Since SPI register access is so expensive, it is worth transferring data
a word at a time if we can. This complicates the driver unfortunately.
Use the byte-swapping feature to avoid having to convert to/from big
endian in software.
This change increases speed from about 2MB/s to about 4.5MB/s.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Rajeshwari S Shinde <rajeshwari.s@samsung.com>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
Accessing SPI registers is slow, but access to the FIFO level register
in particular seems to be extraordinarily expensive (I measure up to
600ns). Perhaps it is required to synchronise with the SPI byte output
logic which might run at 1/8th of the 40MHz SPI speed (just a guess).
Reduce access to this register by filling up and emptying FIFOs
more completely, rather than just one word each time around the inner
loop.
Since the rxfifo value will now likely be much greater that what we read
before we fill the txfifo, we only fill the txfifo halfway. This is
because if the txfifo is empty, but the rxfifo has data in it, then writing
too much data to the txfifo may overflow the rxfifo as data arrives.
This speeds up SPI flash reading from about 1MB/s to about 2MB/s on snow.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Rajeshwari S Shinde <rajeshwari.s@samsung.com>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
For devices that need some time to react after a spi transaction
finishes, add the ability to set a delay.
Implement this as a delay on the first/next transaction to avoid
any delay in the fairly common case where a SPI transaction is
followed by other processing.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Rajeshwari S Shinde <rajeshwari.s@samsung.com>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
The chipselect must be written into the CTRL0 register after the SSP
block is reset, otherwise the block will always use ChipSelect #0.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
Acked-by: Stefano Babic <sbabic@denx.de>
85xx, 86xx PowerPC folders have code variables with CamelCase naming conventions.
because of this code checkpatch script generates "WARNING: Avoid CamelCase".
Convert variables name to normal naming convention and modify board, driver
files with updated the new structure.
Signed-off-by: Prabhakar Kushwaha <prabhakar@freescale.com>
Acked-by: York Sun <yorksun@freescale.com>
Use DIV_ROUND_UP to simplify the code.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Fix a trivial conflict in arch/arm/dts/exynos5250.dtsi about gpio and
serial.
Conflicts:
arch/arm/dts/exynos5250.dtsi
Signed-off-by: Tom Rini <trini@ti.com>
The spi clock divisor is of the form x * (2**y), or x << y, where x is
1 to 16, and y is 0 to 15. Note the similarity with floating point numbers.
Convert the desired divisor to the smallest number which is >= desired divisor,
and can be represented in this form. The previous algorithm chose a divisor
which could be almost twice as large as needed.
Signed-off-by: Troy Kisky <troy.kisky@boundarydevices.com>
Signed-off-by: Dirk Behme <dirk.behme@gmail.com>
Fix two issues with the calculation of pre_div and post_div:
1. pre_div: While the calculation of pre_div looks correct, to set the
CONREG[15-12] bits pre_div needs to be decremented by 1:
The i.MX 6Dual/6Quad Applications Processor Reference Manual (IMX6DQRM
Rev. 0, 11/2012) states:
CONREG[15-12]: PRE_DIVIDER
0000 Divide by 1
0001 Divide by 2
0010 Divide by 3
...
1101 Divide by 14
1110 Divide by 15
1111 Divide by 16
I.e. if we want to divide by 2, we have to write 1 to CONREG[15-12].
2. In case the post divider becomes necessary, pre_div will be divided by
16. So set pre_div to 16, too. And not 15.
Both issues above are tested using the following examples:
clk_src = 60000000 (60MHz, default i.MX6 ECSPI clock)
a) max_hz == 23000000 (23MHz, max i.MX6 ECSPI read clock)
-> pre_div = 3 (divide by 3 => CONREG[15-12] == 2)
-> post_div = 0 (divide by 1 => CONREG[11- 8] == 0)
=> 60MHz / 3 = 20MHz SPI clock
b) max_hz == 2000000 (2MHz)
-> pre_div = 16 (divide by 16 => CONREG[15-12] == 15)
-> post_div = 1 (divide by 2 => CONREG[11- 8] == 1)
=> 60MHz / 32 = 1.875MHz SPI clock
c) max_hz == 1000000 (1MHz)
-> pre_div = 16 (divide by 16 => CONREG[15-12] == 15)
-> post_div = 2 (divide by 4 => CONREG[11- 8] == 2)
=> 60MHz / 64 = 937.5kHz SPI clock
d) max_hz == 500000 (500kHz)
-> pre_div = 16 (divide by 16 => CONREG[15-12] == 15)
-> post_div = 3 (divide by 8 => CONREG[11- 8] == 3)
=> 60MHz / 128 = 468.75kHz SPI clock
Signed-off-by: Dirk Behme <dirk.behme@gmail.com>
This patch adds SPI support for carrying out the cros_ec protocol.
