[PATCH v5 0/3] Broadcom STB AVS CPUfreq driver

From: Markus Mayer
Date: Thu Oct 27 2016 - 17:14:19 EST


From: Markus Mayer <mmayer@xxxxxxxxxxxx>

This series contains the CPUfreq driver for Broadcom SoCs that use "AVS
Firmware" for voltage and frequency scaling. All voltage and frequency
transitions are performed by the firmware and are therefore hidden from
Linux.

The driver provides a standard CPUfreq interface to other kernel
components and to userland on the one hand and communicates with the
AVS co-processor on the other.

Communication between the two processors is via shared mailbox
registers and interrupts (ARM -> AVS to tell the firmware that there is
a command to process and AVS -> ARM to tell the driver that a command
finished executing).

Changes from v4:
- Addressed comments regarding the binding document (see also
https://lkml.org/lkml/2016/10/19/642)
- Added Virsh's Acks
- Rebased on v4.9-rc1
- Patches sent from non-Broadcom account on purpose to prevent series
from being marked as spam (this is currently happening for some
recipients)
- No changes to any code

Changes from v3:
- moved code from brcm_avs_cpufreq_exit() into brcm_avs_cpufreq_remove()
- removed brcm_avs_cpufreq_exit() altogether

Changes from v2:
- removed file Documentation/cpu-freq/brcmstb-avs-cpufreq.txt, instead
adding the text as comment in the driver itself
- wildcard match in MAINTAINERS file (drivers/cpufreq/brcmstb*)
- drivers/cpufreq/Kconfig.arm: have ARM_BRCMSTB_AVS_CPUFREQ default y
for ARCH_BRCMSTB
- fixed naming and Kconfig entry for CONFIG_ARM_BRCMSTB_AVS_CPUFREQ_DEBUG
- reordered and fixed include files
- made sure driver name doesn't exceed size of cpufreq_driver.name
- got rid of device_node pointers in struct private_data; there's no
need for them, as we can call of_node_put() in __map_region()
- don't set driver_data for CPUFREQ_TABLE_END
- have brcm_avs_cpufreq_get() read from the firmware rather than return
policy->cur
- don't update policy->cur in brcm_avs_target_index()
- created new function brcm_avs_prepare_init() to handle compatibility
checks during initialization and moved corresponding code from
brcm_avs_cpufreq_init() to the new function
- brcm_avs_prepare_init() is called from brcm_avs_cpufreq_probe() and
therefore will only be called once, even in the case of an error,
whereas brcm_avs_cpufreq_init() is called for every core if there is
an error
- renamed brcm_avs_cpu_(init|exit) to brcm_avs_cpufreq_(init|exit)
- added platform_set_drvdata(pdev, NULL) upon deregistration
- fixed various style complaints by checkpatch.pl --strict
- fixed other style issues
- updated and improved some comments
- added Viresh's Acks to patches 1/3 and 3/3

Changes from v1:
- renamed binding document
- rewrote the introduction of the binding document
- created a new driver documentation file that contains Linux specific
information that was previously part of the binding document
- renamed the driver (and related config options) to include a reference
to "STB", since this implementation is primarily intended for use on
set-top boxes
- improved comments
- updated function __map_region()
- updated struct private_data
- added code to unmap memory regions in the error and exit paths
- added new sysfs property to report frequency directly from the
co-processor register

Markus Mayer (3):
dt: cpufreq: brcm: New binding document for brcmstb-avs-cpufreq
cpufreq: brcmstb-avs-cpufreq: AVS CPUfreq driver for Broadcom STB SoCs
cpufreq: brcmstb-avs-cpufreq: add debugfs support

.../bindings/cpufreq/brcm,stb-avs-cpu-freq.txt | 78 ++
MAINTAINERS | 8 +
drivers/cpufreq/Kconfig.arm | 21 +
drivers/cpufreq/Makefile | 1 +
drivers/cpufreq/brcmstb-avs-cpufreq.c | 1057 ++++++++++++++++++++
5 files changed, 1165 insertions(+)
create mode 100644 Documentation/devicetree/bindings/cpufreq/brcm,stb-avs-cpu-freq.txt
create mode 100644 drivers/cpufreq/brcmstb-avs-cpufreq.c

--
2.7.4