Re: 2.6.0-test9 and sleeping function called from invalid context

[Manfred Spraul <manfred@colorfullife.com>]

Andrew Morton wrote:

>What protects against concurrent execution of proc_pid_lookup() and
>proc_task_lookup()?  I think nothing, because one is at /proc/42 and the
>other is at /proc/41/42; the parent dir inodes are different.  hmm.
>  
>
Ugs.
/proc/1 and /proc/1/task/1 are two different dentrys.
proc_task_lookup happily overwrites task->proc_dentry. Which means the 
task patch broke tsk->proc_dentry.
I think the cure is simple: proc_task_lookup should not write 
proc_dentry, only proc_pid_lookup should do that.
tsk->proc_dentry is only used by proc_pid_flush: If a task exits, all 
entries below /proc/<pid> are stale, and a shrink_dcache_parent on the 
/proc/<pid> dentry recovers the memory.
There is a race between in proc_pid_lookup between checking that the 
task is still running and setting tsk->proc_dentry, but AFAICS the race 
is not critical: In the worst case, the stale dentries remain around. 
They are never returned to user space, d_revalidate prevents that.

--
    Manfred