Re: [PATCH] x86/bugs: Switch to "auto" when "ibrs" selected on Enhanced IBRS parts

[Josh Poimboeuf]

On Thu, Jul 14, 2022 at 01:42:11PM +0200, Peter Zijlstra wrote:
> On Thu, Jul 14, 2022 at 08:17:26AM -0300, Thadeu Lima de Souza Cascardo wrote:
> > On Wed, Jul 13, 2022 at 10:32:37PM -0700, Pawan Gupta wrote:
> > > Currently spectre_v2=ibrs forces write to MSR_IA32_SPEC_CTRL at every
> > > entry and exit. On Enhanced IBRS parts setting MSR_IA32_SPEC_CTRL[IBRS]
> > > only once at bootup is sufficient. MSR write at every kernel entry/exit
> > > incur unnecessary penalty that can be avoided.
> > > 
> > > When Enhanced IBRS feature is present, switch from "ibrs" to "auto" mode
> > > so that appropriate mitigation is selected.
> > > 
> > > Fixes: 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")
> > > Cc: stable@xxxxxxxxxxxxxxx # 5.10+
> > > Signed-off-by: Pawan Gupta <pawan.kumar.gupta@xxxxxxxxxxxxxxx>
> > > ---
> > >  arch/x86/kernel/cpu/bugs.c | 6 ++++++
> > >  1 file changed, 6 insertions(+)
> > > 
> > > diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c
> > > index 0dd04713434b..7d7ebfdfbeda 100644
> > > --- a/arch/x86/kernel/cpu/bugs.c
> > > +++ b/arch/x86/kernel/cpu/bugs.c
> > > @@ -1303,6 +1303,12 @@ static enum spectre_v2_mitigation_cmd __init spectre_v2_parse_cmdline(void)
> > >  		return SPECTRE_V2_CMD_AUTO;
> > >  	}
> > >  
> > > +	if (cmd == SPECTRE_V2_CMD_IBRS && boot_cpu_has(X86_FEATURE_IBRS_ENHANCED)) {
> > > +		pr_err("%s selected but CPU supports Enhanced IBRS. Switching to AUTO select\n",
> > > +		       mitigation_options[i].option);
> > > +		return SPECTRE_V2_CMD_AUTO;
> > > +	}
> > > +
> > >  	spec_v2_print_cond(mitigation_options[i].option,
> > >  			   mitigation_options[i].secure);
> > >  	return cmd;
> > > 
> > > base-commit: 72a8e05d4f66b5af7854df4490e3135168694b6b
> > > -- 
> > > 2.35.3
> > > 
> > > 
> > 
> > Shouldn't we just use the mitigation the user asked for if it is still
> > possible? We could add the warning advising the user that a different
> > mitigation could be used instead with less penalty, but if the user asked for
> > IBRS and that is available, it should be used.
> > 
> > One of the reasons for that is testing. I know it was useful enough for me and
> > it helped me find some bugs.
> 
> Yeah this; if the user asks for IBRS, we should give him IBRS. I hate
> the 'I know better, let me change that for you' mentality.

eIBRS CPUs don't even have legacy IBRS so I don't see how this is even
possible.

-- 
Josh