Re: [PATCH 3/3 -mm] kexec based hibernation -v6: kexec hibernate/resume

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Monday, 19 of November 2007, Huang, Ying wrote:
> This patch implements kexec based hibernate/resume. This is based on
> the facility provided by kexec_jump. The ACPI methods are called at
> specified environment to conform the ACPI specification. Two new
> reboot commands are added to trigger hibernate/resume.
> 
> Signed-off-by: Huang Ying <ying.huang@intel.com>
> 
> ---
>  include/linux/kexec.h   |    5 +
>  include/linux/reboot.h  |    2 
>  include/linux/suspend.h |    9 ++
>  kernel/power/disk.c     |  155 ++++++++++++++++++++++++++++++++++++++++++++++++
>  kernel/sys.c            |   42 +++++++++++++
>  5 files changed, 212 insertions(+), 1 deletion(-)
> 
> --- a/kernel/power/disk.c
> +++ b/kernel/power/disk.c
> @@ -21,6 +21,7 @@
>  #include <linux/console.h>
>  #include <linux/cpu.h>
>  #include <linux/freezer.h>
> +#include <linux/kexec.h>
>  
>  #include "power.h"
>  
> @@ -438,6 +439,160 @@ int hibernate(void)
>  	return error;
>  }
>  
> +#ifdef CONFIG_KEXEC
> +static void kexec_hibernate_power_down(void)
> +{
> +	switch (hibernation_mode) {
> +	case HIBERNATION_TEST:
> +	case HIBERNATION_TESTPROC:
> +		break;
> +	case HIBERNATION_REBOOT:
> +		machine_restart(NULL);
> +		break;
> +	case HIBERNATION_PLATFORM:
> +		if (!hibernation_ops)
> +			break;
> +		hibernation_ops->enter();

hibernation_platform_enter() should be used here (as of the current mainline).

> +		/* We should never get here */
> +		while (1);
> +		break;
> +	case HIBERNATION_SHUTDOWN:
> +		machine_power_off();
> +		break;
> +	}
> +	machine_halt();
> +	/*
> +	 * Valid image is on the disk, if we continue we risk serious data
> +	 * corruption after resume.
> +	 */
> +	printk(KERN_CRIT "Please power me down manually\n");
> +	while (1);
> +}

Hm, what's the difference between the above function and power_down(),
actually?

> +
> +int kexec_hibernate(struct kimage *image)
> +{
> +	int error;
> +	int platform_mode = (hibernation_mode == HIBERNATION_PLATFORM);
> +	unsigned long cmd_ret;
> +
> +	mutex_lock(&pm_mutex);
> +
> +	pm_prepare_console();
> +	suspend_console();
> +
> +	error = pm_notifier_call_chain(PM_HIBERNATION_PREPARE);
> +	if (error)
> +		goto Resume_console;
> +
> +	error = platform_start(platform_mode);
> +	if (error)
> +		goto Resume_console;
> +
> +	error = device_suspend(PMSG_FREEZE);
> +	if (error)
> +		goto Resume_console;
> +
> +	error = platform_pre_snapshot(platform_mode);
> +	if (error)
> +		goto Resume_devices;
> +
> +	error = disable_nonboot_cpus();
> +	if (error)
> +		goto Resume_devices;

I wonder if it's viable to merge the above with hibernate() and
hibernation_snapshot() somehow, to avoid code duplication?

> +	local_irq_disable();
> +	/* At this point, device_suspend() has been called, but *not*
> +	 * device_power_down(). We *must* device_power_down() now.
> +	 * Otherwise, drivers for some devices (e.g. interrupt
> +	 * controllers) become desynchronized with the actual state of
> +	 * the hardware at resume time, and evil weirdness ensues.
> +	 */
> +	error = device_power_down(PMSG_FREEZE);
> +	if (error)
> +		goto Enable_irqs;
> +
> +	save_processor_state();
> +	error = machine_kexec_jump(image, &cmd_ret,
> +				   KJUMP_CMD_HIBERNATE_WRITE_IMAGE);

It looks like machine_kexec_jump() simply replaces the call to
swsusp_arch_suspend() in create_image().  Is it really necessary to duplicate
all that code?

