Re: [PATCH 2/3] drivers/xen/balloon.c: Various balloon features and fixes

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

On Mon, Dec 20, 2010 at 02:47:24PM +0100, Daniel Kiper wrote:
> Features and fixes:
>   - HVM mode is supported now,
>   - migration from mod_timer() to schedule_delayed_work(),
>   - removal of driver_pages (I do not have seen any
>     references to it),
>   - protect before CPU exhaust by event/x process during
>     errors by adding some delays in scheduling next event,
>   - some other minor fixes.
> 
> Signed-off-by: Daniel Kiper <dkiper@net-space.pl>
> ---
>  arch/x86/xen/mmu.c    |    3 +-
>  drivers/xen/balloon.c |  128 +++++++++++++++++++++++++++++++++----------------
>  2 files changed, 87 insertions(+), 44 deletions(-)
> 
> diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
> index 42086ac..6278650 100644
> --- a/arch/x86/xen/mmu.c
> +++ b/arch/x86/xen/mmu.c
> @@ -75,8 +75,7 @@
>  
>  /*
>   * Protects atomic reservation decrease/increase against concurrent increases.
> - * Also protects non-atomic updates of current_pages and driver_pages, and
> - * balloon lists.
> + * Also protects non-atomic updates of current_pages and balloon lists.
>   */
>  DEFINE_SPINLOCK(xen_reservation_lock);
>  
> diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c
> index 500290b..06dbdad 100644
> --- a/drivers/xen/balloon.c
> +++ b/drivers/xen/balloon.c
> @@ -69,14 +69,11 @@ struct balloon_stats {
>  	/* We aim for 'current allocation' == 'target allocation'. */
>  	unsigned long current_pages;
>  	unsigned long target_pages;
> -	/*
> -	 * Drivers may alter the memory reservation independently, but they
> -	 * must inform the balloon driver so we avoid hitting the hard limit.
> -	 */
> -	unsigned long driver_pages;
>  	/* Number of pages in high- and low-memory balloons. */
>  	unsigned long balloon_low;
>  	unsigned long balloon_high;
> +	unsigned long schedule_delay;
> +	unsigned long max_schedule_delay;
>  };
>  
>  static DEFINE_MUTEX(balloon_mutex);
> @@ -103,15 +100,14 @@ static LIST_HEAD(ballooned_pages);
>  
>  /* Main work function, always executed in process context. */
>  static void balloon_process(struct work_struct *work);
> -static DECLARE_WORK(balloon_worker, balloon_process);
> -static struct timer_list balloon_timer;
> +static DECLARE_DELAYED_WORK(balloon_worker, balloon_process);
>  
>  /* When ballooning out (allocating memory to return to Xen) we don't really
>     want the kernel to try too hard since that can trigger the oom killer. */
>  #define GFP_BALLOON \
>  	(GFP_HIGHUSER | __GFP_NOWARN | __GFP_NORETRY | __GFP_NOMEMALLOC)
>  
> -static void scrub_page(struct page *page)
> +static inline void scrub_page(struct page *page)
>  {
>  #ifdef CONFIG_XEN_SCRUB_PAGES
>  	clear_highpage(page);
> @@ -172,9 +168,29 @@ static struct page *balloon_next_page(struct page *page)
>  	return list_entry(next, struct page, lru);
>  }
>  
> -static void balloon_alarm(unsigned long unused)
> +static void update_schedule_delay(int cmd)
>  {
> -	schedule_work(&balloon_worker);
> +	unsigned long new_schedule_delay;
> +
> +	/*
> +	 * cmd >= 0: balloon_stats.schedule_delay = 1,
> +	 * cmd < 0: increase balloon_stats.schedule_delay but
> +	 *          no more than balloon_stats.max_schedule_delay.
> +	 */
> +
> +	if (cmd >= 0) {
> +		balloon_stats.schedule_delay = 1;
> +		return;
> +	}
> +
> +	new_schedule_delay = balloon_stats.schedule_delay << 1;
> +
> +	if (new_schedule_delay > balloon_stats.max_schedule_delay) {
> +		balloon_stats.schedule_delay = balloon_stats.max_schedule_delay;
> +		return;
