| Age | Commit message (Collapse) | Author |
|---|
|
The domains are boot-time system-level partitions so we should
allow platform support to register domains only before hart
domain assignments are finalized.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Xiang W <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
|
|
The unsigned length may be 4 bytes or 8 bytes, amoadd.w only applies
to 4 bytes hence this patch.
Signed-off-by: Xiang W <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
Replace the test-and-set spinlock implementation with ticket locks
in order to get fairness (in form of FIFO order).
The implementation uses a 32-bit wide struct, which consists of
two 16-bit counters (owner and next). This is inspired by similar
spinlock implementations on other architectures.
This allows that the code works for both, RV32 and RV64.
Signed-off-by: Christoph Muellner <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
Reviewed-by: Xiang W <[email protected]>
|
|
The current implementation of SPIN_LOCK_INIT() provides the spinlock
to be initialized as reference. This does not allow a direct
initialization of the spinlock object at the creation site.
Let's pass the spinlock directly instead (like Linux does as well)
and adjust all users of the macro (in fact there is only one user).
Signed-off-by: Christoph Muellner <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
csr_read_allowed/csr_read_allowed requires trap.case to detect the results,
but if no exception occurs, the value of trap.case will remain unchanged,
which makes the detection results unreliable. Add code to initialize
trap.case to 0.
Signed-off-by: Xiang W <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
There are two copies of the same abnormal exit code, this patch deletes one
Signed-off-by: Xiang W <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
Current fdt_plic_fixup() only does necessary fix-up against the legacy
"riscv,plic0" node. The upstream Linux kernel defines its official DT
bindings which uses "sifive,plic-1.0.0" as the compatible string and
we should check that first, and if not present fall back to legacy.
Signed-off-by: Bin Meng <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
At present fdt_plic_fixup() accepts a 'compat' parameter for PLIC
compatible string. In preparation to support the new DT bindings,
drop this and use "riscv,plic0" directly in fdt_plic_fixup().
Signed-off-by: Bin Meng <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
Linux kernel commit a2770b57d083 ("dt-bindings: timer: Add CLINT bindings")
adds the official DT bindings for CLINT, which uses "sifive,clint0"
as the compatible string. "riscv,clint0" is now legacy and has to
be kept for backward compatibility of legacy systems.
Signed-off-by: Bin Meng <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
GCC has already a predefined macro __ASSEMBLER__ therefore, it can be
used without the need to define a new flag with -D__ASSEMBLY__.
This is useful when adding the library to projects having a build
system such one can build without the need to make changes.
THe build system does not use the Makefile in the sources tree.
Signed-off-by: Marouene Boubakri <[email protected]>
Signed-off-by: Anup Patel <[email protected]>
|
|
The trap handler sbi_emulate_csr_read() invokes hpm_allowed() which reads
CSR 0x306 (mcounteren). The K210 does not support CSR 0x306. While trying
to handle a trap occurring in S-mode code this creates an additional trap
in M-mode. This results in failure to redirect to S-mode and the system
hanging in sbi_hart_hang().
In hart_detect_features() we have already determined if CSR 0x306 is
available and stored that information in the scratch area. We can use this
information to decide if CSR 0x306 shall be accessed in hpm_allowed() and
thus avoid the M-mode trap.
Likewise if CSR scounteren is not available we have to avoid reading CSR
0x106.
Signed-off-by: Heinrich Schuchardt <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
The sbi_platform_ipi_clear() called from wait_for_coldboot() and
sbi_hsm_hart_wait() is redundant because IPI will be automatically
cleared by sbi_platform_ipi_init() called from sbi_ipi_init().
Further, wait_for_coldboot() is common for warm startup and warm
resume path so the sbi_platform_ipi_clear() called in warm resume
path cause resuming HART to miss an IPI injected other HART to
wakeup the HART.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
This patch implements the SBI HSM suspend function. Using this
new SBI call, the S-mode software can put calling HART in platform
specific suspend (i.e. low-power) state. For a successful retentive
suspend, the SBI call will return without errors upon resuming
whereas for a successful non-retentive suspend, the SBI call will
resume from a user provided resume address.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We add sbi_hart_reinit() function which will re-initialize HART CSRs
assuming HART features are already detected. This new function will
be useful in re-initializing HART after it resumes from HSM SUSPENDED
state.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
The sbi_hsm_hart_start() and sbi_hsm_hart_stop() functions should
only return error codes as defined by the SBI specification.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
The sbi_hsm_hart_started() function is only used by sbi_hsm_hart_stop()
for checking state of calling HART and current domain assignment.
