Lokesh Vutla [Tue, 31 Dec 2019 10:19:55 +0000 (15:49 +0530)]
arm: mach-k3: Enable WA for R5F deadlock
On K3 devices there are 2 conditions where R5F can deadlock:
1.When software is performing series of store operations to
cacheable write back/write allocate memory region and later
on software execute barrier operation (DSB or DMB). R5F may
hang at the barrier instruction.
2.When software is performing a mix of load and store operations
within a tight loop and store operations are all writing to
cacheable write back/write allocates memory regions, R5F may
hang at one of the load instruction.
To avoid the above two conditions disable linefill optimization
inside Cortex R5F which will make R5F to only issue up to 2 cache
line fills at any point of time.
Lokesh Vutla [Tue, 31 Dec 2019 10:18:48 +0000 (15:48 +0530)]
arm: dts: k3-j721e: ddr: Update to 0.2 version of DDR config tool
Update the ddr settings to use the DDR reg config tool rev 0.2.0.
This reduces the aging count(in DDRSS_CTL_274_DATA reg) to 15 in-order
to avoid DSS underflow errors.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com> Signed-off-by: Kevin Scholz <k-scholz@ti.com>
include/u-boot/crc.h: pull 'uint' definition on musl targets
The build failure was originally reported on arm64-musl
target at https://bugs.gentoo.org/703132. Here is the amd64-musl
variant:
```
$ LANG=C make CROSS_COMPILE=x86_64-gentoo-linux-musl- tools-only_defconfig -j$(nproc)
$ LANG=C make CROSS_COMPILE=x86_64-gentoo-linux-musl- tools-all -j$(nproc)
...
include/u-boot/crc.h:37:44: error: unknown type name 'uint'; did you mean 'int'?
37 | void crc16_ccitt_wd_buf(const uint8_t *in, uint len,
| ^~~~
| int
```
Note: 'uint' is not defined there.
On glibc 'uint' comes from <sys/types.h> and happens to work on most .c files.
The change imports 'uint' declaration from '<compiler.h>'.
Bug: https://bugs.gentoo.org/703132 Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>
The build failure was originally reported on arm64-musl
target at https://bugs.gentoo.org/703132. Here is the amd64-musl
variant:
```
$ LANG=C make CROSS_COMPILE=x86_64-gentoo-linux-musl- tools-only_defconfig -j$(nproc)
$ LANG=C make CROSS_COMPILE=x86_64-gentoo-linux-musl- tools-all -j$(nproc)
...
In file included from tools/env/../../env/flags.c:7,
from tools/env/env_flags.c:1:
include/env.h:159:1: error: unknown type name 'ulong'; did you mean 'long'?
159 | ulong env_get_ulong(const char *name, int base, ulong default_val);
| ^~~~~
| long
```
Note: 'ulong' is not defined there.
On glibc 'ulong' comes from <sys/types.h>:
```c
/* Old compatibility names for C types. */
typedef unsigned long int ulong;
```
On musl it comes from <sys/types.h> as well but from under different guards:
```c
typedef unsigned long u_long, ulong;
```
The change inlines 'ulong' define similar to 'uint' define.
Bug: https://bugs.gentoo.org/703132 Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>
Tom Rini [Thu, 2 Jan 2020 15:28:26 +0000 (10:28 -0500)]
Merge tag 'u-boot-rockchip-20191231' of https://gitlab.denx.de/u-boot/custodians/u-boot-rockchip
- Fix latest mainline kernel for rk3308
- Update rk3288-evb config to suport OP-TEE
- Fix for firefly-px30 DEBUG_UART channel and make it standalone
- Script make_fit_atf add python3 support
- Fix rk3328 timer with correct COUNTER_FREQUENCY
- Fix rk3328 ATF support with enable spl-fifo-mode
Stephen Warren [Wed, 18 Dec 2019 18:37:21 +0000 (11:37 -0700)]
test_env: don't strip() printenv results
get_env() was originally written to strip() the output of printenv to
isolate the test from any whitespace changes in printenv's output.
However, this throws away any whitespace in the variable value, which can
cause issues when test code expects to see that whitespace. In fact,
printenv never adds any whitespace at all, so there's no need to strip.
The strip causes a practical problem for test_env_echo_exists() if
state_test_env.get_existent_var() happens to choose a U-Boot variable that
contains trailing whitespace. This is true for variable boot_targets.
