This adds comments regarding the ordering and purpose of certain
instructions as I understand them.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Reviewed-by: Leo Liang <ycliang@andestech.com>
This ensures constructs like `if (gd & gd->...) { ... }` work when
accessing the global data pointer. Without this change, it was possible for
a very early trap to cause _exit_trap to directly or indirectly (through
printf) to read arbitrary memory. This could cause a second trap,
preventing show_regs from being printed.
printf (and specifically puts) uses gd to determine what function to print
with. These functions in turn use gd to find the serial device, etc.
However, before accessing gd, puts first checks to see if it is non-NULL.
This indicates an existing (perhaps undocumented) assumption that either gd
is NULL or it is completely valid.
Before this patch, gd either points to unexpected data (because it retains
the value it did from the prior-stage) or points to uninitialized data
(because it has not yet been initialized by board_init_f_init_reserve)
until the hart has acquired available_harts_lock. This can cause two
problems, depending on the value of gd->flags. If GD_FLG_SERIAL_READY is
unset, then some garbage data will be printed to stdout, but there will not
be a second trap. However, if GD_FLG_SERIAL_READY is set, then puts will
try to print with serial_puts, which will likely cause a second trap.
After this patch, gd is zero up until either a hart has set it in
wait_for_gd_init, or until it is set by arch_init_gd. This prevents its
usage before its data is initialized because both handle_trap and puts
ensure that gd is nonzero before using it. After gd has been set, it is OK
to access it because its data has been cleared (and so flags is valid).
XIP cannot use locks because flash is not writable. This leaves it
vulnerable to the same class of bugs regarding already-pending IPIs as
before this series. Fixing that would require finding another method of
synchronization, which is outside the scope of this series.
Fixes: 7c6ca03eae ("riscv: additional crash information")
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
We can reduce the number of instructions needed to use available_harts_lock
by using the aq and rl suffixes for AMOs.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Even though we no longer call smp_function if an IPI was not sent by
U-Boot, we still need to clear any IPIs which were pending from the
execution environment. Otherwise, secondary harts will busy-wait in
secondary_hart_loop, instead of relaxing.
Along with the previous commit ("riscv: Use a valid bit to ignore
already-pending IPIs"), this fixes SMP booting on the Kendryte K210.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Some IPIs may already be pending when U-Boot is started. This could be a
problem if a secondary hart tries to handle an IPI before the boot hart has
initialized the IPI device.
To be specific, the Kendryte K210 ROM-based bootloader does not clear IPIs
before passing control to U-Boot. Without this patch, the secondary hart
jumps to address 0x0 as soon as it enters secondary_hart_loop, and then
hangs in its trap handler.
This commit introduces a valid bit so secondary harts know when and IPI
originates from U-Boot, and it is safe to use the IPI API. The valid bit is
initialized to 0 by board_init_f_init_reserve. Before this, secondary harts
wait in wait_for_gd_init.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Reviewed-by: Leo Liang <ycliang@andestech.com>
Without a matching barrier on the write side, the barrier in handle_ipi
does nothing. It was entirely possible for the boot hart to write to addr,
arg0, and arg1 *after* sending the IPI, because there was no barrier on the
sending side.
Fixes: 90ae281437 ("riscv: add option to wait for ack from secondary harts in smp functions")
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Reviewed-by: Leo Liang <ycliang@andestech.com>
Clearing MIP.MSIP is not guaranteed to do anything by the spec. In
addition, most existing RISC-V hardware does nothing when this bit is set.
The following commits "riscv: Use a valid bit to ignore already-pending
IPIs" and "riscv: Clear pending IPIs on initialization" should implement
the original intent of the reverted commit in a more robust manner.
This reverts commit 9472630337.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
We currently do this in a u-boot specific dts, but hopefully we can get
these bindings added in Linux in the future.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@openfive.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
The interrupt controller property is removed from the clint binding because
the clint is not an interrupt-controller. That is, no other devices have an
interrupt which is controlled by the clint.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
This converts the clint driver from the riscv-specific interface to be a
DM-based UCLASS_TIMER driver. In addition, the SiFive DDR driver previously
implicitly depended on the CLINT to select REGMAP.
Unlike Andes's PLMT/PLIC (which AFAIK never have anything pass it a dtb),
the SiFive CLINT is part of the device tree passed in by qemu. This device
tree doesn't have a clocks or clock-frequency property on clint, so we need
to fall back on the timebase-frequency property. Perhaps in the future we
can get a clock-frequency property added to the qemu dtb.
