linux-stable/arch/arm/mach-omap1/clock_data.c

835 lines
27 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* linux/arch/arm/mach-omap1/clock_data.c
*
* Copyright (C) 2004 - 2005, 2009-2010 Nokia Corporation
* Written by Tuukka Tikkanen <tuukka.tikkanen@elektrobit.com>
* Based on clocks.h by Tony Lindgren, Gordon McNutt and RidgeRun, Inc
*
* To do:
* - Clocks that are only available on some chips should be marked with the
* chips that they are present on.
*/
#include <linux/kernel.h>
#include <linux/io.h>
#include <linux/clk.h>
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
#include <linux/clkdev.h>
#include <linux/clk-provider.h>
#include <linux/cpufreq.h>
#include <linux/delay.h>
#include <linux/soc/ti/omap1-io.h>
#include <asm/mach-types.h> /* for machine_is_* */
#include "soc.h"
#include "hardware.h"
#include "usb.h" /* for OTG_BASE */
#include "iomap.h"
#include "clock.h"
#include "sram.h"
/* Some ARM_IDLECT1 bit shifts - used in struct arm_idlect1_clk */
#define IDL_CLKOUT_ARM_SHIFT 12
#define IDLTIM_ARM_SHIFT 9
#define IDLAPI_ARM_SHIFT 8
#define IDLIF_ARM_SHIFT 6
#define IDLLB_ARM_SHIFT 4 /* undocumented? */
#define OMAP1510_IDLLCD_ARM_SHIFT 3 /* undocumented? */
#define IDLPER_ARM_SHIFT 2
#define IDLXORP_ARM_SHIFT 1
#define IDLWDT_ARM_SHIFT 0
/* Some MOD_CONF_CTRL_0 bit shifts - used in struct clk.enable_bit */
#define CONF_MOD_UART3_CLK_MODE_R 31
#define CONF_MOD_UART2_CLK_MODE_R 30
#define CONF_MOD_UART1_CLK_MODE_R 29
#define CONF_MOD_MMC_SD_CLK_REQ_R 23
#define CONF_MOD_MCBSP3_AUXON 20
/* Some MOD_CONF_CTRL_1 bit shifts - used in struct clk.enable_bit */
#define CONF_MOD_SOSSI_CLK_EN_R 16
/* Some OTG_SYSCON_2-specific bit fields */
#define OTG_SYSCON_2_UHOST_EN_SHIFT 8
/* Some SOFT_REQ_REG bit fields - used in struct clk.enable_bit */
#define SOFT_MMC2_DPLL_REQ_SHIFT 13
#define SOFT_MMC_DPLL_REQ_SHIFT 12
#define SOFT_UART3_DPLL_REQ_SHIFT 11
#define SOFT_UART2_DPLL_REQ_SHIFT 10
#define SOFT_UART1_DPLL_REQ_SHIFT 9
#define SOFT_USB_OTG_DPLL_REQ_SHIFT 8
#define SOFT_CAM_DPLL_REQ_SHIFT 7
#define SOFT_COM_MCKO_REQ_SHIFT 6
#define SOFT_PERIPH_REQ_SHIFT 5 /* sys_ck gate for UART2 ? */
#define USB_REQ_EN_SHIFT 4
#define SOFT_USB_REQ_SHIFT 3 /* sys_ck gate for USB host? */
#define SOFT_SDW_REQ_SHIFT 2 /* sys_ck gate for Bluetooth? */
#define SOFT_COM_REQ_SHIFT 1 /* sys_ck gate for com proc? */
#define SOFT_DPLL_REQ_SHIFT 0
/*
* Omap1 clocks
*/
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk ck_ref = {
.hw.init = CLK_HW_INIT_NO_PARENT("ck_ref", &omap1_clk_rate_ops, 0),
.rate = 12000000,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk ck_dpll1 = {
.hw.init = CLK_HW_INIT("ck_dpll1", "ck_ref", &omap1_clk_rate_ops,
/*
* force recursive refresh of rates of the clock
* and its children when clk_get_rate() is called
*/
CLK_GET_RATE_NOCACHE),
};
/*
* FIXME: This clock seems to be necessary but no-one has asked for its
* activation. [ FIX: SoSSI, SSR ]
*/
static struct arm_idlect1_clk ck_dpll1out = {
.clk = {
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("ck_dpll1out", "ck_dpll1", &omap1_clk_gate_ops, 0),
.ops = &clkops_generic,
.flags = CLOCK_IDLE_CONTROL | ENABLE_REG_32BIT,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT2),
.enable_bit = EN_CKOUT_ARM,
},
.idlect_shift = IDL_CLKOUT_ARM_SHIFT,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk sossi_ck = {
.hw.init = CLK_HW_INIT("ck_sossi", "ck_dpll1out", &omap1_clk_full_ops, 0),
.ops = &clkops_generic,
.flags = CLOCK_NO_IDLE_PARENT | ENABLE_REG_32BIT,
.enable_reg = OMAP1_IO_ADDRESS(MOD_CONF_CTRL_1),
.enable_bit = CONF_MOD_SOSSI_CLK_EN_R,
.recalc = &omap1_sossi_recalc,
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.round_rate = &omap1_round_sossi_rate,
.set_rate = &omap1_set_sossi_rate,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk arm_ck = {
.hw.init = CLK_HW_INIT("arm_ck", "ck_dpll1", &omap1_clk_rate_ops, 0),
.rate_offset = CKCTL_ARMDIV_OFFSET,
.recalc = &omap1_ckctl_recalc,
.round_rate = omap1_clk_round_rate_ckctl_arm,
.set_rate = omap1_clk_set_rate_ckctl_arm,
};
static struct arm_idlect1_clk armper_ck = {
.clk = {
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("armper_ck", "ck_dpll1", &omap1_clk_full_ops,
CLK_IS_CRITICAL),
.ops = &clkops_generic,
.flags = CLOCK_IDLE_CONTROL,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT2),
.enable_bit = EN_PERCK,
.rate_offset = CKCTL_PERDIV_OFFSET,
.recalc = &omap1_ckctl_recalc,
.round_rate = omap1_clk_round_rate_ckctl_arm,
.set_rate = omap1_clk_set_rate_ckctl_arm,
},
.idlect_shift = IDLPER_ARM_SHIFT,
};
/*
* FIXME: This clock seems to be necessary but no-one has asked for its
* activation. [ GPIO code for 1510 ]
*/
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk arm_gpio_ck = {
.hw.init = CLK_HW_INIT("ick", "ck_dpll1", &omap1_clk_gate_ops, CLK_IS_CRITICAL),
.ops = &clkops_generic,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT2),
.enable_bit = EN_GPIOCK,
};
static struct arm_idlect1_clk armxor_ck = {
.clk = {
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("armxor_ck", "ck_ref", &omap1_clk_gate_ops,
CLK_IS_CRITICAL),
.ops = &clkops_generic,
.flags = CLOCK_IDLE_CONTROL,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT2),
.enable_bit = EN_XORPCK,
},
.idlect_shift = IDLXORP_ARM_SHIFT,
};
static struct arm_idlect1_clk armtim_ck = {
.clk = {
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("armtim_ck", "ck_ref", &omap1_clk_gate_ops,
CLK_IS_CRITICAL),
.ops = &clkops_generic,
.flags = CLOCK_IDLE_CONTROL,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT2),
.enable_bit = EN_TIMCK,
},
.idlect_shift = IDLTIM_ARM_SHIFT,
};
static struct arm_idlect1_clk armwdt_ck = {
.clk = {
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("armwdt_ck", "ck_ref", &omap1_clk_full_ops, 0),
.ops = &clkops_generic,
.flags = CLOCK_IDLE_CONTROL,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT2),
.enable_bit = EN_WDTCK,
.fixed_div = 14,
.recalc = &omap_fixed_divisor_recalc,
},
.idlect_shift = IDLWDT_ARM_SHIFT,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk arminth_ck16xx = {
.hw.init = CLK_HW_INIT("arminth_ck", "arm_ck", &omap1_clk_null_ops, 0),
/* Note: On 16xx the frequency can be divided by 2 by programming
* ARM_CKCTL:ARM_INTHCK_SEL(14) to 1
*
* 1510 version is in TC clocks.
