Re: New kernel / new builderror

2009-05-26 Thread Dominique Leuenberger
Dmitry, Marcelo,

>>> On 5/26/2009 at  6:53 , Dmitry Torokhov  wrote:
> modules/linux/vmblock/linux/control.c-#if LINUX_VERSION_CODE < 
> KERNEL_VERSION(2, 6, 24)
> modules/linux/vmblock/linux/control.c:   entry->owner = THIS_MODULE;
> modules/linux/vmblock/linux/control.c-#endif
> 
> I just freshly unpacked the 05.22 tarball and modules compiled against
> 2.6.30-rc7 without any issues.

Sorry, you're absolutely right. All this failing actually had me look at the 
wrong build log for this issue. I have two project s with open-vm-tools,
one with the released version (was still 2009.03.23) and one rolling 
development, to be released. And the compilation of course failed in the older
one.

I'm so very sorry for this. 

Dominique

--
Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT 
is a gathering of tech-side developers & brand creativity professionals. Meet
the minds behind Google Creative Lab, Visual Complexity, Processing, & 
iPhoneDevCamp as they present alongside digital heavyweights like Barbarian 
Group, R/GA, & Big Spaceship. http://p.sf.net/sfu/creativitycat-com 
___
open-vm-tools-devel mailing list
open-vm-tools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open-vm-tools-devel


Re: New kernel / new builderror

2009-05-26 Thread Dmitry Torokhov
Hi Dominique,

On Monday 25 May 2009 07:54:36 Dominique Leuenberger wrote:
> Hi mates,
>
> Just one question:
> Is there already a patch for the open-vm-tools (2009.05.22) for Kernel
> 2.6.30 available?
>
> for now I see it failing with:
> sing 2.6.x kernel build system.
> CC [M]
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/block.o In
> file included from
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/vmblockInt.h:
>40, from
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/block.c:30:
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/include/vm_basic_ty
>pes.h:145:7: warning: "__FreeBSD__" is not defined CC [M]
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.o In
> file included from
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/vmblockInt.h:
>40, from
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c:33:
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/include/vm_basic_ty
>pes.h:145:7: warning: "__FreeBSD__" is not defined
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c: In
> function 'SetupProcDevice':
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c:159
>: error: 'struct proc_dir_entry' has no member named 'owner'
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c:171
>: error: 'struct proc_dir_entry' has no member named 'owner' make[3]: ***
> [/usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.o]
> Error 1 make[2]: ***
> [_module_/usr/src/packages/BUILD/obj/debug/modules/linux/vmblock] Error 2
> make[1]: *** [sub-make] Error 2
> make: *** [all] Error 2
> make: Leaving directory `/usr/src/linux-2.6.30-rc6-git3-3-obj/x86_64/debug'
> error: Bad exit status from /var/tmp/rpm-tmp.62886 (%build)
>

The following patch should help with 'undefined macro' errors.

Thanks,
Dmitry

--- vm_basic_types.h.orig
+++ vm_basic_types.h
@@ -113,12 +113,12 @@ typedef char int8;
 
 #else /* !HAVE_STDINT_H */
 
-#ifdef _MSC_VER
+#if defined(_MSC_VER)
 
 typedef unsigned __int64 uint64;
 typedef signed __int64 int64;
 
-#elif __GNUC__
+#elif defined(__GNUC__)
 /* The Xserver source compiles with -ansi -pendantic */
 #ifndef __STRICT_ANSI__
 #if defined(VM_X86_64)
@@ -128,7 +128,7 @@ typedef long int64;
 typedef unsigned long long uint64;
 typedef long long int64;
 #endif
-#elif __FreeBSD__
+#elif defined(__FreeBSD__)
 typedef unsigned long long uint64;
 typedef long long int64;
 #endif
@@ -261,7 +261,7 @@ typedef int64 VmTimeVirtualClock;  /* Vi
#define FMTSZ "z"
#define FMTPD "l"
#define FMTH ""
-#elif __GNUC__
+#elif defined(__GNUC__)
#define FMTH ""
#if defined(N_PLAT_NLM) || defined(sun) || \
(defined(__FreeBSD__) && (__FreeBSD__ + 0) && ((__FreeBSD__ + 0) < 5))
@@ -600,7 +600,8 @@ typedef void * UserVA;
  * are added the inline-ness should be removed.
  */
 