With Python 2, get_existent_var() never returned boot_targets so this
issue never caused a practical problem.
With Python 3, get_existent_var does sometimes return boot_targets, no
doubt due to Python 3's different dict hash key order implementation,
about 0.5-2% of the time, so this test appears intermittent. With the
strip removed, this intermittency is solved, since the test passes for all
possible U-Boot variables.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Patrik Dahlström [Sat, 21 Dec 2019 16:18:22 +0000 (17:18 +0100)]
omap3_beagle: Change NAND ECC scheme back to OMAP_ECC_HAM1_CODE_HW
The omap3_beagle NAND ECC scheme was changed in 4b37928d357 for unknown
reasons, leading to uncorrectible ecc errors. This commit changes it
back to what it was before.
Adam Ford [Sun, 3 Nov 2019 13:58:40 +0000 (07:58 -0600)]
ARM: imx6q_logic: Cleanup boot sequence check
The board_boot_order() function currenly assumes that the boot source
is MMC/eMMC, but this isn't true for the NAND devices.
This patch cleans up board_boot_order() to check for NAND, SD, ESD,
MMC or EMMC. Anything beyond these are not supported, so it will
default back to the serial downloader if any of those devices are not
available.
Fixes: 9fb50c68daa6 ("ARM: imx6q_logic: Fix MMC2 booting") Signed-off-by: Adam Ford <aford173@gmail.com>
Jagan Teki [Sat, 21 Dec 2019 07:54:30 +0000 (13:24 +0530)]
spi: rk: Limit transfers to (64K - 1) bytes
The Rockchip SPI controller's length register only supports 16-bits,
yielding a maximum length of 64KiB (the CTRLR1 register holds "length -
1"). Trying to transfer more than that (e.g., with a large SPI flash
read) will cause the driver to hang.
Now, it seems that while theoretically we should be able to program
CTRLR1 with 0xffff, and get a 64KiB transfer, but that also seems to
cause the core to choke, so stick with a maximum of 64K - 1 bytes --
i.e., 0xffff.
Note, that the size is further divided into 'minus 1' while writing
into CTRLR1.
This change fixed two different read issues,
1. sf read failure when with > 0x10000
2. Boot from SPI flash failed during spi_flash_read call in
common/spl/spl_spi.c
Observed and Tested in
- Rockpro64 with Gigadevice flash
- ROC-RK3399-PC with Winbond flash
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Static checker warns 'ret' variable may be used uninitialized in
spi_nor_erase() and spi_nor_write() in case of zero length requests.
Fix these warnings by checking for zero length requests and returning
early.
Reported-by: Dan Murphy <dmurphy@ti.com> Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
When compile the curren mainline linux kernel(Linux 5.5-rc3)
with defconfig, the final Image is 29M, it's much
larger than Linux 5.4.
On the current u-boot side on rk3308, the gap between
kernel and fdt is 25M, the fdt will overwrite kernel
Image, so move ftd to a higher memory to give 34M
gab for them.
Signed-off-by: Andy Yan <andy.yan@rock-chips.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Thomas Hebb [Sat, 21 Dec 2019 02:05:22 +0000 (18:05 -0800)]
rockchip: add description for TPL_ROCKCHIP_COMMON_BOARD
SPL_ROCKCHIP_COMMON_BOARD, an almost identical option, has a title but
this one doesn't for some reason. Add a description to make the menu
easier to read.
Signed-off-by: Thomas Hebb <tommyhebb@gmail.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Kever Yang [Fri, 13 Dec 2019 09:41:17 +0000 (17:41 +0800)]
rockchip: px30-firefly add standalone dts
Firefly Core-PX30-JD4 use UART2M1 while PX30 evb using UART2M0, the U-Boot
proper will use the dts setting to do the IOMUX init, and a separate dts
is needed for px30-firefly.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Jack Mitchell [Tue, 10 Dec 2019 13:07:21 +0000 (13:07 +0000)]
rockchip: make_fit_atf: explicitly use python3
On a distribution with no python2 installed and no
python->python3 symlink the script will fail to execute.
Specify python3 explicitly as it's already a requirement
to build u-boot.
Signed-off-by: Jack Mitchell <ml@embed.me.uk> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Roger Quadros [Wed, 6 Nov 2019 14:21:17 +0000 (16:21 +0200)]
phy: ti-pipe3: improve DPLL stability for SATA & USB
For increased DPLL stability use the settings recommended in
the TRM [1] for PHY_RX registers for SATA and USB.
