Ensure that we unmap and free a pending transmit skb when the interface
is stopped. We rearrange the code a little bit to give all places a
similar layout when freeing the skb in both the completion and interface
stop paths - this gives some consistency to the code.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Both the transmit and receive DMA store identical data: the skb, dma
address, and the dma registers. Move this data into its own data
structure. The following replacements were used:
rxskb -> dma_rx.skb
rxbuf_dma -> dma_rx.dma
rxdma -> dma_rx.regs
txskb -> dma_tx.skb
txbuf_dma -> dma_tx.dma
txdma -> dma_tx.regs
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Convert the sa11x0-ir driver to obtain its interrupt number from the
platform device resources, rather than via the asm/irq.h include.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Add the missing SET_NETDEV_DEV() call to set the parent device
correctly for this network interface.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Change the sa1111 device id to be a bitmask. This allows us to
specify the actual device, while allowing a single driver to bind
to both PS2 devices.
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Move the handling of the 5v supply into badge4.c, removing this board
specific detail from the sa1111 ohci driver.
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Add platform hooks to be called when individual sa1111 devices are
enabled and disabled. This will allow us to move some platform
specifics out of the individual drivers.
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Add a shutdown hook to the sa1111_driver structure to allow drivers
to be notified of system reboots and shutdowns.
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Implement the necessary allocation/freeing functionality to support
sparse IRQs with the SA-1111 device. On non-sparse IRQ platforms,
this allows us to dynamically allocate from within the available IRQ
number space.
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Commit 19851c58e6 (sa1111: allow cascaded IRQs to be used by platforms)
moved the IRQ definitions to the .c file, and added an irq_base member
to the private data structure.
The inerrupt demultiplexer uses irq_base, but the interrupt setup code
does not. Also, although the commit adds a private data structure to
pass this data, it isn't even referenced, resulting in irq_base being
zero.
We also copied the IRQ numbers from the device info array into the actual
devices, resulting in wrong interrupt numbers passed to the sub-devices.
The net effect of this is that we always overwrite IRQs 0-54, even if
they are allocated elsewhere in the system.
Add the code necessary to setup the private irq_base, and use it in the
IRQ setup code. Make the SA-1111 probe fail with -EINVAL if there is no
platform data provided.
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Add a .owner initializer to the sa1111 driver structures to allow
allow the modules to be associated with their driver structures.
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
We should not write to the SA1111 registers after setting the SLEEP
bit. Moreover, the manual says that the PWM registers should be
disabled before we enter sleep. So, move the clearing of these
registers earlier in the suspend sequence.
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
We weren't re-enabling the memory request/grant signals on resume,
causing DMA devices on the sa1111 to fail.
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Now that we ioremap() the neponset register space, there's no need
to static map the neponset registers. Get rid of this static mapping.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Move the board specific neponset register definitions to the board
file, rather than mach/neponset.h. However, as the NCR_0 register
definitions are used by some drivers, leave these behind.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
The neponset board is a daughter board for the Assabet. Create the
neponset platform device in assabet.c, where we don't have to wrap
it with machine_is_assabet() stuff. We also create this device
dynamically rather than keeping it as a static device.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Suspend and resume in the _noirq state, so that we're saving the
state of the modem control signals as late as possible, and restoring
them as early as possible. There's nothing to do in thaw/poweroff
methods as we've already saved the necessary state.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Complete the neponset device resources by covering the children's
memory resources in the parent neponset device.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Obtain the parent IRQ from the neponset device resource rather than
hard-coding it into the code.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Implement the necessary allocation/freeing functionality to support
sparse IRQs with the Neponset device. On non-sparse IRQ platforms,
this allows us to dynamically allocate from within the available IRQ
number space.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Use platform_device_register_full() to dynamically create the various
neponset child platform devices, and place them below the neponset
device itself to ensure proper PM ordering and device structure.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Save and restore the modem output control register across a suspend/
resume, as well as the NCR register. Place these in a locally
allocated data structure rather than needing a new static variable.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Rather than having direct register accesses to NCR_0 scattered amongst
the code, provide a function instead. This contains the necessary
race protection for this platform, ensuring that updates to this
register are safe.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Since ARM was converted to genirq, the neponset IRQ implementation has
gradually broken as a result of various subtle changes being introduced
into genirq.
It used to be that simple IRQs did not need an IRQ chip. This is no
longer the case, and genirq barfs in irq_set_handler(). Fix this by
introducing a dummy no-op chip, and registering it along with the flow
handler.
