]> git.karo-electronics.de Git - karo-tx-linux.git/log
karo-tx-linux.git
11 years agoARM: dove: remove legacy pcie and clock init
Sebastian Hesselbarth [Mon, 12 Aug 2013 18:46:55 +0000 (20:46 +0200)]
ARM: dove: remove legacy pcie and clock init

With DT support for PCIe controllers on Dove, we can now remove the
legacy pcie init and the last remaining clock workaround for its clocks.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: add initial DT file for Globalscale D3Plug
Sebastian Hesselbarth [Mon, 12 Aug 2013 18:46:54 +0000 (20:46 +0200)]
ARM: dove: add initial DT file for Globalscale D3Plug

This adds an initial DT file for the Globalscale D3Plug with Dove SoC.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: add PCIe controllers to SoC DT
Sebastian Hesselbarth [Mon, 12 Aug 2013 18:46:53 +0000 (20:46 +0200)]
ARM: dove: add PCIe controllers to SoC DT

This adds a node for the pcie controllers found on Dove SoCs to the
SoC DT include.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: switch to DT probed mbus address windows
Sebastian Hesselbarth [Mon, 29 Jul 2013 12:31:54 +0000 (14:31 +0200)]
ARM: dove: switch to DT probed mbus address windows

With proper mbus ranges and all internal nodes moved over, we can now
switch from static address window allocation to DT probed allocation.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: relocate internal registers device nodes
Sebastian Hesselbarth [Mon, 29 Jul 2013 12:31:53 +0000 (14:31 +0200)]
ARM: dove: relocate internal registers device nodes

With mbus node in place, now relocate all internal device nodes
to internal-regs node with proper address ranges.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: add MBus DT node
Sebastian Hesselbarth [Mon, 29 Jul 2013 12:31:52 +0000 (14:31 +0200)]
ARM: dove: add MBus DT node

This adds a MBus node including ranges and pcie apertures required later.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: add MBUS_ID macro to Dove DT
Sebastian Hesselbarth [Mon, 29 Jul 2013 12:31:51 +0000 (14:31 +0200)]
ARM: dove: add MBUS_ID macro to Dove DT

This adds a macro used for defining address window's target ID and
attribute cells for the MBus ranges entry.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: use preprocessor on device tree files
Sebastian Hesselbarth [Mon, 29 Jul 2013 12:31:50 +0000 (14:31 +0200)]
ARM: dove: use preprocessor on device tree files

This coverts Dove DT board files to preprocessor includes instead
of dtc includes.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: mvebu: add support for Marvell Dove SoCs
Sebastian Hesselbarth [Tue, 13 Aug 2013 12:25:24 +0000 (14:25 +0200)]
PCI: mvebu: add support for Marvell Dove SoCs

This patch adds a compatible for the PCIe controller found on Marvell
Dove SoCs. Binding documentation and Kconfig entry are also updated.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: mvebu: add support for reset on GPIO
Sebastian Hesselbarth [Tue, 13 Aug 2013 12:25:23 +0000 (14:25 +0200)]
PCI: mvebu: add support for reset on GPIO

This patch adds a check for DT passed reset-gpios property and deasserts/
asserts reset pin on probe/remove with configurable delay. Corresponding
binding documentation is also updated.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: mvebu: remove subsys_initcall
Sebastian Hesselbarth [Tue, 13 Aug 2013 12:25:22 +0000 (14:25 +0200)]
PCI: mvebu: remove subsys_initcall

This removes the subsys_initcall from the driver and converts it to
a normal platform_driver. Also, drvdata is set and a remove functions
is added to disable the clock and free resources. As pci driver removal
currently is not supported, set .suppress_bind_attrs to permit unbinding.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: mvebu: increment nports only for registered ports
Sebastian Hesselbarth [Tue, 13 Aug 2013 12:25:21 +0000 (14:25 +0200)]
PCI: mvebu: increment nports only for registered ports

The number of ports is probed by counting the number of available child nodes.
Later on, the registration of a port can fail and cause a mismatch between
the ->nports counter and registered ports. This patch modifies the counting
strategy, to make ->nports represent the number of registered ports instead
of the number of available childs.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: mvebu: move clock enable before register access
Sebastian Hesselbarth [Tue, 13 Aug 2013 12:25:20 +0000 (14:25 +0200)]
PCI: mvebu: move clock enable before register access

The clock passed to PCI controller found on MVEBU SoCs may come from a
clock gate. This requires the clock to be enabled before any registers
are accessed. Therefore, move the clock enable before register iomap to
ensure it is enabled.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoMerge branch 'mvebu/msi_mvebu' into mvebu/dove_pcie-mbus-dt
Jason Cooper [Tue, 13 Aug 2013 18:19:45 +0000 (18:19 +0000)]
Merge branch 'mvebu/msi_mvebu' into mvebu/dove_pcie-mbus-dt

Conflicts:
Documentation/devicetree/bindings/pci/mvebu-pci.txt
arch/arm/boot/dts/armada-370.dtsi
arch/arm/boot/dts/armada-xp-mv78230.dtsi
arch/arm/boot/dts/armada-xp-mv78260.dtsi
arch/arm/boot/dts/armada-xp-mv78460.dtsi
drivers/pci/host/pci-mvebu.c

11 years agoMerge branch 'mvebu/mbus_dt' into mvebu/dove_pcie-mbus-dt
Jason Cooper [Tue, 13 Aug 2013 18:19:21 +0000 (18:19 +0000)]
Merge branch 'mvebu/mbus_dt' into mvebu/dove_pcie-mbus-dt

Conflicts:
arch/arm/boot/dts/kirkwood.dtsi

11 years agoARM: mvebu: link PCIe controllers to the MSI controller
Thomas Petazzoni [Fri, 9 Aug 2013 20:27:15 +0000 (22:27 +0200)]
ARM: mvebu: link PCIe controllers to the MSI controller

This commit adjusts the Armada 370 and Armada XP PCIe controllers
Device Tree informations to reference their MSI controller. In the
case of this platform, the MSI controller is implemented by the MPIC.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Daniel Price <daniel.price@gmail.com>
Acked-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: mvebu: add support for MSI
Thomas Petazzoni [Fri, 9 Aug 2013 20:27:14 +0000 (22:27 +0200)]
PCI: mvebu: add support for MSI

This commit adds support for Message Signaled Interrupts in the
Marvell PCIe host controller. The work is very simple: it simply gets
a reference to the msi_chip associated to the PCIe controller thanks
to the msi-parent DT property, and stores this reference in the
pci_bus structure. This is enough to let the Linux PCI core use the
functions of msi_chip to setup and teardown MSIs.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Thierry Reding <thierry.reding@gmail.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: the MPIC now provides MSI controller features
Thomas Petazzoni [Fri, 9 Aug 2013 20:27:13 +0000 (22:27 +0200)]
ARM: mvebu: the MPIC now provides MSI controller features

Adds the 'msi-controller' property to the main interrupt controller
Device Tree node, to indicate that it can now behave as a MSI
controller.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Thierry Reding <thierry.reding@gmail.com>
Tested-by: Daniel Price <daniel.price@gmail.com>
Acked-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoirqchip: armada-370-xp: implement MSI support
Thomas Petazzoni [Fri, 9 Aug 2013 20:27:11 +0000 (22:27 +0200)]
irqchip: armada-370-xp: implement MSI support

This commit introduces the support for the MSI interrupts in the
armada-370-xp interrupt controller driver. It registers an MSI chip to
the MSI chip registry, which will be used by the Marvell PCIe host
controller driver.

