From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Biggers Subject: [PATCH] lib/mpi: call cond_resched() from mpi_powm() loop Date: Mon, 6 Nov 2017 22:19:51 -0800 Message-ID: <20171107061951.861-1-ebiggers3@gmail.com> Cc: Tudor-Dan Ambarus , Mat Martineau , Salvatore Benedetto , keyrings@vger.kernel.org, linux-kernel@vger.kernel.org, Eric Biggers , stable@vger.kernel.org To: linux-crypto@vger.kernel.org, Herbert Xu Return-path: Received: from mail-pg0-f68.google.com ([74.125.83.68]:57253 "EHLO mail-pg0-f68.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754242AbdKGGVn (ORCPT ); Tue, 7 Nov 2017 01:21:43 -0500 Sender: linux-crypto-owner@vger.kernel.org List-ID: From: Eric Biggers On a non-preemptible kernel, if KEYCTL_DH_COMPUTE is called with the largest permitted inputs (16384 bits), the kernel spends 10+ seconds doing modular exponentiation in mpi_powm() without rescheduling. If all threads do it, it locks up the system. Moreover, it can cause rcu_sched-stall warnings. Notwithstanding the insanity of doing this calculation in kernel mode rather than in userspace, fix it by calling cond_resched() as each bit from the exponent is processed. It's still noninterruptible, but at least it's preemptible now. Cc: stable@vger.kernel.org # v4.12+ Signed-off-by: Eric Biggers --- lib/mpi/mpi-pow.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/lib/mpi/mpi-pow.c b/lib/mpi/mpi-pow.c index e24388a863a7..f089a52dbbdb 100644 --- a/lib/mpi/mpi-pow.c +++ b/lib/mpi/mpi-pow.c @@ -26,6 +26,7 @@ * however I decided to publish this code under the plain GPL. */ +#include #include #include "mpi-internal.h" #include "longlong.h" @@ -263,6 +264,8 @@ int mpi_powm(MPI res, MPI base, MPI exp, MPI mod) break; e = ep[i]; c = BITS_PER_MPI_LIMB; + + cond_resched(); } /* We shifted MOD, the modulo reduction argument, left MOD_SHIFT_CNT -- 2.15.0