For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
will highly increase the memset/memcpy performance. This is able
thanks to the ARM multiple register instructions.
Unfortunatelly the relocation is done without the cache enabled,
so it takes some time, but zeroing the BSS memory takes much more
longer, especially for the configs with big static buffers.
A quick test confirms, that the boot time improvement after using
the arch memcpy for relocation has no significant meaning.
The same test confirms that enable the memset for zeroing BSS,
reduces the boot time.
So this patch enables the arch memset for zeroing the BSS after
the relocation process. For ARM boards, this can be enabled
in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.
This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~1384ms - before this change
- ~888ms - after this change
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Tom Rini <trini@konsulko.com>
The 'nandecc sw' command selects a software-based error correction
algorithm. By default, this is OMAP_ECC_HAM1_CODE_SW but some
platforms use OMAP_ECC_BCH8_CODE_HW_DETECTION_SW as their
software-based correction algorithm. Allow a user to be specific e.g.
# nandecc sw <hamming|bch8>
where 'hamming' is still the default.
Note: we don't just use CONFIG_NAND_OMAP_ECCSCHEME as it might be set
to a hardware-based ECC scheme---a little strange when the user
has requested 'sw' ECC.
Signed-off-by: Ash Charles <ashcharles@gmail.com>
This patch extends OMAP3 support for AM/DM37xx and
introduces the AM3703-based Quipos Cairo board.
Signed-off-by: Albert ARIBAUD (3ADEV) <albert.aribaud@3adev.fr>
Reviewed-by: Simon Glass <sjg@chromium.org>
esbc_validate command uses various IP Blocks: Security Monitor, CAAM block
and SFP registers. Hence the respective CONFIG's are enabled.
Apart from these CONFIG_SHA_PROG_HW_ACCEL and CONFIG_RSA are also enabled.
Signed-off-by: Gaurav Rana <gaurav.rana@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
According to table 2-3 on page 87 of Marvell's latest PXA270
Specification Update Rev. I from 2010.04.19 [1] there exists a breed of
chips with a new CPU ID for PXA270M A1 stepping which our latest
Colibri PXA270 V2.4A modules actually have assembled. This patch helps
in correctly identifying those chips upon boot as well which then looks
as follows:
CPU: Marvell PXA27xM rev. A1
[1] http://www.marvell.com/application-processors/pxa-family/assets/pxa_27x_spec_update.pdf
Acked-by: Marek Vasut <marex@denx.de>
commit d9f43c8f5c sets
get_reset_cause() as static, but this conflicts with mx5
where its prototype is in sys_proto.h.
Drop it from sys_proto.h and drop print_cpuinfo from mx53_loco,
factorizing the call for this board.
Signed-off-by: Stefano Babic <sbabic@denx.de>
CC: Jason Liu <jason.hui@linaro.org>
Import DTS for Arria V development kit and enable support
for DT. The DT is imported from Linux 3.19-rc1 as of commit
97bf6af1f928216fd6c5a66e8a57bfa95a659672 .
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Chin Liang See <clsee@opensource.altera.com>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Acked-by: Pavel Machek <pavel@denx.de>
Reviewed-by: Stefan Roese <sr@denx.de>
Cc: Vince Bridgers <vbridger@opensource.altera.com>
Import DTS for Cyclone V development kit and enable support
for DT. The DT is imported from Linux 3.19-rc1 as of commit
97bf6af1f928216fd6c5a66e8a57bfa95a659672 .
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Chin Liang See <clsee@opensource.altera.com>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Acked-by: Pavel Machek <pavel@denx.de>
Reviewed-by: Stefan Roese <sr@denx.de>
Cc: Vince Bridgers <vbridger@opensource.altera.com>
Add support for the Altera Arria V development kit.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Chin Liang See <clsee@opensource.altera.com>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Pavel Machek <pavel@denx.de>
Cc: Stefan Roese <sr@denx.de>
Cc: Vince Bridgers <vbridger@opensource.altera.com>
Currently in some cases SDRAM init requires global_data to be available
and soon this will not be available prior to board_init_f(). Adjust the
code paths in these cases to be correct. In some cases we had the SPL
stack be in DDR as we might have large stacks (due to Falcon Mode +
Environment). In these cases switch to CONFIG_SPL_STACK_R. In other
cases we had simply been setting CONFIG_SPL_STACK into SRAM. In these
cases we no longer need to (CONFIG_SYS_INIT_SP_ADDR is used and is also
in SRAM) so drop those lines.
