xserver-xorg

xserver-xorg

[Marc Ferland]

Hi all,

I'm having trouble building the xserver-xorg package. The build fails
with:

| render2.c: In function '__glXDisp_Map1d':
| render2.c:104:5: error: the comparison will always evaluate as 'true' for the address of 'u1' will never be NULL [-Werror=address]
| render2.c:105:5: error: the comparison will always evaluate as 'true' for the address of 'u2' will never be NULL [-Werror=address]
| render2.c: In function '__glXDisp_Map2d':
| render2.c:147:5: error: the comparison will always evaluate as 'true' for the address of 'u1' will never be NULL [-Werror=address]
| render2.c:148:5: error: the comparison will always evaluate as 'true' for the address of 'u2' will never be NULL [-Werror=address]
| render2.c:149:5: error: the comparison will always evaluate as 'true' for the address of 'v1' will never be NULL [-Werror=address]
| render2.c:150:5: error: the comparison will always evaluate as 'true' for the address of 'v2' will never be NULL [-Werror=address]
| cc1: some warnings being treated as errors

The target machine is a x86_64. Looks like the compiler doesn't like how
the __GLX_GET_DOUBLE macro tests the address of stack variables against
NULL.

I would like to know if there is a patch upstream or if this is a known
issue? What surprises me the most is that this machine looks a lot like
the meta-sugarbay which builds correctly.

Regards,

Marc

Re: xserver-xorg

[Laurentiu Palcu]

[-- Attachment #1: Type: text/plain, Size: 1884 bytes --]

Hi Marc,

Apparently, these tests were removed upstream. As we are going to
upgrade xserver-xorg package to 1.12.3, those fixes will be included
too. However, if you need the fix now, I attached the patch.

Thanks,
Laurentiu

On 07/16/2012 08:56 PM, Marc Ferland wrote:
> Hi all,
> 
> I'm having trouble building the xserver-xorg package. The build fails
> with:
> 
> | render2.c: In function '__glXDisp_Map1d':
> | render2.c:104:5: error: the comparison will always evaluate as 'true' for the address of 'u1' will never be NULL [-Werror=address]
> | render2.c:105:5: error: the comparison will always evaluate as 'true' for the address of 'u2' will never be NULL [-Werror=address]
> | render2.c: In function '__glXDisp_Map2d':
> | render2.c:147:5: error: the comparison will always evaluate as 'true' for the address of 'u1' will never be NULL [-Werror=address]
> | render2.c:148:5: error: the comparison will always evaluate as 'true' for the address of 'u2' will never be NULL [-Werror=address]
> | render2.c:149:5: error: the comparison will always evaluate as 'true' for the address of 'v1' will never be NULL [-Werror=address]
> | render2.c:150:5: error: the comparison will always evaluate as 'true' for the address of 'v2' will never be NULL [-Werror=address]
> | cc1: some warnings being treated as errors
> 
> The target machine is a x86_64. Looks like the compiler doesn't like how
> the __GLX_GET_DOUBLE macro tests the address of stack variables against
> NULL.
> 
> I would like to know if there is a patch upstream or if this is a known
> issue? What surprises me the most is that this machine looks a lot like
> the meta-sugarbay which builds correctly.
> 
> Regards,
> 
> Marc
> _______________________________________________
> yocto mailing list
> yocto@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/yocto
> 


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-glx-warning-fixes.patch --]
[-- Type: text/x-patch; name="0001-glx-warning-fixes.patch", Size: 1875 bytes --]

From 63ccaec2c39f4b5742383472c951ee2cd35c9e14 Mon Sep 17 00:00:00 2001
From: Adam Jackson <ajax@redhat.com>
Date: Tue, 19 Oct 2010 11:59:23 -0400
Subject: [PATCH] glx: warning fixes
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

render2.c: In function ‘__glXDisp_Map2d’:
render2.c:127: warning: ‘u1’ may be used uninitialized in this function
render2.c: In function ‘__glXDisp_Map1d’:
render2.c:90: warning: ‘u1’ may be used uninitialized in this function

Remove unnecessary test, and change memcpy to memmove as all users were
doing overlapping copies.

Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
---
 glx/unpack.h             |    2 +-
 hw/dmx/glxProxy/unpack.h |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/glx/unpack.h b/glx/unpack.h
index a4e6d7e..738e79d 100644
--- a/glx/unpack.h
+++ b/glx/unpack.h
@@ -47,7 +47,7 @@
 ** Fetch a double from potentially unaligned memory.
 */
 #ifdef __GLX_ALIGN64
-#define __GLX_MEM_COPY(dst,src,n)	if (src != NULL && dst != NULL) memcpy(dst,src,n)
+#define __GLX_MEM_COPY(dst,src,n)	memmove(dst,src,n)
 #define __GLX_GET_DOUBLE(dst,src)	__GLX_MEM_COPY(&dst,src,8)
 #else
 #define __GLX_GET_DOUBLE(dst,src)	(dst) = *((GLdouble*)(src))
diff --git a/hw/dmx/glxProxy/unpack.h b/hw/dmx/glxProxy/unpack.h
index f34b6fd..f4a9572 100644
--- a/hw/dmx/glxProxy/unpack.h
+++ b/hw/dmx/glxProxy/unpack.h
@@ -43,7 +43,7 @@
 ** Fetch a double from potentially unaligned memory.
 */
 #ifdef __GLX_ALIGN64
-#define __GLX_MEM_COPY(dst,src,n)	memcpy(dst,src,n)
+#define __GLX_MEM_COPY(dst,src,n)	memmove(dst,src,n)
 #define __GLX_GET_DOUBLE(dst,src)	__GLX_MEM_COPY(&dst,src,8)
 #else
 #define __GLX_GET_DOUBLE(dst,src)	(dst) = *((GLdouble*)(src))
-- 
1.7.9.5

Re: xserver-xorg

[Marc Ferland]

Laurentiu Palcu <laurentiu.palcu@linux.intel.com> writes:

> Hi Marc,
>
> Apparently, these tests were removed upstream. As we are going to
> upgrade xserver-xorg package to 1.12.3, those fixes will be included
> too. However, if you need the fix now, I attached the patch.
>

Just in case someone else stumbles on this issue, setting the preferred
version of xserver-xorg to 1.11.2 in the machine.conf also solves the
issue.

PREFERRED_VERSION_xserver-xorg ?= "1.11.2"

Marc

Re: xserver-xorg

[Paul Eggleton]

On Tuesday 17 July 2012 09:22:43 Marc Ferland wrote:
> Laurentiu Palcu <laurentiu.palcu@linux.intel.com> writes:
> > Apparently, these tests were removed upstream. As we are going to
> > upgrade xserver-xorg package to 1.12.3, those fixes will be included
> > too. However, if you need the fix now, I attached the patch.
> 
> Just in case someone else stumbles on this issue, setting the preferred
> version of xserver-xorg to 1.11.2 in the machine.conf also solves the
> issue.
> 
> PREFERRED_VERSION_xserver-xorg ?= "1.11.2"

FWIW, setting this kind of thing in the machine config is discouraged. This is 
really distro policy which is something that belongs in your distro 
configuration (or local configuration if it's temporary).

Cheers,
Paul

-- 

Paul Eggleton
Intel Open Source Technology Centre