> +	restore_processor_state();
> +
> +	if (cmd_ret == KJUMP_CMD_HIBERNATE_POWER_DOWN)
> +		kexec_hibernate_power_down();
> +
> +	platform_leave(platform_mode);
> +
> +	/* NOTE:  device_power_up() is just a resume() for devices
> +	 * that suspended with irqs off ... no overall powerup.
> +	 */
> +	device_power_up();
> + Enable_irqs:
> +	local_irq_enable();
> +	enable_nonboot_cpus();
> + Resume_devices:
> +	platform_finish(platform_mode);
> +	device_resume();
> + Resume_console:
> +	pm_notifier_call_chain(PM_POST_HIBERNATION);
> +	resume_console();
> +	pm_restore_console();
> +	mutex_unlock(&pm_mutex);
> +	return error;
> +}
> +
> +int kexec_resume(struct kimage *image)
> +{
> +	int error;
> +	int platform_mode = (hibernation_mode == HIBERNATION_PLATFORM);
> +
> +	mutex_lock(&pm_mutex);
> +
> +	pm_prepare_console();
> +	suspend_console();
> +
> +	error = device_suspend(PMSG_PRETHAW);
> +	if (error)
> +		goto Resume_console;
> +
> +	error = platform_pre_restore(platform_mode);
> +	if (error)
> +		goto Resume_devices;
> +
> +	error = disable_nonboot_cpus();
> +	if (error)
> +		goto Resume_devices;
> +	local_irq_disable();
> +	/* At this point, device_suspend() has been called, but *not*
> +	 * device_power_down(). We *must* device_power_down() now.
> +	 * Otherwise, drivers for some devices (e.g. interrupt controllers)
> +	 * become desynchronized with the actual state of the hardware
> +	 * at resume time, and evil weirdness ensues.
> +	 */
> +	error = device_power_down(PMSG_PRETHAW);
> +	if (error)
> +		goto Enable_irqs;
> +
> +	save_processor_state();
> +	error = machine_kexec_jump(image, NULL, KJUMP_CMD_HIBERNATE_RESUME);

Same here.  Much of the code in this function duplicates the existing code.  Is
this really necessary?

> +	restore_processor_state();
> +
> +	platform_leave(platform_mode);
> +
> +	/* NOTE:  device_power_up() is just a resume() for devices
> +	 * that suspended with irqs off ... no overall powerup.
> +	 */
> +	device_power_up();
> + Enable_irqs:
> +	local_irq_enable();
> +	enable_nonboot_cpus();
> + Resume_devices:
> +	platform_restore_cleanup(platform_mode);
> +	device_resume();
> + Resume_console:
> +	resume_console();
> +	pm_restore_console();
> +	mutex_unlock(&pm_mutex);
> +	return error;
> +}
> +#endif
>  
>  /**
>   *	software_resume - Resume from a saved image.

Apart from the above, there's some new debug code to be added to disk.c
in 2.6.25.  It's in the ACPI test tree right now and you can get it as
individual patches from:
http://www.sisk.pl/kernel/hibernation_and_suspend/2.6.24-rc2/patches/
(patches 10-12).

Please base your changes on top of that.

