[Qemu-devel] Re: [PATCH 5/9] Monitor: Return before exiting with 'quit'

[Jan Kiszka <jan.kiszka@web.de>]

[-- Attachment #1: Type: text/plain, Size: 3734 bytes --]

Anthony Liguori wrote:
> On 04/26/2010 01:53 PM, Luiz Capitulino wrote:
>> On Mon, 26 Apr 2010 13:25:38 -0500
>> Anthony Liguori<anthony@codemonkey.ws>  wrote:
>>
>>   
>>> On 04/26/2010 01:22 PM, Luiz Capitulino wrote:
>>>     
>>>> On Mon, 26 Apr 2010 12:49:40 -0500
>>>> Anthony Liguori<anthony@codemonkey.ws>   wrote:
>>>>
>>>>
>>>>       
>>>>> On 04/26/2010 10:47 AM, Luiz Capitulino wrote:
>>>>>
>>>>>         
>>>>>> The 'quit' Monitor command (implemented by do_quit()) calls
>>>>>> exit() directly, this is problematic under QMP because QEMU
>>>>>> exits before having a chance to send the ok response.
>>>>>>
>>>>>> Clients don't know if QEMU exited because of a problem or
>>>>>> because the 'quit' command has been executed.
>>>>>>
>>>>>> This commit fixes that by moving the exit() call to the main
>>>>>> loop, so that do_quit() requests the system to quit, instead
>>>>>> of calling exit() directly.
>>>>>>
>>>>>>
>>>>>>            
>>>>> Does this also have the effect of printing out a (qemu) prompt after
>>>>> quit before an EOF appears on that socket?
>>>>>
>>>>>          
>>>>    Ah, right..
>>>>
>>>>        
>>> It's not necessarily a bad thing if it does.  I just wanted to raise
>>> that because it's possible that someone depends on the behavior.
>>>      
>>   Yes, and it's also a bit ugly if you do '-monitor stdio':
>>
>> ~/stuff/virt/ ./qemu-qmp -monitor stdio
>> QEMU 0.12.50 monitor - type 'help' for more information
>> (qemu) quit
>> (qemu) ~/stuff/virt/
>>
>>   So, I'd like to fix it.
>>
>>   
>>> I'm not sure it matters to me if we change this behavior though.
>>>      
>>   It's very easy to fix, the following patch does it.
>>
>>   I have (tested and) rebased the for-anthony branch with it.
>>
>>  From 14ecef47d1989f9b646fc0fe7a4bf42c091d1432 Mon Sep 17 00:00:00 2001
>> From: Luiz Capitulino<lcapitulino@redhat.com>
>> Date: Tue, 6 Apr 2010 18:55:54 -0300
>> Subject: [PATCH 5/9] Monitor: Return before exiting with 'quit'
>>
>> The 'quit' Monitor command (implemented by do_quit()) calls
>> exit() directly, this is problematic under QMP because QEMU
>> exits before having a chance to send the ok response.
>>
>> Clients don't know if QEMU exited because of a problem or
>> because the 'quit' command has been executed.
>>
>> This commit fixes that by moving the exit() call to the main
>> loop, so that do_quit() requests the system to quit, instead
>> of calling exit() directly.
>>
>> Signed-off-by: Luiz Capitulino<lcapitulino@redhat.com>
>> ---
>>   monitor.c |    7 ++++++-
>>   sysemu.h  |    2 ++
>>   vl.c      |   18 ++++++++++++++++++
>>   3 files changed, 26 insertions(+), 1 deletions(-)
>>
>> diff --git a/monitor.c b/monitor.c
>> index ef84298..5b0258f 100644
>> --- a/monitor.c
>> +++ b/monitor.c
>> @@ -1017,7 +1017,12 @@ static void do_info_cpu_stats(Monitor *mon)
>>    */
>>   static int do_quit(Monitor *mon, const QDict *qdict, QObject
>> **ret_data)
>>   {
>> -    exit(0);
>> +    if (monitor_ctrl_mode(mon)) {
>> +        qemu_system_exit_request();
>> +    } else {
>> +        exit(0);
>> +    }
>> +
>>    
> 
> Now they have different behaviors though because
> qemu_system_exit_request() actually exits gracefully.
> 
> For instance, now in QMP mode, an ifdown script will be executed when
> the 'quit' command is invoked whereas it won't be executed when done
> through the human monitor.
> 
> I honestly don't mind the weird print outs with -monitor stdio.  I think
> exiting gracefully in both cases is better.

Just suspend the monitor before leaving for termination. That should
have both the proper visual and functional effect.

Jan



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 257 bytes --]