*/
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk dsp_ck = {
.hw.init = CLK_HW_INIT("dsp_ck", "ck_dpll1", &omap1_clk_full_ops, 0),
.ops = &clkops_generic,
.enable_reg = OMAP1_IO_ADDRESS(ARM_CKCTL),
.enable_bit = EN_DSPCK,
.rate_offset = CKCTL_DSPDIV_OFFSET,
.recalc = &omap1_ckctl_recalc,
.round_rate = omap1_clk_round_rate_ckctl_arm,
.set_rate = omap1_clk_set_rate_ckctl_arm,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk dspmmu_ck = {
.hw.init = CLK_HW_INIT("dspmmu_ck", "ck_dpll1", &omap1_clk_rate_ops, 0),
.rate_offset = CKCTL_DSPMMUDIV_OFFSET,
.recalc = &omap1_ckctl_recalc,
.round_rate = omap1_clk_round_rate_ckctl_arm,
.set_rate = omap1_clk_set_rate_ckctl_arm,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk dspper_ck = {
.hw.init = CLK_HW_INIT("dspper_ck", "ck_dpll1", &omap1_clk_full_ops, 0),
.ops = &clkops_dspck,
.enable_reg = DSP_IDLECT2,
.enable_bit = EN_PERCK,
.rate_offset = CKCTL_PERDIV_OFFSET,
.recalc = &omap1_ckctl_recalc_dsp_domain,
.round_rate = omap1_clk_round_rate_ckctl_arm,
.set_rate = &omap1_clk_set_rate_dsp_domain,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk dspxor_ck = {
.hw.init = CLK_HW_INIT("dspxor_ck", "ck_ref", &omap1_clk_gate_ops, 0),
.ops = &clkops_dspck,
.enable_reg = DSP_IDLECT2,
.enable_bit = EN_XORPCK,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk dsptim_ck = {
.hw.init = CLK_HW_INIT("dsptim_ck", "ck_ref", &omap1_clk_gate_ops, 0),
.ops = &clkops_dspck,
.enable_reg = DSP_IDLECT2,
.enable_bit = EN_DSPTIMCK,
};
static struct arm_idlect1_clk tc_ck = {
.clk = {
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("tc_ck", "ck_dpll1", &omap1_clk_rate_ops, 0),
.flags = CLOCK_IDLE_CONTROL,
.rate_offset = CKCTL_TCDIV_OFFSET,
.recalc = &omap1_ckctl_recalc,
.round_rate = omap1_clk_round_rate_ckctl_arm,
.set_rate = omap1_clk_set_rate_ckctl_arm,
},
.idlect_shift = IDLIF_ARM_SHIFT,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk arminth_ck1510 = {
.hw.init = CLK_HW_INIT("arminth_ck", "tc_ck", &omap1_clk_null_ops, 0),
/* Note: On 1510 the frequency follows TC_CK
*
* 16xx version is in MPU clocks.
*/
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk tipb_ck = {
/* No-idle controlled by "tc_ck" */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("tipb_ck", "tc_ck", &omap1_clk_null_ops, 0),
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk l3_ocpi_ck = {
/* No-idle controlled by "tc_ck" */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("l3_ocpi_ck", "tc_ck", &omap1_clk_gate_ops, 0),
.ops = &clkops_generic,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT3),
.enable_bit = EN_OCPI_CK,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk tc1_ck = {
.hw.init = CLK_HW_INIT("tc1_ck", "tc_ck", &omap1_clk_gate_ops, 0),
.ops = &clkops_generic,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT3),
.enable_bit = EN_TC1_CK,
};
/*
* FIXME: This clock seems to be necessary but no-one has asked for its
* activation. [ pm.c (SRAM), CCP, Camera ]
*/
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk tc2_ck = {
.hw.init = CLK_HW_INIT("tc2_ck", "tc_ck", &omap1_clk_gate_ops, CLK_IS_CRITICAL),
.ops = &clkops_generic,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT3),
.enable_bit = EN_TC2_CK,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk dma_ck = {
/* No-idle controlled by "tc_ck" */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("dma_ck", "tc_ck", &omap1_clk_null_ops, 0),
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk dma_lcdfree_ck = {
.hw.init = CLK_HW_INIT("dma_lcdfree_ck", "tc_ck", &omap1_clk_null_ops, 0),
};
static struct arm_idlect1_clk api_ck = {
.clk = {
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("api_ck", "tc_ck", &omap1_clk_gate_ops, 0),
.ops = &clkops_generic,
.flags = CLOCK_IDLE_CONTROL,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT2),
.enable_bit = EN_APICK,
},
.idlect_shift = IDLAPI_ARM_SHIFT,
};
static struct arm_idlect1_clk lb_ck = {
.clk = {
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("lb_ck", "tc_ck", &omap1_clk_gate_ops, 0),
.ops = &clkops_generic,
.flags = CLOCK_IDLE_CONTROL,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT2),
.enable_bit = EN_LBCK,
},
.idlect_shift = IDLLB_ARM_SHIFT,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk rhea1_ck = {
.hw.init = CLK_HW_INIT("rhea1_ck", "tc_ck", &omap1_clk_null_ops, 0),
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk rhea2_ck = {
.hw.init = CLK_HW_INIT("rhea2_ck", "tc_ck", &omap1_clk_null_ops, 0),
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk lcd_ck_16xx = {
.hw.init = CLK_HW_INIT("lcd_ck", "ck_dpll1", &omap1_clk_full_ops, 0),
.ops = &clkops_generic,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT2),
.enable_bit = EN_LCDCK,
.rate_offset = CKCTL_LCDDIV_OFFSET,
.recalc = &omap1_ckctl_recalc,
.round_rate = omap1_clk_round_rate_ckctl_arm,
.set_rate = omap1_clk_set_rate_ckctl_arm,
};
static struct arm_idlect1_clk lcd_ck_1510 = {
.clk = {
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("lcd_ck", "ck_dpll1", &omap1_clk_full_ops, 0),
.ops = &clkops_generic,
.flags = CLOCK_IDLE_CONTROL,
.enable_reg = OMAP1_IO_ADDRESS(ARM_IDLECT2),
.enable_bit = EN_LCDCK,
.rate_offset = CKCTL_LCDDIV_OFFSET,
.recalc = &omap1_ckctl_recalc,
.round_rate = omap1_clk_round_rate_ckctl_arm,
.set_rate = omap1_clk_set_rate_ckctl_arm,
},
.idlect_shift = OMAP1510_IDLLCD_ARM_SHIFT,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
/*
* XXX The enable_bit here is misused - it simply switches between 12MHz
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
* and 48MHz. Reimplement with clk_mux.
*
* XXX does this need SYSC register handling?
*/
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk uart1_1510 = {
/* Direct from ULPD, no real parent */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("uart1_ck", "armper_ck", &omap1_clk_full_ops, 0),
.flags = ENABLE_REG_32BIT | CLOCK_NO_IDLE_PARENT,
.enable_reg = OMAP1_IO_ADDRESS(MOD_CONF_CTRL_0),
.enable_bit = CONF_MOD_UART1_CLK_MODE_R,
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.round_rate = &omap1_round_uart_rate,
.set_rate = &omap1_set_uart_rate,
.recalc = &omap1_uart_recalc,
};
/*
* XXX The enable_bit here is misused - it simply switches between 12MHz
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
* and 48MHz. Reimplement with clk_mux.
*
* XXX SYSC register handling does not belong in the clock framework
*/
static struct uart_clk uart1_16xx = {
.clk = {
.ops = &clkops_uart_16xx,
/* Direct from ULPD, no real parent */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("uart1_ck", "armper_ck", &omap1_clk_full_ops, 0),
.rate = 48000000,
.flags = ENABLE_REG_32BIT | CLOCK_NO_IDLE_PARENT,
.enable_reg = OMAP1_IO_ADDRESS(MOD_CONF_CTRL_0),
.enable_bit = CONF_MOD_UART1_CLK_MODE_R,
},
.sysc_addr = 0xfffb0054,
};
/*
* XXX The enable_bit here is misused - it simply switches between 12MHz
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
* and 48MHz. Reimplement with clk_mux.
*
* XXX does this need SYSC register handling?
*/
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk uart2_ck = {
/* Direct from ULPD, no real parent */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("uart2_ck", "armper_ck", &omap1_clk_full_ops, 0),
.flags = ENABLE_REG_32BIT | CLOCK_NO_IDLE_PARENT,
.enable_reg = OMAP1_IO_ADDRESS(MOD_CONF_CTRL_0),
.enable_bit = CONF_MOD_UART2_CLK_MODE_R,
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.round_rate = &omap1_round_uart_rate,
.set_rate = &omap1_set_uart_rate,
.recalc = &omap1_uart_recalc,
};
/*
* XXX The enable_bit here is misused - it simply switches between 12MHz
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
* and 48MHz. Reimplement with clk_mux.
*
* XXX does this need SYSC register handling?
*/
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk uart3_1510 = {
/* Direct from ULPD, no real parent */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("uart3_ck", "armper_ck", &omap1_clk_full_ops, 0),
.flags = ENABLE_REG_32BIT | CLOCK_NO_IDLE_PARENT,
.enable_reg = OMAP1_IO_ADDRESS(MOD_CONF_CTRL_0),
.enable_bit = CONF_MOD_UART3_CLK_MODE_R,
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.round_rate = &omap1_round_uart_rate,
.set_rate = &omap1_set_uart_rate,
.recalc = &omap1_uart_recalc,
};
/*
* XXX The enable_bit here is misused - it simply switches between 12MHz
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
* and 48MHz. Reimplement with clk_mux.