Signed-off-by: Hung-ying Tyan <tyanh@chromium.org>
Signed-off-by: Randall Spangler <rspangler@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
This change slightly improves readability.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
This change slightly improves readability.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Richard Retanubun <richardretanubun@ruggedcom.com>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
It's done in spi_alloc_slave(), thus remove the redundant code.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Marek Vasut <marex@denx.de>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Marek Vasut <marex@denx.de>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Ajay Bhargav <ajay.bhargav@einfochips.com>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
Support interfaces with a preamble before each received message.
We handle this when the client has requested a SPI_XFER_END, meaning
that we must close of the transaction. In this case we read until we
see the preamble (or a timeout occurs), skipping all data before and
including the preamble. The client will receive only data bytes after
the preamble.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Rajeshwari Shinde <rajeshwari.s@samsung.com>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
There may be dirty data in RDBR, so we should discard invalid data.
This operation also clears RXS bit in STAT register.
Signed-off-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
BF5xx rx dma causes spi flash random read error.
Accually spi controller has problems both on tx and rx dma.
So remove spi dma support in u-boot.
Signed-off-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Using IP version to check whether it has wdrbt bit in mode register
Tested in at91sam9x5ek and at91sam9n12ek.
Signed-off-by: Bo Shen <voice.shen@atmel.com>
[fix warning about incompatible parameter]
Signed-off-by: Josh Wu <josh.wu@atmel.com>
Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
The following headers are moved to a i.MX common location:
- regs-common.h
- regs-apbh.h
- regs-bch.h
- regs-gpmi.h
- dma.h
This way this header can be re-used also by other i.MX platforms.
For example the i.MX6 which will need it for the upcoming NAND
support.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
The glitch in the SPI clock line, which commit 3cea335c34 (spi: mxc_spi: Fix spi
clock glitch durant reset) solved, is back now and itwas re-introduced by
commit d36b39bf0d (spi: mxc_spi: Fix ECSPI reset handling).
Actually the glitch is happening due to always toggling between slave mode
and master mode by configuring the CHANNEL_MODE bits in this reset function.
Since the spi driver only supports master mode, set the mode for all channels
always to master mode in order to have a stable, "glitch-free" SPI clock line.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Reviewing the ECSPI reset handling shows two issues:
1. For the enable/reset bit (MXC_CSPICTRL_EN) in the control reg
(ECSPIx_CONGREG) the i.MX6 technical reference manual states:
-- cut --
ECSPIx_CONREG[0]: EN: Writing zero to this bit disables the block
and resets the internal logic with the exception of the ECSPI_CONREG.
-- cut --
Note the exception mentioned: The CONREG itself isn't reset.
Fix this by manually writing the reset value 0 to the whole register.
This sets the EN bit to zero, too (i.e. includes the old
~MXC_CSPICTRL_EN).
2. We want to reset the whole SPI block here. So it makes no sense
to first read the old value of the CONREG and write it back, later.
This will give us the old (historic/random) value of the CONREG back.
And doesn't reset the CONREG.
To get a clean CONREG after the reset of the block, too, don't use
the old (historic/random) value of the CONREG while doing the reset.
And read the clean CONREG after the reset.
This was found while working on a SPI boot device where the i.MX6 boot
ROM has already initialized the SPI block. The initialization by the
boot ROM might be different to what the U-Boot driver wants to configure.
I.e. we need a clean reset of SPI block, including the CONREG.
Signed-off-by: Dirk Behme <dirk.behme@de.bosch.com>
CC: Stefano Babic <sbabic@denx.de>
CC: Fabio Estevam <fabio.estevam@freescale.com>
'bool' is defined in random places. This patch consolidates them into a
single header file include/linux/types.h, using stdbool.h introduced in C99.
All other #define, typedef and enum are removed. They are all consistent with
true = 1, false = 0.
Replace FALSE, False with false. Replace TRUE, True with true.
Skip *.py, *.php, lib/* files.