> +	}
> +
> +	balloon_stats.schedule_delay = new_schedule_delay;
>  }
>  
>  static unsigned long current_target(void)
> @@ -191,9 +207,9 @@ static unsigned long current_target(void)
>  
>  static int increase_reservation(unsigned long nr_pages)
>  {
> +	int rc, state = 0;
>  	unsigned long  pfn, i, flags;
>  	struct page   *page;
> -	long           rc;
>  	struct xen_memory_reservation reservation = {
>  		.address_bits = 0,
>  		.extent_order = 0,
> @@ -206,8 +222,17 @@ static int increase_reservation(unsigned long nr_pages)
>  	spin_lock_irqsave(&xen_reservation_lock, flags);
>  
>  	page = balloon_first_page();
> +
> +	if (!page) {
> +		state = -ENOMEM;
> +		goto out;
> +	}
> +
>  	for (i = 0; i < nr_pages; i++) {
> -		BUG_ON(page == NULL);
> +		if (!page) {
> +			nr_pages = i;
> +			break;
> +		}
>  		frame_list[i] = page_to_pfn(page);
>  		page = balloon_next_page(page);
>  	}
> @@ -215,8 +240,11 @@ static int increase_reservation(unsigned long nr_pages)
>  	set_xen_guest_handle(reservation.extent_start, frame_list);
>  	reservation.nr_extents = nr_pages;
>  	rc = HYPERVISOR_memory_op(XENMEM_populate_physmap, &reservation);
> -	if (rc < 0)
> -		goto out;
> +	if (rc < nr_pages) {
> +		state = (rc <= 0) ? -ENOMEM : 1;
> +		if (rc <= 0)
> +			goto out;
> +	}
>  
>  	for (i = 0; i < rc; i++) {
>  		page = balloon_retrieve();
> @@ -229,7 +257,7 @@ static int increase_reservation(unsigned long nr_pages)
>  		set_phys_to_machine(pfn, frame_list[i]);
>  
>  		/* Link back into the page tables if not highmem. */
> -		if (pfn < max_low_pfn) {
> +		if (xen_pv_domain() && !PageHighMem(page)) {
>  			int ret;
>  			ret = HYPERVISOR_update_va_mapping(
>  				(unsigned long)__va(pfn << PAGE_SHIFT),
> @@ -249,15 +277,14 @@ static int increase_reservation(unsigned long nr_pages)
>   out:
>  	spin_unlock_irqrestore(&xen_reservation_lock, flags);
>  
> -	return rc < 0 ? rc : rc != nr_pages;
> +	return state;
>  }
>  
>  static int decrease_reservation(unsigned long nr_pages)
>  {
>  	unsigned long  pfn, i, flags;
>  	struct page   *page;
> -	int            need_sleep = 0;
> -	int ret;
> +	int ret, state = 0;
>  	struct xen_memory_reservation reservation = {
>  		.address_bits = 0,
>  		.extent_order = 0,
> @@ -270,7 +297,7 @@ static int decrease_reservation(unsigned long nr_pages)
>  	for (i = 0; i < nr_pages; i++) {
>  		if ((page = alloc_page(GFP_BALLOON)) == NULL) {
>  			nr_pages = i;
> -			need_sleep = 1;
> +			state = -ENOMEM;
>  			break;
>  		}
>  
> @@ -279,7 +306,7 @@ static int decrease_reservation(unsigned long nr_pages)
>  
>  		scrub_page(page);
>  
> -		if (!PageHighMem(page)) {
> +		if (xen_pv_domain() && !PageHighMem(page)) {
>  			ret = HYPERVISOR_update_va_mapping(
>  				(unsigned long)__va(pfn << PAGE_SHIFT),
>  				__pte_ma(0), 0);
> @@ -310,7 +337,7 @@ static int decrease_reservation(unsigned long nr_pages)
>  
>  	spin_unlock_irqrestore(&xen_reservation_lock, flags);
>  
> -	return need_sleep;
> +	return state;
>  }
>  
>  /*
> @@ -321,27 +348,41 @@ static int decrease_reservation(unsigned long nr_pages)
>   */
>  static void balloon_process(struct work_struct *work)
>  {
> -	int need_sleep = 0;
> +	int rc, state = 0;
>  	long credit;
>  
>  	mutex_lock(&balloon_mutex);
>  
>  	do {
>  		credit = current_target() - balloon_stats.current_pages;
> -		if (credit > 0)
> -			need_sleep = (increase_reservation(credit) != 0);
> -		if (credit < 0)
> -			need_sleep = (decrease_reservation(-credit) != 0);
> +
> +		/*
> +		 * state > 0: hungry,
> +		 * state == 0: done or nothing to do,
> +		 * state < 0: error, go to sleep.

