The critical clock handling in __clk_core_init isn't taking the enable lock
before calling clk_core_enable, which in turns triggers the warning in the
lockdep_assert_held call in that function when lockep is enabled.
Add the calls to clk_enable_lock/unlock to make sure it doesn't happen.
Fixes: 32b9b10961 ("clk: Allow clocks to be marked as CRITICAL")
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Unlike devm_clk_register(), devm_clk_hw_register() returns integer.
So, the statement "Clocks returned from this function ..." sounds
odd. Adjust the comment for this new API.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Now that we have a clk registration API that doesn't return
struct clks, we need to have some way to hand out struct clks via
the clk_get() APIs that doesn't involve associating struct clk
pointers with an OF node. Currently we ask the OF provider to
give us a struct clk pointer for some clkspec, turn that struct
clk into a struct clk_hw and then allocate a new struct clk to
return to the caller.
Let's add a clk_hw based OF provider hook that returns a struct
clk_hw directly, so that we skip the intermediate step of
converting from struct clk to struct clk_hw. Eventually when
we've converted all OF clk providers to struct clk_hw based APIs
we can remove the struct clk based ones.
It should also be noted that we change the onecell provider to
have a flex array instead of a pointer for the array of clk_hw
pointers. This allows providers to allocate one structure of the
correct length in one step instead of two.
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
We've largely split the clk consumer and provider APIs along
struct clk and struct clk_hw, but clk_register() still returns a
struct clk pointer for each struct clk_hw that's registered.
Eventually we'd like to only allocate struct clks when there's a
user, because struct clk is per-user now, so clk_register() needs
to change.
Let's add new APIs to register struct clk_hws, but this time
we'll hide the struct clk from the caller by returning an int
error code. Also add an unregistration API that takes the clk_hw
structure that was passed to the registration API. This way
provider drivers never have to deal with a struct clk pointer
unless they're using the clk consumer APIs.
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
This call matches clocks which have been marked as critical in DT
and sets the appropriate flag. These flags can then be used to
mark the clock core flags appropriately prior to registration.
Legacy bindings requiring this feature must add the clock-critical
property to their binding descriptions, as it is not a part of
common-clock binding.
Cc: devicetree@vger.kernel.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/1455225554-13267-4-git-send-email-mturquette@baylibre.com
Critical clocks are those which must not be gated, else undefined
or catastrophic failure would occur. Here we have chosen to
ensure the prepare/enable counts are correctly incremented, so as
not to confuse users with enabled clocks with no visible users.
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/1455225554-13267-2-git-send-email-mturquette@baylibre.com
Russell King recently pointed out a bug in the clk-gpio code
where it fails to register the clk if of_clk_get_parent_count()
returns an error because the "clocks" property isn't present in
the DT node. If we're trying to count parents from DT we'd like
to know the count, not if there is a "clocks" property or not.
Furthermore, some drivers are assigning the return value to their
clk_init_data::num_parents member which is unsigned, leading to
potentially large numbers of parents when the property isn't
present.
Let's change the API to return an unsigned int instead of an int.
All the callers just want to know the count anyway, and this
avoids the bug that was in the clk-gpio driver.
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
of_clk_init() uses for_each_matching_node_and_match() to find clock
providers, which returns all matching device nodes, whether they are
enabled or not. Hence clock providers that are disabled explicitly in DT
using e.g.
"status = "disabled";
are still activated.
Add a check to ignore device nodes that are not enabled, like
of_irq_init() does.
Reported-by: Ramesh Shanmugasundaram <ramesh.shanmugasundaram@bp.renesas.com>
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Let's compare the degrees from clk_set_rate with
clk->core->phase. If the requested degrees is already
there, skip the following steps.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
[sboyd@codeaurora.org: s/drgrees/degrees/ in commit text]
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
__clk_init() was renamed to __clk_core_init() but these comments
weren't updated.
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
This reverts commit 858d588156.
Joachim reports that this commit breaks lpc18xx boot. This is
because the hardware has circular clk topology where PLLs can
feed into dividers and the same dividers can feed into the PLLs.
The hardware is designed this way so that you can choose to put
the divider before the PLL or after the PLL depending on what you
configure to be the parent of the divider and what you configure
to be the parent of the PLL.