Signed-off-by: York Sun <yorksun@freescale.com>
Enclosing process_nodes() and spi_get_config() inside
CONFIG_OF_CONTROL, since they are compiled only for DT systems.
This fixes following warning:
exynos_spi.c:391:12: warning: 'process_nodes' defined but not used [-Wunused-function]
Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Add driver for tegra114 SPI controller. This controller is not
compatible with either the tegra20 or tegra30 controllers, so it
requires a new driver.
Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Add a common interface to fdt based SPI drivers. Each driver is
represented by a table entry in fdt_spi_drivers[]. If there are
multiple SPI drivers in the table, the first driver to return success
from spi_init() will be registered as the SPI driver.
Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Make the tegra20 SPI driver similar to the tegra30 (and soon to be
tegra114) SPI drivers in preparation of common fdt SPI driver front
end.
Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Move register structs from headers into .c files and use common name.
This is in preparation of making common fdt front end for SPI
drivers.
Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Remove non fdt support from tegra20 and tegra30 SPI drivers in
preparation of new common fdt based SPI driver front end.
Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Rename tegra SPI drivers to tegra20_flash and tegra20_slink in
preparation for commonization and addition of tegra114_spi.
Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
This feature was only used for tegra20 seaboard that had a pinmux
conflict on the SPI pins. These boards were never manufactured, so
remove this support to clean up SPI driver.
Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
This SPI controller can only write 64 bytes at a time. Add this restriction
in so that 'sf write' works correct for blocks larger than 64 bytes.
Signed-off-by: Simon Glass <sjg@chromium.org>
This supports Intel ICH7/9. The Intel controller is a little unusual in
that it is mostly intended for use with SPI flash, and has some
optimisations and features specifically for that application. In
particular it is not possible to support ongoing transactions that
continue over many calls with SPI_XFER_BEGIN and SPI_XFER_END.
This driver supports writes of up to 64 bytes at a time, the limit
for the controller. Future work will improve this.
Signed-off-by: Bernie Thompson <bhthompson@chromium.org>
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Signed-off-by: Bill Richardson <wfrichar@chromium.org>
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Signed-off-by: Gabe Black <gabeblack@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Rather than each driver having its own way to allocate a SPI slave,
use the new allocation function everywhere. This will make it easier
to extend the interface without breaking drivers.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present it is difficult to extend the SPI structure since all
drivers allocate it themselves, and few of them zero all fields. Add
a new function spi_alloc_slave() which can be used by SPI drivers
to perform this allocation, and thus ensure that all drivers can
better cope with SPI structure changes.
Signed-off-by: Simon Glass <sjg@chromium.org>
The CONFIG_MXS_SPI_DMA_ENABLE is no longer relevant as the SPI DMA
has proven to work correctly. Remove this configuration option.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
The MX23 has slightly different register layout. Adjust the SPI
driver to match the layout, both the PIO and DMA part.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
This patch streamlines the differences between the MX23 and MX28 by
implementing a few helper functions to handle different DMA channel
mapping, different clock domain for SSP block and fixes a few minor
bugs.
First of all, the DMA channel mapping is now fixed in dma.h by defining
the actual channel map for both MX23 and MX28. Thus, MX23 now does no
longer use MX28 channel map which was wrong. Also, there is a fix for
MX28 DMA channel map, where the last four channels were incorrect.
Next, because correct DMA channel map is in place, the mxs_dma_init_channel()
call now bases the channel ID starting from SSP port #0. This removes the
need for DMA channel offset being added and cleans up the code. For the
same reason, the SSP0 offset can now be used in mxs_dma_desc_append(), thus
no need to adjust dma channel number in the driver either.
Lastly, the SSP clock ID is now retrieved by calling mxs_ssp_clock_by_bus()
which handles the fact that MX23 has shared SSP clock for both ports, while
MX28 has per-port SSP clock.
Finally, the mxs_ssp_bus_id_valid() pulls out two implementations of the
same functionality from MMC and SPI driver into common code.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
Spi driver for bf60x is different from old one, so implement a new
driver for it.
Signed-off-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Bob Liu <lliubbo@gmail.com>
Signed-off-by: Sonic Zhang <sonic.adi@gmail.com>
Add driver for tegra SPI "SLINK" style driver. This controller is
similar to the tegra20 SPI "SFLASH" controller. The difference is
that the SLINK controller is a genernal purpose SPI controller and the
SFLASH controller is special purpose and can only talk to FLASH
devices. In addition there are potentially many instances of an SLINK
controller on tegra and only a single instance of SFLASH. Tegra20 is
currently ths only version of tegra that instantiates an SFLASH
controller.