The MSI interrupts use the 16 high doorbells, and are therefore
notified using IRQ1 of the main interrupt controller.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoirqchip: armada-370-xp: properly request resources
Thomas Petazzoni [Fri, 9 Aug 2013 20:27:10 +0000 (22:27 +0200)]
irqchip: armada-370-xp: properly request resources

Instead of using of_iomap(), we now use of_address_to_resource(),
request_mem_region() and ioremap(). This allows the corresponding I/O
regions to be properly requested and visible in /proc/iomem.

The main motivation for this change is that the introduction of the
MSI support requires us to get the physical address of the main
interrupt controller registers, so we will need the corresponding
'struct resource' anyway.

We also take this opportunity to change a panic() to BUG_ON(), in
order to be consistent with the rest of the driver.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Daniel Price <daniel.price@gmail.com>
Acked-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: msi: add default MSI operations for !HAVE_GENERIC_HARDIRQS platforms
Thomas Petazzoni [Tue, 13 Aug 2013 08:11:42 +0000 (10:11 +0200)]
PCI: msi: add default MSI operations for !HAVE_GENERIC_HARDIRQS platforms

Some platforms (e.g S390) don't use the generic hardirqs code and
therefore do not defined HAVE_GENERIC_HARDIRQS. This prevents using
the irq_set_chip_data() and irq_get_chip_data() functions that are
used for the default implementations of the MSI operations.

So, when CONFIG_GENERIC_HARDIRQS is not enabled, provide another
default implementation of the MSI operations, that simply errors
out. The architecture is responsible for implementing those operations
(which is the case on S390), and cannot use the msi_chip infrastructure.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: add initial DT file for Globalscale D2Plug
Sebastian Hesselbarth [Mon, 29 Jul 2013 12:29:06 +0000 (14:29 +0200)]
ARM: dove: add initial DT file for Globalscale D2Plug

This adds an initial DT file for the Globalscale D2Plug with Dove SoC.
Currently, one LED is missing and I have not been able to get SD8787 driver
working. Those will be taken care of later.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: add GPIO IR receiver node to SolidRun CuBox
Sebastian Hesselbarth [Mon, 29 Jul 2013 12:29:05 +0000 (14:29 +0200)]
ARM: dove: add GPIO IR receiver node to SolidRun CuBox

This adds a node for the IR receiver connected to a GPIO pin on the
SolidRun CuBox.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: add common pinmux functions to DT
Sebastian Hesselbarth [Mon, 29 Jul 2013 12:29:04 +0000 (14:29 +0200)]
ARM: dove: add common pinmux functions to DT

This adds common dedicated and gpio pinmux functions to SoC pinctrl
node. It also relocates pinctrl references to corresponding DT nodes.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: add cpu device tree node
Sebastian Hesselbarth [Mon, 29 Jul 2013 12:29:03 +0000 (14:29 +0200)]
ARM: dove: add cpu device tree node

This adds a node for the Marvell Sheeva PJ4A CPU found on Dove SoCs.
While at it, also move the l2-cache node out of internal registers and
consistently name different nodes.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
Conflicts:
arch/arm/boot/dts/dove.dtsi

11 years agoarch/arm/mach-kirkwood: Avoid using ARRAY_AND_SIZE(e) as a function argument
Julia Lawall [Sun, 11 Aug 2013 16:51:56 +0000 (18:51 +0200)]
arch/arm/mach-kirkwood: Avoid using ARRAY_AND_SIZE(e) as a function argument

Replace ARRAY_AND_SIZE(e) in function argument position to avoid hiding the
arity of the called function.

The semantic match that makes this change is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression e,f;
@@

f(...,
- ARRAY_AND_SIZE(e)
+ e,ARRAY_SIZE(e)
  ,...)
// </smpl>

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: fix DT building and update defconfig
Jason Cooper [Mon, 12 Aug 2013 18:05:30 +0000 (18:05 +0000)]
ARM: kirkwood: fix DT building and update defconfig

commit

  844703c ARM: kirkwood: Remove all remaining trace of DNS-320/325 platform code

removed the last kirkwood DT/platform board in kirkwood_defconfig, this
triggered a rescan of the config symbols, the result of which breaks the
build because KIRKWOOD_DT does not select OF_IRQ.

This commit fixes the regression and updates the defconfig to
explicitly enable SERIAL_OF_PLATFORM and KIRWOOD_DT.

Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: Remove all remaining trace of DNS-320/325 platform code
Jamie Lentin [Sun, 11 Aug 2013 14:26:50 +0000 (15:26 +0100)]
ARM: kirkwood: Remove all remaining trace of DNS-320/325 platform code

No platform code left now, so remove all references to it.

The reset gpio code didn't work to begin with, it will be replaced with
a DT node at a future date. -jc

Signed-off-by: Jamie Lentin <jm@lentin.co.uk>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agomemory: mvebu-devbus: Remove unused variable
Ezequiel Garcia [Sat, 10 Aug 2013 13:05:14 +0000 (10:05 -0300)]
memory: mvebu-devbus: Remove unused variable

This variable is not being used anywhere and it's only forgotten
garbage that should have been removed in the previous commit:

  commit 9b6e4c0a58e24c28bd757c9365824a37e80b751c
  Author: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
  Date:   Fri Jul 26 10:17:38 2013 -0300

  memory: mvebu-devbus: Remove address decoding window workaround

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: pci: add ->add_bus() and ->remove_bus() hooks to hw_pci
Thomas Petazzoni [Fri, 9 Aug 2013 20:27:12 +0000 (22:27 +0200)]
ARM: pci: add ->add_bus() and ->remove_bus() hooks to hw_pci

Some PCI drivers may need to adjust the pci_bus structure after it has
been allocated by the Linux PCI core. The PCI core allows
architectures to implement the pcibios_add_bus() and
pcibios_remove_bus() for this purpose. This commit therefore extends
the hw_pci and pci_sys_data structures of the ARM PCI core to allow
PCI drivers to register ->add_bus() and ->remove_bus() in hw_pci,
which will get called when a bus is added or removed from the system.

This will be used for example by the Marvell PCIe driver to connect a
particular PCI bus with its corresponding MSI chip to handle Message
Signaled Interrupts.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Thierry Reding <thierry.reding@gmail.com>
Acked-by: Russell King <rmk+kernel@arm.linux.org.uk>
Tested-by: Daniel Price <daniel.price@gmail.com>
Tested-by: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoof: pci: add registry of MSI chips
Thomas Petazzoni [Fri, 9 Aug 2013 20:27:09 +0000 (22:27 +0200)]
of: pci: add registry of MSI chips

This commit adds a very basic registry of msi_chip structures, so that
an IRQ controller driver can register an msi_chip, and a PCIe host
controller can find it, based on a 'struct device_node'.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Rob Herring <rob.herring@calxeda.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: Introduce new MSI chip infrastructure
Thierry Reding [Fri, 9 Aug 2013 20:27:08 +0000 (22:27 +0200)]
PCI: Introduce new MSI chip infrastructure

The new struct msi_chip is used to associated an MSI controller with a
PCI bus. It is automatically handed down from the root to its children
during bus enumeration.

