* Re: [PATCH 5/5] libata/drivers: Add pata_macio, driver Apple PowerMac/PowerBook IDE controller
From: Andreas Schwab @ 2009-12-02 11:44 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: tj, linux-ide, linuxppc-dev, jeff
In-Reply-To: <20091202003630.29919100851@ozlabs.org>
Benjamin Herrenschmidt <benh@kernel.crashing.org> writes:
> v2. Better tested now, seems to be reasonably solid.
>
> Addressed Tejun comments and made remove more robust vs. media-bay,
> should also fix Andreas problem.
Thanks, this works fine now, including suspend *and* resume. :-)
Andreas.
--
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5
"And now for something completely different."
^ permalink raw reply
* Can user set timeout for MPC 83x watchdog timer
From: Aswathi C @ 2009-12-02 11:49 UTC (permalink / raw)
To: linuxppc-dev
[-- Attachment #1: Type: text/plain, Size: 219 bytes --]
I would like to know whether user can set timeout from user space for the
watchdog.
I am using MPC 8323E processor and linux-2.6.24 kernel.
If user can set, can you tell me how will it be implemented?
Regards,
Aswathi
[-- Attachment #2: Type: text/html, Size: 266 bytes --]
^ permalink raw reply
* Can user set timeout for MPC 83x watchdog timer
From: Aswathi C @ 2009-12-02 11:13 UTC (permalink / raw)
To: galak; +Cc: linuxppc-dev
[-- Attachment #1: Type: text/plain, Size: 230 bytes --]
Hi Kumar,
I would like to know whether user can set timeout from user space for the
watchdog.
I am using MPC 8323E processor and linux-2.6.24 kernel.
If user can set, can you tell me how will it be implemented?
Regards,
Aswathi
[-- Attachment #2: Type: text/html, Size: 258 bytes --]
^ permalink raw reply
* [v10 PATCH 9/9]: POWER: Enable default_idle when power_save=off
From: Arun R Bharadwaj @ 2009-12-02 10:03 UTC (permalink / raw)
To: Peter Zijlstra, Benjamin Herrenschmidt, Ingo Molnar,
Vaidyanathan Srinivasan, Dipankar Sarma, Balbir Singh,
Venkatesh Pallipadi, Arun Bharadwaj
Cc: linux-arch, linux-acpi, linuxppc-dev, linux-kernel
In-Reply-To: <20091202095427.GA27251@linux.vnet.ibm.com>
* Arun R Bharadwaj <arun@linux.vnet.ibm.com> [2009-12-02 15:24:27]:
This patch enables default_idle when power_save=off kernel boot
option is specified.
Earlier, this was done by setting ppc_md.power_save = NULL and hence
HMT_low() and HMT_very_low() was called. Now this is defined under
default_idle() and hence by setting boot_option_idle_override = 1,
the cpuidle registration stuff does not happen and hence default_idle
is chosen in cpuidle_idle_call.
Signed-off-by: Arun R Bharadwaj <arun@linux.vnet.ibm.com>
---
arch/powerpc/include/asm/processor.h | 2 ++
arch/powerpc/kernel/idle.c | 4 +++-
arch/powerpc/platforms/pseries/processor_idle.c | 5 +++++
3 files changed, 10 insertions(+), 1 deletion(-)
Index: linux.trees.git/arch/powerpc/include/asm/processor.h
===================================================================
--- linux.trees.git.orig/arch/powerpc/include/asm/processor.h
+++ linux.trees.git/arch/powerpc/include/asm/processor.h
@@ -332,6 +332,8 @@ static inline unsigned long get_clean_sp
}
#endif
+extern int boot_option_idle_override;
+
#endif /* __KERNEL__ */
#endif /* __ASSEMBLY__ */
#endif /* _ASM_POWERPC_PROCESSOR_H */
Index: linux.trees.git/arch/powerpc/kernel/idle.c
===================================================================
--- linux.trees.git.orig/arch/powerpc/kernel/idle.c
+++ linux.trees.git/arch/powerpc/kernel/idle.c
@@ -40,9 +40,11 @@
#define cpu_should_die() 0
#endif
+int boot_option_idle_override = 0;
+
static int __init powersave_off(char *arg)
{
- ppc_md.power_save = NULL;
+ boot_option_idle_override = 1;
return 0;
}
__setup("powersave=off", powersave_off);
Index: linux.trees.git/arch/powerpc/platforms/pseries/processor_idle.c
===================================================================
--- linux.trees.git.orig/arch/powerpc/platforms/pseries/processor_idle.c
+++ linux.trees.git/arch/powerpc/platforms/pseries/processor_idle.c
@@ -171,6 +171,11 @@ static int __init pseries_processor_idle
int cpu;
int result;
+ if (boot_option_idle_override) {
+ printk(KERN_DEBUG "Using default idle\n");
+ return 0;
+ }
+
result = cpuidle_register_driver(&pseries_idle_driver);
if (result < 0)
^ permalink raw reply
* [v10 PATCH 8/9]: pSeries: implement pSeries processor idle module
From: Arun R Bharadwaj @ 2009-12-02 10:02 UTC (permalink / raw)
To: Peter Zijlstra, Benjamin Herrenschmidt, Ingo Molnar,
Vaidyanathan Srinivasan, Dipankar Sarma, Balbir Singh,
Venkatesh Pallipadi, Arun Bharadwaj
Cc: linux-arch, linux-acpi, linuxppc-dev, linux-kernel
In-Reply-To: <20091202095427.GA27251@linux.vnet.ibm.com>
* Arun R Bharadwaj <arun@linux.vnet.ibm.com> [2009-12-02 15:24:27]:
This patch creates arch/powerpc/platforms/pseries/processor_idle.c,
which implements the cpuidle infrastructure for pseries.
It implements a pseries_cpuidle_loop() which would be the main idle loop
called from cpu_idle(). It makes decision of entering either
dedicated_snooze_loop or dedicated_cede_loop for dedicated lpar and
shared_cede_loop for shared lpar processor based on the
decision taken by the cpuidle governor.
Signed-off-by: Arun R Bharadwaj <arun@linux.vnet.ibm.com>
---
arch/powerpc/include/asm/system.h | 6
arch/powerpc/kernel/sysfs.c | 2
arch/powerpc/platforms/pseries/Makefile | 1
arch/powerpc/platforms/pseries/processor_idle.c | 196 ++++++++++++++++++++++++
arch/powerpc/platforms/pseries/pseries.h | 6
5 files changed, 211 insertions(+)
Index: linux.trees.git/arch/powerpc/platforms/pseries/Makefile
===================================================================
--- linux.trees.git.orig/arch/powerpc/platforms/pseries/Makefile
+++ linux.trees.git/arch/powerpc/platforms/pseries/Makefile
@@ -26,3 +26,4 @@ obj-$(CONFIG_HCALL_STATS) += hvCall_inst
obj-$(CONFIG_PHYP_DUMP) += phyp_dump.o
obj-$(CONFIG_CMM) += cmm.o
obj-$(CONFIG_DTL) += dtl.o
+obj-$(CONFIG_CPU_IDLE) += processor_idle.o
Index: linux.trees.git/arch/powerpc/platforms/pseries/pseries.h
===================================================================
--- linux.trees.git.orig/arch/powerpc/platforms/pseries/pseries.h
+++ linux.trees.git/arch/powerpc/platforms/pseries/pseries.h
@@ -10,6 +10,8 @@
#ifndef _PSERIES_PSERIES_H
#define _PSERIES_PSERIES_H
+#include <linux/cpuidle.h>
+
extern void __init fw_feature_init(const char *hypertas, unsigned long len);
struct pt_regs;
@@ -40,4 +42,8 @@ extern unsigned long rtas_poweron_auto;
extern void find_udbg_vterm(void);
+DECLARE_PER_CPU(unsigned long, smt_snooze_delay);
+
+extern struct cpuidle_driver pseries_idle_driver;
+
#endif /* _PSERIES_PSERIES_H */
Index: linux.trees.git/arch/powerpc/platforms/pseries/processor_idle.c
===================================================================
--- /dev/null
+++ linux.trees.git/arch/powerpc/platforms/pseries/processor_idle.c
@@ -0,0 +1,196 @@
+/*
+ * processor_idle - idle state cpuidle driver.
+ * Adapted from drivers/acpi/processor_idle.c
+ *
+ * Arun R Bharadwaj <arun@linux.vnet.ibm.com>
+ *
+ * Copyright (C) 2009 IBM Corporation.
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or (at
+ * your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/moduleparam.h>
+#include <linux/cpuidle.h>
+
+#include <asm/paca.h>
+#include <asm/reg.h>
+#include <asm/system.h>
+#include <asm/machdep.h>
+#include <asm/firmware.h>
+
+#include "plpar_wrappers.h"
+#include "pseries.h"
+
+MODULE_AUTHOR("Arun R Bharadwaj");
+MODULE_DESCRIPTION("pSeries Idle State Driver");
+MODULE_LICENSE("GPL");
+
+struct cpuidle_driver pseries_idle_driver = {
+ .name = "pseries_idle",
+ .owner = THIS_MODULE,
+};
+
+DEFINE_PER_CPU(struct cpuidle_device, pseries_dev);
+
+#define IDLE_STATE_COUNT 2
+
+/* pSeries Idle state Flags */
+#define PSERIES_DEDICATED_SNOOZE (0x01)
+#define PSERIES_DEDICATED_CEDE (0x02)
+#define PSERIES_SHARED_CEDE (0x03)
+
+static int pseries_idle_init(struct cpuidle_device *dev)
+{
+ return cpuidle_register_device(dev);
+}
+
+static void shared_cede_loop(void)
+{
+ get_lppaca()->idle = 1;
+ cede_processor();
+ get_lppaca()->idle = 0;
+}
+
+static void dedicated_snooze_loop(void)
+{
+ local_irq_enable();
+ set_thread_flag(TIF_POLLING_NRFLAG);
+ while (!need_resched()) {
+ ppc64_runlatch_off();
+ HMT_low();
+ HMT_very_low();
+ }
+ HMT_medium();
+ clear_thread_flag(TIF_POLLING_NRFLAG);
+ smp_mb();
+ local_irq_disable();
+}
+
+static void dedicated_cede_loop(void)
+{
+ ppc64_runlatch_off();
+ HMT_medium();
+ cede_processor();
+}
+
+static void pseries_cpuidle_loop(struct cpuidle_device *dev,
+ struct cpuidle_state *st)
+{
+ unsigned long in_purr, out_purr;
+
+ get_lppaca()->idle = 1;
+ get_lppaca()->donate_dedicated_cpu = 1;
+ in_purr = mfspr(SPRN_PURR);
+
+ if (st->flags & PSERIES_SHARED_CEDE)
+ shared_cede_loop();
+ else if (st->flags & PSERIES_DEDICATED_SNOOZE)
+ dedicated_snooze_loop();
+ else
+ dedicated_cede_loop();
+
+ out_purr = mfspr(SPRN_PURR);
+ get_lppaca()->wait_state_cycles += out_purr - in_purr;
+ get_lppaca()->donate_dedicated_cpu = 0;
+ get_lppaca()->idle = 0;
+}
+
+static int pseries_setup_cpuidle(struct cpuidle_device *dev, int cpu)
+{
+ int i;
+ struct cpuidle_state *state;
+
+ dev->cpu = cpu;
+
+ if (get_lppaca()->shared_proc) {
+ state = &dev->states[0];
+ snprintf(state->name, CPUIDLE_NAME_LEN, "IDLE");
+ state->enter = pseries_cpuidle_loop;
+ strncpy(state->desc, "shared_cede", CPUIDLE_DESC_LEN);
+ state->flags = PSERIES_SHARED_CEDE;
+ state->exit_latency = 0;
+ state->target_residency = 0;
+ return 0;
+ }
+
+ for (i = 0; i < IDLE_STATE_COUNT; i++) {
+ state = &dev->states[i];
+
+ snprintf(state->name, CPUIDLE_NAME_LEN, "CEDE%d", i);
+ state->enter = pseries_cpuidle_loop;
+
+ switch (i) {
+ case 0:
+ strncpy(state->desc, "snooze", CPUIDLE_DESC_LEN);
+ state->flags = PSERIES_DEDICATED_SNOOZE;
+ state->exit_latency = 0;
+ state->target_residency = 0;
+ break;
+
+ case 1:
+ strncpy(state->desc, "cede", CPUIDLE_DESC_LEN);
+ state->flags = PSERIES_DEDICATED_CEDE;
+ state->exit_latency = 1;
+ state->target_residency =
+ __get_cpu_var(smt_snooze_delay);
+ break;
+ }
+ }
+ dev->state_count = IDLE_STATE_COUNT;
+
+ return 0;
+}
+
+void update_smt_snooze_delay(int snooze)
+{
+ int cpu;
+ for_each_online_cpu(cpu)
+ per_cpu(pseries_dev, cpu).states[0].target_residency = snooze;
+}
+
+static int __init pseries_processor_idle_init(void)
+{
+ int cpu;
+ int result;
+
+ result = cpuidle_register_driver(&pseries_idle_driver);
+
+ if (result < 0)
+ return result;
+
+ printk(KERN_DEBUG "pSeries idle driver registered\n");
+
+ if (!firmware_has_feature(FW_FEATURE_SPLPAR)) {
+ printk(KERN_DEBUG "Using default idle\n");
+ return 0;
+ }
+
+ for_each_online_cpu(cpu) {
+ pseries_setup_cpuidle(&per_cpu(pseries_dev, cpu), cpu);
+ pseries_idle_init(&per_cpu(pseries_dev, cpu));
+ }
+
+ printk(KERN_DEBUG "Using cpuidle idle loop\n");
+
+ return 0;
+}
+
+device_initcall(pseries_processor_idle_init);
Index: linux.trees.git/arch/powerpc/include/asm/system.h
===================================================================
--- linux.trees.git.orig/arch/powerpc/include/asm/system.h
+++ linux.trees.git/arch/powerpc/include/asm/system.h
@@ -549,5 +549,11 @@ extern struct dentry *powerpc_debugfs_ro
void cpu_idle_wait(void);
+#ifdef CONFIG_CPU_IDLE
+extern void update_smt_snooze_delay(int snooze);
+#else
+static inline void update_smt_snooze_delay(int snooze) {}
+#endif
+
#endif /* __KERNEL__ */
#endif /* _ASM_POWERPC_SYSTEM_H */
Index: linux.trees.git/arch/powerpc/kernel/sysfs.c
===================================================================
--- linux.trees.git.orig/arch/powerpc/kernel/sysfs.c
+++ linux.trees.git/arch/powerpc/kernel/sysfs.c
@@ -18,6 +18,7 @@
#include <asm/machdep.h>
#include <asm/smp.h>
#include <asm/pmc.h>
+#include <asm/system.h>
#include "cacheinfo.h"
@@ -51,6 +52,7 @@ static ssize_t store_smt_snooze_delay(st
return -EINVAL;
per_cpu(smt_snooze_delay, cpu->sysdev.id) = snooze;
+ update_smt_snooze_delay(snooze);
return count;
}
^ permalink raw reply
* [v10 PATCH 7/9]: POWER: add a default_idle idle loop for POWER
From: Arun R Bharadwaj @ 2009-12-02 10:01 UTC (permalink / raw)
To: Peter Zijlstra, Benjamin Herrenschmidt, Ingo Molnar,
Vaidyanathan Srinivasan, Dipankar Sarma, Balbir Singh,
Venkatesh Pallipadi, Arun Bharadwaj
Cc: linux-arch, linux-acpi, linuxppc-dev, linux-kernel
In-Reply-To: <20091202095427.GA27251@linux.vnet.ibm.com>
* Arun R Bharadwaj <arun@linux.vnet.ibm.com> [2009-12-02 15:24:27]:
In arch/powerpc/kernel/idle.c create a default_idle() routine by moving
the failover condition of the cpu_idle() idle loop. This is needed by
cpuidle infrastructure to call default_idle when other idle routines
are not yet registered. Functionality remains the same, but the code is
slightly moved around.
Signed-off-by: Arun R Bharadwaj <arun@linux.vnet.ibm.com>
---
arch/powerpc/Kconfig | 3 +++
arch/powerpc/include/asm/system.h | 1 +
arch/powerpc/kernel/idle.c | 6 ++++++
3 files changed, 10 insertions(+)
Index: linux.trees.git/arch/powerpc/Kconfig
===================================================================
--- linux.trees.git.orig/arch/powerpc/Kconfig
+++ linux.trees.git/arch/powerpc/Kconfig
@@ -94,6 +94,9 @@ config ARCH_HAS_ILOG2_U64
config ARCH_HAS_CPU_IDLE_WAIT
def_bool y
+config ARCH_HAS_DEFAULT_IDLE
+ def_bool y
+
config GENERIC_HWEIGHT
bool
default y
Index: linux.trees.git/arch/powerpc/include/asm/system.h
===================================================================
--- linux.trees.git.orig/arch/powerpc/include/asm/system.h
+++ linux.trees.git/arch/powerpc/include/asm/system.h
@@ -218,6 +218,7 @@ extern unsigned long klimit;
extern void *alloc_maybe_bootmem(size_t size, gfp_t mask);
extern void *zalloc_maybe_bootmem(size_t size, gfp_t mask);
+extern void default_idle(void);
extern int powersave_nap; /* set if nap mode can be used in idle loop */
/*
Index: linux.trees.git/arch/powerpc/kernel/idle.c
===================================================================
--- linux.trees.git.orig/arch/powerpc/kernel/idle.c
+++ linux.trees.git/arch/powerpc/kernel/idle.c
@@ -121,6 +121,12 @@ void cpu_idle_wait(void)
}
EXPORT_SYMBOL_GPL(cpu_idle_wait);
+void default_idle(void)
+{
+ HMT_low();
+ HMT_very_low();
+}
+
int powersave_nap;
#ifdef CONFIG_SYSCTL
^ permalink raw reply
* [v10 PATCH 6/9]: pSeries/cpuidle: refactor pseries idle loops
From: Arun R Bharadwaj @ 2009-12-02 10:01 UTC (permalink / raw)
To: Peter Zijlstra, Benjamin Herrenschmidt, Ingo Molnar,
Vaidyanathan Srinivasan, Dipankar Sarma, Balbir Singh,
Venkatesh Pallipadi, Arun Bharadwaj
Cc: linux-arch, linux-acpi, linuxppc-dev, linux-kernel
In-Reply-To: <20091202095427.GA27251@linux.vnet.ibm.com>
* Arun R Bharadwaj <arun@linux.vnet.ibm.com> [2009-12-02 15:24:27]:
This patch removes the routines, pseries_shared_idle_sleep and
pseries_dedicated_idle_sleep, since this is implemented as a part
of arch/powerpc/platform/pseries/processor_idle.c
Also, similar to x86, call cpuidle_idle_call from cpu_idle() idle
loop instead of ppc_md.power_save.
Signed-off-by: Arun R Bharadwaj <arun@linux.vnet.ibm.com>
---
arch/powerpc/kernel/idle.c | 58 ++++++++++-----------
arch/powerpc/platforms/pseries/setup.c | 89 ---------------------------------
2 files changed, 30 insertions(+), 117 deletions(-)
Index: linux.trees.git/arch/powerpc/platforms/pseries/setup.c
===================================================================
--- linux.trees.git.orig/arch/powerpc/platforms/pseries/setup.c
+++ linux.trees.git/arch/powerpc/platforms/pseries/setup.c
@@ -75,9 +75,6 @@ EXPORT_SYMBOL(CMO_PageSize);
int fwnmi_active; /* TRUE if an FWNMI handler is present */
-static void pseries_shared_idle_sleep(void);
-static void pseries_dedicated_idle_sleep(void);
-
static struct device_node *pSeries_mpic_node;
static void pSeries_show_cpuinfo(struct seq_file *m)
@@ -297,18 +294,8 @@ static void __init pSeries_setup_arch(vo
pSeries_nvram_init();
/* Choose an idle loop */
- if (firmware_has_feature(FW_FEATURE_SPLPAR)) {
+ if (firmware_has_feature(FW_FEATURE_SPLPAR))
vpa_init(boot_cpuid);
- if (get_lppaca()->shared_proc) {
- printk(KERN_DEBUG "Using shared processor idle loop\n");
- ppc_md.power_save = pseries_shared_idle_sleep;
- } else {
- printk(KERN_DEBUG "Using dedicated idle loop\n");
- ppc_md.power_save = pseries_dedicated_idle_sleep;
- }
- } else {
- printk(KERN_DEBUG "Using default idle loop\n");
- }
if (firmware_has_feature(FW_FEATURE_LPAR))
ppc_md.enable_pmcs = pseries_lpar_enable_pmcs;
@@ -496,80 +483,6 @@ static int __init pSeries_probe(void)
return 1;
}
-
-DECLARE_PER_CPU(unsigned long, smt_snooze_delay);
-
-static void pseries_dedicated_idle_sleep(void)
-{
- unsigned int cpu = smp_processor_id();
- unsigned long start_snooze;
- unsigned long in_purr, out_purr;
-
- /*
- * Indicate to the HV that we are idle. Now would be
- * a good time to find other work to dispatch.
- */
- get_lppaca()->idle = 1;
- get_lppaca()->donate_dedicated_cpu = 1;
- in_purr = mfspr(SPRN_PURR);
-
- /*
- * We come in with interrupts disabled, and need_resched()
- * has been checked recently. If we should poll for a little
- * while, do so.
- */
- if (__get_cpu_var(smt_snooze_delay)) {
- start_snooze = get_tb() +
- __get_cpu_var(smt_snooze_delay) * tb_ticks_per_usec;
- local_irq_enable();
- set_thread_flag(TIF_POLLING_NRFLAG);
-
- while (get_tb() < start_snooze) {
- if (need_resched() || cpu_is_offline(cpu))
- goto out;
- ppc64_runlatch_off();
- HMT_low();
- HMT_very_low();
- }
-
- HMT_medium();
- clear_thread_flag(TIF_POLLING_NRFLAG);
- smp_mb();
- local_irq_disable();
- if (need_resched() || cpu_is_offline(cpu))
- goto out;
- }
-
- cede_processor();
-
-out:
- HMT_medium();
- out_purr = mfspr(SPRN_PURR);
- get_lppaca()->wait_state_cycles += out_purr - in_purr;
- get_lppaca()->donate_dedicated_cpu = 0;
- get_lppaca()->idle = 0;
-}
-
-static void pseries_shared_idle_sleep(void)
-{
- /*
- * Indicate to the HV that we are idle. Now would be
- * a good time to find other work to dispatch.
- */
- get_lppaca()->idle = 1;
-
- /*
- * Yield the processor to the hypervisor. We return if
- * an external interrupt occurs (which are driven prior
- * to returning here) or if a prod occurs from another
- * processor. When returning here, external interrupts
- * are enabled.
