]> git.karo-electronics.de Git - mv-sheeva.git/log
mv-sheeva.git
16 years ago[ARM] 4820/1: RealView: Select the timer IRQ at run-time
Catalin Marinas [Mon, 4 Feb 2008 16:43:02 +0000 (17:43 +0100)]
[ARM] 4820/1: RealView: Select the timer IRQ at run-time

This patch sets the timer IRQ at run-time by moving the sys_timer
structure and the timer_init function to the realview_eb.c file. This
allows multiple RealView platforms to be compiled in the same kernel
image.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] 4819/1: RealView: Fix entry-macro.S to work with multiple platforms
Catalin Marinas [Mon, 4 Feb 2008 16:41:01 +0000 (17:41 +0100)]
[ARM] 4819/1: RealView: Fix entry-macro.S to work with multiple platforms

This patch modifies the get_irqnr_preamble macro to work with multiple
platforms at run-time by reading the address of the GIC controller from
the gic_cpu_base_addr variable. This variable is defined in core.c and
intialised in realview_eb.c (gic_init_irq).

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] 4818/1: RealView: Add core-tile detection
Catalin Marinas [Mon, 4 Feb 2008 16:39:00 +0000 (17:39 +0100)]
[ARM] 4818/1: RealView: Add core-tile detection

This patch adds the core-tile detection and only enables devices if the
corresponding tile is present. It currently detects the ARM11MPCore via
the core_tile_eb11mp() macro.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] 4817/1: RealView: Move the AMBA resource definitions to realview_eb.c
Catalin Marinas [Mon, 4 Feb 2008 16:36:59 +0000 (17:36 +0100)]
[ARM] 4817/1: RealView: Move the AMBA resource definitions to realview_eb.c

This patch moves the IRQ and DMA definitions from core.h into
realview_eb.c since they are platform-specific. It adds a
realview_eb11mp_fixup function to adjust the IRQ numbers if the
ARM11MPCore tile is fitted. The realview_smc91x_device is also moved
from core.c into realview_eb.c.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] 4816/1: RealView: Move the platform-specific definitions into board-eb.h
Catalin Marinas [Mon, 4 Feb 2008 16:34:58 +0000 (17:34 +0100)]
[ARM] 4816/1: RealView: Move the platform-specific definitions into board-eb.h

This patch moves the platform specific definitions from platform.h into
the board-eb.h file. It drops the INT_* definitions as they are no
longer used in irqs.h (moved to board-eb.h). It renames REALVIEW_*
macros to REALVIEW_EB_* or REALVIEW_EB11MP_* to distinguish between
standard EB and EB + the ARM11MPCore tile. The platform.h file contains
common definitions to the RealView platforms and it is only directly
included in board-*.h files.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] 4815/1: RealView: Add clockevents suport for the local timers
Catalin Marinas [Mon, 4 Feb 2008 16:32:57 +0000 (17:32 +0100)]
[ARM] 4815/1: RealView: Add clockevents suport for the local timers

This patch registers the local timers on ARM11MPCore as clock event
devices. The clock device can be set up as periodic or oneshot.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] 4814/1: RealView: Add broadcasting clockevents support for ARM11MPCore
Catalin Marinas [Mon, 4 Feb 2008 16:30:57 +0000 (17:30 +0100)]
[ARM] 4814/1: RealView: Add broadcasting clockevents support for ARM11MPCore

This patch adds dummy local timers for each CPU so that the board clock
device is used to broadcast events to the other CPUs. The patch also
adds the declaration for the dummy_timer_setup function (the equivalent
of local_timer_setup when CONFIG_LOCAL_TIMERS is not set).

Due to the way clockevents work, the dummy timer on the first CPU has to
be registered before the board timer.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] 4813/1: Add SMP helper functions for clockevents support
Catalin Marinas [Mon, 4 Feb 2008 16:28:56 +0000 (17:28 +0100)]
[ARM] 4813/1: Add SMP helper functions for clockevents support

This patch adds the smp_call_function_single and smp_timer_broadcast
functions and modifies ipi_timer to call the platform-specific function
local_timer_interrupt.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] 4812/1: RealView: clockevents support for the RealView platforms
Catalin Marinas [Mon, 4 Feb 2008 16:26:55 +0000 (17:26 +0100)]
[ARM] 4812/1: RealView: clockevents support for the RealView platforms

The patch updates the RealView code to the clockevents infrastructure.
The SMP support is implemented in subsequent patches. Based on the
Versatile implementation by Kevin Hilman.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] 4811/1: RealView: clocksource support for the RealView platforms
Catalin Marinas [Mon, 4 Feb 2008 16:24:54 +0000 (17:24 +0100)]
[ARM] 4811/1: RealView: clocksource support for the RealView platforms

The patch updates the RealView platform code to use the generic
clocksource infrastructure for basic time keeping. Based on the
Versatile implementation by Kevin Hilman.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years agoMerge git://git.infradead.org/battery-2.6
Linus Torvalds [Sat, 2 Feb 2008 04:13:05 +0000 (15:13 +1100)]
Merge git://git.infradead.org/battery-2.6

* git://git.infradead.org/battery-2.6:
  apm_power: check I.intval for zero value, we use it as the divisor
  MAINTAINERS: remove kernel-discuss@handhelds.org list
  pda_power: implement polling
  pda_power: various cleanups
  apm_power: support using VOLTAGE_* properties for apm calculations
  pda_power: add suspend/resume support
  power_supply: add few more values and props
  pda_power: only register available psu
  power: fix incorrect unregistration in power_supply_create_attrs error path
  power: remove POWER_SUPPLY_PROP_CAPACITY_LEVEL
  [BATTERY] power_supply_leds: use kasprintf
  [BATTERY] Every file should include the headers containing the prototypes for its global functions.

16 years agoMerge branch 'for-linus' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Sat, 2 Feb 2008 03:31:28 +0000 (14:31 +1100)]
Merge branch 'for-linus' of git://linux-nfs.org/~bfields/linux

* 'for-linus' of git://linux-nfs.org/~bfields/linux: (100 commits)
  SUNRPC: RPC program information is stored in unsigned integers
  SUNRPC: Move exported symbol definitions after function declaration part 2
  NLM: tear down RPC clients in nlm_shutdown_hosts
  SUNRPC: spin svc_rqst initialization to its own function
  nfsd: more careful input validation in nfsctl write methods
  lockd: minor log message fix
  knfsd: don't bother mapping putrootfh enoent to eperm
  rdma: makefile
  rdma: ONCRPC RDMA protocol marshalling
  rdma: SVCRDMA sendto
  rdma: SVCRDMA recvfrom
  rdma: SVCRDMA Core Transport Services
  rdma: SVCRDMA Transport Module
  rdma: SVCRMDA Header File
  svc: Add svc_xprt_names service to replace svc_sock_names
  knfsd: Support adding transports by writing portlist file
  svc: Add svc API that queries for a transport instance
  svc: Add /proc/sys/sunrpc/transport files
  svc: Add transport hdr size for defer/revisit
  svc: Move the xprt independent code to the svc_xprt.c file
  ...

