The sparse IRQ framework is preferred nowadays so switch over to it.

Signed-off-by: Jonas Bonn <jonas@southpole.se>

The Wishbone bus is an open source hardware bus intended to let the parts of an
integrated circuit communicate with each other.  This bus is primarily used by
the cores hosted by the OpenCores project.

The Linux device I/O model assumes that devices have fixed endianess; however,
as the endianess of the OpenCores cores is determined at synthesis time, these
devices may exist in both little and big endian versions.  Effectively, the
endianess selected at synthesis time sets an implicit bus endianess to which
drivers for these devices must comply.

This patch adds I/O primitives of the form wb_io[read|write][8|16|32] which map
to the underlying io[read|write] functions of the correct endianess.  It is
intended that devices connecting to a Wishbone bus use these primitives for
device I/O in order that the correct endianess be automatically applied.

Signed-off-by: Jonas Bonn <jonas@southpole.se>

The ethoc driver wants this as it needs to know which io accessor to
use.

Signed-off-by: Jonas Bonn <jonas@southpole.se>

This is a new driver for the character device SCET.  This driver sets
up the device with major number 100 and minor 0.  An IOCTL is provided
for interfacing to the device from user space applications.

Who wrote this driver?  Where does it come from?  Attributed to Adam
for now but this needs to get sorted out before pushing upstream.

Signed-off-by: Jonas Bonn <jonas@southpole.se>

This patch cleans up the driver initialization and device probing routines.
Miscellaneous code cleanups and usage of devres routines for resource
allocation.

Signed-off-by: Jonas Bonn <jonas@southpole.se>

Custodial fixes:

-- take care of some build warnings
-- use standard Linux routins for debug output
-- reindent scet.h

Signed-off-by: Jonas Bonn <jonas@southpole.se>

This matches the corresponding C file.

Platform data declaration moved to include/linux/ohs900.h

The device platform data is now initiated either from the .dts configuration
or through platform_device.

Switch over to use devres routines and clean up probe routine.

Balance spin_lock_irqsave()/spin_unlock_irqsave()
for all code-paths in ohs900h_urb_dequeue()

Signed-off-by: Anders Nystrom <anders.nystrom@southpole.se>

Update ohs900h_hub_descriptor() and make it resemble ehci_hub_descriptor() more.

The prefetch macro is used in out_packet() function (is it really needed?).
However the required include is is missing.

Also removed a double include and two outcommented ones.

Signed-off-by: Anders Nystrom <anders.nystrom@southpole.se>

device_driver.name is "const char *"

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Jonas Bonn <jonas@southpole.se>

This also fixes a compiler warning, as one cast was missing:

    drivers/usb/host/ohs900-hcd.c: In function ‘out_packet’:
    drivers/usb/host/ohs900-hcd.c:256:200: warning: comparison of distinct pointer types lacks a cast [enabled by default]

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Jonas Bonn <jonas@southpole.se>

The KSZ8001L returns an incorrect ID the first time a PHY is probed.  This
leads to a PHY at address 0 on the MDIO bus being missed.  Resetting the
PHY at initialization resolves this.

The OpenRISC processor is big-endian and generally has a big endian
Wishbone bus.  Accordingly, the bus accesses need to be big endian, too.

Signed-off-by: Jonas Bonn <jonas@southpole.se>

Devices from opencores.org require a compatibility string containing version
information.

Signed-off-by: Jonas Bonn <jonas@southpole.se>

Fixes build issues with 'make ARCH=i386 allyesconfig'

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>
Signed-off-by: Jonas Bonn <jonas@southpole.se>

This is a driver for the OpenCores "Simple" SPI controller.

The spi-simple SPI core has both a FIFO for buffering data and a
transfer register that's used for shifting bits in/out.  The WREMPTY
status bit only indicates that the write FIFO is empty, not that
everything has been shifted out because the last byte may be in the
transfer register.

In order to determine that a transfer has completed and a byte is
ready to read, it's better to test the RFEMPTY bit; this bit will
clear when there's a byte ready to be read.