This driver supports basic PIO mode of operation and is configurable
(CONFIG_OF_CONTROL) to be driven off devicetree bindings. Up to 4
devices per controller may be attached, although typically only a
single chip select line is exposed from tegra per controller so in
reality this is usually limited to 1.
To enable this driver, use CONFIG_TEGRA_SLINK
Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Add support for configuring tegra SPI driver from devicetree.
Support is keyed off CONFIG_OF_CONTROL. Add entry in seaboard dts
file for spi controller to describe seaboard spi.
Signed-off-by: Allen Martin <amartin@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Create function which converts SSP bus number to SSP register pointer.
This functionality is reimplemented multiple times in the code, thus
make one common implementation. Moreover, make it a switch(), since the
SSP ports are not mapped in such nice linear fashion on MX23, therefore
having it a switch will simplify things there.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Andy Fleming <afleming@freescale.com>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Stefano Babic <sbabic@denx.de>
This saves stack and code memory for local copy, and
consumes initialized data memory. For 22 of the 29
kirkwood-based boards, this results in a global saving
of about 30 bytes. For 7 of them, it results in an
increase of 6 to 14 bytes.
Signed-off-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
This patch adds FDT support to the SPI driver.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Rajeshwari Shinde <rajeshwari.s@samsung.com>
Acked-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Added posted writes (read after writes) to effect the
change immediately for channel confiuration and channel
enable register
Disable the channel to purge receieve data in TX_ONLY
mode transfer otherwise rx data will get affected by
the next immediate RX_ONLY mode transfer
Wait for the EOT bit to be set after last byte has been
loaded to TX shift register in the the TX_ONLY mode.This
ensures TX data has been completely shifted out
Disable the channel in RX_ONLY mode before reading the
last data from RXX register to prevent the SPI slave
to transmit next word
Signed-off-by: Ajoy Kumar Das <akdas75@yahoo.in>
Cc: Tom Rini <trini@ti.com>
Cc: jacopo mondi <j.mondi@voltaelectronics.com>
The at91sam9m10g45 also support WDRBT bit, add support for it
Signed-off-by: Bo Shen <voice.shen@atmel.com>
Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
Measuring the spi clock line on a scope shows a 'glitch' during the reset of the
spi.
Fix this by toggling only the MXC_CSPICTRL_EN bit, so that the clock line becomes
always stable.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Acked-by: Stefano Babic <sbabic@denx.de>
In decode_cs() function the polarity of the chip select must be taken into
account.
Also, for the case of low active chip select, the CS was activated too early.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
This patch adds SPI driver for EXYNOS.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Padmavathi Venna <padma.v@samsung.com>
Signed-off-by: Gabe Black <gabeblack@google.com>
Signed-off-by: Rajeshwari Shinde <rajeshwari.s@samsung.com>
Signed-off-by: Hatim Ali <hatim.rv@samsung.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: jy0922.shim@samsung.com
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
D0/D1 Swapped or not is a board property, not anything specific to
the am33xx SoC, so add a custom define for it.
At the same time correct the bit handling for the swapped mode
(DPE0 should be cleared and SI/DPE1 set).
Signed-off-by: Peter Korsgaard <peter.korsgaard@barco.com>
Current espi controller driver assumes the command length of write command is
not equal to '1', it was made based on SPANSION SPI flash, but some SPI flash
driver such as SST does use write command length as '1', so write command on
SST SPI flash will not work. And the length check for write command is not
necessary for SPANSION, though it's harmless for SPANSION, it will stop write
operation on flashes like SST, so we remove the check.
Signed-off-by: Shaohui Xie <Shaohui.Xie@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
The move is pretty straight-forward. ap20.h and tegra20.h were renamed to ap.h and tegra.h.
Some files remain in arch-tegra20 but 'include' a file in 'arch-tegra' with #defines & structs
that will be common between T20 and T30 HW. HW-specific #defines, etc. stay in the 'arch-tegra20'
'root' file.
All boards build OK w/MAKEALL -s tegra20. Checkpatch.pl runs clean. Seaboard works OK.
Signed-off-by: Tom Warren <twarren@nvidia.com>
This patch adds a driver for Freescale Colfire Queued SPI bus.