16 years agoMerge branch 'suspend' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux...
Linus Torvalds [Sat, 2 Feb 2008 03:29:57 +0000 (14:29 +1100)]
Merge branch 'suspend' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6

* 'suspend' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6: (38 commits)
  suspend: cleanup reference to swsusp_pg_dir[]
  PM: Remove obsolete /sys/devices/.../power/state docs
  Hibernation: Invoke suspend notifications after console switch
  Suspend: Invoke suspend notifications after console switch
  Suspend: Clean up suspend_64.c
  Suspend: Add config option to disable the freezer if architecture wants that
  ACPI: Print message before calling _PTS
  ACPI hibernation: Call _PTS before suspending devices
  Hibernation: Introduce begin() and end() callbacks
  ACPI suspend: Call _PTS before suspending devices
  ACPI: Separate disabling of GPEs from _PTS
  ACPI: Separate invocations of _GTS and _BFS from _PTS and _WAK
  Suspend: Introduce begin() and end() callbacks
  suspend: fix ia64 allmodconfig build
  ACPI: clear GPE earily in resume to avoid warning
  Suspend: Clean up Kconfig (V2)
  Hibernation: Clean up Kconfig (V2)
  Hibernation: Update messages
  Suspend: Use common prefix in messages
  Hibernation: Remove unnecessary variable declaration
  ...

16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/pci-2.6
Linus Torvalds [Sat, 2 Feb 2008 03:29:33 +0000 (14:29 +1100)]
Merge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/pci-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/pci-2.6: (64 commits)
  PCI: make pci_bus a struct device
  PCI: fix codingstyle issues in include/linux/pci.h
  PCI: fix codingstyle issues in drivers/pci/pci.h
  PCI: PCIE ASPM support
  PCI: Fix fakephp deadlock
  PCI: modify SB700 SATA MSI quirk
  PCI: Run ACPI _OSC method on root bridges only
  PCI ACPI: AER driver should only register PCIe devices with _OSC
  PCI ACPI: Added a function to register _OSC with only PCIe devices.
  PCI: constify function pointer tables
  PCI: Convert drivers/pci/proc.c to use unlocked_ioctl
  pciehp: block new requests from the device before power off
  pciehp: workaround against Bad DLLP during power off
  pciehp: wait for 1000ms before LED operation after power off
  PCI: Remove pci_enable_device_bars() from documentation
  PCI: Remove pci_enable_device_bars()
  PCI: Remove users of pci_enable_device_bars()
  PCI: Add pci_enable_device_{io,mem} intefaces
  PCI: avoid save the same type of cap multiple times
  PCI: correctly initialize a structure for pcie_save_pcix_state()
  ...

16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6
Linus Torvalds [Sat, 2 Feb 2008 03:28:57 +0000 (14:28 +1100)]
Merge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (128 commits)
  USB: fix codingstyle issues in drivers/usb/core/*.c
  USB: fix codingstyle issues in drivers/usb/core/message.c
  USB: fix codingstyle issues in drivers/usb/core/hcd-pci.c
  USB: fix codingstyle issues in drivers/usb/core/devio.c
  USB: fix codingstyle issues in drivers/usb/core/devices.c
  USB: fix codingstyle issues in drivers/usb/core/*.h
  USB: fix codingstyle issues in include/linux/usb/
  USB: fix codingstyle issues in include/linux/usb.h
  USB: mark USB drivers as being GPL only
  USB: use a real vendor and product id for root hubs
  USB: mount options: fix usbfs
  USB: Fix usb_serial_driver structure for Kobil cardreader driver.
  usb: ehci should use u16 for isochronous intervals
  usb: ehci, remove false clear-reset path
  USB: Use menuconfig objects
  usb: ohci-sm501 driver
  usb: dma bounce buffer support
  USB: last abuses of intfdata in close for usb-serial drivers
  USB: kl5kusb105 don't flush to logically disconnected devices
  USB: oti6858: cleanup
  ...

16 years agodebug: softlockup looping fix
Peter Zijlstra [Fri, 1 Feb 2008 23:23:08 +0000 (00:23 +0100)]
debug: softlockup looping fix

Rafael J. Wysocki reported weird, multi-seconds delays during
suspend/resume and bisected it back to:

  commit 82a1fcb90287052aabfa235e7ffc693ea003fe69
  Author: Ingo Molnar <mingo@elte.hu>
  Date:   Fri Jan 25 21:08:02 2008 +0100

      softlockup: automatically detect hung TASK_UNINTERRUPTIBLE tasks

fix it:

 - restore the old wakeup mechanism
 - fix break usage in do_each_thread() { } while_each_thread().
 - fix the hotplug switch stmt, a fall-through case was broken.

Bisected-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Tested-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoapm_power: check I.intval for zero value, we use it as the divisor
Anton Vorontsov [Sat, 12 Jan 2008 23:44:54 +0000 (02:44 +0300)]
apm_power: check I.intval for zero value, we use it as the divisor

Signed-off-by: Anton Vorontsov <cbou@mail.ru>
16 years agoMAINTAINERS: remove kernel-discuss@handhelds.org list
Anton Vorontsov [Sat, 12 Jan 2008 23:44:54 +0000 (02:44 +0300)]
MAINTAINERS: remove kernel-discuss@handhelds.org list

kernel-discuss at handhelds.org is down for months, and nobody
knows why. So remove it for now.

Signed-off-by: Anton Vorontsov <cbou@mail.ru>
16 years agopda_power: implement polling
Anton Vorontsov [Sat, 12 Jan 2008 23:44:20 +0000 (02:44 +0300)]
pda_power: implement polling

Signed-off-by: Anton Vorontsov <cbou@mail.ru>
16 years agopda_power: various cleanups
Anton Vorontsov [Sat, 12 Jan 2008 23:39:17 +0000 (02:39 +0300)]
pda_power: various cleanups

- handle spurious interrupts correctly;
- get rid of pda_power_supplies array, use two variables instead;
- factor out psy_changed() function, it will be used for polling.

Signed-off-by: Anton Vorontsov <cbou@mail.ru>
16 years agoapm_power: support using VOLTAGE_* properties for apm calculations
Dmitry Baryshkov [Sat, 12 Jan 2008 23:39:16 +0000 (02:39 +0300)]
apm_power: support using VOLTAGE_* properties for apm calculations

It's pretty dummy, but useful for batteries for which we can only
get voltages.

Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com>
Signed-off-by: Anton Vorontsov <cbou@mail.ru>
16 years agopda_power: add suspend/resume support
Dmitry Baryshkov [Sat, 12 Jan 2008 23:35:43 +0000 (02:35 +0300)]
pda_power: add suspend/resume support

Add suspend/resume/wakeup support for pda_power.

Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com>
Signed-off-by: Anton Vorontsov <cbou@mail.ru>
16 years agopower_supply: add few more values and props
Dmitry Baryshkov [Mon, 7 Jan 2008 01:12:41 +0000 (04:12 +0300)]
power_supply: add few more values and props

Add LiMn (one of the most common for small non-rechargable batteries)
battery technology and voltage_min/_max properties support.

Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com>
Signed-off-by: Anton Vorontsov <cbou@mail.ru>
16 years agopda_power: only register available psu
Dmitry Baryshkov [Mon, 7 Jan 2008 01:12:39 +0000 (04:12 +0300)]
pda_power: only register available psu

Currently pda-power adds both ac and usb power supply units.
This patch fixes it so that psu are added only if they are enabled.

Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com>
Signed-off-by: Anton Vorontsov <cbou@mail.ru>
16 years agopower: fix incorrect unregistration in power_supply_create_attrs error path
Andres Salomon [Wed, 12 Dec 2007 19:12:59 +0000 (14:12 -0500)]
power: fix incorrect unregistration in power_supply_create_attrs error path

In power_supply_create_attrs(), we create static attributes as referenced
by power_supply_static_attrs[i].  After that, if we fail, we unregister
via power_supply_static_attrs[psy->properties[i]].  This is incorrect, as
psy->properties has absolutely no bearing on static attribs.  This patch
fixes it to unregister the correct attrib.

Another line which was unnecessarily line wrapped is also unwrapped.

Signed-off-by: Andres Salomon <dilinger@debian.org>
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
16 years agopower: remove POWER_SUPPLY_PROP_CAPACITY_LEVEL
Andres Salomon [Wed, 12 Dec 2007 19:12:56 +0000 (14:12 -0500)]
power: remove POWER_SUPPLY_PROP_CAPACITY_LEVEL

The CAPACITY_LEVEL stuff defines various levels of charge; however, what
is the difference between them?  What differentiates between HIGH and NORMAL,
LOW and CRITICAL, etc?

As it appears that these are fairly arbitrary, we end up making such policy
decisions in the kernel (or in hardware).  This is the sort of decision that
should be made in userspace, not in the kernel.

If the hardware does not support _CAPACITY and it cannot be easily calculated,
then perhaps the driver should register a custom CAPACITY_LEVEL attribute;
however, userspace should not become accustomed to looking for such a thing,
and we should certainly not encourage drivers to provide CAPACITY_LEVEL
stubs.

The following removes support for POWER_SUPPLY_PROP_CAPACITY_LEVEL.  The
OLPC battery driver is the only driver making use of this, so it's
removed from there as well.

Signed-off-by: Andres Salomon <dilinger@debian.org>
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
16 years ago[BATTERY] power_supply_leds: use kasprintf
Akinobu Mita [Sat, 17 Nov 2007 10:55:58 +0000 (19:55 +0900)]
[BATTERY] power_supply_leds: use kasprintf

Use kasprintf instead of kmalloc()-strcpy()-strcat().

Cc: Anton Vorontsov <cbou@mail.ru>
Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Anton Vorontsov <cbou@mail.ru>
16 years ago[BATTERY] Every file should include the headers containing the prototypes for its...
Adrian Bunk [Sun, 25 Nov 2007 21:25:45 +0000 (00:25 +0300)]
[BATTERY] Every file should include the headers containing the prototypes for its global functions.

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Anton Vorontsov <cbou@mail.ru>
16 years agosuspend: cleanup reference to swsusp_pg_dir[]
Rafael J. Wysocki [Fri, 1 Feb 2008 14:28:16 +0000 (15:28 +0100)]
suspend: cleanup reference to swsusp_pg_dir[]

swsusp_pg_dir[] is used for suspend, but not for hibernation.
clean-up the ifdefs which worked by accident, while implying the opposite.
Delete the __nosavedata, which also implied the opposite.

Some day we may optimize CONFIG_ACPI_SLEEP to build minimal kernels
for just hibernate or just suspend but not both,
but today isn't that day.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoPM: Remove obsolete /sys/devices/.../power/state docs
David Brownell [Fri, 11 Jan 2008 00:26:57 +0000 (01:26 +0100)]
PM: Remove obsolete /sys/devices/.../power/state docs

The /sys/devices/.../power/state files have been gone for a while
now, but I just noticed some documentation that still refers to
them.  (Fortunately described as DEPRECATED and WILL REMOVE).

Time to remove that obsolete documentation too ...

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Invoke suspend notifications after console switch
Rafael J. Wysocki [Fri, 11 Jan 2008 00:25:21 +0000 (01:25 +0100)]
Hibernation: Invoke suspend notifications after console switch

Following the recent change in the suspend code path, switch consoles before
calling PM notifiers during hibernation.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoSuspend: Invoke suspend notifications after console switch
Johannes Berg [Fri, 11 Jan 2008 00:22:23 +0000 (01:22 +0100)]
Suspend: Invoke suspend notifications after console switch

In order to fix APM emulation it is necessary to enable apm-emulation
notifications for suspends triggered in various ways via the suspend
notifiers.  However, this will cause the systems using APM emulation
to lock up between X being needed to switch away from the VT and X
already waiting for resume in the APM ioctl.

This patch moves the console switch (if enabled) before the suspend
notification (and after the resume notification) to avoid this issue.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoSuspend: Clean up suspend_64.c
Borislav Petkov [Sun, 16 Dec 2007 23:30:22 +0000 (00:30 +0100)]
Suspend: Clean up suspend_64.c

There's a freakishly long comment in suspend_64.c, shorten it.

Signed-off-by: Borislav Petkov <bbpetkov@yahoo.de>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoSuspend: Add config option to disable the freezer if architecture wants that
Johannes Berg [Wed, 16 Jan 2008 04:17:00 +0000 (23:17 -0500)]
Suspend: Add config option to disable the freezer if architecture wants that

This patch makes the freezer optional for suspend to allow the
system to work (or not work) like the original PMU suspend.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoACPI: Print message before calling _PTS
Rafael J. Wysocki [Mon, 7 Jan 2008 23:10:57 +0000 (00:10 +0100)]
ACPI: Print message before calling _PTS

Make acpi_sleep_prepare() static and cause it to print a message
specifying the ACPI system sleep state to be entered (helpful for
debugging the suspend/hibernation code).