For SATA we need to use spread spectrum settings even
though we don't have spread spectrum enabled. The
suggested non-spread spectrum settings don't work.
Roger Quadros [Wed, 6 Nov 2019 14:21:15 +0000 (16:21 +0200)]
phy: ti-pipe3: Use TRM recommended settings for SATA DPLL
The AM572x Technical Reference Manual, SPRUHZ6H,
Revised November 2016 [1], shows recommended settings for the
SATA DPLL in Table 26-8. DPLL CLKDCOLDO Recommended Settings.
Use those settings in the driver. The TRM does not show
a value for 20MHz SYS_CLK so we use something close to the
26MHz setting.
Kever Yang [Thu, 5 Dec 2019 10:11:55 +0000 (18:11 +0800)]
rockchip: rk3288-evb: update config to support OPTEE
Upstream kernel and rockchip kernel has default enable PSCI which needs
OPTEE in trust word, enable OPTEE support for evb by default and SPL_FIT
option to pack OPTEE with U-Boot proper.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
And if enable the DEBUG for everyting in SPL, the log will hang at dwmmc
sending CMD16 for 'uboot' loadables binary because this step need a
large stack cost(about 0x2d00).
Kever Yang [Thu, 5 Dec 2019 10:11:52 +0000 (18:11 +0800)]
rockchip: fit_spl_optee: get text and optee base from build
Instead of hardcode the base address, we can get them from the build
output, eg. get the SYS_TEXT_BASE from .config and get optee base from
DRAM_BASE.
We can use this script for SoCs with DRAM base not from 0x60000000(rk3229
and many other 32bit Rockchip SoCs), eg. rk3288 DRAM base is 0.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
DWC3 support phy interfaces like 8/16-bit UTMI+. phy interface
initialization code would handle them properly along with UNKNOWN
type by default if none of the user/board doesn't need to use the
phy interfaces at all.
The current code is masking the 8/16-bit UTMI+ interface bits globally
which indeed effect the UNKNOWN cases, therefore it effects the platforms
which are not using phy interfaces at all.
So, handle the phy masking bits accordingly on respective interface
type cases.
Fixes: 6b7ebff00190 ("usb: dwc3: Add phy interface for dwc3_uboot") Reported-by: Andy Shevchenko <andy.shevchenko@gmail.com> Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
usb: cdns3: ep0: Fix build warnings related to cache ops
Since, commit 62f9b6544728 ("common: Move older CPU functions to their own header")
cache ops functions are declared in a separate header. Include the same
to avoid build warnings.
fw_setenv: avoid writing environment when nothing has changed
In the case where one deletes an already-non-existing variable, or sets
a variable to the value it already has, there is no point in writing the
environment back, thus reducing wear on the underlying storage
device.
In the case of redundant environments, if the two environments
differ (e.g. because one is corrupt), make sure that any call of
fw_setenv causes the two to become synchronized, even if the fw_setenv
call does not change anything in the good copy.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk> Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Alex Marginean [Thu, 19 Dec 2019 12:35:37 +0000 (14:35 +0200)]
drivers: net: bcm-sf2: pass -1 to phy_connect()
Passing 0 to PHY connect used to trigger a MDIO scan due to a bug fixed in
the meantime. It's unclear if bcm-sf2 wants to connect to PHY @ addr 0 or
is scanning the bus, passing -1 here should keep it functional either way.