*
* XXX SYSC register handling does not belong in the clock framework
*/
static struct uart_clk uart3_16xx = {
.clk = {
.ops = &clkops_uart_16xx,
/* Direct from ULPD, no real parent */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("uart3_ck", "armper_ck", &omap1_clk_full_ops, 0),
.rate = 48000000,
.flags = ENABLE_REG_32BIT | CLOCK_NO_IDLE_PARENT,
.enable_reg = OMAP1_IO_ADDRESS(MOD_CONF_CTRL_0),
.enable_bit = CONF_MOD_UART3_CLK_MODE_R,
},
.sysc_addr = 0xfffb9854,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk usb_clko = { /* 6 MHz output on W4_USB_CLKO */
.ops = &clkops_generic,
/* Direct from ULPD, no parent */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT_NO_PARENT("usb_clko", &omap1_clk_full_ops, 0),
.rate = 6000000,
.flags = ENABLE_REG_32BIT,
.enable_reg = OMAP1_IO_ADDRESS(ULPD_CLOCK_CTRL),
.enable_bit = USB_MCLK_EN_BIT,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk usb_hhc_ck1510 = {
.ops = &clkops_generic,
/* Direct from ULPD, no parent */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT_NO_PARENT("usb_hhc_ck", &omap1_clk_full_ops, 0),
.rate = 48000000, /* Actually 2 clocks, 12MHz and 48MHz */
.flags = ENABLE_REG_32BIT,
.enable_reg = OMAP1_IO_ADDRESS(MOD_CONF_CTRL_0),
.enable_bit = USB_HOST_HHC_UHOST_EN,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk usb_hhc_ck16xx = {
.ops = &clkops_generic,
/* Direct from ULPD, no parent */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT_NO_PARENT("usb_hhc_ck", &omap1_clk_full_ops, 0),
.rate = 48000000,
/* OTG_SYSCON_2.OTG_PADEN == 0 (not 1510-compatible) */
.flags = ENABLE_REG_32BIT,
.enable_reg = OMAP1_IO_ADDRESS(OTG_BASE + 0x08), /* OTG_SYSCON_2 */
.enable_bit = OTG_SYSCON_2_UHOST_EN_SHIFT
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk usb_dc_ck = {
.ops = &clkops_generic,
/* Direct from ULPD, no parent */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT_NO_PARENT("usb_dc_ck", &omap1_clk_full_ops, 0),
.rate = 48000000,
.enable_reg = OMAP1_IO_ADDRESS(SOFT_REQ_REG),
.enable_bit = SOFT_USB_OTG_DPLL_REQ_SHIFT,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk uart1_7xx = {
.ops = &clkops_generic,
/* Direct from ULPD, no parent */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT_NO_PARENT("uart1_ck", &omap1_clk_full_ops, 0),
.rate = 12000000,
.enable_reg = OMAP1_IO_ADDRESS(SOFT_REQ_REG),
.enable_bit = 9,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk uart2_7xx = {
.ops = &clkops_generic,
/* Direct from ULPD, no parent */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT_NO_PARENT("uart2_ck", &omap1_clk_full_ops, 0),
.rate = 12000000,
.enable_reg = OMAP1_IO_ADDRESS(SOFT_REQ_REG),
.enable_bit = 11,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk mclk_1510 = {
.ops = &clkops_generic,
/* Direct from ULPD, no parent. May be enabled by ext hardware. */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT_NO_PARENT("mclk", &omap1_clk_full_ops, 0),
.rate = 12000000,
.enable_reg = OMAP1_IO_ADDRESS(SOFT_REQ_REG),
.enable_bit = SOFT_COM_MCKO_REQ_SHIFT,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk mclk_16xx = {
.ops = &clkops_generic,
/* Direct from ULPD, no parent. May be enabled by ext hardware. */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT_NO_PARENT("mclk", &omap1_clk_full_ops, 0),
.enable_reg = OMAP1_IO_ADDRESS(COM_CLK_DIV_CTRL_SEL),
.enable_bit = COM_ULPD_PLL_CLK_REQ,
.set_rate = &omap1_set_ext_clk_rate,
.round_rate = &omap1_round_ext_clk_rate,
.init = &omap1_init_ext_clk,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk bclk_1510 = {
/* Direct from ULPD, no parent. May be enabled by ext hardware. */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT_NO_PARENT("bclk", &omap1_clk_rate_ops, 0),
.rate = 12000000,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk bclk_16xx = {
.ops = &clkops_generic,
/* Direct from ULPD, no parent. May be enabled by ext hardware. */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT_NO_PARENT("bclk", &omap1_clk_full_ops, 0),
.enable_reg = OMAP1_IO_ADDRESS(SWD_CLK_DIV_CTRL_SEL),
.enable_bit = SWD_ULPD_PLL_CLK_REQ,
.set_rate = &omap1_set_ext_clk_rate,
.round_rate = &omap1_round_ext_clk_rate,
.init = &omap1_init_ext_clk,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk mmc1_ck = {
.ops = &clkops_generic,
/* Functional clock is direct from ULPD, interface clock is ARMPER */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("mmc1_ck", "armper_ck", &omap1_clk_full_ops, 0),
.rate = 48000000,
.flags = ENABLE_REG_32BIT | CLOCK_NO_IDLE_PARENT,
.enable_reg = OMAP1_IO_ADDRESS(MOD_CONF_CTRL_0),
.enable_bit = CONF_MOD_MMC_SD_CLK_REQ_R,
};
/*
* XXX MOD_CONF_CTRL_0 bit 20 is defined in the 1510 TRM as
* CONF_MOD_MCBSP3_AUXON ??
*/
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk mmc2_ck = {
.ops = &clkops_generic,
/* Functional clock is direct from ULPD, interface clock is ARMPER */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("mmc2_ck", "armper_ck", &omap1_clk_full_ops, 0),
.rate = 48000000,
.flags = ENABLE_REG_32BIT | CLOCK_NO_IDLE_PARENT,
.enable_reg = OMAP1_IO_ADDRESS(MOD_CONF_CTRL_0),
.enable_bit = 20,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk mmc3_ck = {
.ops = &clkops_generic,
/* Functional clock is direct from ULPD, interface clock is ARMPER */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
.hw.init = CLK_HW_INIT("mmc3_ck", "armper_ck", &omap1_clk_full_ops, 0),
.rate = 48000000,
.flags = ENABLE_REG_32BIT | CLOCK_NO_IDLE_PARENT,
.enable_reg = OMAP1_IO_ADDRESS(SOFT_REQ_REG),
.enable_bit = SOFT_MMC_DPLL_REQ_SHIFT,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk virtual_ck_mpu = {
/* Is smarter alias for arm_ck */
.hw.init = CLK_HW_INIT("mpu", "arm_ck", &omap1_clk_rate_ops, 0),
.recalc = &followparent_recalc,
.set_rate = &omap1_select_table_rate,
.round_rate = &omap1_round_to_table_rate,
};
/* virtual functional clock domain for I2C. Just for making sure that ARMXOR_CK
remains active during MPU idle whenever this is enabled */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk i2c_fck = {
.hw.init = CLK_HW_INIT("i2c_fck", "armxor_ck", &omap1_clk_gate_ops, 0),
.flags = CLOCK_NO_IDLE_PARENT,
};
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
static struct omap1_clk i2c_ick = {
.hw.init = CLK_HW_INIT("i2c_ick", "armper_ck", &omap1_clk_gate_ops, 0),
.flags = CLOCK_NO_IDLE_PARENT,
};
/*
* clkdev integration
*/
static struct omap_clk omap_clks[] = {
/* non-ULPD clocks */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
CLK(NULL, "ck_ref", &ck_ref.hw, CK_16XX | CK_1510 | CK_310 | CK_7XX),
CLK(NULL, "ck_dpll1", &ck_dpll1.hw, CK_16XX | CK_1510 | CK_310 | CK_7XX),
/* CK_GEN1 clocks */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
CLK(NULL, "ck_dpll1out", &ck_dpll1out.clk.hw, CK_16XX),
CLK(NULL, "ck_sossi", &sossi_ck.hw, CK_16XX),
CLK(NULL, "arm_ck", &arm_ck.hw, CK_16XX | CK_1510 | CK_310),
CLK(NULL, "armper_ck", &armper_ck.clk.hw, CK_16XX | CK_1510 | CK_310),
CLK("omap_gpio.0", "ick", &arm_gpio_ck.hw, CK_1510 | CK_310),
