Pull clk driver fixes from Stephen Boyd:
- Make the regulator state match the GDSC power domain state at boot on
Qualcomm SoCs so that the regulator isn't turned off inadvertently.
- Fix earlycon on i.MX6Q SoCs
* tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux:
clk: qcom: gdsc: Ensure regulator init state matches GDSC state
clk: imx6q: fix uart earlycon unwork
At least one user currently duplicates some functions that are provided
by fractional divider module. Let's export approximation algorithm and
replace the open-coded variant.
As a bonus the exported function will get better documentation in place.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Tested-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Heiko Stuebner <heiko@sntech.de>
Link: https://lore.kernel.org/r/20210812170025.67074-1-andriy.shevchenko@linux.intel.com
[sboyd@kernel.org: Add header guard because why not]
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
We're moving pistachio to generic MIPS kernel. The clk driver
should be avilable to the generic MIPS kernel.
Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Acked-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
FUSE driver now takes care of keeping the clock enabled when necessary.
Remove the CLK_IS_CRITICAL flag from the clock.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
There was an "unknown" firmware variant turning up in the wild
causing problems in the clock driver. Add this missing variant
and clarify that varian 11 and 15 are Samsung variants, as this
is now very well known from released products.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
MSM8939 has 3 a53pll clocks with different frequency table for Cluster0,
Cluster1 and CCI. It adds function qcom_a53pll_get_freq_tbl() to create
pll_freq_tbl from OPP, so that those a53pll frequencies can be defined
in DT with operating-points-v2 bindings rather than being coded in the
driver. In this case, one compatible rather than three would be needed
for these 3 a53pll clocks.
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Link: https://lore.kernel.org/r/20210704024032.11559-5-shawn.guo@linaro.org
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Different from MSM8916 which has only one a53pll/mux clock, MSM8939 gets
three for Cluster0 (little cores), Cluster1 (big cores) and CCI (Cache
Coherent Interconnect). That said, a53pll/mux clock needs to be named
uniquely. Append @unit-address of device node to the clock name, so
that a53pll/mux will be named like below on MSM8939.
a53pll@b016000
a53pll@b116000
a53pll@b1d0000
a53mux@b1d1000
a53mux@b011000
a53mux@b111000
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Link: https://lore.kernel.org/r/20210704024032.11559-3-shawn.guo@linaro.org
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
The clock source for MSM8916 cpu cores is like below.
|\
a53pll --------| \ a53mux +------+
| |------------| cpus |
gpll0_vote --------| / +------+
|/
So a53mux rather than a53pll is actually the parent clock of cpu cores.
It makes more sense to flag a53mux as critical instead, so that when
either a53pll or gpll0_vote is used by cpu cores, the clock will be kept
enabled while the other can be disabled.
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Link: https://lore.kernel.org/r/20210704024032.11559-2-shawn.guo@linaro.org
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
As GDSCs are registered and found to be already enabled gdsc_init()
ensures that 1) the kernel state matches the hardware state, and 2)
votable GDSCs are properly enabled from this master as well.
But as the (optional) supply regulator is enabled deep into
gdsc_toggle_logic(), which is only executed for votable GDSCs, the
kernel's state of the regulator might not match the hardware. The
regulator might be automatically turned off if no other users are
present or the next call to gdsc_disable() would cause an unbalanced
regulator_disable().
Given that the votable case deals with an already enabled GDSC, most of
gdsc_enable() and gdsc_toggle_logic() can be skipped. Reduce it to just
clearing the SW_COLLAPSE_MASK and enabling hardware control to simply
call regulator_enable() in both cases.
The enablement of hardware control seems to be an independent property
from the GDSC being enabled, so this is moved outside that conditional
segment.
Lastly, as the propagation of ALWAYS_ON to GENPD_FLAG_ALWAYS_ON needs to
happen regardless of the initial state this is grouped together with the
other sc->pd updates at the end of the function.