> --- a/include/linux/suspend.h
> +++ b/include/linux/suspend.h
> @@ -207,6 +207,15 @@ static inline void hibernation_set_ops(s
>  static inline int hibernate(void) { return -ENOSYS; }
>  #endif /* CONFIG_HIBERNATION */
>  
> +struct kimage;
> +#if defined(CONFIG_HIBERNATION) && defined(CONFIG_KEXEC)
> +extern int kexec_hibernate(struct kimage *image);
> +extern int kexec_resume(struct kimage *image);
> +#else
> +static inline int kexec_hibernate(struct kimage *image) { return -ENOSYS; }
> +static inline int kexec_resume(struct kimage *image) { return -ENOSYS; }
> +#endif
> +
>  #ifdef CONFIG_PM_SLEEP
>  void save_processor_state(void);
>  void restore_processor_state(void);
> --- a/kernel/sys.c
> +++ b/kernel/sys.c
> @@ -321,6 +321,34 @@ static int kernel_kexec(unsigned long cm
>  	return ret;
>  }
>  
> +static int kernel_kexec_hibernate(void)
> +{
> +	int ret = -ENOSYS;
> +#if defined(CONFIG_KEXEC) && defined(CONFIG_HIBERNATION)
> +	struct kimage *image;
> +	image = xchg(&kexec_image, NULL);
> +	if (!image)
> +		return -EINVAL;
> +	ret = kexec_hibernate(image);
> +	kimage_free(image);
> +#endif
> +	return ret;
> +}
> +
> +static int kernel_kexec_resume(void)
> +{
> +	int ret = -ENOSYS;
> +#if defined(CONFIG_KEXEC) && defined(CONFIG_HIBERNATION)
> +	struct kimage *image;
> +	image = xchg(&kexec_image, NULL);
> +	if (!image)
> +		return -EINVAL;
> +	ret = kexec_resume(image);
> +	kimage_free(image);
> +#endif
> +	return ret;
> +}
> +
>  static void kernel_shutdown_prepare(enum system_states state)
>  {
>  	blocking_notifier_call_chain(&reboot_notifier_list,
> @@ -442,6 +470,20 @@ asmlinkage long sys_reboot(int magic1, i
>  		}
>  #endif
>  
> +	case LINUX_REBOOT_CMD_KEXEC_HIBERNATE:
> +		{
> +			int ret = kernel_kexec_hibernate();
> +			unlock_kernel();
> +			return ret;
> +		}
> +
> +	case LINUX_REBOOT_CMD_KEXEC_RESUME:
> +		{
> +			int ret = kernel_kexec_resume();
> +			unlock_kernel();
> +			return ret;
> +		}
> +
>  	default:
>  		unlock_kernel();
>  		return -EINVAL;
> --- a/include/linux/reboot.h
> +++ b/include/linux/reboot.h
> @@ -33,6 +33,8 @@
>  #define	LINUX_REBOOT_CMD_RESTART2	0xA1B2C3D4
>  #define	LINUX_REBOOT_CMD_SW_SUSPEND	0xD000FCE2
>  #define	LINUX_REBOOT_CMD_KEXEC		0x45584543
> +#define LINUX_REBOOT_CMD_KEXEC_HIBERNATE 0xE4390DF3
> +#define LINUX_REBOOT_CMD_KEXEC_RESUME	0xF5A41C2E
>  
>  
>  #ifdef __KERNEL__
> --- a/include/linux/kexec.h
> +++ b/include/linux/kexec.h
> @@ -129,7 +129,10 @@ extern int machine_kexec_jump(struct kim
>  			      unsigned long cmd);
>  extern int kexec_jump(struct kimage *image, unsigned long *cmd_ret,
>  		      unsigned long cmd);
> -#define KJUMP_CMD_NONE 0
> +#define KJUMP_CMD_NONE			0x0
> +#define KJUMP_CMD_HIBERNATE_WRITE_IMAGE	0x6b630001
> +#define KJUMP_CMD_HIBERNATE_POWER_DOWN	0x6b630002
> +#define KJUMP_CMD_HIBERNATE_RESUME	0x6b630003
>  #ifdef CONFIG_COMPAT
>  extern asmlinkage long compat_sys_kexec_load(unsigned long entry,
>  				unsigned long nr_segments,
> 
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/