AuxXxilium/linux_dsm_epyc7002
1
0
Fork 0
mirror of https://github.com/AuxXxilium/linux_dsm_epyc7002.git synced 2024-12-26 10:15:13 +07:00
Code Issues Actions Packages Projects Releases Wiki Activity
c93764a65b
linux_dsm_epyc7002/kernel/locking/lockdep_internals.h
Waiman Long 810507fe6f locking/lockdep: Reuse freed chain_hlocks entries 
Once a lock class is zapped, all the lock chains that include the zapped
class are essentially useless. The lock_chain structure itself can be
reused, but not the corresponding chain_hlocks[] entries. Over time,
we will run out of chain_hlocks entries while there are still plenty
of other lockdep array entries available.

To fix this imbalance, we have to make chain_hlocks entries reusable
just like the others. As the freed chain_hlocks entries are in blocks of
various lengths. A simple bitmap like the one used in the other reusable
lockdep arrays isn't applicable. Instead the chain_hlocks entries are
put into bucketed lists (MAX_CHAIN_BUCKETS) of chain blocks.  Bucket 0
is the variable size bucket which houses chain blocks of size larger than
MAX_CHAIN_BUCKETS sorted in decreasing size order.  Initially, the whole
array is in one chain block (the primordial chain block) in bucket 0.

The minimum size of a chain block is 2 chain_hlocks entries. That will
be the minimum allocation size. In other word, allocation requests
for one chain_hlocks entry will cause 2-entry block to be returned and
hence 1 entry will be wasted.

Allocation requests for the chain_hlocks are fulfilled first by looking
for chain block of matching size. If not found, the first chain block
from bucket[0] (the largest one) is split. That can cause hlock entries
fragmentation and reduce allocation efficiency if a chain block of size >
MAX_CHAIN_BUCKETS is ever zapped and put back to after the primordial
chain block. So the MAX_CHAIN_BUCKETS must be large enough that this
should seldom happen.

By reusing the chain_hlocks entries, we are able to handle workloads
that add and zap a lot of lock classes without the risk of running out
of chain_hlocks entries as long as the total number of outstanding lock
classes at any time remain within a reasonable limit.

Two new tracking counters, nr_free_chain_hlocks & nr_large_chain_blocks,
are added to track the total number of chain_hlocks entries in the
free bucketed lists and the number of large chain blocks in buckets[0]
respectively. The nr_free_chain_hlocks replaces nr_chain_hlocks.

The nr_large_chain_blocks counter enables to see if we should increase
the number of buckets (MAX_CHAIN_BUCKETS) available so as to avoid to
avoid the fragmentation problem in bucket[0].

An internal nfsd test that ran for more than an hour and kept on
loading and unloading kernel modules could cause the following message
to be displayed.

  [ 4318.443670] BUG: MAX_LOCKDEP_CHAIN_HLOCKS too low!

The patched kernel was able to complete the test with a lot of free
chain_hlocks entries to spare:

  # cat /proc/lockdep_stats
     :
   dependency chains:                   18867 [max: 65536]
   dependency chain hlocks:             74926 [max: 327680]
   dependency chain hlocks lost:            0
     :
   zapped classes:                       1541
   zapped lock chains:                  56765
   large chain blocks:                      1

By changing MAX_CHAIN_BUCKETS to 3 and add a counter for the size of the
largest chain block. The system still worked and We got the following
lockdep_stats data:

   dependency chains:                   18601 [max: 65536]
   dependency chain hlocks used:        73133 [max: 327680]
   dependency chain hlocks lost:            0
     :
   zapped classes:                       1541
   zapped lock chains:                  56702
   large chain blocks:                  45165
   large chain block size:              20165

By running the test again, I was indeed able to cause chain_hlocks
entries to get lost:

   dependency chain hlocks used:        74806 [max: 327680]
   dependency chain hlocks lost:          575
     :
   large chain blocks:                  48737
   large chain block size:                  7

Due to the fragmentation, it is possible that the
"MAX_LOCKDEP_CHAIN_HLOCKS too low!" error can happen even if a lot of
of chain_hlocks entries appear to be free.

Fortunately, a MAX_CHAIN_BUCKETS value of 16 should be big enough that
few variable sized chain blocks, other than the initial one, should
ever be present in bucket 0.

Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lkml.kernel.org/r/20200206152408.24165-7-longman@redhat.com
2020-02-11 13:10:52 +01:00

255 lines
6.9 KiB
C
Raw Blame History

/* SPDX-License-Identifier: GPL-2.0 */
/*
* kernel/lockdep_internals.h
*
* Runtime locking correctness validator
*
* lockdep subsystem internal functions and variables.
*/
/*
* Lock-class usage-state bits:
*/
enum lock_usage_bit {
#define LOCKDEP_STATE(__STATE) \
LOCK_USED_IN_##__STATE, \
LOCK_USED_IN_##__STATE##_READ, \
LOCK_ENABLED_##__STATE, \
LOCK_ENABLED_##__STATE##_READ,
#include "lockdep_states.h"
#undef LOCKDEP_STATE
LOCK_USED,
LOCK_USAGE_STATES
};
#define LOCK_USAGE_READ_MASK 1
#define LOCK_USAGE_DIR_MASK 2
#define LOCK_USAGE_STATE_MASK (~(LOCK_USAGE_READ_MASK | LOCK_USAGE_DIR_MASK))
/*
* Usage-state bitmasks:
*/
#define __LOCKF(__STATE) LOCKF_##__STATE = (1 << LOCK_##__STATE),
enum {
#define LOCKDEP_STATE(__STATE) \
__LOCKF(USED_IN_##__STATE) \
__LOCKF(USED_IN_##__STATE##_READ) \
__LOCKF(ENABLED_##__STATE) \
__LOCKF(ENABLED_##__STATE##_READ)
#include "lockdep_states.h"
#undef LOCKDEP_STATE
__LOCKF(USED)
};
#define LOCKDEP_STATE(__STATE) LOCKF_ENABLED_##__STATE |
static const unsigned long LOCKF_ENABLED_IRQ =
#include "lockdep_states.h"
0;
#undef LOCKDEP_STATE
#define LOCKDEP_STATE(__STATE) LOCKF_USED_IN_##__STATE |
static const unsigned long LOCKF_USED_IN_IRQ =
#include "lockdep_states.h"
0;
#undef LOCKDEP_STATE
#define LOCKDEP_STATE(__STATE) LOCKF_ENABLED_##__STATE##_READ |
static const unsigned long LOCKF_ENABLED_IRQ_READ =
#include "lockdep_states.h"
0;
#undef LOCKDEP_STATE
#define LOCKDEP_STATE(__STATE) LOCKF_USED_IN_##__STATE##_READ |
static const unsigned long LOCKF_USED_IN_IRQ_READ =
#include "lockdep_states.h"
0;
#undef LOCKDEP_STATE
#define LOCKF_ENABLED_IRQ_ALL (LOCKF_ENABLED_IRQ | LOCKF_ENABLED_IRQ_READ)
#define LOCKF_USED_IN_IRQ_ALL (LOCKF_USED_IN_IRQ | LOCKF_USED_IN_IRQ_READ)
#define LOCKF_IRQ (LOCKF_ENABLED_IRQ | LOCKF_USED_IN_IRQ)
#define LOCKF_IRQ_READ (LOCKF_ENABLED_IRQ_READ | LOCKF_USED_IN_IRQ_READ)
/*
* CONFIG_LOCKDEP_SMALL is defined for sparc. Sparc requires .text,
* .data and .bss to fit in required 32MB limit for the kernel. With
* CONFIG_LOCKDEP we could go over this limit and cause system boot-up problems.
* So, reduce the static allocations for lockdeps related structures so that
* everything fits in current required size limit.
*/
#ifdef CONFIG_LOCKDEP_SMALL
/*
* MAX_LOCKDEP_ENTRIES is the maximum number of lock dependencies
* we track.
*
* We use the per-lock dependency maps in two ways: we grow it by adding
* every to-be-taken lock to all currently held lock's own dependency
* table (if it's not there yet), and we check it for lock order
* conflicts and deadlocks.
*/
#define MAX_LOCKDEP_ENTRIES 16384UL
#define MAX_LOCKDEP_CHAINS_BITS 15
#define MAX_STACK_TRACE_ENTRIES 262144UL
#define STACK_TRACE_HASH_SIZE 8192
#else
#define MAX_LOCKDEP_ENTRIES 32768UL
#define MAX_LOCKDEP_CHAINS_BITS 16
/*
* Stack-trace: tightly packed array of stack backtrace
* addresses. Protected by the hash_lock.
*/
#define MAX_STACK_TRACE_ENTRIES 524288UL
#define STACK_TRACE_HASH_SIZE 16384
#endif
/*
* Bit definitions for lock_chain.irq_context
*/
#define LOCK_CHAIN_SOFTIRQ_CONTEXT (1 << 0)
#define LOCK_CHAIN_HARDIRQ_CONTEXT (1 << 1)
#define MAX_LOCKDEP_CHAINS (1UL << MAX_LOCKDEP_CHAINS_BITS)
#define MAX_LOCKDEP_CHAIN_HLOCKS (MAX_LOCKDEP_CHAINS*5)
extern struct list_head all_lock_classes;
extern struct lock_chain lock_chains[];
#define LOCK_USAGE_CHARS (1+LOCK_USAGE_STATES/2)