Neponset IRQs really don't have any masking ability - all we have is a
status register to allow us to decode the source, and a three input OR
gate inside a CPLD.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
GPIO27 is just connected to a CPLD input without any pull-ups or pull-
downs. If GPIO27 is left as an input, it will float around mid-supply,
which for CMOS inputs is the worst place for a pin to be. Ensure that
this pin is driven.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Avoid glitching the GPIO signals during initialization, which can
have undesirable effects. Ensure that the desired pin state is set
before we change the GPIO pin direction to be an output.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Neponset calls sa1110_mb_disable() from __devinit code, but
sa1110_mb_disable() is marked __init, and so causes a section
mismatch warning.
As sa1110_mb_enable() and sa1110_mb_disable() need to be callable
from suspend/resume paths as well, they must not be marked __init
or __devinit.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Sometimes, we get stuck while trying to enter sleep. This seems
to occur if we do not have udelay() in the instruction cache. Avoid
this by requesting a short delay prior to modifying the SDRAM timings.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Hackkit defined its flash memory resource to be 32M + 1 byte.
Jornada defined the Epson video controller resources to be one byte
larger than they should be, and mis-mapped the SA-1111 companion
chip one byte smaller than it should be.
Fix these.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
/*
* This is here for documentation purposes only - until these people
* submit their machine types. It will be gone January 2005.
*/
It's now seven years after that date, so let's remove this.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Commit c4a9f88daf ([MTD] [NOR] fix ctrl-alt-del can't reboot for
intel flash bug) interferes with this work-around, causing MTD to
issue this warning:
Flash device refused suspend due to active operation (state 0)
The commit makes our work-around in the map driver unnecessary, so
let's remove it.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Chris Wilson and me have again stared at funny error states and it's
been pretty clear from the start that something was seriously amiss.
The seqnos last seen by the cpu were a few hundred behind those that
the gpu could have possibly emitted last before it died ...
Chris now tracked it down (hopefully, definit verdict's still out),
but in hindsight we'd have found the bug by simply dumping the cpu
side tracking of the ring head and tail registers.
Fix this and prevent an identical time-waster in the future.
Because the hangs always involved semaphores in one way or another,
we've tried to dump the mbox registers, but couldn't find any
inconsistencies. Still, dump them too.
Reviewed-and-wanted-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The current version of perf detects whether or not the perf.data file is
written in a different endianness using the attr_size field in the
header of the file. This field represents sizeof(struct perf_event_attr)
as known to perf record. If the sizes do not match, then perf tries the
byte-swapped version. If they match, then the tool assumes a different
endianness.
The issue with the approach is that it assumes the size of
perf_event_attr always has to match between perf record and perf report.
However, the kernel perf_event ABI is extensible. New fields can be
added to struct perf_event_attr. Consequently, it is not possible to use
attr_size to detect endianness.
This patch takes another approach by using the magic number written at
the beginning of the perf.data file to detect endianness. The magic
number is an eight-byte signature. It's primary purpose is to identify
(signature) a perf.data file. But it could also be used to encode the
endianness.
The patch introduces a new value for this signature. The key difference
is that the signature is written differently in the file depending on
the endianness. Thus, by comparing the signature from the file with the
tool's own signature it is possible to detect endianness. The new
signature is "PERFILE2".
Backward compatiblity with existing perf.data file is ensured.
Tested-by: David Ahern <dsahern@gmail.com>
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Anshuman Khandual <khandual@linux.vnet.ibm.com>
Cc: Arun Sharma <asharma@fb.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Lin Ming <ming.m.lin@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Roberto Agostino Vitillo <ravitillo@lbl.gov>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Vince Weaver <vweaver1@eecs.utk.edu>
Link: http://lkml.kernel.org/r/1328187288-24395-15-git-send-email-eranian@google.com
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Since we've already got logic to special case immediate teardown of the
stream we may as well use it if the pmdown_time has been set to zero by
the application layer instead of scheduling a work item with zero delay.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@ti.com>
The drm drivers set the fb_info->pixmap fields without setting
fb_info->pixmap.addr. If this is not set the fb core will overwrite
these all fb_info->pixmap fields anyway, so there is not much point
in setting them in the first place.
[airlied: dropped nvidiafb piece - not mine]
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Creating a range property is a common pattern, so create
a convenience function for this and use it where appropriate.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Russell King