- */
- cede_processor();
-
- get_lppaca()->idle = 0;
-}
-
static int pSeries_pci_probe_mode(struct pci_bus *bus)
{
if (firmware_has_feature(FW_FEATURE_LPAR))
Index: linux.trees.git/arch/powerpc/kernel/idle.c
===================================================================
--- linux.trees.git.orig/arch/powerpc/kernel/idle.c
+++ linux.trees.git/arch/powerpc/kernel/idle.c
@@ -25,6 +25,7 @@
#include <linux/cpu.h>
#include <linux/sysctl.h>
#include <linux/tick.h>
+#include <linux/cpuidle.h>
#include <asm/system.h>
#include <asm/processor.h>
@@ -46,6 +47,14 @@ static int __init powersave_off(char *ar
}
__setup("powersave=off", powersave_off);
+#ifndef CONFIG_CPU_IDLE
+void cpuidle_idle_call(void)
+{
+ local_irq_enable();
+ cpu_relax();
+}
+#endif
+
/*
* The body of the idle task.
*/
@@ -60,35 +69,26 @@ void cpu_idle(void)
while (!need_resched() && !cpu_should_die()) {
ppc64_runlatch_off();
- if (ppc_md.power_save) {
- clear_thread_flag(TIF_POLLING_NRFLAG);
- /*
- * smp_mb is so clearing of TIF_POLLING_NRFLAG
- * is ordered w.r.t. need_resched() test.
- */
- smp_mb();
- local_irq_disable();
-
- /* Don't trace irqs off for idle */
- stop_critical_timings();
-
- /* check again after disabling irqs */
- if (!need_resched() && !cpu_should_die())
- ppc_md.power_save();
-
- start_critical_timings();
-
- local_irq_enable();
- set_thread_flag(TIF_POLLING_NRFLAG);
-
- } else {
- /*
- * Go into low thread priority and possibly
- * low power mode.
- */
- HMT_low();
- HMT_very_low();
- }
+ clear_thread_flag(TIF_POLLING_NRFLAG);
+ /*
+ * smp_mb is so clearing of TIF_POLLING_NRFLAG
+ * is ordered w.r.t. need_resched() test.
+ */
+ smp_mb();
+ local_irq_disable();
+
+ /* Don't trace irqs off for idle */
+ stop_critical_timings();
+
+ /* check again after disabling irqs */
+ if (!need_resched() && !cpu_should_die())
+ cpuidle_idle_call();
+
+ start_critical_timings();
+
+ local_irq_enable();
+ set_thread_flag(TIF_POLLING_NRFLAG);
+
}
HMT_medium();
^ permalink raw reply
* [v10 PATCH 5/9]: POWER: enable cpuidle for POWER.
From: Arun R Bharadwaj @ 2009-12-02 10:00 UTC (permalink / raw)
To: Peter Zijlstra, Benjamin Herrenschmidt, Ingo Molnar,
Vaidyanathan Srinivasan, Dipankar Sarma, Balbir Singh,
Venkatesh Pallipadi, Arun Bharadwaj
Cc: linux-arch, linux-acpi, linuxppc-dev, linux-kernel
In-Reply-To: <20091202095427.GA27251@linux.vnet.ibm.com>
* Arun R Bharadwaj <arun@linux.vnet.ibm.com> [2009-12-02 15:24:27]:
This patch enables the cpuidle option in Kconfig for pSeries.
Currently cpuidle infrastructure is enabled only for x86 and ARM.
This code is almost completely borrowed from x86 to enable
cpuidle for pSeries.
Signed-off-by: Arun R Bharadwaj <arun@linux.vnet.ibm.com>
---
arch/powerpc/Kconfig | 9 +++++++++
arch/powerpc/include/asm/system.h | 2 ++
arch/powerpc/kernel/idle.c | 19 +++++++++++++++++++
3 files changed, 30 insertions(+)
Index: linux.trees.git/arch/powerpc/Kconfig
===================================================================
--- linux.trees.git.orig/arch/powerpc/Kconfig
+++ linux.trees.git/arch/powerpc/Kconfig
@@ -91,6 +91,9 @@ config ARCH_HAS_ILOG2_U64
bool
default y if 64BIT
+config ARCH_HAS_CPU_IDLE_WAIT
+ def_bool y
+
config GENERIC_HWEIGHT
bool
default y
@@ -247,6 +250,12 @@ source "kernel/Kconfig.freezer"
source "arch/powerpc/sysdev/Kconfig"
source "arch/powerpc/platforms/Kconfig"
+menu "Power management options"
+
+source "drivers/cpuidle/Kconfig"
+
+endmenu
+
menu "Kernel options"
config HIGHMEM
Index: linux.trees.git/arch/powerpc/include/asm/system.h
===================================================================
--- linux.trees.git.orig/arch/powerpc/include/asm/system.h
+++ linux.trees.git/arch/powerpc/include/asm/system.h
@@ -546,5 +546,7 @@ extern void account_system_vtime(struct
extern struct dentry *powerpc_debugfs_root;
+void cpu_idle_wait(void);
+
#endif /* __KERNEL__ */
#endif /* _ASM_POWERPC_SYSTEM_H */
Index: linux.trees.git/arch/powerpc/kernel/idle.c
===================================================================
--- linux.trees.git.orig/arch/powerpc/kernel/idle.c
+++ linux.trees.git/arch/powerpc/kernel/idle.c
@@ -102,6 +102,25 @@ void cpu_idle(void)
}
}
+static void do_nothing(void *unused)
+{
+}
+
+/*
+ * cpu_idle_wait - Used to ensure that all the CPUs come out of the old
+ * idle loop and start using the new idle loop.
+ * Required while changing idle handler on SMP systems.
+ * Caller must have changed idle handler to the new value before the call.
+ */
+void cpu_idle_wait(void)
+{
+ /* Ensure that new value of idle is set */
+ smp_mb();
+ /* kick all the CPUs so that they exit out of old idle routine */
+ smp_call_function(do_nothing, NULL, 1);
+}
+EXPORT_SYMBOL_GPL(cpu_idle_wait);
+
int powersave_nap;
#ifdef CONFIG_SYSCTL
^ permalink raw reply
* [v10 PATCH 4/9]: x86: refactor x86 idle power management code, remove all instances of pm_idle
From: Arun R Bharadwaj @ 2009-12-02 9:59 UTC (permalink / raw)
To: Peter Zijlstra, Benjamin Herrenschmidt, Ingo Molnar,
Vaidyanathan Srinivasan, Dipankar Sarma, Balbir Singh,
Venkatesh Pallipadi, Arun Bharadwaj
Cc: linux-arch, linux-acpi, linuxppc-dev, linux-kernel
In-Reply-To: <20091202095427.GA27251@linux.vnet.ibm.com>
* Arun R Bharadwaj <arun@linux.vnet.ibm.com> [2009-12-02 15:24:27]:
This patch cleans up x86 of all instances of pm_idle.
pm_idle which was earlier called from cpu_idle() idle loop
is replaced by cpuidle_idle_call.
x86 also registers to cpuidle when the idle routine is selected,
by populating the cpuidle_device data structure for each cpu.
This is replicated for apm module and for xen, which also used pm_idle.
Signed-off-by: Arun R Bharadwaj <arun@linux.vnet.ibm.com>
---
arch/x86/kernel/apm_32.c | 46 +++++++++++++++++++++++-
arch/x86/kernel/process.c | 78 +++++++++++++++++++++++++++++++-----------
arch/x86/kernel/process_32.c | 3 +
arch/x86/kernel/process_64.c | 3 +
arch/x86/xen/setup.c | 30 +++++++++++++++-
drivers/acpi/processor_core.c | 9 +++-
drivers/acpi/processor_idle.c | 44 ++++++++++-------------
7 files changed, 160 insertions(+), 53 deletions(-)
Index: linux.trees.git/arch/x86/kernel/process.c
===================================================================
--- linux.trees.git.orig/arch/x86/kernel/process.c
+++ linux.trees.git/arch/x86/kernel/process.c
@@ -10,6 +10,7 @@
#include <linux/clockchips.h>
#include <linux/random.h>
#include <linux/user-return-notifier.h>
+#include <linux/cpuidle.h>
#include <trace/events/power.h>
#include <linux/hw_breakpoint.h>
#include <asm/system.h>
@@ -241,12 +242,6 @@ int sys_vfork(struct pt_regs *regs)
unsigned long boot_option_idle_override = 0;
EXPORT_SYMBOL(boot_option_idle_override);
-/*
- * Powermanagement idle function, if any..
- */
-void (*pm_idle)(void);
-EXPORT_SYMBOL(pm_idle);
-
#ifdef CONFIG_X86_32
/*
* This halt magic was a workaround for ancient floppy DMA
@@ -326,17 +321,15 @@ static void do_nothing(void *unused)
}
/*
- * cpu_idle_wait - Used to ensure that all the CPUs discard old value of
- * pm_idle and update to new pm_idle value. Required while changing pm_idle
- * handler on SMP systems.
+ * cpu_idle_wait - Required while changing idle routine handler on SMP systems.
*
- * Caller must have changed pm_idle to the new value before the call. Old
- * pm_idle value will not be used by any CPU after the return of this function.
+ * Caller must have changed idle routine to the new value before the call. Old
+ * value will not be used by any CPU after the return of this function.
*/
void cpu_idle_wait(void)
{
smp_mb();
- /* kick all the CPUs so that they exit out of pm_idle */
+ /* kick all the CPUs so that they exit out of idle loop */
smp_call_function(do_nothing, NULL, 1);
}
EXPORT_SYMBOL_GPL(cpu_idle_wait);
@@ -515,15 +508,58 @@ static void c1e_idle(void)
default_idle();
}
+static void (*local_idle)(void);
+
+#ifndef CONFIG_CPU_IDLE
+void cpuidle_idle_call(void)
+{
+ if (local_idle)
+ local_idle();
+ else
+ default_idle();
+}
+#endif
+
+DEFINE_PER_CPU(struct cpuidle_device, idle_devices);
+
+struct cpuidle_driver cpuidle_default_driver = {
+ .name = "cpuidle_default",
+};
+
+static void local_idle_loop(struct cpuidle_device *dev,
+ struct cpuidle_state *st)
+{
+ local_idle();
+}
+
+static int setup_cpuidle_simple(void)
+{
+ struct cpuidle_device *dev;
+ int cpu;
+
+ if (!cpuidle_curr_driver)
+ cpuidle_register_driver(&cpuidle_default_driver);
+
+ for_each_online_cpu(cpu) {
+ dev = &per_cpu(idle_devices, cpu);
+ dev->cpu = cpu;
+ dev->states[0].enter = local_idle_loop;
+ dev->state_count = 1;
+ cpuidle_register_device(dev);
+ }
+ return 0;
+}
+device_initcall(setup_cpuidle_simple);
+
void __cpuinit select_idle_routine(const struct cpuinfo_x86 *c)
{
#ifdef CONFIG_SMP
- if (pm_idle == poll_idle && smp_num_siblings > 1) {
+ if (local_idle == poll_idle && smp_num_siblings > 1) {
printk(KERN_WARNING "WARNING: polling idle and HT enabled,"
" performance may degrade.\n");
}
#endif
- if (pm_idle)
+ if (local_idle)
return;
if (cpu_has(c, X86_FEATURE_MWAIT) && mwait_usable(c)) {
@@ -531,18 +567,20 @@ void __cpuinit select_idle_routine(const
* One CPU supports mwait => All CPUs supports mwait
*/
printk(KERN_INFO "using mwait in idle threads.\n");
- pm_idle = mwait_idle;
+ local_idle = mwait_idle;
} else if (check_c1e_idle(c)) {
printk(KERN_INFO "using C1E aware idle routine\n");
- pm_idle = c1e_idle;
+ local_idle = c1e_idle;
} else
- pm_idle = default_idle;
+ local_idle = default_idle;
+
+ return;
}
void __init init_c1e_mask(void)
{
/* If we're using c1e_idle, we need to allocate c1e_mask. */
- if (pm_idle == c1e_idle)
+ if (local_idle == c1e_idle)
zalloc_cpumask_var(&c1e_mask, GFP_KERNEL);
}
@@ -553,7 +591,7 @@ static int __init idle_setup(char *str)
if (!strcmp(str, "poll")) {
printk("using polling idle threads.\n");
- pm_idle = poll_idle;
+ local_idle = poll_idle;
} else if (!strcmp(str, "mwait"))
force_mwait = 1;
else if (!strcmp(str, "halt")) {
@@ -564,7 +602,7 @@ static int __init idle_setup(char *str)
* To continue to load the CPU idle driver, don't touch
* the boot_option_idle_override.
*/
- pm_idle = default_idle;
+ local_idle = default_idle;
idle_halt = 1;
return 0;
} else if (!strcmp(str, "nomwait")) {
Index: linux.trees.git/arch/x86/kernel/process_32.c
===================================================================
--- linux.trees.git.orig/arch/x86/kernel/process_32.c
+++ linux.trees.git/arch/x86/kernel/process_32.c
@@ -40,6 +40,7 @@
#include <linux/uaccess.h>
#include <linux/io.h>
#include <linux/kdebug.h>
+#include <linux/cpuidle.h>
#include <asm/pgtable.h>
#include <asm/system.h>
@@ -112,7 +113,7 @@ void cpu_idle(void)
local_irq_disable();
/* Don't trace irqs off for idle */
stop_critical_timings();
- pm_idle();
+ cpuidle_idle_call();
start_critical_timings();
}
tick_nohz_restart_sched_tick();
Index: linux.trees.git/arch/x86/kernel/process_64.c
===================================================================
--- linux.trees.git.orig/arch/x86/kernel/process_64.c
+++ linux.trees.git/arch/x86/kernel/process_64.c
@@ -39,6 +39,7 @@
#include <linux/io.h>
#include <linux/ftrace.h>
#include <linux/dmi.h>
+#include <linux/cpuidle.h>
#include <asm/pgtable.h>
#include <asm/system.h>
@@ -141,7 +142,7 @@ void cpu_idle(void)
enter_idle();
/* Don't trace irqs off for idle */
stop_critical_timings();
- pm_idle();
+ cpuidle_idle_call();
start_critical_timings();
/* In many cases the interrupt that ended idle
has already called exit_idle. But some idle
Index: linux.trees.git/arch/x86/kernel/apm_32.c
===================================================================
--- linux.trees.git.orig/arch/x86/kernel/apm_32.c
+++ linux.trees.git/arch/x86/kernel/apm_32.c
@@ -227,6 +227,7 @@
#include <linux/suspend.h>
#include <linux/kthread.h>
#include <linux/jiffies.h>
+#include <linux/cpuidle.h>
#include <asm/system.h>
#include <asm/uaccess.h>
@@ -2255,6 +2256,46 @@ static struct dmi_system_id __initdata a
{ }
};
+DEFINE_PER_CPU(struct cpuidle_device, apm_idle_devices);
+
+struct cpuidle_driver cpuidle_apm_driver = {
+ .name = "cpuidle_apm",
+};
+
+static void apm_idle_loop(struct cpuidle_device *dev, struct cpuidle_state *st)
+{
+ apm_cpu_idle();
+}
+
+static void setup_cpuidle_apm(void)
+{
+ struct cpuidle_device *dev;
+ int cpu;
+
+ if (!cpuidle_curr_driver)
+ cpuidle_register_driver(&cpuidle_apm_driver);
+
+ for_each_online_cpu(cpu) {
+ dev = &per_cpu(apm_idle_devices, cpu);
+ dev->cpu = cpu;
+ dev->states[0].enter = apm_idle_loop;
+ dev->state_count = 1;
+ cpuidle_register_device(dev);
+ }
+}
+
+void exit_cpuidle_apm(void)
+{
+ struct cpuidle_device *dev;
+ int cpu;
+
+ for_each_online_cpu(cpu) {
+ dev = &per_cpu(apm_idle_devices, cpu);
+ cpuidle_unregister_device(dev);
+ }
+}
+
+
/*
* Just start the APM thread. We do NOT want to do APM BIOS
* calls from anything but the APM thread, if for no other reason
@@ -2392,8 +2433,7 @@ static int __init apm_init(void)
if (HZ != 100)
idle_period = (idle_period * HZ) / 100;
if (idle_threshold < 100) {
- original_pm_idle = pm_idle;
- pm_idle = apm_cpu_idle;
+ setup_cpuidle_apm();
set_pm_idle = 1;
}
@@ -2405,7 +2445,7 @@ static void __exit apm_exit(void)
int error;
if (set_pm_idle) {
- pm_idle = original_pm_idle;
+ exit_cpuidle_apm();
/*
* We are about to unload the current idle thread pm callback
* (pm_idle), Wait for all processors to update cached/local
Index: linux.trees.git/arch/x86/xen/setup.c
===================================================================
--- linux.trees.git.orig/arch/x86/xen/setup.c
+++ linux.trees.git/arch/x86/xen/setup.c
@@ -8,6 +8,7 @@
#include <linux/sched.h>
#include <linux/mm.h>
#include <linux/pm.h>
+#include <linux/cpuidle.h>
#include <asm/elf.h>
#include <asm/vdso.h>
@@ -151,6 +152,33 @@ void __cpuinit xen_enable_syscall(void)
#endif /* CONFIG_X86_64 */
}
+DEFINE_PER_CPU(struct cpuidle_device, xen_idle_devices);
+struct cpuidle_driver cpuidle_xen_driver = {
+ .name = "cpuidle_xen",
+};
+
+static void xen_idle_loop(struct cpuidle_device *dev, struct cpuidle_state *st)
+{
+ xen_idle();
+}
+
+static void setup_cpuidle_xen(void)
+{
+ struct cpuidle_device *dev;
+ int cpu;
+
+ if (!cpuidle_curr_driver)
+ cpuidle_register_driver(&cpuidle_xen_driver);
+
+ for_each_online_cpu(cpu) {
+ dev = &per_cpu(xen_idle_devices, cpu);
+ dev->cpu = cpu;
+ dev->states[0].enter = xen_idle_loop;
+ dev->state_count = 1;
+ cpuidle_register_device(dev);
+ }
+}
+
void __init xen_arch_setup(void)
{
struct physdev_set_iopl set_iopl;
@@ -186,7 +214,7 @@ void __init xen_arch_setup(void)
MAX_GUEST_CMDLINE > COMMAND_LINE_SIZE ?
COMMAND_LINE_SIZE : MAX_GUEST_CMDLINE);
- pm_idle = xen_idle;
+ setup_cpuidle_xen();
paravirt_disable_iospace();
Index: linux.trees.git/drivers/acpi/processor_core.c
===================================================================
--- linux.trees.git.orig/drivers/acpi/processor_core.c
+++ linux.trees.git/drivers/acpi/processor_core.c
@@ -1150,9 +1150,12 @@ static int __init acpi_processor_init(vo
* should not use mwait for CPU-states.
*/
dmi_check_system(processor_idle_dmi_table);
- result = cpuidle_register_driver(&acpi_idle_driver);
- if (result < 0)
- goto out_proc;
+
+ if (!boot_option_idle_override) {
+ result = cpuidle_register_driver(&acpi_idle_driver);
+ if (result < 0)
+ goto out_proc;
+ }
result = acpi_bus_register_driver(&acpi_processor_driver);
if (result < 0)
Index: linux.trees.git/drivers/acpi/processor_idle.c
===================================================================
--- linux.trees.git.orig/drivers/acpi/processor_idle.c
+++ linux.trees.git/drivers/acpi/processor_idle.c
@@ -821,18 +821,16 @@ static inline void acpi_idle_do_entry(st
*
* This is equivalent to the HALT instruction.
*/
-static int acpi_idle_enter_c1(struct cpuidle_device *dev,
+static void acpi_idle_enter_c1(struct cpuidle_device *dev,
struct cpuidle_state *state)
{
- ktime_t kt1, kt2;
- s64 idle_time;
struct acpi_processor *pr;
struct acpi_processor_cx *cx = cpuidle_get_statedata(state);
pr = __get_cpu_var(processors);
if (unlikely(!pr))
- return 0;
+ return;
local_irq_disable();
@@ -840,20 +838,15 @@ static int acpi_idle_enter_c1(struct cpu
if (acpi_idle_suspend) {
local_irq_enable();
cpu_relax();
- return 0;
+ return;
}
lapic_timer_state_broadcast(pr, cx, 1);
- kt1 = ktime_get_real();
acpi_idle_do_entry(cx);
- kt2 = ktime_get_real();
- idle_time = ktime_to_us(ktime_sub(kt2, kt1));
local_irq_enable();
cx->usage++;
lapic_timer_state_broadcast(pr, cx, 0);
-
- return idle_time;
}
/**
@@ -861,7 +854,7 @@ static int acpi_idle_enter_c1(struct cpu
* @dev: the target CPU
* @state: the state data
*/
-static int acpi_idle_enter_simple(struct cpuidle_device *dev,
+static void acpi_idle_enter_simple(struct cpuidle_device *dev,
struct cpuidle_state *state)
{
struct acpi_processor *pr;
@@ -873,10 +866,12 @@ static int acpi_idle_enter_simple(struct
pr = __get_cpu_var(processors);
if (unlikely(!pr))
- return 0;
+ return;
- if (acpi_idle_suspend)
- return(acpi_idle_enter_c1(dev, state));
+ if (acpi_idle_suspend) {
+ acpi_idle_enter_c1(dev, state);
+ return;
+ }
local_irq_disable();
current_thread_info()->status &= ~TS_POLLING;
@@ -889,7 +884,7 @@ static int acpi_idle_enter_simple(struct
if (unlikely(need_resched())) {
current_thread_info()->status |= TS_POLLING;
local_irq_enable();
- return 0;
+ return;
}
/*
@@ -920,7 +915,6 @@ static int acpi_idle_enter_simple(struct
lapic_timer_state_broadcast(pr, cx, 0);
cx->time += sleep_ticks;
- return idle_time;
}
static int c3_cpu_count;
@@ -933,7 +927,7 @@ static DEFINE_SPINLOCK(c3_lock);
*
* If BM is detected, the deepest non-C3 idle state is entered instead.
*/
-static int acpi_idle_enter_bm(struct cpuidle_device *dev,
+static void acpi_idle_enter_bm(struct cpuidle_device *dev,
struct cpuidle_state *state)
{
struct acpi_processor *pr;
@@ -946,20 +940,23 @@ static int acpi_idle_enter_bm(struct cpu
pr = __get_cpu_var(processors);
if (unlikely(!pr))
- return 0;
+ return;
- if (acpi_idle_suspend)
- return(acpi_idle_enter_c1(dev, state));
+ if (acpi_idle_suspend) {
+ acpi_idle_enter_c1(dev, state);
+ return;
+ }
if (acpi_idle_bm_check()) {
if (dev->safe_state) {
dev->last_state = dev->safe_state;
- return dev->safe_state->enter(dev, dev->safe_state);
+ dev->safe_state->enter(dev, dev->safe_state);
+ return;
} else {
local_irq_disable();
acpi_safe_halt();
local_irq_enable();
- return 0;
+ return;
}
}
@@ -974,7 +971,7 @@ static int acpi_idle_enter_bm(struct cpu
if (unlikely(need_resched())) {
current_thread_info()->status |= TS_POLLING;
local_irq_enable();
- return 0;
+ return;
}
acpi_unlazy_tlb(smp_processor_id());
@@ -1032,7 +1029,6 @@ static int acpi_idle_enter_bm(struct cpu
lapic_timer_state_broadcast(pr, cx, 0);
cx->time += sleep_ticks;
- return idle_time;
}
struct cpuidle_driver acpi_idle_driver = {
^ permalink raw reply
* [v10 PATCH 3/9]: cpuidle: implement a list based approach to register a set of idle routines
From: Arun R Bharadwaj @ 2009-12-02 9:58 UTC (permalink / raw)
To: Peter Zijlstra, Benjamin Herrenschmidt, Ingo Molnar,
Vaidyanathan Srinivasan, Dipankar Sarma, Balbir Singh,
Venkatesh Pallipadi, Arun Bharadwaj
Cc: linux-arch, linux-acpi, linuxppc-dev, linux-kernel
In-Reply-To: <20091202095427.GA27251@linux.vnet.ibm.com>
* Arun R Bharadwaj <arun@linux.vnet.ibm.com> [2009-12-02 15:24:27]:
Implement a list based registering mechanism for architectures which
have multiple sets of idle routines which are to be registered.
