It has been discovered that on some parts, from time to time, self-refresh
procedure doesn't work as expected. Debugging and investigating it proved
that disabling AC DLL introduce glitches in RAM controllers which
leads to unexpected behavior. This is confirmed as a hardware bug. DLL
bypass disables 3 DLLs: 2 DX DLLs and AC DLL. Thus, keep only DX DLLs
disabled. This introduce 6mA extra current consumption on VDDCORE when
switching to any ULP mode or standby mode but the self-refresh procedure
still works.
Fixes: f0bbf17958 ("ARM: at91: pm: add self-refresh support for sama7g5")
Suggested-by: Frederic Schumacher <frederic.schumacher@microchip.com>
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Tested-by: Cristian Birsan <cristian.birsan@microchip.com>
Link: https://lore.kernel.org/r/20220826083927.3107272-3-claudiu.beznea@microchip.com
In suspend/resume procedure for AT91 architecture different controllers
(PMC, SHDWC, RAM, RAM PHY, SFRBU) are accessed to do the proper settings
for power saving. Commit f0bbf17958 ("ARM: at91: pm: add self-refresh
support for sama7g5") introduced the access to RAMC PHY controller for
SAMA7G5. The access to this controller is done after RAMC ports are
closed, thus any TLB walk necessary for RAMC PHY virtual address will
fail. In the development branch this was not encountered. However, on
current kernel the issue is reproducible.
To solve the issue the previous mechanism of pre-loading the TLB with
the RAMC PHY virtual address has been used. However, only the addition
of this new pre-load breaks the functionality for ARMv5 based
devices (SAM9X60). This behavior has been encountered previously
while debugging this code and using the same mechanism for pre-loading
address for different controllers (e.g. pin controller, the assumption
being that other requested translations are replaced from TLB).
To solve this new issue the TLB flush + the extension of pre-loading
the rest of controllers to TLB (e.g. PMC, RAMC) has been added. The
rest of the controllers should have been pre-loaded previously, anyway.
Fixes: f0bbf17958 ("ARM: at91: pm: add self-refresh support for sama7g5")
Depends-on: e42cbbe5c9 ("ARM: at91: pm: group constants and addresses loading")
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Link: https://lore.kernel.org/r/20210930154219.2214051-4-claudiu.beznea@microchip.com
Adapt at91_pm_backup_init() to work for SAMA7G5. Also, set the LPM pin
to shutdown controller. This will signal to PMIC that it needs to switch
to the state corresponding to backup mode.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Link: https://lore.kernel.org/r/20210415105010.569620-21-claudiu.beznea@microchip.com
SAMA7G5 has 5 master clocks 0..4. MCK0 is controlled differently than
MCK 1..4. MCK 1..4 should also be saved/restored in the last phase of
suspend/resume. Thus, adapt wait_mckrdy to support also MCK1..4.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Link: https://lore.kernel.org/r/20210415105010.569620-10-claudiu.beznea@microchip.com
Replace CONFIG_SOC_SAM9X60 with CONFIG_HAVE_AT91_SAM9X60_PLL as the
SAM9X60's PLL is also present on SAMA7G5.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Link: https://lore.kernel.org/r/20210415105010.569620-9-claudiu.beznea@microchip.com
For the previous AT91 RAM controller and self-refresh procedure this
had no side effects. However, for SAMA7G5 the self-refresh procedure
doesn't allow this anymore as the RAM controller ports are closed
before switching it to self-refresh. This commits prepares the code
for the following ones adding self-refresh and PM support for SAMA7G5.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Link: https://lore.kernel.org/r/20210415105010.569620-8-claudiu.beznea@microchip.com
Compiling with the clang integrated assembler warns about
a recently added instruction:
<instantiation>:14:13: error: unknown token in expression
ldr tmp1, =#0x00020010UL
arch/arm/mach-at91/pm_suspend.S:542:2: note: while in macro instantiation
at91_plla_enable
Remove the extra '#' character that is not used for the 'ldr'
instruction when doing an indirect load of a constant.