This patch provides default (weak) implementations for the architecture-
specific MSI functions (arch_setup_msi_irq(), arch_teardown_msi_irq()
and arch_msi_check_device()) which check if a PCI device's bus has an
attached MSI chip and forward the call appropriately.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Tested-by: Daniel Price <daniel.price@gmail.com>
Tested-by: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: remove ARCH_SUPPORTS_MSI kconfig option
Thomas Petazzoni [Fri, 9 Aug 2013 20:27:07 +0000 (22:27 +0200)]
PCI: remove ARCH_SUPPORTS_MSI kconfig option

Now that we have weak versions for each of the PCI MSI architecture
functions, we can actually build the MSI support for all platforms,
regardless of whether they provide or not architecture-specific
versions of those functions. For this reason, the ARCH_SUPPORTS_MSI
hidden kconfig boolean becomes useless, and this patch gets rid of it.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Tested-by: Daniel Price <daniel.price@gmail.com>
Tested-by: Thierry Reding <thierry.reding@gmail.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: linux390@de.ibm.com
Cc: linux-s390@vger.kernel.org
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: x86@kernel.org
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: linux-ia64@vger.kernel.org
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: linux-mips@linux-mips.org
Cc: David S. Miller <davem@davemloft.net>
Cc: sparclinux@vger.kernel.org
Cc: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: use weak functions for MSI arch-specific functions
Thomas Petazzoni [Fri, 9 Aug 2013 20:27:06 +0000 (22:27 +0200)]
PCI: use weak functions for MSI arch-specific functions

Until now, the MSI architecture-specific functions could be overloaded
using a fairly complex set of #define and compile-time
conditionals. In order to prepare for the introduction of the msi_chip
infrastructure, it is desirable to switch all those functions to use
the 'weak' mechanism. This commit converts all the architectures that
were overidding those MSI functions to use the new strategy.

Note that we keep two separate, non-weak, functions
default_teardown_msi_irqs() and default_restore_msi_irqs() for the
default behavior of the arch_teardown_msi_irqs() and
arch_restore_msi_irqs(), as the default behavior is needed by x86 PCI
code.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Tested-by: Daniel Price <daniel.price@gmail.com>
Tested-by: Thierry Reding <thierry.reding@gmail.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: linux390@de.ibm.com
Cc: linux-s390@vger.kernel.org
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: x86@kernel.org
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: linux-ia64@vger.kernel.org
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: linux-mips@linux-mips.org
Cc: David S. Miller <davem@davemloft.net>
Cc: sparclinux@vger.kernel.org
Cc: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: Relocate PCIe node in Armada 370 RD board
Ezequiel Garcia [Thu, 8 Aug 2013 21:03:09 +0000 (18:03 -0300)]
ARM: mvebu: Relocate PCIe node in Armada 370 RD board

The pcie-controller node needs to be relocated according the MBus
DT binding, since it's now a child of the mbus-compatible node.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: Fix AXP-WiFi-AP DT for MBUS DT binding
Ezequiel Garcia [Tue, 6 Aug 2013 17:09:42 +0000 (14:09 -0300)]
ARM: mvebu: Fix AXP-WiFi-AP DT for MBUS DT binding

The ranges property needs to be changed to use the new MBus DT binding.
Also, the pcie-controller node needs to be relocated as according the MBus
DT binding, it's now a child of the mbus-compatible node.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: add support for the AXP WiFi AP board
Thomas Petazzoni [Tue, 30 Jul 2013 15:44:50 +0000 (17:44 +0200)]
ARM: mvebu: add support for the AXP WiFi AP board

The AXP WiFi AP board is a Marvell platform based on the Armada XP
MV78230 SoC. It has two mini-PCIe connectors, one USB 3.0 port powered
by a USB 3.0 controller on PCIe, two Ethernet ports, 1 GB of RAM, 1 GB
of NAND, 16 MB of SPI flash, one SATA port and one button, two UARTs

Successfully tested: USB 3.0 port, the mini-PCIe connectors, SPI
flash, Ethernet ports, SATA port, button, UART.

Untested: NAND flash, due to lack of mainline support for the Armada
370/XP NAND controller for now.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Seif Mazareeb <seif@marvell.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: use dts pre-processor for mv78230
Jason Cooper [Wed, 7 Aug 2013 20:04:21 +0000 (20:04 +0000)]
ARM: mvebu: use dts pre-processor for mv78230

Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: use dts pre-processor for mv88f6281gtw-ge
Jason Cooper [Tue, 6 Aug 2013 15:02:24 +0000 (15:02 +0000)]
ARM: kirkwood: use dts pre-processor for mv88f6281gtw-ge

Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: convert the mv88f6281gtw_ge board to DT
Thomas Petazzoni [Thu, 1 Aug 2013 15:07:32 +0000 (17:07 +0200)]
ARM: kirkwood: convert the mv88f6281gtw_ge board to DT

This commit converts the mv88f6281gtw_ge Kirkwood board to use a
Device Tree representation, except for the Ethernet devices and the
DSA switch. Even though the mv643xx_eth driver has a DT binding,
converting this board to use it is for now left on the side because it
doesn't use a simple PHY, but a DSA switch instead.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: remove LaCie boards that are supported through DT
Thomas Petazzoni [Thu, 1 Aug 2013 15:07:31 +0000 (17:07 +0200)]
ARM: kirkwood: remove LaCie boards that are supported through DT

The "LaCie Internet Space v2 NAS Board", "LaCie Network Space Max v2
NAS Board" and "LaCie Network Space v2 NAS Board" boards are all
supported through the Device Tree.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: remove support for legacy booting of Sheevaplug
Thomas Petazzoni [Thu, 1 Aug 2013 15:07:30 +0000 (17:07 +0200)]
ARM: kirkwood: remove support for legacy booting of Sheevaplug

The Kirkwood Sheevaplug platform has already been converted to the
Device Tree, so we can remove the legacy booting option for this
platform.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: remove support for legacy booting of Guruplug
Thomas Petazzoni [Thu, 1 Aug 2013 15:07:29 +0000 (17:07 +0200)]
ARM: kirkwood: remove support for legacy booting of Guruplug

The Kirkwood Guruplug platform has already been converted to the
Device Tree, so we can remove the legacy booting option for this
platform.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: remove support for legacy booting of Dockstar
Thomas Petazzoni [Thu, 1 Aug 2013 15:07:28 +0000 (17:07 +0200)]
ARM: kirkwood: remove support for legacy booting of Dockstar

The Kirkwood Dockstar platform has already been converted to the
Device Tree, so we can remove the legacy booting option for this
platform.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: mvebu: Adapt to the new device tree layout
Thomas Petazzoni [Fri, 26 Jul 2013 13:18:07 +0000 (10:18 -0300)]
PCI: mvebu: Adapt to the new device tree layout

The new device tree layout encodes the window's target ID and attribute
in the PCIe controller node's ranges property. This allows to parse
such entries to obtain such information and use the recently introduced
MBus API to create the windows, instead of using the current name based
scheme.

Cc: devicetree@vger.kernel.org
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agobus: mvebu-mbus: Add devicetree binding
Ezequiel Garcia [Fri, 26 Jul 2013 13:18:06 +0000 (10:18 -0300)]
bus: mvebu-mbus: Add devicetree binding

Introduce the devicetree binding for the mvebu MBus driver
avaiable in the mvebu SoCs (Armada 370/XP, Kirkwood, Dove, ...).

