linux-toradex.git/drivers/clk/pxa, branch v4.11

Merge tag 'armsoc-soc' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc

2016-12-15T23:39:02+00:00

Pull ARM SoC platform updates from Arnd Bergmann:
 "These are updates for platform specific code on 32-bit ARM machines,
  essentially anything that can not (yet) be expressed using DT files.

  Noteworthy changes include:

   - Added support for the TI DRA71x family of SoCs in mach-omap2, this
     is an new variant of the the DRA72x/DRA74x automotive infotainment
     chips we already supported for a while.

   - Added support for the ST STM32F746 SoC, the first Cortex-M7 based
     microcontroller we support, related to the smaller STM32F4 family.

   - Renesas adds support for r8a7743 and r8a7745 in mach-shmobile, see
     http://elinux.org/RZ-G

   - SMP is now supported on the OX820 platform

   - A lot of code in mach-omap2 gets removed as a follow-up to removing
     support for board files in the previous release

   - Davinci has some new work to improve USB support

   - For i.MX, the performance monitor now supports profiling the memory
     controller using 'perf'"

* tag 'armsoc-soc' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (95 commits)
  ARM: davinci: da830-evm: use gpio descriptor for mmc pins
  ARM: davinci: da850-evm: use gpio descriptor for mmc pins
  ARM: davinci: hawk: use gpio descriptor for mmc pins
  ARM: ARTPEC-6: add select MFD_SYSCON to MACH_ARTPEC6
  ARM: davinci: da8xx: Fix ohci device name
  ARM: oxnas: Add OX820 config and makefile entry
  ARM: oxnas: Add OX820 SMP support
  ARM: davinci: PM: fix build when da850 not compiled in
  ARM: orion5x: remove legacy support of ls-chl
  ARM: integrator: drop EBI access use syscon
  ARM: BCM5301X: Add back handler ignoring external imprecise aborts
  ARM: davinci: PM: support da8xx DT platforms
  ARM: davinci: PM: cleanup: remove references to pdata
  ARM: davinci: PM: rework init, remove platform device
  ARM: Kconfig: Introduce MACH_STM32F746 flag
  ARM: mach-stm32: Add a new SOC - STM32F746
  ARM: shmobile: document SK-RZG1E board
  ARM: shmobile: r8a7745: basic SoC support
  ARM: imx: mach-imx6ul: add imx6ull support
  ARM: zynq: Reserve correct amount of non-DMA RAM
  ...

clk: pxa: Use __iomem properly and staticize lock variable

2016-11-09T20:43:19+00:00

This function is passed an __iomem pointer but we use a u32
pointer instead which makes checkers like spare complain.
Furthermore, "lock" is a pretty poor variable name for a string
that will go into lockdep reports and the symbol isn't marked
static. Cleanup all this.

Acked-by: Robert Jarzmik 
Signed-off-by: Stephen Boyd

clk: pxa: fix pxa2xx_determine_rate return

2016-11-09T20:04:33+00:00

The new pxa2xx_determine_rate() function seems lacking in a few
regards:

- For an exact match or no match at all, the rate is uninitialized
  as reported by gcc -Wmaybe-unintialized:
   drivers/clk/pxa/clk-pxa.c: In function 'pxa2xx_determine_rate':
   drivers/clk/pxa/clk-pxa.c:243:5: error: 'rate' may be used uninitialized in this function

- If we get a non-exact match, the req->rate output is never set
  to the actual rate but remains at the requested rate.

- We should not attempt to print a rate if none could be found

This rewrites the logic accordingly.

Fixes: 9fe694295098 ("clk: pxa: transfer CPU clock setting from pxa2xx-cpufreq")
Signed-off-by: Arnd Bergmann 
Acked-by: Robert Jarzmik 
Signed-off-by: Stephen Boyd

clk: pxa mark dummy helper as 'inline'

2016-11-08T22:46:56+00:00

The dummy_clk_set_parent function is marked as 'static' but is
no longer referenced from the pxa25x clk driver after the last use
of the RATE_RO_OPS() macro is gone from this file, causing a
harmless build warning:

In file included from drivers/clk/pxa/clk-pxa25x.c:24:0:
drivers/clk/pxa/clk-pxa.h:146:12: error: 'dummy_clk_set_parent' defined but not used [-Werror=unused-function]

This marks the functon as 'inline', which lets the compiler simply
drop it when it gets referenced.

Fixes: 9fe694295098 ("clk: pxa: transfer CPU clock setting from pxa2xx-cpufreq")
Signed-off-by: Arnd Bergmann 
Acked-by: Robert Jarzmik 
Signed-off-by: Stephen Boyd

clk: pxa: transfer CPU clock setting from pxa2xx-cpufreq

2016-11-04T20:25:03+00:00

This is the initial stage to transfer the pxa25x and pxa27x CPU clocks
handling from cpufreq to the clock API. More precisely, the clocks
transferred are :
 - cpll : core pll, known also as the CPU core turbo frequency
 - core : core, known also as the CPU actual frequency, being either the
          CPU core turbo frequency or the CPU core run frequency

This transfer is a prequel to shrink the code in pxa2xx-cpufreq.c, so
that it can become, at least in devicetree builds, the casual cpufreq-dt
driver.

Signed-off-by: Robert Jarzmik 
Signed-off-by: Stephen Boyd

clk: pxa: export core clocks

2016-11-02T00:47:01+00:00

pxaxxx_get_clk_frequency_khz() needs several clocks to be available
through clk_get(), ie. the cpu clocks, system bus clock and memory
clocks.

Add the missing clkdev so that their rate can be acquired.

Signed-off-by: Robert Jarzmik 
Signed-off-by: Stephen Boyd

clk: pxa: b bit of clkcfg means fast bus

2016-11-02T00:46:53+00:00

The meaning of this bit was inverted :
 - when set to 0, system bus clock is half of the CPU run clock
 - when set to 1, system bus clock is the CPU run clock

Signed-off-by: Robert Jarzmik 
Signed-off-by: Stephen Boyd

clk: pxa: core pll is not affected by t bit

2016-11-02T00:46:49+00:00

The t bit of clkfcfg doesn't affect the core pll clock, but it makes core
clock select between core pll clock and core run clock.

As such remove it from the core pll rate reporting function, while it
remains in clk_pxa27x_core_get_parent().

Signed-off-by: Robert Jarzmik 
Signed-off-by: Stephen Boyd

clk: pxa: remove unused variables

2016-11-02T00:46:46+00:00

This is a cleanup patch to remove unused values not used in their
respective functions.

Signed-off-by: Robert Jarzmik 
Signed-off-by: Stephen Boyd

clk: pxa25x: OSTIMER0 clocks from the main oscillator

2016-10-18T07:12:36+00:00

The OSTIMER0 clock ticks at the main oscillator rate, not the 32kHz
oscillator rate.  Ensure that it is parented to the main oscillator.

Signed-off-by: Russell King 
Acked-by: Stephen Boyd 
Tested-by: Robert Jarzmik 
Signed-off-by: Robert Jarzmik