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoACPI hibernation: Call _PTS before suspending devices
Rafael J. Wysocki [Mon, 7 Jan 2008 23:09:58 +0000 (00:09 +0100)]
ACPI hibernation: Call _PTS before suspending devices

The ACPI 1.0 specification wants us to put devices into low power
states after executing the _PTS global control method, while ACPI
2.0 and later want us to do that in the reverse order.  The current
hibernation code follows ACPI 2.0 in that respect which may cause some
ACPI 1.0x systems to hang during hibernation (ref.
http://bugzilla.kernel.org/show_bug.cgi?id=9528).

Make the hibernation code execute _PTS before putting devices into
low power states (ie. in accordance with ACPI 1.0x) with the
possibility to override that using the 'acpi_new_pts_ordering' kernel
command line option.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Introduce begin() and end() callbacks
Rafael J. Wysocki [Mon, 7 Jan 2008 23:08:44 +0000 (00:08 +0100)]
Hibernation: Introduce begin() and end() callbacks

Introduce global hibernation callback .end() and rename global
hibernation callback .start() to .begin(), in analogy with the
recent modifications of the global suspend callbacks.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoACPI suspend: Call _PTS before suspending devices
Rafael J. Wysocki [Mon, 7 Jan 2008 23:07:39 +0000 (00:07 +0100)]
ACPI suspend: Call _PTS before suspending devices

The ACPI 1.0 specification wants us to put devices into low power
states after executing the _PTS global control method, while ACPI
2.0 and later want us to do that in the reverse order.  The current
suspend code follows ACPI 2.0 in that respect which causes some
ACPI 1.0x systems to hang during suspend (ref.
http://bugzilla.kernel.org/show_bug.cgi?id=9528).

Make the suspend code execute _PTS before putting devices into low
power states (ie. in accordance with ACPI 1.0x) and provide a command
line option to override the default if need be.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoACPI: Separate disabling of GPEs from _PTS
Rafael J. Wysocki [Mon, 7 Jan 2008 23:06:16 +0000 (00:06 +0100)]
ACPI: Separate disabling of GPEs from _PTS

The preparation to enter an ACPI system sleep state is now tied to
the disabling of GPEs, but the GPEs should not be disabled before
suspending devices.  Since on ACPI 1.0x systems the _PTS global
control method should be executed before suspending devices, we
need to disable GPEs separately.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoACPI: Separate invocations of _GTS and _BFS from _PTS and _WAK
Rafael J. Wysocki [Mon, 7 Jan 2008 23:05:21 +0000 (00:05 +0100)]
ACPI: Separate invocations of _GTS and _BFS from _PTS and _WAK

The execution of ACPI global control methods _GTS and _BFS is
currently tied to the preparation to enter a sleep state and to the
leaving of the sleep state, respectively.  However, these functions
are called before disabling the nonboot CPUs and after enabling
them, respectively (in fact, on ACPI 1.0x systems the first of them
ought to be called before suspending devices), while according to the
ACPI specification, _GTS is to be executed right prior to entering
the system sleep state and _BFS is to be executed right after the
platfor firmware has returned control to the OS on wake up.

Move the execution of _GTS and _BFS to the right places.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoSuspend: Introduce begin() and end() callbacks
Rafael J. Wysocki [Mon, 7 Jan 2008 23:04:17 +0000 (00:04 +0100)]
Suspend: Introduce begin() and end() callbacks

On ACPI systems the target state set by acpi_pm_set_target() is
reset by acpi_pm_finish(), but that need not be called if the
suspend fails. Â All platforms that use the .set_target() global
suspend callback are affected by analogous issues.

For this reason, we need an additional global suspend callback that
will reset the target state regardless of whether or not the suspend
is successful. Â Also, it is reasonable to rename the .set_target()
callback, since it will be used for a different purpose on ACPI
systems (due to ACPI 1.0x code ordering requirements).

Introduce the global suspend callback .end() to be executed at the
end of the suspend sequence and rename the .set_target() global
suspend callback to .begin().

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agosuspend: fix ia64 allmodconfig build
Rafael J. Wysocki [Fri, 14 Dec 2007 00:07:13 +0000 (01:07 +0100)]
suspend: fix ia64 allmodconfig build

kernel/power/main.c:488: error: â€˜pm_test_attr’ undeclared here (not in a function)

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoACPI: clear GPE earily in resume to avoid warning
Shaohua Li [Wed, 20 Jun 2007 01:17:58 +0000 (09:17 +0800)]
ACPI: clear GPE earily in resume to avoid warning

Wakeup GPE hasn't a handler. If system is waked up by such GPE like a
USB hotplug, I saw a lot of error reporting the GPE hasn't handler.
acpi_leave_sleep_state will clear the GPE but it's too late, we should
do it before interrupt is re-enabled.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoSuspend: Clean up Kconfig (V2)
Johannes Berg [Sat, 8 Dec 2007 01:14:00 +0000 (02:14 +0100)]
Suspend: Clean up Kconfig (V2)

This cleans up the suspend Kconfig and removes the need to
declare centrally which architectures support suspend. All
architectures that currently support suspend are modified
accordingly.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Russell King <rmk@arm.linux.org.uk>
Acked-by: Paul Mackerras <paulus@samba.org>
Acked-by: Ralf Baechle <ralf@linux-mips.org>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Cc: Pavel Machek <pavel@suse.cz>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Clean up Kconfig (V2)
Johannes Berg [Sat, 8 Dec 2007 01:12:39 +0000 (02:12 +0100)]
Hibernation: Clean up Kconfig (V2)

This cleans up the hibernation Kconfig and removes the need to
declare centrally which architectures support hibernation. All
architectures that currently support hibernation are modified
accordingly.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Paul Mackerras <paulus@samba.org>
Cc: Pavel Machek <pavel@suse.cz>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Update messages
Rafael J. Wysocki [Sat, 8 Dec 2007 01:09:43 +0000 (02:09 +0100)]
Hibernation: Update messages

Make hibernation messages start with one common prefix "PM: " and use
the word "hibernation" in the messages as a synonym of "suspend to
disk".

Turn some KERN_INFO messages into debug ones.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoSuspend: Use common prefix in messages
Rafael J. Wysocki [Sat, 8 Dec 2007 01:08:38 +0000 (02:08 +0100)]
Suspend: Use common prefix in messages

Make suspend messages start with one common prefix "PM: ".