CLK(NULL, "armxor_ck", &armxor_ck.clk.hw, CK_16XX | CK_1510 | CK_310 | CK_7XX),
CLK(NULL, "armtim_ck", &armtim_ck.clk.hw, CK_16XX | CK_1510 | CK_310),
CLK("omap_wdt", "fck", &armwdt_ck.clk.hw, CK_16XX | CK_1510 | CK_310),
CLK("omap_wdt", "ick", &armper_ck.clk.hw, CK_16XX),
CLK("omap_wdt", "ick", &dummy_ck.hw, CK_1510 | CK_310),
CLK(NULL, "arminth_ck", &arminth_ck1510.hw, CK_1510 | CK_310),
CLK(NULL, "arminth_ck", &arminth_ck16xx.hw, CK_16XX),
/* CK_GEN2 clocks */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
CLK(NULL, "dsp_ck", &dsp_ck.hw, CK_16XX | CK_1510 | CK_310),
CLK(NULL, "dspmmu_ck", &dspmmu_ck.hw, CK_16XX | CK_1510 | CK_310),
CLK(NULL, "dspper_ck", &dspper_ck.hw, CK_16XX | CK_1510 | CK_310),
CLK(NULL, "dspxor_ck", &dspxor_ck.hw, CK_16XX | CK_1510 | CK_310),
CLK(NULL, "dsptim_ck", &dsptim_ck.hw, CK_16XX | CK_1510 | CK_310),
/* CK_GEN3 clocks */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
CLK(NULL, "tc_ck", &tc_ck.clk.hw, CK_16XX | CK_1510 | CK_310 | CK_7XX),
CLK(NULL, "tipb_ck", &tipb_ck.hw, CK_1510 | CK_310),
CLK(NULL, "l3_ocpi_ck", &l3_ocpi_ck.hw, CK_16XX | CK_7XX),
CLK(NULL, "tc1_ck", &tc1_ck.hw, CK_16XX),
CLK(NULL, "tc2_ck", &tc2_ck.hw, CK_16XX),
CLK(NULL, "dma_ck", &dma_ck.hw, CK_16XX | CK_1510 | CK_310),
CLK(NULL, "dma_lcdfree_ck", &dma_lcdfree_ck.hw, CK_16XX),
CLK(NULL, "api_ck", &api_ck.clk.hw, CK_16XX | CK_1510 | CK_310 | CK_7XX),
CLK(NULL, "lb_ck", &lb_ck.clk.hw, CK_1510 | CK_310),
CLK(NULL, "rhea1_ck", &rhea1_ck.hw, CK_16XX),
CLK(NULL, "rhea2_ck", &rhea2_ck.hw, CK_16XX),
CLK(NULL, "lcd_ck", &lcd_ck_16xx.hw, CK_16XX | CK_7XX),
CLK(NULL, "lcd_ck", &lcd_ck_1510.clk.hw, CK_1510 | CK_310),
/* ULPD clocks */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
CLK(NULL, "uart1_ck", &uart1_1510.hw, CK_1510 | CK_310),
CLK(NULL, "uart1_ck", &uart1_16xx.clk.hw, CK_16XX),
CLK(NULL, "uart1_ck", &uart1_7xx.hw, CK_7XX),
CLK(NULL, "uart2_ck", &uart2_ck.hw, CK_16XX | CK_1510 | CK_310),
CLK(NULL, "uart2_ck", &uart2_7xx.hw, CK_7XX),
CLK(NULL, "uart3_ck", &uart3_1510.hw, CK_1510 | CK_310),
CLK(NULL, "uart3_ck", &uart3_16xx.clk.hw, CK_16XX),
CLK(NULL, "usb_clko", &usb_clko.hw, CK_16XX | CK_1510 | CK_310),
CLK(NULL, "usb_hhc_ck", &usb_hhc_ck1510.hw, CK_1510 | CK_310),
CLK(NULL, "usb_hhc_ck", &usb_hhc_ck16xx.hw, CK_16XX),
CLK(NULL, "usb_dc_ck", &usb_dc_ck.hw, CK_16XX | CK_7XX),
CLK(NULL, "mclk", &mclk_1510.hw, CK_1510 | CK_310),
CLK(NULL, "mclk", &mclk_16xx.hw, CK_16XX),
CLK(NULL, "bclk", &bclk_1510.hw, CK_1510 | CK_310),
CLK(NULL, "bclk", &bclk_16xx.hw, CK_16XX),
CLK("mmci-omap.0", "fck", &mmc1_ck.hw, CK_16XX | CK_1510 | CK_310),
CLK("mmci-omap.0", "fck", &mmc3_ck.hw, CK_7XX),
CLK("mmci-omap.0", "ick", &armper_ck.clk.hw, CK_16XX | CK_1510 | CK_310 | CK_7XX),
CLK("mmci-omap.1", "fck", &mmc2_ck.hw, CK_16XX),
CLK("mmci-omap.1", "ick", &armper_ck.clk.hw, CK_16XX),
/* Virtual clocks */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
CLK(NULL, "mpu", &virtual_ck_mpu.hw, CK_16XX | CK_1510 | CK_310),
CLK("omap_i2c.1", "fck", &i2c_fck.hw, CK_16XX | CK_1510 | CK_310 | CK_7XX),
CLK("omap_i2c.1", "ick", &i2c_ick.hw, CK_16XX),
CLK("omap_i2c.1", "ick", &dummy_ck.hw, CK_1510 | CK_310 | CK_7XX),
CLK("omap1_spi100k.1", "fck", &dummy_ck.hw, CK_7XX),
CLK("omap1_spi100k.1", "ick", &dummy_ck.hw, CK_7XX),
CLK("omap1_spi100k.2", "fck", &dummy_ck.hw, CK_7XX),
CLK("omap1_spi100k.2", "ick", &dummy_ck.hw, CK_7XX),
CLK("omap_uwire", "fck", &armxor_ck.clk.hw, CK_16XX | CK_1510 | CK_310),
CLK("omap-mcbsp.1", "ick", &dspper_ck.hw, CK_16XX),
CLK("omap-mcbsp.1", "ick", &dummy_ck.hw, CK_1510 | CK_310),
CLK("omap-mcbsp.2", "ick", &armper_ck.clk.hw, CK_16XX),
CLK("omap-mcbsp.2", "ick", &dummy_ck.hw, CK_1510 | CK_310),
CLK("omap-mcbsp.3", "ick", &dspper_ck.hw, CK_16XX),
CLK("omap-mcbsp.3", "ick", &dummy_ck.hw, CK_1510 | CK_310),
CLK("omap-mcbsp.1", "fck", &dspxor_ck.hw, CK_16XX | CK_1510 | CK_310),
CLK("omap-mcbsp.2", "fck", &armper_ck.clk.hw, CK_16XX | CK_1510 | CK_310),
CLK("omap-mcbsp.3", "fck", &dspxor_ck.hw, CK_16XX | CK_1510 | CK_310),
};
/*
* init
*/
ARM: OMAP: Fix reprogramming of dpll1 rate Commit a66cb3454f220f49f900646ebdc76cb943319eb7 (ARM: OMAP: Map SRAM later on with ioremap_exec()) moved the SRAM init to happen later to remove a dependency to early SoC detection for map_io. This broke booting on some boards not using Kconfig option for OMAP_CLOCKS_SET_BY_BOOTLOADER as the dpll1 reprogramming would cause the following error: kernel BUG at arch/arm/plat-omap/sram.c:226! Internal error: Oops - undefined instruction: 0 [#1] PREEMPT Modules linked in: CPU: 0 Not tainted (3.2.0-rc1-e3 #9) PC is at omap_sram_reprogram_clock+0x28/0x30 LR is at omap1_select_table_rate+0x88/0xb4 pc : [<c001b0c4>] lr : [<c0019f54>] psr: 600000d3 sp : c035bf10 ip : c035bf20 fp : c035bf1c r10: c035bfd4 r9 : 54029252 r8 : c03f8120 r7 : c0362b50 r6 : 00b71b00 r5 : c03873cc r4 : c0362b40 r3 : 00000000 r2 : c0362b40 r1 : 0000010a r0 : 00002cb0 Flags: nZCv IRQs off FIQs off Mode SVC_32 ISA ARM Segment kernel Control: 0000317f Table: 10004000 DAC: 00000017 Process swapper (pid: 0, stack limit = 0xc035a270) Stack: (0xc035bf10 to 0xc035c000) bf00: c035bf3c c035bf20 c0019f54 c001b0ac bf20: 00001000 00002cb3 00000004 c035ed4c c035bf74 c035bf40 c033ea24 c0019edc bf40: c02f526c 00000002 00000015 bc058c9b 93111a16 c035335c 02000000 c035ed4c bf60: c035ed4c c03f8120 c035bf84 c035bf78 c00194c4 c033e8ec c035bfc4 c035bf88 bf80: c033bc24 c00194a0 c035bf90 c035bf98 00000000 00000000 00000000 00000000 bfa0: 00000001 00000000 c0354678 c035ece4 10004000 103532f4 c035bff4 c035bfc8 bfc0: c0338574 c033b598 00000000 00000000 00000000 c035467c 0000317d c035c03c bfe0: c0354678 c035ece4 00000000 c035bff8 10008040 c0338508 00000000 00000000 Backtrace: [<c001b09c>] (omap_sram_reprogram_clock+0x0/0x30) from [<c0019f54>] (omap1_select_table_rate+0x88/0xb4) [<c0019ecc>] (omap1_select_table_rate+0x0/0xb4) from [<c033ea24>] (omap1_clk_init+0x148/0x334) r7:c035ed4c r6:00000004 r5:00002cb3 r4:00001000 [<c033e8dc>] (omap1_clk_init+0x0/0x334) from [<c00194c4>] (omap1_init_early+0x34/0x48) r8:c03f8120 r7:c035ed4c r6:c035ed4c r5:02000000 r4:c035335c [<c0019490>] (omap1_init_early+0x0/0x48) from [<c033bc24>] (setup_arch+0x69c/0x79c) [<c033b588>] (setup_arch+0x0/0x79c) from [<c0338574>] (start_kernel+0x7c/0x2f4) [<c03384f8>] (start_kernel+0x0/0x2f4) from [<10008040>] (0x10008040) r7:c035ece4 r6:c0354678 r5:c035c03c r4:0000317d Code: 0a000002 e1a0e00f e12fff13 e89da800 (e7f001f2) Fix this by adding omap1_clk_late_init() that only reprograms dpll1 if the bootloader rate is less than 60MHz. This also allows removing of the OMAP_CLOCKS_SET_BY_BOOTLOADER option. Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi> Tested-by: Aaro Koskinen <aaro.koskinen@iki.fi> Signed-off-by: Tony Lindgren <tony@atomide.com>
2011-11-11 18:15:11 +00:00
static void __init omap1_show_rates(void)
{
pr_notice("Clocking rate (xtal/DPLL1/MPU): %ld.%01ld/%ld.%01ld/%ld.%01ld MHz\n",
ck_ref.rate / 1000000, (ck_ref.rate / 100000) % 10,
ck_dpll1.rate / 1000000, (ck_dpll1.rate / 100000) % 10,
