[Spice-devel] [PATCH qxl-win v3] async and s3 queue combined

Alon Levy alevy at redhat.com
Thu Jul 7 02:43:05 PDT 2011


v2/v3 -> v3:
 Same as v2/v3 was intented to be - forgot to commit changes, and after rebasing
 most of the patches were changed, only 3 (asynchronous io port support) got changed
 significantly as per the "updated:" line below.

v2/v3 comment:

This is the whole thing, not splitting it this time. So I apologize
for patches already sent and acked on the list (but I think there are
none).

updated:
 async uses an enum for the async ops.
 flush_surfaces is also async.

You need an updated qemu and spice to test. (see patches on the corresponding
lists - 
 spice: 1309952376-29961-1-git-send-email-alevy at redhat.com
 qemu: http://cgit.freedesktop.org/~alon/qemu/log/?h=s3.v4.async.api (patches sent
  but not back in my email client yet to get the MessageID)

Alon Levy (36):
  miniport: we support rev 3 too now
  export pci revision from miniport to display
  asynchronous io port support (introduced in revision 3 of qxl device)
  miniport+display: export FLUSH_{SURFACES,RELEASE} ports for use by
    AssertModeEnable
  qxl_driver.h: bump QXL_DRIVER_INFO_VERSION (added pci_revision,
    flush_* ports)
  don't create a surface if vram mspace is unmapped
  miniport/*: use QXL_IO_LOG with internal snprintf
  miniport/qxl: store vram_start on video memory map for later debug
    usage
  miniport/qxl: implement SetPowerState:VideoPowerOn
  miniport/qxl: add DebugZeroDeviceMemory and use in DBG for hibernate
    and
  miniport: pass dev as first parameter of DEBUG_PRINT instead of 0
  miniport/qxl: disable DBG zap of video ram on map
  miniport/qxl: SetVideoModeInfo: pass QXLExtension for debug prints
  miniport/qxl: raise loglevel for mode prints
  display/mspace: fix mspace_max_footprint and mspace_footprint
  display/{driver,surface}: use pdev in some DEBUG_PRINTs
  display/driver: add DebugCountAliveSurfaces
  display/*: add debug helpers DUMP_VRAM_MSPACE and DUMP_DEVRAM_MSPACE
  display/driver: DisableQXLPrimarySurface: add hide_mouse parameter
    (reused in DrvAssertModeDisable later)
  driver/res: add debugging prints of released resources types
  (mp+dd) whitespace cleanup
  display/res: fix typo in debug print
  display: add a few debug prints (level=3)
  display/*: s/FreeSurface/FreeSurfaceInfo/
  display/res: add helpers for clearing device memory
  display/res: __AllocMem: verbose debugging (log_level 12)
  display/res: add a debug print level 9
  display/surface: add QXL_SURFACE_HOOKS define
  display/surface: GetSurfaceId: return -1 if \!surf->dhsurf
  display/surface: FreeSurfaceInfo: ignore double frees
  display/surface: add DEVICE_BITMAP_ALLOCATION_TYPE_RAM, cleanup
    surface alloc/free code paths
  display/surface: add surfaces from/to ram
  display/*: add PDev->enabled
  display/driver: reimplement DrvAssertMode for Suspend+Hibernate
    (S3+S4) support
  display: add FlushRelease and FlushSurfaces for alternative V06 (2)
    revision implementation
  merge miniport

Yonit Halperin (2):
  display/res: substitute CursorCacheRemove false "break" with "return"
  display/res: make (Cursor|Palette)CacheRemove always release the
    object

 display/driver.c            |  242 ++++++++++++---
 display/mspace.c            |    6 +-
 display/qxldd.h             |  104 ++++++-
 display/res.c               |  334 ++++++++++++++++-----
 display/res.h               |    4 +
 display/rop.c               |   12 +
 display/surface.c           |  241 +++++++++++++--
 display/surface.h           |   32 ++-
 include/qxl_driver.h        |   14 +-
 miniport/minimal_snprintf.c |  708 +++++++++++++++++++++++++++++++++++++++++++
 miniport/minimal_snprintf.h |    9 +
 miniport/qxl.c              |  332 ++++++++++++++-------
 miniport/qxl.h              |    3 +-
 miniport/qxl.inf            |    4 +
 miniport/sources            |    3 +-
 15 files changed, 1780 insertions(+), 268 deletions(-)
 create mode 100644 miniport/minimal_snprintf.c
 create mode 100644 miniport/minimal_snprintf.h

-- 
1.7.5.4



More information about the Spice-devel mailing list