linux-kernel - Re: [PATCH 2/7] KVM: x86: Function for determining exception type

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:	Tue, 22 Jul 2014 10:08:48 +0200
From:	Paolo Bonzini <pbonzini@...hat.com>
To:	Nadav Amit <nadav.amit@...il.com>,
	Nadav Amit <namit@...technion.ac.il>
CC:	tglx@...utronix.de, mingo@...hat.com, hpa@...or.com,
	x86@...nel.org, gleb@...nel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/7] KVM: x86: Function for determining exception type

Il 21/07/2014 23:30, Nadav Amit ha scritto:
> Few comments to see we are on the same page:
> 
> On 7/21/14, 3:18 PM, Paolo Bonzini wrote:
>> Il 21/07/2014 13:37, Nadav Amit ha scritto:
>>> +int kvm_exception_type(unsigned int nr)
>>
>> The manual calls this the exception class.
> Yes, but it also calls it exception "type" (see table 6-1
> "Protected-Mode Exceptions and Interrupts" on the SDM).
> I called it exception type, since there is a function exception_class
> that is used to handle nested exceptions.

Ok.

>>> +    case VE_VECTOR:
>>> +        return EXCPT_FAULT;
>>> +    case DB_VECTOR:
>>> +        return EXCPT_FAULT_OR_TRAP;
>>
>> It is only a fault for instruction fetch breakpoints.  You can modify
>> kvm_vcpu_check_breakpoint to set RF, add a comment here that fault
>> handling is done elsewhere, and return EXCPT_TRAP.
> Unless I am mistaken, kvm_vcpu_check_breakpoint checks only for
> instruction breakpoint. Since instruction breakpoint should not cause RF
> to be set, this function should not be changed.

You're right ("For any fault-class exception except a debug exception
generated in response to an instruction breakpoint, the
value pushed for RF is 1").  It's the debug exception handler that has
to set RF (and then iretd/iretq will modify RF).

> Anyhow, I would return EXCPT_TRAP on DB_VECTOR.

Yeah, just add a comment that it can be a fault for instruction fetch
breakpoints, but we treat it as a trap for the purposes of (not) setting RF.

Alternatively, just rename the function to exception_sets_rf and return
true/false.

Paolo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/