android_kernel_xiaomi_sm8350/kernel
Ingo Molnar 03cbc358aa [PATCH] lockdep core: improve the lock-chain-hash
With CONFIG_DEBUG_LOCK_ALLOC turned off i was getting sporadic failures in
the locking self-test:

  ------------>
  | Locking API testsuite:
  ----------------------------------------------------------------------------
                                   | spin |wlock |rlock |mutex | wsem | rsem |
    --------------------------------------------------------------------------
                       A-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
                   A-B-B-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
               A-B-B-C-C-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
               A-B-C-A-B-C deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
           A-B-B-C-C-D-D-A deadlock:  ok  |FAILED|  ok  |  ok  |  ok  |  ok  |
           A-B-C-D-B-D-D-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
           A-B-C-D-B-C-D-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |FAILED|

after much debugging it turned out to be caused by accidental chain-hash
key collisions.  The current hash is:

 #define iterate_chain_key(key1, key2) \
	(((key1) << MAX_LOCKDEP_KEYS_BITS/2) ^ \
	((key1) >> (64-MAX_LOCKDEP_KEYS_BITS/2)) ^ \
 	(key2))

where MAX_LOCKDEP_KEYS_BITS is 11.  This hash is pretty good as it will
shift by 5 bits in every iteration, where every new ID 'mixed' into the
hash would have up to 11 bits.  But because there was a 6 bits overlap
between subsequent IDs and their high bits tended to be similar, there was
a chance for accidental chain-hash collision for a low number of locks
held.

the solution is to shift by 11 bits:

 #define iterate_chain_key(key1, key2) \
	(((key1) << MAX_LOCKDEP_KEYS_BITS) ^ \
	((key1) >> (64-MAX_LOCKDEP_KEYS_BITS)) ^ \
 	(key2))

This keeps the hash perfect up to 5 locks held, but even above that the
hash is still good because 11 bits is a relative prime to the total 64
bits, so a complete match will only occur after 64 held locks (which doesnt
happen in Linux).  Even after 5 locks held, entropy of the 5 IDs mixed into
the hash is already good enough so that overlap doesnt generate a colliding
hash ID.

