| Age | Commit message (Collapse) | Author |
|---|
|
Node name should be <name>@<address> which is not how partitions are
described.
Issue was found by running dtbs_check as:
flash@0: 'partition@qspi-device-tree', 'partition@qspi-fsbl-uboot',
'partition@qspi-linux', 'partition@qspi-rootfs'
do not match any of the regexes: ...
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
Sync DP subsystem with the latest state in Xilinx U-Boot repository.
This binding hasn't been approved in mainline Linux but it is much better
than ancient version which this patch removes.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
In kernel 5.4, support has been added for reading MTD devices via
the nvmem API.
For this the mtd devices are registered as read-only NVMEM providers
under sysfs with the same name as the flash partition label property.
So if flash partition label property of multiple flash devices are identical
then the second mtd device fails to get registered as a NVMEM provider.
This patch fixes the issue by having different label property for different
flashes.
Signed-off-by: Amit Kumar Mahapatra <amit.kumar-mahapatra@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
Add clock-cells and clock-output-names for sdhci0 and sdhci1.
These are needed for linux sdhci driver from 5.4 version onwards.
Signed-off-by: Ashok Reddy Soma <ashok.reddy.soma@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
Reset controller is handled via firmware that's why it should be the part
of firmware node. Origin solution hasn't been removed when above change was
applied by commit b07e97b4ba27 ("arm64: zynqmp: Use reset header in
zynqmp.dtsi").
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
Modify dts files to add 'no-1-8-v' property for all the ZynqMP boards.
User can remove this property to enable the UHS mode. This is to keep
the same speed (HS) modes across all the stages of the Linux Boot. Due
to power cycling limitation of some of the ZynqMP boards, some SD cards
don't get power cycled and are failing in Linux.
Signed-off-by: Manish Narani <manish.narani@xilinx.com>
|
|
Sync zynqmp fpga manager with mainline.
Signed-off-by: Nava kishore Manne <nava.manne@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
All boards have been converted to firmware based driver that's why we can
remove this file now.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
None name address should be aligned with address. DTC 1.5.1 is reporting
issues related to that.
arch/arm/boot/dts/zynq-zc770-xm010.dts:106.10-119.4: Warning
(spi_bus_reg): /amba/spi@e0007000/flash@0: SPI bus unit address format
error, expected "1"
arch/arm/boot/dts/zynq-zc770-xm013.dts:101.19-109.4: Warning
(spi_bus_reg): /amba/spi@e0006000/eeprom@0: SPI bus unit address format
error, expected "2"
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
Trivial change.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
The same change has been done for Zynq by commit 1241c72b6db1
("ARM: dts: zynq: replace gpio-key,wakeup with wakeup-source property")
in mainline Linux kernel.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
Most of the legacy "gpio-key,wakeup" boolean property is already
replaced with "wakeup-source". However few occurrences of old property
has popped up again, probably from the remnants in downstream trees.
Replace the legacy properties with the unified "wakeup-source"
property introduced in the Linux kernel commit 700a38b27eef
("Input: gpio_keys - switch to using generic device properties")
Cc: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
The dtc has new checks for I2C and SPI buses.
Fix the warnings in node names and unit-addresses.
Warning from Linux kernel:
arch/arm/boot/dts/zynq-zc702.dts:187.13-190.6: Warning (i2c_bus_reg): /amba/i2c@e0004000/i2c-mux@74/i2c@7/hwmon@52: I2C bus unit address format error, expected "34"
arch/arm/boot/dts/zynq-zc702.dts:191.13-194.6: Warning (i2c_bus_reg): /amba/i2c@e0004000/i2c-mux@74/i2c@7/hwmon@53: I2C bus unit address format error, expected "35"
arch/arm/boot/dts/zynq-zc702.dts:195.13-198.6: Warning (i2c_bus_reg): /amba/i2c@e0004000/i2c-mux@74/i2c@7/hwmon@54: I2C bus unit address format error, expected "36"
Signed-off-by: Quanyang Wang <quanyang.wang@windriver.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
Multi boot register can be used for using different boot images and design
better boot strategy. Let EL3 SPL or U-Boot to read it and print it.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
|
|
Add support for the Amlogic based libretech-pc platform.
This platform comes with 2 variant, based on the s905d or s912 SoC.
Acked-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
[narmstrong: update board/amlogic/q200/MAINTAINERS]
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
|
|
Sync the libretech-pc device tree from Linux v5.6-rc2
11a48a5a18c6 ("Linux 5.6-rc2")
Acked-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
|
|
Sync the device tree and dt-bindings from Linux v5.6-rc2
11a48a5a18c6 ("Linux 5.6-rc2")
The only exception to this is the mmc pinctrl pin bias of gxl SoC family.
