On ARM, the gd pointer is stored in registers r9 / x18. For this the
-ffixed-r9 / -ffixed-x18 flag is used when compiling, but using global
register variables causes errors when building with LTO, and these
errors are very difficult to overcome.
Richard Biener says [1]:
Note that global register vars shouldn't be used with LTO and if they
are restricted to just a few compilation units the recommended fix is
to build those CUs without -flto.
We cannot do this for U-Boot since all CUs use -ffixed-reg flag.
It seems that with LTO we could in fact store the gd pointer differently
and gain performance or size benefit by allowing the compiler to use
r9 / x18. But this would need more work.
So for now, when building with LTO, go the clang way, and instead of
declaring gd a global register variable, we make it a function call via
macro.
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68384
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Simon Glass <sjg@chromium.org>
This commit does the same thing as Linux commit 33def8498fdd.
Use a more generic form for __section that requires quotes to avoid
complications with clang and gcc differences.
Remove the quote operator # from compiler_attributes.h __section macro.
Convert all unquoted __section(foo) uses to quoted __section("foo").
Also convert __attribute__((section("foo"))) uses to __section("foo")
even if the __attribute__ has multiple list entry forms.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
This reverts commit 38d6b7ebda.
struct global_data contains a pointer to the bd_info structure. This
pointer was populated spl_set_bd() to a pre-allocated bd_info in the
".data" section. The referenced commit replaced this mechanism to one
that uses malloc(). That new mechanism is only used if SPL_ALLOC_BD=y.
which very few boards do.
The result is that (struct global_data)->bd is NULL in SPL on most
platforms. This breaks falcon mode, since arch_fixup_fdt() tries to
access (struct global_data)->bd and set the "/memory" node in the
devicetree. The result is that the "/memory" node contains garbage
values, causing linux to panic() as it sets up the page table.
Instead of trying to fix the mess, potentially causing other issues,
revert to the code that worked, while this change is reworked.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
DM_SERIAL needs both the device tree as well as an early heap. Thus, we
have to call spl_early_init() to initialize the memory allocator and the
setup the device tree.
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
To use the debug UART we have to call debug_uart_init() in the SPL. Do
so as soon as possible.
As an example, here is how you can use it on a LS1028A SoC:
CONFIG_DEBUG_UART=y
CONFIG_DEBUG_UART_BASE=0x21c0500
CONFIG_DEBUG_UART_CLOCK=200000000
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
Several architectures had a default board_fit_config_name_match already;
this provides a generic weak version. We default to rejecting all configs.
This will use the FIT's default config, instead of the first config. This
may result in boot failures if there are multiple configurations and the
first config is *not* the default.
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
This board has not been converted to CONFIG_DM_MMC by the deadline.
Remove it.
Cc: Eddy Petrișor <eddy.petrisor@gmail.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
HS400 mode on the LS1028A SoC isn't reliable. The linux driver has a
workaroung for the pulse width detection. Apply this workaround in
u-boot, too.
This will make HS400 mode work reliably on the LS1028A SoC.
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
LS1028A SoCs are restricted in what divider values are allowed for HS400
mode. This is basically a port from the corresponding linux driver.
Signed-off-by: Michael Walle <michael@walle.cc>
Historically, the reset_cpu() function had an `addr` parameter which was
meant to pass in an address of the reset vector location, where the CPU
should reset to. This feature is no longer used anywhere in U-Boot as
all reset_cpu() implementations now ignore the passed value. Generic
code has been added which always calls reset_cpu() with `0` which means
this feature can no longer be used easily anyway.
Over time, many implementations seem to have "misunderstood" the
existence of this parameter as a way to customize/parameterize the reset
(e.g. COLD vs WARM resets). As this is not properly supported, the
code will almost always not do what it is intended to (because all
call-sites just call reset_cpu() with 0).
To avoid confusion and to clean up the codebase from unused left-overs
of the past, remove the `addr` parameter entirely. Code which intends
to support different kinds of resets should be rewritten as a sysreset
driver instead.
This transformation was done with the following coccinelle patch:
@@
expression argvalue;
@@
- reset_cpu(argvalue)
+ reset_cpu()
@@
identifier argname;
type argtype;
@@
- reset_cpu(argtype argname)
+ reset_cpu(void)
{ ... }
Signed-off-by: Harald Seiler <hws@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
At present this function does not accept a size for the FIT. This means
that it must be read from the FIT itself, introducing potential security
risk. Update the function to include a size parameter, which can be
invalid, in which case fit_check_format() calculates it.
For now no callers pass the size, but this can be updated later.
Also adjust the return value to an error code so that all the different
types of problems can be distinguished by the user.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reported-by: Bruce Monroe <bruce.monroe@intel.com>
Reported-by: Arie Haenel <arie.haenel@intel.com>
Reported-by: Julien Lenoir <julien.lenoir@intel.com>
This board has not been converted to CONFIG_DM_MMC by the deadline of
v2019.04, which is almost two years ago. In addition there are other DM
migrations it is also missing. Remove it.
Cc: Prabhakar Kushwaha <prabhakar.kushwaha@nxp.com>
Cc: Priyanka Jain <priyanka.jain@nxp.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
Enable the I-Cache to speed up the boot time, especailly for the NOR
boot, currently it takes about 15 seconds from power up to the U-Boot
prompt, and with the I-Cache enabled it only takes around 2.5 seconds.
Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
Move this out of the common header and include it only where needed. In
a number of cases this requires adding "struct udevice;" to avoid adding
another large header or in other cases replacing / adding missing header
files that had been pulled in, very indirectly. Finally, we have a few
cases where we did not need to include <asm/global_data.h> at all, so
remove that include.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
No version information is used in armv8/fwcall.c therefore do not include
version.h header file. This change prevents recompiling fwcall.o when
SOURCE_DATE_EPOCH changes.
Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
H616 uses different address for reset. Add it.
Signed-off-by: Jernej Skrabec <jernej.skrabec@siol.net>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
So far we did not support the BootROM based FEL USB debug mode on the
64-bit builds for Allwinner SoCs: The BootROM is using AArch32, but the
SPL runs in AArch64.
Returning back to AArch32 was not working as expected, since the RMR
reset into 32-bit mode always starts execution in the BootROM, but not
in the FEL routine.
After some debug and research and with help via IRC, the CPU hotplug
mechanism emerged as a solution: If a certain R_CPUCFG register contains
some magic, the BootROM will immediately branch to an address stored in
some other register. This works well for our purposes.
Enable the FEL feature by providing early AArch32 code to first save the
FEL state, *before* initially entering AArch64.
If we eventually determine that we should return to FEL, we reset back
into AArch32, and use the CPU hotplug mechanism to run some small
AArch32 code snippet that restores the initially saved FEL state.
That allows the normal AArch64 SPL build to be loaded via the sunxi-fel
tool, with it returning into FEL mode, so that other payloads can be
transferred via FEL as well.
Tested on A64, H5 and H6.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Tested-by: Icenowy Zheng <icenowy@aosc.io>
Acked-by: Jagan Teki <jagan@amarulasolutions.com>
Tested-by: Priit Laes <plaes@plaes.org> (on Olimex A64-Olinuxino)
Driver model: Rename U_BOOT_DEVICE et al.
dtoc: Tidy up and add more tests
ns16550 code clean-up
x86 and sandbox minor fixes for of-platdata
dtoc prepration for adding build-time instantiation
-----BEGIN PGP SIGNATURE-----
iQFFBAABCgAvFiEEslwAIq+Gp8wWVbYnfxc6PpAIreYFAl/09LURHHNqZ0BjaHJv
bWl1bS5vcmcACgkQfxc6PpAIrebjwwf/fHRjYsAY/Yj/+y1xgo3L3sphIvQUqTDF
KkLl+kHdV5r8W/HJULxLQcF2r7pcPEI6TAQxuj3qQ5SUvm2HviS8GHGPawDEwyht
HgBp9VD56+HUadMfnbG//DVS73ycbL4XSKlYqpkINEejtnlttsCIawUXX5cTyGM/
59VkgnKrKvJQRUXvYLa8MTugTs4fkPJGDqhActBk/7SP1SImj+rfalNSqA2/dx6y
2RnPCSzB1x2231KSj+B1NgGlR3Xb8P8zgh20ijcEU/hrlXBTZyi7K7f4SJR30Efu
LYkkuj4VbxcV/25RozR0fmknqCs0QyAI+/dql6TNtbTSPC/jAfj0jQ==
=9kN3
-----END PGP SIGNATURE-----
Merge tag 'dm-pull-5jan21' of git://git.denx.de/u-boot-dm into next
Driver model: make some udevice fields private
Driver model: Rename U_BOOT_DEVICE et al.
