The revision of the original support card (rev 3.5, rev 3.6) fits in
the 8 bit width revision register.  When it was extended in a weird
way, it was versioned in the format of "3.6.x" (where it should have
been "3.7", of course).  What is worse, only the sub-level version
"6.x" was recorded in the 8 bit width register, completely ignoring
the compatibility of the revision register format.

This patch saves madly-versioned support cards by assuming the major
version "3" when the MSB 4 bit of the register is read as "6".  With
this, the support card revision that were displayed as "6.10" is now
corrected to "3.6.10".
Signed-off-by: NMasahiro Yamada <yamada.masahiro@socionext.com>

This supports the system reset via PSCI for ARMv7 SoCs.

Because the system reset is not supported on PSCI 0.1, let's define
CONFIG_ARMV7_PSCI_1_0. (it is supported since PSCI 0.2, but there
is no CONFIG to enable it in U-Boot for now.)
Signed-off-by: NMasahiro Yamada <yamada.masahiro@socionext.com>

With this, we can save unnecessary udelay().
Signed-off-by: NMasahiro Yamada <yamada.masahiro@socionext.com>

With the CONFIG_DM_MMC_OPS migration, the .set_ios callback can
return an integer now.  Return an appropriate error value rather
than sudden death by BUG().
Signed-off-by: NMasahiro Yamada <yamada.masahiro@socionext.com>

No more reason to define this function above the ops structure.
Move it near the caller.  Also, change its return type to void
because it never fails.
Signed-off-by: NMasahiro Yamada <yamada.masahiro@socionext.com>

Catch up with the DM migration.

As struct dm_mmc_ops does not have .init callback, call the init
function directly from the probe function.
Signed-off-by: NMasahiro Yamada <yamada.masahiro@socionext.com>

They are both only referenced in this file.
Signed-off-by: NMasahiro Yamada <yamada.masahiro@socionext.com>

Commit f4db6c97 ("arm: mvebu: Add runtime detection of UART (xmodem)
boot-mode") added a change to hdr->destaddr when dynamically patching an
image for UART boot mode.  With this change, kwboot ceases to work on
Kirkwood.

Thus, let's change hdr->destaddr only when we are patching an image with
header version 1 (Orion and Kirkwood use header version 0).
Signed-off-by: NSimon Baatz <gmbnomis@gmail.com>
Fixes: f4db6c97 ("arm: mvebu: Add runtime detection of UART (xmodem) boot-mode")
Cc: Stefan Roese <sr@denx.de>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Cc: Kevin Smith <kevin.smith@elecsyscorp.com>
Signed-off-by: NStefan Roese <sr@denx.de>

Signed-off-by: NChris Packham <judge.packham@gmail.com>
Signed-off-by: NStefan Roese <sr@denx.de>

Ensure appropriate error messages are generated. Previously all errors
indicated that the serdes was already in use. Now appropriate error
messages are given.
Signed-off-by: NChris Packham <judge.packham@gmail.com>
Signed-off-by: NStefan Roese <sr@denx.de>

As of commit 88e34e5f ("spl: replace CONFIG_SPL_SPI_* with
CONFIG_SF_DEFAULT_*") these defines are not used. Remove them to avoid
confusion.
Signed-off-by: NChris Packham <judge.packham@gmail.com>
Signed-off-by: NStefan Roese <sr@denx.de>

The NAND interface on the Armada-38x series is similar to that on the
Armada-XP. The key difference is that the NAND ECC clock ratio is
provided via the DFX Server registers instead of the Core Clock.
Signed-off-by: NChris Packham <chris.packham@alliedtelesis.co.nz>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Cc: Dirk Eibach <eibach@gdsys.de>
Signed-off-by: NStefan Roese <sr@denx.de>

Patch f8a10ed1 [i2c: mvtwsi: Make address length variable] accidentally
inverted the sequence of address bytes sent to the I2C device. This
patch corrects this by sending the highest byte first and the lowest
byte last again.