This is a fix which found its way to u-boot but not Linux yet.
Acked-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
|
|
The card-detect GPIO and any other GPIO access currently doesn't work in
U-Boot SPL on any STM32 platform and crashes the SPL. To work around this
problem on AV96 right before release, remove the cd-gpios from DT. This
patch must be reverted right after release, once the proper fix for the
GPIO driver, "gpio: stm32: support gpio ops in SPL", is applied.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Cc: Tom Rini <trini@konsulko.com>
Acked-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
|
|
- verdin-imx8mm board reST documentation update
- Intel Edison board ACPI table I2C/USB minor updates
- Fix a regression of ns16550 serial driver that breaks Intel Edison
|
|
USB 3 host controller may be described in ACPI to allow users alter
the properties or other features. Describe it for Intel Tangier SoC.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
|
|
There is established way to provide I²C timings, or actually counters,
to the OS via ACPI. Fill them for Intel Merrifield platform.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
|
|
There is no need to have an assignment to NULL for XSDT pointer.
Therefore, no need to assign it when rsdt_address is not set.
Because of above changes we may decrease indentation level as well.
While here, drop unnecessary parentheses.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
|
|
- Add support for Jetson Nano, plus miscellaneous other fixes found
during Nano bringup.
- Add Igor's update_uboot wrapper patches.
|
|
Commit f4dc714aaa2d ("arm64: Turn u-boot.bin back into an ELF file after
relocate-rela")
introduce REMAKE_ELF option to recreate u-boot.elf from u-boot ->
u-boot.bin + DT -> u-boot.elf.
The best is to ilustrate it from make V=1 output
cat u-boot-nodtb.bin dts/dt.dtb > u-boot-dtb.bin
cp u-boot-dtb.bin u-boot.bin
aarch64-linux-gnu-objcopy -I binary -B aarch64 -O elf64-littleaarch64 u-boot.bin u-boot-elf.o
aarch64-linux-gnu-ld.bfd u-boot-elf.o -o u-boot.elf --defsym="_start"=0x8000000 -Ttext=0x8000000
Last command has no explicit linker script passed that's why toolchain
internal linker script is used.
In Binutils 2.32 case it contains SIZEOF_HEADERS symbol which has changed
behavior by commit
https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=64029e93683a266c38d19789e780f3748bd6a188
which result in situation that program headers has changed from
(xilinx_zynqmp_mini_defconfig)
Program Headers:
Type Offset VirtAddr PhysAddr
FileSiz MemSiz Flags Align
LOAD 0x0000000000010000 0x00000000fffc0000 0x00000000fffc0000
0x0000000000018918 0x0000000000018918 RW 0x10000
to
Program Headers:
Type Offset VirtAddr PhysAddr
FileSiz MemSiz Flags Align
LOAD 0x0000000000000000 0x00000000fffb0000 0x00000000fffb0000
0x0000000000028918 0x0000000000028918 RW 0x10000
Xilinx tools like XSDB or Bootgen are using program headers for loading ELF
to the right location and by above binutils change ELF is loaded to
incorrect location.
The patch is explicitly use u-boot-elf.lds (just cat now) for u-boot.elf
recreation which is called when REMAKE_ELF is setup.
By purpose u-boot-elf.lds doesn't contain OUTPUT_FORMAT/OUTPUT_ARCH to be
able to use by all archs.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Tested-By: Álvaro Fernández Rojas <noltari@gmail.com>
|
|
When generating the MAC address based on the boards serial number
the last digit was overwritten with the null termination. That way
boards with serial numbers close to each other would use the same
MAC address.
Signed-off-by: Jan-Christoph Tebbe <Jan-Christoph.Tebbe@ithinx.io>
|
|
The Jetson Nano Developer Kit is a Tegra X1-based development board. It
is similar to Jetson TX1 but it is not pin compatible. It features 4GB
of LPDDR4, a SPI NOR flash for early boot firmware and an SD card slot
used for storage.
HDMI 2.0 or DP 1.2 are available for display, four USB ports (3 USB 2.0
and 1 USB 3.0) can be used to attach a variety of peripherals and a PCI
Ethernet controller provides onboard network connectivity. NVMe support
has also been added. Env save is at the end of QSPI (4MB-8K).