This binding provides an accurate model of the SoC address space,
and allows to declare the address and size of the decoding windows the MBus
needs to access the peripherals, together with the target ID and attribute
for those windows.

The binding is composed of two required nodes: one for the MBus bus
and one for the MBus controller.

Cc: devicetree@vger.kernel.org
Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Acked-by: Rob Herring <rob.herring@calxeda.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: Relocate PCIe device tree nodes
Ezequiel Garcia [Fri, 26 Jul 2013 13:18:05 +0000 (10:18 -0300)]
ARM: kirkwood: Relocate PCIe device tree nodes

Now that mbus has been added to the device tree, it's possible to
move the PCIe nodes out of the ocp node, placing it directly
below the mbus. This is a more accurate representation of the hardware.

Moving the PCIe nodes, we now need to introduce an extra cell to
encode the window target ID and attribute. Since this depends on
the PCIe port, we split the ranges translation entries, to
correspond to each MBus window.

In addition, we encode the PCIe memory and I/O apertures in the MBus
node, according to the MBus DT binding specification. The choice made
is 0xe0000000-0xf0000000 for memory space, and 0xf200000-0xf2100000 for
I/O space. These apertures can be changed in each per-board DT file.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: Introduce MBUS_ID
Ezequiel Garcia [Fri, 26 Jul 2013 13:18:04 +0000 (10:18 -0300)]
ARM: kirkwood: Introduce MBUS_ID

This macro is used to define window's target ID and attribute cells
for the MBus ranges entries.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: Introduce MBus DT node
Ezequiel Garcia [Fri, 26 Jul 2013 13:18:03 +0000 (10:18 -0300)]
ARM: kirkwood: Introduce MBus DT node

Add a minimal MBus node, just to allow the MBus driver to probe.
Follow-up patches will migrate the rest of the nodes appropriately.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: Use the preprocessor on device tree files
Ezequiel Garcia [Fri, 26 Jul 2013 13:18:02 +0000 (10:18 -0300)]
ARM: kirkwood: Use the preprocessor on device tree files

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: Split DT and legacy MBus initialization
Ezequiel Garcia [Fri, 26 Jul 2013 13:18:01 +0000 (10:18 -0300)]
ARM: kirkwood: Split DT and legacy MBus initialization

This commit replaces the legacy MBus initialization with the new
DT-based in Kirkwood. For boards that are not yet converted to DT,
we keep the legacy initialization.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: Relocate Armada 370/XP PCIe device tree nodes
Ezequiel Garcia [Fri, 26 Jul 2013 13:18:00 +0000 (10:18 -0300)]
ARM: mvebu: Relocate Armada 370/XP PCIe device tree nodes

Now that mbus has been added to the device tree, it's possible to
move the PCIe nodes out of internal registers, placing it directly
below the mbus. This is a more accurate representation of the
hardware.

Moving the PCIe nodes, we now need to introduce an extra cell to
encode the window target ID and attribute. Since this depends on
the PCIe port, we split the ranges translation entries, to correspond
to each MBus window.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: Relocate Armada 370/XP DeviceBus device tree nodes
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:59 +0000 (10:17 -0300)]
ARM: mvebu: Relocate Armada 370/XP DeviceBus device tree nodes

Now that mbus has been added to the device tree, it's possible to
move the DeviceBus out of internal registers, placing it directly
below the mbus. This is a more accurate representation of the hardware.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: Add BootROM to Armada 370/XP device tree
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:58 +0000 (10:17 -0300)]
ARM: mvebu: Add BootROM to Armada 370/XP device tree

In order to access the SoC BootROM, we need to declare a mapping
(through a ranges property). The mbus driver will use this property
to allocate a suitable address decoding window.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: Add MBus to Armada 370/XP device tree
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:57 +0000 (10:17 -0300)]
ARM: mvebu: Add MBus to Armada 370/XP device tree

The Armada 370/XP SoC family has a completely configurable address
space handled by the MBus controller.

This patch introduces the device tree layout of MBus, making the
'soc' node as mbus-compatible.
Since every peripheral/controller is a child of this 'soc' node,
this makes all of them sit behind the mbus, thus describing the
hardware accurately.

A translation entry has been added for the internal-regs mapping.
This can't be done in the common armada-370-xp.dtsi because A370
and AXP have different addressing width.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: Use the preprocessor on Armada 370/XP device tree files
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:56 +0000 (10:17 -0300)]
ARM: mvebu: Use the preprocessor on Armada 370/XP device tree files

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: Initialize MBus using the DT binding
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:55 +0000 (10:17 -0300)]
ARM: mvebu: Initialize MBus using the DT binding

Now that the mbus device tree binding has been introduced, we can
switch over to it.

Also, and since the initialization of the mbus driver is quite
fundamental for the system to work properly, this patch adds a BUG()
in case mbus fails to initialize.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mvebu: Remove the harcoded BootROM window allocation
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:54 +0000 (10:17 -0300)]
ARM: mvebu: Remove the harcoded BootROM window allocation

The address decoding window to access the BootROM should not be
allocated programatically, but instead declared in the device tree.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agobus: mvebu-mbus: Factorize Armada 370/XP data structures
Thomas Petazzoni [Fri, 26 Jul 2013 13:17:53 +0000 (10:17 -0300)]
bus: mvebu-mbus: Factorize Armada 370/XP data structures

These structures were only different in the mapping tables.
Now that those tables have been removed, it doesn't make any sense
to keep different structures.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agobus: mvebu-mbus: Update main description
Thomas Petazzoni [Fri, 26 Jul 2013 13:17:52 +0000 (10:17 -0300)]
bus: mvebu-mbus: Update main description

After replacing the MBus name-based by the new ID-based API
let's fix the general description of the driver at the beginning
of the file.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agobus: mvebu-mbus: Remove name -> target, attribute mapping tables
Thomas Petazzoni [Fri, 26 Jul 2013 13:17:51 +0000 (10:17 -0300)]
bus: mvebu-mbus: Remove name -> target, attribute mapping tables

This tables were used together with the name-based MBus window
creation API. Since that's has been removed, we can also remove
the tables.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agobus: mvebu-mbus: Remove the no longer used name-based API
Thomas Petazzoni [Fri, 26 Jul 2013 13:17:50 +0000 (10:17 -0300)]
bus: mvebu-mbus: Remove the no longer used name-based API

Now that every user of the deprecated name-based API has been
converted to using the ID-based API, let's remove the former one.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: mvebu: Check valid base address before port setup
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:49 +0000 (10:17 -0300)]
PCI: mvebu: Check valid base address before port setup

This driver does not fail to probe when it cannot obtain
a port base address. Therefore, add a check for NULL base address
before setting up the port, which prevents a kernel panic in such
cases.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoPCI: mvebu: Adapt to the new device tree layout
Thomas Petazzoni [Fri, 26 Jul 2013 13:17:48 +0000 (10:17 -0300)]
PCI: mvebu: Adapt to the new device tree layout

The new device tree layout encodes the window's target ID and attribute
in the PCIe controller node's ranges property. This allows to parse
such entries to obtain such information and use the recently introduced
MBus API to create the windows, instead of using the current name based
scheme.

Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agobus: mvebu-mbus: Add new API for the PCIe memory and IO aperture
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:47 +0000 (10:17 -0300)]
bus: mvebu-mbus: Add new API for the PCIe memory and IO aperture

We add two optional properties to the MBus DT binding, to encode
the PCIe memory and IO aperture. This allows such information to
be retrieved by -for instance- the pci driver to allocate the
MBus decoding windows.

Correspondingly, and in order to retrieve this information,
we add two new APIs.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agobus: mvebu-mbus: Add static window allocation to the DT binding
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:46 +0000 (10:17 -0300)]
bus: mvebu-mbus: Add static window allocation to the DT binding

This patch adds static window allocation to the device tree binding.
Each first-child of the mbus-compatible node, with a suitable 'ranges'
property, declaring an address translation, will trigger an address
decoding window allocation.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agobus: mvebu-mbus: Introduce device tree binding
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:45 +0000 (10:17 -0300)]
bus: mvebu-mbus: Introduce device tree binding

This patch adds the most fundamental device-tree initialization.
We only introduce what's required to be able to probe the mvebu-mbus
driver from the DT. Follow-up patches will extend the device tree binding,
allowing to describe static address decoding windows.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agobus: mvebu-mbus: Factor out initialization details
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:44 +0000 (10:17 -0300)]
bus: mvebu-mbus: Factor out initialization details

We introduce a common initialization function mvebu_mbus_common_init()
that will be used by both legacy and device-tree initialization code.
This patch is an intermediate step, which will allow to introduce the
DT binding for this driver in a less intrusive way.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: Move to ID based window creation
Thomas Petazzoni [Fri, 26 Jul 2013 13:17:43 +0000 (10:17 -0300)]
ARM: dove: Move to ID based window creation

With the introduction of the ID based MBus API, it's better
to switch to use it instead of the current name based scheme.

This will allow to deprecate the name based API, once every
user is removed.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: orion5x: Move to ID based window creation
Thomas Petazzoni [Fri, 26 Jul 2013 13:17:42 +0000 (10:17 -0300)]
ARM: orion5x: Move to ID based window creation

With the introduction of the ID based MBus API, it's better
to switch to use it instead of the current name based scheme.

This will allow to deprecate the name based API, once every
user is removed.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: mv78xx0: Move to ID based window creation
Thomas Petazzoni [Fri, 26 Jul 2013 13:17:41 +0000 (10:17 -0300)]
ARM: mv78xx0: Move to ID based window creation

With the introduction of the ID based MBus API, it's better
to switch to use it instead of the current name based scheme.

This will allow to deprecate the name based API, once every
user is removed.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: Move to ID based MBus window creation
Thomas Petazzoni [Fri, 26 Jul 2013 13:17:40 +0000 (10:17 -0300)]
ARM: kirkwood: Move to ID based MBus window creation

With the introduction of the ID based MBus API, it's better
to switch to use it instead of the current name based scheme.

This will allow to deprecate the name based API, once every
user is removed.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agobus: mvebu-mbus: Add new API for window creation
Thomas Petazzoni [Fri, 26 Jul 2013 13:17:39 +0000 (10:17 -0300)]
bus: mvebu-mbus: Add new API for window creation

We add an API to create MBus address decoding windows from the target
ID and attribute. This function will be used later and deprecate the
current name based scheme.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agomemory: mvebu-devbus: Remove address decoding window workaround
Ezequiel Garcia [Fri, 26 Jul 2013 13:17:38 +0000 (10:17 -0300)]
memory: mvebu-devbus: Remove address decoding window workaround

Now that mbus device tree binding has been introduced, remove the address
decoding window management from this driver.
A suitable 'ranges' entry should be added to the devbus-compatible node in
the device tree, as described by the mbus binding documentation.

Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: plat-orion: add reg offset to DT irq driver stub
Sebastian Hesselbarth [Tue, 2 Jul 2013 11:03:43 +0000 (13:03 +0200)]
ARM: plat-orion: add reg offset to DT irq driver stub

With irqchip driver for Orion SoCs, reg layout of orion-intc has changed.
This updates irq driver stub implemented before to the new reg layout by
adding an offset to the base address passed by DT node. As orion5x still
uses this stub, it cannot be removed yet.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: remove obsolete SDIO clock gate workaround
Sebastian Hesselbarth [Tue, 2 Jul 2013 11:03:42 +0000 (13:03 +0200)]
ARM: kirkwood: remove obsolete SDIO clock gate workaround

Kirkwood DT boards have been converted to DT SDIO already, we can now
remove the legacy clock gate workaround for SDIO clock.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: convert to DT irqchip and clocksource
Sebastian Hesselbarth [Tue, 2 Jul 2013 13:15:07 +0000 (15:15 +0200)]
ARM: kirkwood: convert to DT irqchip and clocksource

With recent support for true irqchip and clocksource drivers for Orion
SoCs, now make use of it on DT enabled Kirkwood boards.

This also introduces a new Kconfig option for legacy (non-DT) Kirkwood
where old code is moved out to and polishes DT board file a little bit.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: convert to DT irqchip and clocksource
Sebastian Hesselbarth [Tue, 2 Jul 2013 11:03:40 +0000 (13:03 +0200)]
ARM: dove: convert to DT irqchip and clocksource

With recent support for true irqchip and clocksource drivers for Orion
SoCs, now make use of it on DT enabled Dove boards.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: orion5x: update intc device tree node to new reg layout
Sebastian Hesselbarth [Tue, 2 Jul 2013 11:03:39 +0000 (13:03 +0200)]
ARM: orion5x: update intc device tree node to new reg layout

Future irqchip driver for Orion SoCs has a modified reg layout starting
from CAUSE register. The former DT irq driver stub just used what the
non-DT driver wants. Currently, orion5x is not ready for irqchip driver
because of legacy timer and missing DT clk provider. To ease future
conversion and just because orion-intc binding documentation already
reflects new reg layout, we convert orion5x DT now.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: move device tree nodes to DT irqchip and clocksource
Sebastian Hesselbarth [Tue, 2 Jul 2013 11:03:38 +0000 (13:03 +0200)]
ARM: kirkwood: move device tree nodes to DT irqchip and clocksource

With recent support for true irqchip and clocksource drivers for Orion
SoCs, now make use of it on DT enabled Kirkwood boards.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: move device tree nodes to DT irqchip and clocksource
Sebastian Hesselbarth [Tue, 2 Jul 2013 11:03:37 +0000 (13:03 +0200)]
ARM: dove: move device tree nodes to DT irqchip and clocksource

With recent support for true irqchip and clocksource drivers for Orion
SoCs, now make use of it on DT enabled Dove boards.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: orion5x: remove legacy mv643xx_eth board setup
Sebastian Hesselbarth [Tue, 2 Jul 2013 11:00:24 +0000 (13:00 +0200)]
ARM: orion5x: remove legacy mv643xx_eth board setup

With DT support for mv643xx_eth we do not need legacy platform_data
based setup for DT enabled boards. This patch removes eth setup
for all orion5x DT board files.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: remove legacy clk alias for mv643xx_eth
Sebastian Hesselbarth [Tue, 2 Jul 2013 11:00:23 +0000 (13:00 +0200)]
ARM: kirkwood: remove legacy clk alias for mv643xx_eth

With all boards converted to DT enabled mv643xx_eth we can now
remove the clock alias for gbe clocks.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: remove redundant DT board files
Sebastian Hesselbarth [Wed, 3 Jul 2013 16:57:51 +0000 (18:57 +0200)]
ARM: kirkwood: remove redundant DT board files

