---
- branch: riastradh-drm2
  date: Wed Jul 24 03:24:03 UTC 2013
  files:
  - new: 1.1.1.1.2.49
    old: 1.1.1.1.2.48
    path: src/sys/external/bsd/drm2/dist/include/drm/drmP.h
    pathrev: src/sys/external/bsd/drm2/dist/include/drm/drmP.h@1.1.1.1.2.49
    type: modified
  - new: 1.1.2.13
    old: 1.1.2.12
    path: src/sys/external/bsd/drm2/include/linux/pci.h
    pathrev: src/sys/external/bsd/drm2/include/linux/pci.h@1.1.2.13
    type: modified
  - new: 1.1.2.2
    old: 1.1.2.1
    path: src/sys/external/bsd/drm2/pci/drm_pci.c
    pathrev: src/sys/external/bsd/drm2/pci/drm_pci.c@1.1.2.2
    type: modified
  id: 20130724T032403Z.edc00a6ac0bc76e23048ff6141ee8724f7a500b0
  log: |
    Kludge up Linux PCI device shims.

    . New drm_pci_attach/drm_pci_detach simplify initialization of
    autoconf-derived struct pci_dev.

    . New linux_pci_dev_init initializes struct pci_dev so that there's
    one place where all its fields are listed.  The kludge parameter is,
    well, kludgey.  Sorry.

    . Replace pci_kludgey_find_dev by pci_get_bus_and_slot with the same
    interface as Linux's, but some kasserts requiring it to look for the
    one bus/device/function tuple that i915drm is interested in.

    . Add pci_get_class which does similarly, for intel_detect_pch.

    Later pci_get_bus_and_slot and pci_get_class should be fixed once we
    can pass a cookie through PCI bus enumeration and pci_find_device.
  module: src
  subject: 'CVS commit: [riastradh-drm2] src/sys/external/bsd/drm2'
  unixtime: '1374636243'
  user: riastradh