Would it be better to just have #defines for this?

> +		 */
> +
> +		if (credit > 0) {
> +			rc = increase_reservation(credit);
> +			state = (rc < 0) ? rc : state;
> +		}
> +
> +		if (credit < 0) {
> +			rc = decrease_reservation(-credit);
> +			state = (rc < 0) ? rc : state;
> +		}
> +
> +		update_schedule_delay(state);
>  
>  #ifndef CONFIG_PREEMPT
>  		if (need_resched())
>  			schedule();
>  #endif
> -	} while ((credit != 0) && !need_sleep);
> +	} while (credit && state >= 0);
>  
>  	/* Schedule more work if there is some still to be done. */
> -	if (current_target() != balloon_stats.current_pages)
> -		mod_timer(&balloon_timer, jiffies + HZ);
> +	if (state < 0)
> +		schedule_delayed_work(&balloon_worker, balloon_stats.schedule_delay * HZ);
>  
>  	mutex_unlock(&balloon_mutex);
>  }
> @@ -351,7 +392,7 @@ static void balloon_set_new_target(unsigned long target)
>  {
>  	/* No need for lock. Not read-modify-write updates. */
>  	balloon_stats.target_pages = target;
> -	schedule_work(&balloon_worker);
> +	schedule_delayed_work(&balloon_worker, 0);
>  }
>  
>  static struct xenbus_watch target_watch =
> @@ -395,28 +436,28 @@ static struct notifier_block xenstore_notifier;
>  
>  static int __init balloon_init(void)
>  {
> -	unsigned long pfn;
> +	unsigned long pfn, nr_pages;
>  	struct page *page;
>  
> -	if (!xen_pv_domain())
> +	if (!xen_domain())
>  		return -ENODEV;
>  
>  	pr_info("xen_balloon: Initialising balloon driver.\n");
>  
> -	balloon_stats.current_pages = min(xen_start_info->nr_pages, max_pfn);
> +	nr_pages = xen_pv_domain() ? xen_start_info->nr_pages : max_pfn;
> +
> +	balloon_stats.current_pages = min(nr_pages, max_pfn);
>  	balloon_stats.target_pages  = balloon_stats.current_pages;
>  	balloon_stats.balloon_low   = 0;
>  	balloon_stats.balloon_high  = 0;
> -	balloon_stats.driver_pages  = 0UL;
>  
> -	init_timer(&balloon_timer);
> -	balloon_timer.data = 0;
> -	balloon_timer.function = balloon_alarm;
> +	balloon_stats.schedule_delay = 1;
> +	balloon_stats.max_schedule_delay = 32;

How did you arrive at that number?

>  
>  	register_balloon(&balloon_sysdev);
>  
>  	/* Initialise the balloon with excess memory space. */
> -	for (pfn = xen_start_info->nr_pages; pfn < max_pfn; pfn++) {
> +	for (pfn = nr_pages; pfn < max_pfn; pfn++) {
>  		page = pfn_to_page(pfn);
>  		if (!PageReserved(page))
>  			balloon_append(page);
> @@ -452,7 +493,9 @@ module_exit(balloon_exit);
>  BALLOON_SHOW(current_kb, "%lu\n", PAGES2KB(balloon_stats.current_pages));
>  BALLOON_SHOW(low_kb, "%lu\n", PAGES2KB(balloon_stats.balloon_low));
>  BALLOON_SHOW(high_kb, "%lu\n", PAGES2KB(balloon_stats.balloon_high));
> -BALLOON_SHOW(driver_kb, "%lu\n", PAGES2KB(balloon_stats.driver_pages));
> +
> +static SYSDEV_ULONG_ATTR(schedule_delay, 0644, balloon_stats.schedule_delay);
> +static SYSDEV_ULONG_ATTR(max_schedule_delay, 0644, balloon_stats.max_schedule_delay);
>  
>  static ssize_t show_target_kb(struct sys_device *dev, struct sysdev_attribute *attr,
>  			      char *buf)
> @@ -515,23 +558,24 @@ static SYSDEV_ATTR(target, S_IRUGO | S_IWUSR,
>  static struct sysdev_attribute *balloon_attrs[] = {
>  	&attr_target_kb,
>  	&attr_target,
> +	&attr_schedule_delay.attr,
> +	&attr_max_schedule_delay.attr
>  };
>  
>  static struct attribute *balloon_info_attrs[] = {
>  	&attr_current_kb.attr,
>  	&attr_low_kb.attr,
>  	&attr_high_kb.attr,
> -	&attr_driver_kb.attr,
>  	NULL
>  };
>  
>  static struct attribute_group balloon_info_group = {
>  	.name = "info",
> -	.attrs = balloon_info_attrs,
> +	.attrs = balloon_info_attrs
>  };
>  
>  static struct sysdev_class balloon_sysdev_class = {
> -	.name = BALLOON_CLASS_NAME,
> +	.name = BALLOON_CLASS_NAME
>  };
>  
>  static int register_balloon(struct sys_device *sysdev)
> -- 
> 1.4.4.4