timer_list: convert timer list to be a proper seq_file

When running with 4096 cores attemping to read /proc/timer_list will fail
with an ENOMEM condition.  On a sufficantly large systems the total amount
of data is more then 4mb, so it won't fit into a single buffer.  The
failure can also occur on smaller systems when memory fragmentation is
high as reported by Dave Jones.

Convert /proc/timer_list to a proper seq_file with its own iterator.  This
is a little more complex given that we have to make two passes with two
separate headers.

[akpm@linux-foundation.org: whitespace fixlet]
[akpm@linux-foundation.org: fix up comment]
[akpm@linux-foundation.org: fix gcc warnings]
[akpm@linux-foundation.org: fix typo in comment]
Signed-off-by: Nathan Zimmer <nzimmer@sgi.com>
Reported-by: Dave Jones <davej@redhat.com>
Cc: John Stultz <johnstul@us.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

diff --git a/kernel/time/timer_list.c b/kernel/time/timer_list.c
index 380a58977490371f60f4fdb7fde0525d7d124aab..8251f3f7ca9f49a1678bf8376b8015c82d0802d1 100644 (file)
--- a/kernel/time/timer_list.c
+++ b/kernel/time/timer_list.c
@@ -252,38 +252,114 @@ static int timer_list_show(struct seq_file *m, void *v)
        u64 now = ktime_to_ns(ktime_get());
        int cpu;
 
-       SEQ_printf(m, "Timer List Version: v0.7\n");
-       SEQ_printf(m, "HRTIMER_MAX_CLOCK_BASES: %d\n", HRTIMER_MAX_CLOCK_BASES);
-       SEQ_printf(m, "now at %Ld nsecs\n", (unsigned long long)now);
-       SEQ_printf(m, "\n");
-
-       for_each_online_cpu(cpu)
+       if (v == (void *)1) {
+               SEQ_printf(m, "Timer List Version: v0.7\n");
+               SEQ_printf(m, "HRTIMER_MAX_CLOCK_BASES: %d\n",
+                          HRTIMER_MAX_CLOCK_BASES);
+               SEQ_printf(m, "now at %Ld nsecs\n", (unsigned long long)now);
+               SEQ_printf(m, "\n");
+       } else if (v < (void *)(unsigned long)(nr_cpu_ids + 2)) {
+               cpu = (unsigned long)(v - 2);
                print_cpu(m, cpu, now);
-
+       }
 #ifdef CONFIG_GENERIC_CLOCKEVENTS
-       timer_list_show_tickdevices_header(m);
-       for_each_online_cpu(cpu)
+       else if (v == (void *)(unsigned long)nr_cpu_ids + 2) {
+               timer_list_show_tickdevices_header(m);
+       } else {
+               cpu = (unsigned long)(v - 3 - nr_cpu_ids);
                print_tickdevice(m, tick_get_device(cpu), cpu);
+       }
 #endif
-
        return 0;
 }
 
+/*
+ * This itererator really needs some explanation since it is offset and has
+ * two passes, one of which is controlled by a config option.
+ * In a hotplugged system some cpus, including cpu 0, may be missing so we have
+ * to use cpumask_* to iterate over the cpus.
+ * For the first pass:
+ * It returns 1 for the header position.
+ * For cpu 0 it returns 2 and the final possible cpu would be nr_cpu_ids + 1.
+ * On the second pass:
+ * It returnes nr_cpu_ids + 1 for the second header position.
+ * For cpu 0 it returns nr_cpu_ids + 2
+ * The final possible cpu would be nr_cpu_ids + nr_cpu_ids + 2.
+ */
+static void *timer_list_start(struct seq_file *file, loff_t *offset)
+{
+       unsigned long n = *offset;
+
+       if (n == 0)
+               return (void *) 1;
+
+       if (n < nr_cpu_ids + 1) {
+               n--;
+               if (n > 0)
+                       n = cpumask_next(n - 1, cpu_online_mask);
+               else
+                       n = cpumask_first(cpu_online_mask);
+               *offset = n + 1;
+               return (void *)(unsigned long)(n + 2);
+       }
+
+#ifdef CONFIG_GENERIC_CLOCKEVENTS
+       if (n == nr_cpu_ids + 1)
+               return (void *)(unsigned long)(nr_cpu_ids + 2);
+
+       if (n < nr_cpu_ids * 2 + 2) {
+               n -= (nr_cpu_ids + 2);
+               if (n > 0)
+                       n = cpumask_next(n - 1, cpu_online_mask);
+               else
+                       n = cpumask_first(cpu_online_mask);
+               *offset = n + 2 + nr_cpu_ids;
+               return (void *)(unsigned long)(n + 3 + nr_cpu_ids);
+       }
+#endif
+
+       return NULL;
+}
+
+static void *timer_list_next(struct seq_file *file, void *data, loff_t *offset)
+{
+       (*offset)++;
+       return timer_list_start(file, offset);
+}
+
+static void timer_list_stop(struct seq_file *file, void *data)
+{
+}
+
+static const struct seq_operations timer_list_sops = {
+       .start = timer_list_start,
+       .next = timer_list_next,
+       .stop = timer_list_stop,
+       .show = timer_list_show,
+};
+
 void sysrq_timer_list_show(void)
 {
        timer_list_show(NULL, NULL);
 }
 
+static int timer_list_release(struct inode *inode, struct file *filep)
+{
+       seq_release(inode, filep);
+
+       return 0;
+}
+
 static int timer_list_open(struct inode *inode, struct file *filp)
 {
-       return single_open(filp, timer_list_show, NULL);
+       return seq_open(filp, &timer_list_sops);
 }
 
 static const struct file_operations timer_list_fops = {
        .open           = timer_list_open,
        .read           = seq_read,
        .llseek         = seq_lseek,
-       .release        = single_release,
+       .release        = timer_list_release,
 };
 
 static int __init init_timer_list_procfs(void)