public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Sean Christopherson <seanjc@google.com>
To: Hou Wenlong <houwenlong.hwl@antgroup.com>
Cc: kvm@vger.kernel.org, Paolo Bonzini <pbonzini@redhat.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	x86@kernel.org, "H. Peter Anvin" <hpa@zytor.com>,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH] KVM: x86: Mark transfer type as X86_TRANSFER_RET when loading CS in iret emulation
Date: Wed, 12 Oct 2022 16:34:57 +0000	[thread overview]
Message-ID: <Y0bssbjJTQVB+SCg@google.com> (raw)
In-Reply-To: <fcaf1408d2aaaa39b33cdd3b11bf06e7e935d11a.1665565774.git.houwenlong.hwl@antgroup.com>

On Wed, Oct 12, 2022, Hou Wenlong wrote:
> When loading code segment descriptor in iret instruction emulation, the
> checks are same as far return instruction emulation, so transfer type
> should be X86_TRANSFER_RET in __load_segment_descriptor(). Although,
> only iret in real mode is implemented now, and no checks are actually
> needed for real mode, it would still be better to mark transfer type as
> X86_TRANSFER_RET.

It's not strictly a RET though.  The RPL vs. DPL checks in __load_segment_descriptor()
might do the right thing, but there's a rather large pile of stuff IRET can do that
RET can't (ignoring the fact that KVM doesn't even emulate FAR RET to outer privilege
levels).

And __emulate_int_real() also loads CS with X86_TRANSFER_NONE, i.e. KVM still has
a weird path to worry about.

Rather than make the IRET case slightly less wrong, what about adding a sanity
check in __load_segment_descriptor() that KVM doesn't attempt to load CS in Protected
Mode with X86_TRANSFER_NONE?

diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
index 3b27622d4642..fe735e18c419 100644
--- a/arch/x86/kvm/emulate.c
+++ b/arch/x86/kvm/emulate.c
@@ -1641,6 +1641,14 @@ static int __load_segment_descriptor(struct x86_emulate_ctxt *ctxt,
                        goto exception;
                break;
        case VCPU_SREG_CS:
+               /*
+                * KVM uses "none" when loading CS as part of emulating Real
+                * Mode exceptions and IRET (handled above).  In all other
+                * cases, loading CS without a control transfer is a KVM bug.
+                */
+               if (WARN_ON_ONCE(transfer == X86_TRANSFER_NONE))
+                       goto exception;
+
                if (!(seg_desc.type & 8))
                        goto exception;

> 
> No functional change intended.
> 
> Signed-off-by: Hou Wenlong <houwenlong.hwl@antgroup.com>
> ---
>  arch/x86/kvm/emulate.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
> index 3b27622d4642..5052eb480068 100644
> --- a/arch/x86/kvm/emulate.c
> +++ b/arch/x86/kvm/emulate.c
> @@ -2100,6 +2100,7 @@ static int emulate_iret_real(struct x86_emulate_ctxt *ctxt)
>  			     X86_EFLAGS_FIXED;
>  	unsigned long vm86_mask = X86_EFLAGS_VM | X86_EFLAGS_VIF |
>  				  X86_EFLAGS_VIP;
> +	u8 cpl = ctxt->ops->cpl(ctxt);
>  
>  	/* TODO: Add stack limit check */
>  
> @@ -2121,7 +2122,8 @@ static int emulate_iret_real(struct x86_emulate_ctxt *ctxt)
>  	if (rc != X86EMUL_CONTINUE)
>  		return rc;
>  
> -	rc = load_segment_descriptor(ctxt, (u16)cs, VCPU_SREG_CS);
> +	rc = __load_segment_descriptor(ctxt, (u16)cs, VCPU_SREG_CS, cpl,
> +				       X86_TRANSFER_RET, NULL);
>  
>  	if (rc != X86EMUL_CONTINUE)
>  		return rc;
> -- 
> 2.31.1
> 

  reply	other threads:[~2022-10-12 16:35 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-12  9:11 [PATCH] KVM: x86: Mark transfer type as X86_TRANSFER_RET when loading CS in iret emulation Hou Wenlong
2022-10-12 16:34 ` Sean Christopherson [this message]
2022-10-13  6:52   ` Hou Wenlong
2022-10-14  4:16     ` Sean Christopherson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=Y0bssbjJTQVB+SCg@google.com \
    --to=seanjc@google.com \
    --cc=bp@alien8.de \
    --cc=dave.hansen@linux.intel.com \
    --cc=houwenlong.hwl@antgroup.com \
    --cc=hpa@zytor.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=tglx@linutronix.de \
    --cc=x86@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox