mesa: Changes to 'ubuntu'

2014-09-18 Thread Maarten Bernardus Lankhorst
 debian/changelog |6 ++
 debian/control   |3 +--
 debian/rules |2 +-
 3 files changed, 8 insertions(+), 3 deletions(-)

New commits:
commit d77d8908a3dd52b156ada75460a2c37f2ec1b4ad
Author: Maarten Lankhorst 
Date:   Fri Aug 22 15:04:19 2014 +0200

And revert again for now.. (LP: #1360241)

diff --git a/debian/changelog b/debian/changelog
index 58777f3..db40d27 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,9 @@
+mesa (10.2.6-1ubuntu3) utopic; urgency=medium
+
+  * And revert again for now.. (LP: #1360241)
+
+ -- Maarten Lankhorst   Fri, 22 Aug 2014 
15:02:30 +0200
+
 mesa (10.2.6-1ubuntu2) utopic; urgency=medium
 
   * Fixup typo.
diff --git a/debian/control b/debian/control
index 0b0191f..0da13b6 100644
--- a/debian/control
+++ b/debian/control
@@ -38,7 +38,7 @@ Build-Depends:
  libudev-dev [linux-any],
  flex,
  bison,
- llvm-3.5-dev (>= 1:3.5~+rc3) [amd64 i386 kfreebsd-amd64 kfreebsd-i386 armhf],
+ llvm-3.4-dev [amd64 i386 kfreebsd-amd64 kfreebsd-i386 armhf],
  libelf-dev [amd64 i386 kfreebsd-amd64 kfreebsd-i386 armhf],
  libwayland-dev (>= 1.2.0) [linux-any],
  libmirclient-dev [!arm64 !powerpc !ppc64 !ppc64el],
@@ -552,7 +552,6 @@ Priority: optional
 Architecture: any
 Pre-Depends: ${misc:Pre-Depends}
 Depends:
- libllvm3.5 (>= 1:3.5~+rc3) [amd64 i386],
  ${shlibs:Depends},
  ${misc:Depends}
 Recommends: libtxc-dxtn-s2tc0 | libtxc-dxtn0
diff --git a/debian/rules b/debian/rules
index 3c10f52..c4bd048 100755
--- a/debian/rules
+++ b/debian/rules
@@ -110,7 +110,7 @@ else
   ifneq (,$(filter $(DEB_HOST_ARCH),amd64 i386 kfreebsd-amd64 kfreebsd-i386 
armhf))
GALLIUM_DRIVERS += radeonsi
confflags_GALLIUM += --enable-gallium-llvm
-   confflags_GALLIUM += ac_cv_path_LLVM_CONFIG=llvm-config-3.5
+   confflags_GALLIUM += ac_cv_path_LLVM_CONFIG=llvm-config-3.4
   endif
 
confflags_DIRECT_RENDERING = --enable-driglx-direct


-- 
To UNSUBSCRIBE, email to debian-x-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/e1xuynb-0008h4...@moszumanska.debian.org



mesa: Changes to 'ubuntu+1'

2014-09-18 Thread Maarten Bernardus Lankhorst
 debian/changelog|1 
 debian/patches/fix-altivec-intrinsics.patch |  111 
 debian/patches/series   |1 
 3 files changed, 113 insertions(+)

New commits:
commit 3d2d673b8719363a982a30185d1083c97000f6f9
Author: Maarten Lankhorst 
Date:   Thu Sep 18 08:32:58 2014 +0200

Import upstream fix for altivec little endian instructions.

diff --git a/debian/changelog b/debian/changelog
index 09fb41a..676e8a3 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,6 +1,7 @@
 mesa (10.3.0~rc3-2ubuntu1) UNRELEASED; urgency=low
 
   * Merge from released debian-experimental.
+  * Import upstream fix for altivec little endian instructions.
 
