Move the creation of a random ethernet address from devices.c
into davinci_emac.c. This allows me to delete the davinci_emac_init
function and directly call the dmxxx specific function from the
board file.
Signed-off-by: Troy Kisky <troy.kisky@boundarydevices.com>

This reverts commit 1420b5d7.

Use the same mac order in emac_set_type0addr as in
emac_set_type1addr. The byte order was reversed.
Signed-off-by: Troy Kisky <troy.kisky@boundarydevices.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Signed-off-by: Stijn Devriendt <HIGHGuY@gmail.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

The scatterlist traversal is made proper. The patch has some minor comments
modification and also sets the value written to the DAVINCI_MMCTOR register
properly. The timeout calculation is made proper by deriving it from the
clock values.

This patch incorporates most of review comments given on LKLM by
Pierre Ossman.
Signed-off-by: Vipin Bhandari <vipin.bhandari@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Currently the auto generated file mach-types.h is checking for
MACH_DAVINCI_DM6467_EVM, while it is configured as
MACH_DAVINCI_DM646X_EVM in the Kconfigs and defconfigs.
So machine_is_davinci_dm6467_evm() always returns FALSE.
Signed-off-by: Naresh Medisetty <naresh@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Update NAND partitioning for the dm6446 evm, unmasking the hidden
data at the beginning and letting the kernel be updated from Linux.

 - This is boot-compatible with TI's software (U-Boot 1.20 and both
   the 2.6.10 and 2.6.18 kernels), in terms of startup and loading
   kernels from flash.

 - In the same way, it's also boot-compatible with mainline U-Boot,
   which stores U-Boot params in block 0 not block 16.

 - It's not quite compatible with systems that previously used NAND
   partitions to hold (filesystem) data.  The compatibilities are a
   bit different based on which kernel was used previously
     + Users of TI/MV kernels no longer see mtd2 "params"
       (mainline u-boot env is in a different place)
	* Filesystem is now mtd2 ... vs mtd3
     + Users of GIT kernels now see mtd0 and mtd1 partitions
	* Filesystem partition starts 640 KBytes earlier
	* Filesystem is now mtd2 ... vs mtd0
     * Linux now *uses* the flash-resident BBT
	* Removes annoying slowdown/hiccup during boot
	* Potentially ~64KB less space available with TI/MV kernels

If you *used* NAND partitions from Linux, there is no solution that's
fully compatible with all previous kernels in those respects ... ergo
this "best compromise".  It'd be good to back back up the filesystem
data; or, carry your own backwards-compatibility patch for awhile.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

The skb data allocated for packet data received is 4 byte aligned.

Unfortunately, this causes non-32bit aligned accesses in IP stack because
the MAC header is non-word aligned (14 bytes).

The result can be observed by looking at /proc/cpu/alignment while the
device is over network.

Doing an skb_reserve(NET_IP_ALIGN) as other driver do fixed the issue.

A quick performance test over lab network using iperf on DM6446 EVM showed
an increase in bandwidth from 60Mbits/s to 70Mbits/s.
Signed-off-by: Steve Chen <schen@mvista.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

The provision for extra bytes in each rx buffer was meant for allocators
who cannot guarantee cacheline aligned buffer allocation.

cacheline aligned allocation is guaranteed by alloc_skb so this provision is
not required.

This should save 32*128 = 4k of space.

Tested on DM6446 EVM using NFS root, iperf (TCP) and tftp of a gigabyte of data.
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Before this, if a "noevent" channel was in use, the code
would go into an infinite loop.
Signed-off-by: Troy Kisky <troy.kisky@boundarydevices.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Fix the definitions of at24_setup() in two davinci board files
that are incorrect.
Signed-off-by: Mark A. Greer <mgreer@mvista.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

From: David Brownell <dbrownell@users.sourceforge.net>

 - Define new setup() hook to export the accessor
 - Implement accessor methods

Moves some error checking out of the sysfs interface code
into the layer below it, which is now shared by both sysfs
and memory access code.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

In the case of at24, the platform code registers a 'setup' callback
with the at24_platform_data.  When the at24 driver detects an EEPROM,
it fills out the read and write functions of the memory_accessor and
calls the setup callback passing the memory_accessor struct.  The
platform code can then use the read/write functions in the
memory_accessor struct for reading and writing the EEPROM.
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

This patch adds an interface by which other kernel code can read/write
persistent memory such as I2C or SPI EEPROMs, or devices which provide
NVRAM.  Use cases include storage of board-specific configuration data
like Ethernet addresses and sensor calibrations.

