> +
>> r = kvm_init_vcpu(env);
>> if (r < 0) {
>> fprintf(stderr, "kvm_init_vcpu failed: %s\n", strerror(-r));
>
> Spurious whitespace change, this should be dropped from the patch.
>
>> diff --git a/main-loop.h b/main-loop.h
>> index dce1cd9..d8d44a4 100644
>> --- a/main-loop.h
>> +++ b/main-loop.h
>> @@ -323,6 +323,9 @@ void qemu_bh_delete(QEMUBH *bh);
>> int qemu_add_child_watch(pid_t pid);
>> #endif
>>
>> +void qemu_mutex_lock_cpu(void *_env);
>> +void qemu_mutex_unlock_cpu(void *_env);
>
> Why void* instead of CPUArchState*?
>
Because we can hide the CPUArchState from apic.c
Thanks and regards,
pingfan
> Stefan
[Qemu-devel] [PATCH 2/2] kvm: use per-cpu lock to free vcpu thread out of the big lock, Liu Ping Fan, 2012/06/21