  -- Maarten Lankhorst   Tue, 12 Aug 2014 
09:38:49 +0200
 
diff --git a/debian/patches/fix-altivec-intrinsics.patch 
b/debian/patches/fix-altivec-intrinsics.patch
new file mode 100644
index 000..c67784a
--- /dev/null
+++ b/debian/patches/fix-altivec-intrinsics.patch
@@ -0,0 +1,111 @@
+commit 0feb977bbfb0d6bb2c8d3178246acb035a739f37
+Author: Ulrich Weigand 
+Date:   Mon Aug 4 18:41:00 2014 +0200
+
+gallivm: Fix Altivec pack intrinsics for little-endian
+
+This patch fixes use of Altivec pack intrinsics on little-endian PowerPC
+systems.  Since little-endian operation only affects the load and store
+instructions, the semantics of pack (and other) instructions that take
+two input vectors implicitly change: the pack instructions still fill
+a register placing values from the first operand into the "high" parts
+of the register, and values from the second operand into the "low" parts
+of the register, but since vector loads and stores perform an endian swap,
+the high parts end up at high memory addresses.
+
+To still achieve the desired effect, we have to swap the two inputs to
+the pack instruction on little-endian systems.  This is done automatically
+by the back-end for instructions generated by LLVM, but needs to be done
+manually when emitting intrisincs (which still result in that instruction
+being emitted directly).
+
+Signed-off-by: Ulrich Weigand 
+Signed-off-by: Maarten Lankhorst 
+
+diff --git a/src/gallium/auxiliary/gallivm/lp_bld_pack.c 
b/src/gallium/auxiliary/gallivm/lp_bld_pack.c
+index a48a922..cdf6d80 100644
+--- a/src/gallium/auxiliary/gallivm/lp_bld_pack.c
 b/src/gallium/auxiliary/gallivm/lp_bld_pack.c
+@@ -464,6 +464,7 @@ lp_build_pack2(struct gallivm_state *gallivm,
+if((util_cpu_caps.has_sse2 || util_cpu_caps.has_altivec) &&
+src_type.width * src_type.length >= 128) {
+   const char *intrinsic = NULL;
++  boolean swap_intrinsic_operands = FALSE;
+ 
+   switch(src_type.width) {
+   case 32:
+@@ -482,6 +483,9 @@ lp_build_pack2(struct gallivm_state *gallivm,
+} else {
+   intrinsic = "llvm.ppc.altivec.vpkuwus";
+}
++#ifdef PIPE_ARCH_LITTLE_ENDIAN
++   swap_intrinsic_operands = TRUE;
++#endif
+  }
+  break;
+   case 16:
+@@ -490,12 +494,18 @@ lp_build_pack2(struct gallivm_state *gallivm,
+   intrinsic = "llvm.x86.sse2.packsswb.128";
+ } else if (util_cpu_caps.has_altivec) {
+   intrinsic = "llvm.ppc.altivec.vpkshss";
++#ifdef PIPE_ARCH_LITTLE_ENDIAN
++  swap_intrinsic_operands = TRUE;
++#endif
+ }
+  } else {
+ if (util_cpu_caps.has_sse2) {
+   intrinsic = "llvm.x86.sse2.packuswb.128";
+ } else if (util_cpu_caps.has_altivec) {
+ intrinsic = "llvm.ppc.altivec.vpkshus";
++#ifdef PIPE_ARCH_LITTLE_ENDIAN
++  swap_intrinsic_operands = TRUE;
++#endif
+ }
+  }
+  break;
+@@ -504,7 +514,11 @@ lp_build_pack2(struct gallivm_state *gallivm,
+   if (intrinsic) {
+  if (src_type.width * src_type.length == 128) {
+ LLVMTypeRef intr_vec_type = lp_build_vec_type(gallivm, intr_type);
+-res = lp_build_intrinsic_binary(builder, intrinsic, 
intr_vec_type, lo, hi);
++if (swap_intrinsic_operands) {
++   res = lp_build_intrinsic_binary(builder, intrinsic, 
intr_vec_type, hi, lo);
++} else {
++   res = lp_build_intrinsic_binary(builder, intrinsic, 
intr_vec_type, lo, hi);
++}
+ if (dst_vec_type != intr_vec_type) {
+res = LLVMBuildBitCast(builder, res, dst_vec_type, "");
+ }
+@@ -513,6 +527,8 @@ lp_build_pack2(struct gallivm_state *gallivm,
+ int num_split = src_type.width * src_type.length / 128;
+ int i;
+ int nlen = 128 / src_type.width;
++int lo_off = swap_intrinsic_operands ? nlen : 0;
++int hi_off = swap_intrinsic_operands ? 0 : nlen;
+ struct lp_type ndst_type = lp_type_unorm(dst_type.width, 128);
+ struct lp_type nintr_type = lp_type_unorm(intr_type.wi

mesa: Changes to 'ubuntu+1'

2014-09-18 Thread Maarten Bernardus Lankhorst
 debian/changelog |   33 ++-
 debian/compat|2 -
 debian/control   |   13 +-
 debian/libegl1-mesa-drivers.install.linux.in |2 -
 debian/rules |2 -
 5 files changed, 45 insertions(+), 7 deletions(-)

New commits:
commit 6620b4abe0f772b97b697549b231671c4adfd30d
Author: Maarten Lankhorst 
Date:   Thu Sep 18 11:59:11 2014 +0200