Signed-off-by: Alex Marginean <alexandru.marginean@nxp.com> Cc: Jiandong Zheng <jdzheng@broadcom.com> Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Marek Vasut [Wed, 18 Dec 2019 06:48:50 +0000 (07:48 +0100)]
net: dwc_eth_qos: Pass -1 to phy_connect() to scan for all PHYs
PHY address 0 is a valid PHY address, to scan for all PHYs, pass -1 to
phy_connect(). Passing 0 used to work before be accident, but does no
longer.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Joe Hershberger <joe.hershberger@ni.com> Cc: Patrice Chotard <patrice.chotard@st.com> Cc: Patrick Delaunay <patrick.delaunay@st.com> Cc: Ramon Fried <rfried.dev@gmail.com> Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com> Acked-by: Joe Hershberger <joe.hershberger@ni.com>
struct sunxi_prcm_reg is naturally packed. There is no need to define it as
packed. Defining it as packed leads to compilation errors with GCC 9.2.1:
CC arch/arm/lib/reloc_arm_efi.o
arch/arm/cpu/armv7/sunxi/psci.c: In function ‘sunxi_cpu_set_power’:
:qarch/arm/cpu/armv7/sunxi/psci.c:163:21: error: taking address of packed
member of ‘struct sunxi_prcm_reg’ may result in an unaligned pointer value
[-Werror=address-of-packed-member]
163 | sunxi_power_switch(&prcm->cpu_pwr_clamp[cpu], &prcm->cpu_pwroff,
| ^~~~~~~~~~~~~~~~~~~~~~~~~
Remove __packed attribute from struct sunxi_prcm_reg.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Diego Rondini [Tue, 19 Nov 2019 13:43:33 +0000 (14:43 +0100)]
sun8i: h3: Support H3 variant of Orange Pi Zero Plus 2
Orangepi Zero Plus 2 is an open-source single-board computer, available
in two Allwinner SOC variants, H3 and H5. We add support for H3 variant
here, as the H5 is already supported.
H3 Orangepi Zero Plus 2 has:
- Quad-core Cortex-A7
- 512MB DDR3
- microSD slot and 8GB eMMC
- Debug TTL UART
- HDMI
- Wifi + BT
- OTG + power supply
arch/arm/cpu/armv7/sunxi/psci.c: In function ‘sunxi_cpu_set_power’:
arch/arm/cpu/armv7/sunxi/psci.c:144:21: error: taking address of packed
member of ‘struct sunxi_cpucfg_reg’ may result in an unaligned pointer
value [-Werror=address-of-packed-member]
144 | sunxi_power_switch(&cpucfg->cpu1_pwr_clamp, &cpucfg->cpu1_pwroff,
| ^~~~~~~~~~~~~~~~~~~~~~~
arch/arm/cpu/armv7/sunxi/psci.c:144:46: error: taking address of packed
member of ‘struct sunxi_cpucfg_reg’ may result in an unaligned pointer
value [-Werror=address-of-packed-member]
144 | sunxi_power_switch(&cpucfg->cpu1_pwr_clamp, &cpucfg->cpu1_pwroff,
| ^~~~~~~~~~~~~~~~~~~~
Use memcpy() and void* pointers to resolve the problem caused by packing
the struct sunxi_cpucfg_reg.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Acked-by: Jagan Teki <jagan@amarulasolutions.com>
Activate by default the command CLS (clear screen);
this command used in pxe or sysboot command (DISTRO support)
when the "menu background" keyword is present.
This patch avoid the warning "Unknown command 'cls'"
with extlinux.conf:
# Generic Distro Configuration file generated by OpenEmbedded
menu title Select the boot mode
MENU BACKGROUND /splash.bmp
TIMEOUT 20
DEFAULT stm32mp157c-ev1-emmc
LABEL stm32mp157c-ev1-emmc
KERNEL /uImage
FDT /stm32mp157c-ev1.dtb
APPEND root=/dev/mmcblk1p4 rootwait rw console=ttySTM0,115200
...
Retrieving file: /mmc0_stm32mp157c-ev1_extlinux/extlinux.conf
614 bytes read in 36 ms (16.6 KiB/s)
Retrieving file: /splash.bmp
46180 bytes read in 40 ms (1.1 MiB/s)
Unknown command 'cls' - try 'help'
Select the boot mode
1: stm32mp157c-ev1-sdcard
...
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Acked-by: Patrick Delaunay <patrick.delaunay@st.com>
With commit 37db55b7e9db ("linux/types.h: fix typo unchar") we have a
duplicate typedef for uchar. As linux/types.h is included in common.h we
don't need another typedef for uchar there.
Fixes: 37db55b7e9db ("linux/types.h: fix typo unchar") Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
Rasmus Villemoes [Thu, 12 Dec 2019 14:38:23 +0000 (14:38 +0000)]
doc: really get rid of Documentation/ directory
Commit 656d8da9d2 (doc: Remove duplicated documentation directory) got
rid of most of Documentation/. But there's still an obviously useless
.gitignore left behind.
