More Qualcomm driver updates for v6.5
The detection of split/non-split firmware files in the MDT loader is
corrected. The Geni driver is updated to not enable unused interrupts,
in some configurations. The count unit for MSM8998 in BWMON is corrected.
RPM master stats driver is corrected to check for the right return value
of devm_ioremap().
Support for socinfo version 18 and 19 are aded, and IPQ5300 is added to
the list of platforms.
* tag 'qcom-drivers-for-6.5-2' of https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux:
soc: qcom: geni-se: Do not bother about enable/disable of interrupts in secondary sequencer
dt-bindings: sram: qcom,imem: document qdu1000
soc: qcom: icc-bwmon: Fix MSM8998 count unit
dt-bindings: soc: qcom,rpmh-rsc: Require power-domains
soc: qcom: socinfo: Add Soc ID for IPQ5300
dt-bindings: arm: qcom,ids: add SoC ID for IPQ5300
soc: qcom: Fix a IS_ERR() vs NULL bug in probe
soc: qcom: socinfo: Add support for new fields in revision 19
soc: qcom: socinfo: Add support for new fields in revision 18
dt-bindings: firmware: scm: Add compatible for SDX75
soc: qcom: mdt_loader: Fix split image detection
Link: https://lore.kernel.org/r/20230615163104.1461905-1-andersson@kernel.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Qualcomm driver updates for v6.5
Konrad Dybcio is promoted, from reviewer, to co-maintainer.
The mdt_loader gets a fix to the detection of split binaries, where the
previous logic sometimes concluded that the first segments was not
split, in a split image. The unconditional calling of
scm_pas_mem_setup() turns out to cause a regression and is reverted.
The altmode subfunction of pmic_glink is enabled for SM8450.
A new driver for exposing power statistics from the RPM, for debugging
purposes, is introduced.
OCMEM gets a debug prints of the hardware version, QMI helpers are
transitioned to alloc_ordered_workqueue() and an error message in
ramp_controller is improved.
An API is introduced to the SMEM driver to allow other drivers to query
the SoC id, rather than open-coding the parsing of the relevant SMEM
item. This is then used to clean up the Qualcomm NVMEM-based cpufreq
driver.
Socinfo is extended with knowledge about IPQ5018, IPQ5312 and IPQ5302.
* tag 'qcom-drivers-for-6.5' of https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux: (23 commits)
soc: qcom: ocmem: Add OCMEM hardware version print
cpufreq: qcom-nvmem: use helper to get SMEM SoC ID
cpufreq: qcom-nvmem: use SoC ID-s from bindings
soc: qcom: smem: introduce qcom_smem_get_soc_id()
soc: qcom: smem: Switch to EXPORT_SYMBOL_GPL()
soc: qcom: socinfo: move SMEM item struct and defines to a header
soc: qcom: mdt_loader: Fix unconditional call to scm_pas_mem_setup
MAINTAINERS: Add Konrad Dybcio as linux-arm-msm co-maintainer
dt-bindings: sram: qcom,imem: Document MSM8226
soc: qcom: socinfo: Add Soc ID for IPQ5312 and IPQ5302
dt-bindings: arm: qcom,ids: add SoC ID for IPQ5312 and IPQ5302
soc: qcom: socinfo: Add IDs for IPQ5018 family
dt-bindings: arm: qcom,ids: Add IDs for IPQ5018 family
soc: qcom: Introduce RPM master stats driver
dt-bindings: soc: qcom: Add RPM Master stats
soc: qcom: qmi: Use alloc_ordered_workqueue() to create ordered workqueues
soc: qcom: ramp_controller: Improve error message for failure in .remove()
dt-bindings: soc: qcom: smd-rpm: allow MSM8226 over SMD
soc: qcom: rpmpd: use correct __le32 type
dt-bindings: soc: qcom: eud: Fix compatible string in the example
...
