Currently the kernel only supports AMD hostbridges up to Family 11h.
This causes PCI numa_node information to be reported incorrectly
for newer Families with multiple sockets.
This patch series augments the logic to discover newer AMD hostbridge types
(Family 12h, 14h 15h and 16h).
Note:
* Patch 1 and 2 are functional changes for newer AMD Families
* Patch 3 is general clean up and restructuring
* Patch 4 warns when BIOS has not supplied an ACPI _PXM method for a
hostbridge and we "guess" a hostbridge's node value
* Patch 5 removes the now redundant 'quirk_amd_nb_node'
-v2: This series was introduced by Suravee Suthikulanit but seems to have
stalled. I would like to get this moving again and eventually pulled
upstream so I've taken Suravee's series and -
o Rebased the original three patches to v3.15-rc1,
o Added a warn if we "guess" a hostbridge's node value (patch 4),
o I believe with the additional Family support that
'quirk_amd_nb_node' is now redundant so I added patch 5 which
removes it.
Reference:
https://bugzilla.kernel.org/show_bug.cgi?id=72051
Myron Stowe (2):
PCI: Remove redundant 'quirk_amd_nb_node'
ACPI/PCI: Warn if we have to "guess" host bridge node information
Suravee Suthikulpanit (3):
x86/PCI: Miscellaneous code clean up for early_fillup_mp_bus_info
x86/PCI: Support additional MMIO range capabilities
x86/PCI: Add support for generic AMD hostbridges
arch/x86/kernel/quirks.c | 58 -------
arch/x86/pci/acpi.c | 6 +
arch/x86/pci/amd_bus.c | 409 ++++++++++++++++++++++++++++++----------------
3 files changed, 271 insertions(+), 202 deletions(-)
--
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
