Add SMP boot functions for SCU equipped mach-shmobile SoCs.
At this point shmobile_smp_scu_prepare_cpus() controls the SCU and
installs boot fn and arg, while shmobile_smp_scu_boot_secondary()
currently does nothing. In the future the boot function and arg
install code will be reworked, so the empty function is ground
work for that.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Current Marzen is using DT booting,
and r8a7779.dtsi already has GIC settings.
So, we can remove the C version of the GIC setup code,
instead rely on GIC information provided by DT.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Current Armadillo800eva is using DT booting,
and r8a7740.dtsi already has GIC settings.
So, we can remove the C version of the GIC setup code,
instead rely on GIC information provided by DT.
This patch removes r8a7740_init_irq() which has
no user any more
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Remove r8a7779 specific ->smp_init_cpus() callback and
instead of relying on shmobile_smp_init_cpus() simply
use DT for CPU core information.
Based on work for the sh73a0 by Magnus Damm.
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Convert the marzen board to use DT_MACHINE_START. With this in
place all mach-shmobile boards use DT_MACHINE_START. Also, this
makes it possible for the r8a7779 SMP code to use DT for CPUs.
Signed-off-by: Magnus Damm <damm@opensource.se>
[horms+renesas@verge.net.au: Select USE_OF in Kconfig]
[horms+renesas@verge.net.au: Provide dt_compat in DT_MACHINE_START]
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
PFC pinctrl configuration can now be performed via DT. Update
ape6evm-reference to use this.
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski+renesas@gmail.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Remove sh73a0 specific ->smp_init_cpus() callback and
instead of relying on shmobile_smp_init_cpus() simply
use DT for CPU core information.
This assumes that DT_MACHINE_START is used which is
the case when AG5EVM and Kota2 are removed.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
This patch adds a sample DT-based Bock-W "reference" implementation.
The use of platform-specific C-code should be avoided
with this configuration as much as possible.
This patch adds new r8a7778_add_dt_devices() which was same stance
of r8a7790_add_dt_devices()
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Update the Lager DT reference code to get rid of the redundant
irqchip header and also make lager_add_standard_devices() static.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
This is sufficient to allow boot of the Lager board with
a console without boards-lager.c compiled into the kernel.
This is an example of a minimal but still useful initialisation
of the board using DT as much as possible.
As such it is the same as the boot of Lager that can be achieved
without a board file. The intention of adding this file
is to facilitate further work to allow board specific devices to be
initialised via DT.
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Add CPUFreq support to ape6evm-reference, using a max8973 regulator, that
is supplying V_DVFS for the 4 CA15 cores on r8a73a4.
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski+renesas@gmail.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
This patch adds a sample DT-based APE6EVM "reference" implementation. The
use of platform-specific C-code should be avoided with this configuration
as much as possible.
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski+renesas@gmail.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Update the r8a7790 arch timer setup code to configure the
frequency dynamically at boot time. This means that the arch
timer driver will be able to detect a timer frequency that
has been calculated based on the MD pins instead of a fixed
and potentially incorrect 13 MHz.
With this patch applied the Linux kernel will correctly
support the r8a7790 Lager board that uses a 20 Mhz EXTAL.
The arch timer will operate on 10 MHz and the Linux arch
timer driver will be correctly configured to use 10 MHz.
Without this patch the 20 MHz EXTAL will be used to drive
the arch timer at 10 MHz, but the Linux arch timer driver
will believe it is counting at 13 Mhz.
Reported-by: Ulrich Hecht <ulrich.hecht@gmail.com>
Signed-off-by: Magnus Damm <damm@opensource.se>
Tested-by: Ulrich Hecht <ulrich.hecht@gmail.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Break out the r8a7790 boot mode code into a separate
function so it can be shared by multiple users.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Update the EMEV2 SMP code to access the SMU directly
instead of relying on help from the legacy clock code.
This change moves us one step closer to common clocks.
Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Update the r8a73a4 code to allow using other
timers than Arch timer for clock event
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Add clock event support for CMT1 timer channel 0
to the r8a73a4 SoC code. The CMT is used together
with a 32KHz clock in this case.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Update the r8a7790 code to allow using other
timers than Arch timer for clock events.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Add clock event support for CMT0 timer channel 0
to the r8a7790 SoC code. On most ARM mach-shmobile
the CMT is hooked up to a 32KHz clock but on r8a7790
a 31.7KHz clock is instead used.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
The r8a7790 MSTP bits should be kept sorted in the same way
as on other mach-shmobile SoCs. Move the HSCIF and thermal
bits to clean up the current state.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
r8a73a4 SoCs have numerous I2C controllers, of which 9 are compatible with
the i2c-sh_mobile.c driver. This patch adds clock definitions for them.
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski+renesas@gmail.com>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
The Z2 clock on r8a73a4 is used to clock the 4 Cortex A7 cores on the SoC.
Add a definition for this clock to later use it from the arm_big_little
CPUFreq driver.