So let's drop this patch for now because we have hardware that
actually has loops. A future patch could check for circular
parents when we change parents and fail the switch, but that's
probably best left to some debugging Kconfig option so that we
don't suffer the sanity checking cost all the time.
Reported-by: Joachim Eastwood <manabian@gmail.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Export symbol of_clk_get_from_provider so it can be used in
loadable kernel modules
Signed-off-by: Andrew F. Davis <afd@ti.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Before commit b3d192d5121f ("clk: simplify __clk_init_parent()"),
__clk_init_parent() called .get_parent() only for multi-parent
clocks. That commit changed the behavior to call .get_parent()
if available even for single-parent clocks and root clocks.
It turned out a problem because there are some single-parent clocks
that implement .get_parent() callback and return non-zero index.
The SOCFPGA clock is the case; the commit broke the SOCFPGA boards.
To keep the original behavior, invoke .get_parent() only when
num_parents is greater than 1.
Fixes: b3d192d5121f ("clk: simplify __clk_init_parent()")
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reported-by: Dinh Nguyen <dinguyen@opensource.altera.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
We don't use CLK_IS_ROOT but in a few places in the common clk
framework core. Let's replace those checks with a check for the
number of parents a clk has instead of the flag, freeing up one
flag for something else. We don't remove the flag yet so that
things keep building, but we'll remove it once all drivers have
removed their flag usage.
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
If clock is already unregistered, it returns with holding lock.
It needs to be unlocked.
Signed-off-by: Insu Yun <wuninsu@gmail.com>
[sboyd@codeaurora.org: Use goto instead]
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
If clk_fetch_parent_index() fails, p_rate is unused. Move the
assignment after the error checking.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
The clk_core_get_parent_by_index can be used as a helper function
to simplify the implementation of clk_fetch_parent_index().
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
If parent is given with NULL, clk_fetch_parent_index() could return
a positive index value.
Currently, parent is checked by the callers of this function, but
it would be safer to do it in this function.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
This loop can be much simpler. If a new parent is available for
orphan clocks, __clk_init_parent(orphan) can detect it.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Currently, clk_register() never checks a circular parent looping,
but clock providers could register such an insane clock topology.
For example, "clk_a" could have "clk_b" as a parent, and vice versa.
In this case, clk_core_reparent() creates a circular parent list
and __clk_recalc_accuracies() calls itself recursively forever.
The core infrastructure should be kind enough to bail out, showing
an appropriate error message in such a case. This helps to easily
find a bug in clock providers. (uh, I made such a silly mistake
when I was implementing my clock providers first. I was upset
because the kernel did not respond, without any error message.)
This commit adds a new helper function, __clk_is_ancestor(). It
returns true if the second argument is a possible ancestor of the
first one. If a clock core is a possible ancestor of itself, it
would make a loop when it were registered. That should be detected
as an error.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
The translation from the index into clk_core is done by
clk_core_get_parent_by_index(). The if-block for num_parents == 1
case is duplicating the code in the clk_core_get_parent_by_index().
Drop the "if (num_parents == 1)" from the special case. Instead,
set the index to zero if .get_parent() is missing.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
The .get_parent is mandatory for multi-parent clocks. Move the check
to __clk_core_init(), like other callback checkings.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
[sboyd@codeaurora.org: Squashed in error path handling, fix typos
in commit message]
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
These three cases let clk_register() fail. They should be considered
as error messages.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
The core->parents is a cache to save expensive clock parent look-ups.
It will be filled as needed later. We do not have to do it here.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Drop the "if (!core->parents)" case and refactor the function a bit
because core->parents is always allocated. (Strictly speaking, it is
ZERO_SIZE_PTR if core->num_parents == 0, but such a case is omitted
by the if-conditional above.)
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Currently, __clk_core_init() allows failure of the kcalloc() for the
core->parents. So, clk_fetch_parent_index() and __clk_init_parent()
also try to allocate core->parents in case it has not been allocated
yet. Scattering memory allocation here and there makes things
complicated.
Like other clk_core members, allocate core->parents in clk_register()
and let it fail in case of memory shortage. If we cannot allocate
such a small piece of memory, the system is already insane. There is
no point to postpone the memory allocation.
