The K3 R5F remoteproc driver in U-Boot was upstreamed prior to the
equivalent remoteproc driver in the Linux kernel. Some of the DT
properties used in U-Boot got upstreamed using different names
in Linux kernel.
The modified property names include the R5F cluster mode configuration
property "lockstep-mode"; and three different individual R5F core config
properties - "atcm-enable", "btcm-enable" and "loczrama". The property
names were updated as follows:
lockstep-mode => ti,cluster-mode
atcm-enable => ti,atcm-enable
btcm-enable => ti,btcm-enable
loczrama => ti,loczrama
Update the K3 R5F remoteproc driver, the corresponding binding, and
all the existing usage in AM65x, J721E and J7200 dts files all at
once to use the new properties and to not break any bisectability.
Signed-off-by: Suman Anna <s-anna@ti.com>
Define LOG_CATEGORY and remove unneeded pr_fmt macro with the dev
macro as dev->name is displayed and CONFIG_LOGF_FUNC can be
activated for log macro.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
ARMv8's generic timer[1] picks up it's graycode from GTC. However,
the frequency of the GTC is supposed to be programmed in CNTFID0[2]
register prior to enabling the GTC in CNTCR[3] register.
In K3 architecture, GTC provides a central time to many parts of the
SoC including graycode to the generic timer in the ARMv8 subsystem.
However, due to the central nature and the need to enable the counter
early in the boot process, the R5 based u-boot enables GTC and
programs it's frequency based on central needs of the system. This
may not be a constant 200MHz based on the system. The bootloader is
supposed to program the FID0 register with the correct frequency it
has sourced for GTC from the central system controller OR from PLLs
as appropriate, and TF-A is supposed[4] to use that as the frequency for
it's local timer.
Currently we are programming just the CNTCR[3] register to enable the
GTC, however we dont let TF-A know the frequency that GTC is actually
running at. A mismatch in programmed frequency and what we program for
generic timer will, as we can imagine, all kind of weird mayhem.
So, program the CNTFID0 register with the clock frequency. Note:
assigned-clock-rates should have set the clock frequency, so the only
operation we need to explicitly do is to retrieve the frequency and
program it in FID0 register.
Since the valid in K3 for GTC clock frequencies are < U32_MAX, we can
just cast the ulong and continue.
[1] https://developer.arm.com/documentation/100095/0002/generic-timer/generic-timer-register-summary/aarch64-generic-timer-register-summary
[2] https://developer.arm.com/docs/ddi0595/h/external-system-registers/cntfid0
[3] https://developer.arm.com/docs/ddi0595/h/external-system-registers/cntcr
[4] 6a22d9ea3c
Signed-off-by: Nishanth Menon <nm@ti.com>
The current macro is a misnomer since it does not declare a device
directly. Instead, it declares driver_info record which U-Boot uses at
runtime to create a device.
The distinction seems somewhat minor most of the time, but is becomes
quite confusing when we actually want to declare a device, with
of-platdata. We are left trying to distinguish between a device which
isn't actually device, and a device that is (perhaps an 'instance'?)
It seems better to rename this macro to describe what it actually is. The
macros is not widely used, since boards should use devicetree to declare
devices.
Rename it to U_BOOT_DRVINFO(), which indicates clearly that this is
declaring a new driver_info record, not a device.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present flags are stored as part of the device. In preparation for
storing them separately, change the access to go through inline functions.
Signed-off-by: Simon Glass <sjg@chromium.org>
Most drivers use these access methods but a few do not. Update them.
In some cases the access is not permitted, so mark those with a FIXME tag
for the maintainer to check.
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Pratyush Yadav <p.yadav@ti.com>
Fix up the code style for those declarations that should now fit onto one
line, which is all of them that currently do not.
This is needed for dtoc to detect the structs correctly, at present.
Signed-off-by: Simon Glass <sjg@chromium.org>
We use 'priv' for private data but often use 'platdata' for platform data.
We can't really use 'pdata' since that is ambiguous (it could mean private
or platform data).
Rename some of the latter variables to end with 'plat' for consistency.
Signed-off-by: Simon Glass <sjg@chromium.org>
This construct is quite long-winded. In earlier days it made some sense
since auto-allocation was a strange concept. But with driver model now
used pretty universally, we can shorten this to 'auto'. This reduces
verbosity and makes it easier to read.
Coincidentally it also ensures that every declaration is on one line,
thus making dtoc's job easier.
Signed-off-by: Simon Glass <sjg@chromium.org>
The coprocessor is running as soon as the hold boot is de-asserted.