Original idea, review and improvement suggestions by David Brownell.
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Clear any set bits in the 'NEXT' field of the MDCTL register in the
Power and Sleep Controller (PSC) before setting any new bits.
This also allows some minor cleanup by removing some no longer
needed lines of code.
Signed-off-by: Mark A. Greer <mgreer@mvista.com>

CONFIG_CPU_V5 not used anywhere so remove it.
Signed-off-by: Mark A. Greer <mgreer@mvista.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

According the to GAS manual, the ARM 'adr' instruction can only
reference symbols that are in the same section as the code.
Unfortunately, there was a bug in GAS that didn't flag that as
an error and simply used the address of the 'adr' instruction
itself instead of the symbol.  That bug has been reported and
fixed upstream.

Now that GAS is fixed, the code needs to be fixed by keeping
__davinci_start() and davinci_machine_no in the same section.
Signed-off-by: Mark A. Greer <mgreer@mvista.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

The DaVinci SoCs have 3 uarts so enable all three by setting
CONFIG_SERIAL_8250_NR_UARTS and CONFIG_SERIAL_8250_RUNTIME_UARTS
to 3.

The other changes to the defconfig file are artifacts of running
the old defconfig through 'make config'.
Signed-off-by: Mark A. Greer <mgreer@mvista.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

The dm355 leopard board is meant to be a low-cost IP netcam. It has
ethernet, USB OTG and SD support, as well as the ability to use
different sensor boards.
It is being promoted as "beagle buddy" to be used as a companion to the
OMAP3 beagleboard.
Signed-off-by: Koen Kooi <koen@beagleboard.org>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Clock and Pin Mux definitions for video on DM646x SoC

Add clock defintions and pin mux definitions for Video Port Interface(VPIF).
Signed-off-by: Chaithrika U S <chaithrika@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Since duplex is u32, use ~0 instead of -1 for clarity.
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

The duplex bit in MACCONTROL wasn't changed.
Signed-off-by: Troy Kisky <Troy.Kisky@boundarydevices.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

A pointer was being passed as an argument to sizeof instead
of the variable.

No need to send this patch to ALSA list as this fix is already present in that
tree.
Signed-off-by: Sudhakar Rajashekhara <sudhakar.raj@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Fixes the PIO mode for MMC/SD on TI's DA830/OMAP-L137 architecture.

On TI's DA830/OMAP-L137, the DMATRIG bit should be set for PIO
transfers.  If MMC is being used for both PIO and EDMA based
transfers, then it is likely that, an extra MMC EVENT is latched
in the EDMA Event register even when just the CPU and MMC are in
play. To properly switch from PIO to EDMA, clear any extra or
unexpected event latched in the Event register, by doing a write
to the corresponding bits in the ECR register.
Signed-off-by: Sudhakar Rajashekhara <sudhakar.raj@ti.com>

Adds edma_clear_event API support to EDMA.

There has been a change to the MMC interface on TI's DA830/OMAP-L137
architecture that requires setting the DMATRIG bit to trigger
the first read data transfer when doing a read operation using
the CPU, but it has the unintended consequence causing a DMA
event.  Because of this, the next DMA read from the MMC/SD will
fail with an Event Miss. This API will be used to clear the ER.
Signed-off-by: David Griego <dgriego@mvista.com>
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Sudhakar Rajashekhara <sudhakar.raj@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

This bug occurs when the davinci serial code tries to register UART(X) when
UART(X-1) or UART(X-2) is not enabled in the structure uart_config of the board
setup code:

This works: .enabled_uarts = (1 << 0)
This works: .enabled_uarts = (1 << 0) | (1 << 1)
This works: .enabled_uarts = (1 << 0) | (1 << 1) | | (1 << 2)
This fails: .enabled_uarts = (1 << 1)
This fails: .enabled_uarts = (1 << 1) | (1 << 2)
This fails: .enabled_uarts = (1 << 0) | (1 << 2)

The bug is triggered by the fact that the 8250 serial driver stops parsing the
serial_platform_data structure as soon as it sees a zero flags entry. Thus the
davinci serial registration code (serial.c) must <pack> the serial_platform_data
structure and only clear the flags entry when there is no more devices following.

