Re: [Qemu-devel] [PATCH for 2.5 1/1] qga: added another non-interactive gspawn() helper file.

["Denis V. Lunev" <den-lists@parallels.com>]

On 11/26/2015 08:31 PM, Yuri Pudgorodskiy wrote:
> On 11/26/2015 1:21 AM, Michael Roth wrote:
>> Quoting Denis V. Lunev (2015-11-25 13:02:26)
>>> From: Yuri Pudgorodskiy <yur@virtuozzo.com>
>>>
>>> With previous commit we added gspawn-win64-helper-console.exe,
>>> required for gspawn() mingw implementation.
>>> Unfortunatly when running as a service without interactive
>>> desktop, gspawn() also requires another helper app.
>>>
>>> Added gspawn-win64-helper.exe and gspawn-win32-helper.exe
>>> for corresponding architectures.
>>>
>>> Signed-off-by: Yuri Pudgorodskiy <yur@virtuozzo.com>
>>> Signed-off-by: Denis V. Lunev <den@openvz.org>
>>> CC: Michael Roth <mdroth@linux.vnet.ibm.com>
>> Thanks, applied to qga tree with minor whitespace fixup:
>>
>>    https://github.com/mdroth/qemu/commits/qga
>>
>> I noticed something testing this though: if we run qemu-ga
>> from a console, then exec something like ipconfig with
>> capture-output: true, qemu-ga returns that output via
>> guest-exec-status.
>>
>> If we run it as a service however, there's no output.
>>
>> # with qemu-ga started via console
>> {'execute':'guest-exec','arguments':{'path':'/Windows/System32/ipconfig.exe', 
>>
>> 'capture-output':true}}
>> {"return": {"pid": 1644}}
>> {'execute':'guest-exec-status','arguments':{'pid':1644}}
>> {"return": {"exitcode": 0, "out-data":
>> "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", 
>>
>> "exited": true}}
>>
>> # wtih qemu-ga started via windows service
>> {'execute':'guest-exec','arguments':{'path':'/Windows/System32/ipconfig.exe', 
>>
>> 'capture-output':true}}
>> {"return": {"pid": 1176}}
>> {'execute':'guest-exec-status','arguments':{'pid':1176}}
>> {"return": {"exitcode": 0, "exited": true}}
>>
>> Is this expected?
>>
>
> No, we want to fix it somehow - but not now,  because the whole 
> picture is not clear yet.
> Looks like when running gspawn() from a win32 service,  some win32 
> processes open
> its own console and write stdout to it instead of parent's fd 
> inherited from gspawn-helper.
>
> Not sure whether it can be corrected without patching gspawn() and 
> gspawn-helper code or not.
> If so, we may consider implementing platform-specific version of 
> guest-exec using win32 api.
>
> Right now the best what can be suggested, it is a workaround: usage of 
> redirection on guest side.
>
> Executing 3 commands in a row do a trick
>
> cmd /c ipconfig >out.txt
> cmd /c type out.txt
> cmd /c del out.txt
>
> but looks ugly and error-prone.
>
> One more pity notice, it seems that gspawn-helper does incorrect quoting:
>
> cmd /c echo "hello"
>
> produces \"hello\" output instead of expected "hello".
> Again, quoting issue may be another point for implementing better 
> win32-specific guest-exec.
>
> I'll share my visions on this problem later after some experiments.
>
>
no

for me this smells like GLIB bug which should
be addressed separately.

Den