Reported-by: Yann Vernier <yann.vernier@orsoc.se>
Signed-off-by: Jonas Bonn <jonas@southpole.se>

spi_opencores uses cpuinfo to find out the frequency it divides to
produce the spi clock speed. This probably should be moved out into
the device tree since it's not really based on the CPU but a wishbone
clock.

Signed-off-by: Yann Vernier <yann.vernier@orsoc.se>
Signed-off-by: Jonas Bonn <jonas@southpole.se>

ocspi_reset() didn't reset the SPI controller; it only enabled it.
Resetting the SPI controller is simply done by disabling it first.
This can matter for instance after interrupting the boot ROM
during an SPI read, leaving the controller already enabled and
potentially with data in its FIFOs or transfer register.

Signed-off-by: Yann Vernier <yann.vernier@orsoc.se>
Signed-off-by: Jonas Bonn <jonas@southpole.se>

...since it uses the arch-specific cpuinfo struct to determine the
clock frequency.  This may be reworked in the future to make this
driver generic, but for now it's OpenRISC-only.

Signed-off-by: Jonas Bonn <jonas@southpole.se>

This is more inline with the current naming convention of the
other spi drivers and better reflect the name of the core.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

__devinit and __devexit was removed with CONFIG_HOTPLUG.
__init and __exit has never been correct on probe/remove
related functions.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

This is needed by the MODULE_ macros used in this driver.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

This also make the driver reject any bits_per_word not equal
to 8, since the handling of bits_per_word == 16 was non
functional.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Those are probably remainders from spi-orion, removing
them purge a couple of compile warnings.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Replaces the deprecated master->transfer function with
master->transfer_one_message and let the SPI subsystem do the
message queueing instead of doing it locally.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Use the wrapper functions for getting and setting the driver data using
platform_device instead of using dev_{get,set}_drvdata() with &pdev>dev,
so we can directly pass a struct platform_device.

see 24b5a82
spi: use platform_{get,set}_drvdata() for the upstream commit
that does this SPI subsystem wide.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Removes boilerplate folded into the module_platform_driver macro.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

This patch adds generic GPIO support and a simple driver for the GPIO
controller from Julius Baxter, which I'm calling jbtrivial for lack
of a better name!

drivers/gpio/jbtrivial.c:273:15: error: expected declaration specifiers or ‘...’ before string constant
drivers/gpio/jbtrivial.c:273:1: warning: data definition has no type or storage class
drivers/gpio/jbtrivial.c:273:1: warning: type defaults to ‘int’ in declaration of ‘MODULE_AUTHOR’
drivers/gpio/jbtrivial.c:273:15: warning: function declaration isn’t a prototype
drivers/gpio/jbtrivial.c:274:20: error: expected declaration specifiers or ‘...’ before string constant
drivers/gpio/jbtrivial.c:274:1: warning: data definition has no type or storage class
drivers/gpio/jbtrivial.c:274:1: warning: type defaults to ‘int’ in declaration of ‘MODULE_DESCRIPTION’
drivers/gpio/jbtrivial.c:274:20: warning: function declaration isn’t a prototype
drivers/gpio/jbtrivial.c:275:16: error: expected declaration specifiers or ‘...’ before string constant
drivers/gpio/jbtrivial.c:275:1: warning: data definition has no type or storage class
drivers/gpio/jbtrivial.c:275:1: warning: type defaults to ‘int’ in declaration of ‘MODULE_LICENSE’
drivers/gpio/jbtrivial.c:275:16: warning: function declaration isn’t a prototype

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>

The data direction logic was inversed, the correct logic should be:
0 = input, 1 = output

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

commit 54b956b ("Remove __dev* markings
from init.h") removed them, causing compile failures.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>

Add a DTS and defconfig for the Terasic DE0-Nano Development and Education
Board running ORPSoC.  This board contains an Altera Cyclone IV FPGA with
support chips and I/O.

The DTS was derived from an original by Gong Tao <gongtao0607@gmail.com>
with subsequent modifications by Kevin Mehall and Mark Czerski.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Kevin Mehall <contact@kevinmehall.net>
Cc: Marek Czerski <ma.czerski@gmail.com>
Signed-off-by: Jonas Bonn <jonas@southpole.se>

Motivation for this is to be able to print the way information
properly in print_cpuinfo(), instead of hardcoding it to one.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>
Signed-off-by: Jonas Bonn <jonas@southpole.se>

The syscall handler for semctl is written under the assumption that the
toolchain will pass "small" unions as function parameters directly instead
of by reference.  The union semun is "small" and thus fits this description.