Currently, in x86 it is done by merely setting pm_idle = idle_routine
and managing this pm_idle pointer is messy.
To give an example of how this mechanism works:
In x86, initially, idle routine is selected from the set of poll/mwait/
c1e/default idle loops. So the selected idle loop is registered in cpuidle
as one idle state cpuidle devices. Once ACPI comes up, it registers
another set of idle states on top of this state. Again, suppose a module
registers another set of idle loops, it is added to this list.
This provides a clean way of registering and unregistering idle state
routines.
In the current implementation, pm_idle is set as the current idle routine
being used and the old idle routine has to be maintained and when a module
registers/unregisters an idle routine, confusion arises.
Signed-off-by: Arun R Bharadwaj <arun@linux.vnet.ibm.com>
---
drivers/cpuidle/cpuidle.c | 54 ++++++++++++++++++++++++++++++++++++++++------
include/linux/cpuidle.h | 1
2 files changed, 48 insertions(+), 7 deletions(-)
Index: linux.trees.git/drivers/cpuidle/cpuidle.c
===================================================================
--- linux.trees.git.orig/drivers/cpuidle/cpuidle.c
+++ linux.trees.git/drivers/cpuidle/cpuidle.c
@@ -22,6 +22,7 @@
#include "cpuidle.h"
DEFINE_PER_CPU(struct cpuidle_device *, cpuidle_devices);
+DEFINE_PER_CPU(struct list_head, cpuidle_devices_list);
DEFINE_MUTEX(cpuidle_lock);
@@ -129,6 +130,45 @@ void cpuidle_resume_and_unlock(void)
EXPORT_SYMBOL_GPL(cpuidle_resume_and_unlock);
+int cpuidle_add_to_list(struct cpuidle_device *dev)
+{
+ int ret, cpu = dev->cpu;
+ struct cpuidle_device *old_dev;
+
+ if (!list_empty(&per_cpu(cpuidle_devices_list, cpu))) {
+ old_dev = list_first_entry(&per_cpu(cpuidle_devices_list, cpu),
+ struct cpuidle_device, idle_list);
+ cpuidle_remove_state_sysfs(old_dev);
+ }
+
+ list_add(&dev->idle_list, &per_cpu(cpuidle_devices_list, cpu));
+ ret = cpuidle_add_state_sysfs(dev);
+ return ret;
+}
+
+void cpuidle_remove_from_list(struct cpuidle_device *dev)
+{
+ struct cpuidle_device *temp_dev;
+ struct list_head *pos;
+ int ret, cpu = dev->cpu;
+
+ list_for_each(pos, &per_cpu(cpuidle_devices_list, cpu)) {
+ temp_dev = container_of(pos, struct cpuidle_device, idle_list);
+ if (dev == temp_dev) {
+ list_del(&temp_dev->idle_list);
+ cpuidle_remove_state_sysfs(temp_dev);
+ break;
+ }
+ }
+
+ if (!list_empty(&per_cpu(cpuidle_devices_list, cpu))) {
+ temp_dev = list_first_entry(&per_cpu(cpuidle_devices_list, cpu),
+ struct cpuidle_device, idle_list);
+ ret = cpuidle_add_state_sysfs(temp_dev);
+ }
+ cpuidle_kick_cpus();
+}
+
/**
* cpuidle_enable_device - enables idle PM for a CPU
* @dev: the CPU
@@ -153,9 +193,6 @@ int cpuidle_enable_device(struct cpuidle
return ret;
}
- if ((ret = cpuidle_add_state_sysfs(dev)))
- return ret;
-
if (cpuidle_curr_governor->enable &&
(ret = cpuidle_curr_governor->enable(dev)))
goto fail_sysfs;
@@ -174,7 +211,7 @@ int cpuidle_enable_device(struct cpuidle
return 0;
fail_sysfs:
- cpuidle_remove_state_sysfs(dev);
+ cpuidle_remove_from_list(dev);
return ret;
}
@@ -199,8 +236,6 @@ void cpuidle_disable_device(struct cpuid
if (cpuidle_curr_governor->disable)
cpuidle_curr_governor->disable(dev);
-
- cpuidle_remove_state_sysfs(dev);
}
EXPORT_SYMBOL_GPL(cpuidle_disable_device);
@@ -271,6 +306,7 @@ int cpuidle_register_device(struct cpuid
}
cpuidle_enable_device(dev);
+ cpuidle_add_to_list(dev);
mutex_unlock(&cpuidle_lock);
@@ -292,6 +328,7 @@ void cpuidle_unregister_device(struct cp
cpuidle_pause_and_lock();
cpuidle_disable_device(dev);
+ cpuidle_remove_from_list(dev);
per_cpu(cpuidle_devices, dev->cpu) = NULL;
@@ -342,12 +379,15 @@ static inline void latency_notifier_init
*/
static int __init cpuidle_init(void)
{
- int ret;
+ int ret, cpu;
ret = cpuidle_add_class_sysfs(&cpu_sysdev_class);
if (ret)
return ret;
+ for_each_possible_cpu(cpu)
+ INIT_LIST_HEAD(&per_cpu(cpuidle_devices_list, cpu));
+
latency_notifier_init(&cpuidle_latency_notifier);
return 0;
Index: linux.trees.git/include/linux/cpuidle.h
===================================================================
--- linux.trees.git.orig/include/linux/cpuidle.h
+++ linux.trees.git/include/linux/cpuidle.h
@@ -92,6 +92,7 @@ struct cpuidle_device {
struct cpuidle_state_kobj *kobjs[CPUIDLE_STATE_MAX];
struct cpuidle_state *last_state;
+ struct list_head idle_list;
struct kobject kobj;
struct completion kobj_unregister;
void *governor_data;
^ permalink raw reply
* [v10 PATCH 2/9]: cpuidle: cleanup drivers/cpuidle/cpuidle.c
From: Arun R Bharadwaj @ 2009-12-02 9:57 UTC (permalink / raw)
To: Peter Zijlstra, Benjamin Herrenschmidt, Ingo Molnar,
Vaidyanathan Srinivasan, Dipankar Sarma, Balbir Singh,
Venkatesh Pallipadi, Arun Bharadwaj
Cc: linux-arch, linux-acpi, linuxppc-dev, linux-kernel
In-Reply-To: <20091202095427.GA27251@linux.vnet.ibm.com>
* Arun R Bharadwaj <arun@linux.vnet.ibm.com> [2009-12-02 15:24:27]:
This patch cleans up drivers/cpuidle/cpuidle.c
Earlier cpuidle assumed pm_idle as the default idle loop. Break that
assumption and make it more generic. cpuidle_idle_call() which is the
main idle loop of cpuidle is to be called by architectures which have
registered to cpuidle.
Signed-off-by: Arun R Bharadwaj <arun@linux.vnet.ibm.com>
---
drivers/cpuidle/cpuidle.c | 93 +++++++++++----------------------------------
drivers/cpuidle/cpuidle.h | 6 --
drivers/cpuidle/driver.c | 4 -
drivers/cpuidle/governor.c | 13 ++----
drivers/cpuidle/sysfs.c | 34 +++++++++-------
include/linux/cpuidle.h | 10 +++-
6 files changed, 58 insertions(+), 102 deletions(-)
Index: linux.trees.git/drivers/cpuidle/cpuidle.c
===================================================================
--- linux.trees.git.orig/drivers/cpuidle/cpuidle.c
+++ linux.trees.git/drivers/cpuidle/cpuidle.c
@@ -24,10 +24,6 @@
DEFINE_PER_CPU(struct cpuidle_device *, cpuidle_devices);
DEFINE_MUTEX(cpuidle_lock);
-LIST_HEAD(cpuidle_detected_devices);
-static void (*pm_idle_old)(void);
-
-static int enabled_devices;
#if defined(CONFIG_ARCH_HAS_CPU_IDLE_WAIT)
static void cpuidle_kick_cpus(void)
@@ -47,21 +43,20 @@ static int __cpuidle_register_device(str
*
* NOTE: no locks or semaphores should be used here
*/
-static void cpuidle_idle_call(void)
+void cpuidle_idle_call(void)
{
struct cpuidle_device *dev = __get_cpu_var(cpuidle_devices);
struct cpuidle_state *target_state;
int next_state;
+ ktime_t t1, t2;
+ s64 diff;
/* check if the device is ready */
if (!dev || !dev->enabled) {
- if (pm_idle_old)
- pm_idle_old();
- else
#if defined(CONFIG_ARCH_HAS_DEFAULT_IDLE)
- default_idle();
+ default_idle();
#else
- local_irq_enable();
+ local_irq_enable();
#endif
return;
}
@@ -75,7 +70,11 @@ static void cpuidle_idle_call(void)
hrtimer_peek_ahead_timers();
#endif
/* ask the governor for the next state */
- next_state = cpuidle_curr_governor->select(dev);
+ if (dev->state_count > 1)
+ next_state = cpuidle_curr_governor->select(dev);
+ else
+ next_state = 0;
+
if (need_resched()) {
local_irq_enable();
return;
@@ -85,7 +84,18 @@ static void cpuidle_idle_call(void)
/* enter the state and update stats */
dev->last_state = target_state;
- dev->last_residency = target_state->enter(dev, target_state);
+
+ t1 = ktime_get();
+
+ target_state->enter(dev, target_state);
+
+ t2 = ktime_get();
+ diff = ktime_to_us(ktime_sub(t2, t1));
+ if (diff > INT_MAX)
+ diff = INT_MAX;
+
+ dev->last_residency = (int) diff;
+
if (dev->last_state)
target_state = dev->last_state;
@@ -99,35 +109,12 @@ static void cpuidle_idle_call(void)
}
/**
- * cpuidle_install_idle_handler - installs the cpuidle idle loop handler
- */
-void cpuidle_install_idle_handler(void)
-{
- if (enabled_devices && (pm_idle != cpuidle_idle_call)) {
- /* Make sure all changes finished before we switch to new idle */
- smp_wmb();
- pm_idle = cpuidle_idle_call;
- }
-}
-
-/**
- * cpuidle_uninstall_idle_handler - uninstalls the cpuidle idle loop handler
- */
-void cpuidle_uninstall_idle_handler(void)
-{
- if (enabled_devices && pm_idle_old && (pm_idle != pm_idle_old)) {
- pm_idle = pm_idle_old;
- cpuidle_kick_cpus();
- }
-}
-
-/**
* cpuidle_pause_and_lock - temporarily disables CPUIDLE
*/
void cpuidle_pause_and_lock(void)
{
mutex_lock(&cpuidle_lock);
- cpuidle_uninstall_idle_handler();
+ cpuidle_kick_cpus();
}
EXPORT_SYMBOL_GPL(cpuidle_pause_and_lock);
@@ -137,7 +124,6 @@ EXPORT_SYMBOL_GPL(cpuidle_pause_and_lock
*/
void cpuidle_resume_and_unlock(void)
{
- cpuidle_install_idle_handler();
mutex_unlock(&cpuidle_lock);
}
@@ -185,7 +171,6 @@ int cpuidle_enable_device(struct cpuidle
dev->enabled = 1;
- enabled_devices++;
return 0;
fail_sysfs:
@@ -216,30 +201,16 @@ void cpuidle_disable_device(struct cpuid
cpuidle_curr_governor->disable(dev);
cpuidle_remove_state_sysfs(dev);
- enabled_devices--;
}
EXPORT_SYMBOL_GPL(cpuidle_disable_device);
#ifdef CONFIG_ARCH_HAS_CPU_RELAX
-static int poll_idle(struct cpuidle_device *dev, struct cpuidle_state *st)
+static void poll_idle(struct cpuidle_device *dev, struct cpuidle_state *st)
{
- ktime_t t1, t2;
- s64 diff;
- int ret;
-
- t1 = ktime_get();
local_irq_enable();
while (!need_resched())
cpu_relax();
-
- t2 = ktime_get();
- diff = ktime_to_us(ktime_sub(t2, t1));
- if (diff > INT_MAX)
- diff = INT_MAX;
-
- ret = (int) diff;
- return ret;
}
static void poll_idle_init(struct cpuidle_device *dev)
@@ -269,7 +240,6 @@ static void poll_idle_init(struct cpuidl
*/
static int __cpuidle_register_device(struct cpuidle_device *dev)
{
- int ret;
struct sys_device *sys_dev = get_cpu_sysdev((unsigned long)dev->cpu);
if (!sys_dev)
@@ -277,16 +247,9 @@ static int __cpuidle_register_device(str
if (!try_module_get(cpuidle_curr_driver->owner))
return -EINVAL;
- init_completion(&dev->kobj_unregister);
-
poll_idle_init(dev);
per_cpu(cpuidle_devices, dev->cpu) = dev;
- list_add(&dev->device_list, &cpuidle_detected_devices);
- if ((ret = cpuidle_add_sysfs(sys_dev))) {
- module_put(cpuidle_curr_driver->owner);
- return ret;
- }
dev->registered = 1;
return 0;
@@ -308,7 +271,6 @@ int cpuidle_register_device(struct cpuid
}
cpuidle_enable_device(dev);
- cpuidle_install_idle_handler();
mutex_unlock(&cpuidle_lock);
@@ -324,8 +286,6 @@ EXPORT_SYMBOL_GPL(cpuidle_register_devic
*/
void cpuidle_unregister_device(struct cpuidle_device *dev)
{
- struct sys_device *sys_dev = get_cpu_sysdev((unsigned long)dev->cpu);
-
if (dev->registered == 0)
return;
@@ -333,9 +293,6 @@ void cpuidle_unregister_device(struct cp
cpuidle_disable_device(dev);
- cpuidle_remove_sysfs(sys_dev);
- list_del(&dev->device_list);
- wait_for_completion(&dev->kobj_unregister);
per_cpu(cpuidle_devices, dev->cpu) = NULL;
cpuidle_resume_and_unlock();
@@ -387,8 +344,6 @@ static int __init cpuidle_init(void)
{
int ret;
- pm_idle_old = pm_idle;
-
ret = cpuidle_add_class_sysfs(&cpu_sysdev_class);
if (ret)
return ret;
Index: linux.trees.git/drivers/cpuidle/governor.c
===================================================================
--- linux.trees.git.orig/drivers/cpuidle/governor.c
+++ linux.trees.git/drivers/cpuidle/governor.c
@@ -43,16 +43,14 @@ static struct cpuidle_governor * __cpuid
*/
int cpuidle_switch_governor(struct cpuidle_governor *gov)
{
- struct cpuidle_device *dev;
+ int cpu;
if (gov == cpuidle_curr_governor)
return 0;
- cpuidle_uninstall_idle_handler();
-
if (cpuidle_curr_governor) {
- list_for_each_entry(dev, &cpuidle_detected_devices, device_list)
- cpuidle_disable_device(dev);
+ for_each_online_cpu(cpu)
+ cpuidle_disable_device(per_cpu(cpuidle_devices, cpu));
module_put(cpuidle_curr_governor->owner);
}
@@ -61,9 +59,8 @@ int cpuidle_switch_governor(struct cpuid
if (gov) {
if (!try_module_get(cpuidle_curr_governor->owner))
return -EINVAL;
- list_for_each_entry(dev, &cpuidle_detected_devices, device_list)
- cpuidle_enable_device(dev);
- cpuidle_install_idle_handler();
+ for_each_online_cpu(cpu)
+ cpuidle_enable_device(per_cpu(cpuidle_devices, cpu));
printk(KERN_INFO "cpuidle: using governor %s\n", gov->name);
}
Index: linux.trees.git/include/linux/cpuidle.h
===================================================================
--- linux.trees.git.orig/include/linux/cpuidle.h
+++ linux.trees.git/include/linux/cpuidle.h
@@ -41,7 +41,7 @@ struct cpuidle_state {
unsigned long long usage;
unsigned long long time; /* in US */
- int (*enter) (struct cpuidle_device *dev,
+ void (*enter) (struct cpuidle_device *dev,
struct cpuidle_state *state);
};
@@ -92,7 +92,6 @@ struct cpuidle_device {
struct cpuidle_state_kobj *kobjs[CPUIDLE_STATE_MAX];
struct cpuidle_state *last_state;
- struct list_head device_list;
struct kobject kobj;
struct completion kobj_unregister;
void *governor_data;
@@ -112,6 +111,9 @@ static inline int cpuidle_get_last_resid
return dev->last_residency;
}
+extern struct cpuidle_driver *cpuidle_curr_driver;
+extern void cpuidle_idle_call(void);
+
/****************************
* CPUIDLE DRIVER INTERFACE *
@@ -133,6 +135,8 @@ extern void cpuidle_pause_and_lock(void)
extern void cpuidle_resume_and_unlock(void);
extern int cpuidle_enable_device(struct cpuidle_device *dev);
extern void cpuidle_disable_device(struct cpuidle_device *dev);
+extern int common_idle_loop(struct cpuidle_device *dev,
+ struct cpuidle_state *st, void (*idle)(void));
#else
@@ -148,6 +152,8 @@ static inline void cpuidle_resume_and_un
static inline int cpuidle_enable_device(struct cpuidle_device *dev)
{return 0;}
static inline void cpuidle_disable_device(struct cpuidle_device *dev) { }
+static inline int common_idle_loop(struct cpuidle_device *dev,
+ struct cpuidle_state *st, void (*idle)(void)) { }
#endif
Index: linux.trees.git/drivers/cpuidle/cpuidle.h
===================================================================
--- linux.trees.git.orig/drivers/cpuidle/cpuidle.h
+++ linux.trees.git/drivers/cpuidle/cpuidle.h
@@ -9,9 +9,7 @@
/* For internal use only */
extern struct cpuidle_governor *cpuidle_curr_governor;
-extern struct cpuidle_driver *cpuidle_curr_driver;
extern struct list_head cpuidle_governors;
-extern struct list_head cpuidle_detected_devices;
extern struct mutex cpuidle_lock;
extern spinlock_t cpuidle_driver_lock;
@@ -27,7 +25,7 @@ extern int cpuidle_add_class_sysfs(struc
extern void cpuidle_remove_class_sysfs(struct sysdev_class *cls);
extern int cpuidle_add_state_sysfs(struct cpuidle_device *device);
extern void cpuidle_remove_state_sysfs(struct cpuidle_device *device);
-extern int cpuidle_add_sysfs(struct sys_device *sysdev);
-extern void cpuidle_remove_sysfs(struct sys_device *sysdev);
+extern int cpuidle_add_sysfs(struct cpuidle_device *device);
+extern void cpuidle_remove_sysfs(struct cpuidle_device *device);
#endif /* __DRIVER_CPUIDLE_H */
Index: linux.trees.git/drivers/cpuidle/sysfs.c
===================================================================
--- linux.trees.git.orig/drivers/cpuidle/sysfs.c
+++ linux.trees.git/drivers/cpuidle/sysfs.c
@@ -311,6 +311,13 @@ int cpuidle_add_state_sysfs(struct cpuid
int i, ret = -ENOMEM;
struct cpuidle_state_kobj *kobj;
+ init_completion(&device->kobj_unregister);
+
+ ret = cpuidle_add_sysfs(device);
+ if (ret) {
+ module_put(cpuidle_curr_driver->owner);
+ return ret;
+ }
/* state statistics */
for (i = 0; i < device->state_count; i++) {
kobj = kzalloc(sizeof(struct cpuidle_state_kobj), GFP_KERNEL);
@@ -347,35 +354,32 @@ void cpuidle_remove_state_sysfs(struct c
for (i = 0; i < device->state_count; i++)
cpuidle_free_state_kobj(device, i);
+
+ cpuidle_remove_sysfs(device);
}
/**
* cpuidle_add_sysfs - creates a sysfs instance for the target device
- * @sysdev: the target device
+ * @device: the target device
*/
-int cpuidle_add_sysfs(struct sys_device *sysdev)
+int cpuidle_add_sysfs(struct cpuidle_device *device)
{
- int cpu = sysdev->id;
- struct cpuidle_device *dev;
int error;
+ struct sys_device *sysdev = get_cpu_sysdev((unsigned long)device->cpu);
- dev = per_cpu(cpuidle_devices, cpu);
- error = kobject_init_and_add(&dev->kobj, &ktype_cpuidle, &sysdev->kobj,
- "cpuidle");
+ error = kobject_init_and_add(&device->kobj, &ktype_cpuidle,
+ &sysdev->kobj, "cpuidle");
if (!error)
- kobject_uevent(&dev->kobj, KOBJ_ADD);
+ kobject_uevent(&device->kobj, KOBJ_ADD);
return error;
}
/**
* cpuidle_remove_sysfs - deletes a sysfs instance on the target device
- * @sysdev: the target device
+ * @device: the target device
*/
-void cpuidle_remove_sysfs(struct sys_device *sysdev)
+void cpuidle_remove_sysfs(struct cpuidle_device *device)
{
- int cpu = sysdev->id;
- struct cpuidle_device *dev;
-
- dev = per_cpu(cpuidle_devices, cpu);
- kobject_put(&dev->kobj);
+ kobject_put(&device->kobj);
+ wait_for_completion(&device->kobj_unregister);
}
Index: linux.trees.git/drivers/cpuidle/driver.c
===================================================================
--- linux.trees.git.orig/drivers/cpuidle/driver.c
+++ linux.trees.git/drivers/cpuidle/driver.c
@@ -27,10 +27,6 @@ int cpuidle_register_driver(struct cpuid
return -EINVAL;
spin_lock(&cpuidle_driver_lock);
- if (cpuidle_curr_driver) {
- spin_unlock(&cpuidle_driver_lock);
- return -EBUSY;
- }
cpuidle_curr_driver = drv;
spin_unlock(&cpuidle_driver_lock);
^ permalink raw reply
* [v10 PATCH 1/9]: cpuidle: Design documentation patch
From: Arun R Bharadwaj @ 2009-12-02 9:55 UTC (permalink / raw)
To: Peter Zijlstra, Benjamin Herrenschmidt, Ingo Molnar,
Vaidyanathan Srinivasan, Dipankar Sarma, Balbir Singh,
Venkatesh Pallipadi, Arun Bharadwaj
Cc: linux-arch, linux-acpi, linuxppc-dev, linux-kernel
In-Reply-To: <20091202095427.GA27251@linux.vnet.ibm.com>
* Arun R Bharadwaj <arun@linux.vnet.ibm.com> [2009-12-02 15:24:27]:
This patch adds a little information about the redesigned cpuidle
infrastructure in Documentation/cpuidle/core.txt
Signed-off-by: Arun R Bharadwaj <arun@linux.vnet.ibm.com>
---
Documentation/cpuidle/core.txt | 35 +++++++++++++++++++++++++++++++++++
1 file changed, 35 insertions(+)
Index: linux.trees.git/Documentation/cpuidle/core.txt
===================================================================
--- linux.trees.git.orig/Documentation/cpuidle/core.txt
+++ linux.trees.git/Documentation/cpuidle/core.txt
@@ -21,3 +21,38 @@ which can be used to switch governors at
is meant for developer testing only. In normal usage, kernel picks the
best governor based on governor ratings.
