This adds support for initializing DDR3 RAMs on the i.MX53 type
SDRAM controller. The code automatically detects size/layout of
the connected RAM, detects the bus width and which chipselects are
populated.
While I believe this code is not 100% generic, it is far too
sophisticated to stay in a single board directory. I'm sure
other boards could make use of this aswell.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The pll setup function is exported, so it makes sense to export
the convenience wrappers for specific frequencies aswell.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Some boards do not have a DCD table since they initialize everything
in code, so allow to skip it and leave the corresponding pointers
empty.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The code initializing the SDRAM controller is not at the same
place where SDRAM is registered with barebox. To reduce the
risk of registering wrong SDRAM sizes this patch adds a
driver for the ESDCTL which reads back the configured SDRAM
size and registers the memory found with barebox.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
In recent reference manuals the plls were renumbered. PLL8 now is
PLL6 and vice versa. Change the code according to the reference
manual to avoid confusion.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The ethernet PLL has a fixed frequency of 500MHz. What is adjustable
are additional dividers which we better describe separately.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This adds lowlevel debug functions for i.MX. As we have a great variety
of different UARTs on i.MX currently no Kconfig support for chosing the
correct one is added. Instead we expect the user to add the correct
define and issue a compiler warning if he hasn't.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Pass the buffer size to the file detection code. This makes sure we do not
read past the buffer. This is especially useful for ext filesystem detection
as the magic is at byte offset 1080. Also introduce a FILE_TYPE_SAFE_BUFSIZE
define which is set to the minimum bufsize the detection code needs to detect
all known filetypes.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Update help info, it's slightly out of date, and a grammatical fix.
No functional change.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This patch sets HPM (Host power mask bit) to bit 16 according to i.MX
Reference Manual. Falsely it was set to bit 8, but this controls pull-up
Impedance.
Reported-by: Michael Burkey <mdburkey@gmail.com>
Signed-off-by: Christoph Fritz <chf.fritz@googlemail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The same issue was fixed in the Linux kernel in commit
66ddfc6 (mx35: add a missing comma in a pad definition)
for 2.6.33-rc7.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This adds support for an update handler for internal boot. Currently
handled are:
- v1 MMC/SD
- v2 MMC/SD
- v2 NAND
where v1 is found on i.MX25, i.MX35 and i.MX51. v2 is found on i.MX53.
This code intentionally does not use the DCD data compiled into every
i.MX internal boot image. This makes it possible to make a pure second
stage barebox bootable on i.MX internal boot devices later.
This has been tested on the i.MX51 babbage, i.MX53 loco and i.MX53 tx53
board.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The KARO Tx53 board in the revision 8030 has an instable SDRAM
setup. It works as long as the MMU is disabled, but the board
crashes at arbitrary places once the MMU gets enabled. So we
need the PLL setup early. Enable it for pbl.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This redefines the sdram controller registers as offsets to the base
rather than as absolute addresses. All users are fixed to use the
SoC specific base address.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Rather than doing this in the SoC specific code just print
it in imx_set_silicon_revision. This saves some lines of code
and also results in i.MX27 now also having the silicon revision
printed during startup.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This moves the known i.MX bootsource settings to a single file
so that the code can be shared. Also we add a enum for the different
boot sources so that it can be used in C Code and not only on the shell.
The pcm038 board is changed to use it instead of digging in the registers
manually.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
On i.MX we enable all necessary clocks during startup of the clock
controller driver, so we do not need the register hacking in the drivers
anymore.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
- for i.MX1 the register is in the System Control unit, so move
the code to arch/arm/mach-imx/imx1.c
- for the other i.MX the register is in the watchdog unit, so move
the code to drivers/watchdog/imxwd.c
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
of_alias_get_id() returns the number of the gpio bank, so we have
to multiply with 32 to get the gpio base.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
- Add a separate header file for the iomux-v1 just like done for
iomux-v3.
- initialize iomux from SoC code so that we do not depend on IMX_GPIO_BASE
anymore.
- define registers as offset to the base rather than absolute addresses
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
All i.MX SoCs now use the same imx_silicon_revision() function to get
the revision. Add a separate header file for it and a common function
used on all SoCs.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This adds the device id mechanism to the i.MX fec driver and
uses it to determine the fec version. Also adds devicetree
probing support.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
- The i.MX1 timer does not have IPG clock as source, so rename
the define accordingly
- for the i.MX31 timer we want to use the per clock, not the ipg
clock.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The old clock support is now unused. Remove it. The former i.MX clko
command is superseeded by generic clock manipulation commands.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
To get rid of the register definitions in the SoC header files.
platform_device_id is used to distinguish between gpt types.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The FSF address has changed in the past. Instead of updating it
each time the address changes, just drop it completely treewide.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
in commit 2bdc9f57a8 the iomux was synced
with the kernel but this leads to some changes in the PAD_CTRL of some
FEC pins leading to a non working FEC on our cpuimx51 board.
This patch set back the PAD_CTRL of the missing pins to the initial
value.
Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The IMX i2c driver is to be used by SOCs on both the ARM and PPC architetures.
Use a more neutral name for the structure, function names and #define.
The driver name is now "i2c-fsl".
Signed-off-by: Renaud Barbier <renaud.barbier@ge.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Add a function to take the usbphy1 through its initialization procedure.
The imx6_usb_phy1_disable_oc function is needed at least on the sabrelite,
as the power design is faulty. The state on other imx6 boards is unknown,
therefore it is an extra function call at the moment.
Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
tested with a m25p128 flash both in nand boot & spi boot.
Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
With pbl support enabled most boards need a pbl-y for their lowlevel
stuff.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
The i.MX5 does not have a valid function to get the spi clock. This
patch introduces a function for i.MX6, and moves the bogus spi clock
speed to the speed-imx5*.c. Not nice, but preserves the current status
quo for i.MX5
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This fix a clock inaccuracy in get_time_ns (used by sleep, time, etc).
At i.MX53 power-on GPT clock is typically 55500000 Hz, and it will be used
to calc the clock multiplier. After call imx53_init_lowlevel() GPT clock
will changed (e.g. to 66666666 Hz), but multiplier not. To fix this behavior call
clock_notifier_call_chain() after changing clock in imx53_init_lowlevel().
Signed-off-by: Wjatscheslaw Stoljarski <wjatscheslaw.stoljarski@kiwigrid.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Procedure is missing, so remove its declaration.
Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Storing the size instead of the resource end in struct resource was
a mistake. 'size' ranges from 0 to UINT[32|64]_MAX + 1 which obviously
leads to problems. 'end' on the other hand will never exceed
UINT[32|64]_MAX. Also this way we can express a iomem region covering
the whole address space.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>