Signed-off-by: Simon Glass <sjg@chromium.org>
Tested on Beagleboard, Beagleboard xM
Tested-by: Matt Porter <mporter@konsulko.com>
Tested on Beaglebone Black, AM43xx GP EVM, OMAP5 uEVM, OMAP4 Pandaboard
Tested-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
At present SPL uses a single stack, either CONFIG_SPL_STACK or
CONFIG_SYS_INIT_SP_ADDR. Since some SPL features (such as MMC and
environment) require a lot of stack, some boards set CONFIG_SPL_STACK to
point into SDRAM. They then set up SDRAM very early, before board_init_f(),
so that the larger stack can be used.
This is an abuse of lowlevel_init(). That function should only be used for
essential start-up code which cannot be delayed. An example of a valid use is
when only part of the SPL code is visible/executable, and the SoC must be set
up so that board_init_f() can be reached. It should not be used for SDRAM
init, console init, etc.
Add a CONFIG_SPL_STACK_R option, which allows the stack to be moved to a new
address before board_init_r() is called in SPL.
The expected SPL flow (for CONFIG_SPL_FRAMEWORK) is documented in the README.
Signed-off-by: Simon Glass <sjg@chromium.org>
For version 1:
Acked-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Bo Shen <voice.shen@atmel.com>
Acked-by: Bo Shen <voice.shen@atmel.com>
Acked-by: Heiko Schocher <hs@denx.de>
Tested-by: Heiko Schocher <hs@denx.de>
Signed-off-by: Tom Rini <trini@konsulko.com>
Use the full driver model GPIO and serial drivers in SPL now that these are
supported. Since device tree is not available they will use platform data.
Remove the special SPL GPIO function as it is no longer needed.
This is all in one commit to maintain bisectability.
Signed-off-by: Simon Glass <sjg@chromium.org>
This is already set up in crt0.S. We don't need a new structure and don't
really want one in the 'data' section of the image, since it will be empty
and crt0.S's changes will be ignored.
As an interim measure, remove it only if CONFIG_DM is not defined. This
allows us to press ahead with driver model in SPL and allow the stragglers
to catch up.
Signed-off-by: Simon Glass <sjg@chromium.org>
This function has grown into something of a monster. Some boards are setting
up a console and DRAM here in SPL. This requires global_data which should be
set up in one place (crt0.S).
There is no need for SPL to use s_init() for anything since board_init_f()
is called immediately afterwards.
Signed-off-by: Simon Glass <sjg@chromium.org>
This fixes the MMC/SD card detect GPIOs for Apalis T30 which got broken
by the following commit:
2b2b50bc87 "dm: tegra: dts: Use TEGRA_GPIO() macro for all GPIOs"
While at it also re-add the comments describing which particular
Apalis/Colibri pins those GPIOs are on.
Signed-off-by: Marcel Ziswiler <marcel@ziswiler.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Stefan Agner <stefan.agner@toradex.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
All boards with a SPI interface have a suitable spi alias except Apalis
T30. Add these missing aliases just as the following commit did for the
others:
d2f60f9332 "dm: tegra: dts: Add aliases for spi on tegra30 boards"
Signed-off-by: Marcel Ziswiler <marcel@ziswiler.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
This patch incorporates a few fixes from Tom Warren <twarren@nvidia.com>.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Tegra210 has a per-pin option named e_io_hv, which indicates that the
pin's input path should be configured to be 3.3v-tolerant. Add support
for this.
Note that this is very similar to previous chip's rcv_sel option.
However, since the Tegra TRM names this option differently for the
different chips, we support the new name so that the code exactly matches
the naming in the TRM, to avoid confusion.
This patch incorporates a few fixes from Tom Warren <twarren@nvidia.com>.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Tegra210 starts its drive group registers at a different offset from the
APB MISC register block that other SoCs. Update the code to handle this.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
T210 support HSM and Schmitt options in the pinmux register (previous
chips placed these options in the drive group register). Update the
code to handle this.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Tegra210 moves some bits around in the pinmux registers. Update the code
to handle this.