Unlike with the Andes PLMT, the Sifive CLINT is also an IPI controller.
RISCV_SYSCON_CLINT is retained for this purpose.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@openfive.com>
This merges the PLIC initialization code from two functions into one.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
This converts the PLMT driver from the riscv-specific timer interface to be
a DM-based UCLASS_TIMER driver.
The clock-frequency/clocks properties are preferred over timebase-frequency
for two reasons. First, properties which affect a device should be located
near its binding in the device tree. Using timebase-frequency only really
makes sense when the cpu itself is the timer device. This is the case when
we read the time from a CSR, but not when there is a separate device.
Second, it lets the device use the clock subsystem which adds flexibility.
If the device is configured for a different clock speed, the timer can
adjust itself.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Rick Chen <rick@andestech.com>
The riscv-timer driver currently serves as a shim for several riscv timer
drivers. This is not too desirable because it bypasses the usual timer
selection via the driver model. There is no easy way to specify an
alternate timing driver, or have the tick rate depend on the cpu's
configured frequency. The timer drivers also do not have device structs,
and so have to rely on storing parameters in gd_t. Lastly, there is no
initialization call, so driver init is done in the same function which
reads the time. This can result in confusing error messages. To a user, it
looks like the driver failed when trying to read the time, whereas it may
have failed while initializing.
This patch removes the shim functionality from the riscv-timer driver, and
has it instead implement the former rdtime.c timer driver. This is because
existing u-boot users who pass in a device tree (e.g. qemu) do not create a
timer device for S-mode u-boot. The existing behavior of creating the
riscv-timer device in the riscv cpu driver must be kept. The actual reading
of the CSRs has been redone in the style of Linux's get_cycles64.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
The gp register is used to store U-Boot's global data pointer. We should
not assume that an UEFI application leaves the gp register unchanged as
the UEFI specifications does not define who is the owner of the gp and tp
registers.
So the following sequence should be followed in the trap handler:
* save the caller's gp register
* restore the global data pointer
* serve interrupts or print crash dump and reset
* restore the caller's gp register
Cc: Abner Chang <abner.chang@hpe.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Add boolean input argument @no_map to helper function
fdtdec_add_reserved_memory() to add or not "no-map" property
for an added reserved memory node.
Property no-map is used by the Linux kernel to not not map memory
in its static memory mapping. It is needed for example for the|
consistency of system non-cached memory and to prevent speculative
accesses to some firewalled memory.
No functional change. A later change will update to OPTEE library to
add no-map property to OP-TEE reserved memory nodes.
Signed-off-by: Etienne Carriere <etienne.carriere@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Function set_gd() is needed in the UEFI sub-system if the global data
pointer is stored in a register.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Provide a command to display information about the SBI implementation.
The output might look like:
=> sbi
SBI 0.2
OpenSBI
Extensions:
sbi_set_timer
sbi_console_putchar
sbi_console_getchar
sbi_clear_ipi
sbi_send_ipi
sbi_remote_fence_i
sbi_remote_sfence_vma
sbi_remote_sfence_vma_asid
sbi_shutdown
SBI Base Functionality
Timer Extension
IPI Extension
RFENCE Extension
Hart State Management Extension
The command can be used to construct a unit test checking that the
communication with the SEE is working.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Atish Patra <atish.patra@wdc.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Tested-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@openfive.com>
Tested-by: Pragnesh Patel <pragnesh.patel@openfive.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Tested-by: Rick Chen <rick@andestech.com>
Building with CONFIG_SPL_SMP=n results in:
arch/riscv/lib/spl.c: In function ‘jump_to_image_no_args’:
arch/riscv/lib/spl.c:33:6:
error: unused variable ‘ret’ [-Werror=unused-variable]
33 | int ret;
| ^~~
Define the variable ret as __maybe_unused.
Fixes: 191636e448 ("riscv: Introduce SPL_SMP Kconfig option for U-Boot SPL")
Fixes: 8c59f2023c ("riscv: add SPL support")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Rick Chen <rick@andestech.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@openfive.com>
At present fdtdec_get_addr() is used to get L2 cache controller base
address. This only works for a fixed #address-cells and #size-cells.
Change to use fdtdec_get_addr_size_auto_parent() instead.
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@openfive.com>
If an exception occurs, the relocated program counter and return address
are required for an analysis.
