On 3/2/2016 3:37 PM, Peter Zijlstra wrote:
On Wed, Mar 02, 2016 at 03:09:31PM -0500, Chris Metcalf wrote:+void task_isolation_debug(int cpu) +{ + struct task_struct *p; + + if (!task_isolation_possible(cpu)) + return; + + rcu_read_lock(); + p = cpu_curr(cpu); + get_task_struct(p);As I think Oleg keeps reminding me, this is not actually a safe thing to do.
So what's the right solution? The fast path in task_isolation_debug_task basically just uses the new "task_isolation_flags", and "pid" and "comm". I would think those would all have to be safe because of the get_task_struct(). The piece that might be problematic is the eventual call to send_sig_info() using the task_struct pointer (called via task_isolation_debug_task -> task_isolation_interrupt). Clearly this is safe at some level, since that's more or less what sys_kill() does and the process could similarly evaporate half way through sending the signal. Suggestions? Thanks! -- Chris Metcalf, Mellanox Technologies http://www.mellanox.com
