From mboxrd@z Thu Jan 1 00:00:00 1970 From: Xiao Guangrong Subject: Re: index-out-of-range ubsan warnings Date: Tue, 23 Feb 2016 18:44:03 +0800 Message-ID: <56CC37F3.3060308@linux.intel.com> References: <20160223092643.GB4396@kmu-ThinkPad-X230> <56CC2F7B.1060607@siemens.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit Cc: Marcelo Tosatti , Paolo Bonzini To: Jan Kiszka , Mike Krinkin , kvm@vger.kernel.org Return-path: Received: from mga11.intel.com ([192.55.52.93]:52273 "EHLO mga11.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752931AbcBWKwO (ORCPT ); Tue, 23 Feb 2016 05:52:14 -0500 In-Reply-To: <56CC2F7B.1060607@siemens.com> Sender: kvm-owner@vger.kernel.org List-ID: Mike, thank you for the report. On 02/23/2016 06:07 PM, Jan Kiszka wrote: > On 2016-02-23 10:26, Mike Krinkin wrote: >> diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c >> index 95a955d..cc673f1 100644 >> --- a/arch/x86/kvm/mmu.c >> +++ b/arch/x86/kvm/mmu.c >> @@ -2008,7 +2008,7 @@ static void kvm_mmu_pages_init(struct kvm_mmu_page *parent, >> struct mmu_page_path *parents, >> struct kvm_mmu_pages *pvec) >> { >> - parents->parent[parent->role.level-1] = NULL; >> + parents->parent[parent->role.level-2] = NULL; >> pvec->nr = 0; >> } >> This is a known issue and i attached a patch to fix it, please refer to: https://groups.google.com/forum/#!msg/syzkaller/4wAzRPswgQ8/sCMTx9RQFQAJ Could you please try it? >> >> The second one is: >> >> [ 168.791851] ================================================================================ >> [ 168.791862] UBSAN: Undefined behaviour in arch/x86/kvm/paging_tmpl.h:252:15 >> [ 168.791866] index 4 is out of range for type 'u64 [4]' >> [ 168.791871] CPU: 0 PID: 2950 Comm: qemu-system-x86 Tainted: G O L 4.5.0-rc5-next-20160222 #7 >> [ 168.791873] Hardware name: LENOVO 23205NG/23205NG, BIOS G2ET95WW (2.55 ) 07/09/2013 >> [ 168.791876] 0000000000000000 ffff8801cfcaf208 ffffffff81c9f780 0000000041b58ab3 >> [ 168.791882] ffffffff82eb2cc1 ffffffff81c9f6b4 ffff8801cfcaf230 ffff8801cfcaf1e0 >> [ 168.791886] 0000000000000004 0000000000000001 0000000000000000 ffffffffa1981600 >> [ 168.791891] Call Trace: >> [ 168.791899] [] dump_stack+0xcc/0x12c >> [ 168.791904] [] ? _atomic_dec_and_lock+0xc4/0xc4 >> [ 168.791910] [] ubsan_epilogue+0xd/0x8a >> [ 168.791914] [] __ubsan_handle_out_of_bounds+0x15c/0x1a3 >> [ 168.791918] [] ? __ubsan_handle_shift_out_of_bounds+0x2bd/0x2bd >> [ 168.791922] [] ? get_user_pages_fast+0x2bf/0x360 >> [ 168.791954] [] ? kvm_largepages_enabled+0x30/0x30 [kvm] >> [ 168.791958] [] ? __get_user_pages_fast+0x360/0x360 >> [ 168.791987] [] paging64_walk_addr_generic+0x1b28/0x2600 [kvm] >> [ 168.792014] [] ? init_kvm_mmu+0x1100/0x1100 [kvm] >> [ 168.792019] [] ? debug_check_no_locks_freed+0x350/0x350 >> [ 168.792044] [] ? init_kvm_mmu+0x1100/0x1100 [kvm] >> [ 168.792076] [] paging64_gva_to_gpa+0x7d/0x110 [kvm] >> [ 168.792121] [] ? paging64_walk_addr_generic+0x2600/0x2600 [kvm] >> [ 