s390/mcck: isolate SIE instruction when setting CIF_MCCK_GUEST flag

Commit d768bd892fc8 ("s390: add options to change branch prediction
behaviour for the kernel") introduced .Lsie_exit label - supposedly
to fence off SIE instruction. However, the corresponding address
range length .Lsie_crit_mcck_length was not updated, which led to
BPON code potentionally marked with CIF_MCCK_GUEST flag.

Both .Lsie_exit and .Lsie_crit_mcck_length were removed with commit
0b0ed657fe00 ("s390: remove critical section cleanup from entry.S"),
but the issue persisted - currently BPOFF and BPENTER macros might
get wrongly considered by the machine check handler as a guest.

Fixes: d768bd892fc8 ("s390: add options to change branch prediction behaviour for the kernel")
Reviewed-by: Sven Schnelle <svens@linux.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@linux.ibm.com>
Signed-off-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

diff --git a/arch/s390/kernel/entry.S b/arch/s390/kernel/entry.S
index 82faf4d0e4d7253939529971f8927010c4e90162..d2a1f2f4f5b883edac8d4ae4d5c4afa6107bf80d 100644 (file)
--- a/arch/s390/kernel/entry.S
+++ b/arch/s390/kernel/entry.S
@@ -251,6 +251,10 @@ ENTRY(sie64a)
        BPEXIT  __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP|_TIF_ISOLATE_BP_GUEST)
 .Lsie_entry:
        sie     0(%r14)
+# Let the next instruction be NOP to avoid triggering a machine check
+# and handling it in a guest as result of the instruction execution.
+       nopr    7
+.Lsie_leave:
        BPOFF
        BPENTER __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP|_TIF_ISOLATE_BP_GUEST)
 .Lsie_skip:
@@ -547,7 +551,7 @@ ENTRY(mcck_int_handler)
        jno     .Lmcck_panic
 #if IS_ENABLED(CONFIG_KVM)
        OUTSIDE %r9,.Lsie_gmap,.Lsie_done,6f
-       OUTSIDE %r9,.Lsie_entry,.Lsie_skip,4f
+       OUTSIDE %r9,.Lsie_entry,.Lsie_leave,4f
        oi      __LC_CPU_FLAGS+7, _CIF_MCCK_GUEST
        j       5f
 4:     CHKSTG  .Lmcck_panic