Also, there's a copy of the linux kernel's net/ethernet.txt binding
imported from v5.0, while the existing one in doc/ is from 4.0-rc1. So
replace the latter by the former, and making Documentation/ finally
empty.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk> Reviewed-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Tom Rini [Wed, 11 Dec 2019 13:17:19 +0000 (08:17 -0500)]
Merge tag 'fix-for-2020.01' of https://gitlab.denx.de/u-boot/custodians/u-boot-i2c
i2c bugfixes for 2020.01
- i2c: i2c_cdns: fix write timeout on fifo boundary
fixes timout issue when writting number of bytes is multiple
of the FIFO depth.
Tom Rini [Wed, 11 Dec 2019 13:16:16 +0000 (08:16 -0500)]
Merge tag 'u-boot-atmel-fixes-2020.01-a' of https://gitlab.denx.de/u-boot/custodians/u-boot-atmel
- First set of u-boot-atmel fixes for 2020.01 cycle:
This set includes a small fix for gpio bank names, one for removing
unused headers (also touches some other boards), and a fix for the QSPI
env read on one of the boards.
The rpi_4_32b_defconfig states that only one DRAM bank is present. This
leads to a wrong configuration of the available DRAM. Fix this by
setting the DRAM bank config accordingly.
Up to now we only update the DRAM banks when we are define
CONFIG_BCM2711. But our one binary approach uses a config that supports
BCM2837 and BCM2711. As a result we only see one gibibyte of RAM on
Raspberry Pi 4, even if it has more RAM.
Fix this by calling dram_init_banksize.
Fixes: 5694090670 ("ARM: defconfig: add unified config for RPi3 and RPi4") Signed-off-by: Matthias Brugger <mbrugger@suse.com>
To update the dram bank information from device-tree we use
fdtdec_decode_ram_size() which expectes the the size-cells and
address-cells to be defined in the memory node. For normal system RAM
these values are defined in the root node. When the values differ from
the default values defined in the spec, we can end up with wrong RAM
bank information.
Switch to the "standard" way to update the RAM bank information to
avoid this.
Fixes: 9de5b89e4c ("rpi4: enable dram bank initialization") Signed-off-by: Matthias Brugger <mbrugger@suse.com>
Michael Auchter [Mon, 9 Dec 2019 18:16:16 +0000 (18:16 +0000)]
i2c: i2c_cdns: fix write timeout on fifo boundary
This fixes an issue that would cause I2C writes to timeout when the
number of bytes is a multiple of the FIFO depth (i.e. 16 bytes).
Within the transfer loop, after writing the data register with a new
byte to transfer, if the transfer size equals the FIFO depth, the loop
pauses until the INTERRUPT_COMP bit asserts to indicate data has been
sent. This same check is performed after the loop as well to ensure data
has been transferred prior to returning.
In the case where the amount of data to be written is a multiple of the
FIFO depth, the transfer loop would wait for the INTERRUPT_COMP bit to
assert after writing the final byte, and then wait for this bit to
assert once more. However, since the transfer has finished at this
point, no new data has been written to the data register, and hence
INTERRUPT_COMP will never assert.
Fix this by only waiting for INTERRUPT_COMP in the transfer loop if
there's still data to be written.
Signed-off-by: Michael Auchter <michael.auchter@ni.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Tom Rini [Tue, 10 Dec 2019 20:41:15 +0000 (15:41 -0500)]
Merge tag 'fixes-for-2020.01' of https://gitlab.denx.de/u-boot/custodians/u-boot-video
- fix crash and board reset when drawing RLE8 bitmaps
bigger than the framebuffer resolution
- reduce dead code in video and console uclass routines
(tested on mx53cx9020, sama5d2_xplained, stm32mp157c-ev1,
stm32f746-disco, stm32f769-disco and wandboard)
Swapna Gurumani [Mon, 2 Dec 2019 21:06:10 +0000 (21:06 +0000)]
configs: sama5d27_som1_ek: Add default config to read ENV from QSPI
In the initial SPI flash setup, the default bus mode being used was 3,
which is incorrect, causing a CRC error when the ENV was being read from
QSPI. Setting the default bus mode to 0 which is the correct mode.
Tom Rini [Tue, 10 Dec 2019 02:53:23 +0000 (21:53 -0500)]
Merge branch 'master' of https://gitlab.denx.de/u-boot/custodians/u-boot-riscv
- Increase stack size to avoid a stack overflow during distro boot.
- Add hifive-unleashed-a00.dts for SIFIVE FU540.
- Add OF_SEPARATE support for SIFIVE FU540.