This doesn't attempt to address the issues with the group-to-group varying
drive group register layout mentioned earlier. This patch handles the
SoC-to-SoC differences in the mux register layout.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
On some future SoCs, some per-drive-group features became per-pin
features. Move all type definitions early in the header so they can
be enabled irrespective of the setting of TEGRA_PMX_SOC_HAS_DRVGRPS.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
On some future SoCs, some of the per-drive-group features no longer
exist. Add some ifdefs to support this.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Future SoCs have a slightly different combination of pinmux options per
pin. This will be simpler to handle if we simply have one define per
option, rather than grouping various options together, in combinations
that don't align with future chips.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Tegra's drive group registers have a remarkably inconsistent layout. The
current U-Boot driver doesn't take this into account at all. Add a
comment to describe the issue, so at least anyone debugging the driver
will be aware of this. To solve this, we'd need to add a per-drive-group
data structure describing the layout for the individual register. Since
we don't set up too many drive groups in U-Boot at present, this
hopefully isn't causing too much practical issue. Still, we probably need
to fix this sometime.
Wth Tegra210, the register layout becomes almost entirely consistent, so
this problem partially solves itself over time.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
This is needed to correctly apply the new Jetson TK1 pinmux config.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
When the CPU is in non-secure (NS) mode (when running U-Boot under a
secure monitor), certain actions cannot be taken, since they would need
to write to secure-only registers. One example is configuring the ARM
architectural timer's CNTFRQ register.
We could support this in one of two ways:
1) Compile twice, once for secure mode (in which case anything goes) and
once for non-secure mode (in which case certain actions are disabled).
This complicates things, since everyone needs to keep track of
different U-Boot binaries for different situations.
2) Detect NS mode at run-time, and optionally skip any impossible actions.
This has the advantage of a single U-Boot binary working in all cases.
(2) is not possible on ARM in general, since there's no architectural way
to detect secure-vs-non-secure. However, there is a Tegra-specific way to
detect this.
This patches uses that feature to detect secure vs. NS mode on Tegra, and
uses that to:
* Skip the ARM arch timer initialization.
* Set/clear an environment variable so that boot scripts can take
different action depending on which mode the CPU is in. This might be
something like:
if CPU is secure:
load secure monitor code into RAM.
boot secure monitor.
secure monitor will restart (a new copy of) U-Boot in NS mode.
else:
execute normal boot process
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Some systems have so much RAM that the end of RAM is beyond 4GB. An
example would be a Tegra124 system (where RAM starts at 2GB physical)
that has more than 2GB of RAM.
In this case, we want gd->ram_size to represent the actual RAM size, so
that the actual RAM size is passed to the OS. This is useful if the OS
implements LPAE, and can actually use the "extra" RAM.
However, we can't use get_ram_size() to verify the actual amount of RAM
present on such systems, since some of the RAM can't be accesses, which
confuses that function. Avoid calling get_ram_size() when the RAM size
is too large for it to work correctly. It's never actually needed anyway,
since there's no reason for the BCT to report the wrong RAM size.
In systems with >=4GB RAM, we still need to clip the reported RAM size
since U-Boot uses a 32-bit variable to represent the RAM size in bytes.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
size_mb is used to hold a value that's sometimes KB, sometimes MB,
and sometimes bytes. Use separate correctly named variables to avoid
confusion here. Also fix indentation of a conditional statement.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Add basic Xilinx ZynqMP arm64 support.
Serial and SD is supported.
It supports emulation platfrom ep108 and QEMU.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
With a389531 we now call readl() from this file so add <asm/io.h> so
that we have a prototype for the function.
Signed-off-by: Tom Rini <trini@konsulko.com>
Freescale's SEC block has built-in Data Encryption
Key(DEK) Blob Protocol which provides a method for
protecting a DEK for non-secure memory storage.
SEC block protects data in a data structure called
a Secret Key Blob, which provides both confidentiality
and integrity protection.
Every time the blob encapsulation is executed,
a AES-256 key is randomly generated to encrypt the DEK.
This key is encrypted with the OTP Secret key
from SoC. The resulting blob consists of the encrypted
AES-256 key, the encrypted DEK, and a 16-bit MAC.
