Description: Linux kernel race condition with PTRACE_SETREGS and fatal signal
References:
 http://seclists.org/oss-sec/2013/q1/326
Notes:
 There is a similar x86-specific ptrace race fixed by 848e8f5f0ad3169560c516fff6471be65f76e69f, 95cf00fa5d5e2a200a2c044c84bde8389a237e02.
 The fix for this issue at least textually depends on that.
Bugs:
upstream: released (3.8-rc5) [910ffdb18a6408e14febbb6e4b6840fd2c928c82, 9899d11f654474d2d54ea52ceaa2a1f4db3abd68, 9067ac85d533651b98c2ff903182a20cbb361fcb]
2.6.32-upstream-stable: released (2.6.32.61)
sid: released (3.2.39-1)
2.6.32-squeeze-security: released (2.6.32-48squeeze2) [bugfix/all/ptrace-ptrace_resume-shouldnt-wake-up-TASK_TRACED-thread.patch, bugfix/all/ptrace-introduce-signal_wake_up_state-and-ptrace_signal_wake_up.patch, bugfix/all/ptrace-ensure-arch_ptrace-ptrace_request-can-never-race-with-SIGKILL.patch, bugfix/all/wake_up_process-should-be-never-used-to-wakeup-a-TASK_STOPPED-TRACED-task.patch, bugfix/all/fix-ptrace-when-task-is-in-task_is_stopped-state.patch]
3.2-upstream-stable: released (3.2.39) [ptrace-introduce-signal_wake_up_state-and-ptrace_signal_wake_up.patch, ptrace-ensure-arch_ptrace-ptrace_request-can-never-race-with.patch, wake_up_process-should-be-never-used-to-wakeup-a.patch]
3.2-wheezy-security: released (3.2.39-1)