Lockdep reported a possible deadlock between the cpuclk lock and for example
the i2c driver.
CPU0 CPU1
---- ----
lock(clk_lock);
local_irq_disable();
lock(&(&i2c->lock)->rlock);
lock(clk_lock);
<Interrupt>
lock(&(&i2c->lock)->rlock);
*** DEADLOCK ***
The generic clock-types of the core ccf already use spin_lock_irqsave when
touching clock registers, so do the same for the cpuclk.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
[mturquette@linaro.org: removed initialization of "flags"]
Commit 0e5bdb3f9f (clk: rockchip: switch to using the new cpuclk type
for armclk) didn't take into account that the divider used on rk3288
are of the (n+1) type.
The rk3066 and rk3188 socs use more complex divider types making it
necessary for the list-elements to be the real register-values to write.
Therefore reduce divider values in the table accordingly so that they
really are the values that should be written to the registers and match
the dividers actually specified for the rk3288.
Reported-by: Sonny Rao <sonnyrao@chromium.org>
Fixes: 0e5bdb3f9f ("clk: rockchip: switch to using the new cpuclk type for armclk")
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Cc: stable@vger.kernel.org
The bit locations indicating the locking status of the plls on rk3066 are
shifted by one to the right when compared to the rk3188, bits [7:4] instead
of [8:5] on the rk3188, thus indicating the locking state of the wrong pll
or a completely different information in case of the gpll.
The recently introduced pll init code exposed that problem on some rk3066
boards when it tried to bring the boot-pll value in line with the value
from the rate table.
Fix this by defining separate pll definitions for rk3066 with the correct
locking indices.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Fixes: 2c14736c75 ("clk: rockchip: add clock driver for rk3188 and rk3066 clocks")
Tested-by: FUKAUMI Naoki <naobsd@gmail.com>
Cc: stable@vger.kernel.org
Do not disable clock gate "hclk_emem_peri", otherwise EMAC clocks no longer work
and it breaks ethernet on RK3066 and RK3188. It fixes a regression introduced by
commit 78eaf6095c ("clk: rockchip: disable unused clocks").
Signed-off-by: Romain Perier <romain.perier@gmail.com>
Fixes: 78eaf6095c ("clk: rockchip: disable unused clocks")
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
This patch adds CLK_IGNORE_UNUSED flag to hclk_usb_peri, hclk_usbotg0
and hclk_usbotg1 because these clocks must remain enabled to use the
USB controllers in host mode.
This fixes a regression introduced by commit 78eaf6095c
("clk: rockchip: disable unused clocks").
Signed-off-by: Julien CHAUVEAU <julien.chauveau@neo-technologies.fr>
Fixes: 78eaf6095c ("clk: rockchip: disable unused clocks")
Reviewed-by: Romain Perier <romain.perier@gmail.com>
Tested-by: Romain Perier <romain.perier@gmail.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
- pll init to allow syncing to actual rate table values
- some more exported clocks
- fixes for some clocks (typos etc) all of them not yet used
in actual drivers
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABCAAGBQJUd7jIAAoJEPOmecmc0R2BrOkIAJpMe0tqHxtcN1fBSS2rS3/b
bgcu37WfGXt1FkBkGPyoVfy0m6tQ9hHx07o62IowVvaWzmvoQppLmC20+y6gKVMj
nVGlhbfhZ3bDJBbsbWAyzQvH+fJx+pOTIGQJCCNX/aixSYnfvda/FP0u7EVrY4W/
mF/b5mPmHhah9EeyuAwy4eAT5xCCER5DqR9eryObDwj3N9cUluM9BiK1JWXXWdYt
i/p3NS+MEjFG2KrLB+Rzu39yFRsjF8IOBKuYNh752NVcsNy2CuxqbDSpn8NM2niW
bxW07X5J8JSpecpuzJmkuOomauE+4pKXMM3W+XL+XO8iF3h9inXXp0nzX31cQRI=
=st2A
-----END PGP SIGNATURE-----
Merge tag 'v3.19-rockchip-clk2' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip into clk-next
- clock phase setting capability for the rk3288 mmc clocks
- pll init to allow syncing to actual rate table values
- some more exported clocks
- fixes for some clocks (typos etc) all of them not yet used
in actual drivers
This patch adds the 2 physical clocks for the mmc (drive and sample). They're
mostly there for the phase properties, but they also show the true clock
(by dividing by RK3288_MMC_CLKGEN_DIV).