extern void get_usage_chars(struct lock_class *class,
char usage[LOCK_USAGE_CHARS]);
extern const char *__get_key_name(const struct lockdep_subclass_key *key,
char *str);
struct lock_class *lock_chain_get_class(struct lock_chain *chain, int i);
extern unsigned long nr_lock_classes;
extern unsigned long nr_zapped_classes;
extern unsigned long nr_zapped_lock_chains;
extern unsigned long nr_list_entries;
long lockdep_next_lockchain(long i);
unsigned long lock_chain_count(void);
extern unsigned long nr_stack_trace_entries;
extern unsigned int nr_hardirq_chains;
extern unsigned int nr_softirq_chains;
extern unsigned int nr_process_chains;
extern unsigned int nr_free_chain_hlocks;
extern unsigned int nr_lost_chain_hlocks;
extern unsigned int nr_large_chain_blocks;
extern unsigned int max_lockdep_depth;
extern unsigned int max_bfs_queue_depth;
#ifdef CONFIG_PROVE_LOCKING
extern unsigned long lockdep_count_forward_deps(struct lock_class *);
extern unsigned long lockdep_count_backward_deps(struct lock_class *);
#ifdef CONFIG_TRACE_IRQFLAGS
u64 lockdep_stack_trace_count(void);
u64 lockdep_stack_hash_count(void);
#endif
#else
static inline unsigned long
lockdep_count_forward_deps(struct lock_class *class)
{
return 0;
}
static inline unsigned long
lockdep_count_backward_deps(struct lock_class *class)
{
return 0;
}
#endif
#ifdef CONFIG_DEBUG_LOCKDEP
#include <asm/local.h>
/*
* Various lockdep statistics.
* We want them per cpu as they are often accessed in fast path
* and we want to avoid too much cache bouncing.
*/
struct lockdep_stats {
unsigned long chain_lookup_hits;
unsigned int chain_lookup_misses;
unsigned long hardirqs_on_events;
unsigned long hardirqs_off_events;
unsigned long redundant_hardirqs_on;
unsigned long redundant_hardirqs_off;
unsigned long softirqs_on_events;
unsigned long softirqs_off_events;
unsigned long redundant_softirqs_on;
unsigned long redundant_softirqs_off;
int nr_unused_locks;
unsigned int nr_redundant_checks;
unsigned int nr_redundant;
unsigned int nr_cyclic_checks;
unsigned int nr_find_usage_forwards_checks;
unsigned int nr_find_usage_backwards_checks;
/*
* Per lock class locking operation stat counts
*/
unsigned long lock_class_ops[MAX_LOCKDEP_KEYS];
};
DECLARE_PER_CPU(struct lockdep_stats, lockdep_stats);
extern struct lock_class lock_classes[MAX_LOCKDEP_KEYS];
#define __debug_atomic_inc(ptr) \
this_cpu_inc(lockdep_stats.ptr);
#define debug_atomic_inc(ptr) { \
WARN_ON_ONCE(!irqs_disabled()); \
__this_cpu_inc(lockdep_stats.ptr); \
}
#define debug_atomic_dec(ptr) { \
WARN_ON_ONCE(!irqs_disabled()); \
__this_cpu_dec(lockdep_stats.ptr); \
}
#define debug_atomic_read(ptr) ({ \
struct lockdep_stats *__cpu_lockdep_stats; \
unsigned long long __total = 0; \
int __cpu; \
for_each_possible_cpu(__cpu) { \
__cpu_lockdep_stats = &per_cpu(lockdep_stats, __cpu); \
__total += __cpu_lockdep_stats->ptr; \
} \
__total; \
})
static inline void debug_class_ops_inc(struct lock_class *class)
{
int idx;
idx = class - lock_classes;
__debug_atomic_inc(lock_class_ops[idx]);
}
static inline unsigned long debug_class_ops_read(struct lock_class *class)
{
int idx, cpu;
unsigned long ops = 0;
idx = class - lock_classes;
for_each_possible_cpu(cpu)
ops += per_cpu(lockdep_stats.lock_class_ops[idx], cpu);
return ops;
}
#else
# define __debug_atomic_inc(ptr) do { } while (0)
# define debug_atomic_inc(ptr) do { } while (0)
# define debug_atomic_dec(ptr) do { } while (0)
# define debug_atomic_read(ptr) 0
# define debug_class_ops_inc(ptr) do { } while (0)
#endif
Reference in New Issue View Git Blame Copy Permalink