With DT support for mv643xx_eth board specific init for some boards now
is unneccessary. Remove those board files, Kconfig entries, and
corresponding entries in kirkwood_defconfig.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: remove legacy mv643xx_eth setup
Sebastian Hesselbarth [Tue, 2 Jul 2013 11:00:21 +0000 (13:00 +0200)]
ARM: dove: remove legacy mv643xx_eth setup

With DT support for mv643xx_eth we do not need legacy platform_data
based setup for DT enabled boards anymore.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: orion5x: add gigabit ethernet and mvmdio device tree nodes
Sebastian Hesselbarth [Tue, 2 Jul 2013 11:00:20 +0000 (13:00 +0200)]
ARM: orion5x: add gigabit ethernet and mvmdio device tree nodes

This patch adds mv643xx_eth and mvmdio device tree nodes for DT enabled
Orion5x boards. Phy nodes are also added with reg property set on a
per-board basis.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: kirkwood: add gigabit ethernet and mvmdio device tree nodes
Sebastian Hesselbarth [Sun, 7 Jul 2013 20:34:56 +0000 (22:34 +0200)]
ARM: kirkwood: add gigabit ethernet and mvmdio device tree nodes

This patch adds mv643xx_eth and mvmdio device tree nodes for DT enabled
Kirkwood boards. Phy nodes are also added with reg property set on a
per-board basis.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoARM: dove: add gigabit ethernet and mvmdio device tree nodes
Sebastian Hesselbarth [Tue, 2 Jul 2013 11:00:18 +0000 (13:00 +0200)]
ARM: dove: add gigabit ethernet and mvmdio device tree nodes

This patch adds orion-eth and mvmdio device tree nodes for DT enabled
Dove boards. As there is only one ethernet controller on Dove, a default
phy node is also added with a note to set its reg property on a per-board
basis.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
11 years agoLinux 3.11-rc2 v3.11-rc2
Linus Torvalds [Sun, 21 Jul 2013 19:05:29 +0000 (12:05 -0700)]
Linux 3.11-rc2

11 years agoMerge tag 'acpi-video-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Sun, 21 Jul 2013 17:11:04 +0000 (10:11 -0700)]
Merge tag 'acpi-video-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

Pull ACPI video support fixes from Rafael Wysocki:
 "I'm sending a separate pull request for this as it may be somewhat
  controversial.  The breakage addressed here is not really new and the
  fixes may not satisfy all users of the affected systems, but we've had
  so much back and forth dance in this area over the last several weeks
  that I think it's time to actually make some progress.

  The source of the problem is that about a year ago we started to tell
  BIOSes that we're compatible with Windows 8, which we really need to
  do, because some systems shipping with Windows 8 are tested with it
  and nothing else, so if we tell their BIOSes that we aren't compatible
  with Windows 8, we expose our users to untested BIOS/AML code paths.

  However, as it turns out, some Windows 8-specific AML code paths are
  not tested either, because Windows 8 actually doesn't use the ACPI
  methods containing them, so if we declare Windows 8 compatibility and
  attempt to use those ACPI methods, things break.  That occurs mostly
  in the backlight support area where in particular the _BCM and _BQC
  methods are plain unusable on some systems if the OS declares Windows
  8 compatibility.

  [ The additional twist is that they actually become usable if the OS
    says it is not compatible with Windows 8, but that may cause
    problems to show up elsewhere ]

  Investigation carried out by Matthew Garrett indicates that what
  Windows 8 does about backlight is to leave backlight control up to
  individual graphics drivers.  At least there's evidence that it does
  that if the Intel graphics driver is used, so we've decided to follow
  Windows 8 in that respect and allow i915 to control backlight (Daniel
  likes that part).

  The first commit from Aaron Lu makes ACPICA export the variable from
  which we can infer whether or not the BIOS believes that we are
  compatible with Windows 8.

  The second commit from Matthew Garrett prepares the ACPI video driver
  by making it initialize the ACPI backlight even if it is not going to
  be used afterward (that is needed for backlight control to work on
  Thinkpads).

  The third commit implements the actual workaround making i915 take
  over backlight control if the firmware thinks it's dealing with
  Windows 8 and is based on the work of multiple developers, including
  Matthew Garrett, Chun-Yi Lee, Seth Forshee, and Aaron Lu.

  The final commit from Aaron Lu makes us follow Windows 8 by informing
  the firmware through the _DOS method that it should not carry out
  automatic brightness changes, so that brightness can be controlled by
  GUI.

  Hopefully, this approach will allow us to avoid using blacklists of
  systems that should not declare Windows 8 compatibility just to avoid
  backlight control problems in the future.

   - Change from Aaron Lu makes ACPICA export a variable which can be
     used by driver code to determine whether or not the BIOS believes
     that we are compatible with Windows 8.

   - Change from Matthew Garrett makes the ACPI video driver initialize
     the ACPI backlight even if it is not going to be used afterward
     (that is needed for backlight control to work on Thinkpads).

   - Fix from Rafael J Wysocki implements Windows 8 backlight support
     workaround making i915 take over bakclight control if the firmware
     thinks it's dealing with Windows 8.  Based on the work of multiple
     developers including Matthew Garrett, Chun-Yi Lee, Seth Forshee,
     and Aaron Lu.

   - Fix from Aaron Lu makes the kernel follow Windows 8 by informing
     the firmware through the _DOS method that it should not carry out
     automatic brightness changes, so that brightness can be controlled
     by GUI"

* tag 'acpi-video-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI / video: no automatic brightness changes by win8-compatible firmware
  ACPI / video / i915: No ACPI backlight if firmware expects Windows 8
  ACPI / video: Always call acpi_video_init_brightness() on init
  ACPICA: expose OSI version

11 years agoMerge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso...
Linus Torvalds [Sun, 21 Jul 2013 03:11:42 +0000 (20:11 -0700)]
Merge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4

Pull ext[34] tmpfile bugfix from Ted Ts'o:
 "Fix regression caused by commit af51a2ac36d1f which added ->tmpfile()
  support (along with a similar fix for ext3)"

* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext3: fix a BUG when opening a file with O_TMPFILE flag
  ext4: fix a BUG when opening a file with O_TMPFILE flag

11 years agoext3: fix a BUG when opening a file with O_TMPFILE flag
Zheng Liu [Sun, 21 Jul 2013 02:03:20 +0000 (22:03 -0400)]
ext3: fix a BUG when opening a file with O_TMPFILE flag

When we try to open a file with O_TMPFILE flag, we will trigger a bug.
The root cause is that in ext4_orphan_add() we check ->i_nlink == 0 and
this check always fails because we set ->i_nlink = 1 in
inode_init_always().  We can use the following program to trigger it:

int main(int argc, char *argv[])
{
int fd;

fd = open(argv[1], O_TMPFILE, 0666);
if (fd < 0) {
perror("open ");
return -1;
}
close(fd);
return 0;
}