SEE ALSO: sysfs.txt in this directory.
+
+Design:
+
+Cpuidle allows for registration of multiple sets of idle routines.
+The latest registered set is used by cpuidle governors as the current
+active set to choose the right idle state. This set is managed as a
+list and each time the newly registered set is added to the head of the
+list and made the current active set.
+
+An example of how this would work on x86 is shown below.
+
+----------------- -----------------
+| | | |
+| choose b/w | mwait is chosen | mwait |
+| mwait, poll, |-------------------------------------> |(current active|
+| default, c1e | register to cpuidle | set) |
+| | with mwait as the idle routine | |
+----------------- -----------------
+
+
+----------------- -----------------
+| | | c1, c2, c3 |
+| ACPI | register to cpuidle | (current) |
+| discovery |-------------------------------------> |---------------|
+| | with c1, c2, c3 | mwait |
+| | as set of idle routines | |
+----------------- -----------------
+
+With this mechanism, a module can register and unregister its set of
+idle routines at run time in a clean manner.
+
+The main idle routine called inside cpu_idle() of every arch is defined in
+driver/cpuidle/cpuidle.c which would in turn call the idle routine selected
+by the governor. If the CONFIG_CPU_IDLE is disabled, the arch needs to
+provide an alternate definition for cpuidle_idle_call().
^ permalink raw reply
* [v10 PATCH 0/9] cpuidle: cleanup cpuidle/ introduce cpuidle to POWER
From: Arun R Bharadwaj @ 2009-12-02 9:54 UTC (permalink / raw)
To: Peter Zijlstra, Benjamin Herrenschmidt, Ingo Molnar,
Vaidyanathan Srinivasan, Dipankar Sarma, Balbir Singh,
Venkatesh Pallipadi, Arun Bharadwaj
Cc: linux-arch, linux-acpi, linuxppc-dev, linux-kernel
Hi,
This patchset introduces cpuidle infrastructure to POWER, prototyping
for pSeries, and also does a major refactoring of current x86 idle
power management and a cleanup of cpuidle infrastructure.
This patch series has been in discussion for quite a while now and
below are the links to the previous discussions.
Please consider this for inclusion into the -tip tree.
v9 --> http://lkml.org/lkml/2009/10/16/63
v8 --> http://lkml.org/lkml/2009/10/8/82
v7 --> http://lkml.org/lkml/2009/10/6/278
v6 --> http://lkml.org/lkml/2009/9/22/180
v5 --> http://lkml.org/lkml/2009/9/22/26
v4 --> http://lkml.org/lkml/2009/9/1/133
v3 --> http://lkml.org/lkml/2009/8/27/124
v2 --> http://lkml.org/lkml/2009/8/26/233
v1 --> http://lkml.org/lkml/2009/8/19/150
Change in this version:
Pavel noticed that the code which calls the cpuidle's idle
loop was repeated at many places. So this set optimizes it so
that we dont have repetition of code. The rest of the patches
are same as the earlier iteration.
arun
^ permalink raw reply
* How to use mpic timer of mpc8640d
From: Thirumalai @ 2009-12-02 9:12 UTC (permalink / raw)
To: Kumar Gala; +Cc: linuxppc-dev, linux-kernel
Hi kumar,
I just want to know how to use the global timer of the MPC8640D on linux
environment. I am having a Freescale HPCN evaluation board on which
linux-2.6.21 kernel is running. How to use the timer.
Thank you
T.
^ permalink raw reply
* Re: [PATCH] powerpc/mm: setting mmaped page cache property through device tree
From: Li Yang @ 2009-12-02 6:25 UTC (permalink / raw)
To: Segher Boessenkool; +Cc: paulus, linux-kernel, linuxppc-dev
In-Reply-To: <E389CC83-307F-4A7D-B90B-CD928A1D1D1D@kernel.crashing.org>
On Tue, Dec 1, 2009 at 10:35 PM, Segher Boessenkool
<segher@kernel.crashing.org> wrote:
>> The scenario for the first case is that in a multicore system running
>> ASMP which means different OS runs on different cores. =C2=A0They might
>> communicate through a shared memory region. =C2=A0The region on every OS
>> need to be mapped with the same cache perperty to avoid cache paradox.
>
> This isn't true. =C2=A0In ASMP, you cannot usually do coherency between
> the different CPUs at all. =C2=A0Also, in most PowerPC implementations,
Coherency can't be achieved with proper configuration and management? Why =
so?
> it is fine if one CPU maps a memory range as coherent while another
> maps it as non-coherent; sure, you have to be careful or you will
But we do want the shared region to be coherent. So mappings should
have the same cacheability property.
> read stale data, but things won't wedge.
>
>> The scenario for the second case is to pre-allocate some memory to a
>> certain application or device (probably through mem=3DXXX kernel
>> parameter or limit through device tree). =C2=A0The memory is not known t=
o
>> kernel, but fully managed by the application/device. =C2=A0We need being
>> able to map the region cachable for better performance.
>
> So make the memory known to the kernel, just tell the kernel not to
> use it. =C2=A0If it's normal system RAM, just put it in the "memory" node
> and do a memreserve on it (or do something in your platform code); if
> it's some other memory, do a device driver for it, map it there.
Your solution is feasible. But the memory allocation is a software
configuration. IMHO, it should be better and easier addressed by
changing configurations(like mem parameter) rather than the kernel
platform code which should address hardware configuration.
- Leo
^ permalink raw reply
* Re: EP440XS board boot problem(Note:similart to Sequoia)
From: Allan Wang @ 2009-12-02 4:22 UTC (permalink / raw)
To: Wolfgang Denk; +Cc: linuxppc-dev
In-Reply-To: <20091201231022.AB72CE6D391@gemini.denx.de>
Hello, Mr Denk,=0A=0AThank you for your help. My answer to your questions a=
re between the lines. Attached is the dts file I use.=0A=0ABest regards,=0A=
=0AAllan=0A=0A=0A=0A----- Original Message ----=0AFrom: Wolfgang Denk <wd@d=
enx.de>=0ATo: Allan Wang <aiyuanwang@yahoo.com>=0ACc: linuxppc-dev@lists.oz=
labs.org=0ASent: Tue, December 1, 2009 6:10:22 PM=0ASubject: Re: EP440XS bo=
ard boot problem(Note:similart to Sequoia)=0A=0ADear Allan Wang,=0A=0AIn me=
ssage <401793.31718.qm@web52903.mail.re2.yahoo.com> you wrote:=0A>=0A> I am=
new to the list. I am still newbie to embedded linux. I have=0A> bought an=
ep440xs board which is similar to sequoia board. u-boot is=0A> working. I =
am trying to make it run linux. I have the following=0A> problems and need =
your help. Thank you very much:=0A=0AWhich exact version of U-Boot are you =
running on this board?=0A=0A=3D> version=0A=0AU-Boot 1.1.6 (Oct 25 2009 - 1=
3:26:41)=0A=0A=0AI guess this is a out-of-tree port, i. e. not part of the =
mainline=0AU-Boot repository - so are you sure that device tree support is=
=0Aenabled in your version of U-Boot?=0A=0AThe device tree is not enabled i=
n u-boot. I am building cuImage, so the device tree is supported in kernel(=
2.6.30.4).=0A=0A> =0A> I am using linux kernel 2.6.30.4. I encounter 2 prob=
lems:=0A> 1. After console_init() is called inside start_kernel(), the cons=
ole starts printing garbage like:=0A> Serial: 8250/16550 driver, 4 ports, I=
RQ sharing enabled=0A> serial8250.0: ttyS0 at MMIO 0x1ef600300 (irq =3D 16)=
is a 16550A=0A> =EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=
=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=
=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=
=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=0A...=0A=0A=
What is your exact boot command?=0A=0AI tftp the cuImage to address 0x20000=
0 and bootm to boot.=0A=0AWhich boot arguments do you pass to the kernel?=
=0A=0Abootargs=3D$(bootargs) console=3DttyS0,$(baudrate)=0A=0AWhat is ytour=
console baud rate in U-Boot, and what sort of "console=3D"=0Aboot argument=
do you pass to the Linux kernel?=0A=0AThe u-boot sets it to 115.2k. the dt=
s file is set to 115.2k. Here is the boot parameter:=0Abaudrate=3D115200=0A=
=0AWhat does your device tree look like?=0A=0AI attached my dts file.=0A=0A=
> 2. It pakics when mounting the root file system(NFS):=0A> (gdb) bt=0A> #0=
panic (fmt=3D0xc0290e94 "VFS: Unable to mount root fs on %s")=0A=0ASo what=
is the "%s" argument here? Which sort of root file system are=0Ayou trying=
to mount - NFS, ramdisk, UBIFS, ... ?=0A=0AI try to use NFS for the root f=
ile system.=0A=0ABest regards,=0A=0AWolfgang Denk=0A=0A-- =0ADENX Software =
Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel=0AHRB 165235 Munich=
, Office: Kirchenstr.5, D-82194 Groebenzell, Germany=0APhone: (+49)-8142-66=
989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de=0A"He only drinks when he=
gets depressed." "Why does he get depressed?"=0A"Sometimes it's because he=
hasn't had a drink."=0A - Terry Pratch=
ett, _Men at Arms_=0A=0A=0A=0A
^ permalink raw reply
* Re: [PATCH v3 3/3] powerpc/fsl: 85xx: add cache-sram support
From: Kumar Gala @ 2009-12-02 3:26 UTC (permalink / raw)
To: Scott Wood
Cc: Mahajan Vivek-B08308, Wood Scott-B07421, Gala Kumar-B11780,
linuxppc-dev
In-Reply-To: <4B154C82.5060100@freescale.com>
On Dec 1, 2009, at 11:04 AM, Scott Wood wrote:
> Mahajan Vivek-B08308 wrote:
>>> From: Wood Scott-B07421 Sent: Friday, November 20, 2009 11:09 PM
>>>> Cache-sram does not have any device tree entry since it is not a
>>>> hardware as such. Putting it under chosen can be another option.
>>>> I think, Scott (cc'ed) was of the opinion that since 32b
>>> base address
>>>> support is missing; so there is no point in moving this
>>> address to the
>>>> command line and .config should be okay for now for it.
>>> I don't know what you mean by "32b base address support is
>>> missing". I have no objection to putting it on the command line.
>> It was a typo, it should be missing 36b address support. Since the
>> kernel did not run under different environment (i.e 32b / 36b base
>> address); so
>> it was decided that .config may be okay for this.
>
> The kernel *will* run with a 36-bit address map even if some of our
> drivers don't support it right at the moment -- so we should be
> planning for it.
Not only that but I'd like to move any future parts to only support a
36-bit memory map.
- k
^ permalink raw reply
* Re: [PATCH 5/5] libata/drivers: Add pata_macio, driver Apple PowerMac/PowerBook IDE controller
From: Tejun Heo @ 2009-12-02 1:55 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: linux-ide, linuxppc-dev, jeff
In-Reply-To: <1259718678.2076.1057.camel@pasglop>
Hello,
On 12/02/2009 10:51 AM, Benjamin Herrenschmidt wrote:
> I'm not sure what media bay locking problem you mentioned tho...
>
> The way the driver does the locking now is that I block the media-bay
> thread (lock/unlock_media_bay calls) around initial registration and
> driver removal which blocks the callback event. At any other time, the
> callback event will just schedule EH so things should be fine.
>
> I've left the debouncing in the mediabay code as well for both insertion
> and removal so there should be no issue here with glitches.
Ah... okay, I was misunderstanding where the race was. If there's no
race condition against hotplug/unplug events with non-probing EH, no
problem.
--
tejun
^ permalink raw reply
* Re: [PATCH 5/5] libata/drivers: Add pata_macio, driver Apple PowerMac/PowerBook IDE controller
From: Benjamin Herrenschmidt @ 2009-12-02 1:51 UTC (permalink / raw)
To: Tejun Heo; +Cc: linux-ide, linuxppc-dev, jeff
In-Reply-To: <4B15BDCF.5020000@kernel.org>
On Wed, 2009-12-02 at 10:07 +0900, Tejun Heo wrote:
> On 12/02/2009 09:36 AM, Benjamin Herrenschmidt wrote:
> > This is a libata driver for the "macio" IDE controller used on most Apple
> > PowerMac and PowerBooks. It's a libata equivalent of drivers/ide/ppc/pmac.c
> >
> > It supports all the features of its predecessor, including mediabay hotplug
> > and suspend/resume. It should also support module load/unload.
> >
> > The timing calculations have been simplified to use pre-calculated tables
> > compared to drivers/ide/pmac.c and it uses the new mediabay interface
> > provided by a previous patch.
>
> Other than the media bay locking thing I mentioned before, it looks
> good to me and the media bay thing can be dealth with further patches,
> so...
>
> Acked-by: Tejun Heo <tj@kernel.org>
Thanks.
I'm not sure what media bay locking problem you mentioned tho...
The way the driver does the locking now is that I block the media-bay
thread (lock/unlock_media_bay calls) around initial registration and
driver removal which blocks the callback event. At any other time, the
callback event will just schedule EH so things should be fine.
I've left the debouncing in the mediabay code as well for both insertion
and removal so there should be no issue here with glitches.
Cheers,
Ben.
^ permalink raw reply
* Re: [PATCH 5/5] libata/drivers: Add pata_macio, driver Apple PowerMac/PowerBook IDE controller
From: Tejun Heo @ 2009-12-02 1:07 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: linux-ide, linuxppc-dev, jeff
In-Reply-To: <20091202003630.29919100851@ozlabs.org>
On 12/02/2009 09:36 AM, Benjamin Herrenschmidt wrote:
> This is a libata driver for the "macio" IDE controller used on most Apple
> PowerMac and PowerBooks. It's a libata equivalent of drivers/ide/ppc/pmac.c
>
> It supports all the features of its predecessor, including mediabay hotplug
> and suspend/resume. It should also support module load/unload.
>
> The timing calculations have been simplified to use pre-calculated tables
> compared to drivers/ide/pmac.c and it uses the new mediabay interface
> provided by a previous patch.
Other than the media bay locking thing I mentioned before, it looks
good to me and the media bay thing can be dealth with further patches,
so...
Acked-by: Tejun Heo <tj@kernel.org>
Thanks.
--
tejun
^ permalink raw reply
* [PATCH 5/5] libata/drivers: Add pata_macio, driver Apple PowerMac/PowerBook IDE controller
From: Benjamin Herrenschmidt @ 2009-12-02 0:36 UTC (permalink / raw)
To: linux-ide; +Cc: tj, linuxppc-dev, jeff
This is a libata driver for the "macio" IDE controller used on most Apple
PowerMac and PowerBooks. It's a libata equivalent of drivers/ide/ppc/pmac.c
It supports all the features of its predecessor, including mediabay hotplug
and suspend/resume. It should also support module load/unload.
The timing calculations have been simplified to use pre-calculated tables
compared to drivers/ide/pmac.c and it uses the new mediabay interface
provided by a previous patch.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
v2. Better tested now, seems to be reasonably solid.
Addressed Tejun comments and made remove more robust vs. media-bay,
should also fix Andreas problem.
drivers/ata/Kconfig | 10
drivers/ata/Makefile | 1
drivers/ata/pata_macio.c | 1427 +++++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 1438 insertions(+)
--- /dev/null 1970-01-01 00:00:00.000000000 +0000
+++ linux-work/drivers/ata/pata_macio.c 2009-12-02 11:26:51.000000000 +1100
@@ -0,0 +1,1427 @@
+/*
+ * Libata based driver for Apple "macio" family of PATA controllers
+ *
+ * Copyright 2008/2009 Benjamin Herrenschmidt, IBM Corp
+ * <benh@kernel.crashing.org>
+ *
+ * Some bits and pieces from drivers/ide/ppc/pmac.c
+ *
+ */
+
+#undef DEBUG
+#undef DEBUG_DMA
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/blkdev.h>
+#include <linux/ata.h>
+#include <linux/libata.h>
+#include <linux/adb.h>
+#include <linux/pmu.h>
+#include <linux/scatterlist.h>
+#include <linux/of.h>
+
+#include <scsi/scsi.h>
+#include <scsi/scsi_host.h>
+#include <scsi/scsi_device.h>
+
+#include <asm/macio.h>
+#include <asm/io.h>
+#include <asm/dbdma.h>
+#include <asm/pci-bridge.h>
+#include <asm/machdep.h>
+#include <asm/pmac_feature.h>
+#include <asm/mediabay.h>
+
+#ifdef DEBUG_DMA
+#define dev_dbgdma(dev, format, arg...) \
+ dev_printk(KERN_DEBUG , dev , format , ## arg)
+#else
+#define dev_dbgdma(dev, format, arg...) \
+ ({ if (0) dev_printk(KERN_DEBUG, dev, format, ##arg); 0; })
+#endif
+
+#define DRV_NAME "pata_macio"
+#define DRV_VERSION "0.9"
+
+/* Models of macio ATA controller */
+enum {
+ controller_ohare, /* OHare based */
+ controller_heathrow, /* Heathrow/Paddington */
+ controller_kl_ata3, /* KeyLargo ATA-3 */
+ controller_kl_ata4, /* KeyLargo ATA-4 */
+ controller_un_ata6, /* UniNorth2 ATA-6 */
+ controller_k2_ata6, /* K2 ATA-6 */
+ controller_sh_ata6, /* Shasta ATA-6 */
+};
+
+static const char* macio_ata_names[] = {
+ "OHare ATA", /* OHare based */
+ "Heathrow ATA", /* Heathrow/Paddington */
+ "KeyLargo ATA-3", /* KeyLargo ATA-3 (MDMA only) */
+ "KeyLargo ATA-4", /* KeyLargo ATA-4 (UDMA/66) */
+ "UniNorth ATA-6", /* UniNorth2 ATA-6 (UDMA/100) */
+ "K2 ATA-6", /* K2 ATA-6 (UDMA/100) */
+ "Shasta ATA-6", /* Shasta ATA-6 (UDMA/133) */
+};
+
+/*
+ * Extra registers, both 32-bit little-endian
+ */
+#define IDE_TIMING_CONFIG 0x200
+#define IDE_INTERRUPT 0x300
+
+/* Kauai (U2) ATA has different register setup */
+#define IDE_KAUAI_PIO_CONFIG 0x200
+#define IDE_KAUAI_ULTRA_CONFIG 0x210
+#define IDE_KAUAI_POLL_CONFIG 0x220
+
+/*
+ * Timing configuration register definitions
+ */
+
+/* Number of IDE_SYSCLK_NS ticks, argument is in nanoseconds */
+#define SYSCLK_TICKS(t) (((t) + IDE_SYSCLK_NS - 1) / IDE_SYSCLK_NS)
+#define SYSCLK_TICKS_66(t) (((t) + IDE_SYSCLK_66_NS - 1) / IDE_SYSCLK_66_NS)
+#define IDE_SYSCLK_NS 30 /* 33Mhz cell */
+#define IDE_SYSCLK_66_NS 15 /* 66Mhz cell */
+
+/* 133Mhz cell, found in shasta.
+ * See comments about 100 Mhz Uninorth 2...
+ * Note that PIO_MASK and MDMA_MASK seem to overlap, that's just
+ * weird and I don't now why .. at this stage
+ */
+#define TR_133_PIOREG_PIO_MASK 0xff000fff
+#define TR_133_PIOREG_MDMA_MASK 0x00fff800
+#define TR_133_UDMAREG_UDMA_MASK 0x0003ffff
+#define TR_133_UDMAREG_UDMA_EN 0x00000001
+
+/* 100Mhz cell, found in Uninorth 2 and K2. It appears as a pci device
+ * (106b/0033) on uninorth or K2 internal PCI bus and it's clock is
+ * controlled like gem or fw. It appears to be an evolution of keylargo
+ * ATA4 with a timing register extended to 2x32bits registers (one
+ * for PIO & MWDMA and one for UDMA, and a similar DBDMA channel.
+ * It has it's own local feature control register as well.
+ *
+ * After scratching my mind over the timing values, at least for PIO
+ * and MDMA, I think I've figured the format of the timing register,
+ * though I use pre-calculated tables for UDMA as usual...
+ */
+#define TR_100_PIO_ADDRSETUP_MASK 0xff000000 /* Size of field unknown */
+#define TR_100_PIO_ADDRSETUP_SHIFT 24
+#define TR_100_MDMA_MASK 0x00fff000
+#define TR_100_MDMA_RECOVERY_MASK 0x00fc0000
+#define TR_100_MDMA_RECOVERY_SHIFT 18
+#define TR_100_MDMA_ACCESS_MASK 0x0003f000
+#define TR_100_MDMA_ACCESS_SHIFT 12
+#define TR_100_PIO_MASK 0xff000fff
+#define TR_100_PIO_RECOVERY_MASK 0x00000fc0
+#define TR_100_PIO_RECOVERY_SHIFT 6
+#define TR_100_PIO_ACCESS_MASK 0x0000003f
+#define TR_100_PIO_ACCESS_SHIFT 0
+
+#define TR_100_UDMAREG_UDMA_MASK 0x0000ffff
+#define TR_100_UDMAREG_UDMA_EN 0x00000001
+
+
+/* 66Mhz cell, found in KeyLargo. Can do ultra mode 0 to 2 on
+ * 40 connector cable and to 4 on 80 connector one.
+ * Clock unit is 15ns (66Mhz)
+ *
+ * 3 Values can be programmed:
+ * - Write data setup, which appears to match the cycle time. They
+ * also call it DIOW setup.
+ * - Ready to pause time (from spec)
+ * - Address setup. That one is weird. I don't see where exactly
+ * it fits in UDMA cycles, I got it's name from an obscure piece
+ * of commented out code in Darwin. They leave it to 0, we do as
+ * well, despite a comment that would lead to think it has a
+ * min value of 45ns.
+ * Apple also add 60ns to the write data setup (or cycle time ?) on
+ * reads.