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Remove unnecessary variable declaration
Rafael J. Wysocki [Sat, 8 Dec 2007 01:07:40 +0000 (02:07 +0100)]
Hibernation: Remove unnecessary variable declaration

Remove the unnecessary extern declaration of resume_file[]
from kernel/power/swap.c .

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Fix comment in disk.c
Rafael J. Wysocki [Sat, 8 Dec 2007 01:06:57 +0000 (02:06 +0100)]
Hibernation: Fix comment in disk.c

Fix a comment in kernel/power/disk.c so that it doesn't contain lines
longer that 80 characters.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoSuspend: Fix comment in main.c
Rafael J. Wysocki [Sat, 8 Dec 2007 01:06:00 +0000 (02:06 +0100)]
Suspend: Fix comment in main.c

Fix a comment in kernel/power/main.c so that it doesn't contain lines
longer that 80 characters.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Move low level resume to disk.c
Rafael J. Wysocki [Sat, 8 Dec 2007 01:04:21 +0000 (02:04 +0100)]
Hibernation: Move low level resume to disk.c

Move the low level restore code to kernel/power/disk.c , since the
corresponding low level hibernation code is already there.

Make restore fail if device_power_down(PMSG_PRETHAW) returns an
error.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoSuspend: Fix compilation warning for CONFIG_SUSPEND unset
Rafael J. Wysocki [Sat, 8 Dec 2007 01:03:26 +0000 (02:03 +0100)]
Suspend: Fix compilation warning for CONFIG_SUSPEND unset

Suspend: Make debug facility depend on CONFIG_SUSPEND

Make the new suspend debug facility code depend on CONFIG_SUSPEND,
as appropriate, to remove the compiler warning printed when CONFIG_PM is set
and CONFIG_SUSPEND is not set.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoPM: Convert PM notifiers to out-of-line code
Alan Stern [Mon, 19 Nov 2007 22:49:18 +0000 (23:49 +0100)]
PM: Convert PM notifiers to out-of-line code

This patch (as1008b) converts the PM notifier routines from inline
calls to out-of-line code.  It also prevents pm_chain_head from
being created when CONFIG_PM_SLEEP isn't enabled, and EXPORTs the
notifier registration and unregistration routines.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoPM: Make PM_TRACE more architecture independent
Johannes Berg [Mon, 19 Nov 2007 22:46:16 +0000 (23:46 +0100)]
PM: Make PM_TRACE more architecture independent

When trying to debug a suspend failure I started implementing
PM_TRACE for powerpc. I then noticed that I'm debugging a suspend
failure and so PM_TRACE isn't useful at all, but thought that
nonetheless this could be useful in the future.

Basically, to support PM_TRACE, you add a Kconfig option that
selects PM_TRACE and provides the infrastructure as per the
help text of PM_TRACE.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoPM: Suspend/hibernation debug documentation update (rev. 2)
Rafael J. Wysocki [Mon, 19 Nov 2007 22:43:34 +0000 (23:43 +0100)]
PM: Suspend/hibernation debug documentation update (rev. 2)

Update the suspend/hibernation debugging and testing documentation to describe
the newly introduced testing facilities.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: New testing facility (rev. 2)
Rafael J. Wysocki [Mon, 19 Nov 2007 22:42:31 +0000 (23:42 +0100)]
Hibernation: New testing facility (rev. 2)

Make it possible to test the hibernation core code with the help of the
/sys/power/pm_test attribute introduced for suspend testing in the previous
patch.

Writing an appropriate string to this file causes the hibernation code to work
in one of the test modes defined as follows:

freezer
- test the freezing of processes

devices
- test the freezing of processes and suspending of devices

platform
- test the freezing of processes, suspending of devices and platform global
  control methods(*)

processors
- test the freezing of processes, suspending of devices, platform global
  control methods(*) and the disabling of nonboot CPUs

core
- test the freezing of processes, suspending of devices, platform global
  control methods(*), the disabling of nonboot CPUs and suspending of
  platform/system devices

(*) - the platform global control methods are only available on ACPI systems
  Â  Â  and are only tested if the hibernation mode is set to "platform"

Then, if a hibernation is started by normal means, the hibernation core will
perform its normal operations up to the point indicated by given test level.
Next, it will wait for 5 seconds and carry out the resume operations needed to
transition the system back to the fully functional state.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agosuspend: build fix responding to 2.6.25 kset change
Rafael J. Wysocki [Mon, 28 Jan 2008 23:29:06 +0000 (00:29 +0100)]
suspend: build fix responding to 2.6.25 kset change

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoSuspend: Testing facility (rev. 2)
Rafael J. Wysocki [Mon, 19 Nov 2007 22:41:19 +0000 (23:41 +0100)]
Suspend: Testing facility (rev. 2)

Introduce sysfs attribute /sys/power/pm_test allowing one to test the suspend
core code. Â Namely, writing one of the strings:

freezer
devices
platform
processors
core

to this file causes the suspend code to work in one of the test modes defined as
follows:

freezer
- test the freezing of processes

devices
- test the freezing of processes and suspending of devices

platform
- test the freezing of processes, suspending of devices and platform global
  control methods(*)

processors
- test the freezing of processes, suspending of devices, platform global
  control methods and the disabling of nonboot CPUs

core
- test the freezing of processes, suspending of devices, platform global
  control methods, the disabling of nonboot CPUs and suspending of
  platform/system devices

(*) These are ACPI global control methods on ACPI systems

Then, if a suspend is started by normal means, the suspend core will perform
its normal operations up to the point indicated by given test level. Â Next, it
will wait for 5 seconds and carry out the resume operations needed to transition
the system back to the fully functional state.

Writing "none" to /sys/power/pm_test turns the testing off.

When open for reading, /sys/power/pm_test contains a space-separated list of all
available tests (including "none" that represents the normal functionality) in
which the current test level is indicated by square brackets.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Add PM_RESTORE_PREPARE and PM_POST_RESTORE notifiers (rev. 2)
Alan Stern [Mon, 19 Nov 2007 22:38:25 +0000 (23:38 +0100)]
Hibernation: Add PM_RESTORE_PREPARE and PM_POST_RESTORE notifiers (rev. 2)

Add PM_RESTORE_PREPARE and PM_POST_RESTORE notifiers to the PM core, to be used
in analogy with the existing PM_HIBERNATION_PREPARE and PM_POST_HIBERNATION
notifiers.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Move function prototypes to header
Adrian Bunk [Mon, 19 Nov 2007 22:36:20 +0000 (23:36 +0100)]
Hibernation: Move function prototypes to header

This patch moves the prototypes of count_highmem_pages() and
restore_highmem() to kernel/power/power.h

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoACPI: Fix mismerge in acpi_hibernation_finish
Rafael J. Wysocki [Mon, 19 Nov 2007 22:33:59 +0000 (23:33 +0100)]
ACPI: Fix mismerge in acpi_hibernation_finish