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 3)
+#if defined(__GNUC__) && \
+(__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 3))
 /*
  * Starting at version 3.3, gcc does not always inline functions marked
  * 'inline' (it depends on their size). To force gcc to do so, one must use 
the
@@ -633,7 +634,8 @@ typedef void * UserVA;
 
 #ifdef _MSC_VER
 #define NORETURN __declspec(noreturn)
-#elif __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 9)
+#elif defined(__GNUC__) && \
+  (__GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 9))
 #define NORETURN __attribute__((__noreturn__))
 #else
 #define NORETURN
@@ -644,7 +646,8 @@ typedef void * UserVA;
  * Older GCCs don't know about it --hpreg
  */
 
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 2)
+#if defined(__GNUC__) && \
+(__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 2))
 #   define VM_ASM_PLUS 1
 #else
 #   define VM_ASM_PLUS 0
@@ -663,7 +666,7 @@ typedef void * UserVA;
  * all others we don't so we do nothing.
  */
 
-#if (__GNUC__ >= 3)
+#if defined(__GNUC__) && (__GNUC__ >= 3)
 /*
  * gcc3 uses __builtin_expect() to inform the compiler of an expected value.
  * We use this to inform the static branch predictor. The '!!' in LIKELY




--
Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT 
is a gathering of tech-side developers & brand creativity professionals. Meet
the minds behind Google Creative Lab, Visual Complexity, Processing, & 
iPhoneDevCamp as they present alongside digital heavyweights like Barbarian 
Group, R/GA, & Big Spaceship. http://p.sf.net/sfu/creativitycat-com 
___
open-vm-tools-devel mailing list
open-vm-tools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open-vm-tools-devel


Re: New kernel / new builderror

2009-05-26 Thread Dmitry Torokhov
Dominique,

On Tuesday 26 May 2009 00:09:13 Dominique Leuenberger wrote:
> >>> On 5/26/2009 at 8:56, "Dominique Leuenberger"
>
>  Marcelo,
>
> > I think I posted at least three more lines of errors log, showing the
> > error:
> > /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/include/vm_basic_
> >typ es.h:145:7: warning: "__FreeBSD__" is not defined
> > /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c:
> > In function 'SetupProcDevice':
> > /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c:1
> >59: error: 'struct proc_dir_entry' has no member named 'owner'
> > /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c:1
> >71: error: 'struct proc_dir_entry' has no member named 'owner'
> > make[3]: ***
> > [/usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.o]
> > Error 1
> >
> > So the real error is in control.c, lines 159 and 171, where the structure
> > proc_dir_entry seems no longer to contain an 'owner'.
>
> Just to follow up on this:
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h
>=99b76233803beab302123d243eea9e41149804f3
>
> owner WAS removed from that struct.
>

Something must be wrong with the source tree you are using. We do handle
the removal of owner field form proc_dir_entry buy guarding against
certain kernel version (.24 is first released kernel when setting owner
was not really needed; .29 was the first kernel that actually broke
compilation):

modules/linux/vmblock/linux/control.c-#if LINUX_VERSION_CODE < 
KERNEL_VERSION(2, 6, 24)
modules/linux/vmblock/linux/control.c:   entry->owner = THIS_MODULE;
modules/linux/vmblock/linux/control.c-#endif

I just freshly unpacked the 05.22 tarball and modules compiled against
2.6.30-rc7 without any issues.

I will take a look at that warning in a minute.

-- 
Dmitry

--
Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT 
is a gathering of tech-side developers & brand creativity professionals. Meet
the minds behind Google Creative Lab, Visual Complexity, Processing, & 
iPhoneDevCamp as they present alongside digital heavyweights like Barbarian 
Group, R/GA, & Big Spaceship. http://p.sf.net/sfu/creativitycat-com 
___
open-vm-tools-devel mailing list
open-vm-tools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open-vm-tools-devel


Re: New kernel / new builderror

2009-05-26 Thread Dominique Leuenberger
>>> On 5/26/2009 at 8:56, "Dominique Leuenberger"

 Marcelo,
> I think I posted at least three more lines of errors log, showing the error:
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/include/vm_basic_typ
> es.h:145:7: warning: "__FreeBSD__" is not defined
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c: In 
> function 'SetupProcDevice':
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c:159: 
> error: 'struct proc_dir_entry' has no member named 'owner'
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c:171: 
> error: 'struct proc_dir_entry' has no member named 'owner'
> make[3]: *** 
> [/usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.o] 
> Error 1
> 
> So the real error is in control.c, lines 159 and 171, where the structure 
> proc_dir_entry seems no longer to contain an 'owner'.