Fixes: 4fd36e4583 ("ARM: at91: pm: add plla disable/enable support for sam9x60")
Tested-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Reviewed-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Acked-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Link: https://lore.kernel.org/r/20210204160129.2249394-1-arnd@kernel.org'
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
ULP0 fast improves suspend/resume time with few milliseconds the drawback
being the power consumption. The mean values measured for suspend/resume
time are as follows (measured on SAMA5D2 Xplained board), ULP0 compared
with fast ULP0:
- ulp0 fast: suspend time: 169 ms, resume time: 216 ms
- ulp0 : suspend time: 197 ms, resume time: 258 ms
Current consumption while suspended (measured on SAMA5D2 Xplained board):
- ulp0 fast: 730uA
- ulp0 : 270uA
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/1596616610-15460-2-git-send-email-claudiu.beznea@microchip.com
This will be used to differentiate b/w different PLLs settings to be
applied in the final/first steps of the suspend/resume process by doing
PLL specific configurations.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Acked-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/1579522208-19523-5-git-send-email-claudiu.beznea@microchip.com
Add macros for PLLA disable and enable (in disable macro the PLLA
state will also be saved). This prepares the field for PLLA disable/enable
for suspend/resume on SAM9X60.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/1579522208-19523-4-git-send-email-claudiu.beznea@microchip.com
This reverts commit 2725d70aa5
("ARM: at91: pm: do not disable/enable PLLA for ULP modes").
This is because PLLA is the clock source for CPU, PLLA should
be disabled/enabled in the final/first phase of suspend/resume
so that the power consumption in suspend/resume to be minimal
and suspend/resume time to be minimized.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/1579522208-19523-3-git-send-email-claudiu.beznea@microchip.com
SAM9X60's PMC has different master clock register offset than the other
SoCs' PMC. Due to this, specify master clock register offset based
on PMC compatible and pass it to pm_suspend.S since it is also needed
in there. When PM part for SAM9X60 was published the SAM9X60's PMC
(commit f6deae46039c ("clk: at91: add sam9x60 pmc driver")) wasn't
integrated.
Fixes: 01c7031cfa ("ARM: at91: pm: initial PM support for SAM9X60")
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/1579522208-19523-2-git-send-email-claudiu.beznea@microchip.com
<generated/at91_pm_data-offsets.h> is only generated and included by
arch/arm/mach-at91/, so it does not need to reside in the globally
visible include/generated/.
I renamed it to arch/arm/mach-at91/pm_data-offsets.h since the prefix
'at91_' is just redundant in mach-at91/.
My main motivation of this change is to avoid the race condition for
the parallel build (-j) when CONFIG_IKHEADERS is enabled.
When it is enabled, all the headers under include/ are archived into
kernel/kheaders_data.tar.xz and exposed in the sysfs.
In the parallel build, we have no idea in which order files are built.
- If at91_pm_data-offsets.h is built before kheaders_data.tar.xz,
the header will be included in the archive. Probably nobody will
use it, but it is harmless except that it will increase the archive
size needlessly.
- If kheaders_data.tar.xz is built before at91_pm_data-offsets.h,
the header will not be included in the archive. However, in the next
build, the archive will be re-generated to include the newly-found
at91_pm_data-offsets.h. This is not nice from the build system point
of view.
- If at91_pm_data-offsets.h and kheaders_data.tar.xz are built at the
same time, the corrupted header might be included in the archive,
which does not look nice either.