During decapsulation, the reverse process is performed
to get back the original DEK. A caveat to the blob
decapsulation process, is that the DEK is decrypted
in secure-memory and can only be read by FSL SEC HW.
The DEK is used to decrypt data during encrypted boot.
Commands added
--------------
dek_blob - encapsulating DEK as a cryptgraphic blob
Commands Syntax
---------------
dek_blob src dst len
Encapsulate and create blob of a len-bits DEK at
address src and store the result at address dst.
Signed-off-by: Raul Cardenas <Ulises.Cardenas@freescale.com>
Signed-off-by: Nitin Garg <nitin.garg@freescale.com>
Signed-off-by: Ulises Cardenas <ulises.cardenas@freescale.com>
Signed-off-by: Ulises Cardenas-B45798 <Ulises.Cardenas@freescale.com>
Since commit 3ff46cc42b ("arm: relocate the exception vectors") mx35
does not boot anymore.
Add a specific relocate_vectors macro that skips the vector relocation, as the
i.MX35 SoC does not provide RAM at the high vectors address (0xFFFF0000), and
(0x00000000) maps to ROM.
This allows mx35 to boot again.
Cc: Sebastian Priebe <sebastian.priebe@cadcon.de>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Tested-by: Stefano Babic <sbabic@denx.de>
Since commit 3ff46cc42b ("arm: relocate the exception vectors") mx31
does not boot anymore.
Add a specific relocate_vectors macro that skips the vector relocation, as the
i.MX31 SoC does not provide RAM at the high vectors address (0xFFFF0000), and
(0x00000000) maps to ROM.
This allows mx31 to boot again.
Cc: Anatolij Gustschin <agust@denx.de>
Cc: Magnus Lilja <lilja.magnus@gmail.com>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
If CONFIG_ARMV7_PSCI is not defined and CONFIG_ARMV7_SECURE_BASE is defined,
smp_kicl_all_cpus may enable secondary cores and runs into secure_ram_addr(
_smp_pen), before code is relocated to secure ram.
So need relocation to secure ram before enable secondary cores.
Signed-off-by: Peng Fan <Peng.Fan@freescale.com>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Each way of the system cache has 256 entries for PH1-Pro4 and older
SoCs, whereas 512 entries for PH1-Pro5 and newer SoCs. The line
size is still 128 byte. Thus, the way size is 32KB/64KB for old/new
SoCs.
To keep lowlevel_init SoC-independent, set BOOT_RAM_SIZE to the
constant value 32KB. It is large enough for temporary RAM and
should work for all the SoCs of UniPhier family.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
This function was intended for MN2WS0235 (what we call PH1-Pro4TV).
On that SoC, MPLL is already running on the power-on reset and it
makes sense to stop the PLL at early boot-up.
On the other hand, PH1-Pro4(R) does not have SC_MPLLOSCCTL register,
so this function has no point.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
This code is duplicated in ph1-ld4/sg_init.c and ph1-pro4/sg_init.c.
Merge the same code into a new file, memconf.c.
The helper functions no longer have to be placed in the header file.
Also, move them into memconf.c.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Two support card variants are used with UniPhier reference boards:
- 1 chip select support card (original CPLD)
- 3 chip selects support card (ARIMA-compatible CPLD)
Currently, the former is only supported on PH1-Pro4, but it can be
expanded to PH1-LD4, PH1-sLD8 with a little code change.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Each USB port corresponds to the following IP core:
port0: xHCI (0x65a00000) SS+HS
port1: xHCI (0x65c00000) HS (SS PHY is not implemented)
port2: EHCI (0x5a800100) HS
port3: EHCI (0x5a810100) HS
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
EHCI host controllers have a common register interface.
We may wish to implement a generic EHCI driver someday.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Because uniphier_ehci_reset() is only called from ehci-uniphier.c,
it can be a static function there.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Acked-by: Marek Vasut <marex@denx.de>
Now UniPhier platform highly depends on Device Tree configuration
(CONFIG_OF_CONTROL is select'ed by Kconfig). Since the EHCI is only
used on main U-Boot, we can drop platform devices of the EHCI
controllers. We still keep UART platform devices because they might
be useful for SPL.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Acked-by: Marek Vasut <marex@denx.de>
Deassert the reset signal and provide the clock for STDMAC core.