with this change the false positives went away.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 09:18:25 -07:00
..
irq [PATCH] irq: remove a extra line 2006-09-29 09:18:07 -07:00
power Merge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6 2006-09-26 11:49:46 -07:00
time
.gitignore
acct.c [PATCH] audit/accounting: tty locking 2006-09-29 09:18:25 -07:00
audit.c [PATCH] selinux: rename selinux_ctxid_to_string 2006-09-26 08:48:52 -07:00
audit.h [PATCH] audit: AUDIT_PERM support 2006-09-11 13:32:30 -04:00
auditfilter.c [PATCH] selinux: rename selinux_ctxid_to_string 2006-09-26 08:48:52 -07:00
auditsc.c [PATCH] audit/accounting: tty locking 2006-09-29 09:18:25 -07:00
capability.c [PATCH] pidspace: is_init() 2006-09-29 09:18:12 -07:00
compat.c [PATCH] posix-timers: Fix clock_nanosleep() doesn't return the remaining time in compatibility mode 2006-09-29 09:18:15 -07:00
configs.c
cpu.c [PATCH] Disable CPU hotplug during suspend 2006-09-26 08:48:59 -07:00
cpuset.c [PATCH] cpuset: hotunplug cpus and mems in all cpusets 2006-09-29 09:18:21 -07:00
delayacct.c [PATCH] task delay accounting fixes 2006-09-01 11:39:08 -07:00
dma.c
exec_domain.c
exit.c [PATCH] introduce TASK_DEAD state 2006-09-29 09:18:21 -07:00
extable.c
fork.c [PATCH] copy_process: cosmetic ->ioprio tweak 2006-09-29 09:18:18 -07:00
futex_compat.c [PATCH] futex: Apply recent futex fixes to futex_compat 2006-08-06 08:57:49 -07:00
futex.c [PATCH] sys_get_robust_list(): don't take tasklist_lock 2006-09-29 09:18:18 -07:00
hrtimer.c [PATCH] posix-timers: Fix clock_nanosleep() doesn't return the remaining time in compatibility mode 2006-09-29 09:18:15 -07:00
itimer.c
kallsyms.c [PATCH] null-terminate over-long /proc/kallsyms symbols 2006-07-14 21:53:52 -07:00
Kconfig.hz
Kconfig.preempt
kexec.c [PATCH] kexec warning fix 2006-09-29 09:18:15 -07:00
kfifo.c [PATCH] memory ordering in __kfifo primitives 2006-09-29 09:18:13 -07:00
kmod.c [PATCH] Fix ____call_usermodehelper errors being silently ignored 2006-09-29 09:18:16 -07:00
kprobes.c [PATCH] IA64: kprobe invalidate icache of jump buffer 2006-07-31 13:28:38 -07:00
ksysfs.c
kthread.c [PATCH] remove kernel/kthread.c:kthread_stop_sem() 2006-07-14 21:53:52 -07:00
lockdep_internals.h [PATCH] lockdep: double the number of stack-trace entries 2006-09-13 07:32:14 -07:00
lockdep_proc.c
lockdep.c [PATCH] lockdep core: improve the lock-chain-hash 2006-09-29 09:18:25 -07:00
Makefile [PATCH] per-task-delay-accounting: taskstats interface 2006-07-14 21:53:56 -07:00
module.c [PATCH] /sys/modules: allow full length section names 2006-09-29 09:18:23 -07:00
mutex-debug.c [PATCH] lockdep: prove mutex locking correctness 2006-07-03 15:27:04 -07:00
mutex-debug.h
mutex.c [PATCH] lockdep: prove mutex locking correctness 2006-07-03 15:27:04 -07:00
mutex.h [PATCH] lockdep: prove mutex locking correctness 2006-07-03 15:27:04 -07:00
panic.c [PATCH] Add the __stack_chk_fail() function 2006-09-26 10:52:39 +02:00
params.c [PATCH] module_subsys: initialize earlier 2006-09-29 09:18:08 -07:00
pid.c [PATCH] pid: remove temporary debug code in attach_pid 2006-09-27 08:26:19 -07:00
posix-cpu-timers.c [PATCH] posix-timers: Fix the flags handling in posix_cpu_nsleep() 2006-09-29 09:18:15 -07:00
posix-timers.c [PATCH] posix-timers: Fix clock_nanosleep() doesn't return the remaining time in compatibility mode 2006-09-29 09:18:15 -07:00
printk.c [PATCH] PM: make it possible to disable console suspending 2006-09-26 08:49:03 -07:00
profile.c [PATCH] Profiling: require buffer allocation on the correct node 2006-09-26 08:48:50 -07:00
ptrace.c [PATCH] pidspace: is_init() 2006-09-29 09:18:12 -07:00
rcupdate.c [PATCH] rcu_do_batch: make ->qlen decrement irq safe 2006-09-13 07:32:14 -07:00
rcutorture.c [PATCH] rcu: add lock annotations to rcu{,_bh}_torture_read_{lock,unlock} 2006-09-29 09:18:08 -07:00
relay.c [PATCH] kernel-doc for relay interface 2006-09-29 09:18:06 -07:00
resource.c Resources: insert identical resources above existing resources 2006-09-26 17:43:52 -07:00
rtmutex_common.h
rtmutex-debug.c [PATCH] sched: cleanup, remove task_t, convert to struct task_struct 2006-07-03 15:27:11 -07:00
rtmutex-debug.h
rtmutex-tester.c [PATCH] Add try_to_freeze() to rt-test kthreads 2006-07-14 21:53:53 -07:00
rtmutex.c [PATCH] clean up and remove some extra spinlocks from rtmutex 2006-09-29 09:18:09 -07:00
rtmutex.h
rwsem.c [PATCH] lockdep: prove rwsem locking correctness 2006-07-03 15:27:04 -07:00
sched.c [PATCH] introduce TASK_DEAD state 2006-09-29 09:18:21 -07:00
seccomp.c
signal.c [PATCH] __dequeue_signal() cleanup 2006-09-29 09:18:15 -07:00
softirq.c [PATCH] check return value of cpu_callback 2006-09-29 09:18:14 -07:00
softlockup.c [PATCH] check return value of cpu_callback 2006-09-29 09:18:14 -07:00
spinlock.c [PATCH] remove generic__raw_read_trylock() 2006-09-29 09:18:03 -07:00
stacktrace.c
stop_machine.c [PATCH] stop_machine.c copyright 2006-09-29 09:18:24 -07:00
sys_ni.c
sys.c [PATCH] kill extraneous printk in kernel_restart() 2006-09-29 09:18:16 -07:00
sysctl.c [PATCH] pidspace: is_init() 2006-09-29 09:18:12 -07:00
taskstats.c [NETLINK]: Extend netlink messaging interface 2006-09-22 14:53:43 -07:00
time.c
timer.c [PATCH] simplify update_times (avoid jiffies/jiffies_64 aliasing problem) 2006-09-29 09:18:15 -07:00
uid16.c
unwind.c [PATCH] unwind: fix unused variable warning when !CONFIG_MODULES 2006-09-29 09:18:11 -07:00
user.c
wait.c [PATCH] uninline init_waitqueue_head() 2006-07-10 13:24:25 -07:00
workqueue.c [PATCH] workqueue: remove lock_cpu_hotplug() 2006-08-14 12:54:29 -07:00