Just to follow up on this:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=99b76233803beab302123d243eea9e41149804f3

owner WAS removed from that struct.


--
Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT
is a gathering of tech-side developers & brand creativity professionals. Meet
the minds behind Google Creative Lab, Visual Complexity, Processing, & 
iPhoneDevCamp asthey present alongside digital heavyweights like Barbarian
Group, R/GA, & Big Spaceship. http://www.creativitycat.com 
___
open-vm-tools-devel mailing list
open-vm-tools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open-vm-tools-devel


Re: New kernel / new builderror

2009-05-25 Thread Dominique Leuenberger
Marcelo,

>>> On 5/26/2009 at 1:09, Marcelo Vanzin  wrote: 
>> Just one question: 
>> Is there already a patch for the open-vm-tools (2009.05.22) for Kernel 
>> 2.6.30 
> available?
> 
> Did you try with any other kernel (just to make sure this is really 
> something 
> caused by changes in 2.6.30)?
> 

Yes, the package is at any time build for openSUSE 11.1 (based on 2.6.27) and 
openSUSE Factory (11.2 to be). It was 2.6.29 until just a day ago and
it built, the 2.6.30 landed and I got the build failures.

>> for now I see it failing with:
>> sing 2.6.x kernel build system.
>> CC [M] /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/block.o
>> In file included from 
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/vmblockInt.h:40
> ,
>> from 
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/block.c:30:
>> 
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/include/vm_basic_type
> s.h:145:7: warning: "__FreeBSD__" is not defined
> 
> This is weird. Could you share some more details about how you're staging 
> these 
> files? Line 145 in vm_basic_types.h in my git tree is a comment...


I think I posted at least three more lines of errors log, showing the error:
/usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/include/vm_basic_types.h:145:7:
 warning: "__FreeBSD__" is not defined
/usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c: In 
function 'SetupProcDevice':
/usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c:159: 
error: 'struct proc_dir_entry' has no member named 'owner'
/usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.c:171: 
error: 'struct proc_dir_entry' has no member named 'owner'
make[3]: *** 
[/usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/control.o] Error 
1

So the real error is in control.c, lines 159 and 171, where the structure 
proc_dir_entry seems no longer to contain an 'owner'.

Dominique

--
Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT
is a gathering of tech-side developers & brand creativity professionals. Meet
the minds behind Google Creative Lab, Visual Complexity, Processing, & 
iPhoneDevCamp asthey present alongside digital heavyweights like Barbarian
Group, R/GA, & Big Spaceship. http://www.creativitycat.com 
___
open-vm-tools-devel mailing list
open-vm-tools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open-vm-tools-devel


Re: New kernel / new builderror

2009-05-25 Thread Marcelo Vanzin
Hi Dominique,

Dominique Leuenberger wrote:
> Just one question: 
> Is there already a patch for the open-vm-tools (2009.05.22) for Kernel 2.6.30 
> available?

Did you try with any other kernel (just to make sure this is really something 
caused by changes in 2.6.30)?

> for now I see it failing with:
> sing 2.6.x kernel build system.
> CC [M] /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/block.o
> In file included from 
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/vmblockInt.h:40,
> from /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/linux/block.c:30:
> /usr/src/packages/BUILD/obj/debug/modules/linux/vmblock/include/vm_basic_types.h:145:7:
>  warning: "__FreeBSD__" is not defined

This is weird. Could you share some more details about how you're staging these 
files? Line 145 in vm_basic_types.h in my git tree is a comment...

-- 
- Marcelo

--
Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT
is a gathering of tech-side developers & brand creativity professionals. Meet
the minds behind Google Creative Lab, Visual Complexity, Processing, & 
iPhoneDevCamp asthey present alongside digital heavyweights like Barbarian
Group, R/GA, & Big Spaceship. http://www.creativitycat.com 
___
open-vm-tools-devel mailing list
open-vm-tools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open-vm-tools-devel