Some code in acpi_hibernation_finish() was moved to acpi_hibernation_leave(),
but the old copy had been left (it's harmless, but also useless).  Remove it.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Introduce exportable suspend ioctls header (rev. 2)
Rafael J. Wysocki [Thu, 25 Oct 2007 23:05:05 +0000 (01:05 +0200)]
Hibernation: Introduce exportable suspend ioctls header (rev. 2)

Move the definitions of hibernation ioctls to a separate header file in
include/linux, which can be exported to the user space.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Correct definitions of some ioctls (rev. 2)
Rafael J. Wysocki [Thu, 25 Oct 2007 23:03:33 +0000 (01:03 +0200)]
Hibernation: Correct definitions of some ioctls (rev. 2)

Three ioctl numbers belonging to the hibernation userland interface,
SNAPSHOT_ATOMIC_SNAPSHOT, SNAPSHOT_AVAIL_SWAP, SNAPSHOT_GET_SWAP_PAGE,
are defined in a wrong way (eg. not portable).  Provide new ioctl numbers for
these ioctls and mark the existing ones as deprecated.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Mark SNAPSHOT_SET_SWAP_FILE ioctl as deprecated (rev. 2)
Rafael J. Wysocki [Thu, 25 Oct 2007 23:02:15 +0000 (01:02 +0200)]
Hibernation: Mark SNAPSHOT_SET_SWAP_FILE ioctl as deprecated (rev. 2)

Mark the SNAPSHOT_SET_SWAP_FILE ioctl belonging to the hibernation userland
interface as deprecated.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Rework platform support ioctls (rev. 2)
Rafael J. Wysocki [Thu, 25 Oct 2007 23:01:10 +0000 (01:01 +0200)]
Hibernation: Rework platform support ioctls (rev. 2)

Modify the hibernation userland interface by adding two new ioctls to it,
SNAPSHOT_PLATFORM_SUPPORT and SNAPSHOT_POWER_OFF, that can be used,
respectively, to switch the hibernation platform support on/off and to make the
kernel transition the system to the hibernation state (eg. ACPI S4) using the
platform (eg. ACPI) driver.

These ioctls are intended to replace the misdesigned SNAPSHOT_PMOPS ioctl,
which from now is regarded as obsolete and will be removed in the future.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoHibernation: Introduce SNAPSHOT_GET_IMAGE_SIZE ioctl
Rafael J. Wysocki [Thu, 25 Oct 2007 22:59:31 +0000 (00:59 +0200)]
Hibernation: Introduce SNAPSHOT_GET_IMAGE_SIZE ioctl

Add a new ioctl, SNAPSHOT_GET_IMAGE_SIZE, returning the size of the (just
created) hibernation image, to the hibernation userland interface.

This ioctl is necessary so that the userland utilities using the interface need
not access the hibernation image header, owned by the kernel, in order to obtain
the size of the image.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Len Brown <len.brown@intel.com>
16 years agoPCI: make pci_bus a struct device
Greg Kroah-Hartman [Wed, 23 May 2007 02:47:54 +0000 (22:47 -0400)]
PCI: make pci_bus a struct device

This moves the pci_bus class device to be a real struct device and at
the same time, place it in the device tree in the correct location.

Note, the old "bridge" symlink is now gone, but this was a non-standard
link and no userspace program used it.  If you need to determine the
device that the bus is on, follow the standard device symlink, or walk
up the device tree.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: fix codingstyle issues in include/linux/pci.h
Greg Kroah-Hartman [Wed, 30 Jan 2008 23:21:33 +0000 (15:21 -0800)]
PCI: fix codingstyle issues in include/linux/pci.h

Fixes a number of coding style issues in pci.h.
It's a tad more readable now...

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: fix codingstyle issues in drivers/pci/pci.h
Greg Kroah-Hartman [Wed, 30 Jan 2008 23:21:33 +0000 (15:21 -0800)]
PCI: fix codingstyle issues in drivers/pci/pci.h

Fixes a few coding style issues in the internal pci.h file

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: PCIE ASPM support
Shaohua Li [Thu, 24 Jan 2008 02:21:57 +0000 (10:21 +0800)]
PCI: PCIE ASPM support

PCI Express ASPM defines a protocol for PCI Express components in the D0
state to reduce Link power by placing their Links into a low power state
and instructing the other end of the Link to do likewise. This
capability allows hardware-autonomous, dynamic Link power reduction
beyond what is achievable by software-only controlled power management.
However, The device should be configured by software appropriately.
Enabling ASPM will save power, but will introduce device latency.

This patch adds ASPM support in Linux. It introduces a global policy for
ASPM, a sysfs file /sys/module/pcie_aspm/parameters/policy can control
it. The interface can be used as a boot option too. Currently we have
below setting:
        -default, BIOS default setting
        -powersave, highest power saving mode, enable all available ASPM
state
and clock power management
        -performance, highest performance, disable ASPM and clock power
management
By default, the 'default' policy is used currently.

In my test, power difference between powersave mode and performance mode
is about 1.3w in a system with 3 PCIE links.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: Fix fakephp deadlock
Ian Abbott [Fri, 25 Jan 2008 16:23:56 +0000 (16:23 +0000)]
PCI: Fix fakephp deadlock

If the fakephp driver is used to emulate removal of a PCI device by
writing text string "0" to the "power" sysfs attribute file, this causes
its parent directory and its contents (including the "power" file) to be
deleted before the write operation returns.  Unfortunately, it ends up
in a deadlock waiting for itself to complete.

The deadlock is as follows: sysfs_write_file calls flush_write_buffer
which calls sysfs_get_active_two before calling power_write_file in
pci_hotplug_core.c via the sysfs store operation. The power_write_file
function calls disable_slot in fakephp.c via the slot operation.  The
disable_slot function calls remove_slot which calls pci_hp_deregister
(back in pci_hotplug_core.c) which calls fs_remove_slot which calls
sysfs_remove_file to remove the "power" file. The sysfs_remove_file
function calls sysfs_hash_and_remove which calls sysfs_addrm_finish
which calls sysfs_deactivate. The sysfs_deactivate function sees that
something has an active reference on the sysfs_dirent (from the
previous call to sysfs_get_active_two back up the call stack somewhere)
so waits for the active reference to go away, which is of course
impossible.

The problem has been present since 2.6.21.

This patch breaks the deadlock by queuing work queue items on a single-
threaded work queue to remove a slot from sysfs, and to rescan the PCI
buses.  There is also some protection against disabling a slot that is
already being removed.

Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Cc: Kristen Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: modify SB700 SATA MSI quirk
Shane Huang [Fri, 25 Jan 2008 06:46:24 +0000 (15:46 +0900)]
PCI: modify SB700 SATA MSI quirk

SB700 SATA MSI bug will be fixed in SB700 revision A21 at hardware
level, but the SB700 revision older than A21 will also be found in the
market.  This patch modify the original quirk commit
bc38b411fe696fad32b261f492cb4afbf1835256 instead of withdrawing it.
The patch also removes quirk to 0x4395 because 0x4395 is SB800 device
ID.

Signed-off-by: Shane Huang <shane.huang@amd.com>
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: Run ACPI _OSC method on root bridges only
Andrew Patterson [Wed, 23 Jan 2008 00:18:27 +0000 (17:18 -0700)]
PCI: Run ACPI _OSC method on root bridges only

According to the PCI Firmware Specification Revision 3.0 section 4.5, _OSC
should only be called on a root brdige.  Here is the relevant passage: "The
_OSC interface defined in this section applies only to Host Bridge ACPI
devices that originate PCI, PCI-X, or PCI Express hierarchies". Changed the
code to find the parent root bridge of the device and call _OSC on that.

Signed-off-by: Andrew Patterson <andrew.patterson@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI ACPI: AER driver should only register PCIe devices with _OSC
Andrew Patterson [Wed, 23 Jan 2008 00:18:17 +0000 (17:18 -0700)]
PCI ACPI: AER driver should only register PCIe devices with _OSC

AER is only used with PCIe devices so we should only check PCIe devices for
_OSC support.

Signed-off-by: Andrew Patterson <andrew.patterson@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI ACPI: Added a function to register _OSC with only PCIe devices.
Andrew Patterson [Wed, 23 Jan 2008 00:18:12 +0000 (17:18 -0700)]
PCI ACPI: Added a function to register _OSC with only PCIe devices.

The function pci_osc_support_set() traverses every root bridge when
checking for _OSC support for a capability.  It quits as soon as it finds a
device/bridge that doesn't support the requested capability. This won't
work for systems that have mixed PCI and PCIe bridges when checking for
PCIe features.  I split this function into two -- pci_osc_support_set() and
pcie_osc_support_set(). The latter is used when only PCIe devices should be
traversed.

Signed-off-by: Andrew Patterson <andrew.patterson@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: constify function pointer tables
Jan Engelhardt [Tue, 22 Jan 2008 19:53:43 +0000 (20:53 +0100)]
PCI: constify function pointer tables

Signed-off-by: Jan Engelhardt <jengelh@computergmbh.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: Convert drivers/pci/proc.c to use unlocked_ioctl
Mathieu Segaud [Thu, 10 Jan 2008 13:27:12 +0000 (14:27 +0100)]
PCI: Convert drivers/pci/proc.c to use unlocked_ioctl

Change access to inode thru file->f_dentry->d_inode, and add explicit
lock/unlock_kernel() calls.

Signed-off-by: Mathieu Segaud <mathieu.segaud@regala.cx>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agopciehp: block new requests from the device before power off
Kenji Kaneshige [Thu, 20 Dec 2007 10:46:33 +0000 (19:46 +0900)]
pciehp: block new requests from the device before power off

Disable Bus Master, SERR# and INTx to ensure that no new Requests will
be generated from the device before turning power off, in accordance
with the specification.

Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agopciehp: workaround against Bad DLLP during power off
Kenji Kaneshige [Thu, 20 Dec 2007 10:45:09 +0000 (19:45 +0900)]
pciehp: workaround against Bad DLLP during power off

Set Bad DLLP Mask bit in Correctable Error Mask Register during
turning power off the slot.

This is the workaround against Bad DLLP error that sometimes happen
during turning power off on the slot which conforms to PCI Express
1.0a spec. The cause of this error seems that PCI Express 1.0a spec
doesn't have the following consideration that was added to PCI Express
1.1 spec.

    "If the port is associated with a hot-pluggable slot (Hot-Plug
    Capable bit in the Slot Capabilities register set to 1b), and
    Power Controller Control bit in Slot Control register is 1b(Off),
    then any transition to DL Inactive must not be considered an
    error."

Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agopciehp: wait for 1000ms before LED operation after power off
Kenji Kaneshige [Thu, 20 Dec 2007 10:43:56 +0000 (19:43 +0900)]
pciehp: wait for 1000ms before LED operation after power off

After turning power off, we must wait for at least 1 second *before*
LED operation.

Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: Remove pci_enable_device_bars() from documentation
Grant Grundler [Mon, 24 Dec 2007 07:08:51 +0000 (00:08 -0700)]
PCI: Remove pci_enable_device_bars() from documentation

Patch below removes pci_enable_device_bars() from Documentation/pci.txt .

Signed-off-by: Grant Grundler <grundler@parisc-linux.org>
Acked-by: Alan Cox <alan@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: Remove pci_enable_device_bars()
Benjamin Herrenschmidt [Thu, 20 Dec 2007 04:28:10 +0000 (15:28 +1100)]
PCI: Remove pci_enable_device_bars()

Now that all in-tree users are gone, this removes pci_enable_device_bars()
completely.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: Remove users of pci_enable_device_bars()
Benjamin Herrenschmidt [Thu, 20 Dec 2007 04:28:09 +0000 (15:28 +1100)]
PCI: Remove users of pci_enable_device_bars()

This patch converts users of pci_enable_device_bars() to the new
pci_enable_device_{io,mem} interface.

The new API fits nicely, except maybe for the QLA case where a bit of
code re-organization might be a good idea but I prefer sticking to the
simple patch as I don't have hardware to test on.

I'll also need some feedback on the cs5520 change.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: Add pci_enable_device_{io,mem} intefaces
Benjamin Herrenschmidt [Thu, 20 Dec 2007 04:28:08 +0000 (15:28 +1100)]
PCI: Add pci_enable_device_{io,mem} intefaces

The pci_enable_device_bars() interface isn't well suited to PCI
because you can't actually enable/disable BARs individually on
a device. So for example, if a device has 2 memory BARs 0 and 1,
and one of them (let's say 1) has not been successfully allocated
by the firmware or the kernel, then enabling memory decoding
shouldn't be permitted for the entire device since it will decode
whatever random address is still in that BAR 1.

So a device must be either fully enabled for IO, for Memory, or
for both. Not on a per-BAR basis.