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski+renesas@gmail.com>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
clk_set_rate() should only be called with exact rates, returned by
clk_round_rate(). However, it is still good to verify, that the value,
passed to clock's .set_rate() method is at least valid. This patch adds
such a check for the Z-clock on r8a73a4.
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski+renesas@gmail.com>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
This patch adds support for the Z-clock on r8a73a4 SoCs, which is driving
the Cortex A15 core, and a "cpufreq-cpu0" platform device. Adding an
"operating-points" property to the CPU0 DT node and a regulator, this
patch allows platforms to use the generic cpufreq-cpu0 driver to use
SoC's DVFS capabilities.
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski+renesas@gmail.com>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
To reconfigure clocks, controlled by FRQCRA and FRQCRB, a kick bit has to
be set and to make sure the setting has taken effect, it has to be read
back repeatedly until it is cleared by the hardware. This patch adds the
waiting part, that was missing until now.
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski+renesas@gmail.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
The current temperature may be read using:
cat /sys/class/thermal/thermal_zone0/temp
Based on similar work for the r8a73a4 by Kuninori Morimoto.
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Add a DT reference implementation for the KZM9D board.
Only DT devices are used in this case. UART, STI, GPIO
and SMP / GIC are all provided by emev2.dtsi.
There is still a board specific C file used for enabling
legacy SH clocks. This file will be removed after we have
moved over to common clocks.
Signed-off-by: Magnus Damm <damm@opensource.se>
[horms+renesas@verge.net.au: Do not include trailing blank line in
board-kzm9d-reference.c ]
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Update the EMEV2 struct map_desc to exclude the SMU
and also include the ->map_io() callback in the DT
version of EMEV2 board support.
The EMEV2 SMP code can these days perform ioremap()
early on without the SMU information in the io_desc[].
To correctly support SMP in case of DT-only board
support then the ->map_io() callback is needed.
Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Rework the EMEV2 SMP code to rely on DT for CPU information
instead of reading out number of CPU cores from the SCU.
Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Remove the C version of the EMEV2 GIC setup code,
instead rely on GIC information provided by DT.
Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Remove mach-shmobile mach/dma.h since it only
seems to be used on non-mach-shmobile platforms.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
This patch removes r8a7778_init_irq(), since no-one is using it.
And now, there is no reason to have r8a7778_init_irq_common().
r8a7778_init_irq_dt() includes it.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Update the STI timer frequency comment to 32.768 kHz to
fix 37 kHz typo.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Convert the EMEV2 SoC device setup code from using very
verbose resources and static platform devices to the
same style as more recent boards. This reduces the size
of the code.
Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Remove EMEV2 early console code from KZM9D and instead
rely on console setup during regular platform device
probe time. This makes the DT code and the KZM9D board
support behave the same.
Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Replace the SoC-specific callback init_machine() with
a NULL to use the default code. This cleans up the code
and reduces the number of lines.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Pass NULL to of_platform_populate instead of passing
an empty list. This cleans up the code and reduces the
number of lines.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Pass NULL to of_platform_populate instead of passing
an empty list. This cleans up the code and reduces the
number of lines.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Pass NULL to of_platform_populate instead of passing
an empty list. This cleans up the code and reduces the
number of lines.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Replace the SoC-specific callback init_machine() with
a NULL to use the default code. This cleans up the code
and reduces the number of lines.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Pass NULL to of_platform_populate instead of passing
an empty list. This cleans up the code and reduces the
number of lines.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Replace the SoC-specific callback init_machine() with
a NULL to use the default code. This cleans up the code
and reduces the number of lines.
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Remove redundant irqchip_init() callback. The default case
of NULL will result in invoking irqchip_init() anyway.
Signed-off-by: Magnus Damm <damm@opensource.se>
[ horms+renesas@verge.net.au: Trimmed patch to remove portion
that updates the sh73a0 SoC and altered the subject to
use the same format as the patch that updates the sh73a0. ]
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Commit ebafed7a ("ARM: irq: Call irqchip_init if no init_irq function is
specified") removed the need to explictly setup the init_irq field in
the machine description when using only irqchip_init. Remove that
declaration for shmobile as well.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Remove redundant irqchip_init() callback. The default case
of NULL will result in invoking irqchip_init() anyway.
Signed-off-by: Magnus Damm <damm@opensource.se>
[ horms+renesas@verge.net.au: Trimmed patch to remove portion
that updates the r8a7790 SoC and altered the subject to
use the same format as the patch that updates the r8a7790. ]
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Commit ebafed7a ("ARM: irq: Call irqchip_init if no init_irq function is
specified") removed the need to explictly setup the init_irq field in
the machine description when using only irqchip_init. Remove that
declaration for shmobile as well.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Remove redundant irqchip_init() callback. The default case
of NULL will result in invoking irqchip_init() anyway.
Signed-off-by: Magnus Damm <damm@opensource.se>
[ horms+renesas@verge.net.au: Trimmed patch to remove portion
that updates the r8a73a4 SoC and altered the subject to
use the same format as the patch that updates the r8a73a4. ]
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