Tested on DM6446 and DM355 custom boards.
Signed-off-by: Hugo Villeneuve <hugo@hugovil.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

Add support for Texas Instuments Common Platform Interrupt Controller
(cp_intc) used on DA830/OMAP-L137.
Signed-off-by: Steve Chen <schen@mvista.com>
Signed-off-by: Mark Greer <mgreer@mvista.com>
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>

With a sufficiently new compiler and binutils, code which wasn't
previously generating .eh_frame sections has begun to.  Certain
architectures (powerpc, in this case) may generate unexpected relocation
formats in response to this, preventing modules from loading.

While the new relocation types should probably be handled, revert to the
previous behaviour with regards to generation of .eh_frame sections.

(This was reported against Fedora, which appears to be the only distro
doing any building against gcc-4.4 at present: RH bz#486545.)
Signed-off-by: Kyle McMartin <kyle@redhat.com>
Acked-by: Roland McGrath <roland@redhat.com>
Cc: Alexandre Oliva <aoliva@redhat.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Revert the change to the orphan dates of Windows 95, DOS, compression.
Add a new orphan date for OS/2.
Signed-off-by: Jody McIntyre <scjody@sun.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (32 commits)
  ucc_geth: Fix oops when using fixed-link support
  dm9000: locking bugfix
  net: update dnet.c for bus_id removal
  dnet: DNET should depend on HAS_IOMEM
  dca: add missing copyright/license headers
  nl80211: Check that function pointer != NULL before using it
  sungem: missing net_device_ops
  be2net: fix to restore vlan ids into BE2 during a IF DOWN->UP cycle
  be2net: replenish when posting to rx-queue is starved in out of mem conditions
  bas_gigaset: correctly allocate USB interrupt transfer buffer
  smsc911x: reset last known duplex and carrier on open
  sh_eth: Fix mistake of the address of SH7763
  sh_eth: Change handling of IRQ
  netns: oops in ip[6]_frag_reasm incrementing stats
  net: kfree(napi->skb) => kfree_skb
  net: fix sctp breakage
  ipv6: fix display of local and remote sit endpoints
  net: Document /proc/sys/net/core/netdev_budget
  tulip: fix crash on iface up with shirq debug
  virtio_net: Make virtio_net support carrier detection
  ...

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  sparc64: Fix crash with /proc/iomem
  sparc64: Reschedule KGDB capture to a software interrupt.
  sbus: Auto-load openprom module when device opened.

This patch fixes bug #12208:

  Bug-Entry       : http://bugzilla.kernel.org/show_bug.cgi?id=12208
  Subject         : uml is very slow on 2.6.28 host

This turned out to be not a scheduler regression, but an already
existing problem in ptrace being triggered by subtle scheduler
changes.

The problem is this:

 - task A is ptracing task B
 - task B stops on a trace event
 - task A is woken up and preempts task B
 - task A calls ptrace on task B, which does ptrace_check_attach()
 - this calls wait_task_inactive(), which sees that task B is still on the runq
 - task A goes to sleep for a jiffy
 - ...

Since UML does lots of the above sequences, those jiffies quickly add
up to make it slow as hell.

This patch solves this by not rescheduling in read_unlock() after
ptrace_stop() has woken up the tracer.

Thanks to Oleg Nesterov and Ingo Molnar for the feedback.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
CC: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/galak/powerpc:
  powerpc/mm: Fix Respect _PAGE_COHERENT on classic ppc32 SW TLB load machines

Grant picked up the wrong version of "Respect _PAGE_COHERENT on classic
ppc32 SW" (commit a4bd6a93)

It was missing the code to actually deal with the fixup of
_PAGE_COHERENT based on the CPU feature.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>