Web lists-archives.com

[PATCH 4.4 246/266] x86/speculation/mds: Print SMT vulnerable on MSBDS with mitigations off




From: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>

commit e2c3c94788b08891dcf3dbe608f9880523ecd71b upstream.

This code is only for CPUs which are affected by MSBDS, but are *not*
affected by the other two MDS issues.

For such CPUs, enabling the mds_idle_clear mitigation is enough to
mitigate SMT.

However if user boots with 'mds=off' and still has SMT enabled, we should
not report that SMT is mitigated:

$cat /sys//devices/system/cpu/vulnerabilities/mds
Vulnerable; SMT mitigated

But rather:
Vulnerable; SMT vulnerable

Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>
Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Reviewed-by: Tyler Hicks <tyhicks@xxxxxxxxxxxxx>
Reviewed-by: Josh Poimboeuf <jpoimboe@xxxxxxxxxx>
Link: https://lkml.kernel.org/r/20190412215118.294906495@localhost.localdomain
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
 arch/x86/kernel/cpu/bugs.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

--- a/arch/x86/kernel/cpu/bugs.c
+++ b/arch/x86/kernel/cpu/bugs.c
@@ -1091,7 +1091,8 @@ static ssize_t mds_show_state(char *buf)
 
 	if (boot_cpu_has(X86_BUG_MSBDS_ONLY)) {
 		return sprintf(buf, "%s; SMT %s\n", mds_strings[mds_mitigation],
-			       sched_smt_active() ? "mitigated" : "disabled");
+			       (mds_mitigation == MDS_MITIGATION_OFF ? "vulnerable" :
+			        sched_smt_active() ? "mitigated" : "disabled"));
 	}
 
 	return sprintf(buf, "%s; SMT %s\n", mds_strings[mds_mitigation],