Coded to work with 8 bits per transfer to use with SPI flash.
CPOL, CPHA, and CS_ACTIVE_HIGH can be configured.
Tested with MCF5270 which have 4 chip selects.
Activate by #define CONFIG_CF_QSPI in board config.
Signed-off-by: Richard Retanubun <richardretanubun@ruggedcom.com>
The MPC8308RDB Reference Manual states that no bits in the SPMODE
register are allowed to change while the enable (EN) bit is set.
This driver changes the character length bits (LEN) while the enable
(EN) bit is set. Clearing the EN bit while changing the LEN bits makes
the driver work correctly on MPC8308RDB.
Signed-off-by: Ira W. Snyder <iws@ovro.caltech.edu>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Warning is:
xilinx_spi.c: In function 'spi_setup_slave':
xilinx_spi.c:81: warning: unused variable 'regs'
Signed-off-by: Stephan Linz <linz@li-pro.net>
CC: Michal Simek <monstr@monstr.eu>
Convert TEGRA20_ defines to either TEGRA_ or NV_PA_ where appropriate.
Convert tegra20_ source file and function names to tegra_, also.
Upcoming Tegra30 port will use common code/defines/names where possible.
Signed-off-by: Tom Warren <twarren@nvidia.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
It turns out that in order for the SPI DMA to properly support
continuous transfers longer than 65280 bytes, there are some very
important parts that were left out from the documentation.
Firstly, the XFER_SIZE register is not written with the whole length
of a transfer, but is written by each and every chained descriptor
with the length of the descriptors data buffer.
Next, unlike the demo code supplied by FSL, which only writes one PIO
word per descriptor, this does not apply if the descriptors are chained,
since the XFER_SIZE register must be written. Therefore, it is essential
to use four PIO words, CTRL0, CMD0, CMD1, XFER_SIZE. CMD0 and CMD1 are
written with zero, since they don't apply. The DMA programs the PIO words
in an incrementing order, so four PIO words.
Finally, unlike the demo code supplied by FSL, the SSP_CTRL0_IGNORE_CRC
must not be set during the whole transfer, but it must be set only on the
last descriptor in the chain.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
This patch fixes dcache-related problem. The problem manifested
when dcache was enabled and the following command issued twice:
mw 0x42000000 0 0x4000 ; sf probe ; sf read 0x42000000 0x0 0x10000 ; sha1sum 0x42000000 0x10000
The SHA1 checksum was correct during the first call. Yet with
every subsequent call of the above command, it differed and was
wrong.
It turns out this was because of a race condition. On the first
time the command was called, no cacheline contained any data from
the destination memory location. The DMA transfered data into the
location and the cache above the location was invalidated. Then the
checksum was computed, but that meant the data were loaded into data
cache.
On any subsequent call, the DMA again transfered data into the same
destination. Yet during the transfer, some of the DCache lines were
evicted and written back into the main memory. Once the DMA transfer
completed, the data cache was invalidated over the memory location as
usual. But the data that were to be loaded back into the data cache
by subsequent SHA1 checksuming were corrupted.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
The computation was not correct with low clock values: setting a 1MHz
clock would result in an overlap that would then configure a 25Mhz
clock.
This patch implements a correct computation method according to the
kirkwood functionnal spec. table 600 (Serial Memory Interface
Configuration Register).
Signed-off-by: Valentin Longchamp <valentin.longchamp@keymile.com>
cc: Holger Brunck <holger.brunck@keymile.com>
cc: Prafulla Wadaskar <prafulla@marvell.com>
Acked-by: Prafulla Wadaskar <Prafulla@marvell.com>
Signed-off-by: Prafulla Wadaskar <prafulla@marvell.com>
The atmel at91sam9x5 series spi has feature to avoid receive overren
Using the patch to enable it
Signed-off-by: Bo Shen <voice.shen@atmel.com>
Acked-by: Andreas Bießmann <andreas.devel@googlemail.com>
Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
This change implements DMA chaining into SPI driver. This allows
the transfers to go much faster, while also fixing SF issues.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
Use calloc() instead of malloc() to allocate the mxs_spi_slave structure.
Clearing the memory is necessary since most of the time this gets done
super early in boot, but on warm reboots, and when SPI probing is done
long after the init stages it could actually pick up previously used memory,
and things like the chipselect polarity and other data end up being filled
with trash data if not explicitly set by the board files.
