lkml.org 
[lkml]   [2012]   [Sep]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH] lockdep: Check if nested lock is actually held
From
Date
On Thu, 2012-09-13 at 11:39 +0200, Maarten Lankhorst wrote:
> It is considered good form to lock the lock you claim to be nested in.

Uhm yeah.. cute. You actually found a site where this triggered?

> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
> ---
> diff --git a/kernel/lockdep.c b/kernel/lockdep.c
> index ea9ee45..7175447 100644
> --- a/kernel/lockdep.c
> +++ b/kernel/lockdep.c
> @@ -2998,6 +2998,43 @@ EXPORT_SYMBOL_GPL(lockdep_init_map);
>
> struct lock_class_key __lockdep_no_validate__;
>
> +static int
> +print_lock_nested_lock_not_held(struct task_struct *curr,
> + struct held_lock *lock,
> + struct lockdep_map *nest,
> + unsigned long ip)
> +{
> + if (!debug_locks_off())
> + return 0;
> + if (debug_locks_silent)
> + return 0;
> +
> + printk("\n");
> + printk("==================================\n");
> + printk("[ BUG: Nested lock was not taken ]\n");
> + print_kernel_ident();
> + printk("----------------------------------\n");
> +
> + printk("%s/%d is trying to lock:\n", curr->comm, task_pid_nr(curr));
> + print_lock(lock);
> +
> + printk("\nbut this task is not holding:\n");
> + printk("%s\n", nest->name);
> +
> + printk("\nstack backtrace:\n");
> + dump_stack();
> +
> + printk("\nother info that might help us debug this:\n");
> + lockdep_print_held_locks(curr);
> +
> + printk("\nstack backtrace:\n");
> + dump_stack();
> +
> + return 0;
> +}
> +
> +static int __lock_is_held(struct lockdep_map *lock);
> +
> /*
> * This gets called for every mutex_lock*()/spin_lock*() operation.
> * We maintain the dependency maps and validate the locking attempt:
> @@ -3139,6 +3176,10 @@ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass,
> }
> chain_key = iterate_chain_key(chain_key, id);
>
> + if (nest_lock && !__lock_is_held(nest_lock))
> + return print_lock_nested_lock_not_held(curr, hlock,
> + nest_lock, ip);

At this time we've already set hlock->nest_lock, so I've shortened the
argument list here a little.

> if (!validate_chain(curr, lock, hlock, chain_head, chain_key))
> return 0;
>
>



\
 
 \ /
  Last update: 2012-09-13 12:41    [W:0.042 / U:8.928 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site