So indicate this running state and save the resource table even
if the protective assert, to avoid autonomous reboot, is failed.
This error case should never occurs.
Cc: Fabien DESSENNE <fabien.dessenne@st.com>
Cc: Arnaud POULIQUEN <arnaud.pouliquen@st.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Use the reset function to handle the hold boot bit in RCC
with device tree handle with MCU_HOLD_BOOT identifier.
This generic reset allows to remove the two specific properties:
- st,syscfg-holdboot
- st,syscfg-tz
This patch prepares alignment with kernel device tree.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Fabien DESSENNE <fabien.dessenne@st.com>
Cc: Arnaud POULIQUEN <arnaud.pouliquen@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Usually we can get a device from the current core, but some dev_dbg calls
have been converted to debug, since we are called on a cluster.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Tested-by: Patrick Delaunay <patrick.delaunay@st.com>
This function is never used anywhere, and it also tries to log with a
nonexistant device.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Tested-by: Patrick Delaunay <patrick.delaunay@st.com>
The K3 J7200 SoC family has a revised R5F sub-system and contains a
subset of the R5F clusters present on J721E SoCs. The integration of
these clusters is very much similar to J721E SoCs otherwise.
The revised IP has the following two new features:
1. TCMs are auto-initialized during module power-up, and the behavior
is programmable through a MMR bit controlled by System Firmware.
2. The LockStep-mode allows the Core1 TCMs to be combined with the
Core0 TCMs effectively doubling the amount of TCMs available.
The LockStep-mode on previous SoCs could only use the Core0 TCMs.
This combined TCMs appear contiguous at the respective Core0 TCM
addresses.
Add the support to these clusters in the K3 R5F remoteproc driver
using J7200 specific compatibles and revised logic accounting for
the above IP features/differences.
Signed-off-by: Suman Anna <s-anna@ti.com>
Move this header out of the common header. Network support is used in
quite a few places but it still does not warrant blanket inclusion.
Note that this net.h header itself has quite a lot in it. It could be
split into the driver-mode support, functions, structures, checksumming,
etc.
Signed-off-by: Simon Glass <sjg@chromium.org>
The R5F subsystem/cluster on K3 SoCs can support both LockStep and
Split-modes (superset) or just Split-mode depending on an eFUSE
capability register. The LockStep configuration bit is Read-only
though on Split-mode _only_ devices and as such the System Firmware
does not allow the LockStep mode bit to be configured on such devices.
The current logic in k3_r5f_rproc_configure() fails on Split-mode
devices because of this unconditional programming of the LockStep
mode bit, and results in the probe failure shown during the
"rproc init" step at U-Boot prompt.
Fix this by limiting the LockStep mode bit clear configuration only on
devices supporting both LockStep/Split-modes.
Fixes: 4c850356a8 ("remoteproc: Introduce K3 remoteproc driver for R5F subsystem")
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Andreas Dannenberg <dannenberg@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
The Main R5FSS0 cluster is also enabled to probe the R5F remoteproc
driver within R5 SPL for booting the Core0 very early. This results
in a ti_sci_power_domain_on failure during the probe from the A72
U-Boot when "rproc init" is executed at U-Boot prompt, and doesn't
enumerate all the rproc devices.
Fix this by suppressing the power_domain_on altogether using the
flag DM_FLAG_DEFAULT_PD_CTRL_OFF added in commit af94ad418d
("dm: core: Allow for not controlling the power-domain by DM framework").
Fixes: fac6aa817a ("configs: j721e_evm_r5: Enable R5F remoteproc support")
Signed-off-by: Suman Anna <s-anna@ti.com>
The resets for the DSP processors on K3 SoCs are managed through the
Power and Sleep Controller (PSC) module. Each DSP typically has two
resets - a global module reset for powering on the device, and a local
reset that affects only the CPU while allowing access to the other
sub-modules within the DSP processor sub-systems.
The C66x DSPs have two levels of internal RAMs that can be used to
boot from, and the firmware loading into these RAMs require the
local reset to be asserted with the device powered on/enabled using
the module reset. Enhance the K3 DSP remoteproc driver to add support
for loading into the internal RAMs. The local reset is deasserted on
SoC power-on-reset, so logic has to be added in probe in remoteproc
mode to balance the remoteproc state-machine.
Note that the local resets are a no-op on C71x cores, and the hardware
does not supporting loading into its internal RAMs.