New upstream release candidate. (LP: #1364003)

diff --git a/debian/changelog b/debian/changelog
index c7d6c72..fc28472 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,5 +1,6 @@
 mesa (10.3.0~rc3-3ubuntu1) UNRELEASED; urgency=low
 
+  * New upstream release candidate. (LP: #1364003)
   * Merge from released debian-experimental.
   * Import upstream fix for altivec little endian instructions.
 

commit a2bab76b4a260fa536d2de5619b2f74c9ec7f6da
Author: Julien Cristau 
Date:   Mon Sep 15 22:39:36 2014 +0200

Upload to experimental

diff --git a/debian/changelog b/debian/changelog
index fd75f04..c159591 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,4 +1,4 @@
-mesa (10.3.0~rc3-3) UNRELEASED; urgency=medium
+mesa (10.3.0~rc3-3) experimental; urgency=medium
 
   [ Andreas Boll ]
   * Bump llvm-3.5-dev and libclang-3.5-dev to 1:3.5-1, should fix FTBFS on
@@ -11,7 +11,10 @@ mesa (10.3.0~rc3-3) UNRELEASED; urgency=medium
   [ Dima Kogan ]
   * All -dev packages are now Multi-Arch: same (Closes: #689088, #678040).
 
- -- Andreas Boll   Mon, 15 Sep 2014 14:19:30 +0200
+  [ Julien Cristau ]
+  * Upload clean source without partial Ubuntu alternatives patches (closes: 
#761678)
+
+ -- Julien Cristau   Mon, 15 Sep 2014 22:39:15 +0200
 
 mesa (10.3.0~rc3-2) experimental; urgency=medium
 

commit 3803b01079d10f7ff3370e5566ba0502d96f676f
Author: Andreas Boll 
Date:   Mon Sep 15 20:17:03 2014 +0200

Drop redundant libegl1-mesa-drivers.install.linux.in

It's the same as libegl1-mesa-drivers.install.in

diff --git a/debian/changelog b/debian/changelog
index 5877a8b..fd75f04 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -6,6 +6,7 @@ mesa (10.3.0~rc3-3) UNRELEASED; urgency=medium
   * Bump compat to v9 for changed dh_strip behavior:
 - dh_strip puts separated debug symbols in a location based on their
   build-id (Closes: #755921).
+  * Drop redundant libegl1-mesa-drivers.install.linux.in.
 
   [ Dima Kogan ]
   * All -dev packages are now Multi-Arch: same (Closes: #689088, #678040).
diff --git a/debian/libegl1-mesa-drivers.install.linux.in 
b/debian/libegl1-mesa-drivers.install.linux.in
deleted file mode 100644
index 741f962..000
--- a/debian/libegl1-mesa-drivers.install.linux.in
+++ /dev/null
@@ -1,2 +0,0 @@
-# OS-independent part (from libegl1-mesa-drivers.install.in):
-dri/usr/lib/${DEB_HOST_MULTIARCH}/egl/egl_gallium.so 
usr/lib/${DEB_HOST_MULTIARCH}/egl

commit 4e8ef10eb3d85c3b79991c2189d4e7f4745e5d20
Author: Dima Kogan 
Date:   Sun Feb 16 01:31:01 2014 -0800

All -dev packages are now Multi-Arch: same

Closes #689088, #678040

diff --git a/debian/changelog b/debian/changelog
index 0c9bfb3..5877a8b 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,11 +1,15 @@
 mesa (10.3.0~rc3-3) UNRELEASED; urgency=medium
 
+  [ Andreas Boll ]
   * Bump llvm-3.5-dev and libclang-3.5-dev to 1:3.5-1, should fix FTBFS on
 kfreebsd-*.
   * Bump compat to v9 for changed dh_strip behavior:
 - dh_strip puts separated debug symbols in a location based on their
   build-id (Closes: #755921).
 
+  [ Dima Kogan ]
+  * All -dev packages are now Multi-Arch: same (Closes: #689088, #678040).
+
  -- Andreas Boll   Mon, 15 Sep 2014 14:19:30 +0200
 
