Re: [Xen-devel] Errors in compilation // xl_cmdimpl.c:2733:11 ...

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

On 24/08/12 20:12, Ian Campbell wrote:
> On Fri, 2012-08-24 at 19:42 +0100, Andrew Cooper wrote:
>> On 24/08/12 17:43, p.d@xxxxxx wrote:
>>> nice time,
>>>
>>> Ian, I'm not sure, but I think after Your patch:
>>> http://xenbits.xen.org/hg/xen-unstable.hg/rev/4ca40e0559c3
>>>
>>> xen-tools (+qemu+seabios) will not be maked :)
>>>
>>> Here are last lines of "make -j7":
>>> ==============================================================
>>> gcc  -O1 -fno-omit-frame-pointer -m64 -g -fno-strict-aliasing -std=gnu99 
>>> -Wall -Wstrict-prototypes -Wdeclaration-after-statement 
>>> -Wno-unused-but-set-variable   -D__XEN_TOOLS__ -MMD -MF 
>>> ._libxl_save_msgs_callout.o.d  -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE 
>>> -fno-optimize-sibling-calls  -Werror -Wno-format-zero-length 
>>> -Wmissing-declarations -Wno-declaration-after-statement -Wformat-nonliteral 
>>> -I. -fPIC -pthread 
>>> -I/usr/src/xen_2012_08_24_rev_25779/tools/libxl/../../tools/libxc 
>>> -I/usr/src/xen_2012_08_24_rev_25779/tools/libxl/../../tools/include 
>>> -I/usr/src/xen_2012_08_24_rev_25779/tools/libxl/../../tools/libxc 
>>> -I/usr/src/xen_2012_08_24_rev_25779/tools/libxl/../../tools/include 
>>> -I/usr/src/xen_2012_08_24_rev_25779/tools/libxl/../../tools/xenstore 
>>> -I/usr/src/xen_2012_08_24_rev_25779/tools/libxl/../../tools/include 
>>> -I/usr/src/xen_2012_08_24_rev_25779/tools/libxl/../../tools/blktap2/control 
>>> -I/usr/src/xen_2012_08_24_rev_25779/tools/libxl/../../tools/blktap2/include 
>>> -I/usr/src/xen_2012_08_24_rev_25779/tools/libxl/../../tools/include 
>>> -include /usr/src/xen_2012_08_24_rev_25779/tools/libxl/../../tools/config.h 
>>>  -c -o _libxl_save_msgs_callout.o _libxl_save_msgs_callout.c 
>>> xl_cmdimpl.c: In function âmain_listâ:
>>> xl_cmdimpl.c:2733:11: error: âhandâ may be used uninitialized in this 
>>> function [-Werror=uninitialized]
>>> xl_cmdimpl.c:2689:14: note: âhandâ was declared here
>>> <snip>
>> Please try the attached patch.  I have fixed the error, and also
>> future-proofed the logic.
>>
>> @Ian: the patch can be slimed down if default_output_format can be
>> guaranteed not to change across the duration of this function call, but
>> my cursory glance at this otherwise-unfamilar codebase cant say for certain.
> It can only change during argument parsing.
>
> gcc is being a bit dumb here since default_output_format is a statically
> initialised enum whose only values are OUTPUT_FORMAT_JSON and
> OUTPUT_FORMAT_SXP.
>
> I suspect that now you have re-written it so that hand is only touched
> iff format == JSON and format is const that initialisation of hand is no
> longer necessary. But nonetheless:
> Acked-by: Ian Campbell <ian.campbell@xxxxxxxxxx>
>
> Ian.
>

Sadly, hand being touched iff format == JSON was still not enough to
satisfy my version of GCC.

Attached is a far slimmed version which explicitly sets hand to NULL at
the top, and future-proofs the use of hand in the middle of the domain loop.

-- 
Andrew Cooper - Dom0 Kernel Engineer, Citrix XenServer
T: +44 (0)1223 225 900, http://www.citrix.com

Attachment: fix-cmdimpl-v2.patch
Description: Text Data

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel