It has been noticed on MT7628/88 platforms, that booting the RAM image
does not work reliably. Sometimes it works and sometimes not. Debugging
showed that this "might" be a cache related issue as very strange
errors occurred (e.g. output corrupted etc).
This patch adds a cache flush for the complete SDRAM area to the go cmd
before jumping to the entry point for the MIPS architecture. The
complete area is flushed as we don't know at this point, how big the
area of the "application" really is.
Signed-off-by: Stefan Roese <sr@denx.de>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Tested-by: Mauro Condarelli <mc5686@mclink.it>
Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Cc: Mauro Condarelli <mc5686@mclink.it>
Cc: Weijie Gao <weijie.gao@mediatek.com>
This patch fixes an stability issue seen on some vcoreiii targets,
which was root caused to a cache inconsistency situation.
The inconsistency was caused by having kuseg pointing to NOR area but
used as a stack/gd/heap area during initialization, while only
relatively late remapping the RAM area into kuseg position.
The fix is to initialize the DDR right after the TLB setup, and then
remapping it into position before gd/stack/heap usage.
Reported-by: Ramin Seyed-Moussavi <ramin.moussavi@yacoub.de>
Reviewed-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Reviewed-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: Lars Povlsen <lars.povlsen@microchip.com>
This patch ajust CONFIG_SYS_MEMTEST_END for vcoreiii-based systems to
avoid overwriting the relocated u-boot. The former setting was too
agressive with networking etc. enabled on some platforms.
Reviewed-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Reviewed-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: Lars Povlsen <lars.povlsen@microchip.com>
Code sections should be syntax highlighted as bash.
Comment lines in code should start with a hash sign ('#') but code lines
should not.
Most commands can be executed as normal users. Prepend those commands
requiring elevated authorization with 'sudo'.
dd does not have a parameter cout.
sfdisk does not have a -C parameter on Debian Buster.
Provide the necessary input to sfdisk.
Creating a partition of length zero makes no sense.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Allow SoC or board layers with reconfigurable cpu clocks
capabilties to do implementation specific lookups and service
get_tbclk() requests.
Signed-off-by: Alex Nemirovsky <alex.nemirovsky@cortina-access.com>
Useful in custom HW designs which have a need to flush dcache
range in a completely non standard way.
Signed-off-by: Alex Nemirovsky <alex.nemirovsky@cortina-access.com>
Add a test for the dm drivers command. Also fix a null pointer dereference
revealed by said test.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
- sync all Amlogic DT from Linux v5.6-rc2
- MMC clock fixups
- add support for Libre Computer AML-S905D-PC and AML-S912-PC
-----BEGIN PGP SIGNATURE-----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=PiOt
-----END PGP SIGNATURE-----
Merge tag 'u-boot-amlogic-20200406' of https://gitlab.denx.de/u-boot/custodians/u-boot-amlogic into next
- clk: meson-g12a: missing break
- sync all Amlogic DT from Linux v5.6-rc2
- MMC clock fixups
- add support for Libre Computer AML-S905D-PC and AML-S912-PC
pd_dma_* nodes should be accessible during pre-relocation stage of
U-Boot proper for properly handling power domains.
This fixes the issue with permanent failing of invocation of
power_domain_get_by_index() in the common code of DM power domain
uclass (drivers/power/domain/power-domain-uclass.c).
Fixes: f0cc4eae9a ("core: device: use dev_power_domain_on")
Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
Reviewed-by: Oleksandr Suvorov <oleksandr.suvorov@toradex.com>
pd_dma_* nodes should be accessible during pre-relocation stage of
U-Boot proper for properly handling power domains.
This fixes the issue with permanent failing of invocation of
power_domain_get_by_index() in the common code of DM power domain
uclass (drivers/power/domain/power-domain-uclass.c).
Fixes: f0cc4eae9a ("core: device: use dev_power_domain_on")
Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
Reviewed-by: Oleksandr Suvorov <oleksandr.suvorov@toradex.com>
For non-SPL/TPL setups dm-spl, dm-tpl, dm-pre-proper, dm-pre-reloc are
handled equally, forcing the nodes with these properties
to be accessible and device being probed
before pre-relocation of U-Boot proper (drivers/core/util.c):
bool ofnode_pre_reloc(ofnode node)
{
/* for SPL and TPL the remaining nodes after the fdtgrep 1st pass
* had property dm-pre-reloc or u-boot,dm-spl/tpl.
