Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752962AbbDFXpn (ORCPT ); Mon, 6 Apr 2015 19:45:43 -0400 Received: from userp1040.oracle.com ([156.151.31.81]:44061 "EHLO userp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752673AbbDFXpk (ORCPT ); Mon, 6 Apr 2015 19:45:40 -0400 Message-ID: <55231A82.6010505@oracle.com> Date: Mon, 06 Apr 2015 19:45:06 -0400 From: Sasha Levin User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.5.0 MIME-Version: 1.0 To: Borislav Petkov CC: Linus Torvalds , LKML , Rusty Russell , Dave Jones , Michal Hocko , the arch/x86 maintainers Subject: Re: Hang on large copy_from_user with PREEMPT_NONE References: <552204B8.40605@oracle.com> <5522D9A8.5080104@oracle.com> <20150406193622.GF4078@pd.tnic> In-Reply-To: <20150406193622.GF4078@pd.tnic> Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-Source-IP: userv0021.oracle.com [156.151.31.71] Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4296 Lines: 72 On 04/06/2015 03:36 PM, Borislav Petkov wrote: > On Mon, Apr 06, 2015 at 03:08:24PM -0400, Sasha Levin wrote: >> > Your patch just makes it hang in memset instead: >> > >> > [ 963.104556] NMI watchdog: BUG: soft lockup - CPU#1 stuck for 22s! [trinity-c224:9845] >> > [ 963.104556] Modules linked in: >> > [ 963.104556] irq event stamp: 3773324 >> > [ 963.104556] hardirqs last enabled at (3773323): [] restore_args+0x0/0x34 >> > [ 963.104556] hardirqs last disabled at (3773324): [] apic_timer_interrupt+0x6e/0x80 >> > [ 963.104556] softirqs last enabled at (3773322): [] __do_softirq+0x709/0xd40 >> > [ 963.104556] softirqs last disabled at (3773317): [] irq_exit+0x29d/0x320 >> > [ 963.104556] CPU: 1 PID: 9845 Comm: trinity-c224 Not tainted 4.0.0-rc6-next-20150402-sasha-00039-ge0bdae3-dirty #2130 >> > [ 963.104556] task: ffff8802a3560000 ti: ffff8802a3568000 task.ti: ffff8802a3568000 >> > [ 963.104556] RIP: 0010:[] [] memset_orig+0x33/0xb0 > memset_orig? > > Hmmm, your box doesn't set X86_FEATURE_REP_GOOD. And looking at your > dmesg I can see why - that's a kvm guest and you're emulating some qemu > -cpu kvm64 or so thing, AFAICT. > > Can you give me your full command line? I'm running kvmtool rather than qemu. I've changed the cpu that the guest sees, and it just moved to the optimized memset: [ 843.403237] NMI watchdog: BUG: soft lockup - CPU#0 stuck for 22s! [trinity-c123:9780] [ 843.403265] Modules linked in: [ 843.403265] irq event stamp: 7340472 [ 843.403265] hardirqs last enabled at (7340471): [] restore_args+0x0/0x34 [ 843.403265] hardirqs last disabled at (7340472): [] apic_timer_interrupt+0x6e/0x80 [ 843.403265] softirqs last enabled at (7340470): [] __do_softirq+0x709/0xd40 [ 843.403265] softirqs last disabled at (7340455): [] irq_exit+0x29d/0x320 [ 843.403265] CPU: 0 PID: 9780 Comm: trinity-c123 Not tainted 4.0.0-rc6-next-20150402-sasha-00039-ge0bda e3-dirty #2130 [ 843.403265] task: ffff88034ffb0000 ti: ffff88034ffb8000 task.ti: ffff88034ffb8000 [ 843.403265] RIP: 0010:[] [] __memset+0x24/0x30 [ 843.403265] RSP: 0000:ffff88034ffbfdf0 EFLAGS: 00010206 [ 843.403265] RAX: 0000000000000000 RBX: 00000000fffff000 RCX: 0000000004a8a000 [ 843.403265] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffc900ecdbd000 [ 843.403265] RBP: ffff88034ffbfe18 R08: 000000001ffffe00 R09: ffffc9001220e000 [ 843.403265] R10: ffffc9011220cfff R11: fffff52022441a00 R12: 000000001ffffe00 [ 843.403265] R13: fffff52022399988 R14: 000000001ffffe00 R15: fffff520224419ff [ 843.403265] FS: 00007f0600ec4700(0000) GS:ffff880050600000(0000) knlGS:0000000000000000 [ 843.403265] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [ 843.403265] CR2: 00000000ff355000 CR3: 000000034f73f000 CR4: 00000000000007f0 [ 843.403265] Stack: [ 843.403265] ffffffffa563b4ac ffffffffafafd69c 00000000fffff000 ffff88034ffbfe78 [ 843.403265] ffff88034ffbff18 ffff88034ffbfe38 ffffffffa6fbd759 0000000000010246 [ 843.403265] 1ffff10069ff7fcb ffff88034ffbff48 ffffffffa53c5ef0 ffffc9001220e000 [ 843.403265] Call Trace: [ 843.403265] [] ? memset+0x2c/0x40 [ 843.403265] [] ? bad_to_user+0x66/0x1391 [ 843.403265] [] copy_user_handle_tail+0x69/0x80 [ 843.403265] [] SyS_init_module+0x150/0x210 [ 843.403265] [] ? load_module+0x96a0/0x96a0 [ 843.403265] [] ia32_do_call+0x13/0x13 [ 843.403265] Code: 90 90 90 90 90 90 90 66 66 90 66 90 49 89 f9 48 89 d1 83 e2 07 48 c1 e9 03 40 0f b6 f6 48 b8 01 01 01 01 01 01 01 01 48 0f af c6 48 ab 89 d1 f3 aa 4c 89 c8 c3 90 49 89 f9 40 88 f0 48 89 d1 > > Btw, maybe you should try to trigger this on real metal, just in case, > to rule out virt issues. Or have you, already? I'll give it a go tomorrow. Thanks, Sasha -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/