This provides two new functions, pci_enable_device_io() and
pci_enable_device_mem() to replace pci_enable_device_bars(). The
implementation internally builds a BAR mask in order to be able
to use existing arch infrastructure.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: avoid save the same type of cap multiple times
Shaohua Li [Tue, 18 Dec 2007 01:57:09 +0000 (09:57 +0800)]
PCI: avoid save the same type of cap multiple times

Avoid adding the same type of cap multiple times, otherwise we will see dead loop.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: correctly initialize a structure for pcie_save_pcix_state()
Shaohua Li [Tue, 18 Dec 2007 01:56:56 +0000 (09:56 +0800)]
PCI: correctly initialize a structure for pcie_save_pcix_state()

save_state->cap_nr should be correctly set, otherwise we can't find the
saved cap at resume.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: fix typo in pci_save_pcix_state
Shaohua Li [Tue, 18 Dec 2007 01:56:47 +0000 (09:56 +0800)]
PCI: fix typo in pci_save_pcix_state

pci_save/store_state has multiple bugs, which will cause cap can't be
saved/restored correctly. Below 3 patches fix them.

fix the typo in pci_save_pcix_state

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: use dev_printk in x86 quirk messages
bjorn.helgaas@hp.com [Mon, 17 Dec 2007 21:09:40 +0000 (14:09 -0700)]
PCI: use dev_printk in x86 quirk messages

Convert quirk printks to dev_printk().

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: use dev_printk in quirk messages
bjorn.helgaas@hp.com [Mon, 17 Dec 2007 21:09:39 +0000 (14:09 -0700)]
PCI: use dev_printk in quirk messages

Convert quirk printks to dev_printk().

I made the MSI disable messages a little more consistent:

    - always use "disabled", not "deactivated"
    - specify "device MSI disabled" or "subordinate MSI disabled" when
      disabling MSI for only a specific device or subordinate bus

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: print quirk name in debug messages
bjorn.helgaas@hp.com [Mon, 17 Dec 2007 21:09:38 +0000 (14:09 -0700)]
PCI: print quirk name in debug messages

Instead of printing this:

    PCI: Calling quirk c023b250 for 0000:00:00.0

we can print this:

    pci 0000:00:00.0: calling quirk 0xc023b270: quirk_cardbus_legacy+0x0/0x30()

The address is superfluous because sprint_symbol() includes the
address if the symbol lookup fails, but this is the same style used
in do_initcalls() and pnp_fixup_device().

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: fix for quirk_e100_interrupt()
Ivan Kokshaysky [Mon, 17 Dec 2007 21:39:27 +0000 (00:39 +0300)]
PCI: fix for quirk_e100_interrupt()

Check that the e100 is in the D0 power state. If it's not, it won't
respond to MMIO accesses and we end up with master-abort machine
checks on some platforms.

Signed-off-by: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Auke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: Spelling fixes
Joe Perches [Mon, 17 Dec 2007 19:40:11 +0000 (11:40 -0800)]
PCI: Spelling fixes

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: Kconfig help: don't refer to the PCI-HOWTO
Adrian Bunk [Thu, 13 Dec 2007 18:45:17 +0000 (19:45 +0100)]
PCI: Kconfig help: don't refer to the PCI-HOWTO

A HOWTO that hasn't been updated for half a dozen years no longer
"contains valuable information about which PCI hardware does work under
Linux and which doesn't".

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: drivers/pci/msi.c: move arch hooks to the top
Adrian Bunk [Tue, 11 Dec 2007 22:19:41 +0000 (23:19 +0100)]
PCI: drivers/pci/msi.c: move arch hooks to the top

This patch fixes the following problem present with older gcc versions:

<--  snip  -->

...
  CC      drivers/pci/msi.o
/home/bunk/linux/kernel-2.6/git/linux-2.6/drivers/pci/msi.c:692: warning: weak declaration of `arch_msi_check_device' after first use results in unspecified behavior
/home/bunk/linux/kernel-2.6/git/linux-2.6/drivers/pci/msi.c:704: warning: weak declaration of `arch_setup_msi_irqs' after first use results in unspecified behavior
/home/bunk/linux/kernel-2.6/git/linux-2.6/drivers/pci/msi.c:724: warning: weak declaration of `arch_teardown_msi_irqs' after first use results in unspecified behavior
...

<--  snip  -->

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: quirk: enable MSI Mapping on HT1000
Sebastien Dugue [Fri, 14 Dec 2007 00:09:25 +0000 (16:09 -0800)]
PCI: quirk: enable MSI Mapping on HT1000

Add a quirk to enable the MSI mapping capability on HyperTransport bridges.

Wire Broadcom's HT1000 to use the quirk.

Signed-off-by: Sebastien Dugue <sebastien.dugue@bull.net>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Andy Currid <acurrid@nvidia.com>
Cc: Peer Chen <pchen@nvidia.com>
Cc: Prakash Punnoor <prakash@punnoor.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Michael Chan <mchan@broadcom.com>
Cc: Greg KH <greg@kroah.com>
Cc: Andy Currid <acurrid@nvidia.com>
Cc: Peer Chen <pchen@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: Fix warning in setup-res.c on 32-bit platforms with 64-bit resources
Benjamin Herrenschmidt [Mon, 10 Dec 2007 06:32:16 +0000 (17:32 +1100)]
PCI: Fix warning in setup-res.c on 32-bit platforms with 64-bit resources

This adds appropriate casts to avoid a warning and print the correct
values in pr_debug.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: Fix bus resource assignment on 32 bits with 64b resources
Benjamin Herrenschmidt [Mon, 10 Dec 2007 06:32:15 +0000 (17:32 +1100)]
PCI: Fix bus resource assignment on 32 bits with 64b resources

The current pci_assign_unassigned_resources() code doesn't work properly
on 32 bits platforms with 64 bits resources. The main reason is the use
of unsigned long in various places instead of resource_size_t.

This is a pre-requisite for making powerpc use the generic code instead of
its own half-useful implementation.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI Hotplug: PCIeHP: Fix some whitespace damage
Rolf Eike Beer [Tue, 6 Nov 2007 12:20:17 +0000 (13:20 +0100)]
PCI Hotplug: PCIeHP: Fix some whitespace damage

PCIeHP: Fix some whitespace damage

Signed-off-by: Rolf Eike Beer <eike-hotplug@sf-tec.de>
Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI Hotplug: pciehp: use generic function to find ext capability
Kenji Kaneshige [Fri, 9 Nov 2007 08:28:11 +0000 (17:28 +0900)]
PCI Hotplug: pciehp: use generic function to find ext capability

Remove duplicated code to find an extend capability in PCIEHP driver.

Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI Hotplug: pciehp: remove needless hp_slot calculation
Kenji Kaneshige [Fri, 9 Nov 2007 08:29:42 +0000 (17:29 +0900)]
PCI Hotplug: pciehp: remove needless hp_slot calculation

Remove needless hp_slot calculation. This has no functional changes.

Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>