Signed-off-by: Suman Anna <s-anna@ti.com>
The DSP remote processors on K3 SoCs require a boot register to be
programmed with a boot address, and these boot addresses need to be
aligned on certain address boundaries. The current code does not have
any error checks, and relies on the System Firmware to perform the
checking. Add logic to perform this sanity check within the remoteproc
driver itself to detect these anomalies specifically, and print a
meaningful trace. This avoids the cumbersome debug of root-causing
such failures from the corresponding TI-SCI failure.
The C66x and C71x DSP cores have different alignment needs and are
as follows:
C66x DSP = 1 KB (0x400)
C71x DSP = 2 MB (0x200000)
Signed-off-by: Suman Anna <s-anna@ti.com>
The global module reset is deasserted through the ti_sci_power_domain_on()
call in k3_dsp_start(), but is not asserted back if the local module reset
fails. Fix this.
While at this, remove the stale comment about assigned-clock-rates that
seems to have been copied from the K3 ARM64 Remoteproc driver.
Fixes: ab827b3857 ("remoteproc: Introduce K3 C66 and C71 remoteproc driver")
Signed-off-by: Suman Anna <s-anna@ti.com>
At present dm/device.h includes the linux-compatible features. This
requires including linux/compat.h which in turn includes a lot of headers.
One of these is malloc.h which we thus end up including in every file in
U-Boot. Apart from the inefficiency of this, it is problematic for sandbox
which needs to use the system malloc() in some files.
Move the compatibility features into a separate header file.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present devres.h is included in all files that include dm.h but few
make use of it. Also this pulls in linux/compat which adds several more
headers. Drop the automatic inclusion and require files to include devres
themselves. This provides a good indication of which files use devres.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Anatolij Gustschin <agust@denx.de>
Fix an issue where some sections are never loaded : if p_type is
different from PT_LOAD the phdr pointer must be incremented.
Signed-off-by: Fabien Dessenne <fabien.dessenne@st.com>
Acked-by: Suman Anna <s-anna@ti.com>
Load the optional resource table from the firmware, and write its
address in the dedicated backup register.
Signed-off-by: Fabien Dessenne <fabien.dessenne@st.com>
Acked-by: Patrick Delaunay <patrick.delaunay@st.com>
Update the dedicated backup register to track the coprocessor state and
rely on that register to compute the .is_running() value (which expects
a return value of 0 -not 1- if the processor is running).
Signed-off-by: Fabien Dessenne <fabien.dessenne@st.com>
Acked-by: Patrick Delaunay <patrick.delaunay@st.com>
Add rproc_elf_load_rsc_table(), which searches for a resource table in
an elf64/elf32 image, and if found, copies it to device memory.
Add also the elf32 and elf64 variants of this API.
Add a test for this.
Signed-off-by: Fabien Dessenne <fabien.dessenne@st.com>
Acked-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
These functions are CPU-related and do not use driver model. Move them to
cpu_func.h
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Enable DM_FLAG_DEFAULT_PD_CTRL_OFF for arm64 remote core so that
pd can be enabled after loading the image.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Certain SoCs with K3 architecture have integrated a C66 Corepac DSP
subsystem and an advanced C71 DSPs. Introduce a remoteproc driver
that that does take care of loading an elf to any of the specified
DSPs and start it.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
SoCs with K3 architecture have an integrated Arm Cortex-R5F subsystem
that is comprised of dual-core Arm Cortex-R5F processor cores. This R5
subsytem can be configured at boot time to be either run in a LockStep
mode or in an Asymmetric Multi Processing (AMP) fashion in Split-mode.
This subsystem has each Tightly-Coupled Memory (TCM) internal memories
for each core split between two banks - TCMA and TCMB.
Add a remoteproc driver to support this subsystem to be able to load
and boot the R5 cores primarily in LockStep mode or split mode.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
Power domain for the remote cores needs to be handled in a right
sequence as mandated by the spec. Introduce tisci helper apis
that can control power-domains of remote cores. TISCI clients
can use this api and control the remote cores power domain instead
of hooking it to power-domain layer.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Introduce rproc_elf_get_boot_addr() that returns the entry point of
the elf file. This api auto detects the 64/32 bit elf file and returns
the boot addr accordingly.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Introduce a common remoteproc elf loader and checker functions that
automatically detects the 64 bit elf file or 32 bit elf file and
loads/checks the sections accordingly.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Fabien Dessenne <fabien.dessenne@st.com>
The current rproc-elf-loader supports loading of only 32 bit elf files.
Introduce support for loading of 64 bit elf files in rproc-elf-loader.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Fabien Dessenne <fabien.dessenne@st.com>
rproc_elf32_load_image() rely on user to send a valid address for elf loading.