arm_ck.rate / 1000000, (arm_ck.rate / 100000) % 10);
ARM: OMAP: Fix reprogramming of dpll1 rate Commit a66cb3454f220f49f900646ebdc76cb943319eb7 (ARM: OMAP: Map SRAM later on with ioremap_exec()) moved the SRAM init to happen later to remove a dependency to early SoC detection for map_io. This broke booting on some boards not using Kconfig option for OMAP_CLOCKS_SET_BY_BOOTLOADER as the dpll1 reprogramming would cause the following error: kernel BUG at arch/arm/plat-omap/sram.c:226! Internal error: Oops - undefined instruction: 0 [#1] PREEMPT Modules linked in: CPU: 0 Not tainted (3.2.0-rc1-e3 #9) PC is at omap_sram_reprogram_clock+0x28/0x30 LR is at omap1_select_table_rate+0x88/0xb4 pc : [<c001b0c4>] lr : [<c0019f54>] psr: 600000d3 sp : c035bf10 ip : c035bf20 fp : c035bf1c r10: c035bfd4 r9 : 54029252 r8 : c03f8120 r7 : c0362b50 r6 : 00b71b00 r5 : c03873cc r4 : c0362b40 r3 : 00000000 r2 : c0362b40 r1 : 0000010a r0 : 00002cb0 Flags: nZCv IRQs off FIQs off Mode SVC_32 ISA ARM Segment kernel Control: 0000317f Table: 10004000 DAC: 00000017 Process swapper (pid: 0, stack limit = 0xc035a270) Stack: (0xc035bf10 to 0xc035c000) bf00: c035bf3c c035bf20 c0019f54 c001b0ac bf20: 00001000 00002cb3 00000004 c035ed4c c035bf74 c035bf40 c033ea24 c0019edc bf40: c02f526c 00000002 00000015 bc058c9b 93111a16 c035335c 02000000 c035ed4c bf60: c035ed4c c03f8120 c035bf84 c035bf78 c00194c4 c033e8ec c035bfc4 c035bf88 bf80: c033bc24 c00194a0 c035bf90 c035bf98 00000000 00000000 00000000 00000000 bfa0: 00000001 00000000 c0354678 c035ece4 10004000 103532f4 c035bff4 c035bfc8 bfc0: c0338574 c033b598 00000000 00000000 00000000 c035467c 0000317d c035c03c bfe0: c0354678 c035ece4 00000000 c035bff8 10008040 c0338508 00000000 00000000 Backtrace: [<c001b09c>] (omap_sram_reprogram_clock+0x0/0x30) from [<c0019f54>] (omap1_select_table_rate+0x88/0xb4) [<c0019ecc>] (omap1_select_table_rate+0x0/0xb4) from [<c033ea24>] (omap1_clk_init+0x148/0x334) r7:c035ed4c r6:00000004 r5:00002cb3 r4:00001000 [<c033e8dc>] (omap1_clk_init+0x0/0x334) from [<c00194c4>] (omap1_init_early+0x34/0x48) r8:c03f8120 r7:c035ed4c r6:c035ed4c r5:02000000 r4:c035335c [<c0019490>] (omap1_init_early+0x0/0x48) from [<c033bc24>] (setup_arch+0x69c/0x79c) [<c033b588>] (setup_arch+0x0/0x79c) from [<c0338574>] (start_kernel+0x7c/0x2f4) [<c03384f8>] (start_kernel+0x0/0x2f4) from [<10008040>] (0x10008040) r7:c035ece4 r6:c0354678 r5:c035c03c r4:0000317d Code: 0a000002 e1a0e00f e12fff13 e89da800 (e7f001f2) Fix this by adding omap1_clk_late_init() that only reprograms dpll1 if the bootloader rate is less than 60MHz. This also allows removing of the OMAP_CLOCKS_SET_BY_BOOTLOADER option. Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi> Tested-by: Aaro Koskinen <aaro.koskinen@iki.fi> Signed-off-by: Tony Lindgren <tony@atomide.com>
2011-11-11 18:15:11 +00:00
}
u32 cpu_mask;
int __init omap1_clk_init(void)
{
struct omap_clk *c;
u32 reg;
#ifdef CONFIG_DEBUG_LL
/* Make sure UART clocks are enabled early */
if (cpu_is_omap16xx())
omap_writel(omap_readl(MOD_CONF_CTRL_0) |
CONF_MOD_UART1_CLK_MODE_R |
CONF_MOD_UART3_CLK_MODE_R, MOD_CONF_CTRL_0);
#endif
/* USB_REQ_EN will be disabled later if necessary (usb_dc_ck) */
reg = omap_readw(SOFT_REQ_REG) & (1 << 4);
omap_writew(reg, SOFT_REQ_REG);
if (!cpu_is_omap15xx())
omap_writew(0, SOFT_REQ_REG2);
/* By default all idlect1 clocks are allowed to idle */
arm_idlect1_mask = ~0;
cpu_mask = 0;
if (cpu_is_omap1710())
cpu_mask |= CK_1710;
if (cpu_is_omap16xx())
cpu_mask |= CK_16XX;
if (cpu_is_omap1510())
cpu_mask |= CK_1510;
if (cpu_is_omap310())
cpu_mask |= CK_310;
/* Pointers to these clocks are needed by code in clock.c */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
api_ck_p = &api_ck.clk;
ck_dpll1_p = &ck_dpll1;
ck_ref_p = &ck_ref;
pr_info("Clocks: ARM_SYSST: 0x%04x DPLL_CTL: 0x%04x ARM_CKCTL: 0x%04x\n",
omap_readw(ARM_SYSST), omap_readw(DPLL_CTL),
omap_readw(ARM_CKCTL));
/* We want to be in synchronous scalable mode */
omap_writew(0x1000, ARM_SYSST);
ARM: OMAP: Fix reprogramming of dpll1 rate Commit a66cb3454f220f49f900646ebdc76cb943319eb7 (ARM: OMAP: Map SRAM later on with ioremap_exec()) moved the SRAM init to happen later to remove a dependency to early SoC detection for map_io. This broke booting on some boards not using Kconfig option for OMAP_CLOCKS_SET_BY_BOOTLOADER as the dpll1 reprogramming would cause the following error: kernel BUG at arch/arm/plat-omap/sram.c:226! Internal error: Oops - undefined instruction: 0 [#1] PREEMPT Modules linked in: CPU: 0 Not tainted (3.2.0-rc1-e3 #9) PC is at omap_sram_reprogram_clock+0x28/0x30 LR is at omap1_select_table_rate+0x88/0xb4 pc : [<c001b0c4>] lr : [<c0019f54>] psr: 600000d3 sp : c035bf10 ip : c035bf20 fp : c035bf1c r10: c035bfd4 r9 : 54029252 r8 : c03f8120 r7 : c0362b50 r6 : 00b71b00 r5 : c03873cc r4 : c0362b40 r3 : 00000000 r2 : c0362b40 r1 : 0000010a r0 : 00002cb0 Flags: nZCv IRQs off FIQs off Mode SVC_32 ISA ARM Segment kernel Control: 0000317f Table: 10004000 DAC: 00000017 Process swapper (pid: 0, stack limit = 0xc035a270) Stack: (0xc035bf10 to 0xc035c000) bf00: c035bf3c c035bf20 c0019f54 c001b0ac bf20: 00001000 00002cb3 00000004 c035ed4c c035bf74 c035bf40 c033ea24 c0019edc bf40: c02f526c 00000002 00000015 bc058c9b 93111a16 c035335c 02000000 c035ed4c bf60: c035ed4c c03f8120 c035bf84 c035bf78 c00194c4 c033e8ec c035bfc4 c035bf88 bf80: c033bc24 c00194a0 c035bf90 c035bf98 00000000 00000000 00000000 00000000 bfa0: 00000001 00000000 c0354678 c035ece4 10004000 103532f4 c035bff4 c035bfc8 bfc0: c0338574 c033b598 00000000 00000000 00000000 c035467c 0000317d c035c03c bfe0: c0354678 c035ece4 00000000 c035bff8 10008040 c0338508 00000000 00000000 Backtrace: [<c001b09c>] (omap_sram_reprogram_clock+0x0/0x30) from [<c0019f54>] (omap1_select_table_rate+0x88/0xb4) [<c0019ecc>] (omap1_select_table_rate+0x0/0xb4) from [<c033ea24>] (omap1_clk_init+0x148/0x334) r7:c035ed4c r6:00000004 r5:00002cb3 r4:00001000 [<c033e8dc>] (omap1_clk_init+0x0/0x334) from [<c00194c4>] (omap1_init_early+0x34/0x48) r8:c03f8120 r7:c035ed4c r6:c035ed4c r5:02000000 r4:c035335c [<c0019490>] (omap1_init_early+0x0/0x48) from [<c033bc24>] (setup_arch+0x69c/0x79c) [<c033b588>] (setup_arch+0x0/0x79c) from [<c0338574>] (start_kernel+0x7c/0x2f4) [<c03384f8>] (start_kernel+0x0/0x2f4) from [<10008040>] (0x10008040) r7:c035ece4 r6:c0354678 r5:c035c03c r4:0000317d Code: 0a000002 e1a0e00f e12fff13 e89da800 (e7f001f2) Fix this by adding omap1_clk_late_init() that only reprograms dpll1 if the bootloader rate is less than 60MHz. This also allows removing of the OMAP_CLOCKS_SET_BY_BOOTLOADER option. Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi> Tested-by: Aaro Koskinen <aaro.koskinen@iki.fi> Signed-off-by: Tony Lindgren <tony@atomide.com>
2011-11-11 18:15:11 +00:00
/*
* Initially use the values set by bootloader. Determine PLL rate and
* recalculate dependent clocks as if kernel had changed PLL or
* divisors. See also omap1_clk_late_init() that can reprogram dpll1
* after the SRAM is initialized.