* They are removed in final dtb (fdtgrep 2nd pass)
*/
return true;
if (ofnode_read_bool(node, "u-boot,dm-pre-reloc"))
return true;
if (ofnode_read_bool(node, "u-boot,dm-pre-proper"))
return true;
/*
* In regular builds individual spl and tpl handling both
* count as handled pre-relocation for later second init.
*/
if (ofnode_read_bool(node, "u-boot,dm-spl") ||
ofnode_read_bool(node, "u-boot,dm-tpl"))
return true;
return false;
}
Howewer, to avoid confusion in future, replace dm-spl
`%s/dm-spl/dm-pre-proper/g` properties to dm-pre-proper
to explicitly state that they are handled during pre-relocation
stage of U-Boot proper.
Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
Reviewed-by: Oleksandr Suvorov <oleksandr.suvorov@toradex.com>
For non-SPL/TPL setups dm-spl, dm-tpl, dm-pre-proper, dm-pre-reloc are
handled equally, forcing the nodes with these properties
to be accessible and device being probed
before pre-relocation of U-Boot proper (drivers/core/util.c):
bool ofnode_pre_reloc(ofnode node)
{
/* for SPL and TPL the remaining nodes after the fdtgrep 1st pass
* had property dm-pre-reloc or u-boot,dm-spl/tpl.
* They are removed in final dtb (fdtgrep 2nd pass)
*/
return true;
if (ofnode_read_bool(node, "u-boot,dm-pre-reloc"))
return true;
if (ofnode_read_bool(node, "u-boot,dm-pre-proper"))
return true;
/*
* In regular builds individual spl and tpl handling both
* count as handled pre-relocation for later second init.
*/
if (ofnode_read_bool(node, "u-boot,dm-spl") ||
ofnode_read_bool(node, "u-boot,dm-tpl"))
return true;
return false;
}
Howewer, to avoid confusion in future, replace dm-spl
`%s/dm-spl/dm-pre-proper/g` properties to dm-pre-proper
to explicitly state that they are handled during pre-relocation
stage of U-Boot proper.
Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
Reviewed-by: Oleksandr Suvorov <oleksandr.suvorov@toradex.com>
This fixes the issue when PHY auto negotiation never completes.
Fixes: 431cd76dd8("colibri_imx6: migrate to DM_ETH")
Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
Reviewed-by: Oleksandr Suvorov <oleksandr.suvorov@toradex.com>
The commit a673025535 ("serial: zynq: Initialize uart only before
relocation") introduced code which detects relocation which is working for
single uart instance. With multiple instances in place there is a need to
enable and setup every instance. That's why detect if TX is enabled. If it
is then don't initialize uart again.
In post probe setbrg is called to setup baudrate but values should be the
same.
As a side effect of this change is that DECLARE_GLOBAL_DATA_PTR can be
removed completely.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Add support for Xilinx window watchdog, which can be found on
Versal platforms.
Signed-off-by: Ashok Reddy Soma <ashok.reddy.soma@xilinx.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Similar change was done in past by commit 3b644a3c2f
("arm64: zynqmp: Provide a config to not map DDR region in MMU table").
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
If ATF doesn't exist generate u-boot.itb without it and let U-Boot run in
EL3. Still keep warning to let user know that ATF/BL31 is missing.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Parameter order is not correct based on zynqmp_mmio_write() declaration.
Fixes: be52372ff1 ("arm64: zynqmp: Use zynqmp_mmio_read/write functions")
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
U-Boot proper can still run in EL3 without using firmware interface wired
via ATF. For supporting this use case there is a need to check EL level
where U-Boot runs and based on that choose the way how to talk to firmware.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Delay required for clock propagation is tighly coupled with initialization
done in psu_init(). That's why call it also for u-boot proper with
CONFIG_ZYNQMP_PSU_INIT_ENABLED enabled.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
board_early_init_f() is the right location where debug uart can be
configurated (after MIO initialization).
The patch is taking this call from SPL to also make it available for U-Boot
proper.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Base autodetection is failing for this case:
if test 257 -gt 3ae; then echo first; else echo second; fi
It is because base for 3ae is recognized by _parse_integer_fixup_radix() as
10. The code detects the first char which is not between 'a'/'A' or 'f'/'F'
to change base from dec to hex.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Shiril Tichkule <shirilt@xlinx.com>
When local-mac-address DT property is specified it is reported the same way
as address read from eeprom/ROM. Show properly if mac address is coming
from DT or ROM.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
I found this issue when was running py/test.py on zcu102 which is for me by
default setup to SD boot mode without any way to change boot mode.