With this patch you get:
=> exception undefined
Unhandled exception: Illegal instruction
EPC: 0000000080595908 RA: 000000008059c0c6 TVAL: 000000008030c01e
EPC: 0000000080007908 RA: 000000008000e0c6 reloc adjusted
We can use the relocated addresses to find the involved functions in
u.boot.map:
.text.do_undefined
0x0000000080007908 0x8 cmd/built-in.o
.text.cmd_process
0x000000008000dfcc 0x11a common/built-in.o
0x000000008000dfcc cmd_process
If an exception occurs in an UEFI binary additionally the load addresses of
the UEFI binaries are needed. With this patch:
=> setenv efi_selftest exception
=> bootefi selftest
Unhandled exception: Illegal instruction
EPC: 000000008042e18a RA: 000000008042e18a TVAL: 000000008030c01e
EPC: 000000007fea018a RA: 000000007fea018a reloc adjusted
UEFI image [0x0000000000000000:0xffffffffffffffff] '/\selftest'
UEFI image [0x000000008042e000:0x000000008042e43f] pc=0x18a '/bug.efi'
The value pc=0x18a matches the position of the illegal instruction in
efi_selftest_miniapp_exception.efi (loaded as /bug.efi);
asm volatile (".word 0xffffffff\n");
00000180 93 85 C5 11 1C 64 22 85 82 97 FF FF FF FF 1C 64
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
Tested-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
We should not initialize a variable if the value is overwritten before
being read.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Tested-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Reviewed-by: Rick Chen <rick@andestech.com>
After
if (ret) return ret;
we know that ret is zero. Don't check it again.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
All FU540 driver related options should be in the SoC level Kconfig.
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Tested-by: Pragnesh Patel <pragnesh.patel@sifive.com>
spl_soc_init() seems to be a better name, as all SPL functions
names start from the spl_ prefix.
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Tested-by: Pragnesh Patel <pragnesh.patel@sifive.com>
The generic SPL version of board_init_f() should give a call to
board specific codes to initialize board in the SPL phase.
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Tested-by: Pragnesh Patel <pragnesh.patel@sifive.com>
PRCI module within SiFive SoC's has register with which we can
reset the sub-systems within the SoC. The resets to DDR and ethernet
sub systems within FU540-C000 SoC are active low, and are hold low
by default on power-up. Currently these are directly asserted within
prci driver via register read/write.
With the DM based reset driver support here, we bind the reset
driver with clock (prci) driver and assert the reset signals of
both sub-system's appropriately.
Signed-off-by: Sagar Shrikant Kadam <sagar.kadam@sifive.com>
Reviewed-by: Pragnesh Patel <Pragnesh.patel@sifive.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Tested-by: Bin Meng <bin.meng@windriver.com>
The resets to DDR and ethernet sub-system are connected to
PRCI device reset control register, these reset signals
are active low and are held low at power-up. Add these reset
producer and consumer details needed by the reset driver.
Signed-off-by: Sagar Shrikant Kadam <sagar.kadam@sifive.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Make memory node available to SPL in prepration to updates to SiFive
DDR RAM driver to read memory information from DT.
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Switch off SMP support when building u-boot-spl would cause linking error as follow:
undefined reference to 'secondary hart relocate' and 'smp_call_function'.
Add macro to wrap up proper code region that needs SMP configuration on.
Signed-off by: Leo Liang <ycliang@andestech.com>
Cc: rick@andestech.com
Reviewed-by: Bin Meng <bin.meng@windriver.com>
This reverts commit 958a3f464c.
A more appropriate change below is already in mainline.
Commit fd31e4fd18 ("riscv: Do not build reset.c if SYSRESET is on")
Revert this patch, so that U-Boot can be built successfully for
SiFive Fu540 board.
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Leo Liang <ycliang@andestech.com>
Add booting from SPI for SiFive Unleashed board.
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Tested-by: Bin Meng <bin.meng@windriver.com>
Commit 40686c394e ("riscv: Clean up IPI initialization code")
caused U-Boot failed to boot on SiFive HiFive Unleashed board.
The codes inside arch_cpu_init_dm() may call U-Boot timer APIs
before the call to riscv_init_ipi(). At that time the timer register
base (e.g.: the SiFive CLINT device in this case) is unknown yet.
It might be the name riscv_init_ipi() that misleads people to only
consider it is related to IPI, but in fact the timer capability is
provided by the same SiFive CLINT device that provides the IPI.
Timer capability is needed for both UP and SMP.
Considering that the original refactor does have benefits, that it
makes the IPI code more similar to U-Boot initialization idioms.
It also removes some quite ugly macros. Let's do the minimal revert
instead of a complete revert, plus a fixes to arch_cpu_init_dm() to
consider the SPL case.
