[Mesa-stable] Picks to 10.0 release branch

[Ian Romanick]

I believe that I have picked over all of the patches marked for the 10.0
release branch.  I have included the full list far below.  There are a
few patches that were sent to the mesa-stable mailing list that do not
appear to have landed on master yet.  I have not picked these over.
These are:

radeon/compute: Unconditionally inline all functions
i965/gen7: Emit workaround flush when changing GS enable state.
r600g/sb: work around hw issues with stack on eg/cm
r600g/compute: Fix handling of global buffers in r600_resource_copy_region()
docs: update 10.0 relnotes for GL_NV_vdpau_interop

If a patch that you believe should be on the release branch is not on
one of these two lists, please send (or resend) it to mesa-stable... or
at least send me links to the archive because I accidentally deleted the
message. :(

I am currently planning to do Mesa 10.0-rc1 later today.

3fd32619d7143ed7db21fa402d38e681520f82c2 radeon/llvm: Free elf_buffer after use
7a87dc278e5b521ed6d4e30afa24bcc54df72d2e r600/llvm: Free binary.code/binary.config in r600_llvm_compile
f843604b6a71a9809db8839d36750fbbdec25bc1 r600/llvm: initialize radeon_llvm_binary
e9f8b782788f4f1b27fb2015b1cf226660a17271 svga: mark dest image as defined in svga_surface_copy()
dfff838429060986a0813d74a87ca81b0406df95 svga: do primitive trimming in translate_indices()
11982ca08dd8ef5111bf2155e52c3de8d2840431 gallium/pipe_loader: un-reference udev resources when we're done with them.
713966c82f8051575d53c8da918ec9365b087dda radeonsi/compute: Dispose of LLVM module after compiling kernels
3a98fc6abecc7bce02b3de4861f49db38daff775 radeonsi/compute: Free program and program.kernels on shutdown
531637feee3c4fa378f2fe25399cff833592b269 radeon/llvm: Free created llvm memory buffer
02807c06b89ccd185e007580b5b50da5bf399e27 radeon/llvm: Free libelf resources
9ed045274061ec744b74116c7f49ce7d2960531c radeon/llvm: fix spelling error
ef8fcfc9cf55d11f5fc0b93ec52fa8d4d1a73135 clover: Support multiple devices in clCreateContextFromType() v2
1b45f255b592692bdcd25b1f487a2f72c469ac63 glsl: Rework interface block linking.
1a163c0b347a089c4db59bbc38d6fc87e2da9e50 i965: Fix vertical alignment for multisampled buffers.
53e681f2fed2ca1f9e288b776d6d74186b988d4f main: Fix MaxUniformComponents for geometry shaders.
10c25e58cab58acceee36baa46d1cacf0172a4d6 mesa: Fix derived vertex state not being updated in glCallList()
0558e10160c1601cdf347f94ea8d9359c62812ac dri: Change value param to unsigned
1e51d3a668a3032c028a3a35e04088c7e362adfb i965: Use drm_intel_get_aperture_sizes instead of hard-coded 2GiB
e5839c2397a0f99b9d5035b5989019feb48b856b i915: Use drm_intel_get_aperture_sizes instead of drmAgpSize
7d2187176a3708e3bc00114a8172ec7cc73f39d3 dri: Remove redundant createNewContext function from __DRIimageDriverExtension
329a75511fe2ab454e487855a65f9a3943a11ec1 wayland: Use __DRIimage based getBuffers implementation when available
76434775e07abe9f1f372284c0fd46d6fffccfa6 gbm: Add support for __DRIimage based getBuffers when available
236524430228e2cc090d89652a0a46591ebe7ac0 dri/i915, dri/i965: Fix support for planar images
3e6f200250f3a4e892a18e089dc9faecf6d9c7b8 i965/fs: Try a different pre-scheduling heuristic if the first spills.
99c62ff2ea488046231acec463ed96efadd1b65c i965/fs: Do instruction pre-scheduling just before register allocation.
a5a6ef970282ac1050d6cc310ed5fcd0bff4fa0a i965/fs: Ignore actual latency pre-reg-alloc.
6640147463640bcbb49d4bbdb8efdad0ff6a5558 i965/fs: Fix message setup for SIMD8 spills.
229ee204605b5f024662fb7e2748a40630d60790 i965/fs: Prefer things we know reduce reg pressure when pre-scheduling.
dbddd86cc2b8b21a3d0460bb3d9262a86a5e39f6 i965: Fix undefined value usage in ABO setup.
c702f5eeade9c1cb2417c4b546407bb101d35d2d clover: Fix the const variant of adaptor_range::end to deal with mismatching range sizes.
c4cc166abc038fc7ac4b4c1514fa462f0bfb1d26 gallium: fix build on GNU/Hurd due to missing PIPE_OS_HURD detection
2a3dcece7257722cdaa71978e489a1600849b2dd meta: enable vertex attributes in the context of the newly created array object
accc276df2d32b46d086bc5beb9d5d5f5224a7a5 mesa: call update_array_format() after error checking
afbdcdcaaf11fd513ba44e474913b88774016b9b nouveau/video: mark bitstream-level acceleration as unsupported
6f2877c40d931b693369d1d00186ec23d5ede1e0 nouveau/video: don't try on nv3x
02d9e1be87f0bb03eae829ae3f56a6c0bee33826 egl-static:  Only export necessary symbols v3
095d583e528f7c59c3471df50ba136f776f9bef9 configure.ac: Don't require shared LLVM when building OpenCL
8af132fca9308913235f65a3bd34681a813f409b pipe-loader: Only export necessary symbols v3
ade312cd8a42c0ccb37c91d02c88e0a5672ded9b radeonsi/compute: Add Sea Islands support (cherry picked from commit a859131003fc79094cf99c92cc6a11a0f0c659f7)
f9a74a0b4c3dc36a3e9437ba13124b5ebf6583c3 tests: Fix make check for out of tree builds.
0e3f5999b925aab627b21bca6670a9e1806715b2 osmesa: fix broken triangle/line drawing when using float color buffer
ebc460bc5f0100106c74a9cee82cf187ab998d35 i965: convert brw_lower_offset_array_visitor to ir_rvalue_visitor
0010bdd54a490ecf45652cfdf5b8d8a275b7118f glsl: fix missing breaks in equals(ir_texture,..)
b8a631295a6cfc7a26f529980202f1198701e136 i965/fs: Don't perform CSE on inst HW_REG dests (unless it's null)
c94ed272eb81bb005403811f075ba57f13dee18b Add .cherry-ignore file
47139b02335dee675fdcf14b9a766c4cdcec890f glx: Back DRI3 enablement out of the stable branch.
94251281b4e914ec7b29e811ed6a565128745d0c glx: change query_renderer_integer() value param to unsigned
03a29306b50a0a61965e80323cb5cf2b10c26b8e glx: Fix scons build.
d37ea6dfec64301953486d21a15e744d3d3bbb50 swrast: add missing notify_reset parameter to dri_create_context()
84ee00c1b2d4fb155dea4a92fe814fbdc9579d02 scons: Add dri2_query_renderer.c to sources. (cherry picked from commit cb3c57df3acfa237641aa004545a786f5624f198)
3ffcc96abcfeb709cb431a7555aca078d710b3b9 st/dri: Fix dri_create_context declaration prototype. (cherry picked from commit caf1d96862553b169aa9d8fac63fcbd37a59d099)