linux-kernel - Re: [PATCH] KVM: x86: fix missed hardware breakpoints

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-prev] [thread-next>] [day] [month] [year] [list]

Date:	Fri, 26 Feb 2016 19:42:36 +0800
From:	Xiao Guangrong <guangrong.xiao@...ux.intel.com>
To:	Nadav Amit <nadav.amit@...il.com>
Cc:	Paolo Bonzini <pbonzini@...hat.com>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	kvm <kvm@...r.kernel.org>, oleg@...hat.com,
	Nadav Amit <namit@...technion.ac.il>, avagin@...il.com,
	stable@...r.kernel.org
Subject: Re: [PATCH] KVM: x86: fix missed hardware breakpoints



On 02/26/2016 07:28 PM, Nadav Amit wrote:
> Xiao Guangrong <guangrong.xiao@...ux.intel.com> wrote:
>
>> On 02/19/2016 06:56 PM, Paolo Bonzini wrote:
>>> Sometimes when setting a breakpoint a process doesn't stop on it.
>>> This is because the debug registers are not loaded correctly on
>>> VCPU load.
>>>
>>>
>>> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
>>> index 4244c2baf57d..f4891f2ece23 100644
>>> --- a/arch/x86/kvm/x86.c
>>> +++ b/arch/x86/kvm/x86.c
>>> @@ -2752,6 +2752,7 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
>>>   	}
>>>
>>>   	kvm_make_request(KVM_REQ_STEAL_UPDATE, vcpu);
>>> +	vcpu->arch.switch_db_regs |= KVM_DEBUGREG_RELOAD;
>>
>> Er, i do not understand how it works. The BP is enabled in this test case so
>> the debug registers are always reloaded before entering guest as
>> KVM_DEBUGREG_BP_ENABLED bit is always set on switch_db_regs. What did i miss?
>
> Note that KVM_DEBUGREG_BP_ENABLED does not have to be set, since
> kvm_update_dr7() is not called once the guest has KVM_DEBUGREG_WONT_EXIT.

Ah, yes. So that we will enter guest without reloading debug registers under this
case. Is it safe? What about if gdb/perf attached to QEMU and reset the debug0－３
(in this case, the vcpu is interrupted by IPI and it is not rescheduled)？