[vbox-dev] Announcement: VirtualBox 5.1.24 released

Larry Finger Larry.Finger at lwfinger.net
Fri Jul 21 16:12:08 GMT 2017


On 07/21/2017 05:54 AM, Sérgio Basto wrote:
> On Thu, 2017-07-20 at 19:59 -0500, Larry Finger wrote:
>> On 07/18/2017 06:30 AM, Klaus Espenlaub wrote:
>>> Hi,
>>>
>>> today Oracle released VirtualBox 5.1.24, a maintenance release of
>>> VirtualBox 5.1 which improves stability and fixes regressions.
>>>
>>> See the Changelog at
>>>
>>>     https://www.virtualbox.org/wiki/Changelog
>>>
>>> for a complete list of all changes.
>>>
>>> You can download the binaries here:
>>>
>>>     https://www.virtualbox.org/wiki/Downloads
>>
>> For gcc7 builds for i586, I find the source needs an additional
>> patch:
>>
>> Index:
>> VirtualBox-
>> 5.1.22/src/VBox/Devices/PC/ipxe/src/arch/i386/interface/pxe/pxe_prebo
>> ot.c
>> ===================================================================
>> ---
>> VirtualBox-
>> 5.1.22.orig/src/VBox/Devices/PC/ipxe/src/arch/i386/interface/pxe/pxe_
>> preboot.c
>> +++
>> VirtualBox-
>> 5.1.22/src/VBox/Devices/PC/ipxe/src/arch/i386/interface/pxe/pxe_prebo
>> ot.c
>> @@ -262,6 +262,7 @@ pxenv_restart_tftp ( struct s_PXENV_TFTP
>>
>>           /* Restart NBP */
>>           rmlongjmp ( pxe_restart_nbp, PXENV_RESTART_TFTP );
>> +       return 0;
>>    }
>>
>>    /* PXENV_START_UNDI
>>
>>
>> Without this patch gcc7 errors with a "non-void function returns a
>> random
>> value". Obviously, the compiler does not know how to process that
>> rmlongjmp()
>> call; however, adding that extra line has little effect on the object
>> code.
>>
>> Larry
> 
> I also got issues with latest gcc , glibc etc .
> 
> http://koji.rpmfusion.org/kojifiles/work/tasks/6178/146178/build.log
> 
> http://koji.rpmfusion.org/koji/taskinfo?taskID=146178
> 
> In file included from /builddir/build/BUILD/VirtualBox-
> 5.1.24/src/VBox/ExtPacks/VBoxDTrace/onnv/uts/common/sys/dtrace.h:55:0,
>                   from /builddir/build/BUILD/VirtualBox-
> 5.1.24/src/VBox/ExtPacks/VBoxDTrace/onnv/lib/libdtrace/common/dtrace.h:
> 38,
>                   from /builddir/build/BUILD/VirtualBox-
> 5.1.24/src/VBox/ExtPacks/VBoxDTrace/onnv/cmd/dtrace/dtrace.c:37:
> /builddir/build/BUILD/VirtualBox-
> 5.1.24/src/VBox/ExtPacks/VBoxDTrace/include/VBoxDTraceTypes.h:59:37:
> error: conflicting types for 'greg_t'
>   typedef RTCCINTREG                  greg_t;
>                                       ^~~~~~
> In file included from /usr/include/signal.h:302:0,
>                   from /usr/include/sys/wait.h:36,
>                   from /builddir/build/BUILD/VirtualBox-
> 5.1.24/src/VBox/ExtPacks/VBoxDTrace/onnv/cmd/dtrace/dtrace.c:34:
> /usr/include/sys/ucontext.h:37:37: note: previous declaration of
> 'greg_t' was here
>   __extension__ typedef long long int greg_t;
>                                       ^~~~~~
> kmk: *** [/builddir/build/BUILD/VirtualBox-
> 5.1.24/obj/obj/VBoxDTraceCmd/src/VBox/ExtPacks/VBoxDTrace/onnv/cmd/dtra
> ce/dtrace.o] Error 1
> 
> 
> Any feedback is welcome

I have not seen that problem. Obviously your build system differs.

You might try the statement "#undef greg_t" just before the typedef for greg_t 
in src/VBox/ExtPacks/VBoxDTrace/include/VBoxDTraceTypes.h.

Interesting that the file contains the statement
"#  include <sys/ucontext.h> /* avoid greg_t trouble */". It seems that it is 
not working.

Larry

Larry





More information about the vbox-dev mailing list