+ */
+#define TR_66_UDMA_MASK 0xfff00000
+#define TR_66_UDMA_EN 0x00100000 /* Enable Ultra mode for DMA */
+#define TR_66_PIO_ADDRSETUP_MASK 0xe0000000 /* Address setup */
+#define TR_66_PIO_ADDRSETUP_SHIFT 29
+#define TR_66_UDMA_RDY2PAUS_MASK 0x1e000000 /* Ready 2 pause time */
+#define TR_66_UDMA_RDY2PAUS_SHIFT 25
+#define TR_66_UDMA_WRDATASETUP_MASK 0x01e00000 /* Write data setup time */
+#define TR_66_UDMA_WRDATASETUP_SHIFT 21
+#define TR_66_MDMA_MASK 0x000ffc00
+#define TR_66_MDMA_RECOVERY_MASK 0x000f8000
+#define TR_66_MDMA_RECOVERY_SHIFT 15
+#define TR_66_MDMA_ACCESS_MASK 0x00007c00
+#define TR_66_MDMA_ACCESS_SHIFT 10
+#define TR_66_PIO_MASK 0xe00003ff
+#define TR_66_PIO_RECOVERY_MASK 0x000003e0
+#define TR_66_PIO_RECOVERY_SHIFT 5
+#define TR_66_PIO_ACCESS_MASK 0x0000001f
+#define TR_66_PIO_ACCESS_SHIFT 0
+
+/* 33Mhz cell, found in OHare, Heathrow (& Paddington) and KeyLargo
+ * Can do pio & mdma modes, clock unit is 30ns (33Mhz)
+ *
+ * The access time and recovery time can be programmed. Some older
+ * Darwin code base limit OHare to 150ns cycle time. I decided to do
+ * the same here fore safety against broken old hardware ;)
+ * The HalfTick bit, when set, adds half a clock (15ns) to the access
+ * time and removes one from recovery. It's not supported on KeyLargo
+ * implementation afaik. The E bit appears to be set for PIO mode 0 and
+ * is used to reach long timings used in this mode.
+ */
+#define TR_33_MDMA_MASK 0x003ff800
+#define TR_33_MDMA_RECOVERY_MASK 0x001f0000
+#define TR_33_MDMA_RECOVERY_SHIFT 16
+#define TR_33_MDMA_ACCESS_MASK 0x0000f800
+#define TR_33_MDMA_ACCESS_SHIFT 11
+#define TR_33_MDMA_HALFTICK 0x00200000
+#define TR_33_PIO_MASK 0x000007ff
+#define TR_33_PIO_E 0x00000400
+#define TR_33_PIO_RECOVERY_MASK 0x000003e0
+#define TR_33_PIO_RECOVERY_SHIFT 5
+#define TR_33_PIO_ACCESS_MASK 0x0000001f
+#define TR_33_PIO_ACCESS_SHIFT 0
+
+/*
+ * Interrupt register definitions. Only present on newer cells
+ * (Keylargo and later afaik) so we don't use it.
+ */
+#define IDE_INTR_DMA 0x80000000
+#define IDE_INTR_DEVICE 0x40000000
+
+/*
+ * FCR Register on Kauai. Not sure what bit 0x4 is ...
+ */
+#define KAUAI_FCR_UATA_MAGIC 0x00000004
+#define KAUAI_FCR_UATA_RESET_N 0x00000002
+#define KAUAI_FCR_UATA_ENABLE 0x00000001
+
+
+/* Allow up to 256 DBDMA commands per xfer */
+#define MAX_DCMDS 256
+
+/* Don't let a DMA segment go all the way to 64K */
+#define MAX_DBDMA_SEG 0xff00
+
+
+/*
+ * Wait 1s for disk to answer on IDE bus after a hard reset
+ * of the device (via GPIO/FCR).
+ *
+ * Some devices seem to "pollute" the bus even after dropping
+ * the BSY bit (typically some combo drives slave on the UDMA
+ * bus) after a hard reset. Since we hard reset all drives on
+ * KeyLargo ATA66, we have to keep that delay around. I may end
+ * up not hard resetting anymore on these and keep the delay only
+ * for older interfaces instead (we have to reset when coming
+ * from MacOS...) --BenH.
+ */
+#define IDE_WAKEUP_DELAY_MS 1000
+
+struct pata_macio_timing;
+
+struct pata_macio_priv {
+ int kind;
+ int aapl_bus_id;
+ int mediabay : 1;
+ struct device_node *node;
+ struct macio_dev *mdev;
+ struct pci_dev *pdev;
+ struct device *dev;
+ int irq;
+ u32 treg[2][2];
+ void __iomem *tfregs;
+ void __iomem *kauai_fcr;
+ struct dbdma_cmd * dma_table_cpu;
+ dma_addr_t dma_table_dma;
+ struct ata_host *host;
+ const struct pata_macio_timing *timings;
+};
+
+/* Previous variants of this driver used to calculate timings
+ * for various variants of the chip and use tables for others.
+ *
+ * Not only was this confusing, but in addition, it isn't clear
+ * whether our calculation code was correct. It didn't entirely
+ * match the darwin code and whatever documentation I could find
+ * on these cells
+ *
+ * I decided to entirely rely on a table instead for this version
+ * of the driver. Also, because I don't really care about derated
+ * modes and really old HW other than making it work, I'm not going
+ * to calculate / snoop timing values for something else than the
+ * standard modes.
+ */
+struct pata_macio_timing {
+ int mode;
+ u32 reg1; /* Bits to set in first timing reg */
+ u32 reg2; /* Bits to set in second timing reg */
+};
+
+static const struct pata_macio_timing pata_macio_ohare_timings[] = {
+ { XFER_PIO_0, 0x00000526, 0, },
+ { XFER_PIO_1, 0x00000085, 0, },
+ { XFER_PIO_2, 0x00000025, 0, },
+ { XFER_PIO_3, 0x00000025, 0, },
+ { XFER_PIO_4, 0x00000025, 0, },
+ { XFER_MW_DMA_0, 0x00074000, 0, },
+ { XFER_MW_DMA_1, 0x00221000, 0, },
+ { XFER_MW_DMA_2, 0x00211000, 0, },
+ { -1, 0, 0 }
+};
+
+static const struct pata_macio_timing pata_macio_heathrow_timings[] = {
+ { XFER_PIO_0, 0x00000526, 0, },
+ { XFER_PIO_1, 0x00000085, 0, },
+ { XFER_PIO_2, 0x00000025, 0, },
+ { XFER_PIO_3, 0x00000025, 0, },
+ { XFER_PIO_4, 0x00000025, 0, },
+ { XFER_MW_DMA_0, 0x00074000, 0, },
+ { XFER_MW_DMA_1, 0x00221000, 0, },
+ { XFER_MW_DMA_2, 0x00211000, 0, },
+ { -1, 0, 0 }
+};
+
+static const struct pata_macio_timing pata_macio_kl33_timings[] = {
+ { XFER_PIO_0, 0x00000526, 0, },
+ { XFER_PIO_1, 0x00000085, 0, },
+ { XFER_PIO_2, 0x00000025, 0, },
+ { XFER_PIO_3, 0x00000025, 0, },
+ { XFER_PIO_4, 0x00000025, 0, },
+ { XFER_MW_DMA_0, 0x00084000, 0, },
+ { XFER_MW_DMA_1, 0x00021800, 0, },
+ { XFER_MW_DMA_2, 0x00011800, 0, },
+ { -1, 0, 0 }
+};
+
+static const struct pata_macio_timing pata_macio_kl66_timings[] = {
+ { XFER_PIO_0, 0x0000038c, 0, },
+ { XFER_PIO_1, 0x0000020a, 0, },
+ { XFER_PIO_2, 0x00000127, 0, },
+ { XFER_PIO_3, 0x000000c6, 0, },
+ { XFER_PIO_4, 0x00000065, 0, },
+ { XFER_MW_DMA_0, 0x00084000, 0, },
+ { XFER_MW_DMA_1, 0x00029800, 0, },
+ { XFER_MW_DMA_2, 0x00019400, 0, },
+ { XFER_UDMA_0, 0x19100000, 0, },
+ { XFER_UDMA_1, 0x14d00000, 0, },
+ { XFER_UDMA_2, 0x10900000, 0, },
+ { XFER_UDMA_3, 0x0c700000, 0, },
+ { XFER_UDMA_4, 0x0c500000, 0, },
+ { -1, 0, 0 }
+};
+
+static const struct pata_macio_timing pata_macio_kauai_timings[] = {
+ { XFER_PIO_0, 0x08000a92, 0, },
+ { XFER_PIO_1, 0x0800060f, 0, },
+ { XFER_PIO_2, 0x0800038b, 0, },
+ { XFER_PIO_3, 0x05000249, 0, },
+ { XFER_PIO_4, 0x04000148, 0, },
+ { XFER_MW_DMA_0, 0x00618000, 0, },
+ { XFER_MW_DMA_1, 0x00209000, 0, },
+ { XFER_MW_DMA_2, 0x00148000, 0, },
+ { XFER_UDMA_0, 0, 0x000070c1, },
+ { XFER_UDMA_1, 0, 0x00005d81, },
+ { XFER_UDMA_2, 0, 0x00004a61, },
+ { XFER_UDMA_3, 0, 0x00003a51, },
+ { XFER_UDMA_4, 0, 0x00002a31, },
+ { XFER_UDMA_5, 0, 0x00002921, },
+ { -1, 0, 0 }
+};
+
+static const struct pata_macio_timing pata_macio_shasta_timings[] = {
+ { XFER_PIO_0, 0x0a000c97, 0, },
+ { XFER_PIO_1, 0x07000712, 0, },
+ { XFER_PIO_2, 0x040003cd, 0, },
+ { XFER_PIO_3, 0x0500028b, 0, },
+ { XFER_PIO_4, 0x0400010a, 0, },
+ { XFER_MW_DMA_0, 0x00820800, 0, },
+ { XFER_MW_DMA_1, 0x0028b000, 0, },
+ { XFER_MW_DMA_2, 0x001ca000, 0, },
+ { XFER_UDMA_0, 0, 0x00035901, },
+ { XFER_UDMA_1, 0, 0x000348b1, },
+ { XFER_UDMA_2, 0, 0x00033881, },
+ { XFER_UDMA_3, 0, 0x00033861, },
+ { XFER_UDMA_4, 0, 0x00033841, },
+ { XFER_UDMA_5, 0, 0x00033031, },
+ { XFER_UDMA_6, 0, 0x00033021, },
+ { -1, 0, 0 }
+};
+
+static const struct pata_macio_timing *pata_macio_find_timing(
+ struct pata_macio_priv *priv,
+ int mode)
+{
+ int i;
+
+ for (i = 0; priv->timings[i].mode > 0; i++) {
+ if (priv->timings[i].mode == mode)
+ return &priv->timings[i];
+ }
+ return NULL;
+}
+
+
+static void pata_macio_apply_timings(struct ata_port *ap, unsigned int device)
+{
+ struct pata_macio_priv *priv = ap->private_data;
+ void __iomem *rbase = ap->ioaddr.cmd_addr;
+
+ if (priv->kind == controller_sh_ata6 ||
+ priv->kind == controller_un_ata6 ||
+ priv->kind == controller_k2_ata6) {
+ writel(priv->treg[device][0], rbase + IDE_KAUAI_PIO_CONFIG);
+ writel(priv->treg[device][1], rbase + IDE_KAUAI_ULTRA_CONFIG);
+ } else
+ writel(priv->treg[device][0], rbase + IDE_TIMING_CONFIG);
+}
+
+static void pata_macio_dev_select(struct ata_port *ap, unsigned int device)
+{
+ ata_sff_dev_select(ap, device);
+
+ /* Apply timings */
+ pata_macio_apply_timings(ap, device);
+}
+
+static void pata_macio_set_timings(struct ata_port *ap,
+ struct ata_device *adev)
+{
+ struct pata_macio_priv *priv = ap->private_data;
+ const struct pata_macio_timing *t;
+
+ dev_dbg(priv->dev, "Set timings: DEV=%d,PIO=0x%x (%s),DMA=0x%x (%s)\n",
+ adev->devno,
+ adev->pio_mode,
+ ata_mode_string(ata_xfer_mode2mask(adev->pio_mode)),
+ adev->dma_mode,
+ ata_mode_string(ata_xfer_mode2mask(adev->dma_mode)));
+
+ /* First clear timings */
+ priv->treg[adev->devno][0] = priv->treg[adev->devno][1] = 0;
+
+ /* Now get the PIO timings */
+ t = pata_macio_find_timing(priv, adev->pio_mode);
+ if (t == NULL) {
+ dev_warn(priv->dev, "Invalid PIO timing requested: 0x%x\n",
+ adev->pio_mode);
+ t = pata_macio_find_timing(priv, XFER_PIO_0);
+ }
+ BUG_ON(t == NULL);
+
+ /* PIO timings only ever use the first treg */
+ priv->treg[adev->devno][0] |= t->reg1;
+
+ /* Now get DMA timings */
+ t = pata_macio_find_timing(priv, adev->dma_mode);
+ if (t == NULL || (t->reg1 == 0 && t->reg2 == 0)) {
+ dev_dbg(priv->dev, "DMA timing not set yet, using MW_DMA_0\n");
+ t = pata_macio_find_timing(priv, XFER_MW_DMA_0);
+ }
+ BUG_ON(t == NULL);
+
+ /* DMA timings can use both tregs */
+ priv->treg[adev->devno][0] |= t->reg1;
+ priv->treg[adev->devno][1] |= t->reg2;
+
+ dev_dbg(priv->dev, " -> %08x %08x\n",
+ priv->treg[adev->devno][0],
+ priv->treg[adev->devno][1]);
+
+ /* Apply to hardware */
+ pata_macio_apply_timings(ap, adev->devno);
+}
+
+/*
+ * Blast some well known "safe" values to the timing registers at init or
+ * wakeup from sleep time, before we do real calculation
+ */
+static void pata_macio_default_timings(struct pata_macio_priv *priv)
+{
+ unsigned int value, value2 = 0;
+
+ switch(priv->kind) {
+ case controller_sh_ata6:
+ value = 0x0a820c97;
+ value2 = 0x00033031;
+ break;
+ case controller_un_ata6:
+ case controller_k2_ata6:
+ value = 0x08618a92;
+ value2 = 0x00002921;
+ break;
+ case controller_kl_ata4:
+ value = 0x0008438c;
+ break;
+ case controller_kl_ata3:
+ value = 0x00084526;
+ break;
+ case controller_heathrow:
+ case controller_ohare:
+ default:
+ value = 0x00074526;
+ break;
+ }
+ priv->treg[0][0] = priv->treg[1][0] = value;
+ priv->treg[0][1] = priv->treg[1][1] = value2;
+}
+
+static int pata_macio_cable_detect(struct ata_port *ap)
+{
+ struct pata_macio_priv *priv = ap->private_data;
+
+ /* Get cable type from device-tree */
+ if (priv->kind == controller_kl_ata4 ||
+ priv->kind == controller_un_ata6 ||
+ priv->kind == controller_k2_ata6 ||
+ priv->kind == controller_sh_ata6) {
+ const char* cable = of_get_property(priv->node, "cable-type",
+ NULL);
+ struct device_node *root = of_find_node_by_path("/");
+ const char *model = of_get_property(root, "model", NULL);
+
+ if (cable && !strncmp(cable, "80-", 3)) {
+ /* Some drives fail to detect 80c cable in PowerBook
+ * These machine use proprietary short IDE cable
+ * anyway
+ */
+ if (!strncmp(model, "PowerBook", 9))
+ return ATA_CBL_PATA40_SHORT;
+ else
+ return ATA_CBL_PATA80;
+ }
+ }
+
+ /* G5's seem to have incorrect cable type in device-tree.
+ * Let's assume they always have a 80 conductor cable, this seem to
+ * be always the case unless the user mucked around
+ */
+ if (of_device_is_compatible(priv->node, "K2-UATA") ||
+ of_device_is_compatible(priv->node, "shasta-ata"))
+ return ATA_CBL_PATA80;
+
+ /* Anything else is 40 connectors */
+ return ATA_CBL_PATA40;
+}
+
+static void pata_macio_qc_prep(struct ata_queued_cmd *qc)
+{
+ unsigned int write = (qc->tf.flags & ATA_TFLAG_WRITE);
+ struct ata_port *ap = qc->ap;
+ struct pata_macio_priv *priv = ap->private_data;
+ struct scatterlist *sg;
+ struct dbdma_cmd *table;
+ unsigned int si, pi;
+
+ dev_dbgdma(priv->dev, "%s: qc %p flags %lx, write %d dev %d\n",
+ __func__, qc, qc->flags, write, qc->dev->devno);
+
+ if (!(qc->flags & ATA_QCFLAG_DMAMAP))
+ return;
+
+ table = (struct dbdma_cmd *) priv->dma_table_cpu;
+
+ pi = 0;
+ for_each_sg(qc->sg, sg, qc->n_elem, si) {
+ u32 addr, sg_len, len;
+
+ /* determine if physical DMA addr spans 64K boundary.
+ * Note h/w doesn't support 64-bit, so we unconditionally
+ * truncate dma_addr_t to u32.
+ */
+ addr = (u32) sg_dma_address(sg);
+ sg_len = sg_dma_len(sg);
+
+ while (sg_len) {
+ /* table overflow should never happen */
+ BUG_ON (pi++ >= MAX_DCMDS);
+
+ len = (sg_len < MAX_DBDMA_SEG) ? sg_len : MAX_DBDMA_SEG;
+ st_le16(&table->command, write ? OUTPUT_MORE: INPUT_MORE);
+ st_le16(&table->req_count, len);
+ st_le32(&table->phy_addr, addr);
+ table->cmd_dep = 0;
+ table->xfer_status = 0;
+ table->res_count = 0;
+ addr += len;
+ sg_len -= len;
+ ++table;
+ }
+ }
+
+ /* Should never happen according to Tejun */
+ BUG_ON(!pi);
+
+ /* Convert the last command to an input/output */
+ table--;
+ st_le16(&table->command, write ? OUTPUT_LAST: INPUT_LAST);
+ table++;
+
+ /* Add the stop command to the end of the list */
+ memset(table, 0, sizeof(struct dbdma_cmd));
+ st_le16(&table->command, DBDMA_STOP);
+
+ dev_dbgdma(priv->dev, "%s: %d DMA list entries\n", __func__, pi);
+}
+
+
+static void pata_macio_freeze(struct ata_port *ap)
+{
+ struct dbdma_regs __iomem *dma_regs = ap->ioaddr.bmdma_addr;
+
+ if (dma_regs) {
+ unsigned int timeout = 1000000;
+
+ /* Make sure DMA controller is stopped */
+ writel((RUN|PAUSE|FLUSH|WAKE|DEAD) << 16, &dma_regs->control);
+ while (--timeout && (readl(&dma_regs->status) & RUN))
+ udelay(1);
+ }
+
+ ata_sff_freeze(ap);
+}
+
+
+static void pata_macio_bmdma_setup(struct ata_queued_cmd *qc)
+{
+ struct ata_port *ap = qc->ap;
+ struct pata_macio_priv *priv = ap->private_data;
+ struct dbdma_regs __iomem *dma_regs = ap->ioaddr.bmdma_addr;
+ int dev = qc->dev->devno;
+
+ dev_dbgdma(priv->dev, "%s: qc %p\n", __func__, qc);
+
+ /* Make sure DMA commands updates are visible */
+ writel(priv->dma_table_dma, &dma_regs->cmdptr);
+
+ /* On KeyLargo 66Mhz cell, we need to add 60ns to wrDataSetup on
+ * UDMA reads
+ */
+ if (priv->kind == controller_kl_ata4 &&
+ (priv->treg[dev][0] & TR_66_UDMA_EN)) {
+ void __iomem *rbase = ap->ioaddr.cmd_addr;
+ u32 reg = priv->treg[dev][0];
+
+ if (!(qc->tf.flags & ATA_TFLAG_WRITE))
+ reg += 0x00800000;
+ writel(reg, rbase + IDE_TIMING_CONFIG);
+ }
+
+ /* issue r/w command */
+ ap->ops->sff_exec_command(ap, &qc->tf);
+}
+
+static void pata_macio_bmdma_start(struct ata_queued_cmd *qc)
+{
+ struct ata_port *ap = qc->ap;
+ struct pata_macio_priv *priv = ap->private_data;
+ struct dbdma_regs __iomem *dma_regs = ap->ioaddr.bmdma_addr;
+
+ dev_dbgdma(priv->dev, "%s: qc %p\n", __func__, qc);
+
+ writel((RUN << 16) | RUN, &dma_regs->control);
+ /* Make sure it gets to the controller right now */
+ (void)readl(&dma_regs->control);
+}
+
+static void pata_macio_bmdma_stop(struct ata_queued_cmd *qc)
+{
+ struct ata_port *ap = qc->ap;
+ struct pata_macio_priv *priv = ap->private_data;
+ struct dbdma_regs __iomem *dma_regs = ap->ioaddr.bmdma_addr;
+ unsigned int timeout = 1000000;
+
+ dev_dbgdma(priv->dev, "%s: qc %p\n", __func__, qc);
+
+ /* Stop the DMA engine and wait for it to full halt */
+ writel (((RUN|WAKE|DEAD) << 16), &dma_regs->control);
+ while (--timeout && (readl(&dma_regs->status) & RUN))
+ udelay(1);
+}
+
+static u8 pata_macio_bmdma_status(struct ata_port *ap)
+{
+ struct pata_macio_priv *priv = ap->private_data;
+ struct dbdma_regs __iomem *dma_regs = ap->ioaddr.bmdma_addr;
+ u32 dstat, rstat = ATA_DMA_INTR;
+ unsigned long timeout = 0;
+
+ dstat = readl(&dma_regs->status);
+
+ dev_dbgdma(priv->dev, "%s: dstat=%x\n", __func__, dstat);
+
+ /* We have two things to deal with here:
+ *
+ * - The dbdma won't stop if the command was started
+ * but completed with an error without transferring all
+ * datas. This happens when bad blocks are met during
+ * a multi-block transfer.
+ *
+ * - The dbdma fifo hasn't yet finished flushing to
+ * to system memory when the disk interrupt occurs.
+ *
+ */
+
+ /* First check for errors */
+ if ((dstat & (RUN|DEAD)) != RUN)
+ rstat |= ATA_DMA_ERR;
+
+ /* If ACTIVE is cleared, the STOP command has been hit and
+ * the transfer is complete. If not, we have to flush the
+ * channel.
+ */
+ if ((dstat & ACTIVE) == 0)
+ return rstat;
+
+ dev_dbgdma(priv->dev, "%s: DMA still active, flushing...\n", __func__);
+
+ /* If dbdma didn't execute the STOP command yet, the
+ * active bit is still set. We consider that we aren't
+ * sharing interrupts (which is hopefully the case with
+ * those controllers) and so we just try to flush the
+ * channel for pending data in the fifo
+ */
+ udelay(1);
+ writel((FLUSH << 16) | FLUSH, &dma_regs->control);
+ for (;;) {
+ udelay(1);
+ dstat = readl(&dma_regs->status);
+ if ((dstat & FLUSH) == 0)
+ break;
+ if (++timeout > 1000) {
+ dev_warn(priv->dev, "timeout flushing DMA\n");
+ rstat |= ATA_DMA_ERR;
+ break;
+ }
+ }
+ return rstat;
+}
+
+/* port_start is when we allocate the DMA command list */
+static int pata_macio_port_start(struct ata_port *ap)
+{
+ struct pata_macio_priv *priv = ap->private_data;
+
+ if (ap->ioaddr.bmdma_addr == NULL)
+ return 0;
+
+ /* Allocate space for the DBDMA commands.
