Re: [PATCH v9 00/16] Get rid of the ACPI PCI subdriver mechanism

From: Bjorn Helgaas
Date: Fri Apr 12 2013 - 18:18:05 EST


On Fri, Apr 12, 2013 at 9:44 AM, Jiang Liu <liuj97@xxxxxxxxx> wrote:
> PCI, ACPI: remove ACPI PCI subdriver mechanism
>
> This patch set is directly derived from two sources:
> 1) '[PATCH 00/15] PCI/ACPI: Remove "pci_root" sub-driver support' at
> https://lkml.org/lkml/2012/12/7/11 from Myron Stowe
> The major goal is to resolve any potential sequencing inter-dependencies
> by converting sub-driver functionality to being only supported as
> statically built-in to the kernel as part of the "pci_root" driver
> itself.
>
> 2) 'introduce PCI bus notifier chain to get rid of the ACPI PCI subdriver'
> at http://lwn.net/Articles/533547/
> The major goal is to update PCI slots(pci_slot), ACPI based PCI hotplug
> slots(acpiphp), PCIe AER(aer) etc when hot-plugging PCI devices and P2P
> bridges, and eventually get rid of the ACPI PCI subdriver interfaces.
>
> This patchset applies to Bjorn's pci-next branch. Patch 1-4 are minor
> fixups for 3.9-rc1. Patch 5-8 introduce two hooks into PCI core which
> will be called when creating/destroying PCI buses. Patch 9-12 replace
> ACPI PCI subdriver interfaces with the new hook mechanism for pci_slot
> and acpiphp. Patch 13 removes the ACPI PCI subdriver interfaces.
>
> v7->v8:
> Hook directly into the PCI core to replace pci_bus notification
> Remove ACPI PCI subdriver related code
>
> v8->v9:
> Introduce kernel boot option "acpiphp.disable" to disable the
> acpiphp driver.
>
> Jiang Liu (13):
> PCI: do not check is_added flag in pci_remove_bus()
> pci: clean up usages of pci_bus->is_added
> PCI/acpiphp: don't rely on function 0 in disable_device()
> ACPI/acpiphp: replace local macros with standard ACPI macros
> PCI: introduce platform dependent hooks for creating/destroying PCI
> busses
> PCI, ACPI: prepare stub functions to handle ACPI PCI (hotplug) slots
> PCI, IA64: implement pcibios_{add|remove}_bus() hooks
> PCI, x86: implement pcibios_{add|remove}_bus() hooks
> PCI, ACPI: handle PCI slot devices when creating/destroying PCI
> busses
> PCI/acpiphp: convert acpiphp as a builtin driver
> PCI/acpiphp: do not use ACPI PCI subdriver mechanism
> PCI/acpiphp: protect acpiphp data structures from concurrent updating
> PCI/acpiphp: introduce a kernel option to disable the acpiphp driver
>
> Myron Stowe (1):
> PCI, ACPI: remove support of ACPI PCI subdrivers
>
> Yijing Wang (2):
> PCI/acpiphp: use list_for_each_entry_safe() in acpiphp_sanitize_bus()
> PCI/acpiphp: use normal list to simplify implementation
>
> Documentation/kernel-parameters.txt | 3 +
> arch/ia64/pci/pci.c | 11 +
> arch/x86/pci/common.c | 11 +
> drivers/acpi/pci_root.c | 48 +---
> drivers/acpi/pci_slot.c | 170 ++------------
> drivers/acpi/scan.c | 1 -
> drivers/pci/bus.c | 11 +-
> drivers/pci/hotplug/Kconfig | 7 +-
> drivers/pci/hotplug/acpiphp.h | 13 +-
> drivers/pci/hotplug/acpiphp_core.c | 29 +--
> drivers/pci/hotplug/acpiphp_glue.c | 415 ++++++++++++-----------------------
> drivers/pci/pci-acpi.c | 30 +++
> drivers/pci/probe.c | 15 +-
> drivers/pci/remove.c | 4 +-
> include/linux/acpi.h | 9 -
> include/linux/pci-acpi.h | 26 +++
> include/linux/pci.h | 2 +
> 17 files changed, 280 insertions(+), 525 deletions(-)

I applied these to my pci/jiang-subdrivers branch with minor tweaks.
The most significant is that I folded in the acpiphp.disable option to
the patch that makes the driver builtin-only. That way there's no
window between removing the "edit modules.conf" workaround and adding
the kernel parameter.

Take a look and make sure it's what you want:
http://git.kernel.org/cgit/linux/kernel/git/helgaas/pci.git/log/?h=pci/jiang-subdrivers

Bjorn
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/