168.792130] [] ? debug_lockdep_rcu_enabled+0x7b/0x90 >> [ 168.792178] [] emulator_read_write_onepage+0x27a/0x1150 [kvm] >> [ 168.792208] [] ? __kvm_read_guest_page+0x54/0x70 [kvm] >> [ 168.792234] [] ? kvm_task_switch+0x160/0x160 [kvm] >> [ 168.792238] [] ? debug_lockdep_rcu_enabled+0x7b/0x90 >> [ 168.792263] [] emulator_read_write+0xe7/0x6d0 [kvm] >> [ 168.792290] [] ? em_cr_write+0x230/0x230 [kvm] >> [ 168.792314] [] emulator_write_emulated+0x15/0x20 [kvm] >> [ 168.792340] [] segmented_write+0xf8/0x130 [kvm] >> [ 168.792367] [] ? em_lgdt+0x20/0x20 [kvm] >> [ 168.792374] [] ? vmx_read_guest_seg_ar+0x42/0x1e0 [kvm_intel] >> [ 168.792400] [] writeback+0x3f2/0x700 [kvm] >> [ 168.792424] [] ? em_sidt+0xa0/0xa0 [kvm] >> [ 168.792449] [] ? x86_decode_insn+0x1b3d/0x4f70 [kvm] >> [ 168.792474] [] x86_emulate_insn+0x572/0x3010 [kvm] >> [ 168.792499] [] x86_emulate_instruction+0x3bd/0x2110 [kvm] >> [ 168.792524] [] ? reexecute_instruction.part.110+0x2e0/0x2e0 [kvm] >> [ 168.792532] [] handle_ept_misconfig+0x61/0x460 [kvm_intel] >> [ 168.792539] [] ? handle_pause+0x450/0x450 [kvm_intel] >> [ 168.792546] [] vmx_handle_exit+0xd6a/0x1ad0 [kvm_intel] >> [ 168.792572] [] ? kvm_arch_vcpu_ioctl_run+0xbdc/0x6090 [kvm] >> [ 168.792597] [] kvm_arch_vcpu_ioctl_run+0xd3d/0x6090 [kvm] >> [ 168.792621] [] ? kvm_arch_vcpu_ioctl_run+0xbdc/0x6090 [kvm] >> [ 168.792627] [] ? __ww_mutex_lock_interruptible+0x1630/0x1630 >> [ 168.792651] [] ? kvm_arch_vcpu_runnable+0x4f0/0x4f0 [kvm] >> [ 168.792656] [] ? preempt_notifier_unregister+0x190/0x190 >> [ 168.792681] [] ? kvm_arch_vcpu_load+0x127/0x650 [kvm] >> [ 168.792704] [] kvm_vcpu_ioctl+0x553/0xda0 [kvm] >> [ 168.792727] [] ? vcpu_put+0x40/0x40 [kvm] >> [ 168.792732] [] ? debug_check_no_locks_freed+0x350/0x350 >> [ 168.792735] [] ? _raw_spin_unlock+0x27/0x40 >> [ 168.792740] [] ? handle_mm_fault+0x1673/0x2e40 >> [ 168.792744] [] ? trace_hardirqs_on_caller+0x478/0x6c0 >> [ 168.792747] [] ? trace_hardirqs_on+0xd/0x10 >> [ 168.792751] [] ? debug_lockdep_rcu_enabled+0x7b/0x90 >> [ 168.792756] [] do_vfs_ioctl+0x1b0/0x12b0 >> [ 168.792759] [] ? ioctl_preallocate+0x210/0x210 >> [ 168.792763] [] ? __fget+0x273/0x4a0 >> [ 168.792766] [] ? __fget+0x50/0x4a0 >> [ 168.792770] [] ? __fget_light+0x96/0x2b0 >> [ 168.792773] [] SyS_ioctl+0x79/0x90 >> [ 168.792777] [] entry_SYSCALL_64_fastpath+0x23/0xc1 >> [ 168.792780] ================================================================================ >> >> the following fix solved problem for me: >> >> diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h >> index 6c9fed9..2ce4f05 100644 >> --- a/arch/x86/kvm/paging_tmpl.h >> +++ b/arch/x86/kvm/paging_tmpl.h >> @@ -249,7 +249,7 @@ static int FNAME(update_accessed_dirty_bits)(struct kvm_vcpu *vcpu, >> return ret; >> >> kvm_vcpu_mark_page_dirty(vcpu, table_gfn); >> - walker->ptes[level] = pte; >> + walker->ptes[level - 1] = pte; >> } >> return 0; >> } > This fix looks correct to me. > CC'ing some KVM people with MMU background. Thank you, Jan!