Age | Commit message (Collapse) | Author | Files | Lines |
|
If the ecall SBI_EXT_HSM_HART_START is called it might try to wake the
secondary hart using sbi_ipi_raw_send() to send an IPI to the hart.
This can fail if there is no IPI device but no error is returned from
sbi_ipi_raw_send() so the ecall returns as if the action completed and
the caller continues without noticing (in the case of Linux it just hangs
waiting for the secondary hart to become active)
Fix this by changing sbi_ipi_raw_send() to return and error, and if an
error is returned, then return it via SBI_EXT_HSM_HART_START call.
Signed-off-by: Ben Dooks <ben.dooks@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
Motivation: Suppose a peripheral needs to be configured to transmit
data. There is an SFR bit which indicates that the peripheral is ready
to transmit. The firmware should check the bit and will only transmit
data when the peripheral is ready. When the firmware starts polling the
SFR, the peripheral could be busy transmitting/receiving other data so
the firmware must wait till that completes. Assuming that there is no
other way, the firmware shouldn't wait indefinitely.
The function sbi_timer_waitms_until() will constantly check whether a
certain condition is satisfied, or timeout occurs. It can be used for
the cases when a timeout is required.
Signed-off-by: Adnan Rahman Chowdhury <adnan.chowdhury@sifive.com>
Reviewed-by: Xiang W <wxjstz@126.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
This patch updates OpenSBI version to 1.1 as part of
release preparation.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
|
|
OpenSBI explicitly assumes that there is no pmu hardware counter with
index 1: hardware uses that bit for TM control. So OpenSBI filters
out that index in sanity checks. However OpenSBI also excludes that
counter when reports total amount of hardware counters to Linux. As
a result, Linux uses incomplete counters mask excluding the highest
available counter.
Return accurate number of counters, update the firmware counter
starting index, fix range checks that include num_hw_ctrs.
The simple test is to make sure that there is no counter multiplexing
in the following command:
$ perf stat -e \
r8000000000000000,r8000000000000001,r8000000000000002,r8000000000000003, \
r8000000000000004,r8000000000000005,r8000000000000006,r8000000000000007, \
r8000000000000008,r8000000000000009,r800000000000000a,r800000000000000b, \
r800000000000000c,r800000000000000d,r800000000000000e,r800000000000000f \
ls
Note that 16 firmware events with 16 counters won't require multiplexing.
Signed-off-by: Sergey Matyukevich <geomatsi@gmail.com>
Signed-off-by: Atish Patra <atishp@rivosinc.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Now that SBI v1.0 specification is ratified, we change spec verion
implemented by OpenSBI to v1.0.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
The allwinner,sun20i-d1-clint compatible string is not documented in any
official binding, so it should not be used by drivers.
The MSWI in the D1 CLINT is compatible with the ACLINT specification, so
it can take advantage of generic driver support. However, that is only
possible if the MSWI and MTIMER are split into separate DT nodes. This
means the final binding for this device is likely to be incompatible
with what is implemented here.
Remove this compatible string from the driver to prevent it from
appearing in a stable version and causing future issues.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
The second invocation of plic_context_init() incorrectly calls the
function with m_cntx_id instead of s_cntx_id. This breaks systems which
only have 1 external interrupt per hart.
Fixes: 8c362e7 ("lib: irqchip/plic: Factor out a context init function")
Signed-off-by: Jan Remes <jan.remes@codasip.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Reviewed-by: Samuel Holland <samuel@sholland.org>
|
|
The CLINT in the Allwinner D1 SoC apparently does not support 64-bit
MMIO access. A property was added to support this quirk (and that
property was copied to the ACLINT MTIMER code). However, since this
difference in behavior makes the D1 CLINT incompatible with the SiFive
CLINT's programming interface, a better solution is to use a separate
compatible string.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
Based on architecture review committee feedback, the [m|s|vs]seteienum,
[m|s|vs]clreienum, [m|s|vs]seteipnum, and [m|s|vs]clreipnum CSRs are
removed in the latest AIA draft v0.3.0 specification.
(Refer, https://github.com/riscv/riscv-aia/releases/tag/0.3.0-draft.31)
These CSRs were mostly for software convenience and software can always
use [m|s|vs]iselect and [m|s|vs]ireg CSRs to update the IMSIC interrupt
file bits.
We update the IMSIC programming as-per above to match the latest AIA
draft specification.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
We should not change trap->tval to mepc because mtval already points to
the faulting portion of the emulated instruction fetch, which is also
what stval is expected to be.
