[PATCH v3] Enable serial output for Oxford Semiconductor PCIe cards and fix bugs.

[Konrad Rzeszutek Wilk <konrad@kernel.org>]

Since v1 [http://mid.gmane.org/1394040334-16278-1-git-send-email-konrad.wilk@oracle.com]:
  - Update it per Jan and Andrew comments
  - Tackle on two bug-fixes.
v2: [http://mid.gmane.org/1394218904-21535-1-git-send-email-konrad.wilk@oracle.com]:
 - Squash two patches in,
 - Add appropiate tags,
 - Rework bug-fixes per Jan's recommendations.

Hey,

This is the third posting of the changes to the ns16550 driver
to also support other chipsets - like the Oxford 16C950 one - popular on
the inexpensive Rosewill RC-300E.

There are also two bug-fixes that I discovered while testing my code.
One of them was found through code inspection while the other was
when I was testing my code. Aravind graciously had tested the patches -
so I affixed Tested-by on the patches that did not have much (or any)
code churn since v2 posting.

Of interest is:

[PATCH 3/6] serial: Support OXPCIe952 aka Oxford Semiconductor Ltd

  that enables the functionality on the PCIe card I have - while the rest
of the patches flesh out the code to support more of PCI and PCIe type cards.

 xen/arch/x86/oprofile/op_model_athlon.c |    4 +-
 xen/arch/x86/x86_64/mmconf-fam10h.c     |    1 +
 xen/arch/x86/x86_64/mmconfig-shared.c   |    1 +
 xen/arch/x86/x86_64/mmconfig.h          |    4 -
 xen/drivers/char/ns16550.c              |  255 ++++++++++++++++++++++++++++---
 xen/include/xen/pci_ids.h               |    9 +
 6 files changed, 249 insertions(+), 25 deletions(-)

Konrad Rzeszutek Wilk (6):
      serial: Skip over PCIe device which have no quirks (fix AMT regression).
      serial: Fix COM1 assumption if pci_uart_config did not find the AMT card.
      serial: Support OXPCIe952 aka Oxford Semiconductor Ltd Device c138 (1415:c138)
      serial: Seperate the PCI device ids and parameters (v1)
      pci: Use #defines for PCI vendors.
      serial: Expand the PCI serial quirks for OXPCIe200 and OXPCIe952 1 Native UART