2010-06-23 01:22:17 +08:00
|
|
|
#ifndef _LINUX_U64_STATS_SYNC_H
|
|
|
|
#define _LINUX_U64_STATS_SYNC_H
|
|
|
|
|
|
|
|
/*
|
|
|
|
* To properly implement 64bits network statistics on 32bit and 64bit hosts,
|
|
|
|
* we provide a synchronization point, that is a noop on 64bit or UP kernels.
|
|
|
|
*
|
|
|
|
* Key points :
|
|
|
|
* 1) Use a seqcount on SMP 32bits, with low overhead.
|
|
|
|
* 2) Whole thing is a noop on 64bit arches or UP kernels.
|
|
|
|
* 3) Write side must ensure mutual exclusion or one seqcount update could
|
|
|
|
* be lost, thus blocking readers forever.
|
|
|
|
* If this synchronization point is not a mutex, but a spinlock or
|
|
|
|
* spinlock_bh() or disable_bh() :
|
|
|
|
* 3.1) Write side should not sleep.
|
|
|
|
* 3.2) Write side should not allow preemption.
|
|
|
|
* 3.3) If applicable, interrupts should be disabled.
|
|
|
|
*
|
|
|
|
* 4) If reader fetches several counters, there is no guarantee the whole values
|
|
|
|
* are consistent (remember point 1) : this is a noop on 64bit arches anyway)
|
|
|
|
*
|
|
|
|
* 5) readers are allowed to sleep or be preempted/interrupted : They perform
|
|
|
|
* pure reads. But if they have to fetch many values, it's better to not allow
|
|
|
|
* preemptions/interruptions to avoid many retries.
|
|
|
|
*
|
2010-06-24 08:04:38 +08:00
|
|
|
* 6) If counter might be written by an interrupt, readers should block interrupts.
|
|
|
|
* (On UP, there is no seqcount_t protection, a reader allowing interrupts could
|
|
|
|
* read partial values)
|
|
|
|
*
|
2010-06-24 08:54:06 +08:00
|
|
|
* 7) For softirq uses, readers can use u64_stats_fetch_begin_bh() and
|
|
|
|
* u64_stats_fetch_retry_bh() helpers
|
|
|
|
*
|
2010-06-23 01:22:17 +08:00
|
|
|
* Usage :
|
|
|
|
*
|
|
|
|
* Stats producer (writer) should use following template granted it already got
|
|
|
|
* an exclusive access to counters (a lock is already taken, or per cpu
|
|
|
|
* data is used [in a non preemptable context])
|
|
|
|
*
|
|
|
|
* spin_lock_bh(...) or other synchronization to get exclusive access
|
|
|
|
* ...
|
|
|
|
* u64_stats_update_begin(&stats->syncp);
|
|
|
|
* stats->bytes64 += len; // non atomic operation
|
|
|
|
* stats->packets64++; // non atomic operation
|
|
|
|
* u64_stats_update_end(&stats->syncp);
|
|
|
|
*
|
|
|
|
* While a consumer (reader) should use following template to get consistent
|
|
|
|
* snapshot for each variable (but no guarantee on several ones)
|
|
|
|
*
|
|
|
|
* u64 tbytes, tpackets;
|
|
|
|
* unsigned int start;
|
|
|
|
*
|
|
|
|
* do {
|
|
|
|
* start = u64_stats_fetch_begin(&stats->syncp);
|
|
|
|
* tbytes = stats->bytes64; // non atomic operation
|
|
|
|
* tpackets = stats->packets64; // non atomic operation
|
2010-06-24 08:04:38 +08:00
|
|
|
* } while (u64_stats_fetch_retry(&stats->syncp, start));
|
2010-06-23 01:22:17 +08:00
|
|
|
*
|
|
|
|
*
|
|
|
|
* Example of use in drivers/net/loopback.c, using per_cpu containers,
|
|
|
|
* in BH disabled context.
