From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============1730302628781912990==" MIME-Version: 1.0 From: Harris, James R Subject: [SPDK] Re: SPDK v19.10/DPDK v19.05.0 eal_adjust_config issue, Illegal instruction causing application to abort Date: Thu, 23 Apr 2020 19:40:34 +0000 Message-ID: <6A9F2DBB-0523-4D26-8D9D-4CA48BEDBFA9@intel.com> In-Reply-To: 20200423175443.2763.39298@ml01.vlan13.01.org List-ID: To: spdk@lists.01.org --===============1730302628781912990== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable That's interesting, but it doesn't make sense to me. DPDK #defines RTE_CPU= _AND directly to CPU_AND on Linux, and CPU_AND is provided by system sched.= h. Are you able to get the exact instruction that this crashes on via gdb?= Not just the instruction, but also the values of whatever registers used = by that instruction. Your original e-mail indicated "illegal operand" in t= he log, so that's a little bit of clue. -Jim =EF=BB=BFOn 4/23/20, 10:56 AM, "asprasannavenkatesh(a)gmail.com" wrote: Hi Jim, Thanks will try out your suggestions. = Before that, I wanted to update you, on the only time I was able to hit= this issue without the -g option, it was crashing in the following macro i= nside the compute_ctrl_threads_cpuset line 1466. (eal_adjust_config -> compute_ctrl_threads_cpuset) = In compute_ctrl_threads_cpuset , line 1466 1466 RTE_CPU_AND(cpuset, cpuset, &default_set); @ 00007f800a86c38f (unknown) @ 00007f800bc0e3dd eal_adjust_config /dpdk/dpdk-src/lib/librte_eal/common/e= al_common_options.c:1466 @ 00007f800bbf4f15 rte_eal_init /dpdk/dpdk-src/lib/librte_eal/linux/ea= l/eal.c:846 @ 00007f800bbd52b6 spdk_env_init /spdk/spdk-src/lib/env_dpdk/init.c:456 _______________________________________________ SPDK mailing list -- spdk(a)lists.01.org To unsubscribe send an email to spdk-leave(a)lists.01.org --===============1730302628781912990==--