*/
{
unsigned pll_ctl_val = omap_readw(DPLL_CTL);
ck_dpll1.rate = ck_ref.rate; /* Base xtal rate */
if (pll_ctl_val & 0x10) {
/* PLL enabled, apply multiplier and divisor */
if (pll_ctl_val & 0xf80)
ck_dpll1.rate *= (pll_ctl_val & 0xf80) >> 7;
ck_dpll1.rate /= ((pll_ctl_val & 0x60) >> 5) + 1;
} else {
/* PLL disabled, apply bypass divisor */
switch (pll_ctl_val & 0xc) {
case 0:
break;
case 0x4:
ck_dpll1.rate /= 2;
break;
default:
ck_dpll1.rate /= 4;
break;
}
}
}
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
/* Amstrad Delta wants BCLK high when inactive */
if (machine_is_ams_delta())
omap_writel(omap_readl(ULPD_CLOCK_CTRL) |
(1 << SDW_MCLK_INV_BIT),
ULPD_CLOCK_CTRL);
/* Turn off DSP and ARM_TIMXO. Make sure ARM_INTHCK is not divided */
omap_writew(omap_readw(ARM_CKCTL) & 0x0fff, ARM_CKCTL);
/* Put DSP/MPUI into reset until needed */
omap_writew(0, ARM_RSTCT1);
omap_writew(1, ARM_RSTCT2);
omap_writew(0x400, ARM_IDLECT1);
/*
* According to OMAP5910 Erratum SYS_DMA_1, bit DMACK_REQ (bit 8)
* of the ARM_IDLECT2 register must be set to zero. The power-on
* default value of this bit is one.
*/
omap_writew(0x0000, ARM_IDLECT2); /* Turn LCD clock off also */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
for (c = omap_clks; c < omap_clks + ARRAY_SIZE(omap_clks); c++) {
if (!(c->cpu & cpu_mask))
continue;
if (c->lk.clk_hw->init) { /* NULL if provider already registered */
const struct clk_init_data *init = c->lk.clk_hw->init;
const char *name = c->lk.clk_hw->init->name;
int err;
err = clk_hw_register(NULL, c->lk.clk_hw);
if (err < 0) {
pr_err("failed to register clock \"%s\"! (%d)\n", name, err);
/* may be tried again, restore init data */
c->lk.clk_hw->init = init;
continue;
}
}
clk_hw_register_clkdev(c->lk.clk_hw, c->lk.con_id, c->lk.dev_id);
}
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
omap1_show_rates();
return 0;
}
ARM: OMAP: Fix reprogramming of dpll1 rate Commit a66cb3454f220f49f900646ebdc76cb943319eb7 (ARM: OMAP: Map SRAM later on with ioremap_exec()) moved the SRAM init to happen later to remove a dependency to early SoC detection for map_io. This broke booting on some boards not using Kconfig option for OMAP_CLOCKS_SET_BY_BOOTLOADER as the dpll1 reprogramming would cause the following error: kernel BUG at arch/arm/plat-omap/sram.c:226! Internal error: Oops - undefined instruction: 0 [#1] PREEMPT Modules linked in: CPU: 0 Not tainted (3.2.0-rc1-e3 #9) PC is at omap_sram_reprogram_clock+0x28/0x30 LR is at omap1_select_table_rate+0x88/0xb4 pc : [<c001b0c4>] lr : [<c0019f54>] psr: 600000d3 sp : c035bf10 ip : c035bf20 fp : c035bf1c r10: c035bfd4 r9 : 54029252 r8 : c03f8120 r7 : c0362b50 r6 : 00b71b00 r5 : c03873cc r4 : c0362b40 r3 : 00000000 r2 : c0362b40 r1 : 0000010a r0 : 00002cb0 Flags: nZCv IRQs off FIQs off Mode SVC_32 ISA ARM Segment kernel Control: 0000317f Table: 10004000 DAC: 00000017 Process swapper (pid: 0, stack limit = 0xc035a270) Stack: (0xc035bf10 to 0xc035c000) bf00: c035bf3c c035bf20 c0019f54 c001b0ac bf20: 00001000 00002cb3 00000004 c035ed4c c035bf74 c035bf40 c033ea24 c0019edc bf40: c02f526c 00000002 00000015 bc058c9b 93111a16 c035335c 02000000 c035ed4c bf60: c035ed4c c03f8120 c035bf84 c035bf78 c00194c4 c033e8ec c035bfc4 c035bf88 bf80: c033bc24 c00194a0 c035bf90 c035bf98 00000000 00000000 00000000 00000000 bfa0: 00000001 00000000 c0354678 c035ece4 10004000 103532f4 c035bff4 c035bfc8 bfc0: c0338574 c033b598 00000000 00000000 00000000 c035467c 0000317d c035c03c bfe0: c0354678 c035ece4 00000000 c035bff8 10008040 c0338508 00000000 00000000 Backtrace: [<c001b09c>] (omap_sram_reprogram_clock+0x0/0x30) from [<c0019f54>] (omap1_select_table_rate+0x88/0xb4) [<c0019ecc>] (omap1_select_table_rate+0x0/0xb4) from [<c033ea24>] (omap1_clk_init+0x148/0x334) r7:c035ed4c r6:00000004 r5:00002cb3 r4:00001000 [<c033e8dc>] (omap1_clk_init+0x0/0x334) from [<c00194c4>] (omap1_init_early+0x34/0x48) r8:c03f8120 r7:c035ed4c r6:c035ed4c r5:02000000 r4:c035335c [<c0019490>] (omap1_init_early+0x0/0x48) from [<c033bc24>] (setup_arch+0x69c/0x79c) [<c033b588>] (setup_arch+0x0/0x79c) from [<c0338574>] (start_kernel+0x7c/0x2f4) [<c03384f8>] (start_kernel+0x0/0x2f4) from [<10008040>] (0x10008040) r7:c035ece4 r6:c0354678 r5:c035c03c r4:0000317d Code: 0a000002 e1a0e00f e12fff13 e89da800 (e7f001f2) Fix this by adding omap1_clk_late_init() that only reprograms dpll1 if the bootloader rate is less than 60MHz. This also allows removing of the OMAP_CLOCKS_SET_BY_BOOTLOADER option. Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi> Tested-by: Aaro Koskinen <aaro.koskinen@iki.fi> Signed-off-by: Tony Lindgren <tony@atomide.com>
2011-11-11 18:15:11 +00:00
#define OMAP1_DPLL1_SANE_VALUE 60000000
void __init omap1_clk_late_init(void)
{
unsigned long rate = ck_dpll1.rate;
ARM: OMAP: Fix reprogramming of dpll1 rate Commit a66cb3454f220f49f900646ebdc76cb943319eb7 (ARM: OMAP: Map SRAM later on with ioremap_exec()) moved the SRAM init to happen later to remove a dependency to early SoC detection for map_io. This broke booting on some boards not using Kconfig option for OMAP_CLOCKS_SET_BY_BOOTLOADER as the dpll1 reprogramming would cause the following error: kernel BUG at arch/arm/plat-omap/sram.c:226! Internal error: Oops - undefined instruction: 0 [#1] PREEMPT Modules linked in: CPU: 0 Not tainted (3.2.0-rc1-e3 #9) PC is at omap_sram_reprogram_clock+0x28/0x30 LR is at omap1_select_table_rate+0x88/0xb4 pc : [<c001b0c4>] lr : [<c0019f54>] psr: 600000d3 sp : c035bf10 ip : c035bf20 fp : c035bf1c r10: c035bfd4 r9 : 54029252 r8 : c03f8120 r7 : c0362b50 r6 : 00b71b00 r5 : c03873cc r4 : c0362b40 r3 : 00000000 r2 : c0362b40 r1 : 0000010a r0 : 00002cb0 Flags: nZCv IRQs off FIQs off Mode SVC_32 ISA ARM Segment kernel Control: 0000317f Table: 10004000 DAC: 00000017 Process swapper (pid: 0, stack limit = 0xc035a270) Stack: (0xc035bf10 to 0xc035c000) bf00: c035bf3c c035bf20 c0019f54 c001b0ac bf20: 00001000 00002cb3 00000004 c035ed4c c035bf74 c035bf40 c033ea24 c0019edc bf40: c02f526c 00000002 00000015 bc058c9b 93111a16 c035335c 02000000 c035ed4c bf60: c035ed4c c03f8120 c035bf84 c035bf78 c00194c4 c033e8ec c035bfc4 c035bf88 bf80: c033bc24 c00194a0 c035bf90 c035bf98 00000000 