The atomic_cmpxchg() called by sbi_hsm_hart_stop() will check state of
calling hart anyway and domain assignment can be checked by other domain
function such as sbi_domain_is_assigned_hart().
This means sbi_hsm_hart_started() is redundant and can be removed.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
A hart can take interrupt in the new HSM states introduced by the
SBI HSM suspend function (such as SUSPENDED state) so we rename
sbi_hsm_hart_started_mask() to something more generic such as
sbi_hsm_hart_interruptible_mask().
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We replace the use of SBI_STATE_xyz defines with SBI_HSM_STATE_xyz
defines because the HSM state defines are complete enough to implement
HSM state machine in OpenSBI. As a result of this, we can now remove
sbi_hsm_hart_state_to_status() function because it is now redundant
and sbi_hsm_hart_get_state() can directly return HSM state or error.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We simplify HSM state define names so that these defines can directly
replace SBI_HART_xyz defines used by SBI HSM implementation.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
Currently, the root domain memory regions are fixed in generic
code but some of the platforms may want to explicitly define
memory regions for the root domain.
This patch adds optional domains_root_regions() platform callback
which platforms can use to provide platform specific root domain
memory regions. Due to this changes, the root domain should also
undergo all sanity checks (just like regular domain) so we use
sbi_domain_register() to register root domain.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
Fix uart_putc implementation.
Due to a bug in the IP, this went unnoticed.
Use macros instead of magic numbers to make the code
more readable.
Signed-off-by: Vijai Kumar K <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
Currently, we have fixed TLB flush types supported by the
remote TLB library. This approach is not flexible and does
not allow custom local TLB flush function. For example,
after updating PMP entries on a set of HARTs at runtime,
we have to flush TLB on these HARTs as well.
To support custom local TLB flush function, we replace the
"type" field of "struct sbi_tlb_info" with a local TLB flush
function pointer. We also provide definitions of standard TLB
flush operations (such as fence_i, sfence.vma, hfence.vvma,
hfence.gvma, etc).
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We introduce sbi_trap_exit() API which can help non-firmware
(i.e. generic or platform) code to force exit trap/interrupt
handling and resume execution at context pointed by parameter
"const struct sbi_trap_regs *regs".
This new sbi_trap_exit() API will help Keystone Enclave project
to resume execution of enclave from custom SBI call handler.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We had added args pointer in ecall handler to ensure that ecall
handler only implements functionality and does not deal with
SBI calling convention. This also helped us to keep SBI calling
convention related code in one place at sbi_ecall_handler().
The Keystone Enclavce project needs access to the trap regsiters
in their ecall handler so that they can context switch enclaves
in custom SBI calls. To help the Keystone Enclave project, we
replace the args pointer in ecall handler parameter with a const
pointer to trap registers.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
The misaligned load emulation does not sign-extend values correctly
due to missing sign typecast in value passed to the SET_RD() macro.
A very easy way to reproduce this issue is to load 16-bit value
0xff1e from a byte aligned address using LH instruction on hardware
lacking misaligned load/store.
This patch fixes sbi_misaligned_load_handler() for above issue.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
The domain_get() platform callback function is now redundant
because fdt_domain_populate() register new domain explicitly
using the sbi_domain_register() function.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
The fdt_domain_get() function is now redundant because the
fdt_domains_populate() function can explicitly register new
domains using the sbi_domain_register() function.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We add sbi_domain_register() function to help platform code register
a new domain explicitly instead of depending on domain_get() platform
callback. This function will eventually help us completely remove the
domain_get() platform callback.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We extend fdt_iterate_each_domain() and fdt_iterate_each_memregion()
functions to allow underlying iteration function to fail. This will
help us catch more domain misconfiguration issues at boot time.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
Sync with libfdt v1.6.0 release source codes.
Signed-off-by: Dimitri John Ledkov <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
We add various helper routines to populate domains, iterate domains,
iterate domain memregions, and parse HART to domain assignment from
the FDT.
