blob: c9c399af7c0871c6cca5718f536d1c9335a9c1dd [file] [log] [blame]
Mike Turquette69fe8a82012-03-15 23:11:18 -07001 The Common Clk Framework
2 Mike Turquette <mturquette@ti.com>
3
4This document endeavours to explain the common clk framework details,
5and how to port a platform over to this framework. It is not yet a
6detailed explanation of the clock api in include/linux/clk.h, but
7perhaps someday it will include that information.
8
9 Part 1 - introduction and interface split
10
11The common clk framework is an interface to control the clock nodes
12available on various devices today. This may come in the form of clock
13gating, rate adjustment, muxing or other operations. This framework is
14enabled with the CONFIG_COMMON_CLK option.
15
16The interface itself is divided into two halves, each shielded from the
17details of its counterpart. First is the common definition of struct
18clk which unifies the framework-level accounting and infrastructure that
19has traditionally been duplicated across a variety of platforms. Second
20is a common implementation of the clk.h api, defined in
21drivers/clk/clk.c. Finally there is struct clk_ops, whose operations
22are invoked by the clk api implementation.
23
24The second half of the interface is comprised of the hardware-specific
25callbacks registered with struct clk_ops and the corresponding
26hardware-specific structures needed to model a particular clock. For
27the remainder of this document any reference to a callback in struct
28clk_ops, such as .enable or .set_rate, implies the hardware-specific
29implementation of that code. Likewise, references to struct clk_foo
30serve as a convenient shorthand for the implementation of the
31hardware-specific bits for the hypothetical "foo" hardware.
32
33Tying the two halves of this interface together is struct clk_hw, which
34is defined in struct clk_foo and pointed to within struct clk. This
Sachin Kamat13541952013-06-10 10:02:39 +053035allows for easy navigation between the two discrete halves of the common
Mike Turquette69fe8a82012-03-15 23:11:18 -070036clock interface.
37
38 Part 2 - common data structures and api
39
40Below is the common struct clk definition from
41include/linux/clk-private.h, modified for brevity:
42
43 struct clk {
44 const char *name;
45 const struct clk_ops *ops;
46 struct clk_hw *hw;
47 char **parent_names;
48 struct clk **parents;
49 struct clk *parent;
50 struct hlist_head children;
51 struct hlist_node child_node;
52 ...
53 };
54
55The members above make up the core of the clk tree topology. The clk
56api itself defines several driver-facing functions which operate on
57struct clk. That api is documented in include/linux/clk.h.
58
59Platforms and devices utilizing the common struct clk use the struct
60clk_ops pointer in struct clk to perform the hardware-specific parts of
61the operations defined in clk.h:
62
63 struct clk_ops {
64 int (*prepare)(struct clk_hw *hw);
65 void (*unprepare)(struct clk_hw *hw);
66 int (*enable)(struct clk_hw *hw);
67 void (*disable)(struct clk_hw *hw);
68 int (*is_enabled)(struct clk_hw *hw);
69 unsigned long (*recalc_rate)(struct clk_hw *hw,
70 unsigned long parent_rate);
71 long (*round_rate)(struct clk_hw *hw, unsigned long,
72 unsigned long *);
James Hogan71472c02013-07-29 12:25:00 +010073 long (*determine_rate)(struct clk_hw *hw,
74 unsigned long rate,
75 unsigned long *best_parent_rate,
76 struct clk **best_parent_clk);
Mike Turquette69fe8a82012-03-15 23:11:18 -070077 int (*set_parent)(struct clk_hw *hw, u8 index);
78 u8 (*get_parent)(struct clk_hw *hw);
79 int (*set_rate)(struct clk_hw *hw, unsigned long);
Stephen Boyd3fa22522014-01-15 10:47:22 -080080 int (*set_rate_and_parent)(struct clk_hw *hw,
81 unsigned long rate,
82 unsigned long parent_rate, u8 index);
Boris BREZILLON5279fc42013-12-21 10:34:47 +010083 unsigned long (*recalc_accuracy)(struct clk_hw *hw,
84 unsigned long parent_accuracy);
Mike Turquette69fe8a82012-03-15 23:11:18 -070085 void (*init)(struct clk_hw *hw);
86 };
87
88 Part 3 - hardware clk implementations
89
90The strength of the common struct clk comes from its .ops and .hw pointers
91which abstract the details of struct clk from the hardware-specific bits, and
92vice versa. To illustrate consider the simple gateable clk implementation in
93drivers/clk/clk-gate.c:
94
95struct clk_gate {
96 struct clk_hw hw;
97 void __iomem *reg;
98 u8 bit_idx;
99 ...