The drive and sample phases are generated by dividing an upstream parent clock
by 2, this allows us to adjust the phase by 90 deg.
There's also an option to have up to 255 delay elements (40-80 picoseconds long).
This driver uses those elements (under the assumption that they're 60 ps long)
to generate approximate 22.5 degrees options. 67.5 (22.5*3) might be as high as
90 deg if the delay elements are as big as 80 ps, so a finer division (smaller
than 22.5) was not picked because the phase might not be monotonic anymore.
Suggested-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Alexandru M Stan <amstan@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
This exposes the clock that comes out of the i2s block which generally
goes to the audio codec.
Signed-off-by: Sonny Rao <sonnyrao@chromium.org>
[removed CLK_SET_RATE_PARENT from original patch]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The DMC clocks need to be turned off at runtime. Use the newly
assigned clock IDs to export them.
Signed-off-by: Jeff Chen <cym@rock-chips.com>
[dianders: split into two patches; adjusted commit msg]
Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Add the new flag to gpll and cpll on rk3188 and similar and to
gpll, cpll and npll on rk3288.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
In some cases firmware brings up plls with different parameters than the ones
noted in the rate table for the specific frequency. These firmware-selected
parameters are worse than the tested ones in the pll rate tables but cannot
be changed by a simple clk_set_rate call when the rate stays the same.
Therefore add a ROCKCHIP_PLL_SYNC_RATE flag and implement an init callback
that checks the runtime-parameters against the matching rate table entry
and adjusts them to the table-ones if necessary.
If no rate table is set or the current rate does not match any rate-table
entry no changes are made.
Being able to limit this adjustment to specific plls is necessary to not
touch the ones supplying core components like the apll and dpll supplying
the armcores and dram.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
In some cases we might need to access the data of the pll mux before the actual
mux gets registered - like in the following patch adding an init-callback.
Therefore populate pll_mux before registering the core pll-clock.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
This adds a flag parameter to plls that allows us to create
special flags to tweak the behaviour of the plls if necessary.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
The USB HSIC PHY clock divider is set in the register RK2928_CLKSEL_CON(11).
Signed-off-by: Julien CHAUVEAU <julien.chauveau@neo-technologies.fr>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
In rk3188 clock branches, spdif_pre gate was set to RK2928_CLKGATE_CON(13) bit 13.
This appears to be a copy-paste error because such a register does not exist.
We correct it to RK2928_CLKGATE_CON(0) and find out that the rk3188 spdif clock
is the same as the rk3066 spdif clock, so we move it to the common clock branches.
Signed-off-by: Julien CHAUVEAU <julien.chauveau@neo-technologies.fr>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The parent clock for hclk_lcdc1 was set to aclk_cpu instead of hclk_cpu.
Signed-off-by: Julien CHAUVEAU <julien.chauveau@neo-technologies.fr>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Commit 79c6ab5095 (clk: divider: add CLK_DIVIDER_READ_ONLY flag) in
v3.16 introduced the CLK_DIVIDER_READ_ONLY flag which caused the
recalc_rate() and round_rate() clock callbacks to be omitted.
However using this flag has the unfortunate side effect of causing the
clock recalculation code when a clock rate change is attempted to always
treat it as a pass-through clock, i.e. with a fixed divide of 1, which
may not be the case. Child clock rates are then recalculated using the
wrong parent rate.
Therefore instead of dropping the recalc_rate() and round_rate()
callbacks, alter clk_divider_bestdiv() to always report the current
divider as the best divider so that it is never altered.
For me the read only clock was the system clock, which divided the PLL
rate by 2, from which both the UART and the SPI clocks were divided.
Initial setting of the UART rate set it correctly, but when the SPI
clock was set, the other child clocks were miscalculated. The UART clock
was recalculated using the PLL rate as the parent rate, resulting in a
UART new_rate of double what it should be, and a UART which spewed forth
garbage when the rate changes were propagated.