Link: https://lore.kernel.org/r/20230611010044.2481875-1-andersson@kernel.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
PCI: tegra: Changes for v6.5-rc1
This contains updates to the PCI driver for Tegra194 and later devices
that depend on the memory controller interconnect changes.
* tag 'tegra-for-6.5-pci' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux:
PCI: tegra194: Add interconnect support in Tegra234
PCI: tegra194: Fix possible array out of bounds access
Link: https://lore.kernel.org/r/20230609193620.2275240-5-thierry.reding@gmail.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
memory: tegra: Changes for v6.5-rc1
This introduces an interconnect provider for the memory controller and
external memory controller found on Tegra234 chips that will eventually
be used to dynamically scale the EMC frequency based on a device's
bandwidth needs.
* tag 'tegra-for-6.5-memory' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux:
memory: tegra: Make CPU cluster BW request a multiple of MC channels
memory: tegra: Add software memory clients in Tegra234
memory: tegra: Add memory clients for Tegra234
memory: tegra: Add interconnect support for DRAM scaling in Tegra234
dt-bindings: tegra: Add ICC IDs for dummy memory clients
dt-bindings: tegra: Document compatible for IGX
Link: https://lore.kernel.org/r/20230609193620.2275240-4-thierry.reding@gmail.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
firmware: tegra: Changes for v6.5-rc1
This adds support for using system memory as shared memory between the
CPU and the BPMP, which will be needed for Tegra264 support.
* tag 'tegra-for-6.5-firmware' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux:
firmware: tegra: bpmp: Add support for DRAM MRQ GSCs
Link: https://lore.kernel.org/r/20230609193620.2275240-2-thierry.reding@gmail.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
soc/tegra: Changes for v6.5-rc1
This adds initial support for identifying the Tegra264 SoC family and
fixes potential issues when reading from the FUSE block. A new software
wake event for the AON cluster is added on Tegra234 and the debugfs
initialization is drastically simplified.
* tag 'tegra-for-6.5-soc' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux:
soc/tegra: pmc: Use devm_clk_notifier_register()
soc/tegra: pmc: Simplify debugfs initialization
soc/tegra: fuse: Fix Tegra234 fuse size
soc/tegra: pmc: Add AON SW Wake support for Tegra234
soc/tegra: fuse: Add support for Tegra264
Link: https://lore.kernel.org/r/20230609193620.2275240-1-thierry.reding@gmail.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
The select_fifo/dma_mode() functions in geni driver enable/disable
interrupts (secondary included) conditionally for non-uart modes, while
uart is supposed to manage this internally.
However, only uart uses secondary IRQs while spi, i2c do not care about
these at all making their enablement (or disablement) totally unnecessary
for these protos.
Similarly, select_gpi_mode() also does disable s_irq and its useless again.
Drop enabling/disabling secondary IRQs.
This doesn't solve any observed problem but only gets rid of code pieces
that are not required.
Signed-off-by: Vijaya Krishna Nivarthi <quic_vnivarth@quicinc.com>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/1686742087-30731-1-git-send-email-quic_vnivarth@quicinc.com
Add support for below fields coming in socinfo structure under v19:
* num_func_clusters: number of clusters with at least one functional core
* boot_cluster: cluster selected as boot cluster
* boot_core: core selected as boot core
While at it, rename some variables to align them with their
functionalities.
Signed-off-by: Naman Jain <quic_namajain@quicinc.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20230606134626.18790-3-quic_namajain@quicinc.com
Add support for below fields coming in socinfo structure under v18:
* num_kvps: number of key value pairs (KVP)
* kvps_offset: the offset of the KVP table from the base address of
socinfo structure in SMEM
KVP table has boolean values for certain feature flags, used to determine
hardware configuration.
Signed-off-by: Naman Jain <quic_namajain@quicinc.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20230606134626.18790-2-quic_namajain@quicinc.com
The enhanced detection introduced in commit '210d12c8197a ("soc: qcom:
mdt_loader: Enhance split binary detection")' requires that all segments
lies within the file on disk.