This is necessary for the USB 2.0 host controllers.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
For all the UniPhier SoCs so far, the reset signal of the NAND core
is automatically deasserted after the PLL gets stabled.
(The bit 2 of SC_RSTCTRL is default to one.)
This causes a fatal problem on the NAND controller of PH1-LD4.
For that SoC, the NAND I/O pins are not set up yet at the power-on
reset except the NAND boot mode. As a result, the NAND controller
begins automatic device scanning with wrong I/O pins and finally
hangs up.
Actually, U-Boot dies after printing "NAND:" on the console unless
the boot mode latch detected the NAND boot mode.
To work around this problem, reset the NAND core in SPL for non-NAND
boot modes. If CONFIG_NAND_DENALI is enabled, the reset signal is
deasserted again in U-Boot proper. At this time, I/O pins have been
correctly set up, the device scanning should succeed.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Split the current clkrst_init() into two functions:
- early_clkrst_init(): called from SPL
Deassert the reset signals of the memory controller and some other
basic cores.
- clkrst_init(): called from main U-boot
Deassert the reset signals that are necessary for the access to
peripherals etc.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Now UniPhier SoCs only work with CONFIG_SPL and the function
sbc_init() is called from SPL.
The conditional #if !defined(CONFIG_SPL_BUILD) has no point
any more.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Since commit 0e7368c6c4 (kbuild: prepare for moving headers into
mach-*/include/mach), we can replace #include <asm/arch/*.h> with
<mach/*.h> so we do not need to create the symbolic link during the
build.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
It was found that the L2 cache timings that we had before could cause
freezes and hangs. We should make things more robust with better
timings. Currently the production ChromeOS kernel applies these
timings, but it's nice to fixup firmware too (and upstream probably
won't take our kernel hacks).
This also provides a big cleanup of the L2 cache init code avoiding
some duplication. The way things used to work:
* low_power_start() was installed by the SPL (both at boot and resume
time) and left resident in iRAM for the kernel to use when bringing
up additional CPUs. It used configure_l2_ctlr() and
configure_l2_actlr() when it detected it was on an A15. This was
needed (despite the L2 cache registers being shared among all A15s)
because we might have been the first man in after the whole A15
cluster was shutdown.
* secondary_cores_configure() was called on at boot time and at resume
time. Strangely this called configure_l2_ctlr() but not
configure_l2_actlr() which was almost certainly wrong. Given that
we'll call both (see next bullet) later in the boot process it
didn't matter for normal boot, but I guess this is how L2 cache
settings got set on 5420/5800 (but not 5250?) at resume time.
* exynos5_set_l2cache_params() was called as part of cache enablement.
This should happen at boot time (normally in the SPL except for USB
boot where it happens in main U-Boot).
Note that the old code wasn't setting ECC/parity in the cache
enablement code but we happened to get it anyway because we'd call
secondary_cores_configure() at boot time. For resume time we'd get it
anyway when the 2nd A15 core came up.
Let's make this a whole lot simpler. Now we always set these
parameters in the same place for all boots and use the same code for
setting up secondary CPUs.
Intended net effects of this change (other than cleanup):
* Timings go from before:
data: 0 cycle setup, 3 cycles (0x2) latency
tag: 0 cycle setup, 3 cycles (0x2) latency
after:
data: 1 cycle setup, 4 cycles (0x3) latency
tag: 1 cycle setup, 4 cycles (0x3) latency
* L2ACTLR is properly initted on 5420/5800 in all cases.
One note is that we're still relying on luck to keep low_power_start()
working. The compiler is being nice and not storing anything on the
stack.
Another note is that on its own this patch won't help to fix cache
settings in an RW U-Boot update where we still have the RO SPL. The
plan for that is:
* Have RW U-Boot re-init the cache right before calling the kernel
(after it has turned the L2 cache off). This is why the functions
are in a header file instead of lowlevel_init.c.
* Have the kernel save the L2 cache settings of the boot CPU and apply
them to all other CPUs. We get a little lucky here because the old
code was using "|=" to modify the registers and all of the bits that
it's setting are also present in the new settings (!). That means
that when the 2nd CPU in the A15 cluster comes up it doesn't
actually mess up the settings of the 1st CPU in the A15 cluster. An
alternative option is to have the kernel write its own
low_power_start() code.
Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Akshay Saraswat <akshay.s@samsung.com>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
On warm reset, all cores jump to the low_power_start function because iRAM
data is retained and because while executing iROM code all cores find
the jump flag 0x02020028 set. In low_power_start, cores check the reset
status and if true they clear the jump flag and jump back to 0x0.
The A7 cores do jump to 0x0 but consider following instructions as a Thumb
instructions which in turn makes them loop inside the iROM code instead of
jumping to power_down_core.
This issue is fixed by replacing the "mov pc" instruction with a "bx"
instruction which switches state along with the jump to make the execution
unit consider the branch target as an ARM instruction.
Signed-off-by: Akshay Saraswat <akshay.s@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
When compiled SPL for Thumb secondary cores failed to boot
at the kernel boot up. Only one core came up out of 4.
This was happening because the code relocated to the
address 0x02073000 by the primary core was an ARM asm
code which was executed by the secondary cores as if it
was a thumb code.
This patch fixes the issue of secondary cores considering
relocated code as Thumb instructions and not ARM instructions
by jumping to the relocated with the help of "bx" ARM instruction.
"bx" instruction changes the 5th bit of CPSR which allows
execution unit to consider the following instructions as ARM
instructions.
Signed-off-by: Akshay Saraswat <akshay.s@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
This patch does 3 things:
1. Enables ECC by setting 21st bit of L2CTLR.
2. Restore data and tag RAM latencies to 3 cycles because iROM sets
0x3000400 L2CTLR value during switching.
3. Disable clean/evict push to external by setting 3rd bit of L2ACTLR.
We need to restore this here due to switching.
Signed-off-by: Abhilash Kesavan <a.kesavan@samsung.com>
Signed-off-by: Akshay Saraswat <akshay.s@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
L2 Auxiliary Control Register provides configuration
and control options for the L2 memory system. Bit 3
of L2ACTLR stands for clean/evict push to external.
Setting bit 3 disables clean/evict which is what
this patch intends to do.
Signed-off-by: Akshay Saraswat <akshay.s@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
iROM logic provides undesired jump address for CPU2.
This patch adds a programmable susbstitute for a part of
iROM logic which wakes up cores and provides jump addresses.
This patch creates a logic to make all secondary cores jump
to a particular address which evades the possibility of CPU2
jumping to wrong address and create undesired results.
Logic of the workaround:
Step-1: iROM code checks value at address 0x2020028.
Step-2: If value is 0xc9cfcfcf, it jumps to the address (0x202000+CPUid*4),
else, it continues executing normally.
Step-3: Primary core puts secondary cores in WFE and store 0xc9cfcfcf in
0x2020028 and jump address (pointer to function low_power_start)
in (0x202000+CPUid*4).
Step-4: When secondary cores recieve event signal they jump to this address
and continue execution.
Signed-off-by: Kimoon Kim <kimoon.kim@samsung.com>
Signed-off-by: Akshay Saraswat <akshay.s@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
This patch adds workaround for the ARM errata 799270 which says
"If the L2 cache logic clock is stopped because of L2 inactivity,
setting or clearing the ACTLR.SMP bit might not be effective. The bit is
modified in the ACTLR, meaning a read of the register returns the
updated value. However the logic that uses that bit retains the previous
value."
Signed-off-by: Kimoon Kim <kimoon.kim@samsung.com>
Signed-off-by: Akshay Saraswat <akshay.s@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
This patch adds workaround for ARM errata 798870 which says
"If back-to-back speculative cache line fills (fill A and fill B) are
issued from the L1 data cache of a CPU to the L2 cache, the second
request (fill B) is then cancelled, and the second request would have
detected a hazard against a recent write or eviction (write B) to the
same cache line as fill B then the L2 logic might deadlock."
Signed-off-by: Kimoon Kim <kimoon.kim@samsung.com>
Signed-off-by: Akshay Saraswat <akshay.s@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
This patch adds code to shutdown secondary cores.
When U-boot comes up, all secondary cores appear powered on,
which is undesirable and causes side effects while
initializing these cores in kernel.
Secondary core power down happens in following steps:
Step-1: After Exynos power-on, primary core starts executing first.
Step-2: In iROM code every core has to check 2 flags i.e.
addresses 0x02020028 & 0x02020004.