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Thomas Abraham <thomas.ab@samsung.com>
Cc: Tomasz Figa <t.figa@samsung.com>
Cc: Max Schwarz <max.schwarz@online.de>
Cc: <stable@vger.kernel.org> # v3.16+
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
According to rk3288 trm, the mux selector locate at bit[12:11]
of CRU_CLKSEL13_CON shows:
2'b00: select HOST0 USB pll clock (clk_otgphy1)
2'b01: select HOST1 USB pll clock (clk_otgphy2)
2'b10: select OTG USB pll clock (clk_otgphy0)
The clock map is in Fig. 3-4 CRU Clock Architecture Diagram 3
- clk_otgphy0 -> USB PHY OTG
- clk_otgphy1 -> USB PHY host0
- clk_otgphy2 -> USB PHY host1
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
According to rk3288 trm, the clk_usbphy480m_gate is located at
bit 14 of CRU_CLKGATE5_CON register.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Currently there is no driver owning these clocks and they have to stay
up for the system to function properly, so let's mark them as
CLK_IGNORE_UNUSED.
Without this patch we have trouble with suspend/resume and we have
trouble turning the eDP back on if it ever idles off.
Signed-off-by: Dmitry Torokhov <dtor@chromium.org>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
save and restore some clks, which might be changed in suspend.
Signed-off-by: Tony Xie <xxx@rock-chips.com>
Signed-off-by: Chris Zhong <zyw@rock-chips.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The arguments to COMPOSITE_FRAC for hsadc_frac were mangled, leaving out the
the general clock flags argument. This results in strange effects, as only
sometimes a zero-division is reported as the wrong register is read.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The rockchip clock driver use CLK_IGNORE_UNUSED flag to make sure
all the clocks are available like default power on state.
We have implement the clock manage in most of rockchip drivers,
it is time to remove it for power save.
Instead we add CLK_IGNORE_UNUSED for some clock nodes which should
be on during boot or no module driver in kernel will initialize it.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
dclk_vop0/1 is the source of HDMI TMDS clock in rk3288, usually we
use 594MHz for clock source of dclk_vop0/1.
HDMI CTS 7-9 require TMDS Clock jitter is lower than 0.25*Tbit:
TMDS clock(MHz) CTS require jitter (ps)
297 84.2
148.5 168
74.25 336
27 1247
PLL BW and VCO frequency effects the jitter of PLL output clock,
clock jitter is better if BW is lower or VCO frequency is higher.
If PLL use default setting of RK3066_PLL_RATE( 594000000, 2, 198, 4),
the TMDS Clock jitter is higher than 250ps, which means we can't
pass the test when TMDS clock is 297MHz or 148.5MHz.
If we use RK3066_PLL_RATE_BWADJ(594000000, 1, 198, 8, 1),
the TMDS Clock jitter is about 60ps and we can pass all test case.
So we need this patch to make hdmi si test pass.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The possible sources for the rk3288-gpu-clock also include the npll,
making it the same list of sources as for uart0.
This patch make a common source for uart0 pll src and sclk_gpu,
so that gpu can get its clock from npll.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Removing the CLK_SET_RATE_PARENT from i2s_clkout, to limit i2s0_clkout
to select between its two parent without being able influence the core
i2s clock.
Tested on rk3288 board, suggested by Heiko.
Signed-off-by: Jianqun <jay.xu@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
This patch add 400MHz and 500MHz to clock rate table for rk3288.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
We'd like to be able to call clk_set_rate() on aclk_cpu (a gate) at
bootup. In order for this to have any effect we need its parent
(aclk_cpu_pre) to percolate the rate change to _its_ parent
(aclk_cpu_src). Add CLK_SET_RATE_PARENT to make this happen.
Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The parent should be spdif_8ch_pre not spdif_8ch_src, which doesn't
exist and looks to be a typo. The TRM also confirms this.
Signed-off-by: Sonny Rao <sonnyrao@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Add infrastructure to write the correct value to the restart register and
register the restart notifier for both rk3188 (including rk3066) and rk3288.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The relation of i2s nodes as follows:
i2s_src 0 0 594000000 0
i2s_frac 0 0 11289600 0
i2s_pre 0 0 11289600 0
sclk_i2s0 0 0 11289600 0
i2s0_clkout 0 0 11289600 0
hclk_i2s0 1 1 99000000 0
sclk_i2s0 is the master clock, when to set rate of sclk_i2s0, should
allow to set its parent's rate, by add flag CLK_SET_RATE_PARENT for
"i2s_frac", "i2s_pre", "i2s0_clkout" and "sclk_i2s0".
Tested on rk3288 board using max98090, with command "aplay <music.wav>"
Change-Id: I12faad082566532b65a7de8c0a6845e1c17870e6
Signed-off-by: Jianqun <jay.xu@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
This adds the necessary soc-specific divider values and switches the armclk
to use the newly introduced cpuclk type.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
When changing the armclk on Rockchip SoCs it is supposed to be reparented
to an alternate parent before changing the underlying pll and back after
the change. Additionally there exist clocks that are very tightly bound to
the armclk whose divider values are set according to the armclk rate.