But the Qualcomm firmware files consistently has a BSS-like segment at
the end, with a p_offset aligned to the next 4k boundary. As the p_size
is 0 and there's nothing to load, the image is not padded to cover this
(empty) segment.
Ignore zero-sized segments when determining if the image is split, to
avoid this problem.
Fixes: 210d12c819 ("soc: qcom: mdt_loader: Enhance split binary detection")
Signed-off-by: Bjorn Andersson <quic_bjorande@quicinc.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Tested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> # qrb5165-rb5
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20230612215804.1883458-1-quic_bjorande@quicinc.com
Implement support for DRAM MRQ GSCs.
Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
[treding@nvidia.com: drop unnecessary discrimination enum]
Signed-off-by: Thierry Reding <treding@nvidia.com>
Move to the device-managed version of clk_notifier_register() to remove
the need for manual cleanup. This fixes a potential issue where the
clock notifier would stick around after the driver fails to probe at a
later point.
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
debugfs calls should generally not be error-checked to simplify the case
where debugfs is disabled. Since this driver is built-in and has the
sysfs bind/unbind attributes disabled, it cannot be unloaded, so there
is no need to hold onto a reference to the debugfs files that are
created.
We can further simplify this by moving the debugfs file creation to a
later stage to avoid any cleanup we might have to do during error unwind
operations. This is also a little cleaner because the debugfs file
relies on data structures that are created at a later point than when
the file was previously created.
Suggested-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
arm64: ZynqMP SoC changes for v6.5
soc-power
- Use of_property_present() instead of of_find_property()
soc-event
- Avoid use after free eve_data in event manager code
firmware:
- Extend zynqmp_pm_fpga_load() interface
MAINTAINERS:
- Clean xilinx records
* tag 'zynqmp-soc-for-v6.5' of https://github.com/Xilinx/linux-xlnx:
MAINTAINERS: Switch to @amd.com emails
MAINTAINERS: Remove Hyun and Anurag from maintainer list
firmware: xilinx: Update the zynqmp_pm_fpga_load() API
driver: soc: xilinx: use _safe loop iterator to avoid a use after free
soc: xilinx: Use of_property_present() for testing DT property presence
Link: https://lore.kernel.org/r/411aee78-e7e8-5966-cbe8-40ff45e27ba2@monstr.eu
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Non-urgent ti-sysc driver fix for v6.5 merge window
A fix for masking bool variables that can wait for the merge window
no problem.
* tag 'omap-for-v6.5/ti-sysc-signed' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap:
bus: ti-sysc: Fix dispc quirk masking bool variables
Link: https://lore.kernel.org/r/pull-1685700720-242492@atomide.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Renesas driver updates for v6.5 (take two)
- Convert the R-Mobile SYSC driver to readl_poll_timeout_atomic().
* tag 'renesas-drivers-for-v6.5-tag2' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/renesas-devel:
soc: renesas: rmobile-sysc: Convert to readl_poll_timeout_atomic()
iopoll: Do not use timekeeping in read_poll_timeout_atomic()
iopoll: Call cpu_relax() in busy loops
Link: https://lore.kernel.org/r/cover.1686304612.git.geert+renesas@glider.be
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
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>
Wake-up signal 83 on Tegra234 is triggered by software writing to
WAKE_AOWAKE_SW_WAKE_TIER0_TRIGGER_0 register. This wake-up is mapped
to CPU interrupt 179 and is used by the Sensor Processing Engine (SPE)
in the Always-on (AON) power domain for waking up the system.
Signed-off-by: Viswanath L <viswanathl@nvidia.com>
Signed-off-by: Kartik <kkartik@nvidia.com>
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Use readl_poll_timeout_atomic() instead of open-coding the same
operation.
1. rmobile_pd_power_down(): as typically less than 20 retries are
needed, PSTR_RETRIES (100) µs is a suitable timeout value.