|
|
|
|
*/
|
|
|
|
#include <linux/seqlock.h>
|
|
|
|
|
|
|
|
struct u64_stats_sync {
|
2010-06-24 08:54:06 +08:00
|
|
|
#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
|
2010-06-23 01:22:17 +08:00
|
|
|
seqcount_t seq;
|
2010-06-24 08:54:06 +08:00
|
|
|
#endif
|
2010-06-23 01:22:17 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
static void inline u64_stats_update_begin(struct u64_stats_sync *syncp)
|
|
|
|
{
|
2010-06-24 08:54:06 +08:00
|
|
|
#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
|
2010-06-23 01:22:17 +08:00
|
|
|
write_seqcount_begin(&syncp->seq);
|
2010-06-24 08:54:06 +08:00
|
|
|
#endif
|
2010-06-23 01:22:17 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void inline u64_stats_update_end(struct u64_stats_sync *syncp)
|
|
|
|
{
|
2010-06-24 08:54:06 +08:00
|
|
|
#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
|
2010-06-23 01:22:17 +08:00
|
|
|
write_seqcount_end(&syncp->seq);
|
2010-06-24 08:54:06 +08:00
|
|
|
#endif
|
2010-06-23 01:22:17 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static unsigned int inline u64_stats_fetch_begin(const struct u64_stats_sync *syncp)
|
|
|
|
{
|
2010-06-24 08:54:06 +08:00
|
|
|
#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
|
2010-06-23 01:22:17 +08:00
|
|
|
return read_seqcount_begin(&syncp->seq);
|
2010-06-24 08:54:06 +08:00
|
|
|
#else
|
|
|
|
#if BITS_PER_LONG==32
|
|
|
|
preempt_disable();
|
|
|
|
#endif
|
|
|
|
return 0;
|
|
|
|
#endif
|
2010-06-23 01:22:17 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static bool inline u64_stats_fetch_retry(const struct u64_stats_sync *syncp,
|
|
|
|
unsigned int start)
|
|
|
|
{
|
2010-06-24 08:54:06 +08:00
|
|
|
#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
|
2010-06-23 01:22:17 +08:00
|
|
|
return read_seqcount_retry(&syncp->seq, start);
|
|
|
|
#else
|
2010-06-24 08:54:06 +08:00
|
|
|
#if BITS_PER_LONG==32
|
|
|
|
preempt_enable();
|
|
|
|
#endif
|
|
|
|
return false;
|
|
|
|
#endif
|
2010-06-23 01:22:17 +08:00
|
|
|
}
|
|
|
|
|
2010-06-24 08:54:06 +08:00
|
|
|
/*
|
|
|
|
* In case softirq handlers can update u64 counters, readers can use following helpers
|
|
|
|
* - SMP 32bit arches use seqcount protection, irq safe.
|
|
|
|
* - UP 32bit must disable BH.
|
|
|
|
* - 64bit have no problem atomically reading u64 values, irq safe.
|
|
|
|
*/
|
|
|
|
static unsigned int inline u64_stats_fetch_begin_bh(const struct u64_stats_sync *syncp)
|
2010-06-23 01:22:17 +08:00
|
|
|
{
|
2010-06-24 08:54:06 +08:00
|
|
|
#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
|
|
|
|
return read_seqcount_begin(&syncp->seq);
|
|
|
|
#else
|
|
|
|
#if BITS_PER_LONG==32
|
|
|
|
local_bh_disable();
|
|
|
|
#endif
|
2010-06-23 01:22:17 +08:00
|
|
|
return 0;
|
2010-06-24 08:54:06 +08:00
|
|
|
#endif
|
2010-06-23 01:22:17 +08:00
|
|
|
}
|
|
|
|
|
2010-06-24 08:54:06 +08:00
|
|
|
static bool inline u64_stats_fetch_retry_bh(const struct u64_stats_sync *syncp,
|
2010-06-23 01:22:17 +08:00
|
|
|
unsigned int start)
|
|
|
|
{
|
2010-06-24 08:54:06 +08:00
|
|
|
#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
|
|
|
|
return read_seqcount_retry(&syncp->seq, start);
|
|
|
|
#else
|
|
|
|
#if BITS_PER_LONG==32
|
|
|
|
local_bh_enable();
|
|
|
|
#endif
|
2010-06-23 01:22:17 +08:00
|
|
|
return false;
|
|
|
|
#endif
|
2010-06-24 08:54:06 +08:00
|
|
|
}
|
2010-06-23 01:22:17 +08:00
|
|
|
|
|
|
|
#endif /* _LINUX_U64_STATS_SYNC_H */
|