This commit fixes the race.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Link: https://lore.kernel.org/r/20190823024346.591-1-yamada.masahiro@socionext.com
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Based on 2 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license version 2 as
published by the free software foundation
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license version 2 as
published by the free software foundation #
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-only
has been chosen to replace the boilerplate/reference in 4122 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Enrico Weigelt <info@metux.net>
Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org>
Reviewed-by: Allison Randal <allison@lohutok.net>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190604081206.933168790@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is no need to disable/enable PLLA when switching to one of the
ULP modes. The PLLA consumers should take care of this.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Add ULP1 support for SAM9X60. In pm_suspend.S enable RC oscillator in
PMC if it is not enabled. At resume the state before suspend is
restored.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Acked-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Ludovic Desroches <ludovic.desroches@microchip.com>
The SAMA5D2 NRST input signal is resynchronized with the SLCK clock and it
can take up to 2 SLCK cycles (about 90us) for the internal reset to be
effective. During this delay, the VDDCORE current consumption may still be
high (application-dependent) with the VDDCORE regulator already OFF. Under
such conditions, VDDCORE may operate below its operating range leading to
potential register corruption.
To prevent such situation, it is recommended to decrease significantly the
power consumption of the device once the voltage regulator is turned-off.
This can be achieved by operating the device at a much lower low frequency.
To solve this switch the master clock to slock clock just before writing
shutdown command to shutdown controller.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Suggested-by: Patrice Vilchez <patrice.vilchez@microchip.com>
Acked-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
In the ULP1 mode, in order to achieve the lowest power consumption
with the system in retention mode and be able to resume on the wake
up events, all the clocks are shut off, inclusive the embedded 12MHz
RC oscillator, and the number of wake up sources is limited as well.
When the wake up event is asserted, the embedded 12MHz RC oscillator
restarts automatically.
The ULP1 (Ultra Low-power mode 1) is introduced by SAMA5D2.
The previous size of pm_suspend.o was 2148 bytes. With the addition of
ULP1 mode the new size of pm_suspend.o raised at 2456 bytes.
Signed-off-by: Wenyou Yang <wenyou.yang@atmel.com>
Signed-off-by: Ludovic Desroches <ludovic.desroches@microchip.com>
[claudiu.beznea@microchip.com: aligned with 4.18-rc1]
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Switch to use ULP0 naming instead of slow clock naming for power modes, to
be as closed as possible to datasheet. This commit does the necessary
renaming and macro addition to be as close as possible to the namings
from [1].
[1] https://lore.kernel.org/lkml/1470650705-31418-3-git-send-email-wenyou.yang@atmel.com
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
The sama5d2 has a mode were it is possible to cut power to the SoC while
keeping the RAM in self refresh.
Resuming from that mode needs support in the firmware/bootloader.
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Wenyou Yang <wenyou.yang@atmel.com>
Add a missing bracket at the end of Anti's email
Acked-by: Wenyou Yang <wenyou.yang@atmel.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
The number of register we can safely pass to at91_pm_suspend_in_sram is
limited. Instead, pass the address to the at91_pm_data structure.
The offsets are automatically generated to avoid hardcoding them.
Acked-by: Wenyou Yang <wenyou.yang@atmel.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
fncpy() requires that the source and the destination are both 8-byte
aligned.
Signed-off-by: Patrick Doyle <pdoyle@irobot.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Fixes: d94e688cae ("ARM: at91/pm: move the copying the sram function to the sram initialization phase")
Cc: <stable@vger.kernel.org> # 4.1+
Use the defines from atmel-mc.h instead of at91rm9200_sdramc.h
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
hardware.h is now mostyl unused, move the remaining declarations to pm.c and
remove it.
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Add the WFI instruction to make the cpu to the idle state.
In the meanwhile, disable the processor's clock.
Signed-off-by: Wenyou Yang <wenyou.yang@atmel.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Tested-by: Sylvain Rochet <sylvain.rochet@finsecur.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
As the file name is renamed, rename the function name at91_slow_clock()
--> at91_pm_suspend_sram_fn(), rename the function handler's name at the
same time.
Signed-off-by: Wenyou Yang <wenyou.yang@atmel.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Tested-by: Sylvain Rochet <sylvain.rochet@finsecur.com>
[nicolas.ferre@atmel.com: little update of the commit message]
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Because the sram function is used for both suspend to memory and the suspend
to standby mode, renaming is more elegant.
Signed-off-by: Wenyou Yang <wenyou.yang@atmel.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Tested-by: Sylvain Rochet <sylvain.rochet@finsecur.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
