commit c87f820bc4 upstream.
On Tegra platforms using ACPI, the SMCCC driver already registers the
SoC device. This makes the registration performed by the Tegra fuse
driver redundant.
When booted via ACPI, skip registering the SoC device and suppress
printing SKU information from the Tegra fuse driver, as this information
is already provided by the SMCCC driver.
Fixes: 972167c690 ("soc/tegra: fuse: Add ACPI support for Tegra194 and Tegra234")
Cc: stable@vger.kernel.org
Signed-off-by: Kartik Rajput <kkartik@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
GCC 14.1 complains about the argument usage of kmemdup_array():
drivers/soc/tegra/fuse/fuse-tegra.c:130:65: error: 'kmemdup_array' sizes specified with 'sizeof' in the earlier argument and not in the later argument [-Werror=calloc-transposed-args]
130 | fuse->lookups = kmemdup_array(fuse->soc->lookups, sizeof(*fuse->lookups),
| ^
drivers/soc/tegra/fuse/fuse-tegra.c:130:65: note: earlier argument should specify number of elements, later size of each element
The annotation introduced by commit 7d78a77733 ("string: Add
additional __realloc_size() annotations for "dup" helpers") lets the
compiler think that kmemdup_array() follows the same format as calloc(),
with the number of elements preceding the size of one element. So we
could simply swap the arguments to __realloc_size() to get rid of that
warning, but it seems cleaner to instead have kmemdup_array() follow the
same format as krealloc_array(), memdup_array_user(), calloc() etc.
Fixes: 7d78a77733 ("string: Add additional __realloc_size() annotations for "dup" helpers")
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Link: https://lore.kernel.org/r/20240606144608.97817-2-jean-philippe@linaro.org
Signed-off-by: Kees Cook <kees@kernel.org>
If all the other SoCs are disabled, the driver fails to build:
drivers/soc/tegra/fuse/fuse-tegra30.c:684:17: error: 'tegra30_fuse_read' undeclared here (not in a function); did you mean 'tegra_fuse_readl'?
684 | .read = tegra30_fuse_read,
| ^~~~~~~~~~~~~~~~~
| tegra_fuse_readl
drivers/soc/tegra/fuse/fuse-tegra30.c:694:17: error: 'tegra30_fuse_init' undeclared here (not in a function); did you mean 'tegra_fuse_info'?
694 | .init = tegra30_fuse_init,
| ^~~~~~~~~~~~~~~~~
Fix the list of SoCs using this function to include the newly added one.
Fixes: dee509eb9cd5 ("soc/tegra: fuse: Add support for Tegra241")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Tegra241 SoC data uses tegra194_soc_attr_group, which is only defined
if config CONFIG_ARCH_TEGRA_194_SOC or CONFIG_ARCH_TEGRA_234_SOC or
both are enabled. This causes a build failure if both of these configs
are disabled and CONFIG_ARCH_TEGRA_241_SOC is enabled.
Define tegra194_soc_attr_group if CONFIG_ARCH_TEGRA_241_SOC is enabled.
Signed-off-by: Kartik <kkartik@nvidia.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org> # build-tested
Signed-off-by: Thierry Reding <treding@nvidia.com>
Add ACPI support for Tegra194 & Tegra243 SoC's. This requires
following modifications to the probe when ACPI boot is used:
- Initialize soc data.
- Add nvmem lookups.
- Register soc device.
- use devm_clk_get_optional() instead of devm_clk_get() to get
fuse->clk, as fuse clocks are not required when using ACPI boot.
Also, drop '__init' keyword for tegra_soc_device_register() as this is also
used by tegra_fuse_probe() and use dev_err_probe() wherever applicable.
Signed-off-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Add helper function tegra_fuse_print_sku_info() to print Tegra SKU
information. So, it can be shared between tegra_fuse_init() and
ACPI probe which is to be introduced later.
Signed-off-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Add helper function tegra_fuse_add_lookups() to register Tegra fuse
nvmem lookups. So, this can be shared between tegra_fuse_init() and
ACPI probe, which is to be introduced later.
Use kmemdup_array to duplicate fuse->soc->lookups.
Signed-off-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
In preparation to ACPI support in Tegra fuse driver add function
tegra_acpi_init_apbmisc() to initialize tegra-apbmisc driver.
Also, document the reason of calling tegra_init_apbmisc() at early init.
Note that function tegra_acpi_init_apbmisc() is not placed in the __init
section, because it will be called during probe.