+ *
+ * The +2 is +1 for the stop command and +1 to allow for
+ * aligning the start address to a multiple of 16 bytes.
+ */
+ priv->dma_table_cpu =
+ dmam_alloc_coherent(priv->dev,
+ (MAX_DCMDS + 2) * sizeof(struct dbdma_cmd),
+ &priv->dma_table_dma, GFP_KERNEL);
+ if (priv->dma_table_cpu == NULL) {
+ dev_err(priv->dev, "Unable to allocate DMA command list\n");
+ ap->ioaddr.bmdma_addr = NULL;
+ }
+ return 0;
+}
+
+static void pata_macio_irq_clear(struct ata_port *ap)
+{
+ struct pata_macio_priv *priv = ap->private_data;
+
+ /* Nothing to do here */
+
+ dev_dbgdma(priv->dev, "%s\n", __func__);
+}
+
+static void pata_macio_reset_hw(struct pata_macio_priv *priv, int resume)
+{
+ dev_dbg(priv->dev, "Enabling & resetting... \n");
+
+ if (priv->mediabay)
+ return;
+
+ if (priv->kind == controller_ohare && !resume) {
+ /* The code below is having trouble on some ohare machines
+ * (timing related ?). Until I can put my hand on one of these
+ * units, I keep the old way
+ */
+ ppc_md.feature_call(PMAC_FTR_IDE_ENABLE, priv->node, 0, 1);
+ } else {
+ int rc;
+
+ /* Reset and enable controller */
+ rc = ppc_md.feature_call(PMAC_FTR_IDE_RESET,
+ priv->node, priv->aapl_bus_id, 1);
+ ppc_md.feature_call(PMAC_FTR_IDE_ENABLE,
+ priv->node, priv->aapl_bus_id, 1);
+ msleep(10);
+ /* Only bother waiting if there's a reset control */
+ if (rc == 0) {
+ ppc_md.feature_call(PMAC_FTR_IDE_RESET,
+ priv->node, priv->aapl_bus_id, 0);
+ msleep(IDE_WAKEUP_DELAY_MS);
+ }
+ }
+
+ /* If resuming a PCI device, restore the config space here */
+ if (priv->pdev && resume) {
+ int rc;
+
+ pci_restore_state(priv->pdev);
+ rc = pcim_enable_device(priv->pdev);
+ if (rc)
+ dev_printk(KERN_ERR, &priv->pdev->dev,
+ "Failed to enable device after resume (%d)\n", rc);
+ else
+ pci_set_master(priv->pdev);
+ }
+
+ /* On Kauai, initialize the FCR. We don't perform a reset, doesn't really
+ * seem necessary and speeds up the boot process
+ */
+ if (priv->kauai_fcr)
+ writel(KAUAI_FCR_UATA_MAGIC |
+ KAUAI_FCR_UATA_RESET_N |
+ KAUAI_FCR_UATA_ENABLE, priv->kauai_fcr);
+}
+
+/* Hook the standard slave config to fixup some HW related alignment
+ * restrictions
+ */
+static int pata_macio_slave_config(struct scsi_device *sdev)
+{
+ struct ata_port *ap = ata_shost_to_port(sdev->host);
+ struct pata_macio_priv *priv = ap->private_data;
+ struct ata_device *dev;
+ u16 cmd;
+ int rc;
+
+ /* First call original */
+ rc = ata_scsi_slave_config(sdev);
+ if (rc)
+ return rc;
+
+ /* This is lifted from sata_nv */
+ dev = &ap->link.device[sdev->id];
+
+ /* OHare has issues with non cache aligned DMA on some chipsets */
+ if (priv->kind == controller_ohare) {
+ blk_queue_update_dma_alignment(sdev->request_queue, 31);
+ blk_queue_update_dma_pad(sdev->request_queue, 31);
+
+ /* Tell the world about it */
+ ata_dev_printk(dev, KERN_INFO, "OHare alignment limits applied\n");
+ return 0;
+ }
+
+ /* We only have issues with ATAPI */
+ if (dev->class != ATA_DEV_ATAPI)
+ return 0;
+
+ /* Shasta and K2 seem to have "issues" with reads ... */
+ if (priv->kind == controller_sh_ata6 || priv->kind == controller_k2_ata6) {
+ /* Allright these are bad, apply restrictions */
+ blk_queue_update_dma_alignment(sdev->request_queue, 15);
+ blk_queue_update_dma_pad(sdev->request_queue, 15);
+
+ /* We enable MWI and hack cache line size directly here, this
+ * is specific to this chipset and not normal values, we happen
+ * to somewhat know what we are doing here (which is basically
+ * to do the same Apple does and pray they did not get it wrong :-)
+ */
+ BUG_ON(!priv->pdev);
+ pci_write_config_byte(priv->pdev, PCI_CACHE_LINE_SIZE, 0x08);
+ pci_read_config_word(priv->pdev, PCI_COMMAND, &cmd);
+ pci_write_config_word(priv->pdev, PCI_COMMAND,
+ cmd | PCI_COMMAND_INVALIDATE);
+
+ /* Tell the world about it */
+ ata_dev_printk(dev, KERN_INFO,
+ "K2/Shasta alignment limits applied\n");
+ }
+
+ return 0;
+}
+
+#ifdef CONFIG_PM
+
+static int pata_macio_do_suspend(struct pata_macio_priv *priv, pm_message_t mesg)
+{
+ int rc;
+
+ /* First, core libata suspend to do most of the work */
+ rc = ata_host_suspend(priv->host, mesg);
+ if (rc)
+ return rc;
+
+ /* Restore to default timings */
+ pata_macio_default_timings(priv);
+
+ /* Mask interrupt. Not strictly necessary but old driver did
+ * it and I'd rather not change that here */
+ disable_irq(priv->irq);
+
+ /* The media bay will handle itself just fine */
+ if (priv->mediabay)
+ return 0;
+
+ /* Kauai has bus control FCRs directly here */
+ if (priv->kauai_fcr) {
+ u32 fcr = readl(priv->kauai_fcr);
+ fcr &= ~(KAUAI_FCR_UATA_RESET_N | KAUAI_FCR_UATA_ENABLE);
+ writel(fcr, priv->kauai_fcr);
+ }
+
+ /* For PCI, save state and disable DMA. No need to call
+ * pci_set_power_state(), the HW doesn't do D states that
+ * way, the platform code will take care of suspending the
+ * ASIC properly
+ */
+ if (priv->pdev) {
+ pci_save_state(priv->pdev);
+ pci_disable_device(priv->pdev);
+ }
+
+ /* Disable the bus on older machines and the cell on kauai */
+ ppc_md.feature_call(PMAC_FTR_IDE_ENABLE, priv->node,
+ priv->aapl_bus_id, 0);
+
+ return 0;
+}
+
+static int pata_macio_do_resume(struct pata_macio_priv *priv)
+{
+ /* Reset and re-enable the HW */
+ pata_macio_reset_hw(priv, 1);
+
+ /* Sanitize drive timings */
+ pata_macio_apply_timings(priv->host->ports[0], 0);
+
+ /* We want our IRQ back ! */
+ enable_irq(priv->irq);
+
+ /* Let the libata core take it from there */
+ ata_host_resume(priv->host);
+
+ return 0;
+}
+
+#endif /* CONFIG_PM */
+
+static struct scsi_host_template pata_macio_sht = {
+ ATA_BASE_SHT(DRV_NAME),
+ .sg_tablesize = MAX_DCMDS,
+ /* We may not need that strict one */
+ .dma_boundary = ATA_DMA_BOUNDARY,
+ .slave_configure = pata_macio_slave_config,
+};
+
+static struct ata_port_operations pata_macio_ops = {
+ .inherits = &ata_sff_port_ops,
+
+ .freeze = pata_macio_freeze,
+ .set_piomode = pata_macio_set_timings,
+ .set_dmamode = pata_macio_set_timings,
+ .cable_detect = pata_macio_cable_detect,
+ .sff_dev_select = pata_macio_dev_select,
+ .qc_prep = pata_macio_qc_prep,
+ .mode_filter = ata_bmdma_mode_filter,
+ .bmdma_setup = pata_macio_bmdma_setup,
+ .bmdma_start = pata_macio_bmdma_start,
+ .bmdma_stop = pata_macio_bmdma_stop,
+ .bmdma_status = pata_macio_bmdma_status,
+ .port_start = pata_macio_port_start,
+ .sff_irq_clear = pata_macio_irq_clear,
+};
+
+static void __devinit pata_macio_invariants(struct pata_macio_priv *priv)
+{
+ const int *bidp;
+
+ /* Identify the type of controller */
+ if (of_device_is_compatible(priv->node, "shasta-ata")) {
+ priv->kind = controller_sh_ata6;
+ priv->timings = pata_macio_shasta_timings;
+ } else if (of_device_is_compatible(priv->node, "kauai-ata")) {
+ priv->kind = controller_un_ata6;
+ priv->timings = pata_macio_kauai_timings;
+ } else if (of_device_is_compatible(priv->node, "K2-UATA")) {
+ priv->kind = controller_k2_ata6;
+ priv->timings = pata_macio_kauai_timings;
+ } else if (of_device_is_compatible(priv->node, "keylargo-ata")) {
+ if (strcmp(priv->node->name, "ata-4") == 0) {
+ priv->kind = controller_kl_ata4;
+ priv->timings = pata_macio_kl66_timings;
+ } else {
+ priv->kind = controller_kl_ata3;
+ priv->timings = pata_macio_kl33_timings;
+ }
+ } else if (of_device_is_compatible(priv->node, "heathrow-ata")) {
+ priv->kind = controller_heathrow;
+ priv->timings = pata_macio_heathrow_timings;
+ } else {
+ priv->kind = controller_ohare;
+ priv->timings = pata_macio_ohare_timings;
+ }
+
+ /* XXX FIXME --- setup priv->mediabay here */
+
+ /* Get Apple bus ID (for clock and ASIC control) */
+ bidp = of_get_property(priv->node, "AAPL,bus-id", NULL);
+ priv->aapl_bus_id = bidp ? *bidp : 0;
+
+ /* Fixup missing Apple bus ID in case of media-bay */
+ if (priv->mediabay && bidp == 0)
+ priv->aapl_bus_id = 1;
+}
+
+static void __devinit pata_macio_setup_ios(struct ata_ioports *ioaddr,
+ void __iomem * base,
+ void __iomem * dma)
+{
+ /* cmd_addr is the base of regs for that port */
+ ioaddr->cmd_addr = base;
+
+ /* taskfile registers */
+ ioaddr->data_addr = base + (ATA_REG_DATA << 4);
+ ioaddr->error_addr = base + (ATA_REG_ERR << 4);
+ ioaddr->feature_addr = base + (ATA_REG_FEATURE << 4);
+ ioaddr->nsect_addr = base + (ATA_REG_NSECT << 4);
+ ioaddr->lbal_addr = base + (ATA_REG_LBAL << 4);
+ ioaddr->lbam_addr = base + (ATA_REG_LBAM << 4);
+ ioaddr->lbah_addr = base + (ATA_REG_LBAH << 4);
+ ioaddr->device_addr = base + (ATA_REG_DEVICE << 4);
+ ioaddr->status_addr = base + (ATA_REG_STATUS << 4);
+ ioaddr->command_addr = base + (ATA_REG_CMD << 4);
+ ioaddr->altstatus_addr = base + 0x160;
+ ioaddr->ctl_addr = base + 0x160;
+ ioaddr->bmdma_addr = dma;
+}
+
+static void __devinit pmac_macio_calc_timing_masks(struct pata_macio_priv *priv,
+ struct ata_port_info *pinfo)
+{
+ int i = 0;
+
+ pinfo->pio_mask = 0;
+ pinfo->mwdma_mask = 0;
+ pinfo->udma_mask = 0;
+
+ while (priv->timings[i].mode > 0) {
+ unsigned int mask = 1U << (priv->timings[i].mode & 0x0f);
+ switch(priv->timings[i].mode & 0xf0) {
+ case 0x00: /* PIO */
+ pinfo->pio_mask |= (mask >> 8);
+ break;
+ case 0x20: /* MWDMA */
+ pinfo->mwdma_mask |= mask;
+ break;
+ case 0x40: /* UDMA */
+ pinfo->udma_mask |= mask;
+ break;
+ }
+ i++;
+ }
+ dev_dbg(priv->dev, "Supported masks: PIO=%lx, MWDMA=%lx, UDMA=%lx\n",
+ pinfo->pio_mask, pinfo->mwdma_mask, pinfo->udma_mask);
+}
+
+static int __devinit pata_macio_common_init(struct pata_macio_priv *priv,
+ resource_size_t tfregs,
+ resource_size_t dmaregs,
+ resource_size_t fcregs,
+ unsigned long irq)
+{
+ struct ata_port_info pinfo;
+ const struct ata_port_info *ppi[] = { &pinfo, NULL };
+ void __iomem *dma_regs = NULL;
+
+ /* Fill up privates with various invariants collected from the
+ * device-tree
+ */
+ pata_macio_invariants(priv);
+
+ /* Make sure we have sane initial timings in the cache */
+ pata_macio_default_timings(priv);
+
+ /* Not sure what the real max is but we know it's less than 64K, let's
+ * use 64K minus 256
+ */
+ dma_set_max_seg_size(priv->dev, MAX_DBDMA_SEG);
+
+ /* Allocate libata host for 1 port */
+ memset(&pinfo, 0, sizeof(struct ata_port_info));
+ pmac_macio_calc_timing_masks(priv, &pinfo);
+ pinfo.flags = ATA_FLAG_SLAVE_POSS | ATA_FLAG_MMIO |
+ ATA_FLAG_NO_LEGACY;
+ pinfo.port_ops = &pata_macio_ops;
+ pinfo.private_data = priv;
+
+ priv->host = ata_host_alloc_pinfo(priv->dev, ppi, 1);
+ if (priv->host == NULL) {
+ dev_err(priv->dev, "Failed to allocate ATA port structure\n");
+ return -ENOMEM;
+ }
+
+ /* Setup the private data in host too */
+ priv->host->private_data = priv;
+
+ /* Map base registers */
+ priv->tfregs = devm_ioremap(priv->dev, tfregs, 0x100);
+ if (priv->tfregs == NULL) {
+ dev_err(priv->dev, "Failed to map ATA ports\n");
+ return -ENOMEM;
+ }
+ priv->host->iomap = &priv->tfregs;
+
+ /* Map DMA regs */
+ if (dmaregs != 0) {
+ dma_regs = devm_ioremap(priv->dev, dmaregs,
+ sizeof(struct dbdma_regs));
+ if (dma_regs == NULL)
+ dev_warn(priv->dev, "Failed to map ATA DMA registers\n");
+ }
+
+ /* If chip has local feature control, map those regs too */
+ if (fcregs != 0) {
+ priv->kauai_fcr = devm_ioremap(priv->dev, fcregs, 4);
+ if (priv->kauai_fcr == NULL) {
+ dev_err(priv->dev, "Failed to map ATA FCR register\n");
+ return -ENOMEM;
+ }
+ }
+
+ /* Setup port data structure */
+ pata_macio_setup_ios(&priv->host->ports[0]->ioaddr,
+ priv->tfregs, dma_regs);
+ priv->host->ports[0]->private_data = priv;
+
+ /* hard-reset the controller */
+ pata_macio_reset_hw(priv, 0);
+ pata_macio_apply_timings(priv->host->ports[0], 0);
+
+ /* Enable bus master if necessary */
+ if (priv->pdev && dma_regs)
+ pci_set_master(priv->pdev);
+
+ dev_info(priv->dev, "Activating pata-macio chipset %s, Apple bus ID %d\n",
+ macio_ata_names[priv->kind], priv->aapl_bus_id);
+
+ /* Start it up */
+ priv->irq = irq;
+ return ata_host_activate(priv->host, irq, ata_sff_interrupt, 0,
+ &pata_macio_sht);
+}
+
+static int __devinit pata_macio_attach(struct macio_dev *mdev,
+ const struct of_device_id *match)
+{
+ struct pata_macio_priv *priv;
+ resource_size_t tfregs, dmaregs = 0;
+ unsigned long irq;
+ int rc;
+
+ /* Check for broken device-trees */
+ if (macio_resource_count(mdev) == 0) {
+ dev_err(&mdev->ofdev.dev,
+ "No addresses for controller\n");
+ return -ENXIO;
+ }
+
+ /* Enable managed resources */
+ macio_enable_devres(mdev);
+
+ /* Allocate and init private data structure */
+ priv = devm_kzalloc(&mdev->ofdev.dev,
+ sizeof(struct pata_macio_priv), GFP_KERNEL);
+ if (priv == NULL) {
+ dev_err(&mdev->ofdev.dev,
+ "Failed to allocate private memory\n");
+ return -ENOMEM;
+ }
+ priv->node = of_node_get(mdev->ofdev.node);
+ priv->mdev = mdev;
+ priv->dev = &mdev->ofdev.dev;
+
+ /* Request memory resource for taskfile registers */
+ if (macio_request_resource(mdev, 0, "pata-macio")) {
+ dev_err(&mdev->ofdev.dev,
+ "Cannot obtain taskfile resource\n");
+ return -EBUSY;
+ }
+ tfregs = macio_resource_start(mdev, 0);
+
+ /* Request resources for DMA registers if any */
+ if (macio_resource_count(mdev) >= 2) {
+ if (macio_request_resource(mdev, 1, "pata-macio-dma"))
+ dev_err(&mdev->ofdev.dev,
+ "Cannot obtain DMA resource\n");
+ else
+ dmaregs = macio_resource_start(mdev, 1);
+ }
+
+ /*
+ * Fixup missing IRQ for some old implementations with broken
+ * device-trees.
+ *
+ * This is a bit bogus, it should be fixed in the device-tree itself,
+ * via the existing macio fixups, based on the type of interrupt
+ * controller in the machine. However, I have no test HW for this case,
+ * and this trick works well enough on those old machines...
+ */
+ if (macio_irq_count(mdev) == 0) {
+ dev_warn(&mdev->ofdev.dev,
+ "No interrupts for controller, using 13\n");
+ irq = irq_create_mapping(NULL, 13);
+ } else
+ irq = macio_irq(mdev, 0);
+
+ /* Prevvent media bay callbacks until fully registered */
+ lock_media_bay(priv->mdev->media_bay);
+
+ /* Get register addresses and call common initialization */
+ rc = pata_macio_common_init(priv,
+ tfregs, /* Taskfile regs */
+ dmaregs, /* DBDMA regs */
+ 0, /* Feature control */
+ irq);
+ unlock_media_bay(priv->mdev->media_bay);
+
+ return rc;
+}
+
+static int __devexit pata_macio_detach(struct macio_dev *mdev)
+{
+ struct ata_host *host = macio_get_drvdata(mdev);
+ struct pata_macio_priv *priv = host->private_data;
+
+ lock_media_bay(priv->mdev->media_bay);
+
+ /* Make sure the mediabay callback doesn't try to access
+ * dead stuff
+ */
+ priv->host->private_data = NULL;
+
+ ata_host_detach(host);
+
+ unlock_media_bay(priv->mdev->media_bay);
+
+ return 0;
+}
+
+#ifdef CONFIG_PM
+
+static int pata_macio_suspend(struct macio_dev *mdev, pm_message_t mesg)
+{
+ struct ata_host *host = macio_get_drvdata(mdev);
+
+ return pata_macio_do_suspend(host->private_data, mesg);
+}
+
+static int pata_macio_resume(struct macio_dev *mdev)
+{
+ struct ata_host *host = macio_get_drvdata(mdev);
+
+ return pata_macio_do_resume(host->private_data);
+}
+
+#endif /* CONFIG_PM */
+
+#ifdef CONFIG_PMAC_MEDIABAY
+static void pata_macio_mb_event(struct macio_dev* mdev, int mb_state)
+{
+ struct ata_host *host = macio_get_drvdata(mdev);
+ struct ata_port *ap;
+ struct ata_eh_info *ehi;
+ struct ata_device *dev;
+ unsigned long flags;
+
+ if (!host || !host->private_data)
+ return;
+ ap = host->ports[0];
+ spin_lock_irqsave(ap->lock, flags);
+ ehi = &ap->link.eh_info;
+ if (mb_state == MB_CD) {
+ ata_ehi_push_desc(ehi, "mediabay plug");
+ ata_ehi_hotplugged(ehi);
+ ata_port_freeze(ap);
+ } else {
+ ata_ehi_push_desc(ehi, "mediabay unplug");
+ ata_for_each_dev(dev, &ap->link, ALL)
+ dev->flags |= ATA_DFLAG_DETACH;
+ ata_port_abort(ap);
+ }
+ spin_unlock_irqrestore(ap->lock, flags);
+
+}
+#endif /* CONFIG_PMAC_MEDIABAY */
+
+
+static int __devinit pata_macio_pci_attach(struct pci_dev *pdev,
+ const struct pci_device_id *id)
+{
+ struct pata_macio_priv *priv;
+ struct device_node *np;
+ resource_size_t rbase;
+
+ /* We cannot use a MacIO controller without its OF device node */
+ np = pci_device_to_OF_node(pdev);
+ if (np == NULL) {
+ dev_err(&pdev->dev,
+ "Cannot find OF device node for controller\n");
+ return -ENODEV;
+ }
+
+ /* Check that it can be enabled */
+ if (pcim_enable_device(pdev)) {
+ dev_err(&pdev->dev,
+ "Cannot enable controller PCI device\n");
+ return -ENXIO;
+ }
+
+ /* Allocate and init private data structure */
+ priv = devm_kzalloc(&pdev->dev,
+ sizeof(struct pata_macio_priv), GFP_KERNEL);
+ if (priv == NULL) {
+ dev_err(&pdev->dev,
+ "Failed to allocate private memory\n");
+ return -ENOMEM;
+ }
+ priv->node = of_node_get(np);
+ priv->pdev = pdev;
+ priv->dev = &pdev->dev;
+
+ /* Get MMIO regions */
+ if (pci_request_regions(pdev, "pata-macio")) {
+ dev_err(&pdev->dev,
+ "Cannot obtain PCI resources\n");
+ return -EBUSY;
+ }
+
+ /* Get register addresses and call common initialization */
+ rbase = pci_resource_start(pdev, 0);
+ if (pata_macio_common_init(priv,
+ rbase + 0x2000, /* Taskfile regs */
+ rbase + 0x1000, /* DBDMA regs */
+ rbase, /* Feature control */
+ pdev->irq))
+ return -ENXIO;
+
+ return 0;
+}
+
+static void __devexit pata_macio_pci_detach(struct pci_dev *pdev)
+{
+ struct ata_host *host = dev_get_drvdata(&pdev->dev);
+
+ ata_host_detach(host);
+}
+
+#ifdef CONFIG_PM
+
+static int pata_macio_pci_suspend(struct pci_dev *pdev, pm_message_t mesg)
+{
+ struct ata_host *host = dev_get_drvdata(&pdev->dev);
+
+ return pata_macio_do_suspend(host->private_data, mesg);
+}
+
+static int pata_macio_pci_resume(struct pci_dev *pdev)
+{
+ struct ata_host *host = dev_get_drvdata(&pdev->dev);
+
+ return pata_macio_do_resume(host->private_data);
+}
+
+#endif /* CONFIG_PM */
+
+static struct of_device_id pata_macio_match[] =
+{
+ {
+ .name = "IDE",
+ },
+ {
+ .name = "ATA",
+ },
+ {
+ .type = "ide",
+ },
+ {
+ .type = "ata",
+ },
+ {},
+};
+
+static struct macio_driver pata_macio_driver =
+{
+ .name = "pata-macio",
+ .match_table = pata_macio_match,
+ .probe = pata_macio_attach,
+ .remove = pata_macio_detach,
+#ifdef CONFIG_PM
+ .suspend = pata_macio_suspend,
+ .resume = pata_macio_resume,
+#endif
+#ifdef CONFIG_PMAC_MEDIABAY
+ .mediabay_event = pata_macio_mb_event,
+#endif
+ .driver = {
+ .owner = THIS_MODULE,
+ },
+};
+
+static const struct pci_device_id pata_macio_pci_match[] = {
+ { PCI_VDEVICE(APPLE, PCI_DEVICE_ID_APPLE_UNI_N_ATA), 0 },
+ { PCI_VDEVICE(APPLE, PCI_DEVICE_ID_APPLE_IPID_ATA100), 0 },
+ { PCI_VDEVICE(APPLE, PCI_DEVICE_ID_APPLE_K2_ATA100), 0 },
+ { PCI_VDEVICE(APPLE, PCI_DEVICE_ID_APPLE_SH_ATA), 0 },
+ { PCI_VDEVICE(APPLE, PCI_DEVICE_ID_APPLE_IPID2_ATA), 0 },
+ {},
+};
+
+static struct pci_driver pata_macio_pci_driver = {
+ .name = "pata-pci-macio",
+ .id_table = pata_macio_pci_match,
+ .probe = pata_macio_pci_attach,
+ .remove = pata_macio_pci_detach,
+#ifdef CONFIG_PM
+ .suspend = pata_macio_pci_suspend,
+ .resume = pata_macio_pci_resume,
+#endif
+ .driver = {
+ .owner = THIS_MODULE,
+ },
+};
+MODULE_DEVICE_TABLE(pci, pata_macio_pci_match);
+
+
+static int __init pata_macio_init(void)
+{
+ int rc;
+
+ if (!machine_is(powermac))
+ return -ENODEV;
+
+ rc = pci_register_driver(&pata_macio_pci_driver);
+ if (rc)
+ return rc;
+ rc = macio_register_driver(&pata_macio_driver);
+ if (rc) {
+ pci_unregister_driver(&pata_macio_pci_driver);
+ return rc;
+ }
+ return 0;
+}
+
+static void __exit pata_macio_exit(void)
+{
+ macio_unregister_driver(&pata_macio_driver);
+ pci_unregister_driver(&pata_macio_pci_driver);
+}
+
+module_init(pata_macio_init);
+module_exit(pata_macio_exit);
+
+MODULE_AUTHOR("Benjamin Herrenschmidt");
+MODULE_DESCRIPTION("Apple MacIO PATA driver");
+MODULE_LICENSE("GPL");
+MODULE_VERSION(DRV_VERSION);
Index: linux-work/drivers/ata/Kconfig
===================================================================
--- linux-work.orig/drivers/ata/Kconfig 2009-12-02 11:24:34.000000000 +1100
+++ linux-work/drivers/ata/Kconfig 2009-12-02 11:26:51.000000000 +1100
@@ -781,5 +781,15 @@ config PATA_BF54X
If unsure, say N.
