Re: [Qemu-devel] [PATCH 0/3] vnc: trivial problem and memory leak fix

[Gonglei]

On 2014/10/23 16:08, Gerd Hoffmann wrote:

> On Do, 2014-10-23 at 13:39 +0800, address@hidden wrote:
>> From: Gonglei <address@hidden>
>>
>> Beside those problems, I also found another issue, see below pls.
>>
>> Qemu command line:
>> $ ./qemu-system-x86_64 -enable-kvm -m 4096 -smp 4 -name redhat6.2 -drive 
>> file=/home/redhat.img,if=none,id=drive-ide0-0-0 -device 
>> ide-hd,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0,bootindex=1 -drive 
>> file=/mnt/sdb/gonglei/iso/rhel-server-7.0-x86_64-dvd.iso,if=none,id=drive-ide0-0-1
>>  -device 
>> ide-cd,bus=ide.0,unit=1,drive=drive-ide0-0-1,id=ide0-0-1,bootindex=4 -qmp 
>> unix:/tmp/qmp,server,nowait -monitor stdio -vnc 0.0.0.0:10,password,sasl
>> QEMU 2.1.50 monitor - type 'help' for more information
>> (qemu) change vnc password 123
>> (qemu) change vnc abc
>> error parsing address 'abc'     ---> an intended result because of I set a 
>> invalid value 'abc'
>> (qemu) change vnc password 456
>> If you want use passwords please enable password auth using '-vnc 
>> ${dpy},password'.Could not set password  --> Oops. I shouldn't get this 
>> output.
>> (qemu) 
>>
>>
>> Though we call qmp_change_vnc() failed, the content of global variable 
>> vnc_display
>> still will change, such as 'vs->auth = VNC_AUTH_NONE' now. I think this is 
>> not
>> reasonable, but I have not good idea to address this issue.
> 
> I think the current behavior is fine.  Better be on the safe side (no
> connects possible) in case "change vnc $display" failed.  You can fix
> that using "change vnc 0.0.0.0:10,password,sasl"
> 

Yes. we can do this. But is it reasonable that an API return false,
but its content changed?

Best regards,
-Gonglei