Tested on theadorable Armada-XP board.
Signed-off-by: NStefan Roese <sr@denx.de>
Cc: Mario Six <mario.six@gdsys.cc>
Cc: Heiko Schocher <hs@denx.de>

In tegra20_slink.c, the set_mode() function may be executed before the
SPI bus is claimed the first time, and hence the clocks to the SPI
controller may not be running. If so, any register read/write at this
time will hang the CPU. Fix this by ensuring the clock is running as soon
as the driver is probed. This is observed on the Tegra30 Beaver board.

Apply the same clock initialization fix to all other Tegra SPI drivers so
that if set_mode() is ever implemented there, the same bug will not appear.
Note that tegra114_spi.c already operates in this fashion.

The clock manipulation code is copied from claim_bus() to probe() rather
than moved. This ensures that any calls to set_speed() take effect; the
clock can't be set once during probe and left unchanged.

Fixes: 5cb1b7b3 ("spi: tegra20: Add support for mode selection")
Cc: Mirza Krak <mirza.krak@hostmobility.com>
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Reviewed-by: NSimon Glass <sjg@chromium.org>
Signed-off-by: NTom Warren <twarren@nvidia.com>

The nvidia,bpmp property is left over from an old BPMP I2C binding, and
shouldn't be present. Remove it from the SoC DT file, and update the
I2C driver not to parse it; the value wasn't used for anything any more
anyway.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NTom Warren <twarren@nvidia.com>

The Tegra SDHCI binding dictates that the reseet name for the Tegra SDHCI
clock be "sdhci" not "sdmmc", and that the clock is accessed by index
rather than by name. Fix the Tegra186 DT and MMC driver to honor this.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NTom Warren <twarren@nvidia.com>

The Tegra I2C binding dictates that the clock name for the Tegra I2C clock
be "div-clk" not "i2c". Fix the Tegra186 DT and I2C driver to honor this.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NTom Warren <twarren@nvidia.com>

Signed-off-by: NTom Rini <trini@konsulko.com>

The fs-test.sh script expected there to be a \n\r style newline at the
end of the output. This is no longer the case, so use 'tr' to remove the
\r that we get.

Fixes: (c5917b4b "dm: serial-uclass: Move a carriage return before a
        line feed")
Signed-off-by: NTom Rini <trini@konsulko.com>

When phy autoneg on, the link speed and duplex should be
determined by phy advertising register and
phy link partner ability register.
Check phy advertising register when geting phy link speed and
duplex if autoneg on.
Signed-off-by: NDongpo Li <lidongpo@hisilicon.com>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: NJoe Hershberger <joe.hershberger@ni.com>

The spatch series converting legacy drivers from miiphy_register to
mdio_register changed the return convention of the davinci_emac internal
MDIO accessors, making the internal code relying on it misbehaving:
no mdiodev get registered and U-Boot crashes when using net cmds in the
context of the old legacy net API.

ATM davinci_emac_initialize and cpu_eth_init don't return a proper value
in that case but fixing them would not avoid the crash.

This change is just a follow-up to the spatch pass, the MDIO accessors
of the mdiodev introduced by the spatch pass retain their proper values.
Signed-off-by: NKarl Beldan <karl.beldan+oss@gmail.com>
Acked-by: NJoe Hershberger <joe.hershberger@ni.com>

The QMan is not used in FMan IM mode, so no QMI enqueue or QMI
dequeue are performed.
Signed-off-by: NHou Zhiqiang <Zhiqiang.Hou@nxp.com>
Acked-by: NJoe Hershberger <joe.hershberger@ni.com>

ATM when receiving a packet the whole buffer is invalidated, this change
optimizes this behaviour.
Signed-off-by: NKarl Beldan <karl.beldan+oss@gmail.com>
Acked-by: NJoe Hershberger <joe.hershberger@ni.com>
Reviewed-by: NTom Rini <trini@konsulko.com>
Reviewed-by: NMugunthan V N <mugunthanvnm@ti.com>

check_cache_range() warns that the top boundaries are not properly
aligned when flushing or invalidating the buffers and make these
operations fail.