Also, allocate core->parents regardless of core->num_parents. We want
it even if core->num_parents == 1 because clk_fetch_parent_index()
might be called against the clk_core with a single parent.
If core->num_parents == 0, core->parents is set to ZERO_SIZE_PTR. It
is harmless because no access happens to core->parents in such a case.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Now, the clock parent is not "struct clk *", but "struct clk_core *".
Of course, the size of a pointer is always same, but strictly speaking,
sizeof(struct clk *) should be sizeof(struct clk_core *) here.
This mismatch happened when we split the structure into struct clk
and struct clk_core. For the potential possibility of future renaming,
sizeof(*core->parents) would be better.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
This if-block has been here since the introduction of the common
clock framework. Now no clock drivers are statically initialized.
core->parent is always NULL at this point. Drop the redundant
check and the confusing comment.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Now this function takes clk_core as its argument. __clk_core_init()
would be more suitable for the name of this function.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
The argument clk_user is used only for the clk_user->core. The rest
of this function only takes care of clk_core.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
The "struct device *dev" is not used at all in this function.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Some clocks need to be enabled to accept rate changes. This patch adds a
new flag CLK_SET_RATE_UNGATE that lets clk_change_rate enable the clock
before trying to change the rate and disable it again afterwards.
This of course doesn't effect clocks that are already running at that
point, as their refcount will only temporarily increase.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Currently, of_clk_get_parent_name() returns a wrong parent clock name
when "clock-indices" property exists and the target index is not
found in the property. In this case, NULL should be returned.
For example,
oscillator {
compatible = "myclocktype";
#clock-cells = <1>;
clock-indices = <1>, <3>;
clock-output-names = "clka", "clkb";
};
consumer {
compatible = "myclockconsumer";
clocks = <&oscillator 0>, <&oscillator 1>;
};
Currently, of_clk_get_parent_name(consumer_np, 0) returns "clka"
(and of_clk_get_parent_name(consumer_np, 1) also returns "clka",
this is correct). Because the "clock-indices" in the clock parent
does not contain <0>, of_clk_get_parent_name(consumer_np, 0) should
return NULL.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
The word "cases" is doubled. Keep decent forms for the following
lines.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
This minor refactoring does not change the function behavior.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
This if-block can be dropped because the of_parse_phandle_with_args()
in the following line returns -EINVAL for negative index.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Add clk_hw_is_enabled() to the provider APIs so clk providers can
use a struct clk_hw instead of a struct clk to check if a clk is
enabled or not.
Signed-off-by: Joachim Eastwood <manabian@gmail.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
for_each_matching_node_and_match performs an of_node_get on each iteration,
so a break out of the loop requires an of_node_put.
A simplified version of the semantic patch that fixes this problem is as
follows (http://coccinelle.lip6.fr):
// <smpl>
@@
expression e1,e2,e;
local idexpression np;
@@
for_each_matching_node_and_match(np, e1, e2) {
... when != of_node_put(np)
when != e = np
(
return np;
|
+ of_node_put(np);
? return ...;
)
...
}
// </smpl>
Besides the problem identified by the semantic patch, this patch adds an
of_node_get in front of saving np in a field of parent, to account for the
fact that this value will be put on going on to the next element in the
iteration, and then adds of_node_puts in the two loops where the parent
pointer can be freed.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
When calling __clk_get_name() on a const clock:
warning: passing argument 1 of '__clk_get_name' discards 'const' qualifier from pointer target type
include/linux/clk-provider.h:613:13: note: expected 'struct clk *' but argument is of type 'const struct clk *'
__clk_get_name() does not modify the passed clock, hence make it const.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
If a clock provider has #clock-cells = 1 and we call
of_clk_get_parent_name() on it we may end up returning the name
of the provider node if the provider doesn't have a
clock-output-names property. This doesn't make sense, especially
when you consider that calling of_clk_get_parent_name() on such a
node with different indices will return the same name each time.
Let's try getting the clock from the framework via of_clk_get()
instead, and only fallback to the node name if we have a provider
with #clock-cells = 0. This way, we can't hand out the same name
for different clocks when we don't actually know their names.
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