Instead do a sanity check on the address passed by user. This will help
all rproc elf users to not call sanity_check explicitly before calling
elf_loading.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Fabien Dessenne <fabien.dessenne@st.com>
Introduce a new parameter "size" that accepts size of the region to
remoteproc ops callback device_to_virt(). This can enforce more checks
on the region that device_to_virt() is dealing with.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Tested-by: Fabien Dessenne <fabien.dessenne@st.com>
Reviewed-by: Fabien Dessenne <fabien.dessenne@st.com>
Alignment with kernel patch proposal for binding:
[PATCH v4 0/8] stm32 m4 remoteproc on STM32MP157c
https://lkml.org/lkml/2019/5/14/159
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
k3_rproc driver is specifically meant for controlling an arm64
core using TISCI protocol. So rename the driver, Kconfig symbol,
compatible and functions accordingly.
While at it drop this remoteproc selection for a53 defconfig.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Update the k3_rproc driver to use the generic ti_sci_proc helper
apis which simplifies the driver a bit.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Texas Instruments' K3 generation SoCs has specific modules/register
spaces used for configuring the various aspects of a remote processor.
These include power, reset, boot vector and other configuration features
specific to each compute processor present on the SoC. These registers
are managed by the System Controller such as DMSC on K3 AM65x SoCs.
The Texas Instrument's System Control Interface (TI-SCI) Message Protocol
is used to communicate to the System Controller from various compute
processors to invoke specific services provided by the firmware running
on the System Controller.
Add a common processor control interface header file that can be used by
multiple remoteproc drivers. The helper functions within this header file
abstract the various TI SCI protocol ops for the remoteproc drivers, and
allow them to request the System Controller to be able to program and
manage various remote processors on the SoC. The common macros required
by the R5 remoteproc driver were also added. The remoteproc drivers are
expected to manage the life-cycle of their ti_sci_proc_dev local
structures.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
This patch introduces support of Cortex-M4 remote processor for STM32
MCU and MPU families.
Signed-off-by: Loic Pallardy <loic.pallardy@st.com>
Signed-off-by: Fabien Dessenne <fabien.dessenne@st.com>
The current implementation supports only binary file load.
Add helpers to support ELF32 format (sanity check, and load).
Note that since an ELF32 image is built for the remote processor, the
load function uses the device_to_virt ops to translate the addresses.
Implement a basic translation for sandbox_testproc.
Add related tests. Test result:
=> ut dm remoteproc_elf
Test: dm_test_remoteproc_elf: remoteproc.c
Test: dm_test_remoteproc_elf: remoteproc.c (flat tree)
Failures: 0
Signed-off-by: Loic Pallardy <loic.pallardy@st.com>
Signed-off-by: Fabien Dessenne <fabien.dessenne@st.com>
Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
There is one case where 400ms is not sufficient for loading the
system firmware:
- System firmware is not signed with rsa degenerate key.
- ROM loading the sysfw directly from SPI flash which is in memory
mapped mode.
The above scenario is definitely not desired in production use cases
as it effects boot time. But still keeping this support as this is
a valid boot scenario.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Add support for K3 based remoteproc driver that
communicates with TISCI to start start a remote processor.
Reviewed-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
K3 specific SoCs have a dedicated microcontroller for doing
resource management. Any HLOS/firmware on compute clusters should
load a firmware to this microcontroller before accessing any resource.
Adding support for loading this firmware.
After the K3 system controller got loaded with firmware and started
up it sends out a boot notification message through the secure proxy
facility using the TI SCI protocol. Intercept and receive this message
through the rproc start operation which will need to get invoked
explicitly after the firmware got loaded.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Andreas Dannenberg <dannenberg@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Existing rproc_init() api tries to initialize all available
remoteproc devices. This will fail when there is dependency
among available remoteprocs. So introduce a separate api
that allows to initialize remoteprocs individually based
on id.
Reviewed-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
When U-Boot started using SPDX tags we were among the early adopters and
there weren't a lot of other examples to borrow from. So we picked the
area of the file that usually had a full license text and replaced it
with an appropriate SPDX-License-Identifier: entry. Since then, the
Linux Kernel has adopted SPDX tags and they place it as the very first
line in a file (except where shebangs are used, then it's second line)
and with slightly different comment styles than us.
In part due to community overlap, in part due to better tag visibility
and in part for other minor reasons, switch over to that style.
This commit changes all instances where we have a single declared
license in the tag as both the before and after are identical in tag
contents. There's also a few places where I found we did not have a tag
and have introduced one.
Signed-off-by: Tom Rini <trini@konsulko.com>
