Skip to content

GitLab

Switch branch/tag
  1. 11 Jan, 2019 10 commits
  3. 20 Nov, 2018 1 commit
  5. 14 Nov, 2018 3 commits
  7. 26 Oct, 2018 1 commit
    • Rob Clark's avatar
      freedreno: import libdrm_freedreno + redesign submit · f3cc0d27
      Rob Clark authored 
      

In the pursuit of lowering driver overhead, it became clear that some
amount of redesign of how libdrm_freedreno constructs the submit ioctl
would be needed.  In particular, as the gallium driver is starting to
make heavier use of CP_SET_DRAW_STATE state groups/objects, the over-
head of tracking cmd buffers and relocs becomes too much.  And for
"streaming" state, which isn't ever reused (like uniform uploads) the
overhead of allocating/freeing ringbuffer[1] objects is too high.

This redesign makes two main changes:

 1) Introduces a fd_submit object for tracking bos and cmds table
    for the submit ioctl, making ringbuffer objects more light-
    weight.  This was previously done in the ringbuffer.  But we
    have many ringbuffer instances involved in a submit (gmem +
    draw + potentially 1000's of state-group rbs), and only need
    a single bos and cmds table.  (Reloc table is still per-rb)

    The submit is also a convenient place for a slab allocator for
    ringbuffer objects.  Other options would have required locking
    because, while we can guarantee allocations will only happen on
    a single thread, free's could happen either on the application
    thread or the flush_queue thread.  With the slab allocator in
    the submit object, any frees that happen on the flush_queue
    thread happen after we know that the application thread is done
    with the submit.

 2) Introduce a new "softpin" msm_ringbuffer_sp implementation that
    does not use relocs and only has cmds table entries for IB1 (ie.
    the cmdstream buffers that kernel needs to CP_INDIRECT_BUFFER
    to from the RB).  To do this properly will require some updates
    on the kernel side, so whether you get the softpin or legacy
    submit/ringbuffer implementation at runtime depends on your
    kernel version.

To make all these changes in libdrm would basically require adding a
libdrm_freedreno2, so this is a good point to just pull the libdrm code
into mesa.  Plus it allows for using mesa's hashtable, slab allocator,
etc.  And it lets us have asserts enabled for debug mesa buids but
omitted for release builds.  And it makes life easier if further API
changes become necessary.

At this point I haven't tried to pull in the kgsl backend.  Although
I left the level of vfunc indirection which would make it possible
to have other backends.  (And this was convenient to keep to allow
for the "softpin" ringbuffer to coexist.)

NOTE: if bisecting a build error takes you here, try a clean build.
There are a bunch of ways things can go wrong if you still have
libdrm_freedreno cflags.

[1] "ringbuffer" is probably a bad name, the only level of cmdstream
    buffer that is actually a ring is RB managed by kernel.  User-
    space cmdstream is all IB1/IB2 and state-groups.
Reviewed-by: default avatarKristian H. Kristensen <hoegsberg@chromium.org>
Reviewed-by: Eric Engestrom's avatarEric Engestrom <eric.engestrom@intel.com>
Signed-off-by: Rob Clark's avatarRob Clark <robdclark@gmail.com>
      f3cc0d27
  9. 23 Oct, 2018 1 commit
  11. 22 Oct, 2018 1 commit
  13. 19 Oct, 2018 1 commit
  15. 17 Oct, 2018 1 commit
  17. 10 Oct, 2018 1 commit
    • Dylan Baker's avatar
      meson: Don't allow building EGL on Windows or MacOS · b8521704
      Dylan Baker authored 
      

Currently mesa only supports EGL on Unix like systems, cygwin, and
haiku. Meson should actually enforce this. This fixes the default build
on MacOS.

v2: - invert the condition, mark darwin and windows as not supported
      instead of trying to mark what is supported.
v3: - add missing )
v3: - Update comment to reflect condition change in v2

CC: 18.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Eric Engestrom's avatarEric Engestrom <eric.engestrom@intel.com>
      b8521704
  19. 03 Oct, 2018 1 commit
  21. 25 Sep, 2018 2 commits
  23. 24 Sep, 2018 1 commit
  25. 21 Sep, 2018 1 commit
  27. 20 Sep, 2018 1 commit
  29. 17 Sep, 2018 1 commit
    • Dylan Baker's avatar
      move pthread_setaffinity_np check to the build system · 3acc18fc
      Dylan Baker authored 
      Rather than trying to encode all of the rules in a header, lets just put
them in the build system where they belong. This fixes the build on
FreeBSD, which does have pthraed_setaffinity_np, but it's in a
pthread_np.h, not behind _GNU_SOURCE. FreeBSD also implements cpu_set
slightly differently, so additional changes would be required to get it
working right there anyway.

v2: - fix #define in autotools

Fixes: 9f1bbbdb


       ("util: try to fix the Android and MacOS build")
Cc: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: default avatarMarek Olšák <marek.olsak@amd.com>
Reviewed-by: Eric Engestrom's avatarEric Engestrom <eric.engestrom@intel.com>
      3acc18fc
  31. 06 Sep, 2018 2 commits
  33. 05 Sep, 2018 1 commit
    • Dylan Baker's avatar
      meson: Print a message about why a libdrm version was selected · d25a27ec
      Dylan Baker authored 
      We require a single version of libdrm for all of our libdrm
dependencies (core and driver), but the way this is structured can make
the error message less than helpful, as one driver might be the one
setting the libdrm requirement, while another might be the one that
generates the version failure.

This adds a simple message to the output announcing which libdrm module
set the version, which might be more helpful.

v2: - Use message suggested by Eric Engstrom

Fixes: c445b1d5


       ("meson: Use the same version for all libdrm checks")
Reviewed-by: Eric Engestrom's avatarEric Engestrom <eric.engestrom@intel.com>
      d25a27ec
  35. 27 Aug, 2018 2 commits
  37. 20 Aug, 2018 1 commit
  39. 16 Aug, 2018 3 commits
  41. 10 Aug, 2018 1 commit
  43. 08 Aug, 2018 1 commit
  45. 07 Aug, 2018 2 commits