The oops message looks like this:

kernel: kernel BUG at fs/ext3/namei.c:1992!
kernel: invalid opcode: 0000 [#1] SMP
kernel: Modules linked in: ext4 jbd2 crc16 cpufreq_ondemand ipv6 dm_mirror dm_region_hash dm_log dm_mod parport_pc parport serio_raw sg dcdbas pcspkr i2c_i801 ehci_pci ehci_hcd button acpi_cpufreq mperf e1000e ptp pps_core ttm drm_kms_helper drm hwmon i2c_algo_bit i2c_core ext3 jbd sd_mod ahci libahci libata scsi_mod uhci_hcd
kernel: CPU: 0 PID: 2882 Comm: tst_tmpfile Not tainted 3.11.0-rc1+ #4
kernel: Hardware name: Dell Inc. OptiPlex 780 /0V4W66, BIOS A05 08/11/2010
kernel: task: ffff880112d30050 ti: ffff8801124d4000 task.ti: ffff8801124d4000
kernel: RIP: 0010:[<ffffffffa00db5ae>] [<ffffffffa00db5ae>] ext3_orphan_add+0x6a/0x1eb [ext3]
kernel: RSP: 0018:ffff8801124d5cc8  EFLAGS: 00010202
kernel: RAX: 0000000000000000 RBX: ffff880111510128 RCX: ffff8801114683a0
kernel: RDX: 0000000000000000 RSI: ffff880111510128 RDI: ffff88010fcf65a8
kernel: RBP: ffff8801124d5d18 R08: 0080000000000000 R09: ffffffffa00d3b7f
kernel: R10: ffff8801114683a0 R11: ffff8801032a2558 R12: 0000000000000000
kernel: R13: ffff88010fcf6800 R14: ffff8801032a2558 R15: ffff8801115100d8
kernel: FS:  00007f5d172b5700(0000) GS:ffff880117c00000(0000) knlGS:0000000000000000
kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
kernel: CR2: 00007f5d16df15d0 CR3: 0000000110b1d000 CR4: 00000000000407f0
kernel: Stack:
kernel: 000000000000000c ffff8801048a7dc8 ffff8801114685a8 ffffffffa00b80d7
kernel: ffff8801124d5e38 ffff8801032a2558 ffff88010ce24d68 0000000000000000
kernel: ffff88011146b300 ffff8801124d5d44 ffff8801124d5d78 ffffffffa00db7e1
kernel: Call Trace:
kernel: [<ffffffffa00b80d7>] ? journal_start+0x8c/0xbd [jbd]
kernel: [<ffffffffa00db7e1>] ext3_tmpfile+0xb2/0x13b [ext3]
kernel: [<ffffffff821076f8>] path_openat+0x11f/0x5e7
kernel: [<ffffffff821c86b4>] ? list_del+0x11/0x30
kernel: [<ffffffff82065fa2>] ?  __dequeue_entity+0x33/0x38
kernel: [<ffffffff82107cd5>] do_filp_open+0x3f/0x8d
kernel: [<ffffffff82112532>] ? __alloc_fd+0x50/0x102
kernel: [<ffffffff820f9296>] do_sys_open+0x13b/0x1cd
kernel: [<ffffffff820f935c>] SyS_open+0x1e/0x20
kernel: [<ffffffff82398c02>] system_call_fastpath+0x16/0x1b
kernel: Code: 39 c7 0f 85 67 01 00 00 0f b7 03 25 00 f0 00 00 3d 00 40 00 00 74 18 3d 00 80 00 00 74 11 3d 00 a0 00 00 74 0a 83 7b 48 00 74 04 <0f> 0b eb fe 49 8b 85 50 03 00 00 4c 89 f6 48 c7 c7 c0 99 0e a0
kernel: RIP  [<ffffffffa00db5ae>] ext3_orphan_add+0x6a/0x1eb [ext3]
kernel: RSP <ffff8801124d5cc8>

Here we couldn't call clear_nlink() directly because in d_tmpfile() we
will call inode_dec_link_count() to decrease ->i_nlink.  So this commit
tries to call d_tmpfile() before ext4_orphan_add() to fix this problem.

Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: Jan Kara <jack@suse.cz>
Cc: Al Viro <viro@zeniv.linux.org.uk>
11 years agoext4: fix a BUG when opening a file with O_TMPFILE flag
Zheng Liu [Sun, 21 Jul 2013 01:58:38 +0000 (21:58 -0400)]
ext4: fix a BUG when opening a file with O_TMPFILE flag

When we try to open a file with O_TMPFILE flag, we will trigger a bug.
The root cause is that in ext4_orphan_add() we check ->i_nlink == 0 and
this check always fails because we set ->i_nlink = 1 in
inode_init_always().  We can use the following program to trigger it:

int main(int argc, char *argv[])
{
int fd;

fd = open(argv[1], O_TMPFILE, 0666);
if (fd < 0) {
perror("open ");
return -1;
}
close(fd);
return 0;
}

The oops message looks like this:

kernel BUG at fs/ext4/namei.c:2572!
invalid opcode: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
Modules linked in: dlci bridge stp hidp cmtp kernelcapi l2tp_ppp l2tp_netlink l2tp_core sctp libcrc32c rfcomm tun fuse nfnetli
nk can_raw ipt_ULOG can_bcm x25 scsi_transport_iscsi ipx p8023 p8022 appletalk phonet psnap vmw_vsock_vmci_transport af_key vmw_vmci rose vsock atm can netrom ax25 af_rxrpc ir
da pppoe pppox ppp_generic slhc bluetooth nfc rfkill rds caif_socket caif crc_ccitt af_802154 llc2 llc snd_hda_codec_realtek snd_hda_intel snd_hda_codec serio_raw snd_pcm pcsp
kr edac_core snd_page_alloc snd_timer snd soundcore r8169 mii sr_mod cdrom pata_atiixp radeon backlight drm_kms_helper ttm
CPU: 1 PID: 1812571 Comm: trinity-child2 Not tainted 3.11.0-rc1+ #12
Hardware name: Gigabyte Technology Co., Ltd. GA-MA78GM-S2H/GA-MA78GM-S2H, BIOS F12a 04/23/2010
task: ffff88007dfe69a0 ti: ffff88010f7b6000 task.ti: ffff88010f7b6000
RIP: 0010:[<ffffffff8125ce69>]  [<ffffffff8125ce69>] ext4_orphan_add+0x299/0x2b0
RSP: 0018:ffff88010f7b7cf8  EFLAGS: 00010202
RAX: 0000000000000000 RBX: ffff8800966d3020 RCX: 0000000000000000
RDX: 0000000000000000 RSI: ffff88007dfe70b8 RDI: 0000000000000001
RBP: ffff88010f7b7d40 R08: ffff880126a3c4e0 R09: ffff88010f7b7ca0
R10: 0000000000000000 R11: 0000000000000000 R12: ffff8801271fd668
R13: ffff8800966d2f78 R14: ffff88011d7089f0 R15: ffff88007dfe69a0
FS:  00007f70441a3740(0000) GS:ffff88012a800000(0000) knlGS:00000000f77c96c0
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000002834000 CR3: 0000000107964000 CR4: 00000000000007e0
DR0: 0000000000780000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000600
Stack:
 0000000000002000 00000020810b6dde 0000000000000000 ffff88011d46db00
 ffff8800966d3020 ffff88011d7089f0 ffff88009c7f4c10 ffff88010f7b7f2c
 ffff88007dfe69a0 ffff88010f7b7da8 ffffffff8125cfac ffff880100000004
Call Trace:
 [<ffffffff8125cfac>] ext4_tmpfile+0x12c/0x180
 [<ffffffff811cba78>] path_openat+0x238/0x700
 [<ffffffff8100afc4>] ? native_sched_clock+0x24/0x80
 [<ffffffff811cc647>] do_filp_open+0x47/0xa0
 [<ffffffff811db73f>] ? __alloc_fd+0xaf/0x200
 [<ffffffff811ba2e4>] do_sys_open+0x124/0x210
 [<ffffffff81010725>] ? syscall_trace_enter+0x25/0x290
 [<ffffffff811ba3ee>] SyS_open+0x1e/0x20
 [<ffffffff816ca8d4>] tracesys+0xdd/0xe2
 [<ffffffff81001001>] ? start_thread_common.constprop.6+0x1/0xa0
Code: 04 00 00 00 89 04 24 31 c0 e8 c4 77 04 00 e9 43 fe ff ff 66 25 00 d0 66 3d 00 80 0f 84 0e fe ff ff 83 7b 48 00 0f 84 04 fe ff ff <0f> 0b 49 8b 8c 24 50 07 00 00 e9 88 fe ff ff 0f 1f 84 00 00 00