Signed-off-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
To prepare for adding ACPI support to the tegra-apbmisc driver,
relocate the code responsible for mapping memory resources from
the function ‘tegra_init_apbmisc’ to the function
‘tegra_init_apbmisc_resources.’ This adjustment will allow the
code to be shared between ‘tegra_init_apbmisc’ and the upcoming
‘tegra_acpi_init_apbmisc’ function.
Signed-off-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Currently, in tegra_fuse_probe() if clock/reset get fails, then the
driver prints an error if the error is not caused by -EPROBE_DEFER.
This can be improved by using dev_err_probe() instead.
So, return dev_err_probe() if clock/reset get fails.
Signed-off-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Convert platform_get_resource(), devm_ioremap_resource() to a single
call to devm_platform_get_and_ioremap_resource(), as this is exactly
what this function does.
Signed-off-by: Yangtao Li <frank.li@vivo.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
The DT of_device.h and of_platform.h date back to the separate
of_platform_bus_type before it as merged into the regular platform bus.
As part of that merge prepping Arm DT support 13 years ago, they
"temporarily" include each other. They also include platform_device.h
and of.h. As a result, there's a pretty much random mix of those include
files used throughout the tree. In order to detangle these headers and
replace the implicit includes with struct declarations, users need to
explicitly include the correct includes.
Signed-off-by: Rob Herring <robh@kernel.org>
Signed-off-by: Thierry Reding <treding@nvidia.com>
The Tegra234 fuse size is incorrectly defined as 0x98c. The actual size
of the Tegra234 fuses is 0xf8c and so update the size of the Tegra234
fuses to fix this.
Signed-off-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Tegra pre-silicon platforms do not have chip revisions. This makes the
revision SoC attribute meaningless on these platforms.
Instead, populate the revision SoC attribute with a combination of the
platform name and the chip revision for silicon platforms, and simply
with the platform name on pre-silicon platforms.
Signed-off-by: Kartik <kkartik@nvidia.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
On Tegra186 and later, various FUSE offsets are restricted and cannot be
accessed from CCPLEX. Currently nvmem binary interface allows reading
such offsets from userspace, which results in RAS errors.
Add nvmem keepout lists to avoid any reads to restricted offsets.
Signed-off-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Tegra FUSE block size, availability and offsets can vary from one SoC
generation to another.
Signed-off-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
soc/tegra: cbb: Changes for v6.1-rc1
This introduces the CBB driver that is used to provide (a lot of)
information about SErrors when things go wrong, instead of the kernel
just crashing or hanging.
* tag 'tegra-for-6.1-cbb' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux:
soc/tegra: cbb: Add support for Tegra241 (Grace)
soc/tegra: cbb: Add driver for Tegra234 CBB 2.0
soc/tegra: cbb: Add CBB 1.0 driver for Tegra194
soc/tegra: Set ERD bit to mask inband errors
Link: https://lore.kernel.org/r/20220916101957.1635854-2-thierry.reding@gmail.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
In tegra_init_apbmisc(), of_find_matching_node() will return a node
pointer with refcount incremented. We should use of_node_put() in each
failure path or when it is not used anymore.
Signed-off-by: Liang He <windhl@126.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
In this function, of_find_matching_node() will return a node pointer
with refcount incremented. We should use of_node_put() when the "np"
pointer is not used anymore.
Signed-off-by: Liang He <windhl@126.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Add a function to set the ERD (Error Response Disable) bit in the
MISCREG_CCROC_ERR_CONFIG register from the Control Backbone (CBB) error
handler driver.
ERD bit allows masking of SError due to inband errors which are caused
by illegal register accesses through CBB. When the bit is set, interrupt
is used for reporting errors and magic code '0xdead2003' is returned.
This change is only required for Tegra194 SoC as the config is moved to
CBB register space for future SoC's. Also, remove unmapping the
apbmisc_base as it's required to get the base address for accessing the
misc register.
Signed-off-by: Sumit Gupta <sumitg@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Add nvmem cell lookup entries for below FUSE cells:
- gcplex-config-fuse
- pdi0
- pdi1
Also include the device name prefix "gpu-" in the names of the GPU FUSE
cells in nvmem_cell_info.
Signed-off-by: Sagar Kamble <skamble@nvidia.com>
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
sparse is picky about casts between different address spaces. A cast to
plain void * needs to be accompanied by a __force modifier and casting
back to void __iomem * needs to be explicit to avoid warnings.