This gets rid of the remaining warnings:
CACHE: Misaligned operation at range
Signed-off-by: NKarl Beldan <karl.beldan+oss@gmail.com>
Acked-by: NJoe Hershberger <joe.hershberger@ni.com>
Reviewed-by: NTom Rini <trini@konsulko.com>
Reviewed-by: NMugunthan V N <mugunthanvnm@ti.com>

ATM the rx and tx descriptors are handled as cached memory while they
lie in a dedicated RAM of the SoCs, which is an uncached area.
Removing the said dcache ops, while optimizing the logic and clarifying
the code, also gets rid of most of the check_cache_range() incurred
warnings:
CACHE: Misaligned operation at range
Signed-off-by: NKarl Beldan <karl.beldan+oss@gmail.com>
Acked-by: NJoe Hershberger <joe.hershberger@ni.com>

We use an empty hostname, so remove all the "processing" of the
known-to-be-empty hostname and just write 0's where needed.
Signed-off-by: NJoe Hershberger <joe.hershberger@ni.com>

Instead of always allocating a huge temporary buffer on the stack and
then memcpy()ing the result into the transmit buffer, simply figure out
where in the transmit buffer the bytes will belong and write them there
directly as each message is built.
Signed-off-by: NJoe Hershberger <joe.hershberger@ni.com>

Much of the information is verbose and derived directly from the
environment. Only output in debug mode. This also saves about 300 bytes
from the code size.
Signed-off-by: NJoe Hershberger <joe.hershberger@ni.com>

Use the same name throughout the nfs code and use the same member of the
union to avoid casts.
Signed-off-by: NJoe Hershberger <joe.hershberger@ni.com>

The buffer is of 32-bit elements, not bytes.
Signed-off-by: NJoe Hershberger <joe.hershberger@ni.com>

Instead of repeating the same large snippet for dealing with attributes
it should be shared with a helper function.
Signed-off-by: NJoe Hershberger <joe.hershberger@ni.com>

Fix complaints from checkpatch.pl.
Signed-off-by: NJoe Hershberger <joe.hershberger@ni.com>

The type of the buffer is uint32_t, but the parameter used to size it
is referring to bytes. Divide by the size of the array elements.

Strictly speaking, this shouldn't be needed at all... It could just be 1
just like the request.
Signed-off-by: NJoe Hershberger <joe.hershberger@ni.com>

The v3 handles can be larger than v2, but that doesn't mean we need a
separate buffer. Reuse the same (larger) buffer for both.
Signed-off-by: NJoe Hershberger <joe.hershberger@ni.com>

This patch enables NFSv3 support.
If NFSv2 is available use it as usual.
If NFSv2 is not available, but NFSv3 is available, use NFSv3.
If NFSv2 and NFSv3 are not available, print an error message since NFSv4 is not supported.

Tested on iMX6 sabrelite with 4 Linux NFS servers:
  * NFSv2 + NFSv3 + NFSv4 server: use NFSv2 protocol
  * NFSv2 + NFSv3 server: use NFSv2 protocol
  * NFSv3 + NFSv4 server: use NFSv3 protocol
  * NFSv3 server: use NFSv3 protocol
Signed-off-by: NGuillaume GARDET <guillaume.gardet@free.fr>
Cc: Tom Rini <trini@konsulko.com>
Cc: joe.hershberger@ni.com
Acked-by: NJoe Hershberger <joe.hershberger@ni.com>

Unreferenced, so remove the noise.
Signed-off-by: NJoe Hershberger <joe.hershberger@ni.com>

There is no reason to store the default filename in a separate buffer
only to immediately copy it to the main name buffer. Just write it there
directly and remove the other buffer.
Signed-off-by: NJoe Hershberger <joe.hershberger@ni.com>