 mesa (10.3.0~rc3-2) experimental; urgency=medium
diff --git a/debian/control b/debian/control
index cc021cc..857f2a5 100644
--- a/debian/control
+++ b/debian/control
@@ -143,6 +143,7 @@ Depends:
 Provides: libgl-dev, mesag-dev, libgl1-mesa-swrast-dev
 Conflicts: mesa-dev, libgl-dev, mesag3 (<< 3.1-1), nvidia-glx-dev, mesag-dev, 
libgl1-mesa-swrast-dev
 Replaces: libgl-dev, mesag-dev, libgl1-mesa-swrast-dev
+Multi-Arch: same
 Description: free implementation of the OpenGL API -- development files
  This package provides the development environment required for
  compiling programs with Mesa.  For a complete description of Mesa,
@@ -189,6 +190,7 @@ Architecture: linux-any
 Depends:
  libxatracker2 (= ${binary:Version}),
  ${misc:Depends},
+Multi-Arch: same
 Description: X acceleration library -- development files
  This package contains the XA (X acceleration) library.  It is used exclusively
  by the X server to do render, copy and video acceleration.
@@ -237,6 +239,7 @@ Depends:
  libgbm1 (= ${binary:Version}),
  libudev-dev,
  ${misc:Depends},
+Multi-Arch: same
 Description: generic buffer management API -- development files
  This package contains the GBM buffer management library.  It provides a
  mechanism for allocating buffers 

Bug#762047: xserver-xorg-video-ati: Black screen with mouse cursor on PowerPC

2014-09-18 Thread Bill Chatfield
It would seem so, but it is the right information. I double checked that it is 
right by booting without the video kernel parameter and running dmesg. I got 
the same radeonfb lines that you show in your message. I think that's part of 
the problem, radeonfb is compiled into the kernel and active when drm/radeon 
are being used? On a normal boot with a default configuration, radeonfb is 
active. It is blacklisted in /etc/modprobe.d/fbdev-blacklist.conf, but it's not 
compiled as an external kernel module (it's in the kernel itself) so the 
blacklisting doesn't work.



On Thu, 9/18/14, Michel Dänzer  wrote:

 Subject: Re: Bug#762047: xserver-xorg-video-ati: Black screen with mouse 
cursor on PowerPC
 To: "Bill Chatfield" 
 Cc: 762...@bugs.debian.org
 Date: Thursday, September 18, 2014, 2:24 AM
 
 On 18.09.2014 13:23, Bill Chatfield
 wrote:
 > [    0.997986] radeonfb :00:10.0:
 enabling device (0006 -> 0007)
 > [    1.193788] radeonfb :00:10.0: Invalid
 ROM contents
 > [    1.193813] radeonfb (:00:10.0):
 Invalid ROM signature 303 should be 0xaa55
 > [    1.193823] radeonfb: Retrieved PLL infos
 from Open Firmware
 > [    1.193842] radeonfb: Reference=27.00 MHz
 (RefDiv=12) Memory=203.00 Mhz, System=392.00 MHz
 > [    1.193864] radeonfb: PLL min 12000 max
 35000
 > [    1.289994] i2c i2c-2: unable to read EDID
 block.
 > [    1.441986] i2c i2c-2: unable to read EDID
 block.
 > [    1.593986] i2c i2c-2: unable to read EDID
 block.
 > [    1.949865] radeonfb: Monitor 1 type LCD
 found
 > [    1.949881] radeonfb: EDID probed
 > [    1.949893] radeonfb: Monitor 2 type no
 found
 > [    1.949915] radeonfb: Using Firmware
 dividers 0x0002008e from PPLL 0
 > [    1.949981] radeonfb: Dynamic Clock Power
 Management enabled
 > [    1.982573] Console: switching to colour
 frame buffer device 160x53
 > [    2.004457] radeonfb: Backlight
 initialized (radeonbl0)
 > [    2.004635] radeonfb (:00:10.0): ATI
 Radeon 4e50 "NP"
 
 This shows that radeonfb is still enabled, so it doesn't
 provide the 
 information corresponding to the bug.
 
 
 -- 
 Earthling Michel Dänzer         
   |             
     http://www.amd.com
 Libre software enthusiast         
 |               
 Mesa and X developer



--
To UNSUBSCRIBE, email to debian-x-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/1411050789.6500.yahoomailba...@web124705.mail.ne1.yahoo.com



Bug#761491: Link to another upstream bug report

2014-09-18 Thread Yvan Masson
Hi,

After searching for my initial problem on bugs.freedesktop.org, I think
this bug should be also linked with
https://bugs.freedesktop.org/show_bug.cgi?id=83851

Thanks


-- 
To UNSUBSCRIBE, email to debian-x-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/1411059279.1441.2.camel@yvan-macbook



Bug#761491: Link to another upstream bug report

2014-09-18 Thread Yvan Masson
Hi,

Sorry for previous mail, it was the wrong bug report. The right one is :
https://bugs.freedesktop.org/show_bug.cgi?id=81438

Thanks


-- 
To UNSUBSCRIBE, email to debian-x-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/1411060552.5834.2.camel@yvan-macbook