In addition, htinst is only allowed to be zero for instruction access
faults or page faults, and is only allowed to be zero or a
psuedoinstruction for instruction guest-page faults. Fix trap->tinst for
these cases.
Signed-off-by: dramforever <dramforever@live.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
If there is an exception while emulating a misaligned load/store, fixup
uptrap.tinst before redirecting. Otherwise, HS-mode software may receive
an htinst describing the lbu/sb instruction that faulted during
emulation[1].
[1]: https://github.com/riscv-software-src/opensbi/issues/258
Signed-off-by: dramforever <dramforever@live.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
Add psuedoinstruction encodings written to mtinst/htinst for faults
caused by implicit memory access for VS-stage address translation
Signed-off-by: dramforever <dramforever@live.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
Allwinner D1 contains a "PPU" power domain controller which can
automatically power down/up the CPU power domain. This power domain
includes the C906 core along with its CLINT and PLIC.
This HSM implementation supports non-retentive hart suspend by:
1) Saving/restoring state that is lost during hart suspend,
2) Performing cache maintenance before/after hart suspend,
3) Configuring wakeup sources before hart suspend, and
4) Asking the PPU to power down the hart when it enters WFI.
Since this HSM implementation is for a single-core SoC, it does not need
to worry about concurrency or saving multiple instances of state.
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Samuel Holland <samuel@sholland.org>
|
|
The delegation bit is lost along with the rest of the PLIC state when
the CPU power domain in the Allwinner D1 is powered down, so the PLIC
needs to be re-delegated to S-mode during the hart resume path.
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Samuel Holland <samuel@sholland.org>
|
|
These functions save/restore the state of the PLIC associated with the
current hart. The context save/restore functions only manipulate a
single context, since most likely the M-mode context is unused and does
not need to be saved.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
These can be used by platform code to save the PLIC priority state, if
it would otherwise be lost during non-retentive suspend. The platform
is responsible for allocating all necessary storage.
As a space optimization, store the saved priority values as 8-bit
integers, since that is large enough to hold any priority value on the
relevant platforms.
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Samuel Holland <samuel@sholland.org>
|
|
These can be used by platform code to save the PLIC context state, if
it would otherwise be lost during non-retentive suspend. The platform
is responsible for allocating all necessary storage.
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Samuel Holland <samuel@sholland.org>
|
|
This simplifies both the callers and the callees by removing duplicated
code and consolidating the error handling. It also fixes two bugs in the
process:
1) ie_words was one too large when plic->num_src was a multiple of 32.
2) plic_set_ie takes a 32-bit mask, not a Boolean value, so the FPGA
platforms previously only enabled one out of every 32 interrupts.
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Samuel Holland <samuel@sholland.org>
|
|
None of the functions modify the passed-in plic_data, so mark it const.
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Samuel Holland <samuel@sholland.org>
|
|
The suspend code needs to know the resume address for two reasons:
1) Programming some hardware register or management firmware. Here we
assume the hardware/firmware maintains its state between suspends,
so it only needs to be programmed once at startup.
2) When a non-retentive suspend request ends up being retentive, due
to lack of hardware support, pending interrupt, or for some other
reason. However, the behavior here is not platform-dependent, and
this can be handled in the generic hart suspend function.
Since neither situation requires the platform-level suspend function to
know the resume address, stop passing it to that function. Instead,
handle the non-retentive to retentive situation generically.
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Samuel Holland <samuel@sholland.org>
|
|
Non-retentive suspend states may require platform-specific actions
during resume. For example, firmware may need to save and restore the
values of custom CSRs. Add a hook to support this.
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Samuel Holland <samuel@sholland.org>
|
|
Allow the vendor_ext_check and vendor_ext_provider APIs of the
generic platform to be overridden by other platforms
Signed-off-by: Mayuresh Chitale <mchitale@ventanamicro.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
HS-mode software can choose what exceptions to delegate to VS-mode using
the hedeleg CSR. Synthetic VS/VU-mode exceptions should also honor
hedeleg. They should be redirected to VS-mode if and only if delegated
by HS-mode.
Signed-off-by: dramforever <dramforever@live.com>
Reviewed-by: Xiang W <wxjstz@126.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
While OpenC906 appears to properly decode `fence.tso` as a fence
instruction[1], the version of the C906 taped out in the Allwinner D1
does not, and raises illegal instruction.
Handle this errata by emulating `fence.tso` as `fence rw, rw`.