2. __rmobile_pd_power_up(): the old method of first polling some
cycles with a 1 µs delay, followed by more polling cycles without
any delay didn't make much sense, as the latter was insignificant
compared to the former. Furthermore, typically no retries are
needed. Hence just retain the polling with delay.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/ae4bf03ab8fd5a557c683086958d6764babc0723.1685692810.git.geert+renesas@glider.be
read_poll_timeout_atomic() uses ktime_get() to implement the timeout
feature, just like its non-atomic counterpart. However, there are
several issues with this, due to its use in atomic contexts:
1. When called in the s2ram path (as typically done by clock or PM
domain drivers), timekeeping may be suspended, triggering the
WARN_ON(timekeeping_suspended) in ktime_get():
WARNING: CPU: 0 PID: 654 at kernel/time/timekeeping.c:843 ktime_get+0x28/0x78
Calling ktime_get_mono_fast_ns() instead of ktime_get() would get
rid of that warning. However, that would break timeout handling,
as (at least on systems with an ARM architectured timer), the time
returned by ktime_get_mono_fast_ns() does not advance while
timekeeping is suspended.
Interestingly, (on the same ARM systems) the time returned by
ktime_get() does advance while timekeeping is suspended, despite
the warning.
2. Depending on the actual clock source, and especially before a
high-resolution clocksource (e.g. the ARM architectured timer)
becomes available, time may not advance in atomic contexts, thus
breaking timeout handling.
Fix this by abandoning the idea that one can rely on timekeeping to
implement timeout handling in all atomic contexts, and switch from a
global time-based to a locally-estimated timeout handling. In most
(all?) cases the timeout condition is exceptional and an error
condition, hence any additional delays due to underestimating wall clock
time are irrelevant.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Tony Lindgren <tony@atomide.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/3d2a2f4e553489392d871108797c3be08f88300b.1685692810.git.geert+renesas@glider.be
It is considered good practice to call cpu_relax() in busy loops, see
Documentation/process/volatile-considered-harmful.rst. This can not
only lower CPU power consumption or yield to a hyperthreaded twin
processor, but also allows an architecture to mitigate hardware issues
(e.g. ARM Erratum 754327 for Cortex-A9 prior to r2p0) in the
architecture-specific cpu_relax() implementation.
In addition, cpu_relax() is also a compiler barrier. It is not
immediately obvious that the @op argument "function" will result in an
actual function call (e.g. in case of inlining).
Where a function call is a C sequence point, this is lost on inlining.
Therefore, with agressive enough optimization it might be possible for
the compiler to hoist the:
(val) = op(args);
"load" out of the loop because it doesn't see the value changing. The
addition of cpu_relax() would inhibit this.
As the iopoll helpers lack calls to cpu_relax(), people are sometimes
reluctant to use them, and may fall back to open-coded polling loops
(including cpu_relax() calls) instead.
Fix this by adding calls to cpu_relax() to the iopoll helpers:
- For the non-atomic case, it is sufficient to call cpu_relax() in
case of a zero sleep-between-reads value, as a call to
usleep_range() is a safe barrier otherwise. However, it doesn't
hurt to add the call regardless, for simplicity, and for similarity
with the atomic case below.
- For the atomic case, cpu_relax() must be called regardless of the
sleep-between-reads value, as there is no guarantee all
architecture-specific implementations of udelay() handle this.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Tony Lindgren <tony@atomide.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/45c87bec3397fdd704376807f0eec5cc71be440f.1685692810.git.geert+renesas@glider.be
Introduce a helper to return the SoC SMEM ID, which is used to identify the
exact SoC model as there may be differences in the same SoC family.
Currently, cpufreq-nvmem does this completely in the driver and there has
been more interest expresed for other drivers to use this information so
lets expose a common helper to prevent redoing it in individual drivers
since this field is present on every SMEM table version.
Signed-off-by: Robert Marko <robimarko@gmail.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20230526204802.3081168-3-robimarko@gmail.com