100};
101
102struct clk_gate contains struct clk_hw hw as well as hardware-specific
103knowledge about which register and bit controls this clk's gating.
104Nothing about clock topology or accounting, such as enable_count or
105notifier_count, is needed here. That is all handled by the common
106framework code and struct clk.
107
108Let's walk through enabling this clk from driver code:
109
110 struct clk *clk;
111 clk = clk_get(NULL, "my_gateable_clk");
112
113 clk_prepare(clk);
114 clk_enable(clk);
115
116The call graph for clk_enable is very simple:
117
118clk_enable(clk);
119 clk->ops->enable(clk->hw);
120 [resolves to...]
121 clk_gate_enable(hw);
122 [resolves struct clk gate with to_clk_gate(hw)]
123 clk_gate_set_bit(gate);
124
125And the definition of clk_gate_set_bit:
126
127static void clk_gate_set_bit(struct clk_gate *gate)
128{
129 u32 reg;
130
131 reg = __raw_readl(gate->reg);
132 reg |= BIT(gate->bit_idx);
133 writel(reg, gate->reg);
134}
135
136Note that to_clk_gate is defined as:
137
138#define to_clk_gate(_hw) container_of(_hw, struct clk_gate, clk)
139
140This pattern of abstraction is used for every clock hardware
141representation.
142
143 Part 4 - supporting your own clk hardware
144
145When implementing support for a new type of clock it only necessary to
146include the following header:
147
148#include <linux/clk-provider.h>
149
150include/linux/clk.h is included within that header and clk-private.h
151must never be included from the code which implements the operations for
152a clock. More on that below in Part 5.
153
154To construct a clk hardware structure for your platform you must define
155the following:
156
157struct clk_foo {
158 struct clk_hw hw;
159 ... hardware specific data goes here ...
160};
161
162To take advantage of your data you'll need to support valid operations
163for your clk:
164
165struct clk_ops clk_foo_ops {
166 .enable = &clk_foo_enable;
167 .disable = &clk_foo_disable;
168};
169
170Implement the above functions using container_of:
171
172#define to_clk_foo(_hw) container_of(_hw, struct clk_foo, hw)
173
174int clk_foo_enable(struct clk_hw *hw)
175{
176 struct clk_foo *foo;
177
178 foo = to_clk_foo(hw);
179
180 ... perform magic on foo ...
181
182 return 0;
183};
184
185Below is a matrix detailing which clk_ops are mandatory based upon the
Eduardo Valentina368a6a2013-02-28 09:59:07 -0400186hardware capabilities of that clock. A cell marked as "y" means
Mike Turquette69fe8a82012-03-15 23:11:18 -0700187mandatory, a cell marked as "n" implies that either including that
Eduardo Valentina368a6a2013-02-28 09:59:07 -0400188callback is invalid or otherwise unnecessary. Empty cells are either
Mike Turquette69fe8a82012-03-15 23:11:18 -0700189optional or must be evaluated on a case-by-case basis.
