Skip to content


Subversion checkout URL

You can clone with
Download ZIP
Branch: master
Commits on Jan 4, 2011
  1. pscnv: install the nv04/c0_pushbuf.h headers

    Christoph Bumiller authored
  2. pscnv: fix nouveau_grobj_autobind pushbuf commands

    Christoph Bumiller authored
  3. pscnv: import split pushbuf headers from nouveau

    Christoph Bumiller authored
  4. Merge remote branch 'fdo/master'

    Christoph Bumiller authored
    Not really conflicts, something went wrong.
Commits on Dec 21, 2010
  1. nouveau: fix up reloc_emit() to accept NULL target buffer

    Ben Skeggs authored
    The nvc0 gallium drivers passes NULL here to indicate to the memory manager
    that a buffer is being used, but without creating an actual reloc.
    Signed-off-by: Ben Skeggs <>
  2. nouveau: split pushbuf macros specific to nv04-nv50 out, and add nvc0

    Ben Skeggs authored
    Signed-off-by: Ben Skeggs <>
  3. nouveau: nvc0 drm has no concept of "notifier block"

    Ben Skeggs authored
    Signed-off-by: Ben Skeggs <>
Commits on Dec 19, 2010
  1. @ickle

    intel: Export CONSTANT_BUFFER addressing mode

    ickle authored
    Signed-off-by: Chris Wilson <>
Commits on Dec 10, 2010
  1. @ickle

    configure: Bump version to 2.4.23

    ickle authored
    To export new kernel API for Intel's 2010Q4 release.
    Signed-off-by: Chris Wilson <>
  2. @ickle

    tests: Update for ENOENT returns from unknown handles

    ickle authored
    Signed-off-by: Chris Wilson <>
Commits on Dec 8, 2010
  1. pscnv: use bits ff00 for memory storage type on nvc0 as well

    Christoph Bumiller authored

    Christoph Bumiller authored
Commits on Dec 7, 2010
  1. @ickle

    intel: Reorder need_fence vs fenced_command to avoid fences on gen4

    ickle authored
    gen4+ hardware doesn't use fences for GPU access and the older kernel
    doesn't expect userspace to make such a mistake. So don't.
    Signed-off-by: Chris Wilson <>
  2. drm: don't do the create the node ourselves if we have udev.

    Dave Airlie authored
    this can remove nodes it shouldn't, let udev run the show.
    this is needed for reliably GPU switch.
    Signed-off-by: Dave Airlie <>
Commits on Dec 3, 2010
  1. @ickle

    intel: If the command is fenced inform the kernel

    ickle authored
    ... but only account for a fenced used if the object is tiled.
    Signed-off-by: Chris Wilson <>
Commits on Dec 2, 2010
  1. @marekolsak
Commits on Nov 25, 2010
  1. @ickle

    intel: Add a forward declaration of struct drm_clip_rect

    ickle authored
    ... so that intel_bufmgr.h can be compiled standalone.
    Signed-off-by: Chris Wilson <>
Commits on Nov 22, 2010
  1. @ickle

    intel: Compute in-aperture size for relaxed fenced objects

    ickle authored
    For relaxed fencing the object may only consume the small set of active
    pages, but still requires a fence region once bound into the aperture.
    This is the size we need to use when computing the maximum possible
    aperture space that could be used by a single batchbuffer and so avoid
    hitting ENOSPC.
    Reported-by: Daniel Vetter <>
    Signed-off-by: Chris Wilson <>
  2. @curro

    nouveau: Add implicit pushbuf flush before gpuobj destruction.

    curro authored
    It makes sure that GPU object destruction is executed in order with
    respect to the previous FIFO commands.
    Signed-off-by: Francisco Jerez <>
    Acked-by: Ben Skeggs <>
Commits on Nov 9, 2010
  1. @anholt

    intel: Fix drm_intel_gem_bo_wait_rendering to wait for read-only usag…

    anholt authored
    …e too.
    Both the consumers of this API (sync objects and client throttling)
    were expecting this behavior.  The kernel used to actually behave the
    desired (but incorrect) way for us anyway, but that got fixed a while
Commits on Nov 7, 2010
  1. @ickle

    intel: initialize bufmgr.bo_mrb_exec unconditionally

    Albert Damen authored ickle committed
    If bufmgr.bo_mrb_exec is not set, drm_intel_bo_mrb_exec returns ENODEV
    even though drm_intel_gem_bo_mrb_exec2 will work fine for the RENDER ring.
    Fixes xf86-video-intel after commit 'add BLT ring support' (5bed685f76)
    with kernels without BSD or BLT ring support (2.6.34 and before).
    Signed-off-by: Albert Damen <>
    Signed-off-by: Chris Wilson <>
Commits on Nov 2, 2010
  1. @anholt

    intel: Drop silly asserts on mappings present at unmap time.

    anholt authored
    The intent of these was to catch mismatched map/unmap.  What it
    actually did was check whether there was ever a mapping of that type
    (including in a previous life of the buffer through the userland BO
    cache), not whether they were mismatched.  We don't even actually want
    to catch mismatched map/unmap, unless we also do refcounting, since at
    one point Mesa would do map/map/use/unmap/unmap.  Just remove this
    code instead.
  2. @anholt

    intel: Remove gratuitous assert on bo_reference.

    anholt authored
    This couldn't be triggered except by overflow, since there's an assert
    in unreference to catch the usual failure of over-unreferencing.
  3. pscnv: s/libdrm_nouveau/libdrm_pscnv in

    Christoph Bumiller authored
  4. pscnv: increase spin count on fence wait

    Christoph Bumiller authored
  5. nvc0: add an IB call marker into the trace file

    Christoph Bumiller authored
  6. pscnv: silence IB PUT debug message

    Christoph Bumiller authored
  7. pscnv: add pushbuf_size param to nouveau_channel_alloc

    Christoph Bumiller authored
    Regain API compatiblity.
  8. pscnv: use NOUVEAU_BO_TILE_SCANOUT like nouveau

    Christoph Bumiller authored
  9. Reestablish compatibility with fdo libdrm.

    Christoph Bumiller authored
Commits on Nov 1, 2010
  1. @anholt
  2. @anholt

    intel: Remove stale comment.

    anholt authored
Commits on Oct 31, 2010
  1. @curro

    nouveau: Avoid unnecessary call to CPU_FINI.

    curro authored
    nouveau_bo_unmap called the CPU_FINI IOCTL even if it was a NOSYNC
    mapping. It caused no harmful effects (actually CPU_FINI is a no-op on
    recent enough kernels) besides the precious CPU cycles being wasted.
    Signed-off-by: Francisco Jerez <>
Commits on Oct 29, 2010
  1. @ickle

    intel: enable relaxed fence allocation for i915

    ickle authored
    The kernel has always allowed userspace to underallocate objects
    supplied for fencing. However, the kernel only allocated the object size
    for the fence in the GTT and so caused tiling corruption. More recently
    the kernel does allocate the full fence region in the GTT for an
    under-sized object and so advertises that clients may finally make use
    of this feature. The biggest benefit is for texture-heavy GL games on
    i945 such as World of Padman which go from needing over 1GiB of RAM to
    play to fitting in the GTT!
    Signed-off-by: Chris Wilson <>
Commits on Oct 27, 2010
  1. @nwnk


    nwnk authored
    _DRM_MALLOC hasn't been a relevant concern since we split libdrm out
    from xserver.
    Signed-off-by: Adam Jackson <>
Something went wrong with that request. Please try again.