dtoc: Tidy up and add more tests
ns16550 code clean-up
x86 and sandbox minor fixes for of-platdata
dtoc prepration for adding build-time instantiation
Most drivers use these access methods but a few do not. Update them.
In some cases the access is not permitted, so mark those with a FIXME tag
for the maintainer to check.
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Pratyush Yadav <p.yadav@ti.com>
This uses up space in the SPL binary but it always starts as zero. Also
some boards cannot support data in TPL (e.g. Intel Apollo Lake).
Use malloc() to allocate this structure instead, by moving the init a
little later, after malloc() is inited. Make this function optional since
it pulls in malloc().
This reduces the TPL binary size on coral by about 64 bytes
Signed-off-by: Simon Glass <sjg@chromium.org>
The USB support is not by default enabled on all designs, so it does not
make seance to have USB specific erratum's enabled on such a designs.
On our internal Hitachi-Powergrids design not using the USB controller
there is a crash when accessing those specific memory locations selected
by the erratum flags.
Signed-off-by: Aleksandar Gerasimovski <aleksandar.gerasimovski@hitachi-powergrids.com>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
When booting from TF-A there is a logic that attempt to detect if the
default environment is used, if this is the case it then set the
`bootcmd` and `mcinitcmd` depending of the device we booted from.
This detection logic is dubious as it access internals of the env
implementation and it doesn't always work correctly.
First of all it detect any valid environment as not being the
default, so after running `env default -a && saveenv` the board
doesn't boot anymore as `bootcmd` is then empty.
But it also fails in some other ways, for example it always detect a
default environment when redundant env is enabled on MMC, so in that
case `bootcmd` is overwritten on every boot.
Instead of increasing the complexity of the detection just check if
`bootcmd` and `mcinitcmd` are set in the environment and set them if
they are not.
Signed-off-by: Alban Bedel <alban.bedel@aerq.com>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
Check for NULL return value from fdt_getprop() in
fdt_fixup_remove_jr()
Signed-off-by: Priyanka Singh <priyanka.singh@nxp.com>
[Fixed checkpatch errors/warnings]
Signed-off-by: Priyanka Jain <priyanka.jain@nxp.com>
LX2162 is LX2160 based SoC, it has same die as of LX2160
with different packaging.
LX2162A support 64-bit 2.9GT/s DDR4 memory, i2c, micro-click module,
microSD card, eMMC support, serial console, qspi nor flash, qsgmii,
sgmii, 25g, 40g, 50g network interface, one usb 3.0 and serdes
interface to support three PCIe gen3 interface.
Signed-off-by: Meenakshi Aggarwal <meenakshi.aggarwal@nxp.com>
[Fixed whitespace errors]
Signed-off-by: Priyanka Jain <priyanka.jain@nxp.com>
The BL31 expects the GIC to be uninitialized. Thus, if we are loading
the BL31 by the SPL we must not initialize it. If u-boot is loaded by
the SPL directly, it will initialize the GIC again (in the same
lowlevel_init()).
This was tested on a custom board with SPL loading the BL31 and jumping
to u-boot as BL33 as well as loading u-boot directly by the SPL. In case
the ATF BL1/BL2 is used, this patch won't change anything, because no
SPL is used at all.
Signed-off-by: Michael Walle <michael@walle.cc>
There is SPL_ARMV8_SEC_FIRMWARE_SUPPORT and ARMV8_SEC_FIRMWARE_SUPPORT.
