rseq: Avoid infinite recursion when delivering SIGSEGV
authorWill Deacon <will.deacon@arm.com>
Fri, 22 Jun 2018 10:45:07 +0000 (11:45 +0100)
committerThomas Gleixner <tglx@linutronix.de>
Fri, 22 Jun 2018 17:04:22 +0000 (19:04 +0200)
commit784e0300fe9fe4aa81bd7df9d59e138f56bb605b
tree442642dae3c8e1b50ab68460c2e7662684d9e194
parent9a789fcfe8605417f7a1a970355f5efa4fe88c64
rseq: Avoid infinite recursion when delivering SIGSEGV

When delivering a signal to a task that is using rseq, we call into
__rseq_handle_notify_resume() so that the registers pushed in the
sigframe are updated to reflect the state of the restartable sequence
(for example, ensuring that the signal returns to the abort handler if
necessary).

However, if the rseq management fails due to an unrecoverable fault when
accessing userspace or certain combinations of RSEQ_CS_* flags, then we
will attempt to deliver a SIGSEGV. This has the potential for infinite
recursion if the rseq code continuously fails on signal delivery.

Avoid this problem by using force_sigsegv() instead of force_sig(), which
is explicitly designed to reset the SEGV handler to SIG_DFL in the case
of a recursive fault. In doing so, remove rseq_signal_deliver() from the
internal rseq API and have an optional struct ksignal * parameter to
rseq_handle_notify_resume() instead.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: peterz@infradead.org
Cc: paulmck@linux.vnet.ibm.com
Cc: boqun.feng@gmail.com
Link: https://lkml.kernel.org/r/1529664307-983-1-git-send-email-will.deacon@arm.com
arch/arm/kernel/signal.c
arch/powerpc/kernel/signal.c
arch/x86/entry/common.c
arch/x86/kernel/signal.c
include/linux/sched.h
kernel/rseq.c