This solves a semi-random, almost unreproducable error whereby SPI devices
act very, very strangly on boot.
Signed-off-by: Matt Sealey <matt@genesi-usa.com>
Acked-by: Stefano Babic <sbabic@denx.de>
Use calloc() instead of malloc() to allocate the mxc_spi_slave structure.
Clearing the memory is necessary since most of the time this gets done
super early in boot, but on warm reboots, and when SPI probing is done
long after the init stages it could actually pick up previously used memory,
and things like the chipselect polarity and other data end up being filled
with trash data if not explicitly set by the board files.
This solves a semi-random, almost unreproducable error whereby SPI devices
act very, very strangly on boot. Tested on Efika MX over several years..
Signed-off-by: Matt Sealey <matt@genesi-usa.com>
Acked-by: Stefano Babic <sbabic@denx.de>
Since the input frequency of the API is a maximum that should not be exceeded in
order for the devices to operate properly, the SPI clock divider should be
rounded up, not truncated.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Stefano Babic <sbabic@denx.de>
Acked-by: Stefano Babic <sbabic@denx.de>
The sys_proto.h functions (except the boot modes) are compatible with
i.MX233 and i.MX28 so we use 'mxs' prefix for its methods.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
This is make naming consistent with the kernel and devicetree and in
preparation of pulling out the common tegra20 code.
Signed-off-by: Allen Martin <amartin@nvidia.com>
Acked-by: Stephen Warren <swarren@wwwdotorg.org>
Tested-by: Thierry Reding <thierry.reding@avionic-design.de>
Signed-off-by: Tom Warren <twarren@nvidia.com>
The DMA transfers happen only if the transfered data are larger
than 512 bytes. Otherwise PIO is used. This is a small speed
optimization.
The DMA transfer doesn't work if unaligned transfer is requested
due to the limitation of the DMA controller. This has to be fixed
by introducing generic bounce buffer. Therefore the DMA feature
is now disabled by default.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Wolfgang Denk <wd@denx.de>
Pull out all the PIO transfer logic into separate function,
so DMA can be added.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Wolfgang Denk <wd@denx.de>
This makes it easier to adapt for addition of DMA support.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Wolfgang Denk <wd@denx.de>
The patch "spi: tegra2: rename tegra2_spi.* to tegra_spi.*"
(sha1: edffa63d3d)
renamed tegra2_spi.c to tegra_spi.c
and the patch "Merge branch 'master' of git://git.denx.de/u-boot-microblaze"
(sha1: d978780b2e)
has wrongly resolved confict.
This patch fix it.
Signed-off-by: Michal Simek <monstr@monstr.eu>
* 'master' of git://git.denx.de/u-boot-microblaze:
microblaze: Wire up SPI driver
spi: microblaze: Adds driver for Xilinx SPI controller
microblaze: intc: Clear interrupt code
microblaze: Call serial multi initialization
microblaze: Move __udelay implementation
microblaze: Remove extern from board.c
microblaze: Wire up dts configuration
fdt: Add board specific dts inclusion
microblaze: Move individual board linker scripts to common script in cpu tree.
microblaze: Add gpio.h
microblaze: Add missing undefs for UBI and UBIFS
microblaze: Expand and correct configuration comments
microblaze: Enable ubi support
microblaze: Avoid compile error on systems without cfi flash
microblaze: Remove wrong define CONFIG_SYS_FLASH_PROTECTION
Conflicts:
drivers/spi/Makefile
Signed-off-by: Wolfgang Denk <wd@denx.de>
This is an improved version of the driver patch original
submitted by Graeme Smecher <graeme.smecher@mail.mcgill.ca>
The changes are:
- remove hard coded Xilinx BSP defines (XPAR_SPI_*) and
use CONFIG_SYS_SPI_BASE from config.h instead
- add extensive register struct definitions
- remove offset calculation for register access and
use the new register struct instead
- move default SPI controller configuration from
spi_setup_slave() to spi_claim_bus()
- add spi_set_speed()
- insert SPI controller deactivation in spi_release_bus()
- protect while loops in spi_xfer() with counter / timeouts
- support SPI mode flags: LSB_FIRST, CPHA, CPOL, LOOP
Come from:
http://patchwork.ozlabs.org/patch/71797/
Signed-off-by: Stephan Linz <linz@li-pro.net>
Tested-by: Michal Simek <monstr@monstr.eu>