Thus use CONFIG_IS_ENABLED() instead of the simple #ifdef.
Signed-off-by: Michael Walle <michael@walle.cc>
Acked-by: Michal Simek <michal.simek@xilinx.com>
On all newer Layerscape SoCs, only the export-controlled ciphers of the
crypto module are disabled on non-E parts. Thus it doesn't make sense to
completely remove the node. Linux will figure out what is there and what
is not.
Just remove it for older SoCs, where the module is indeed completely
disabled on non-E parts.
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Horia Geanta <horia.geanta@nxp.com>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
As per hardware documentation, ECx_PMUX has precedence
over SerDes protocol.
For LX2160/LX2162 if DPMACs 17 and 18 are enabled as SGMII
through SerDes protocol but ECx_PMUX configured them as RGMII,
then the ports will be configured as RGMII and not SGMII.
Signed-off-by: Razvan Ionut Cirjan <razvanionut.cirjan@nxp.com>
[Rebased]
Signed-off-by: Priyanka Jain <priyanka.jain@nxp.com>
In the current implementation, u-boot creates iommu mappings only
for PCI devices enumarated at boot time thus does not take into
account more dynamic scenarios such as SR-IOV or PCI hot-plug.
Add an u-boot env var and a device tree property (to be used for
example in more static scenarios such as hardwired PCI endpoints
that get initialized later in the system setup) that would allow
two things:
- for a SRIOV capable PCI EP identified by its B.D.F specify
the maximum number of VFs that will ever be created for it
- for hot-plug case, specify the B.D.F with which the device
will show up on the PCI bus
More details can be found in the included documentation:
arch/arm/cpu/armv8/fsl-layerscape/doc/README.pci_iommu_extra
Signed-off-by: Laurentiu Tudor <laurentiu.tudor@nxp.com>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
When the actual offset between link and runtime address is zero, there
is no need for patching up U-Boot early when running with
CONFIG_POSITION_INDEPENDENT.
Skip the whole routine when the distance is 0.
This helps when U-Boot is loaded into ROM, or in otherwise sensitive
memory locations.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Make sure that SW_RST_REQ and RST_REQ_MSK are cleared
before triggering hardware reset request.
Signed-off-by: Thirupathaiah Annapureddy <thiruan@linux.microsoft.com>
Signed-off-by: Meenakshi Aggarwal <meenakshi.aggarwal@nxp.com>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
The workaround of LPI one-way reset issue is broken by the series:
https://patchwork.ozlabs.org/project/uboot/list/?series=192398
This patch is to add DT node for GIC RD tables and create corresponding
reserved-memory node in kernel DT to fix it.
Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
Configure DWC3’s cache type to ‘cacheable’ for better
performance. Actually related register definition and values are SoC
specific, which means this setting is only applicable to Layerscape SoC,
not generic for all platforms which have integrated DWC3 IP.
Signed-off-by: Ran Wang <ran.wang_1@nxp.com>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
Linking a U-Boot larger than 1MB fails with PIE enabled:
u-boot/arch/arm/cpu/armv8/start.S:71:(.text+0x3c): relocation
truncated to fit: R_AARCH64_ADR_PREL_LO21 against symbol `__rel_dyn_end'
defined in .bss_start section in u-boot.
This extends the supported range by using adrp & add to load symbols
early while starting up.
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
PIE requires a 4K aligned load address. If this is not met, trap
the startup sequence in a WFI loop rather than running into obscure
failures.
Tested-by: Michal Simek <michal.simek@xilinx.com>
Suggested-by: André Przywara <andre.przywara@arm.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Introduce a minimal Xen guest board running as a virtual
machine under Xen Project's hypervisor [1], [2].
Part of the code is ported from Xen mini-os and also uses
work initially done by different authors from NXP: please see
relevant files for their copyrights.
[1] https://xenbits.xen.org
[2] https://wiki.xenproject.org/
Signed-off-by: Andrii Anisov <andrii_anisov@epam.com>
Signed-off-by: Oleksandr Andrushchenko <oleksandr_andrushchenko@epam.com>
Signed-off-by: Anastasiia Lukianenko <anastasiia_lukianenko@epam.com>
We still have some platforms that only implements functionalities in
PSCI 0.1 (e.g. Allwinner ARMv7 SoCs).
