ptrace: the killed tracee should not enter the syscall

Another old/known problem.  If the tracee is killed after it reports
syscall_entry, it starts the syscall and debugger can't control this.
This confuses the users and this creates the security problems for
ptrace jailers.

Change tracehook_report_syscall_entry() to return non-zero if killed,
this instructs syscall_trace_enter() to abort the syscall.

Reported-by: Chris Evans <[email protected]>
Tested-by: Indan Zupancic <[email protected]>
Signed-off-by: Oleg Nesterov <[email protected]>
Cc: Denys Vlasenko <[email protected]>
Cc: Tejun Heo <[email protected]>
Cc: Pedro Alves <[email protected]>
Cc: Jan Kratochvil <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

0 files changed, 0 insertions, 0 deletions