linux-kernel - Re: [PATCHv5, REBASED 3/4] x86/tdx: Dynamically disable SEPT violations from causing #VEs

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]

Message-ID: <fs4n5t3ylzhboxcdrnuhlm6rdsprt7xaeeoae3cbyapw6y4cha@kqm5cwjavs3n>
Date: Tue, 27 Aug 2024 13:04:19 +0300
From: "Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>
To: "Huang, Kai" <kai.huang@...el.com>
Cc: Dave Hansen <dave.hansen@...ux.intel.com>, 
	Thomas Gleixner <tglx@...utronix.de>, Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>, 
	x86@...nel.org, "H. Peter Anvin" <hpa@...or.com>, linux-coco@...ts.linux.dev, 
	linux-kernel@...r.kernel.org, stable@...r.kernel.org
Subject: Re: [PATCHv5, REBASED 3/4] x86/tdx: Dynamically disable SEPT
 violations from causing #VEs

On Wed, Aug 21, 2024 at 01:52:49PM +1200, Huang, Kai wrote:
> > + * attribute is no longer reliable. It reflects the initial state of the
> > + * control for the TD, but it will not be updated if someone (e.g. bootloader)
> > + * changes it before the kernel starts. Kernel must check TDCS_TD_CTLS bit to
> > + * determine if SEPT #VEs are enabled or disabled.
> > + */
> > +static void disable_sept_ve(u64 td_attr)
> > +{
> > +	const char *msg = "TD misconfiguration: SEPT #VE has to be disabled";
> 
> The original msg was:
> 
> 	"TD misconfiguration: SEPT_VE_DISABLE attribute must be set."
> 
> Any reason to change?

Because the attribute is not the only way to control if #VE is going to be
injected.

> 
> 
> > +	bool debug = td_attr & ATTR_DEBUG;
> > +	u64 config, controls;
> > +
> > +	/* Is this TD allowed to disable SEPT #VE */
> > +	tdg_vm_rd(TDCS_CONFIG_FLAGS, &config);
> > +	if (!(config & TDCS_CONFIG_FLEXIBLE_PENDING_VE)) {
> 
> Does this field ID exist in TDX1.0?  I.e., whether it can fail here and
> should we check the return value first?

See TDG.VM.RD definition:

R8  Contents of the field
    In case of no success, as indicated by RAX, R8 returns 0.

No need in error checking here.

> > diff --git a/arch/x86/include/asm/shared/tdx.h b/arch/x86/include/asm/shared/tdx.h
> > index 7e12cfa28bec..fecb2a6e864b 100644
> > --- a/arch/x86/include/asm/shared/tdx.h
> > +++ b/arch/x86/include/asm/shared/tdx.h
> > @@ -19,9 +19,17 @@
> >   #define TDG_VM_RD			7
> >   #define TDG_VM_WR			8
> > -/* TDCS fields. To be used by TDG.VM.WR and TDG.VM.RD module calls */
> > +/* TDX TD-Scope Metadata. To be used by TDG.VM.WR and TDG.VM.RD */
> 
> I am not sure whether this change is necessary.

It is more in-line with spec json dump.

> > +#define TDCS_CONFIG_FLAGS		0x1110000300000016
> > +#define TDCS_TD_CTLS			0x1110000300000017
> 
> The TDX 1.5 spec 'td_scope_metadata.json' says they are 0x9110000300000016
> and 0x9110000300000017.

The spec is broken. It is going to be fixed. I use correct values.

> I know the bit 63 is ignored by the TDX module, but since (IIUC) those two
> fields are introduced in TDX1.5, it's just better to follow what TDX1.5 spec
> says.

Newer modules will ignore this bit and both values are going to
acceptable.

-- 
  Kiryl Shutsemau / Kirill A. Shutemov