Since it is assumed that the union will be passed directly and not by
reference, it is safe to access the union members without going via
get_user.

The OpenRISC architecture, however, passes all unions by reference, thus
breaking the above assumption.

The technically correct fix here is to mark the union as being transparent
so that the ABI of the union's first element determines the parameter
passing method and thus make explicit what's already implied in the function
definition.

Signed-off-by: Jonas Bonn <jonas@southpole.se>

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Jonas Bonn <jonas@southpole.se>

By slightly reorganizing the code, the number of registers
used in the tlb miss handlers can be reduced by two,
thus removing the need to save them to memory.

Also, some dead and commented out code is removed.

No functional change.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>
Signed-off-by: Jonas Bonn <jonas@southpole.se>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

On OpenRISC, with its 8k pages, PAGE_SHIFT is defined to be 13.
That makes the expression (1UL << (PAGE_SHIFT-2)) evaluate
to 2048.
The correct value for PTRS_PER_PGD should be 256.

Correcting the PTRS_PER_PGD define unveiled a bug in map_ram(),
where PTRS_PER_PGD was used when the intent was to iterate
over a set of page table entries.
This patch corrects that issue as well.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Return to userspace via _resume_userspace instead of via syscall return
path for the rt_sigreturn syscall.

I'll rework this comment more later, but this patch needs testing.

Old comment from previous patch:

The sigreturn syscall is more like a context switch than a function call;
it entails a return from one context (the signal handler) to another
(the process in question).  For a context switch like this there are
effectively no call-saved regs that remain constant across the transition.

This patch restores the call-saved regs from pt_regs before returning from
the syscall, effectively restoring the context that the process had before
being interrupted by the signal handler.  Restoring the call-saved regs
in this way allows us to return to userspace via the usual syscall fast
path.

Reported-by: Sebastian Macke <sebastian@macke.de>
Signed-off-by: Jonas Bonn <jonas@southpole.se>

Historically OpenRISC GCC has reserved r10 which we now use to hold
the thread pointer for thread-local storage (TLS).

Signed-off-by: Christian Svensson <blue@cmd.nu>

This is a follow up patch to Christian Svensson's TLS patch.
The original patch [1] haven't been applied yet, so this should
be fixed there.
This patch just serves as a reminder of that...

[1] http://lists.openrisc.net/pipermail/linux/2014-January/000548.html

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

In addition to consolidating the or1k-pic with other interrupt
controllers, this makes OpenRISC less tied to its on-cpu
interrupt controller.

All or1k-pic specific parts are moved out of irq.c and into
drivers/irqchip/irq-or1k-pic.c

In that transition, the funtionality have been divided into
three chip variants.
One that handles level triggered interrupts, one that handles edge
triggered interrupts and one that handles the interrupt
controller that is present in the or1200 OpenRISC cpu
implementation.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Previously, the area between 0x0-0x100 have been used as a
"scratch" memory area to temporarily store regs during
exception entry.
In a multi-core environment, this will not work.

This uses the SPR_ISR hack introduced by Stefan Wallentowitz
as a temporary solution until a proper "scratch spr" solution
have been designed.

Currently only the "critical" temp load/stores are covered,
the EMERGENCY_PRINT ones are left as is (when they are used,
it's game over anyway), they need to be handled as well in the
future.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

This utilize the load-link/store-conditional l.lwa and l.swa
instructions to implement the atomic bitops.
When those instructions are not available, a fallback to the
generic implementation is provided.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Optimized version that use the l.lwa and l.swa instructions

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Using the l.lwa and l.swa instructions

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Heavily based on the ARM implementation, this adds
ticket spinlock support for OpenRISC.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Simple enough to be compatible with simulation environments,
such as verilated systems etc.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Work-in-progress, contains stubs for the IPI stuff.
Apart from that secondary  bringup should be "working",
no guarantees though.
Boot loader spinning of second cpu is not supported yet,
it's assumed that Linux is booted straight after cpu reset.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

The reason why this is created as an array, and not with
DEFINE_PER_CPU is that doing it this way, we'll save a load
in the tlb-miss handler (the load from __per_cpu_offset).

