On 26/06/2014 10:17, lig.f...@cn.fujitsu.com wrote:
>> -----邮件原件-----
>> 发件人: Frederic Konrad [mailto:fred.kon...@greensocs.com]
>> 发送时间: 2014年6月26日 15:28
>> 收件人: Li, Guang/李 广; qemu-devel@nongnu.org
>> 抄送: peter.mayd...@linaro.org; quint...@redhat.com;
>> mark.bur...@greensocs.com; dgilb...@redhat.com; amit.s...@redhat.com;
>> pbonz...@redhat.com; vilan...@ac.upc.edu
>> 主题: Re: 答复: [Qemu-devel] [RFC PATCH v5 09/13] introduce reverse
>> execution mechanism.
>>
>> On 26/06/2014 04:31, lig.f...@cn.fujitsu.com wrote:
>>> Hi,
>> Hi,
>>
>>>> -----邮件原件-----
>>>> 发件人: qemu-devel-bounces+lig.fnst=cn.fujitsu....@nongnu.org
>>>> [mailto:qemu-devel-bounces+lig.fnst=cn.fujitsu....@nongnu.org] 代表
>>>> fred.kon...@greensocs.com
>>>> 发送时间: 2014年6月25日 23:20
>>>> 收件人: qemu-devel@nongnu.org
>>>> 抄送: peter.mayd...@linaro.org; quint...@redhat.com;
>>>> mark.bur...@greensocs.com; dgilb...@redhat.com;
>> amit.s...@redhat.com;
>>>> pbonz...@redhat.com; vilan...@ac.upc.edu; fred.kon...@greensocs.com
>>>> 主题: [Qemu-devel] [RFC PATCH v5 09/13] introduce reverse execution
>>>> mechanism.
>>>>
>>>> From: KONRAD Frederic <fred.kon...@greensocs.com>
>>>>
>>>> This introduces the basic reverse-execution mechanism.
>>>>
>>>> Signed-off-by: KONRAD Frederic <fred.kon...@greensocs.com>
>>>> ---
>>>> Makefile.target | 1 +
>>>> cpus.c | 6 +
>>>> include/reverse-execution.h | 41 ++++++
>>>> reverse-execution.c | 306
>>>> ++++++++++++++++++++++++++++++++++++++++++++
>>>> vl.c | 7 +-
>>>> 5 files changed, 360 insertions(+), 1 deletion(-) create mode 100644
>>>> include/reverse-execution.h create mode 100644 reverse-execution.c
>> [...]
>>>> +
>>>> +static QLIST_HEAD(, snapshot_entry) snapshot =
>>>> +QLIST_HEAD_INITIALIZER(snapshot);
>>>> +
>>>> +QEMUTimer *snap_timer;
>>> s/snap_/snapshot_/g ?
>>>
>>>> +QEMUTimer *stop_timer;
>>>> +
>>>> +struct cexe_state {
>>>> + int stepping_back;
>>>> + int continue_backward_mode;
>>>> + int singlestep_was_enabled;
>>> boolize these 3 vars?
>>>
>>>> + bool step_done;
>> Thanks for having looking into this,
>> I'll do these change.
>>
>>>> + saved_vm_running = runstate_is_running();
>>>> + vm_stop(RUN_STATE_SAVE_VM);
>>>> + snap->time = qemu_clock_get_ns(QEMU_CLOCK_ICOUNT);
>>>> + sprintf(filename, ".save%04u", snap->id);
>>> File name should be more readable
>> What do you mean by more readable?
> e.g.
> gdb-reverse-step-saved-01
> gdb-reverse-step-saved-02
> ...
Ok understood, it makes sense.
Thanks,
Fred
>>>> +
>>>> + trace_snap_callback(snap->time, filename);
>>>> +
>>>> + file = qemu_fopen(filename, "wb");
>>>> + qemu_savevm_state(file);
>>>> + qemu_fclose(file);
>>>> +
>>>> + if (saved_vm_running) {
>>>> + vm_start();
>>>> + }
>>>> + timer_mod_ns(snap_timer, snap->time + 100000000);
>>>> + }
>>>> +}
>>>> +
>>>> +/*
>>>> + * Timer callback called when the VM have to stop.
>>>> + */
>>>> +static void stop_callback(void *opaque) {
>>>> + trace_stop_callback(qemu_clock_get_ns(QEMU_CLOCK_ICOUNT));
>>>> + cexe_state.stop_requested = true;
>>>> +}
>>>> +
>>>> +void cexe_setup(void)
>>> Sorry, but, what's the mean of cexe here and there?
>> This means "exec" in reverse..
> good!
> But seems a little hard to get it.
>
> Thanks!
>
>