+config PATA_MACIO
+ tristate "Apple PowerMac/PowerBook internal 'MacIO' IDE"
+ depends on PPC_PMAC
+ help
+ Most IDE capable PowerMacs have IDE busses driven by a variant
+ of this controller which is part of the Apple chipset used on
+ most PowerMac models. Some models have multiple busses using
+ different chipsets, though generally, MacIO is one of them.
+
+
endif # ATA_SFF
endif # ATA
Index: linux-work/drivers/ata/Makefile
===================================================================
--- linux-work.orig/drivers/ata/Makefile 2009-12-02 11:24:34.000000000 +1100
+++ linux-work/drivers/ata/Makefile 2009-12-02 11:26:51.000000000 +1100
@@ -18,6 +18,7 @@ obj-$(CONFIG_SATA_MV) += sata_mv.o
obj-$(CONFIG_SATA_INIC162X) += sata_inic162x.o
obj-$(CONFIG_PDC_ADMA) += pdc_adma.o
obj-$(CONFIG_SATA_FSL) += sata_fsl.o
+obj-$(CONFIG_PATA_MACIO) += pata_macio.o
obj-$(CONFIG_PATA_ALI) += pata_ali.o
obj-$(CONFIG_PATA_AMD) += pata_amd.o
^ permalink raw reply
* [PATCH 4/5] libata/sff: Use ops->bmdma_stop instead of ata_bmdma_stop()
From: Benjamin Herrenschmidt @ 2009-12-02 0:36 UTC (permalink / raw)
To: linux-ide; +Cc: tj, linuxppc-dev, jeff
In libata-sff, ata_sff_post_internal_cmd() directly calls ata_bmdma_stop()
instead of ap->ops->bmdma_stop(). This can be a problem for controllers
that use their own bmdma_stop for which the generic sff one isn't suitable
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
drivers/ata/libata-sff.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- linux-work.orig/drivers/ata/libata-sff.c 2009-12-01 17:48:27.000000000 +1100
+++ linux-work/drivers/ata/libata-sff.c 2009-12-01 17:48:48.000000000 +1100
@@ -2384,7 +2384,7 @@ void ata_sff_post_internal_cmd(struct at
ap->hsm_task_state = HSM_ST_IDLE;
if (ap->ioaddr.bmdma_addr)
- ata_bmdma_stop(qc);
+ ap->ops->bmdma_stop(qc);
spin_unlock_irqrestore(ap->lock, flags);
}
^ permalink raw reply
* [PATCH 3/5] powerpc/macio: Rework hotplug media bay support
From: Benjamin Herrenschmidt @ 2009-12-02 0:36 UTC (permalink / raw)
To: linux-ide; +Cc: tj, linuxppc-dev, jeff
The hotplug mediabay has tendrils deep into drivers/ide code
which makes a libata port reather difficult. In addition it's
ugly and could be done better.
This reworks the interface between the mediabay and the rest
of the world so that:
- Any macio_driver can now have a mediabay_event callback
which will be called when that driver sits on a mediabay and
it's been either plugged or unplugged. The device type is
passed as an argument. We can now move all the IDE cruft
into the IDE driver itself
- A check_media_bay() function can be used to take a peek
at the type of device currently in the bay if any, a cleaner
variant of the previous function with the same name.
- A pair of lock/unlock functions are exposed to allow the
IDE driver to block the hotplug callbacks during the initial
setup and probing of the bay in order to avoid nasty race
conditions.
- The mediabay code no longer needs to spin on the status
register of the IDE interface when it detects an IDE device,
this is done just fine by the IDE code itself
Overall, less code, simpler, and allows for another driver
than our old drivers/ide based one.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
v2. Fix compile on 64-bit
arch/powerpc/include/asm/macio.h | 3
arch/powerpc/include/asm/mediabay.h | 29 +--
drivers/block/swim3.c | 39 +---
drivers/ide/pmac.c | 90 ++++++---
drivers/macintosh/mediabay.c | 328 +++++++++++++-----------------------
5 files changed, 213 insertions(+), 276 deletions(-)
--- linux-work.orig/arch/powerpc/include/asm/macio.h 2009-12-01 17:57:00.000000000 +1100
+++ linux-work/arch/powerpc/include/asm/macio.h 2009-12-01 18:00:28.000000000 +1100
@@ -134,6 +134,9 @@ struct macio_driver
int (*resume)(struct macio_dev* dev);
int (*shutdown)(struct macio_dev* dev);
+#ifdef CONFIG_PMAC_MEDIABAY
+ void (*mediabay_event)(struct macio_dev* dev, int mb_state);
+#endif
struct device_driver driver;
};
#define to_macio_driver(drv) container_of(drv,struct macio_driver, driver)
Index: linux-work/arch/powerpc/include/asm/mediabay.h
===================================================================
--- linux-work.orig/arch/powerpc/include/asm/mediabay.h 2009-12-01 17:57:00.000000000 +1100
+++ linux-work/arch/powerpc/include/asm/mediabay.h 2009-12-01 18:00:28.000000000 +1100
@@ -17,26 +17,31 @@
#define MB_POWER 6 /* media bay contains a Power device (???) */
#define MB_NO 7 /* media bay contains nothing */
-/* Number of bays in the machine or 0 */
-extern int media_bay_count;
+struct macio_dev;
-#ifdef CONFIG_BLK_DEV_IDE_PMAC
-#include <linux/ide.h>
+#ifdef CONFIG_PMAC_MEDIABAY
-int check_media_bay_by_base(unsigned long base, int what);
-/* called by IDE PMAC host driver to register IDE controller for media bay */
-int media_bay_set_ide_infos(struct device_node *which_bay, unsigned long base,
- int irq, ide_hwif_t *hwif);
-
-int check_media_bay(struct device_node *which_bay, int what);
+/* Check the content type of the bay, returns MB_NO if the bay is still
+ * transitionning
+ */
+extern int check_media_bay(struct macio_dev *bay);
+
+/* The ATA driver uses the calls below to temporarily hold on the
+ * media bay callbacks while initializing the interface
+ */
+extern void lock_media_bay(struct macio_dev *bay);
+extern void unlock_media_bay(struct macio_dev *bay);
#else
-static inline int check_media_bay(struct device_node *which_bay, int what)
+static inline int check_media_bay(struct macio_dev *bay)
{
- return -ENODEV;
+ return MB_NO;
}
+static inline void lock_media_bay(struct macio_dev *bay) { }
+static inline void unlock_media_bay(struct macio_dev *bay) { }
+
#endif
#endif /* __KERNEL__ */
Index: linux-work/drivers/macintosh/mediabay.c
===================================================================
--- linux-work.orig/drivers/macintosh/mediabay.c 2009-12-01 17:57:00.000000000 +1100
+++ linux-work/drivers/macintosh/mediabay.c 2009-12-01 18:00:28.000000000 +1100
@@ -33,15 +33,6 @@
#include <linux/adb.h>
#include <linux/pmu.h>
-
-#define MB_DEBUG
-
-#ifdef MB_DEBUG
-#define MBDBG(fmt, arg...) printk(KERN_INFO fmt , ## arg)
-#else
-#define MBDBG(fmt, arg...) do { } while (0)
-#endif
-
#define MB_FCR32(bay, r) ((bay)->base + ((r) >> 2))
#define MB_FCR8(bay, r) (((volatile u8 __iomem *)((bay)->base)) + (r))
@@ -76,28 +67,14 @@ struct media_bay_info {
int index;
int cached_gpio;
int sleeping;
+ int user_lock;
struct mutex lock;
-#ifdef CONFIG_BLK_DEV_IDE_PMAC
- ide_hwif_t *cd_port;
- void __iomem *cd_base;
- int cd_irq;
- int cd_retry;
-#endif
-#if defined(CONFIG_BLK_DEV_IDE_PMAC)
- int cd_index;
-#endif
};
#define MAX_BAYS 2
static struct media_bay_info media_bays[MAX_BAYS];
-int media_bay_count = 0;
-
-#ifdef CONFIG_BLK_DEV_IDE_PMAC
-/* check the busy bit in the media-bay ide interface
- (assumes the media-bay contains an ide device) */
-#define MB_IDE_READY(i) ((readb(media_bays[i].cd_base + 0x70) & 0x80) == 0)
-#endif
+static int media_bay_count = 0;
/*
* Wait that number of ms between each step in normal polling mode
@@ -130,21 +107,11 @@ int media_bay_count = 0;
/*
* Wait this many ticks after an IDE device (e.g. CD-ROM) is inserted
- * (or until the device is ready) before waiting for busy bit to disappear
+ * (or until the device is ready) before calling into the driver
*/
#define MB_IDE_WAIT 1000
/*
- * Timeout waiting for busy bit of an IDE device to go down
- */
-#define MB_IDE_TIMEOUT 5000
-
-/*
- * Max retries of the full power up/down sequence for an IDE device
- */
-#define MAX_CD_RETRIES 3
-
-/*
* States of a media bay
*/
enum {
@@ -153,7 +120,6 @@ enum {
mb_enabling_bay, /* enable bits set, waiting MB_RESET_DELAY */
mb_resetting, /* reset bit unset, waiting MB_SETUP_DELAY */
mb_ide_resetting, /* IDE reset bit unser, waiting MB_IDE_WAIT */
- mb_ide_waiting, /* Waiting for BUSY bit to go away until MB_IDE_TIMEOUT */
mb_up, /* Media bay full */
mb_powering_down /* Powering down (avoid too fast down/up) */
};
@@ -373,12 +339,12 @@ static inline void set_mb_power(struct m
if (onoff) {
bay->ops->power(bay, 1);
bay->state = mb_powering_up;
- MBDBG("mediabay%d: powering up\n", bay->index);
+ pr_debug("mediabay%d: powering up\n", bay->index);
} else {
/* Make sure everything is powered down & disabled */
bay->ops->power(bay, 0);
bay->state = mb_powering_down;
- MBDBG("mediabay%d: powering down\n", bay->index);
+ pr_debug("mediabay%d: powering down\n", bay->index);
}
bay->timer = msecs_to_jiffies(MB_POWER_DELAY);
}
@@ -387,107 +353,118 @@ static void poll_media_bay(struct media_
{
int id = bay->ops->content(bay);
- if (id == bay->last_value) {
- if (id != bay->content_id) {
- bay->value_count += msecs_to_jiffies(MB_POLL_DELAY);
- if (bay->value_count >= msecs_to_jiffies(MB_STABLE_DELAY)) {
- /* If the device type changes without going thru
- * "MB_NO", we force a pass by "MB_NO" to make sure
- * things are properly reset
- */
- if ((id != MB_NO) && (bay->content_id != MB_NO)) {
- id = MB_NO;
- MBDBG("mediabay%d: forcing MB_NO\n", bay->index);
- }
- MBDBG("mediabay%d: switching to %d\n", bay->index, id);
- set_mb_power(bay, id != MB_NO);
- bay->content_id = id;
- if (id == MB_NO) {
-#ifdef CONFIG_BLK_DEV_IDE_PMAC
- bay->cd_retry = 0;
-#endif
- printk(KERN_INFO "media bay %d is empty\n", bay->index);
- }
- }
- }
- } else {
+ static char *mb_content_types[] = {
+ "a floppy drive",
+ "a floppy drive",
+ "an unsuported audio device",
+ "an ATA device",
+ "an unsupported PCI device",
+ "an unknown device",
+ };
+
+ if (id != bay->last_value) {
bay->last_value = id;
bay->value_count = 0;
+ return;
}
-}
+ if (id == bay->content_id)
+ return;
-#ifdef CONFIG_BLK_DEV_IDE_PMAC
-int check_media_bay(struct device_node *which_bay, int what)
-{
- int i;
-
- for (i=0; i<media_bay_count; i++)
- if (media_bays[i].mdev && which_bay == media_bays[i].mdev->ofdev.node) {
- if ((what == media_bays[i].content_id) && media_bays[i].state == mb_up)
- return 0;
- media_bays[i].cd_index = -1;
- return -EINVAL;
+ bay->value_count += msecs_to_jiffies(MB_POLL_DELAY);
+ if (bay->value_count >= msecs_to_jiffies(MB_STABLE_DELAY)) {
+ /* If the device type changes without going thru
+ * "MB_NO", we force a pass by "MB_NO" to make sure
+ * things are properly reset
+ */
+ if ((id != MB_NO) && (bay->content_id != MB_NO)) {
+ id = MB_NO;
+ pr_debug("mediabay%d: forcing MB_NO\n", bay->index);
}
- return -ENODEV;
+ pr_debug("mediabay%d: switching to %d\n", bay->index, id);
+ set_mb_power(bay, id != MB_NO);
+ bay->content_id = id;
+ if (id >= MB_NO || id < 0)
+ printk(KERN_INFO "mediabay%d: Bay is now empty\n", bay->index);
+ else
+ printk(KERN_INFO "mediabay%d: Bay contains %s\n",
+ bay->index, mb_content_types[id]);
+ }
}
-EXPORT_SYMBOL(check_media_bay);
-int check_media_bay_by_base(unsigned long base, int what)
+int check_media_bay(struct macio_dev *baydev)
{
- int i;
+ struct media_bay_info* bay;
+ int id;
- for (i=0; i<media_bay_count; i++)
- if (media_bays[i].mdev && base == (unsigned long) media_bays[i].cd_base) {
- if ((what == media_bays[i].content_id) && media_bays[i].state == mb_up)
- return 0;
- media_bays[i].cd_index = -1;
- return -EINVAL;
- }
+ if (baydev == NULL)
+ return MB_NO;
- return -ENODEV;
+ /* This returns an instant snapshot, not locking, sine
+ * we may be called with the bay lock held. The resulting
+ * fuzzyness of the result if called at the wrong time is
+ * not actually a huge deal
+ */
+ bay = macio_get_drvdata(baydev);
+ if (bay == NULL)
+ return MB_NO;
+ id = bay->content_id;
+ if (bay->state != mb_up)
+ return MB_NO;
+ if (id == MB_FD1)
+ return MB_FD;
+ return id;
}
-EXPORT_SYMBOL_GPL(check_media_bay_by_base);
+EXPORT_SYMBOL_GPL(check_media_bay);
-int media_bay_set_ide_infos(struct device_node* which_bay, unsigned long base,
- int irq, ide_hwif_t *hwif)
+void lock_media_bay(struct macio_dev *baydev)
{
- int i;
+ struct media_bay_info* bay;
- for (i=0; i<media_bay_count; i++) {
- struct media_bay_info* bay = &media_bays[i];
+ if (baydev == NULL)
+ return;
+ bay = macio_get_drvdata(baydev);
+ if (bay == NULL)
+ return;
+ mutex_lock(&bay->lock);
+ bay->user_lock = 1;
+}
+EXPORT_SYMBOL_GPL(lock_media_bay);
- if (bay->mdev && which_bay == bay->mdev->ofdev.node) {
- int timeout = 5000, index = hwif->index;
-
- mutex_lock(&bay->lock);
-
- bay->cd_port = hwif;
- bay->cd_base = (void __iomem *) base;
- bay->cd_irq = irq;
-
- if ((MB_CD != bay->content_id) || bay->state != mb_up) {
- mutex_unlock(&bay->lock);
- return 0;
- }
- printk(KERN_DEBUG "Registered ide%d for media bay %d\n", index, i);
- do {
- if (MB_IDE_READY(i)) {
- bay->cd_index = index;
- mutex_unlock(&bay->lock);
- return 0;
- }
- mdelay(1);
- } while(--timeout);
- printk(KERN_DEBUG "Timeount waiting IDE in bay %d\n", i);
- mutex_unlock(&bay->lock);
- return -ENODEV;
- }
+void unlock_media_bay(struct macio_dev *baydev)
+{
+ struct media_bay_info* bay;
+
+ if (baydev == NULL)
+ return;
+ bay = macio_get_drvdata(baydev);
+ if (bay == NULL)
+ return;
+ if (bay->user_lock) {
+ bay->user_lock = 0;
+ mutex_unlock(&bay->lock);
}
+}
+EXPORT_SYMBOL_GPL(unlock_media_bay);
- return -ENODEV;
+static int mb_broadcast_hotplug(struct device *dev, void *data)
+{
+ struct media_bay_info* bay = data;
+ struct macio_dev *mdev;
+ struct macio_driver *drv;
+ int state;
+
+ if (dev->bus != &macio_bus_type)
+ return 0;
+
+ state = bay->state == mb_up ? bay->content_id : MB_NO;
+ if (state == MB_FD1)
+ state = MB_FD;
+ mdev = to_macio_device(dev);
+ drv = to_macio_driver(dev->driver);
+ if (dev->driver && drv->mediabay_event)
+ drv->mediabay_event(mdev, state);
+ return 0;
}
-EXPORT_SYMBOL_GPL(media_bay_set_ide_infos);
-#endif /* CONFIG_BLK_DEV_IDE_PMAC */
static void media_bay_step(int i)
{
@@ -497,8 +474,8 @@ static void media_bay_step(int i)
if (bay->state != mb_powering_down)
poll_media_bay(bay);
- /* If timer expired or polling IDE busy, run state machine */
- if ((bay->state != mb_ide_waiting) && (bay->timer != 0)) {
+ /* If timer expired run state machine */
+ if (bay->timer != 0) {
bay->timer -= msecs_to_jiffies(MB_POLL_DELAY);
if (bay->timer > 0)
return;
@@ -508,100 +485,50 @@ static void media_bay_step(int i)
switch(bay->state) {
case mb_powering_up:
if (bay->ops->setup_bus(bay, bay->last_value) < 0) {
- MBDBG("mediabay%d: device not supported (kind:%d)\n", i, bay->content_id);
+ pr_debug("mediabay%d: device not supported (kind:%d)\n",
+ i, bay->content_id);
set_mb_power(bay, 0);
break;
}
bay->timer = msecs_to_jiffies(MB_RESET_DELAY);
bay->state = mb_enabling_bay;
- MBDBG("mediabay%d: enabling (kind:%d)\n", i, bay->content_id);
+ pr_debug("mediabay%d: enabling (kind:%d)\n", i, bay->content_id);
break;
case mb_enabling_bay:
bay->ops->un_reset(bay);
bay->timer = msecs_to_jiffies(MB_SETUP_DELAY);
bay->state = mb_resetting;
- MBDBG("mediabay%d: waiting reset (kind:%d)\n", i, bay->content_id);
+ pr_debug("mediabay%d: releasing bay reset (kind:%d)\n",
+ i, bay->content_id);
break;
case mb_resetting:
if (bay->content_id != MB_CD) {
- MBDBG("mediabay%d: bay is up (kind:%d)\n", i, bay->content_id);
+ pr_debug("mediabay%d: bay is up (kind:%d)\n", i,
+ bay->content_id);
bay->state = mb_up;
+ device_for_each_child(&bay->mdev->ofdev.dev,
+ bay, mb_broadcast_hotplug);
break;
}
-#ifdef CONFIG_BLK_DEV_IDE_PMAC
- MBDBG("mediabay%d: waiting IDE reset (kind:%d)\n", i, bay->content_id);
+ pr_debug("mediabay%d: releasing ATA reset (kind:%d)\n",
+ i, bay->content_id);
bay->ops->un_reset_ide(bay);
bay->timer = msecs_to_jiffies(MB_IDE_WAIT);
bay->state = mb_ide_resetting;
-#else
- printk(KERN_DEBUG "media-bay %d is ide (not compiled in kernel)\n", i);
- set_mb_power(bay, 0);
-#endif /* CONFIG_BLK_DEV_IDE_PMAC */
break;
-#ifdef CONFIG_BLK_DEV_IDE_PMAC
+
case mb_ide_resetting:
- bay->timer = msecs_to_jiffies(MB_IDE_TIMEOUT);
- bay->state = mb_ide_waiting;
- MBDBG("mediabay%d: waiting IDE ready (kind:%d)\n", i, bay->content_id);
+ pr_debug("mediabay%d: bay is up (kind:%d)\n", i, bay->content_id);
+ bay->state = mb_up;
+ device_for_each_child(&bay->mdev->ofdev.dev,
+ bay, mb_broadcast_hotplug);
break;
- case mb_ide_waiting:
- if (bay->cd_base == NULL) {
- bay->timer = 0;
- bay->state = mb_up;
- MBDBG("mediabay%d: up before IDE init\n", i);
- break;
- } else if (MB_IDE_READY(i)) {
- bay->timer = 0;
- bay->state = mb_up;
- if (bay->cd_index < 0) {
- printk("mediabay %d, registering IDE...\n", i);
- pmu_suspend();
- ide_port_scan(bay->cd_port);
- if (bay->cd_port->present)
- bay->cd_index = bay->cd_port->index;
- pmu_resume();
- }
- if (bay->cd_index == -1) {
- /* We eventually do a retry */