Alternative software bootmode selection to JTAG is not working because JTAG
mode is 0 which also reset value for it. That's why saying SPL to take
u-boot.itb from RAM instead of SD in SD boot mode is not possible via
alternative bootmode selection.
That's why setup third boot mode to JTAG(BOOT_DEVICE_RAM) as final
fallback.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Not all IPs have private MDIO bus and MDIO bus should be shared between
several IPs. In past one patch tried to implement it
(https://lists.denx.de/pipermail/u-boot/2018-February/319285.html)
in pretty raw way but it is not the cleanest solution.
This patch is just taking the part of that solution to be able to handle it
over releases without conflicts.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
This information is shown already that's why there is no reason to print it
again via custom prints.
U-Boot SPL 2020.01-03080-ga6214d033bd0 (Mar 05 2020 - 09:59:05 +0100)
mmc boot
Trying to boot from MMC1
or
U-Boot SPL 2020.01-03080-ga6214d033bd0 (Mar 05 2020 - 10:49:46 +0100)
qspi boot
Trying to boot from SPI
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
ZynqMP mini configurations works without PMU firmware that's why there is
no reason to enable the driver and check if it was probed properly.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Enable support for Gigadevice/ISSI flash parts for Versal platform.
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
All configs are using mmc as default fat interface. That's why make it
default for everybody.
The reason for this patch is to make it default for Xilinx Zynq platform
which is not listed there.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
With multi defconfig NOR flash information about NOR should be taken from
DT that's why there is no reason to specify address and sizes via fixed
config.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
With multi defconfig targeting several board configurations bug report like
below is so verbose.
Flash: ## Unknown flash on Bank 1 - Size = 0x00000000 = 0 MB
0 Bytes
Do not report that message and simply say "Flash: 0 Bytes" because most of
Zynq boards are using different type of flashes than NOR.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
There is no reason to continue when DT status property indicates that NAND
flash is disabled. But that means that NOR flash should be present that's
why try it find it out.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
xnand structure is private data structure and it is handled by core and
probe shouldn't touch it.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
There is no real need to include full DT when only some nodes are enough to
use. It will save some space.
Retested with FSBL for initial SoC setup. SPL didn't work.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Add cache flush to zynq_gem_free_pkt. This is necessary
because some net routines would modify this buffer in place.
The cache_invalidate in the zynq_gem_recv function would cause
the modifications to the buffer to overwrite the DMA from the GEM,
if cache coherency is not enabled in the GEM, the next time the
buffer is in use.
Flushing the cache when the buffer is no longer in use by the
net functions ensures that the GEM DMA is going to take place
into a clean buffer.
Signed-off-by: Ashok Reddy Soma <ashok.reddy.soma@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
There are a lot of zynq configurations which can be merged together and use
only one for all. The similar change has been done for ZynqMP by commit
be1b6c32d9 ("arm64: zynqmp: Use zynqmp_virt platform")
Build SPL with u-boot.img for zc706 like this.
export DEVICE_TREE=zynq-zc706 && make xilinx_zynq_virt_defconfig && make -j8
u-boot.img is generic for all boards.
Tested on Zybo, zc702, zc706, zc770-xm011-x16, cc108 and microzed.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Users have option to overwrite default device tree
(CONFIG_DEFAULT_DEVICE_TREE) via environment variable DEVICE_TREE.
Feature has been added long time ago by commit 74de8c9a16
("dts/Makefile: Build the user specified dts") for a little bit different
reason.
But this variable can be also used for different purpose like choosing
proper configuration from FIT image in SPL.
And this is the functionality I would like to use on Xilinx Zynq devices
that current u-boot.img can be composed in the same way based on OF_LIST
and different configuration is taken based on platform specific SPL.
SPL requires low level ps7_init_gpl configuration that's why different
boards require different SPL with fixed board_fit_config_name_match().
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Instead of symlink include origin file and just change model description.
Difference is not in DT but in ps7_init configurations which is taken based
on device tree name that's why the same DT can't be used.
Also update model and update comments to match configurations.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
This patch will add default values for ENV_OFFSET
and ENV_SECT_SIZE for Microblaze.
Signed-off-by: Varalaxmi Bingi <varalaxmi.bingi@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Pinctrl is handled via firmare interface that's why move it there without
reg property and new compatible string.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