- Add SPL support for Andes AX25 AE350.
- Improve U-Boot SPL / OpenSBI smp boot flow for RISC-V.
Lukas Auer [Sun, 8 Dec 2019 22:28:52 +0000 (23:28 +0100)]
spl: opensbi: wait for ack from secondary harts before entering OpenSBI
At the start, OpenSBI relocates itself to its link address. If the link
address ranges of U-Boot SPL and OpenSBI overlap, the relocation can
lead to code corruption if a hart is still running U-Boot SPL during
relocation. To avoid this problem, the main hart is specified as the
preferred boot hart to perform the relocation. This fixes the code
corruption problems based on the assumption that since the main hart
schedules the secondary harts to enter OpenSBI, it will be the last to
enter OpenSBI. However it was reported that this assumption is not
always correct.
To make sure the assumption always holds true, wait for all secondary
harts to acknowledge the call-function request before entering OpenSBI
on the main hart.
Reported-by: Rick Chen <rick@andestech.com> Signed-off-by: Lukas Auer <lukas.auer@aisec.fraunhofer.de> Reviewed-by: Rick Chen <rick@andestech.com> Tested-by: Rick Chen <rick@andestech.com> Reviewed-by: Anup Patel <anup.patel@wdc.com>
Lukas Auer [Sun, 8 Dec 2019 22:28:51 +0000 (23:28 +0100)]
riscv: add option to wait for ack from secondary harts in smp functions
Add a wait option to smp_call_function() to wait for the secondary harts
to acknowledge the call-function request. The request is considered to
be acknowledged once each secondary hart has cleared the corresponding
IPI.
As part of the call-function request, the secondary harts invalidate the
instruction cache after clearing the IPI. This adds a delay between
acknowledgment (clear IPI) and fulfillment (call function) of the
request. We want to use the acknowledgment to be able to judge when the
request has been completed. Remove the delay by clearing the IPI after
cache invalidation and just before calling the function from the
request.
Signed-off-by: Lukas Auer <lukas.auer@aisec.fraunhofer.de> Reviewed-by: Rick Chen <rick@andestech.com> Tested-by: Rick Chen <rick@andestech.com> Reviewed-by: Anup Patel <anup.patel@wdc.com>
Lukas Auer [Sun, 8 Dec 2019 22:28:50 +0000 (23:28 +0100)]
riscv: add functions for reading the IPI status
Add the function riscv_get_ipi() for reading the pending status of IPIs.
The supported controllers are Andes' Platform Level Interrupt Controller
(PLIC), the Supervisor Binary Interface (SBI), and SiFive's Core Local
Interruptor (CLINT).
Signed-off-by: Lukas Auer <lukas.auer@aisec.fraunhofer.de> Reviewed-by: Rick Chen <rick@andestech.com>
Lukas Auer [Sun, 8 Dec 2019 22:28:49 +0000 (23:28 +0100)]
spl: opensbi: specify main hart as preferred boot hart
OpenSBI uses a relocation lottery to determine the hart to relocate
OpenSBI to its link address. In the U-Boot SPL boot flow, the main hart
schedules the secondary harts to enter OpenSBI before doing so itself.
One of the secondary harts will therefore always be the winner of the
relocation lottery. This is problematic if the link address ranges of
OpenSBI and U-Boot SPL overlap. OpenSBI will be relocated and therefore
overwrite U-Boot SPL while some harts may still run it, leading to code
corruption.
Avoid this problem by specifying the main hart as the preferred boot
hart to perform the OpenSBI relocation. The main hart will be the last
hart to enter OpenSBI, relocation can therefore occur safely.
The boot hart field was added to version 2 of the OpenSBI FW_DYNAMIC
info structure. The header file include/opensbi.h is synchronized with
include/sbi/fw_dynamic.h from the OpenSBI project to update the info
structure. The header file is recent as of commit 7a13beb21326 ("firmware: Add preferred boot HART field in struct
fw_dynamic_info").
Reported-by: Rick Chen <rick@andestech.com> Suggested-by: Anup Patel <Anup.Patel@wdc.com> Signed-off-by: Lukas Auer <lukas.auer@aisec.fraunhofer.de> Reviewed-by: Rick Chen <rick@andestech.com> Tested-by: Rick Chen <rick@andestech.com> Reviewed-by: Anup Patel <anup.patel@wdc.com>
projects / u-boot.git / log