- bay->cd_retry++;
- printk("IDE register error\n");
- set_mb_power(bay, 0);
- } else {
- printk(KERN_DEBUG "media-bay %d is ide%d\n", i, bay->cd_index);
- MBDBG("mediabay %d IDE ready\n", i);
- }
- break;
- } else if (bay->timer > 0)
- bay->timer -= msecs_to_jiffies(MB_POLL_DELAY);
- if (bay->timer <= 0) {
- printk("\nIDE Timeout in bay %d !, IDE state is: 0x%02x\n",
- i, readb(bay->cd_base + 0x70));
- MBDBG("mediabay%d: nIDE Timeout !\n", i);
- set_mb_power(bay, 0);
- bay->timer = 0;
- }
- break;
-#endif /* CONFIG_BLK_DEV_IDE_PMAC */
+
case mb_powering_down:
bay->state = mb_empty;
-#ifdef CONFIG_BLK_DEV_IDE_PMAC
- if (bay->cd_index >= 0) {
- printk(KERN_DEBUG "Unregistering mb %d ide, index:%d\n", i,
- bay->cd_index);
- ide_port_unregister_devices(bay->cd_port);
- bay->cd_index = -1;
- }
- if (bay->cd_retry) {
- if (bay->cd_retry > MAX_CD_RETRIES) {
- /* Should add an error sound (sort of beep in dmasound) */
- printk("\nmedia-bay %d, IDE device badly inserted or unrecognised\n", i);
- } else {
- /* Force a new power down/up sequence */
- bay->content_id = MB_NO;
- }
- }
-#endif /* CONFIG_BLK_DEV_IDE_PMAC */
- MBDBG("mediabay%d: end of power down\n", i);
+ device_for_each_child(&bay->mdev->ofdev.dev,
+ bay, mb_broadcast_hotplug);
+ pr_debug("mediabay%d: end of power down\n", i);
break;
}
}
@@ -676,11 +603,6 @@ static int __devinit media_bay_attach(st
bay->last_value = bay->ops->content(bay);
bay->value_count = msecs_to_jiffies(MB_STABLE_DELAY);
bay->state = mb_empty;
- do {
- msleep(MB_POLL_DELAY);
- media_bay_step(i);
- } while((bay->state != mb_empty) &&
- (bay->state != mb_up));
/* Mark us ready by filling our mdev data */
macio_set_drvdata(mdev, bay);
@@ -725,7 +647,7 @@ static int media_bay_resume(struct macio
set_mb_power(bay, 0);
msleep(MB_POWER_DELAY);
if (bay->ops->content(bay) != bay->content_id) {
- printk("mediabay%d: content changed during sleep...\n", bay->index);
+ printk("mediabay%d: Content changed during sleep...\n", bay->index);
mutex_unlock(&bay->lock);
return 0;
}
@@ -733,9 +655,6 @@ static int media_bay_resume(struct macio
bay->last_value = bay->content_id;
bay->value_count = msecs_to_jiffies(MB_STABLE_DELAY);
bay->timer = msecs_to_jiffies(MB_POWER_DELAY);
-#ifdef CONFIG_BLK_DEV_IDE_PMAC
- bay->cd_retry = 0;
-#endif
do {
msleep(MB_POLL_DELAY);
media_bay_step(bay->index);
@@ -823,9 +742,6 @@ static int __init media_bay_init(void)
for (i=0; i<MAX_BAYS; i++) {
memset((char *)&media_bays[i], 0, sizeof(struct media_bay_info));
media_bays[i].content_id = -1;
-#ifdef CONFIG_BLK_DEV_IDE_PMAC
- media_bays[i].cd_index = -1;
-#endif
}
if (!machine_is(powermac))
return 0;
Index: linux-work/drivers/ide/pmac.c
===================================================================
--- linux-work.orig/drivers/ide/pmac.c 2009-12-01 17:57:00.000000000 +1100
+++ linux-work/drivers/ide/pmac.c 2009-12-02 11:24:46.000000000 +1100
@@ -43,10 +43,7 @@
#include <asm/pmac_feature.h>
#include <asm/sections.h>
#include <asm/irq.h>
-
-#ifndef CONFIG_PPC64
#include <asm/mediabay.h>
-#endif
#define DRV_NAME "ide-pmac"
@@ -59,13 +56,14 @@ typedef struct pmac_ide_hwif {
int irq;
int kind;
int aapl_bus_id;
- unsigned mediabay : 1;
unsigned broken_dma : 1;
unsigned broken_dma_warn : 1;
struct device_node* node;
struct macio_dev *mdev;
u32 timings[4];
volatile u32 __iomem * *kauai_fcr;
+ ide_hwif_t *hwif;
+
/* Those fields are duplicating what is in hwif. We currently
* can't use the hwif ones because of some assumptions that are
* beeing done by the generic code about the kind of dma controller
@@ -854,6 +852,11 @@ sanitize_timings(pmac_ide_hwif_t *pmif)
pmif->timings[2] = pmif->timings[3] = value2;
}
+static int on_media_bay(pmac_ide_hwif_t *pmif)
+{
+ return pmif->mdev && pmif->mdev->media_bay != NULL;
+}
+
/* Suspend call back, should be called after the child devices
* have actually been suspended
*/
@@ -866,7 +869,7 @@ static int pmac_ide_do_suspend(pmac_ide_
disable_irq(pmif->irq);
/* The media bay will handle itself just fine */
- if (pmif->mediabay)
+ if (on_media_bay(pmif))
return 0;
/* Kauai has bus control FCRs directly here */
@@ -889,7 +892,7 @@ static int pmac_ide_do_suspend(pmac_ide_
static int pmac_ide_do_resume(pmac_ide_hwif_t *pmif)
{
/* Hard reset & re-enable controller (do we really need to reset ? -BenH) */
- if (!pmif->mediabay) {
+ if (!on_media_bay(pmif)) {
ppc_md.feature_call(PMAC_FTR_IDE_RESET, pmif->node, pmif->aapl_bus_id, 1);
ppc_md.feature_call(PMAC_FTR_IDE_ENABLE, pmif->node, pmif->aapl_bus_id, 1);
msleep(10);
@@ -950,13 +953,11 @@ static void pmac_ide_init_dev(ide_drive_
pmac_ide_hwif_t *pmif =
(pmac_ide_hwif_t *)dev_get_drvdata(hwif->gendev.parent);
- if (pmif->mediabay) {
-#ifdef CONFIG_PMAC_MEDIABAY
- if (check_media_bay_by_base(pmif->regbase, MB_CD) == 0) {
+ if (on_media_bay(pmif)) {
+ if (check_media_bay(pmif->mdev->media_bay) == MB_CD) {
drive->dev_flags &= ~IDE_DFLAG_NOPROBE;
return;
}
-#endif
drive->dev_flags |= IDE_DFLAG_NOPROBE;
}
}
@@ -1072,26 +1073,22 @@ static int __devinit pmac_ide_setup_devi
writel(KAUAI_FCR_UATA_MAGIC |
KAUAI_FCR_UATA_RESET_N |
KAUAI_FCR_UATA_ENABLE, pmif->kauai_fcr);
-
- pmif->mediabay = 0;
/* Make sure we have sane timings */
sanitize_timings(pmif);
+ /* If we are on a media bay, wait for it to settle and lock it */
+ lock_media_bay(pmif->mdev->media_bay);
+
host = ide_host_alloc(&d, hws, 1);
- if (host == NULL)
- return -ENOMEM;
- hwif = host->ports[0];
+ if (host == NULL) {
+ rc = -ENOMEM;
+ goto bail;
+ }
+ hwif = pmif->hwif = host->ports[0];
-#ifndef CONFIG_PPC64
- /* XXX FIXME: Media bay stuff need re-organizing */
- if (np->parent && np->parent->name
- && strcasecmp(np->parent->name, "media-bay") == 0) {
-#ifdef CONFIG_PMAC_MEDIABAY
- media_bay_set_ide_infos(np->parent, pmif->regbase, pmif->irq,
- hwif);
-#endif /* CONFIG_PMAC_MEDIABAY */
- pmif->mediabay = 1;
+ if (on_media_bay(pmif)) {
+ /* Fixup bus ID for media bay */
if (!bidp)
pmif->aapl_bus_id = 1;
} else if (pmif->kind == controller_ohare) {
@@ -1100,9 +1097,7 @@ static int __devinit pmac_ide_setup_devi
* units, I keep the old way
*/
ppc_md.feature_call(PMAC_FTR_IDE_ENABLE, np, 0, 1);
- } else
-#endif
- {
+ } else {
/* This is necessary to enable IDE when net-booting */
ppc_md.feature_call(PMAC_FTR_IDE_RESET, np, pmif->aapl_bus_id, 1);
ppc_md.feature_call(PMAC_FTR_IDE_ENABLE, np, pmif->aapl_bus_id, 1);
@@ -1112,17 +1107,20 @@ static int __devinit pmac_ide_setup_devi
}
printk(KERN_INFO DRV_NAME ": Found Apple %s controller (%s), "
- "bus ID %d%s, irq %d\n", model_name[pmif->kind],
- pmif->mdev ? "macio" : "PCI", pmif->aapl_bus_id,
- pmif->mediabay ? " (mediabay)" : "", hw->irq);
+ "bus ID %d%s, irq %d\n", model_name[pmif->kind],
+ pmif->mdev ? "macio" : "PCI", pmif->aapl_bus_id,
+ on_media_bay(pmif) ? " (mediabay)" : "", hw->irq);
rc = ide_host_register(host, &d, hws);
- if (rc) {
- ide_host_free(host);
- return rc;
- }
+ if (rc)
+ pmif->hwif = NULL;
- return 0;
+ unlock_media_bay(pmif->mdev->media_bay);
+
+ bail:
+ if (rc && host)
+ ide_host_free(host);
+ return rc;
}
static void __devinit pmac_ide_init_ports(struct ide_hw *hw, unsigned long base)
@@ -1362,6 +1360,25 @@ pmac_ide_pci_resume(struct pci_dev *pdev
return rc;
}
+#ifdef CONFIG_PMAC_MEDIABAY
+static void pmac_ide_macio_mb_event(struct macio_dev* mdev, int mb_state)
+{
+ pmac_ide_hwif_t *pmif =
+ (pmac_ide_hwif_t *)dev_get_drvdata(&mdev->ofdev.dev);
+
+ switch(mb_state) {
+ case MB_CD:
+ if (!pmif->hwif->present)
+ ide_port_scan(pmif->hwif);
+ break;
+ default:
+ if (pmif->hwif->present)
+ ide_port_unregister_devices(pmif->hwif);
+ }
+}
+#endif /* CONFIG_PMAC_MEDIABAY */
+
+
static struct of_device_id pmac_ide_macio_match[] =
{
{
@@ -1386,6 +1403,9 @@ static struct macio_driver pmac_ide_maci
.probe = pmac_ide_macio_attach,
.suspend = pmac_ide_macio_suspend,
.resume = pmac_ide_macio_resume,
+#ifdef CONFIG_PMAC_MEDIABAY
+ .mediabay_event = pmac_ide_macio_mb_event,
+#endif
};
static const struct pci_device_id pmac_ide_pci_match[] = {
Index: linux-work/drivers/block/swim3.c
===================================================================
--- linux-work.orig/drivers/block/swim3.c 2009-12-01 17:57:00.000000000 +1100
+++ linux-work/drivers/block/swim3.c 2009-12-01 18:00:28.000000000 +1100
@@ -200,7 +200,7 @@ struct floppy_state {
int ejected;
wait_queue_head_t wait;
int wanted;
- struct device_node* media_bay; /* NULL when not in bay */
+ struct macio_dev *mdev;
char dbdma_cmd_space[5 * sizeof(struct dbdma_cmd)];
};
@@ -303,14 +303,13 @@ static int swim3_readbit(struct floppy_s
static void do_fd_request(struct request_queue * q)
{
int i;
- for(i=0;i<floppy_count;i++)
- {
-#ifdef CONFIG_PMAC_MEDIABAY
- if (floppy_states[i].media_bay &&
- check_media_bay(floppy_states[i].media_bay, MB_FD))
+
+ for(i=0; i<floppy_count; i++) {
+ struct floppy_state *fs = &floppy_states[i];
+ if (fs->mdev->media_bay &&
+ check_media_bay(fs->mdev->media_bay) != MB_FD)
continue;
-#endif /* CONFIG_PMAC_MEDIABAY */
- start_request(&floppy_states[i]);
+ start_request(fs);
}
}
@@ -849,10 +848,9 @@ static int floppy_ioctl(struct block_dev
if ((cmd & 0x80) && !capable(CAP_SYS_ADMIN))
return -EPERM;
-#ifdef CONFIG_PMAC_MEDIABAY
- if (fs->media_bay && check_media_bay(fs->media_bay, MB_FD))
+ if (fs->mdev->media_bay &&
+ check_media_bay(fs->mdev->media_bay) != MB_FD)
return -ENXIO;
-#endif
switch (cmd) {
case FDEJECT:
@@ -876,10 +874,9 @@ static int floppy_open(struct block_devi
int n, err = 0;
if (fs->ref_count == 0) {
-#ifdef CONFIG_PMAC_MEDIABAY
- if (fs->media_bay && check_media_bay(fs->media_bay, MB_FD))
+ if (fs->mdev->media_bay &&
+ check_media_bay(fs->mdev->media_bay) != MB_FD)
return -ENXIO;
-#endif
out_8(&sw->setup, S_IBM_DRIVE | S_FCLK_DIV2);
out_8(&sw->control_bic, 0xff);
out_8(&sw->mode, 0x95);
@@ -963,10 +960,9 @@ static int floppy_revalidate(struct gend
struct swim3 __iomem *sw;
int ret, n;
-#ifdef CONFIG_PMAC_MEDIABAY
- if (fs->media_bay && check_media_bay(fs->media_bay, MB_FD))
+ if (fs->mdev->media_bay &&
+ check_media_bay(fs->mdev->media_bay) != MB_FD)
return -ENXIO;
-#endif
sw = fs->swim3;
grab_drive(fs, revalidating, 0);
@@ -1009,7 +1005,6 @@ static const struct block_device_operati
static int swim3_add_device(struct macio_dev *mdev, int index)
{
struct device_node *swim = mdev->ofdev.node;
- struct device_node *mediabay;
struct floppy_state *fs = &floppy_states[index];
int rc = -EBUSY;
@@ -1036,9 +1031,7 @@ static int swim3_add_device(struct macio
}
dev_set_drvdata(&mdev->ofdev.dev, fs);
- mediabay = (strcasecmp(swim->parent->type, "media-bay") == 0) ?
- swim->parent : NULL;
- if (mediabay == NULL)
+ if (mdev->media_bay == NULL)
pmac_call_feature(PMAC_FTR_SWIM3_ENABLE, swim, 0, 1);
memset(fs, 0, sizeof(*fs));
@@ -1068,7 +1061,7 @@ static int swim3_add_device(struct macio
fs->secpercyl = 36;
fs->secpertrack = 18;
fs->total_secs = 2880;
- fs->media_bay = mediabay;
+ fs->mdev = mdev;
init_waitqueue_head(&fs->wait);
fs->dma_cmd = (struct dbdma_cmd *) DBDMA_ALIGN(fs->dbdma_cmd_space);
@@ -1093,7 +1086,7 @@ static int swim3_add_device(struct macio
init_timer(&fs->timeout);
printk(KERN_INFO "fd%d: SWIM3 floppy controller %s\n", floppy_count,
- mediabay ? "in media bay" : "");
+ mdev->media_bay ? "in media bay" : "");
return 0;
^ permalink raw reply
* [PATCH 2/5] powerpc/macio: Add dma_parms support to macio
From: Benjamin Herrenschmidt @ 2009-12-02 0:36 UTC (permalink / raw)
To: linux-ide; +Cc: tj, linuxppc-dev, jeff
This adds dma_parms to macio devices and initializes them with
default values. This will allow pata_macio to setup the appropriate
max segment size for the block layer.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
arch/powerpc/include/asm/macio.h | 1 +
drivers/macintosh/macio_asic.c | 5 +++++
2 files changed, 6 insertions(+)
--- linux-work.orig/arch/powerpc/include/asm/macio.h 2009-11-27 14:38:05.000000000 +1100
+++ linux-work/arch/powerpc/include/asm/macio.h 2009-11-27 14:38:20.000000000 +1100
@@ -39,6 +39,7 @@ struct macio_dev
struct macio_bus *bus; /* macio bus this device is on */
struct macio_dev *media_bay; /* Device is part of a media bay */
struct of_device ofdev;
+ struct device_dma_parameters dma_parms; /* ide needs that */
int n_resources;
struct resource resource[MACIO_DEV_COUNT_RESOURCES];
int n_interrupts;
Index: linux-work/drivers/macintosh/macio_asic.c
===================================================================
--- linux-work.orig/drivers/macintosh/macio_asic.c 2009-11-27 14:39:24.000000000 +1100
+++ linux-work/drivers/macintosh/macio_asic.c 2009-11-27 16:37:39.000000000 +1100
@@ -379,6 +379,11 @@ static struct macio_dev * macio_add_one_
dev->ofdev.dev.parent = parent;
dev->ofdev.dev.bus = &macio_bus_type;
dev->ofdev.dev.release = macio_release_dev;
+ dev->ofdev.dev.dma_parms = &dev->dma_parms;
+
+ /* Standard DMA paremeters */
+ dma_set_max_seg_size(&dev->ofdev.dev, 65536);
+ dma_set_seg_boundary(&dev->ofdev.dev, 0xffffffff);
#ifdef CONFIG_PCI
/* Set the DMA ops to the ones from the PCI device, this could be
^ permalink raw reply
* [PATCH 1/5] powerpc/macio: Add devres support to macio_device
From: Benjamin Herrenschmidt @ 2009-12-02 0:36 UTC (permalink / raw)
To: linux-ide; +Cc: tj, linuxppc-dev, jeff
This adds some basic devres support. When enabled via macio_enable_devres()
resources requested by drivers will be automatically released.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
arch/powerpc/include/asm/macio.h | 2 +
drivers/macintosh/macio_asic.c | 47 +++++++++++++++++++++++++++++++++++++++
2 files changed, 49 insertions(+)
--- linux-work.orig/arch/powerpc/include/asm/macio.h 2009-11-30 12:48:33.000000000 +1100
+++ linux-work/arch/powerpc/include/asm/macio.h 2009-11-30 14:16:55.000000000 +1100
@@ -78,6 +78,8 @@ static inline unsigned long macio_resour
return res->end - res->start + 1;
}
+extern int macio_enable_devres(struct macio_dev *dev);
+
extern int macio_request_resource(struct macio_dev *dev, int resource_no, const char *name);
extern void macio_release_resource(struct macio_dev *dev, int resource_no);
extern int macio_request_resources(struct macio_dev *dev, const char *name);
Index: linux-work/drivers/macintosh/macio_asic.c
===================================================================
--- linux-work.orig/drivers/macintosh/macio_asic.c 2009-11-30 12:41:12.000000000 +1100
+++ linux-work/drivers/macintosh/macio_asic.c 2009-11-30 14:17:19.000000000 +1100
@@ -538,6 +538,42 @@ void macio_unregister_driver(struct maci
driver_unregister(&drv->driver);
}
+/* Managed MacIO resources */
+struct macio_devres {
+ u32 res_mask;
+};
+
+static void maciom_release(struct device *gendev, void *res)
+{
+ struct macio_dev *dev = to_macio_device(gendev);
+ struct macio_devres *dr = res;
+ int i, max;
+
+ max = min(dev->n_resources, 32);
+ for (i = 0; i < max; i++) {
+ if (dr->res_mask & (1 << i))
+ macio_release_resource(dev, i);
+ }
+}
+
+int macio_enable_devres(struct macio_dev *dev)
+{
+ struct macio_devres *dr;
+
+ dr = devres_find(&dev->ofdev.dev, maciom_release, NULL, NULL);
+ if (!dr) {
+ dr = devres_alloc(maciom_release, sizeof(*dr), GFP_KERNEL);
+ if (!dr)
+ return -ENOMEM;
+ }
+ return devres_get(&dev->ofdev.dev, dr, NULL, NULL) != NULL;
+}
+
+static struct macio_devres * find_macio_dr(struct macio_dev *dev)
+{
+ return devres_find(&dev->ofdev.dev, maciom_release, NULL, NULL);
+}
+
/**
* macio_request_resource - Request an MMIO resource
* @dev: pointer to the device holding the resource
@@ -555,6 +591,8 @@ void macio_unregister_driver(struct maci
int macio_request_resource(struct macio_dev *dev, int resource_no,
const char *name)
{
+ struct macio_devres *dr = find_macio_dr(dev);
+
if (macio_resource_len(dev, resource_no) == 0)
return 0;
@@ -562,6 +600,9 @@ int macio_request_resource(struct macio_
macio_resource_len(dev, resource_no),
name))
goto err_out;
+
+ if (dr && resource_no < 32)
+ dr->res_mask |= 1 << resource_no;
return 0;
@@ -582,10 +623,14 @@ err_out:
*/
void macio_release_resource(struct macio_dev *dev, int resource_no)
{
+ struct macio_devres *dr = find_macio_dr(dev);
+
if (macio_resource_len(dev, resource_no) == 0)
return;
release_mem_region(macio_resource_start(dev, resource_no),
macio_resource_len(dev, resource_no));
+ if (dr && resource_no < 32)
+ dr->res_mask &= ~(1 << resource_no);
}
/**
@@ -744,3 +789,5 @@ EXPORT_SYMBOL(macio_request_resource);
EXPORT_SYMBOL(macio_release_resource);
EXPORT_SYMBOL(macio_request_resources);
EXPORT_SYMBOL(macio_release_resources);
+EXPORT_SYMBOL(macio_enable_devres);
+
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox