lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date:   Fri, 29 Jul 2022 13:46:38 -0500
From:   Coleman Dietsch <dietschc@....edu>
To:     kvm@...r.kernel.org
Cc:     Coleman Dietsch <dietschc@....edu>,
        Sean Christopherson <seanjc@...gle.com>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        Dave Hansen <dave.hansen@...ux.intel.com>, x86@...nel.org,
        "H . Peter Anvin" <hpa@...or.com>, linux-kernel@...r.kernel.org,
        skhan@...uxfoundation.org, Pavel Skripkin <paskripkin@...il.com>,
        linux-kernel-mentees@...ts.linuxfoundation.org
Subject: [PATCH v2 0/2] KVM: x86/xen: Prevent xen timer init when running

This patch series was created to address the following bug found by syzbot
WARNING: ODEBUG bug in kvm_xen_vcpu_set_attr.

When running the syzbot reproducer code, the following crash dump occurs:

ODEBUG: init active (active state 0)
object type: hrtimer hint: xen_timer_callbac0
RIP: 0010:debug_print_object+0x16e/0x250 lib/debugobjects.c:502
Call Trace:
__debug_object_init
debug_hrtimer_init
debug_init
hrtimer_init
kvm_xen_init_timer
kvm_xen_vcpu_set_attr
kvm_arch_vcpu_ioctl
kvm_vcpu_ioctl
vfs_ioctl

The ODEBUG bug crash appears to be happening when vcpu->arch.xen.timer is
already set and kvm_xen_init_timer() is called, which appears to be the
result of two separate issues.

The first issue is that kvm_xen_init_timer() is run "every"
KVM_XEN_VCPU_ATTR_TYPE_TIMER. This is addressed in patch 1.

The second issue is that the stop xen timer code should be run before
changing the IRQ vector. This is addressed in patch 2 with some cleanup.

version 2 changes (mostly feedback from Sean Christopherson)
-split patch into 2 patches
-fix changelogs to be more descriptive
-fix formatting issues
-add check for existing xen timer before trying to initialize another one
-removed conditional for kvm_xen_stop_timer() so that it always runs
-ensure that xen timer is stopped before changing IRQ vector
-streamlined switch case KVM_XEN_VCPU_ATTR_TYPE_TIMER a bit

Coleman Dietsch (2):
  KVM: x86/xen: Initialize Xen timer only once
  KVM: x86/xen: Stop Xen timer before changing the IRQ vector

 arch/x86/kvm/xen.c | 32 +++++++++++++++++---------------
 1 file changed, 17 insertions(+), 15 deletions(-)

-- 
2.34.1

Powered by blists - more mailing lists