Step-3: Initially 0x02020028 is 0 for all cores and 0x02020004 has a
jump address for primary core and 0 for all secondary cores.
Step-4: Therefore, primary core follows normal iROM execution and jumps
to BL1 eventually, whereas all secondary cores enter WFE.
Step-5: When primary core comes into function secondary_cores_configure,
it puts pointer to function power_down_core into 0x02020004
and provides DSB and SEV for all cores so that they may come out
of WFE and jump to power_down_core function.
Step-6: And ultimately because of power_down_core all
secondary cores shut-down.
Signed-off-by: Kimoon Kim <kimoon.kim@samsung.com>
Signed-off-by: Akshay Saraswat <akshay.s@samsung.com>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
ED Mini V2 is based on Orion 5x which boots at fixed
address 0xFFFF0000 in NOR Flash. Place SPL there, and
switch U-Boot from .bin to .img format, stored in
NOR Flash at 0xFFF90000.
Note: this patch was tested on HW and works, i.e.
it boots U-Boot properly, but SPL console output
currently does not appear, due to GD being trashed
by arch/arm/lib/spl.c. This trashing is soon to be
removed, and then ED Mini V2 SPL console output will
become visible.
Signed-off-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
Porter is an entry level development board based on R-Car M2 SoC (R8A7791)
This commit supports the following peripherals:
- SCIF, I2C, Ethernet, QSPI, SD, USB Host
Signed-off-by: Vladimir Barinov <vladimir.barinov@cogentembedded.com>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Lager board has two SDHI port as SDHI0 and SDHI2.
This adds GPIO configuration and initialization function of SDHI, and
enables MMC command.
Signed-off-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj@renesas.com>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Alt board has two SDHI port.
This adds GPIO configuration and initialization function of SDHI, and
enables MMC command.
Signed-off-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj@renesas.com>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
This is still a non-generic board.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Acked-by: Sughosh Ganu <urwithsughosh@gmail.com>
Cc: Syed Mohammed Khasim <sm.khasim@gmail.com>
Acked-by: Marek Vasut <marex@denx.de>
This is still a non-generic board.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Chan-Taek Park <c-park@ti.com>
Acked-by: Marek Vasut <marex@denx.de>
This is still a non-generic board.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Po-Yu Chuang <ratbert@faraday-tech.com>
Acked-by: Marek Vasut <marex@denx.de>
These are still non-generic boards.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Greg Ungerer <greg.ungerer@opengear.com>
Acked-by: Marek Vasut <marex@denx.de>
This is still a non-generic board.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Lei Wen <leiwen@marvell.com>
Acked-by: Marek Vasut <marex@denx.de>
This is still a non-generic board.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Cc: Matthias Weisser <weisserm@arcor.de>
Acked-by: Marek Vasut <marex@denx.de>
Now CONFIG_SPL_BUILD is not defined in Kconfig, so
"!depends on SPL_BUILD" and "if !SPL_BUILD" are redundant.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
When Kconfig for U-boot was examined, one of the biggest issues was
how to support multiple images (Normal, SPL, TPL). There were
actually two options, "single .config" and "multiple .config".
After some discussions and thought experiments, I chose the latter,
i.e. to create ".config", "spl/.config", "tpl/.config" for Normal,
SPL, TPL, respectively.
It is true that the "multiple .config" strategy provided us the
maximum flexibility and helped to avoid duplicating CONFIGs among
Normal, SPL, TPL, but I have noticed some fatal problems:
[1] It is impossible to share CONFIG options across the images.
If you change the configuration of Main image, you often have to
adjust some SPL configurations correspondingly. Currently, we
cannot handle the dependencies between them. It means one of the
biggest advantages of Kconfig is lost.
[2] It is too painful to change both ".config" and "spl/.config".
Sunxi guys started to work around this problem by creating a new
configuration target. Commit cbdd9a9737 (sunxi: kconfig: Add
%_felconfig rule to enable FEL build of sunxi platforms.) added
"make *_felconfig" to enable CONFIG_SPL_FEL on both images.
Changing the configuration of multiple images in one command is a
generic demand. The current implementation cannot propose any
good solution about this.
[3] Kconfig files are getting ugly and difficult to understand.
