Commit cd81acc6 authored by Nicholas Piggin's avatar Nicholas Piggin Committed by Michael Ellerman
Browse files

powerpc/64s/exception: KVM Fix for host DSI being taken in HPT guest MMU context 



Commit 2284ffea ("powerpc/64s/exception: Only test KVM in SRR
interrupts when PR KVM is supported") removed KVM guest tests from
interrupts that do not set HV=1, when PR-KVM is not configured.

This is wrong for HV-KVM HPT guest MMIO emulation case which attempts
to load the faulting instruction word with MSR[DR]=1 and MSR[HV]=1 with
the guest MMU context loaded. This can cause host DSI, DSLB interrupts
which must test for KVM guest. Restore this and add a comment.

Fixes: 2284ffea ("powerpc/64s/exception: Only test KVM in SRR interrupts when PR KVM is supported")
Cc: stable@vger.kernel.org # v5.7+
Signed-off-by: default avatarNicholas Piggin <npiggin@gmail.com>
Signed-off-by: default avatarMichael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201117135617.3521127-1-npiggin@gmail.com
parent e02152ba

arch/powerpc/kernel/exceptions-64s.S

+7 −4
Original line number Diff line number Diff line
@@ -1410,6 +1410,11 @@ END_FTR_SECTION_IFSET(CPU_FTR_HVMODE) 
 *   If none is found, do a Linux page fault. Linux page faults can happen in

 *   kernel mode due to user copy operations of course.

 *

 *   KVM: The KVM HDSI handler may perform a load with MSR[DR]=1 in guest

 *   MMU context, which may cause a DSI in the host, which must go to the

 *   KVM handler. MSR[IR] is not enabled, so the real-mode handler will

 *   always be used regardless of AIL setting.

 *

 * - Radix MMU

 *   The hardware loads from the Linux page table directly, so a fault goes

 *   immediately to Linux page fault.
@@ -1420,10 +1425,8 @@ INT_DEFINE_BEGIN(data_access) 
	IVEC=0x300

	IDAR=1

	IDSISR=1

#ifdef CONFIG_KVM_BOOK3S_PR_POSSIBLE

	IKVM_SKIP=1

	IKVM_REAL=1

#endif

INT_DEFINE_END(data_access)



EXC_REAL_BEGIN(data_access, 0x300, 0x80)
@@ -1462,6 +1465,8 @@ ALT_MMU_FTR_SECTION_END_IFCLR(MMU_FTR_TYPE_RADIX) 
 *   ppc64_bolted_size (first segment). The kernel handler must avoid stomping

 *   on user-handler data structures.

 *

 *   KVM: Same as 0x300, DSLB must test for KVM guest.

 *

 * A dedicated save area EXSLB is used (XXX: but it actually need not be

 * these days, we could use EXGEN).

 */
@@ -1470,10 +1475,8 @@ INT_DEFINE_BEGIN(data_access_slb) 
	IAREA=PACA_EXSLB

	IRECONCILE=0

	IDAR=1

#ifdef CONFIG_KVM_BOOK3S_PR_POSSIBLE

	IKVM_SKIP=1

	IKVM_REAL=1

#endif

INT_DEFINE_END(data_access_slb)



EXC_REAL_BEGIN(data_access_slb, 0x380, 0x80)

CRA Git | Maintained and supported by SUSTech CRA and CCSE