A 40-pin header on the board can be used to extend the capabilities and
exposed interfaces of the Jetson Nano.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Tested-by: Peter Robinson <pbrobinson@gmail.com>
|
|
According to the HW team, for some reason the normal clock select code
picks what appears to be a perfectly valid 375KHz SD card clock, based
on the CAR clock source and SDMMC1 controller register settings (CAR =
408MHz PLLP0 divided by 68 for 6MHz, then a SD Clock Control register
divisor of 16 = 375KHz). But the resulting SD card clock, as measured by
the HW team, is 700KHz, which is out-of-spec. So the WAR is to use the
values given in the TRM PLLP table to generate a 400KHz SD-clock (CAR
clock of 24.7MHz, SD Clock Control divisor of 62) only for SDMMC1 on
T210 when the requested clock is <= 400KHz. Note that as far as I can
tell, the other requests for clocks in the Tegra MMC driver result in
valid SD clocks.
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
|
|
As per the T210 TRM, when running at 3.3v, the SDMMC1 tap/trim and
autocal values need to be set to condition the signals correctly before
talking to the SD-card. This is the same as what's being done in CBoot,
but it gets reset when the SDMMC1 HW is soft-reset during SD driver
init, so needs to be repeated here. Also set autocal and tap/trim for
SDMMC3, although no T210 boards use it for SD-card at this time.
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
|
|
Fix VI_I2C clock source type. Will be needed by VI_I2C driver.
Also added use of INTERNAL_ID macro in two places, needed to keep
the id returned to 8 bits.
Signed-off-by: Tom Warren <twarren@nvidia.com>
|
|
T210 CBoot is now doing the full pinmux and GPIO init, based on the DTB
tables. Remove pinmux/GPIO init tables & code from all T210-based builds
below:
p2371-2180 aka TX1
p2371-0000
e2220-1170
p2571
Signed-off-by: Tom Warren <twarren@nvidia.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
|
|
This commit removes the programming sequence that enables PLLE and UPHY
PLL hardware power sequencers. Per TRM, boot software should enable PLLE
and UPHY PLLs in software controlled power-on state and should power
down PLL before jumping into kernel or the next stage boot software.
Adds call to board_cleanup_before_linux to facilitate this.
Signed-off-by: JC Kuo <jckuo@nvidia.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
|
|
This adds to the DT the I2C controllers that connect to the board ID EEPROM,
etc. With this change, you can now probe all I2C devices on a TX1 board.
Signed-off-by: Tom Warren <twarren@nvidia.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
|
|
Enable pre console buffer for rk3399 platform.
This would help to capture the console messages prior to
the console being initialised. Enabling this would help
to capture all the console messages on video output source
like HDMI. So we can find the full console messages of
U-Boot proper on HDMI display when enabled it for RK3399
platform boards.
Buffer address used for pre console is 0x0f200000 which is
ram base plus 240MiB. right now the Allwinner SoC is using
similar computation.
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Tested-by: Peter Robinson <pbrobinson@gmail.com>
|
|
VOP display endpoint pipeline configuration differs
between rk3288 vs rk3399.
These VOP pipeline configuration depends on how the
different display interfaces connected in sequence to
IN and OUT ports like for,
RK3288:
vopb_out: port {
#address-cells = <1>;
#size-cells = <0>;
vopb_out_edp: endpoint@0 {
reg = <0>;
remote-endpoint = <&edp_in_vopb>;
};
vopb_out_hdmi: endpoint@1 {
reg = <1>;
remote-endpoint = <&hdmi_in_vopb>;
};
vopb_out_lvds: endpoint@2 {
reg = <2>;
remote-endpoint = <&lvds_in_vopb>;
};
vopb_out_mipi: endpoint@3 {
reg = <3>;
remote-endpoint = <&mipi_in_vopb>;
};
};
RK3399:
vopb_out: port {
#address-cells = <1>;
#size-cells = <0>;
vopb_out_edp: endpoint@0 {
reg = <0>;
remote-endpoint = <&edp_in_vopb>;
};
vopb_out_mipi: endpoint@1 {
reg = <1>;
remote-endpoint = <&mipi_in_vopb>;
};
vopb_out_hdmi: endpoint@2 {
reg = <2>;
remote-endpoint = <&hdmi_in_vopb>;
};
vopb_out_mipi1: endpoint@3 {
reg = <3>;
remote-endpoint = <&mipi1_in_vopb>;
};
vopb_out_dp: endpoint@4 {
reg = <4>;
remote-endpoint = <&dp_in_vopb>;
};
};
here, HDMI interface has endpoint 1 in rk3288 and 2 in rk3399.
The rockchip vop driver often depends on this determined endpoint
number and stored in vop_mode. So based on this vop_mode the bpp
and pin polarity would configure on detected display interface.
Since, the existing driver using rk3288 vop mode settings enabling
the same will result wrong display interface configuration for rk3399.