Fixes: 40686c394e ("riscv: Clean up IPI initialization code")
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
Tested-by: Leo Liang <ycliang@andestech.com>
Replace printf() and debug() by log_err() and log_debug().
"No reserved memory region found in source FDT\n" is not an error but a
debug information.
%s/can not/cannot/ - use the more common spelling.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Atish Patra <atish.patra@wdc.com>
Add L2 cache node to enable all cache ways from U-Boot proper.
Signed-off-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
fdtdec_get_addr_size_no_parent is not an optimized version if parent
node is already available with the caller.
Use fdtdec_get_addr_size_auto_parent to read the "reg" property
Signed-off-by: Atish Patra <atish.patra@wdc.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Not all errors are fatal. If a reserved memory node already exists in the
destination device tree, we can continue to boot without failing.
Signed-off-by: Atish Patra <atish.patra@wdc.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Starting from OpenSBI v0.7, the SBI firmware inserts/fixes up the
reserved memory node for PMP protected memory regions. All RISC-V
boards need to copy the reserved memory node from the device tree
provided by the firmware to the device tree used by U-Boot.
Turn on CONFIG_OF_BOARD_FIXUP by default for OF_SEPARATE.
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Atish Patra <atish.patra@wdc.com>
Reviewed-by: Rick Chen <rick@andestech.com>
The FDT blob might not have sufficient space to hold a copy of
reserved memory node. Expand it before the copy.
Reported-by: Rick Chen <rick@andestech.com>
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Atish Patra <atish.patra@wdc.com>
Reviewed-by: Rick Chen <rick@andestech.com>
The copy of reserved memory node from source dtb to destination dtb
can be avoided if they point to the same place. This is useful when
OF_PRIOR_STAGE is used.
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Reviewed-by: Atish Patra <atish.patra@wdc.com>
Add cpu aliases to U-Boot specific dtsi for hifive-unleashed.
Without aliases we see that the CPU device sequence numbers are set
randomly and the cpu list/detail command will show it as follows:
=> cpu list
1: cpu@1 rv64imafdc
2: cpu@2 rv64imafdc
3: cpu@3 rv64imafdc
0: cpu@4 rv64imafdc
Seems like CPU probing with dm-model also relies on aliases as observed
in case spi. The fu540-c000-u-boot.dtsi has cpu nodes and so adding
corresponding aliases we can ensure that cpu devices are assigned
proper sequence as follows:
=> cpu list
1: cpu@1 rv64imafdc
2: cpu@2 rv64imafdc
3: cpu@3 rv64imafdc
4: cpu@4 rv64imafdc
Signed-off-by: Sagar Shrikant Kadam <sagar.kadam@sifive.com>
Reviewed-by: Pragnesh Patel <pragnesh.patel@sifive.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
The Sipeed Maix series is a collection of boards built around the RISC-V
Kendryte K210 processor. This processor contains several peripherals to
accelerate neural network processing and other "ai" tasks. This includes a
"KPU" neural network processor, an audio processor supporting beamforming
reception, and a digital video port supporting capture and output at VGA
resolution. Other peripherals include 8M of sram (accessible with and
without caching); remappable pins, including 40 GPIOs; AES, FFT, and SHA256
accelerators; a DMA controller; and I2C, I2S, and SPI controllers. Maix
peripherals vary, but include spi flash; on-board usb-serial bridges; ports
for cameras, displays, and sd cards; and ESP32 chips. Currently, only the
Sipeed Maix Bit V2.0 (bitm) is supported, but the boards are fairly
similar.
Documentation for Maix boards is located at
<http://dl.sipeed.com/MAIX/HDK/>. Documentation for the Kendryte K210 is
located at <https://kendryte.com/downloads/>. However, hardware details are
rather lacking, so most technical reference has been taken from the
standalone sdk located at
<https://github.com/kendryte/kendryte-standalone-sdk>.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Where possible, I have tried to find compatible drivers based on the layout
of registers. However, many devices remain untested. All untested devices
have been left disabled, but some tentative properties (such as compatible
strings, and clocks, interrupts, and resets properties) have been added.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Currently, one cannot use a reset driver on RISC-V. Follow the MIPS
example, and disable the default reset handler when the sysreset driver is
enabled.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Some older processors (notably the Kendryte K210) use an older version of
the RISC-V privileged specification. The primary changes between the old
and new are in virtual memory, and in the merging of three separate counter
enable CSRs. Using the new CSR on an old processor causes an illegal
instruction exception. This patch adds an option to use the old CSRs
instead of the new one.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>