mesa-19.0.1

[Douglas R. Reno]

New point version

Hi List,

I'd like to announce the availability of mesa 19.0.1.

This is the first bug fix release of the 19.0 branch. It's been a
pretty calm cycle, and there's not too much here. I think things are
looking pretty good overall.

Dylan

shortlog:
Andres Gomez (4):
      glsl: correctly validate component layout qualifier for dvec{3,4}
      glsl/linker: don't fail non static used inputs without matching outputs
      glsl/linker: simplify xfb_offset vs xfb_stride overflow check
      Revert "glsl: relax input->output validation for SSO programs"

Bas Nieuwenhuizen (2):
      radv: Use correct image view comparison for fast clears.
      ac/nir: Return frag_coord as integer.

Danylo Piliaiev (2):
      anv: Treat zero size XFB buffer as disabled
      glsl: Cross validate variable's invariance by explicit invariance only

Dave Airlie (1):
      softpipe: fix texture view crashes

Dylan Baker (6):
      docs: Add SHA256 sums for 19.0.0
      cherry-ignore: Add commit that doesn't apply
      bin/install_megadrivers.py: Correctly handle DESTDIR=''
      bin/install_megadrivers.py: Fix regression for set DESTDIR
      bump version for 19.0.1
      Add release notes for 19.0.1

Eric Anholt (1):
      v3d: Fix leak of the renderonly struct on screen destruction.

Jason Ekstrand (6):
      glsl/lower_vector_derefs: Don't use a temporary for TCS outputs
      glsl/list: Add a list variant of insert_after
      anv/pass: Flag the need for a RT flush for resolve attachments
      nir/builder: Add a vector extract helper
      nir: Add a new pass to lower array dereferences on vectors
      intel/nir: Lower array-deref-of-vector UBO and SSBO loads

Józef Kucia (2):
      radv: Fix driverUUID
      mesa: Fix GL_NUM_DEVICE_UUIDS_EXT

Kenneth Graunke (1):
      intel/fs: Fix opt_peephole_csel to not throw away saturates.

Kevin Strasser (1):
      egl/dri: Avoid out of bounds array access

Mark Janes (1):
      mesa: properly report the length of truncated log messages

Plamena Manolova (1):
      i965: Disable ARB_fragment_shader_interlock for platforms prior to GEN9

Samuel Pitoiset (3):
      radv: set the maximum number of IBs per submit to 192
      radv: always initialize HTILE when the src layout is UNDEFINED
      radv: fix binding transform feedback buffers

Sergii Romantsov (1):
      d3d: meson: do not prefix user provided d3d-drivers-path

Tapani Pälli (2):
      isl: fix automake build when sse41 is not supported
      anv/radv: release memory allocated by glsl types during spirv_to_nir

Seems the DESTDIR bug is fixed!

[ken@…]

A report that running a game under wine is broken, but that is probably not new breakage since 19.0 (gamers seem to be very vocal about their use case ;-)

I'd like to take this, in the context of fixing the meson build, but for the moment I have not *installed* it.

[Douglas R. Reno]

Replying to ken@…:

A report that running a game under wine is broken, but that is probably not new breakage since 19.0 (gamers seem to be very vocal about their use case ;-)

Off-topic, but replying to Ken's comment:

That's not surprising. There's a new program available called Proton that emulates the Windows API directly, and then translates it to calls to Linux OpenGL libraries. It's got a 99% compatibility rate, with most Windows Games on services like Steam and GOG functioning perfectly on Linux, as well as stuff from Blizzard. A majority of gamers were tied to Windows, even with their data-mining practices, and now they are able to go to Linux without problems.

Same applies to Microsoft Office and Visual Studio actually (it's even used upstream as a support GitLab/GitHub CI configuration), I'm just too lazy to create a Multilib system just to get away from having to use Windows for Office and Visual Studio.

​https://www.protondb.com/ - a majority of those reports are because of Proton.

[ken@…]

Some directories/files in 19.0.0 were not noted:

/usr/share/drirc.d/

This contains 00-mesa-defaults.conf which has workarounds for various applications (particularly, browsers and games) and sometimes for specific drivers.

/usr/share/vulkan/icd.d/

This has json files for Vulkan Installable Client Drivers (intel and radeon) : These appear to specify how vulkan can find drivers.

[ken@…]

Bruce has posted some (reformatted) output from 'meson configure' on -dev, I need to find time to read, understand, and then digest all that information.

[ken@…]

If people do want to limit the drivers they build, then r600 needs to be added to the list of drivers which cover "most modern video cards". My ryzen3 is using a Caicos card - low power, not latest, but a quick check suggests I could buy another one tomorrow if I wished, so it is no older than some i915 devices.

Using

LIBGL_DEBUG=verbose glxgears 2>&1 | tee glxlog

I can see that it uses /usr/lib/dri/r600_dri.so. Omitting r600, I have no r600 or radeon (except radeonsi) drivers.

If I add r600 to the gallium drivers, this reinstates /usr/lib/libXvMCr600.so, /usr/lib/dri/r600_dri.so, /usr/lib/dri/r600_drv_video.so and /usr/lib/vdpau/libvdpau_r600.so.

That means that only the r200 and r300 dri drivers, and the r300 vdpau driver are omitted.

For people who know exactly which video hardware they have, omitting things they do no need makes some sense. But the amount of space and time saved is tiny unless people remove other drivers - in this case, omitting r200, r300 saved only 8MB and 0.1 SBU. with r600 also omitted, it saved perhaps 10 MB more (totals for build and DESTDIR install).

[ken@…]

Side note: oddly, running that same verbose command on my intel haswell tells me nothing about which driver it is using, although I see it is using DRI2

[ken@…]

Measurements:

Full release build, 473 MB build and DESTDIR install with demos and docs, 2.4 SBU.

With added -Dbuild-tests=true add 141 MB and 0.4 SBU : 70 tests, all reported OK. Actually running the tests took a handful of seconds, the bulk of the time was in the (main) build.

Although my first attempt to run the tests appeared to be using git to create a hash before running them, subsequent builds with git hidden ran fine. Digging down to try to see what it had run, I found a message

   YOU HAVE 5 DISABLED TESTS

from build/src/compiler/nir/nir_vars_test but no further details. The tests appear to all be written in c++ and live in the main src/ directory.

NB mis-spelling a define name (without the hyphen in this case) does nothing and does not give any error messages.

[ken@…]

I had been intending to drop the current variables for auto builds, and adding things like -Dgallium-drivers="i915,swrast" -Ddri-drivers="" for people who wished to limit what they build. But in best mesa tradition, putting those into an otherwise empty variable barfs on the quotes (whether single or double). So, for the drivers I intend to keep the existing variables, but with a default value of 'auto'.

Will try to remember to note that these defines can be omitted by people who only want to use auto.

For platforms, I can't see why anyone would want to use anything except 'auto', so I am inclined not to mention that define.

[ken@…]

Updated at r21425.

In the end, I did not specifically mention omitting -Dgallium-drivers and -Ddri-drivers to force 'auto'.

[Bruce Dubbs]

Milestone renamed