These helper routines can be used by platform support code and FDT
fixup code.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We introduce domains_init() platform operation which can be used by
platform support to initialize/populate domains in the coldboot path.
The domains_init() is called late in the coldboot sequence from the
sbi_domain_finalize() so sbi_printf() can be used by platform support
to print errors/warnings at time of populating domains.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We add error prints in sbi_domain_finalize() and sanitize_domain()
to help debug domain configuration issues.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
If coldboot HART is assigned to a domain then we should use coldboot
HART as the boot HART of the domain. This allows coldboot HART to
continue boot sequence even when it is assigned to some non-root domain.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
The sbi_domain_finalize() should auto start a domain only if the
boot HART is withing limits (i.e. less than SBI_HARTMASK_MAX_BITS).
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
The domain_count check in sbi_domain_finalize() for newly discovered
domain is incorrect and should be reversed.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
There has been typo in automatic domain boot hart startup
for non-root domains so this patch fixes it.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We should check and access N-extension CSRs in sbi_hart_switch_mode()
when next_mode is u-mode because N-extension is optional.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
The SBI SRST extension has been accepted and merged in the latest
SBI v0.3-draft specification.
(Refer, https://github.com/riscv/riscv-sbi-doc)
It allows to S-mode software to request system shutdown, cold reboot,
and warm reboot.
This patch implements SBI SRST extension as a replacement of the
legacy sbi_shutdown() call of SBI v0.1 specification.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
To implement the SBI SRST extension, we need two platform operations
for system reset:
1) system_reset_check() - This operation will check whether given
reset type and reason are supported by the platform
2) system_reset() - This operation will do the actual platform
system reset and it will not return if reset type and reason
are supported by the platform
This patch updates system reset related code everywhere as-per above.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We can now use the standard SBI SRST extension reset types instead
of the opensbi specific (SBI_PLATFORM_RESET_xyz) reset types hence
remove related opensbi specific defines. The "platform_" prefix of
the reset type parameter of sbi_system_reset() function should also
be removed.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Alistair Francis <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
Currently, we have all boot prints at the end of cold boot sequence
which means if there is any failure in cold boot sequence before
boot prints then we don't get any print.
This patch improves boot prints in cold boot sequence as follows:
1. We divide the boot prints into multiple parts and print it
from different locations after sbi_console_init()
2. We throw an error print if there is any failure in cold boot
sequence after sbi_console_init()
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
We know about pmp granularity and number of bits supported by PMP.
Show those information in the boot time info print
Signed-off-by: Atish Patra <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
As per RISC-V privilege specification, a platform may choose to implement
a coarser granularity scheme for PMP addresses. In that case, we shouldn't
allow any pmp region size smaller than the platform supports. A platform
may not also implement all the bits for a PMP address specified in the priv
specification.
The pmp range granularity and address bits should be detected dynamically
before detecing PMP regions. Any pmp modification request beyond these detected
value must not succeed.
Signed-off-by: Atish Patra <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
The "reg" property in a device node may not be the correct address always.
If a parent node defines a "ranges" property, the child address need to be
translated with respect to parents address. If the ranges property is not
present, it will just use 1:1 translation.
Signed-off-by: Atish Patra <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
|
|
We extend boot prints to display details of each domain. In the
process, we remove sbi_hart_pmp_dump() because it shows redundant
information which domain details already show.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
The PMP configuration on each HART should be only based on the memory
regions of the assigned domain because each domain includes a memory
region to protect the underlying firmware.
This patch updates the sbi_hart_pmp_configure() function accordingly.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
The sbi_hart_pmp_get() and sbi_hart_pmp_check_addr() functions
are not used anymore hence this patch removes these functions.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
The fdt_cpu_fixup() should disable a HART in DT if the HART
is not assigned to the current HART domain. This patch updates
fdt_cpu_fixup() accordingly.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
|
Now that each HART is mapped to a domain having a set of memory
regions, we update fdt_reserved_memory_fixup() to use domain memory
regions for adding reserved memory nodes in device tree.
We also change reserved memory node name prefix from "mmode_pmp"
to "mmode_resv" because domain memory regions can impact other
hardware configurations (such as IOPMP, etc) along with PMP.
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
|