190
James Hogan71472c02013-07-29 12:25:00 +0100191 clock hardware characteristics
192 -----------------------------------------------------------
193 | gate | change rate | single parent | multiplexer | root |
194 |------|-------------|---------------|-------------|------|
195.prepare | | | | | |
196.unprepare | | | | | |
197 | | | | | |
198.enable | y | | | | |
199.disable | y | | | | |
200.is_enabled | y | | | | |
201 | | | | | |
202.recalc_rate | | y | | | |
203.round_rate | | y [1] | | | |
204.determine_rate | | y [1] | | | |
205.set_rate | | y | | | |
206 | | | | | |
207.set_parent | | | n | y | n |
208.get_parent | | | n | y | n |
209 | | | | | |
Boris BREZILLON5279fc42013-12-21 10:34:47 +0100210.recalc_accuracy| | | | | |
211 | | | | | |
James Hogan71472c02013-07-29 12:25:00 +0100212.init | | | | | |
213 -----------------------------------------------------------
214[1] either one of round_rate or determine_rate is required.
Mike Turquette69fe8a82012-03-15 23:11:18 -0700215
216Finally, register your clock at run-time with a hardware-specific
217registration function. This function simply populates struct clk_foo's
218data and then passes the common struct clk parameters to the framework
219with a call to:
220
221clk_register(...)
222
223See the basic clock types in drivers/clk/clk-*.c for examples.
224
225 Part 5 - static initialization of clock data
226
227For platforms with many clocks (often numbering into the hundreds) it
228may be desirable to statically initialize some clock data. This
229presents a problem since the definition of struct clk should be hidden
230from everyone except for the clock core in drivers/clk/clk.c.
231
232To get around this problem struct clk's definition is exposed in
233include/linux/clk-private.h along with some macros for more easily
234initializing instances of the basic clock types. These clocks must
235still be initialized with the common clock framework via a call to
236__clk_init.
237
238clk-private.h must NEVER be included by code which implements struct
239clk_ops callbacks, nor must it be included by any logic which pokes
240around inside of struct clk at run-time. To do so is a layering
241violation.
242
243To better enforce this policy, always follow this simple rule: any
244statically initialized clock data MUST be defined in a separate file
245from the logic that implements its ops. Basically separate the logic
246from the data and all is well.
Olof Johansson1e435252013-04-27 14:10:18 -0700247
248 Part 6 - Disabling clock gating of unused clocks
249
250Sometimes during development it can be useful to be able to bypass the
251default disabling of unused clocks. For example, if drivers aren't enabling
252clocks properly but rely on them being on from the bootloader, bypassing
253the disabling means that the driver will remain functional while the issues
254are sorted out.
255
256To bypass this disabling, include "clk_ignore_unused" in the bootargs to the
257kernel.
Laurent Pinchart843bad82014-02-28 13:40:56 +0100258
259 Part 7 - Locking
260
261The common clock framework uses two global locks, the prepare lock and the
262enable lock.
263
264The enable lock is a spinlock and is held across calls to the .enable,
265.disable and .is_enabled operations. Those operations are thus not allowed to
266sleep, and calls to the clk_enable(), clk_disable() and clk_is_enabled() API
267functions are allowed in atomic context.
268
269The prepare lock is a mutex and is held across calls to all other operations.
270All those operations are allowed to sleep, and calls to the corresponding API
271functions are not allowed in atomic context.
272
273This effectively divides operations in two groups from a locking perspective.
274
275Drivers don't need to manually protect resources shared between the operations
276of one group, regardless of whether those resources are shared by multiple
277clocks or not. However, access to resources that are shared between operations
278of the two groups needs to be protected by the drivers. An example of such a
279resource would be a register that controls both the clock rate and the clock
280enable/disable state.
281
282The clock framework is reentrant, in that a driver is allowed to call clock
283framework functions from within its implementation of clock operations. This
284can for instance cause a .set_rate operation of one clock being called from
285within the .set_rate operation of another clock. This case must be considered
286in the driver implementations, but the code flow is usually controlled by the
287driver in that case.
288
289Note that locking must also be considered when code outside of the common
290clock framework needs to access resources used by the clock operations. This
291is considered out of scope of this document.