linux-kernel - Re: [PATCH] KVM: x86: use positive error values for msr emulation that causes #GP

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] [day] [month] [year] [list]

Date:   Thu, 5 Nov 2020 11:00:53 +0100
From:   Pankaj Gupta <pankaj.gupta.linux@...il.com>
To:     Maxim Levitsky <mlevitsk@...hat.com>
Cc:     kvm@...r.kernel.org, Borislav Petkov <bp@...en8.de>,
        "open list:X86 ARCHITECTURE (32-BIT AND 64-BIT)" 
        <linux-kernel@...r.kernel.org>,
        "maintainer:X86 ARCHITECTURE (32-BIT AND 64-BIT)" <x86@...nel.org>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Vitaly Kuznetsov <vkuznets@...hat.com>,
        Joerg Roedel <joro@...tes.org>,
        "H. Peter Anvin" <hpa@...or.com>,
        Sean Christopherson <sean.j.christopherson@...el.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Ingo Molnar <mingo@...hat.com>, Qian Cai <cai@...hat.com>
Subject: Re: [PATCH] KVM: x86: use positive error values for msr emulation
 that causes #GP

> > This looks good to me. This should solve "-EPERM" return by "__kvm_set_msr" .
> >
> > A question I have, In the case of "kvm_emulate_rdmsr()",  for "r" we
> > are injecting #GP.
> > Is there any possibility of this check to be hit and still result in #GP?
>
> When I wrote this patch series I assumed that msr reads usually don't have
> side effects so they shouldn't fail, and fixed only the msr write code path
> to deal with negative errors. Now that you put this in this light,
> I do think that you are right and I should have added code for both msr reads and writes
> especially to catch cases in which negative errors are returned by mistake
> like this one (my mistake in this case since my patch series was merged
> after the userspace msrs patch series).
>
> What do you think?
>
> I can prepare a separate patch for this, which should go to the next
> kernel version since this doesn't fix a regression.

Patch on the top should be okay. I think.

Thanks,
Pankaj