Cc: stable@vger.kernel.org
Fixes: 37416e5549 ("clk: qcom: gdsc: Handle GDSC regulator supplies")
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20210721224056.3035016-1-bjorn.andersson@linaro.org
[sboyd@kernel.org: Rephrase commit text]
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
.determine_rate is meant to replace .round_rate. The former comes with a
benefit which is especially relevant on 32-bit systems: since
.determine_rate uses an "unsigned long" (compared to a "signed long"
which is used by .round_rate) the maximum value on 32-bit systems
increases from 2^31 (or approx. 2.14GHz) to 2^32 (or approx. 4.29GHz).
Implement .determine_rate in addition to .round_rate so drivers that are
using clk_divider_{ro_,}ops can benefit from this by default. Keep the
.round_rate callback for now since some drivers rely on
clk_divider_ops.round_rate being implemented.
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Link: https://lore.kernel.org/r/20210702225145.2643303-2-martin.blumenstingl@googlemail.com
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Pull clk fixes from Stephen Boyd:
"A collection of clk driver fixes and one core clk API fix:
- Fix stm32 clk data to avoid a crash early on
- Fix a randconfig build error in HiSilicon clk driver
- Avoid an oops at boot on Qualcomm MSM8936 SoCs due to
an improper consolidation of structs
- Fix imbalanced disabling of the unused MMC clock on
Tegra210 Jetson Nano
- Plug a memory leak in devm_clk_bulk_get_all() unwind
path"
* tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux:
clk: fix leak on devm_clk_bulk_get_all() unwind
clk: tegra: Implement disable_unused() of tegra_clk_sdmmc_mux_ops
clk: qcom: smd-rpm: Fix MSM8936 RPM_SMD_PCNOC_A_CLK
clk: hisilicon: hi3559a: select RESET_HISI
clk: stm32f4: fix post divisor setup for I2S/SAI PLLs
Currently, no driver support for DDR memory controller (DMC) is present,
as a result, no driver is explicitly consuming the ddrphy clock. This means
that VPLL1 (parent of ddr clock) will be shutdown if we enable
and then disable any child clock of VPLL1 (e.g. SCLK_I2S0_8CH_TX).
If VPLL1 is disabled, the whole system will freeze, because the DDR
controller will lose its clock. So, it's necessary to prevent VPLL1 from
shutting down, by marking the ddrphy4x CLK_IS_CRITICAL.
This bug was discovered when I was porting rockchip_i2s_tdm driver to
mainline kernel from Rockchip 4.4 kernel. I guess that other Rockchip
SoCs without DMC driver may need the same patch. If this applies to
other devices, please let us know.
Signed-off-by: Yunhao Tian <t123yh@outlook.com>
Link: https://lore.kernel.org/r/BYAPR20MB24886765F888A9705CBEB70789E39@BYAPR20MB2488.namprd20.prod.outlook.com
[adapted subject, changed to add the clock to the critical list]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The rk3036/rk3328 pll types were converted to checking the lock status
via the internal register in january 2020, so don't need the grf
reference since then.
But it was forgotten to remove grf check when deciding between the
pll rate ops (read-only vs. read-write), so a clock driver without
the needed grf reference might've been put into the read-only mode
just because the grf reference was missing.
This affected the rk356x that needs to reclock certain plls at boot.
Fix this by removing the check for the grf for selecting the utilized
operations.
Suggested-by: Heiko Stuebner <heiko@sntech.de>
Fixes: 7f6ffbb885 ("clk: rockchip: convert rk3036 pll type to use internal lock status")
Signed-off-by: Peter Geis <pgwipeout@gmail.com>
[adjusted the commit message, adjusted the fixes tag]
Link: https://lore.kernel.org/r/20210728180034.717953-3-pgwipeout@gmail.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
I found that COMMON_CLK_MT8167* do not depend on COMMON_CLK_MT8167,
so it is possible to config:
CONFIG_COMMON_CLK_MT8167=n
CONFIG_COMMON_CLK_MT8167_*=y
Although it does not cause build breaks with such configuration,
I think it is clearer to make COMMON_CLK_MT8167* depend on
COMMON_CLK_MT8167.
Signed-off-by: Miles Chen <miles.chen@mediatek.com>
Link: https://lore.kernel.org/r/20210716051732.3422-1-miles.chen@mediatek.com
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