Reported-by: kernel test robot <lkp@intel.com>
Fixes: 88724b78a8 ("soc/tegra: fuse: Use resource-managed helpers")
Signed-off-by: Thierry Reding <treding@nvidia.com>
Reviewed-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Update tegra_fuse_cells with below entries:
- gcplex-config-fuse:
Configuration bits for GPU, used to enable/disable write protected
region used for storing GPU firmware.
- pdi0:
Unique per chip public identifier.
- pdi1:
Unique per chip public identifier.
Signed-off-by: Prathamesh Shete <pshete@nvidia.com>
Signed-off-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Pull ARM SoC driver updates from Arnd Bergmann:
"There are cleanups and minor bugfixes across several SoC specific
drivers, for Qualcomm, Samsung, NXP i.MX, AT91, Tegra, Keystone,
Renesas, ZynqMP
Noteworthy new features are:
- The op-tee firmware driver gains support for asynchronous
notifications from secure-world firmware.
- Qualcomm platforms gain support for new SoC types in various
drivers: power domain, cache controller, RPM sleep, soc-info
- Samsung SoC drivers gain support for new SoCs in ChipID and PMU, as
well as a new USIv2 driver that handles various types of serial
communiction (uart, i2c, spi)
- Renesas adds support for R-Car S4-8 (R8A779F0) in multiple drivers,
as well as memory controller support for RZ/G2L (R9A07G044).
- Apple M1 gains support for the PMGR power management driver"
* tag 'drivers-5.17' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (94 commits)
soc: qcom: rpmh-rsc: Fix typo in a comment
soc: qcom: socinfo: Add SM6350 and SM7225
dt-bindings: arm: msm: Don't mark LLCC interrupt as required
dt-bindings: firmware: scm: Add SM6350 compatible
dt-bindings: arm: msm: Add LLCC for SM6350
soc: qcom: rpmhpd: Sort power-domain definitions and lists
soc: qcom: rpmhpd: Remove mx/cx relationship on sc7280
soc: qcom: rpmhpd: Rename rpmhpd struct names
soc: qcom: rpmhpd: sm8450: Add the missing .peer for sm8450_cx_ao
soc: qcom: socinfo: add SM8450 ID
soc: qcom: rpmhpd: Add SM8450 power domains
dt-bindings: power: rpmpd: Add SM8450 to rpmpd binding
soc: qcom: smem: Update max processor count
dt-bindings: arm: qcom: Document SM8450 SoC and boards
dt-bindings: firmware: scm: Add SM8450 compatible
dt-bindings: arm: cpus: Add kryo780 compatible
soc: qcom: rpmpd: Add support for sm6125
dt-bindings: qcom-rpmpd: Add sm6125 power domains
soc: qcom: aoss: constify static struct thermal_cooling_device_ops
PM: AVS: qcom-cpr: Use div64_ul instead of do_div
...
Use resource-managed helpers to make code cleaner and more correct,
properly releasing all resources in case of driver probe error.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
The FUSE controller is enabled at a boot time. Reset it in order to put
hardware and clock into clean and disabled state.
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
A new warning in clang points out two instances where boolean
expressions are being used with a bitwise OR instead of logical OR:
drivers/soc/tegra/fuse/speedo-tegra20.c:72:9: warning: use of bitwise '|' with boolean operands [-Wbitwise-instead-of-logical]
reg = tegra_fuse_read_spare(i) |
^~~~~~~~~~~~~~~~~~~~~~~~~~
||
drivers/soc/tegra/fuse/speedo-tegra20.c:72:9: note: cast one or both operands to int to silence this warning
drivers/soc/tegra/fuse/speedo-tegra20.c:87:9: warning: use of bitwise '|' with boolean operands [-Wbitwise-instead-of-logical]
reg = tegra_fuse_read_spare(i) |
^~~~~~~~~~~~~~~~~~~~~~~~~~
||
drivers/soc/tegra/fuse/speedo-tegra20.c:87:9: note: cast one or both operands to int to silence this warning
2 warnings generated.
The motivation for the warning is that logical operations short circuit
while bitwise operations do not.
In this instance, tegra_fuse_read_spare() is not semantically returning
a boolean, it is returning a bit value. Use u32 for its return type so
that it can be used with either bitwise or boolean operators without any
warnings.
