Patchwork [Quantal] pull-request: Add support for Haswell graphics

login
register
mail settings
Submitter Leann Ogasawara
Date Sept. 17, 2012, 6:31 p.m.
Message ID <50576C85.7040109@canonical.com>
Download mbox
Permalink /patch/184521/
State New
Headers show

Pull-request

git://kernel.ubuntu.com/ogasawara/ubuntu-quantal.git lp1011440

Comments

Leann Ogasawara - Sept. 17, 2012, 6:31 p.m.
Hi All,

The following pull request cherry-picks/backports upstream patches from
v3.6 needed for Haswell graphics support in our Quantal v3.5 based
kernel.  I'd like to get a quick additional review of these patches
before applying.  I'd ideally also like to see these applied and
uploaded before our Beta-2 freeze this Thurs Sep 20 for more widespread
testing.  I would also note that I've received initial testing feedback
on 2 development platforms with Haswell graphics.   Results are as follows:

For each, normally the monitor (Iiyama ProLite E2209HDS) can handle
1920x1080 and 1280x1024 VGA.

Laptop Platform (Beta HW):
 * HDMI: 1920x1080, 1280x1024, 1024x768, 800x600 - all work fine
 * DisplayPort: 1920x1080, 1280x1024, 1024x768, 800x600 - all work fine
 * VGA: 1024x768, 800x600, works fine
    - 1280x1024, not stretched fill width on monitor
    - 1920x1080, edges fall off the monitor, can't sync up
 * A S3 test cycle running glxgears, the machine locked up on the 1st
iteration and on reboot can't get any USB keyboards to currently work. 
suspect the USB keyboard problem is a completely different issue though.

Desktop Platform (Alpha HW):
 * HDMI: 1920x1080, 1280x1024, 1024x768, 800x600 - all work fine
 * Display Port: No output at all (even on UEFI boot, this is probably
an alpha board issue, not a kernel problem).
 * VGA: Same as Laptop Platform:
    - 1024x768, 800x600, works fine   
    - 1280x1024, not stretched fill width on monitor
    - 1920x1080, edges fall off the monitor, can't sync up.  The image
is not very stable.  Again, saw this issue on the early laptop board, so
it's a silicon/board issue probably.
 * And S3 reboots the machine on 1st iteration, so the same as the
laptop SDP - however, this time USB works fine (thankfully!).

So while not completely perfect feedback from initial non-exhaustive
testing, I do feel it is worthwhile applying these patches while in the
process of fixing the remaining outlying issues.  I've also tested on
non-Haswell kit to confirm no regressions.  For anyone else interested
in testing, I've posted a test kernel at the following location:

http://people.canonical.com/~ogasawara/lp1011440/
<http://people.canonical.com/%7Eogasawara/lp1011440/>

Please review the pull request below.  I'll then push and upload.

Thanks,
Leann

The following changes since commit c72b8bc73efec4092bfa4ee9e66560855fe6cef3:

  eCryptfs: Call lower ->flush() from ecryptfs_flush() (2012-09-17
11:12:05 -0700)

are available in the git repository at:

  git://kernel.ubuntu.com/ogasawara/ubuntu-quantal.git lp1011440

for you to fetch changes up to ed6eb51c45f60a35bd133927ef6e9970f0c37e57:

  drm/i915: add more Haswell PCI IDs (2012-09-17 11:12:49 -0700)

----------------------------------------------------------------
Chris Wilson (1):
      drm/i915: Implement w/a for sporadic read failures on waking from rc6

Daniel Vetter (1):
      drm/i915: clarify IBX dp workaround

Eugeni Dodonov (6):
      drm/i915: support Haswell force waking
      drm/i915: add RPS configuration for Haswell
      drm/i915: enable RC6 by default on Haswell
      drm/i915: introduce haswell_init_clock_gating
      drm/i915: enable RC6 workaround on Haswell
      drm/i915: re-initialize DDI buffer translations after resume

Paulo Zanoni (3):
      drm/i915: fix PIPE_DDI_PORT_MASK
      drm/i915: try to train DP even harder
      drm/i915: add more Haswell PCI IDs

 drivers/char/agp/intel-agp.h         |   39 +++++++++++--
 drivers/char/agp/intel-gtt.c         |   60 ++++++++++++++++++-
 drivers/gpu/drm/i915/i915_drv.c      |   76 ++++++++++++++++++++++--
 drivers/gpu/drm/i915/i915_reg.h      |   13 ++++-
 drivers/gpu/drm/i915/intel_display.c |    4 +-
 drivers/gpu/drm/i915/intel_dp.c      |    4 +-
 drivers/gpu/drm/i915/intel_pm.c      |  107
+++++++++++++++++++++++++++++-----
 7 files changed, 272 insertions(+), 31 deletions(-)
Tim Gardner - Sept. 17, 2012, 7:29 p.m.
Regression tested on Atom w/i915 GPU.