00000000 00000000 00000000 bfa0: 00000001 00000000 c0354678 c035ece4 10004000 103532f4 c035bff4 c035bfc8 bfc0: c0338574 c033b598 00000000 00000000 00000000 c035467c 0000317d c035c03c bfe0: c0354678 c035ece4 00000000 c035bff8 10008040 c0338508 00000000 00000000 Backtrace: [<c001b09c>] (omap_sram_reprogram_clock+0x0/0x30) from [<c0019f54>] (omap1_select_table_rate+0x88/0xb4) [<c0019ecc>] (omap1_select_table_rate+0x0/0xb4) from [<c033ea24>] (omap1_clk_init+0x148/0x334) r7:c035ed4c r6:00000004 r5:00002cb3 r4:00001000 [<c033e8dc>] (omap1_clk_init+0x0/0x334) from [<c00194c4>] (omap1_init_early+0x34/0x48) r8:c03f8120 r7:c035ed4c r6:c035ed4c r5:02000000 r4:c035335c [<c0019490>] (omap1_init_early+0x0/0x48) from [<c033bc24>] (setup_arch+0x69c/0x79c) [<c033b588>] (setup_arch+0x0/0x79c) from [<c0338574>] (start_kernel+0x7c/0x2f4) [<c03384f8>] (start_kernel+0x0/0x2f4) from [<10008040>] (0x10008040) r7:c035ece4 r6:c0354678 r5:c035c03c r4:0000317d Code: 0a000002 e1a0e00f e12fff13 e89da800 (e7f001f2) Fix this by adding omap1_clk_late_init() that only reprograms dpll1 if the bootloader rate is less than 60MHz. This also allows removing of the OMAP_CLOCKS_SET_BY_BOOTLOADER option. Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi> Tested-by: Aaro Koskinen <aaro.koskinen@iki.fi> Signed-off-by: Tony Lindgren <tony@atomide.com>
2011-11-11 18:15:11 +00:00
/* Find the highest supported frequency and enable it */
ARM: OMAP1: clock: Convert to CCF OMAP1 still uses its own implementation of standard clock API defined in include/linux/clk.h. Internals of that implementation are not visible outside OMAP1 directory. As a consequence, device drivers are not able to register clocks potentially provided by peripheral devices. Drop OMAP1 implementation of the clock API and enable common clock framework. Modify the remaining low level code to be compatible with clock provider API and register the clocks with CCF. Move initialisation of clocks to omap1_timer_init() to avoid memory allocation issues at early setup phase from where omap1_init_early() is called. Register the clocks after initialization of clock I/O registers, local clock pointers used by OMAP1 clock ops, and local .rate fields of clocks with no local implementation of .recalc ops, so CCF structures are populated with correct data during clock registration. Instead of enabling some of the registered clocks, flag them for CCF as critical. Introduce .is_enabled op using code that verifies hardware status of clock enablement, split out from implementation of .disable_unused op, so the latter is actually called by CCF for not requested but hardware enabled clocks. Add .round_rate ops where missing so .set_rate ops are called by CCF as expected. Since CCF allows parallel execution of .enable/.disable and .set_rate ops, protect registers shared among those groups of ops from concurrent access with spinlocks. Drop local debugfs support in favor of that provided by CCF. v2: flag tc2_ck as CLK_IS_CRITICAL (Aaro) v3: rebase on top of soc/omap1-multiplatform-5.18, - drop no longer needed includes from arch/arm/mach-omap1/io.c Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2022-04-10 13:07:57 +00:00
if (omap1_select_table_rate(&virtual_ck_mpu, ~0, arm_ck.rate)) {
ARM: OMAP: Fix reprogramming of dpll1 rate Commit a66cb3454f220f49f900646ebdc76cb943319eb7 (ARM: OMAP: Map SRAM later on with ioremap_exec()) moved the SRAM init to happen later to remove a dependency to early SoC detection for map_io. This broke booting on some boards not using Kconfig option for OMAP_CLOCKS_SET_BY_BOOTLOADER as the dpll1 reprogramming would cause the following error: kernel BUG at arch/arm/plat-omap/sram.c:226! Internal error: Oops - undefined instruction: 0 [#1] PREEMPT Modules linked in: CPU: 0 Not tainted (3.2.0-rc1-e3 #9) PC is at omap_sram_reprogram_clock+0x28/0x30 LR is at omap1_select_table_rate+0x88/0xb4 pc : [<c001b0c4>] lr : [<c0019f54>] psr: 600000d3 sp : c035bf10 ip : c035bf20 fp : c035bf1c r10: c035bfd4 r9 : 54029252 r8 : c03f8120 r7 : c0362b50 r6 : 00b71b00 r5 : c03873cc r4 : c0362b40 r3 : 00000000 r2 : c0362b40 r1 : 0000010a r0 : 00002cb0 Flags: nZCv IRQs off FIQs off Mode SVC_32 ISA ARM Segment kernel Control: 0000317f Table: 10004000 DAC: 00000017 Process swapper (pid: 0, stack limit = 0xc035a270) Stack: (0xc035bf10 to 0xc035c000) bf00: c035bf3c c035bf20 c0019f54 c001b0ac bf20: 00001000 00002cb3 00000004 c035ed4c c035bf74 c035bf40 c033ea24 c0019edc bf40: c02f526c 00000002 00000015 bc058c9b 93111a16 c035335c 02000000 c035ed4c bf60: c035ed4c c03f8120 c035bf84 c035bf78 c00194c4 c033e8ec c035bfc4 c035bf88 bf80: c033bc24 c00194a0 c035bf90 c035bf98 00000000 00000000 00000000 00000000 bfa0: 00000001 00000000 c0354678 c035ece4 10004000 103532f4 c035bff4 c035bfc8 bfc0: c0338574 c033b598 00000000 00000000 00000000 c035467c 0000317d c035c03c bfe0: c0354678 c035ece4 00000000 c035bff8 10008040 c0338508 00000000 00000000 Backtrace: [<c001b09c>] (omap_sram_reprogram_clock+0x0/0x30) from [<c0019f54>] (omap1_select_table_rate+0x88/0xb4) [<c0019ecc>] (omap1_select_table_rate+0x0/0xb4) from [<c033ea24>] (omap1_clk_init+0x148/0x334) r7:c035ed4c r6:00000004 r5:00002cb3 r4:00001000 [<c033e8dc>] (omap1_clk_init+0x0/0x334) from [<c00194c4>] (omap1_init_early+0x34/0x48) r8:c03f8120 r7:c035ed4c r6:c035ed4c r5:02000000 r4:c035335c [<c0019490>] (omap1_init_early+0x0/0x48) from [<c033bc24>] (setup_arch+0x69c/0x79c) [<c033b588>] (setup_arch+0x0/0x79c) from [<c0338574>] (start_kernel+0x7c/0x2f4) [<c03384f8>] (start_kernel+0x0/0x2f4) from [<10008040>] (0x10008040) r7:c035ece4 r6:c0354678 r5:c035c03c r4:0000317d Code: 0a000002 e1a0e00f e12fff13 e89da800 (e7f001f2) Fix this by adding omap1_clk_late_init() that only reprograms dpll1 if the bootloader rate is less than 60MHz. This also allows removing of the OMAP_CLOCKS_SET_BY_BOOTLOADER option. Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi> Tested-by: Aaro Koskinen <aaro.koskinen@iki.fi> Signed-off-by: Tony Lindgren <tony@atomide.com>
2011-11-11 18:15:11 +00:00
pr_err("System frequencies not set, using default. Check your config.\n");
/*
* Reprogramming the DPLL is tricky, it must be done from SRAM.