[1]: https://github.com/T-head-Semi/openc906/blob/30827e7f/C906_RTL_FACTORY/gen_rtl/idu/rtl/aq_idu_id_decd.v#L2097
Signed-off-by: Samuel Holland <samuel@sholland.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
This table does not need to be modified at runtime.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
The device's timer_value callback is already the right prototype to use
for the get_time_val function pointer.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
Match data stores hardware attributes which do not change at runtime, so
it does not need to be mutable. Make it const.
Reviewed-by: Guo Ren <guoren@kernel.org>
Reviewed-by: Xiang W <wxjstz@126.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Samuel Holland <samuel@sholland.org>
|
|
If the csr's operation comes from M mode, it should not be forwarded
to low-privilege processing, this patch fixes this problem.
Signed-off-by: Xiang W <wxjstz@126.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
The function returns the highest hart-id of the harts actually used in
the system (enabled). Change the name to reflect this fact.
Signed-off-by: Jan Remes <jan.remes@codasip.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
Ignore CPU nodes in FDT which are not enabled.
Signed-off-by: Jan Remes <jan.remes@codasip.com>
Reviewed-by: Xiang W <wxjstz@126.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
Ignore CPU nodes in FDT that are not enabled.
Signed-off-by: Jan Remes <jan.remes@codasip.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
If an FDT node contains a "status" property and this property is not
"ok" or "okay", this node should be ignored. Introduce a function that
checks this.
Signed-off-by: Jan Remes <jan.remes@codasip.com>
Reviewed-by: Xiang W <wxjstz@126.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
|
|
Fix typo in irqchip_imsic_update_hartid_table() when checking for CPU
node.
Signed-off-by: Jan Remes <jan.remes@codasip.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
|
|
Let us move SiFive platform overrides for FU540 and FU740 into a separate
directory so better maintainability. Other SoC vendors can also create
their own directory under platform/generic.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Instead of having platform override module list hard-coded in the C source,
we generate it using carray.sh at compile-time.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Instead of having FDT gpio driver list hard-coded in the C source,
we generate it using carray.sh at compile-time.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Instead of having FDT i2c adapter driver list hard-coded in the C source,
we generate it using carray.sh at compile-time.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Instead of having FDT ipi driver list hard-coded in the C source,
we generate it using carray.sh at compile-time.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Instead of having FDT irqchip driver list hard-coded in the C source,
we generate it using carray.sh at compile-time.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Instead of having FDT timer driver list hard-coded in the C source,
we generate it using carray.sh at compile-time.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Instead of having FDT serial driver list hard-coded in the C source,
we generate it using carray.sh at compile-time.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Instead of having FDT reset driver list hard-coded in the C source,
we generate it using carray.sh at compile-time.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Generating C array at compile time based on details provided by
objects.mk is a very useful feature which will help us compile
only a subset of drivers or modules.
We add a bash script (carray.sh) which takes array details and
object/variable list from command-line to generate a C source
containing array of object/variable pointers. We also extend
top-level makefile to use carray.sh whenever specified through
objects.mk.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
The generated C source could be anywhere within build directory so
let us update the make rule to comple generated C source accordingly.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
We add platform specific extensions_init() callback which allows
platforms to populate HART extensions for each HART. For example,
the generic platform can populate HART extensions from HART ISA
string described in DeviceTree.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
We add sbi_hart_update_extension() function which allow platforms
to enable/disable hart extensions.
Signed-off-by: Anup Patel <anup@brainfault.org>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Currently, the hart_detect_features() is called everytime a hart
is stopped and started again which is unnecessary work.
We update hart_detect_features() to detect hart features only
once for each hart.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
Since past few years, we have been using "hart features" in OpenSBI
to represent all optionalities and multi-letter extensions defined
by the RISC-V specifications.
The RISC-V profiles specification has taken a different approach and
started assigning extension names for all optionalities which did not
have any extension name previously.
(Refer, https://github.com/riscv/riscv-profiles/blob/main/profiles.adoc)
Inspired from the RISC-V profiles specification, we convert OpenSBI
hart features into hart extensions. Going forward, we align the
extension naming with RISC-V profiles specification. Currently, only
"time CSR" and "AIA CSR" have not been assigned extension name but
for everything else we have a name.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
The AIA feature detection uses unnecessary goto which is not need
and AIA case in sbi_hart_feature_id2string() does not break. This
patch fixes both issues in AIA feature detection.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|
|
If a hart implements privileged spec v1.12 (or higher) then we can
safely assume that menvcfg CSR is present and we don't need MENVCFG
as a hart feature.
Signed-off-by: Anup Patel <apatel@ventanamicro.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
|