Add a Kconfig option for exporting only PSCI 0.1. The code to export
PSCI 0.1 is still available and gets activated by this patch.
In addition, default ARCH_SUNXI U-Boot PSCI implementation to export
PSCI 0.1, to fix poweroff/reboot regression on Allwinner multi-core
ARMv7 SoCs.
Signed-off-by: Icenowy Zheng <icenowy@aosc.io>
Use device tree and UCLASS_SYSCON driver to get
Generic Interrupt Controller (GIC) lpi address and
maximum GIC redistributors count.
Also update Kconfig to select REGMAP and SYSCON when
GIC_V3_ITS is enabled.
Signed-off-by: Rayagonda Kokatanur <rayagonda.kokatanur@broadcom.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Add API to save boot parameters passed from BL31
Use assembly implementation of save_boot_params instead of c function.
Because generally ATF does not set up SP_EL2 on exiting.
Thus, usage of a C function immediately after exiting with no stack
setup done by ATF explicitly, may cause SP_EL2 to be not sane,
which in turn causes a crash if this boot was not lucky to get
an SP_EL2 in valid range. Replace C implementation with assembly one
which does not use stack this early, and let u-boot to set up its stack
later.
Signed-off-by: Abhishek Shah <abhishek.shah@broadcom.com>
Signed-off-by: Rajesh Ravi <rajesh.ravi@broadcom.com>
Signed-off-by: Vladimir Olovyannikov <vladimir.olovyannikov@broadcom.com>
Signed-off-by: Rayagonda Kokatanur <rayagonda.kokatanur@broadcom.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Changes in relation to FriendlyARM's U-Boot nanopi2-v2016.01:
- SPL not supported yet --> no spl-dir in arch/arm/cpu/armv7/s5p4418/.
Appropriate line in Makefile removed.
- cpu.c: '#include <cpu_func.h>' added.
- arch/arm/cpu/armv7/s5p4418/u-boot.lds removed, is not required
anylonger.
- "obj-$(CONFIG_ARCH_NEXELL) += s5p-common/" added to
arch/arm/cpu/armv7/Makefile since s5p-common/pwm.c is used instead
of drivers/pwm/pwm-nexell.c.
- s5p4418.dtsi: '#include "../../../include/generated/autoconf.h"'
removed, is not necessary, error at out-of-tree building.
'#ifdef CONFIG_CPU_NXP4330'-blocks (2x) removed. Some minor changes
regarding mmc. 'u-boot,dm-pre-reloc' added to dp0 because of added
DM_VIDEO support.
- board/s5p4418/ renamed to board/friendlyarm/
- All s5p4418-boards except nanopi2 removed because there is no
possibility to test the other boards.
- Kconfig: Changes to have a structure like mach-bcm283x (RaspberryPi),
e.g. "config ..." entries moved from/to other Kconfig.
- "CONFIG_" removed from several s5p4418/nanopi2 specific defines
because the appropriate values do not need to be configurable.
- nanopi2/board.c: All getenv(), getenv_ulong(), setenv() and saveenv()
renamed to env_get(), env_get_ulong(), env_set() and env_save(),
respectively. MACH_TYPE_S5P4418 is not defined anymore, therefore
appropriate code removed (not necessary for DT-kernels).
- nanopi2/onewire.c: All crc8() renamed to crc8_ow() because crc8() is
already defined in lib/crc8.c (with different parameters).
- dts: "nexell,s5pxx18-i2c" used instead of "i2c-gpio", i2c0 and
i2c1 added. gmac-, ehci- and dwc2otg-entries removed because the
appropriate functionality is not supported yet. New mmc-property
"mmcboost" added.
s5p4418-pinctrl.dtsi: gmac-entries removed, mmc- and i2c-entries
added.
- '#ifdef CONFIG...' changed to 'if (IS_ENABLED(CONFIG...))' where
possible (and similar).
Signed-off-by: Stefan Bosch <stefan_b@posteo.net>
