bacct_add_tsk: fix unsafe and wrong parent/group_leader dereference
PATCH 2/6 bacct_add_tsk: fix unsafe and wrong parent/group_leader dereference
1. ts = timespec_sub(uptime, current->group_leader->start_time);
It is possible that current != tsk. Probably it was supposed to be 'tsk->group_leader->start_time. But why we are reading group_leader's start_time ? This accounting is per thread, not per procees, I changed this to 'tsk->start_time. Please corect me.
2. stats->ac_ppid = (tsk->parent) ? tsk->parent->pid : 0;
tsk->parent never == NULL, and it is unsafe to dereference it. Both the task and it's parent may exit after the caller unlocks tasklist_lock, the memory could be unmapped (DEBUG_SLAB). (And we should use ->real_parent->tgid in fact).
Q: I don't understand the 'if (thread_group_leader(tsk))' check. Why it is needed ?
kernel/tsacct.c renamed from kernel/tsacct.c~2_par_fix +5 -2
@@ -36,7 +36,7 @@ void bacct_add_tsk(struct taskstats *sta/* calculate task elapsed time in timespec */ do_posix_clock_monotonic_gettime (&uptime); - ts =timespec_sub (uptime, current->group_leader->start_time); + ts =timespec_sub (uptime, tsk->start_time);/* rebase elapsed time to usec */ ac_etime =timespec_to_ns (&ts);do_div (ac_etime,NSEC_PER_USEC );
@@ -58,7 +58,10 @@ void bacct_add_tsk(struct taskstats *sta stats->ac_uid = tsk->uid; stats->ac_gid = tsk->gid; stats->ac_pid = tsk->pid; - stats->ac_ppid = (tsk->parent) ? tsk->parent->pid :0 ; +rcu_read_lock (); + stats->ac_ppid =pid_alive (tsk) ? +rcu_dereference (tsk->real_parent)->tgid :0 ; +rcu_read_unlock (); stats->ac_utime =cputime_to_msecs (tsk->utime) *USEC_PER_MSEC ; stats->ac_stime =cputime_to_msecs (tsk->stime) *USEC_PER_MSEC ; stats->ac_minflt = tsk->min_flt;
Quoted 7 lines · Oleg Nesterov
1. ts =
timespec_sub(uptime, current->group_leader->start_time);It is possible that current != tsk. Probably it was supposed to be '
tsk->group_leader->start_time. But why we are reading group_leader'sstart_time? This accounting is per thread, not per procees, I changed this to 'tsk->start_time. Please corect me.
This is right. Thanks!
Quoted 10 lines
2.
stats->ac_ppid= (tsk->parent) ?tsk->parent->pid: 0;
tsk->parentnever == NULL, and it is unsafe to dereference it. Both the task and it's parent may exit after the caller unlockstasklist_lock, the memory could be unmapped (DEBUG_SLAB). (And we should use ->real_parent->tgid in fact).Q: I don't understand the 'if (
thread_group_leader(tsk))' check. Why it is needed ?
The code was borrowed from kernel/acct.c. The CSA code was extended from the BSD accounting, so we just borrowed the basic accounting stuff from the code.
Thanks,
- jay
Quoted 27 lines
Signed-off-by Oleg Nesterov <[email protected]>kernel/tsacct.c renamed from kernel/tsacct.c~2_par_fix +5
-2@@ -36,7 +36,7 @@ void bacct_add_tsk(struct taskstats *sta/* calculate task elapsed time in timespec */ do_posix_clock_monotonic_gettime (&uptime); - ts =timespec_sub (uptime, current->group_leader->start_time); + ts =timespec_sub (uptime, tsk->start_time);/* rebase elapsed time to usec */ ac_etime =timespec_to_ns (&ts);do_div (ac_etime,NSEC_PER_USEC );@@ -58,7 +58,10 @@ void bacct_add_tsk(struct taskstats *sta stats->ac_uid = tsk->uid; stats->ac_gid = tsk->gid; stats->ac_pid = tsk->pid; - stats->ac_ppid = (tsk->parent) ? tsk->parent->pid :0 ; +rcu_read_lock (); + stats->ac_ppid =pid_alive (tsk) ? +rcu_dereference (tsk->real_parent)->tgid :0 ; +rcu_read_unlock (); stats->ac_utime =cputime_to_msecs (tsk->utime) *USEC_PER_MSEC ; stats->ac_stime =cputime_to_msecs (tsk->stime) *USEC_PER_MSEC ; stats->ac_minflt = tsk->min_flt;