Add the patch for fixing these vop modes for rk3399.
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Tested-by: Peter Robinson <pbrobinson@gmail.com>
|
|
https://gitlab.denx.de/u-boot/custodians/u-boot-stm
- Fix device tree of Avenger96 board from Arrow Electronics and add
compatibility with stm32mp15_dhcom_basic_defconfig
|
|
The core and vdd PMIC buck regulators were misconfigured, which caused
instability of the board and malfunction of high-speed interfaces, like
the RGMII. Configure the PMIC correctly to repair these problems. Also,
model the missing Enpirion EP53A8LQI on the DHCOR SoM as a fixed regulator.
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
|
|
Add PHY reset GPIO on AV96 ethernet PHY.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
|
|
The AV96 RGMII uses different pinmux for ETH_RGMII_TXD0, ETH_RGMII_RXD2
and ETH_RGMII_TX_CTL. Use the correct pinmux to make ethernet operational.
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
|
|
Add another mux option for DWMAC RGMII, this is used on AV96 board.
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
|
|
The board has an EEPROM on the same I2C bus as PMIC, at address 0x53.
The EEPROM contains the board MAC address.
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Change-Id: I340a0675c11e4599968b2e3ef0515fb8da8d7b42
|
|
Use DT /aliases node to establish a stable phandle to the configuration
EEPROM. This permits the configuration EEPROM to be moved e.g. to a
different address or a different bus. Adjust the board code to handle
new phandle lookup.
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
|
|
The DH Electronics DHCOR SOM has QSPI NOR on the SoM itself, add it
into the DT.
Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Change-Id: Ia7c454c496f50e3fc4851ec1154f3641c416e98e
|
|
The eMMC uses different pinmux for the top four data lines, use such
a pinmux, otherwise it takes a very long time until the test for 8bit
operation times out. And this is the correct pinmux per schematic too.
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
|
|
Add another mux option for SDMMC2 pins 4..7, this is used on AV96 board.
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
|
|
The SD uses different pinmux for the D123DIRline, use such a pinmux,
otherwise there is a pinmux collision on the AV96. Add missing SD
voltage regulator switch and enable SDR104 operation.
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
|
|
Add another mux option for SDMMC1 direction pins, in particular
SDMMC1_D123DIR, this is used on AV96 board.
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
|
|
The sdmmc1_dir_pins_a: sdmmc1-dir-0 layout changed in commit 35a54d41d9d4
("ARM: dts: stm32mp1: sync device tree with v5.2-rc4") such that pins{};
became pins1{};pins2{};, however the SPL extras were not updated to reflect
that change. Fix this.
This fixes booting from SD1 X9 slot on the AV96 board.
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Fixes: 35a54d41d9d4 ("ARM: dts: stm32mp1: sync device tree with v5.2-rc4")
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
|
|
|
|
https://gitlab.denx.de/u-boot/custodians/u-boot-arc
This last minute pull-request is intended to fix some drivers
when used on ARC boards. The problem was introduced by
https://gitlab.denx.de/u-boot/u-boot/-/commit/07906b3dad157bd58411664bcc6a2a7976d5e0a9
What happened while doing one pretty simple improvement to make
U-Boot port more flexible and portable (by switching accessors from
assembly-written to plain C version) we implicitly added 2 problems:
1. Downgraded accessors from being volatile which signalled to
the compiler that it's now possible to do all kinds of optimizations
which may easily include merge of subsequent byte reads/writes into
word operations. Which is OK for accessing mormal memory but
breaks operation of peripherals if we access its memory-mapped regs
in such a "creative" manner.
2. As a part of assembly-written implementation we had compiler barriers
in form of the following construction 'asm volatile("" : : : "memory")',
and we dropped it in C implemntation. This in its turn enabled compiler
to mess with instruction ordering. Guess what it gives us in the end :)
So with all that we had in some corner-cases veeery funny instruction flows
generated. And in particular it broke DW SPI functionality when we were
writing large amount of data. Funny enough our tests which were writing
small amount of data still worked and only by the chance we caught that
breakage and unrolled that quite interesting loop of unexpected
problems.
The road to hell is paved with good intentions. Amen :)
|
|
We add memory barriers for __raw_readX / __raw_writeX accessors same
way as it is done for readX and writeX accessors as lots of U-boot
driver uses __raw_readX / __raw_writeX instead of proper accessor
with barrier.
It will save us from lot's of debugging in the future and it is OK
as U-Boot is not that performance oriented as real run-time
software like OS or user bare-metal app so we may afford being not
super fast as we only being executed once.
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
|