Here we couldn't call clear_nlink() directly because in d_tmpfile() we
will call inode_dec_link_count() to decrease ->i_nlink.  So this commit
tries to call d_tmpfile() before ext4_orphan_add() to fix this problem.

Reported-by: Dave Jones <davej@redhat.com>
Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
Tested-by: Darrick J. Wong <darrick.wong@oracle.com>
Tested-by: Dave Jones <davej@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Acked-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agoMerge tag 'staging-3.11-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sat, 20 Jul 2013 22:42:38 +0000 (15:42 -0700)]
Merge tag 'staging-3.11-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging

Pull staging tree fixes from Greg KH:
 "Here are a few iio driver fixes for 3.11-rc2.  They are still spread
  across drivers/iio and drivers/staging/iio so they are coming in
  through this tree.

  I've also removed the drivers/staging/csr/ driver as the developers
  who originally sent it to me have moved on to other companies, and CSR
  still will not send us the specs for the device, making the driver
  pretty much obsolete and impossible to fix up.  Deleting it now
  prevents people from sending in lots of tiny codingsyle fixes that
  will never go anywhere.

  It also helps to offset the large lustre filesystem merge that
  happened in 3.11-rc1 in the overall 3.11.0 diffstat.  :)"

* tag 'staging-3.11-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  staging: csr: remove driver
  iio: lps331ap: Fix wrong in_pressure_scale output value
  iio staging: fix lis3l02dq, read error handling
  staging:iio:ad7291: add missing .driver_module to struct iio_info
  iio: ti_am335x_adc: add missing .driver_module to struct iio_info
  iio: mxs-lradc: Remove useless check in read_raw
  iio: mxs-lradc: Fix misuse of iio->trig
  iio: inkern: fix iio_convert_raw_to_processed_unlocked
  iio: Fix iio_channel_has_info
  iio:trigger: device_unregister->device_del to avoid double free
  iio: dac: ad7303: fix error return code in ad7303_probe()

11 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Sat, 20 Jul 2013 17:50:01 +0000 (10:50 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

Pull vfs fixes from Al Viro:
 "The sget() one is a long-standing bug and will need to go into -stable
  (in fact, it had been originally caught in RHEL6), the other two are
  3.11-only"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  vfs: constify dentry parameter in d_count()
  livelock avoidance in sget()
  allow O_TMPFILE to work with O_WRONLY

11 years agoMerge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso...
Linus Torvalds [Sat, 20 Jul 2013 17:48:59 +0000 (10:48 -0700)]
Merge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4

Pull ext4 bugfixes from Ted Ts'o:
 "Fixes for 3.11-rc2, sent at 5pm, in the professoinal style.  :-)"

I'm not sure I like this new level of "professionalism".
9-5, people, 9-5.

* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: call ext4_es_lru_add() after handling cache miss
  ext4: yield during large unlinks
  ext4: make the extent_status code more robust against ENOMEM failures
  ext4: simplify calculation of blocks to free on error
  ext4: fix error handling in ext4_ext_truncate()

11 years agoMerge tag 'nfs-for-3.11-3' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Linus Torvalds [Sat, 20 Jul 2013 17:48:24 +0000 (10:48 -0700)]
Merge tag 'nfs-for-3.11-3' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client bugfixes from Trond Myklebust:
 - Fix a regression against NFSv4 FreeBSD servers when creating a new
   file
 - Fix another regression in rpc_client_register()

* tag 'nfs-for-3.11-3' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  NFSv4: Fix a regression against the FreeBSD server
  SUNRPC: Fix another issue with rpc_client_register()

11 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/josef/btrfs...
Linus Torvalds [Sat, 20 Jul 2013 17:47:38 +0000 (10:47 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/josef/btrfs-next

Pull btrfs fixes from Josef Bacik:
 "I'm playing the role of Chris Mason this week while he's on vacation.
  There are a few critical fixes for btrfs here, all regressions and
  have been tested well"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/josef/btrfs-next:
  Btrfs: fix wrong write offset when replacing a device
  Btrfs: re-add root to dead root list if we stop dropping it
  Btrfs: fix lock leak when resuming snapshot deletion
  Btrfs: update drop progress before stopping snapshot dropping

11 years agovfs: constify dentry parameter in d_count()
Peng Tao [Thu, 18 Jul 2013 14:09:08 +0000 (22:09 +0800)]
vfs: constify dentry parameter in d_count()

so that it can be used in places like d_compare/d_hash
without causing a compiler warning.

Signed-off-by: Peng Tao <tao.peng@emc.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolivelock avoidance in sget()
Al Viro [Fri, 19 Jul 2013 23:13:55 +0000 (03:13 +0400)]
livelock avoidance in sget()

Eric Sandeen has found a nasty livelock in sget() - take a mount(2) about
to fail.  The superblock is on ->fs_supers, ->s_umount is held exclusive,
->s_active is 1.  Along comes two more processes, trying to mount the same
thing; sget() in each is picking that superblock, bumping ->s_count and
trying to grab ->s_umount.  ->s_active is 3 now.  Original mount(2)
finally gets to deactivate_locked_super() on failure; ->s_active is 2,
superblock is still ->fs_supers because shutdown will *not* happen until
->s_active hits 0.  ->s_umount is dropped and now we have two processes
chasing each other:
s_active = 2, A acquired ->s_umount, B blocked
A sees that the damn thing is stillborn, does deactivate_locked_super()
s_active = 1, A drops ->s_umount, B gets it
A restarts the search and finds the same superblock.  And bumps it ->s_active.
s_active = 2, B holds ->s_umount, A blocked on trying to get it
... and we are in the earlier situation with A and B switched places.

The root cause, of course, is that ->s_active should not grow until we'd
got MS_BORN.  Then failing ->mount() will have deactivate_locked_super()
shut the damn thing down.  Fortunately, it's easy to do - the key point
is that grab_super() is called only for superblocks currently on ->fs_supers,
so it can bump ->s_count and grab ->s_umount first, then check MS_BORN and
bump ->s_active; we must never increment ->s_count for superblocks past
->kill_sb(), but grab_super() is never called for those.

The bug is pretty old; we would've caught it by now, if not for accidental
exclusion between sget() for block filesystems; the things like cgroup or
e.g. mtd-based filesystems don't have anything of that sort, so they get
bitten.  The right way to deal with that is obviously to fix sget()...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>