*/
omap_sram_reprogram_clock(0x2290, 0x0005);
ARM: OMAP: Fix reprogramming of dpll1 rate Commit a66cb3454f220f49f900646ebdc76cb943319eb7 (ARM: OMAP: Map SRAM later on with ioremap_exec()) moved the SRAM init to happen later to remove a dependency to early SoC detection for map_io. This broke booting on some boards not using Kconfig option for OMAP_CLOCKS_SET_BY_BOOTLOADER as the dpll1 reprogramming would cause the following error: kernel BUG at arch/arm/plat-omap/sram.c:226! Internal error: Oops - undefined instruction: 0 [#1] PREEMPT Modules linked in: CPU: 0 Not tainted (3.2.0-rc1-e3 #9) PC is at omap_sram_reprogram_clock+0x28/0x30 LR is at omap1_select_table_rate+0x88/0xb4 pc : [<c001b0c4>] lr : [<c0019f54>] psr: 600000d3 sp : c035bf10 ip : c035bf20 fp : c035bf1c r10: c035bfd4 r9 : 54029252 r8 : c03f8120 r7 : c0362b50 r6 : 00b71b00 r5 : c03873cc r4 : c0362b40 r3 : 00000000 r2 : c0362b40 r1 : 0000010a r0 : 00002cb0 Flags: nZCv IRQs off FIQs off Mode SVC_32 ISA ARM Segment kernel Control: 0000317f Table: 10004000 DAC: 00000017 Process swapper (pid: 0, stack limit = 0xc035a270) Stack: (0xc035bf10 to 0xc035c000) bf00: c035bf3c c035bf20 c0019f54 c001b0ac bf20: 00001000 00002cb3 00000004 c035ed4c c035bf74 c035bf40 c033ea24 c0019edc bf40: c02f526c 00000002 00000015 bc058c9b 93111a16 c035335c 02000000 c035ed4c bf60: c035ed4c c03f8120 c035bf84 c035bf78 c00194c4 c033e8ec c035bfc4 c035bf88 bf80: c033bc24 c00194a0 c035bf90 c035bf98 00000000 00000000 00000000 00000000 bfa0: 00000001 00000000 c0354678 c035ece4 10004000 103532f4 c035bff4 c035bfc8 bfc0: c0338574 c033b598 00000000 00000000 00000000 c035467c 0000317d c035c03c bfe0: c0354678 c035ece4 00000000 c035bff8 10008040 c0338508 00000000 00000000 Backtrace: [<c001b09c>] (omap_sram_reprogram_clock+0x0/0x30) from [<c0019f54>] (omap1_select_table_rate+0x88/0xb4) [<c0019ecc>] (omap1_select_table_rate+0x0/0xb4) from [<c033ea24>] (omap1_clk_init+0x148/0x334) r7:c035ed4c r6:00000004 r5:00002cb3 r4:00001000 [<c033e8dc>] (omap1_clk_init+0x0/0x334) from [<c00194c4>] (omap1_init_early+0x34/0x48) r8:c03f8120 r7:c035ed4c r6:c035ed4c r5:02000000 r4:c035335c [<c0019490>] (omap1_init_early+0x0/0x48) from [<c033bc24>] (setup_arch+0x69c/0x79c) [<c033b588>] (setup_arch+0x0/0x79c) from [<c0338574>] (start_kernel+0x7c/0x2f4) [<c03384f8>] (start_kernel+0x0/0x2f4) from [<10008040>] (0x10008040) r7:c035ece4 r6:c0354678 r5:c035c03c r4:0000317d Code: 0a000002 e1a0e00f e12fff13 e89da800 (e7f001f2) Fix this by adding omap1_clk_late_init() that only reprograms dpll1 if the bootloader rate is less than 60MHz. This also allows removing of the OMAP_CLOCKS_SET_BY_BOOTLOADER option. Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi> Tested-by: Aaro Koskinen <aaro.koskinen@iki.fi> Signed-off-by: Tony Lindgren <tony@atomide.com>
2011-11-11 18:15:11 +00:00
ck_dpll1.rate = OMAP1_DPLL1_SANE_VALUE;
}
propagate_rate(&ck_dpll1);
omap1_show_rates();
loops_per_jiffy = cpufreq_scale(loops_per_jiffy, rate, ck_dpll1.rate);
ARM: OMAP: Fix reprogramming of dpll1 rate Commit a66cb3454f220f49f900646ebdc76cb943319eb7 (ARM: OMAP: Map SRAM later on with ioremap_exec()) moved the SRAM init to happen later to remove a dependency to early SoC detection for map_io. This broke booting on some boards not using Kconfig option for OMAP_CLOCKS_SET_BY_BOOTLOADER as the dpll1 reprogramming would cause the following error: kernel BUG at arch/arm/plat-omap/sram.c:226! Internal error: Oops - undefined instruction: 0 [#1] PREEMPT Modules linked in: CPU: 0 Not tainted (3.2.0-rc1-e3 #9) PC is at omap_sram_reprogram_clock+0x28/0x30 LR is at omap1_select_table_rate+0x88/0xb4 pc : [<c001b0c4>] lr : [<c0019f54>] psr: 600000d3 sp : c035bf10 ip : c035bf20 fp : c035bf1c r10: c035bfd4 r9 : 54029252 r8 : c03f8120 r7 : c0362b50 r6 : 00b71b00 r5 : c03873cc r4 : c0362b40 r3 : 00000000 r2 : c0362b40 r1 : 0000010a r0 : 00002cb0 Flags: nZCv IRQs off FIQs off Mode SVC_32 ISA ARM Segment kernel Control: 0000317f Table: 10004000 DAC: 00000017 Process swapper (pid: 0, stack limit = 0xc035a270) Stack: (0xc035bf10 to 0xc035c000) bf00: c035bf3c c035bf20 c0019f54 c001b0ac bf20: 00001000 00002cb3 00000004 c035ed4c c035bf74 c035bf40 c033ea24 c0019edc bf40: c02f526c 00000002 00000015 bc058c9b 93111a16 c035335c 02000000 c035ed4c bf60: c035ed4c c03f8120 c035bf84 c035bf78 c00194c4 c033e8ec c035bfc4 c035bf88 bf80: c033bc24 c00194a0 c035bf90 c035bf98 00000000 00000000 00000000 00000000 bfa0: 00000001 00000000 c0354678 c035ece4 10004000 103532f4 c035bff4 c035bfc8 bfc0: c0338574 c033b598 00000000 00000000 00000000 c035467c 0000317d c035c03c bfe0: c0354678 c035ece4 00000000 c035bff8 10008040 c0338508 00000000 00000000 Backtrace: [<c001b09c>] (omap_sram_reprogram_clock+0x0/0x30) from [<c0019f54>] (omap1_select_table_rate+0x88/0xb4) [<c0019ecc>] (omap1_select_table_rate+0x0/0xb4) from [<c033ea24>] (omap1_clk_init+0x148/0x334) r7:c035ed4c r6:00000004 r5:00002cb3 r4:00001000 [<c033e8dc>] (omap1_clk_init+0x0/0x334) from [<c00194c4>] (omap1_init_early+0x34/0x48) r8:c03f8120 r7:c035ed4c r6:c035ed4c r5:02000000 r4:c035335c [<c0019490>] (omap1_init_early+0x0/0x48) from [<c033bc24>] (setup_arch+0x69c/0x79c) [<c033b588>] (setup_arch+0x0/0x79c) from [<c0338574>] (start_kernel+0x7c/0x2f4) [<c03384f8>] (start_kernel+0x0/0x2f4) from [<10008040>] (0x10008040) r7:c035ece4 r6:c0354678 r5:c035c03c r4:0000317d Code: 0a000002 e1a0e00f e12fff13 e89da800 (e7f001f2) Fix this by adding omap1_clk_late_init() that only reprograms dpll1 if the bootloader rate is less than 60MHz. This also allows removing of the OMAP_CLOCKS_SET_BY_BOOTLOADER option. Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi> Tested-by: Aaro Koskinen <aaro.koskinen@iki.fi> Signed-off-by: Tony Lindgren <tony@atomide.com>
2011-11-11 18:15:11 +00:00
}