Linux kernel mirror (for testing) git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel os linux

perf: Fix perf_lock_task_context() vs RCU

Jiri managed to trigger this warning:

[] ======================================================
[] [ INFO: possible circular locking dependency detected ]
[] 3.10.0+ #228 Tainted: G W
[] -------------------------------------------------------
[] p/6613 is trying to acquire lock:
[] (rcu_node_0){..-...}, at: [<ffffffff810ca797>] rcu_read_unlock_special+0xa7/0x250
[]
[] but task is already holding lock:
[] (&ctx->lock){-.-...}, at: [<ffffffff810f2879>] perf_lock_task_context+0xd9/0x2c0
[]
[] which lock already depends on the new lock.
[]
[] the existing dependency chain (in reverse order) is:
[]
[] -> #4 (&ctx->lock){-.-...}:
[] -> #3 (&rq->lock){-.-.-.}:
[] -> #2 (&p->pi_lock){-.-.-.}:
[] -> #1 (&rnp->nocb_gp_wq[1]){......}:
[] -> #0 (rcu_node_0){..-...}:

Paul was quick to explain that due to preemptible RCU we cannot call
rcu_read_unlock() while holding scheduler (or nested) locks when part
of the read side critical section was preemptible.

Therefore solve it by making the entire RCU read side non-preemptible.

Also pull out the retry from under the non-preempt to play nice with RT.

Reported-by: Jiri Olsa <jolsa@redhat.com>
Helped-out-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: <stable@kernel.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>

authored by

Peter Zijlstra and committed by
Ingo Molnar
058ebd0e 06f41796

+14 -1
+14 -1
kernel/events/core.c
··· 947 947 { 948 948 struct perf_event_context *ctx; 949 949 950 - rcu_read_lock(); 951 950 retry: 951 + /* 952 + * One of the few rules of preemptible RCU is that one cannot do 953 + * rcu_read_unlock() while holding a scheduler (or nested) lock when 954 + * part of the read side critical section was preemptible -- see 955 + * rcu_read_unlock_special(). 956 + * 957 + * Since ctx->lock nests under rq->lock we must ensure the entire read 958 + * side critical section is non-preemptible. 959 + */ 960 + preempt_disable(); 961 + rcu_read_lock(); 952 962 ctx = rcu_dereference(task->perf_event_ctxp[ctxn]); 953 963 if (ctx) { 954 964 /* ··· 974 964 raw_spin_lock_irqsave(&ctx->lock, *flags); 975 965 if (ctx != rcu_dereference(task->perf_event_ctxp[ctxn])) { 976 966 raw_spin_unlock_irqrestore(&ctx->lock, *flags); 967 + rcu_read_unlock(); 968 + preempt_enable(); 977 969 goto retry; 978 970 } 979 971 ··· 985 973 } 986 974 } 987 975 rcu_read_unlock(); 976 + preempt_enable(); 988 977 return ctx; 989 978 } 990 979