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

This imitates the arm64 implementation by having a
smp_cross_call callback that can be set by set_smp_cross_call
to initiate an IPI and a handle_IPI function that is expected
to be called from an IPI irqchip driver.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

The SMP implementation of flush_tlb_* works by sending out
a function-call IPI to all the non-local cpus by using the
generic on_each_cpu() function.

Currently, all flush_tlb_* functions will result in a
flush_tlb_all(), which has always been the behaviour in the
UP case.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

In the SMP case, SPR_COREID is used to load the
correct cpu specific current_thread_info.
In the UP case, current_thread_info_set[0] is unconditionally
loaded.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Only IPI support is implemented so far and the whole driver
should be seen as tentative at this point.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

the opcode for l.swa is 0x33

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

And replace it with using a shadow regster file if such is
available.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Not all OpenRISC cpus have support for the l.lwa and l.swa,
this adds a config knob to opt them in and out.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

This utilize the load-link/store-conditional l.lwa and l.swa
instructions to implement the atomic bitops.
When those instructions are not available, a fallback to the
generic implementation is provided.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Optimized version that make use of the l.lwa and l.swa instructions
if available.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Optimized version that make use of the l.lwa and l.swa instructions
if available.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

The way to determine if a data pagefault was caused
by a write is to read the instruction at the PC that caused
the fault and check if it's a store instruction.
To recognize pagefaults caused by the l.swa instruction with
opcode 0x33, has to be included in the check.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Support for the futex_atomic_* operations by using the
load-link/store-conditional l.lwa/l.swa instructions.

A fallback to the generic implementation is provided when those
implementations aren't available.
The generic implementation however only provides stubs for the
functions, so to obtain proper futex support, the implementation
that is added here is needed.

Signed-off-by: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>

Conflicts:
	arch/openrisc/Kconfig
	arch/openrisc/include/asm/atomic.h

Linux 3.16

Commit a71e3c3 ("net: phy: Set the driver when registering an MDIO bus
device") caused the following regression on the fec driver:

root@imx6qsabresd:~# echo mem > /sys/power/state
PM: Syncing filesystems ... done.
Freezing user space processes ... (elapsed 0.003 seconds) done.
Freezing remaining freezable tasks ... (elapsed 0.002 seconds) done.
Unable to handle kernel NULL pointer dereference at virtual address 0000002c
pgd = bcd14000
[0000002c] *pgd=4d9e0831, *pte=00000000, *ppte=00000000
Internal error: Oops: 17 [#1] SMP ARM
Modules linked in:
CPU: 0 PID: 617 Comm: sh Not tainted 3.16.0 torvalds#17
task: bc0c4e00 ti: bceb6000 task.ti: bceb6000
PC is at fec_suspend+0x10/0x70
LR is at dpm_run_callback.isra.7+0x34/0x6c
pc : [<803f8a98>]    lr : [<80361f44>]    psr: 600f0013
sp : bceb7d70  ip : bceb7d88  fp : bceb7d84
r10: 8091523c  r9 : 00000000  r8 : bd88f478
r7 : 803f8a88  r6 : 81165988  r5 : 00000000  r4 : 00000000
r3 : 00000000  r2 : 00000000  r1 : bd88f478  r0 : bd88f478
Flags: nZCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment user
Control: 10c5387d  Table: 4cd1404a  DAC: 00000015
Process sh (pid: 617, stack limit = 0xbceb6240)
Stack: (0xbceb7d70 to 0xbceb8000)
....

The problem with the original commit is explained by Russell King:

"It has the effect (as can be seen from the oops) of attaching the MDIO bus
device (itself is a bus-less device) to the platform driver, which means
that if the platform driver supports power management, it will be called
to power manage the MDIO bus device.

Moreover, drivers do not expect to be called for power management
operations for devices which they haven't probed, and certainly not for
devices which aren't part of the same bus that the driver is registered
against."

This reverts commit a71e3c3.

Cc: <stable@vger.kernel.org> #3.16
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>