Fixes: 25cd5a3914 ("ARM: tegra: Add speedo-based process identification")
Link: https://github.com/ClangBuiltLinux/linux/issues/1488
Suggested-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
The FUSE clock should be enabled during suspend on Tegra124. Currently
clk driver enables it on all SoCs, but FUSE may require a higher core
voltage on Tegra30 while enabled. Move the quirk into the FUSE driver
and make it specific to Tegra124.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
The Tegra FUSE belongs to the core power domain and we're going to enable
GENPD support for the core domain. Now FUSE device must be resumed using
runtime PM API in order to initialize the FUSE power state. Add runtime PM
support to the FUSE driver.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
The fuse->clk must be cleared if FUSE driver fails to probe, otherwise
tegra_fuse_readl() will crash. It's unlikely to happen in practice,
nevertheless let's correct it for completeness.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
fuse->base can not be unmapped if allocate lookups failed in
tegra_init_fuse(), because it is an early_initcall, the driver
will be loaded anyway and fuse->base will be accessed by other
functions later, so remove the return -ENOMEM after allocating
lookups failed to make less confusing.
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
[treding@nvidia.com: drop error message, out-of-memory is noisy anyway]
Signed-off-by: Thierry Reding <treding@nvidia.com>
soc/tegra: Changes for v5.11-rc1
This contains a couple of warning fixes and a fix for a mostly harmless
bug in the process ID retrieval code.
* tag 'tegra-for-5.11-soc' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux:
soc/tegra: fuse: Fix index bug in get_process_id
soc: tegra: fuse: speedo-tegra210: Remove a group of set but unused variables
soc: tegra: fuse: speedo-tegra124: Remove some set but unused variables
Link: https://lore.kernel.org/r/20201127144329.124891-2-thierry.reding@gmail.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes the following W=1 kernel build warning(s):
drivers/soc/tegra/fuse/speedo-tegra210.c: In function ‘tegra210_init_speedo_data’:
drivers/soc/tegra/fuse/speedo-tegra210.c:105:56: warning: variable ‘soc_iddq’ set but not used [-Wunused-but-set-variable]
drivers/soc/tegra/fuse/speedo-tegra210.c:105:46: warning: variable ‘gpu_iddq’ set but not used [-Wunused-but-set-variable]
drivers/soc/tegra/fuse/speedo-tegra210.c:105:36: warning: variable ‘cpu_iddq’ set but not used [-Wunused-but-set-variable]
Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Jonathan Hunter <jonathanh@nvidia.com>
Cc: linux-tegra@vger.kernel.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
[treding@nvidia.com: remove unnecessary reads altogether]
Signed-off-by: Thierry Reding <treding@nvidia.com>
Fixes the following W=1 kernel build warning(s):
drivers/soc/tegra/fuse/speedo-tegra124.c: In function ‘tegra124_init_speedo_data’:
drivers/soc/tegra/fuse/speedo-tegra124.c:105:38: warning: variable ‘soc_iddq_value’ set but not used [-Wunused-but-set-variable]
drivers/soc/tegra/fuse/speedo-tegra124.c:105:22: warning: variable ‘gpu_iddq_value’ set but not used [-Wunused-but-set-variable]
drivers/soc/tegra/fuse/speedo-tegra124.c:105:6: warning: variable ‘cpu_iddq_value’ set but not used [-Wunused-but-set-variable]
Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Jonathan Hunter <jonathanh@nvidia.com>
Cc: linux-tegra@vger.kernel.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
[treding@nvidia.com: remove unnecessary reads altogether]
Signed-off-by: Thierry Reding <treding@nvidia.com>
The MISC block is largely similar to that found on earlier chips, but
not completely compatible. Allow binding to the instantiation found on
Tegra234.
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Add support for FUSE block found on the Tegra234 SoC, which is largely
similar to the IP found on previous generations.
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
This function can be used by drivers to determine whether code is
running on silicon or on a simulation platform.
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
This function extracts the PRE_SI_PLATFORM field from the HIDREV
register and can be used to determine which platform the kernel runs on
(silicon, simulation, ...). Note that while only Tegra194 and later
define this field, it should be safe to call this on prior generations
as well since this field should read as 0, indicating silicon.
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Currently the SoC revision attribute for Tegra devices displays the
value of the enum associated with a particular revision. This is not
very useful because to obtain the actual revision you need to
use the tegra_revision enumeration to translate the value.
It is more meaningful to display a name for the revision, such as
'A01', than the enumarated value and therefore, update the revision
attribute to display a name. This change does alter the ABI, which
is unfortunate, but this is more meaningful and maintable.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