Commit b724bd7d63 (dm: Kconfig: Move CONFIG_SYS_MALLOC_F_LEN to
Kconfig) has sprinkled "if !SPL_BUILD" over the Kconfig files.
[4] The build system got more complicated than it should be.
To adjust Linux-originated Kconfig to U-Boot, the helper script
"scripts/multiconfig.sh" was introduced. Writing a complicated
text processor is a shell script sometimes caused problems.
Now I believe the "single .config" will serve us better. With it,
all the problems above would go away. Instead, we will have to add
some CONFIG_SPL_* (and CONFIG_TPL_*) options such as CONFIG_SPL_DM,
but we will not have much. Anyway, this is what we do now in
scripts/Makefile.spl.
I admit my mistake with my apology and this commit switches to the
single .config configuration.
It is not so difficult to do that:
- Remove unnecessary processings from scripts/multiconfig.sh
This file will remain for a while to support the current defconfig
format. It will be removed after more cleanups are done.
- Adjust some makefiles and Kconfigs
- Add some entries to include/config_uncmd_spl.h and the new file
scripts/Makefile.uncmd_spl. Some CONFIG options that are not
supported on SPL must be disabled because one .config is shared
between SPL and U-Boot proper going forward. I know this is not
a beautiful solution and I think we can do better, but let's see
how much we will have to describe them.
- update doc/README.kconfig
More cleaning up patches will follow this.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
It is true that malloc is necessary for Driver Model before
relocation, but there is no good reason to reserve the malloc
space more than enough. The default value 0x400 works well.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
There 4 JRs, 4 RTICs and 8 DECOs, and set them the same stream id
for using the same SMMU3 on LS1021A.
Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: Alison Wang <alison.wang@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
The RCPM FSM may not be reset after power-on, for example,
in the cases of cold boot and wakeup from deep sleep.
It causes cache coherency problem and may block deep sleep.
Therefore, reset them if they are not be reset.
Signed-off-by: Chenhui Zhao <chenhui.zhao@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
LS1021A's PCIe1 region begins 0x40_00000000; PCIe2 begins
0x48_00000000. In order to access PCIe device, we must create
TLB to map the 40bit physical address to 32bit virtual address.
This patch will enable MMU after DDR is available and creates MMU
table in DRAM to map all 4G space; then, re-use the reserved space
to map PCIe region. The following the mapping layout.
VA mapping:
------- <---- 0GB
| |
| |
|-------| <---- 0x24000000
|///////| ===> 192MB VA map for PCIe1 with offset 0x40_0000_0000
|-------| <---- 0x300000000
| |
|-------| <---- 0x34000000
|///////| ===> 192MB VA map for PCIe2 with offset 0x48_0000_0000
|-------| <---- 0x40000000
| |
|-------| <---- 0x80000000 DDR0 space start
|\\\\\\\|
|\\\\\\\| ===> 2GB VA map for 2GB DDR0 Memory space
|\\\\\\\|
------- <---- 4GB DDR0 space end
Signed-off-by: Minghuan Lian <Minghuan.Lian@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
This patch is to define default values for some CCSR macros
to make header files cleaner.
Signed-off-by: Alison Wang <alison.wang@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
Upgrade Manage Complex (MC) flib API to 0.5.2. Rename directory
fsl_mc to fsl-mc. Change the fsl-mc node in Linux device tree
from "fsl,dprcr" to "fsl-mc". Print MC version info when
appropriate.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Lijun Pan <Lijun.Pan@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
FSL-LSCH3 platforms can have multiple DDR clocks. LS2085A has one clock for
general DDR controlers, and another clock for DP-DDR. DDR driver needs to
change to support multiple clocks.
Signed-off-by: York Sun <yorksun@freescale.com>
This patch adds the fdt-fixup logic for the clock frequency of the
NS16550A related device tree nodes.
Signed-off-by: Bhupesh Sharma <bhupesh.sharma@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
Flushing L3 cache in CCN-504 requries d-cache to be disabled. Using
assembly function to guarantee stack is not used before flushing is
completed. Timeout is needed for simualtor on which CCN-504 is not
implemented. Return value can be checked for timeout situation.
Change bootm.c to disable dcache instead of simply flushing, required
by flushing L3.
Signed-off-by: York Sun <yorksun@freescale.com>