Add a special clock-type to handle all that. The rate table and divider
values will be supplied from the soc-specific clock controllers.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
On a rk3288-board:
Tested-by: Doug Anderson <dianders@chromium.org>
Rockchip SoCs contain clocks tightly bound to the armclk, where the best
rate / divider is supplied by the vendor after careful measuring.
Often this ideal rate may be greater than the current rate.
Therefore prevent the ccf from trying to set these dividers itself by
setting them to read-only.
In the case of the rk3066, this also includes the aclk_cpu, which makes it
necessary to also split its direct child-clocks (pclk_cpu, hclk_cpu, ...)
into individual definitions for rk3066 and rk3188.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
aclk_cpu_pre on the rk3188 can either be sourced from the armclk or the gpll.
To reduce complexity on apll changes caused by cpufreq, reparent it always
to the gpll source.
If really necessary it could be reparented back on a per board level using
the assigned-clocks mechanism.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
In RK3288, APLL lock status bit is in GRF_SOC_STATUS1,
but in RK3188, is GRFSOC_STATUS0.
Signed-off-by: Jianqun <jay.xu@rock-chips.com>
Also name the constant accordingly as GRF_SOC_STATUS1
to prevent confusion.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
The register providing the pll lock status is at a different address on the
rk3066. The error became apparent while working on cpufreq support for
the rockchip SoCs.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The Rockchip PLL code switches into slow mode (AKA bypass more AKA
24MHz mode) before actually changing the PLL. This keeps anyone from
using the PLL while it's changing. However, in all known Rockchip
SoCs nobody should ever see the 24MHz when changing the PLL supplying
the armclk because we should reparent children to an alternate
(faster than 24MHz) PLL.
One problem is that the code to switch to an alternate parent was
running in PRE_RATE_CHANGE. ...and the code to switch to slow mode
was _also_ running in PRE_RATE_CHANGE. That meant there was no real
guarantee that we would switch to an alternate parent before switching
to 24MHz mode.
Let's move the switch to "slow mode" straight into
rockchip_rk3066_pll_set_rate(). That means we're guaranteed that the
24MHz is really a last-resort.
Note that without this change on real systems we were the code to
switch to an alternate parent at 24MHz. In some older versions of
that code we'd appy a (temporary) / 5 to the 24MHz causing us to run
at 4.8MHz. That wasn't enough to service USB interrupts in some cases
and could lead to a system hang.
Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
This patch add the clock node in PD_VIDEO
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
This patch use the new defined clock ID to initial the clock nodes.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
The npll on rk3288 is exactly the same pll type as the other 4. Yet it
was missing the link to the rate table, making rate changes impossible.
Change that by setting the table.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
The rk3288 actually has 12 softresets, so fix the register count.
Signed-off-by: Mark yao <mark.yao@rock-chips.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
The dwc2 usb controller also uses agressive clock gating, which in this
case leads to hclk_peri getting disabled and hanging the system.
Therefore move it to the critical clocks until we also control that
part of the system.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
The clocks for i2c1 and i2c2 are flipped. The clock tree matched the
Technical Reference Manual (TRM) but the TRM was wrong. Swap them in
the clock tree. This was determined experimentally (by Addy) and
confirmed by the Rockchip IC team.
Signed-off-by: Doug Anderson <dianders@chromium.org>
Reported-by: Addy Ke <addy.ke@rock-chips.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
The clock-tree contains clocks that should never get disabled automatically.
One example are the base ACLKs, the base supplies for all peripherals.
Therefore add a structure similar to the sunxi clock-tree to protect these
special clocks from being disabled.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Doug Anderson <dianders@chromium.org>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
It is only used locally in clk/rockchip/clk.c and thus can be static.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
Rockchip SoCs may provide fraction dividers for some clocks, mostly for
i2s and uarts. In contrast to the other registers, these do not use
the hiword-mask paradigm, but instead split the register into the upper
16 bit for the nominator and the lower 16 bit for the denominator.
The common clock framework got a generic fractional divider clock type
recently that can accomodate this setting easily. All currently known
fraction dividers have a separate gate too, therefore implement the
divider as composite using the ops-struct from fractional_divider clock
and add the gate if necessary.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
Heiko Stübner