On 2/14/25 2:13 AM, Steven Price wrote:
MMIO emulation for a realm cannot be done directly with the VM's
registers as they are protected from the host. However, for emulatable
data aborts, the RMM uses GPRS[0] to provide the read/written value.
We can transfer this from/to the equivalent VCPU's register entry and
then depend on the generic MMIO handling code in KVM.
For a MMIO read, the value is placed in the shared RecExit structure
during kvm_handle_mmio_return() rather than in the VCPU's register
entry.
Signed-off-by: Steven Price <steven.price@xxxxxxx>
---
Changes since v5:
* Inject SEA to the guest is an emulatable MMIO access triggers a data
abort.
* kvm_handle_mmio_return() - disable kvm_incr_pc() for a REC (as the PC
isn't under the host's control) and move the REC_ENTER_EMULATED_MMIO
flag setting to this location (as that tells the RMM to skip the
instruction).
---
arch/arm64/kvm/inject_fault.c | 4 +++-
arch/arm64/kvm/mmio.c | 16 ++++++++++++----
arch/arm64/kvm/rme-exit.c | 6 ++++++
3 files changed, 21 insertions(+), 5 deletions(-)
One nitpick below, with it addressed:
Reviewed-by: Gavin Shan <gshan@xxxxxxxxxx>
diff --git a/arch/arm64/kvm/inject_fault.c b/arch/arm64/kvm/inject_fault.c
index a640e839848e..2a9682b9834f 100644
--- a/arch/arm64/kvm/inject_fault.c
+++ b/arch/arm64/kvm/inject_fault.c
@@ -165,7 +165,9 @@ static void inject_abt32(struct kvm_vcpu *vcpu, bool is_pabt, u32 addr)
*/
void kvm_inject_dabt(struct kvm_vcpu *vcpu, unsigned long addr)
{
- if (vcpu_el1_is_32bit(vcpu))
+ if (unlikely(vcpu_is_rec(vcpu)))
+ vcpu->arch.rec.run->enter.flags |= REC_ENTER_FLAG_INJECT_SEA;
+ else if (vcpu_el1_is_32bit(vcpu))
inject_abt32(vcpu, false, addr);
else
inject_abt64(vcpu, false, addr);
diff --git a/arch/arm64/kvm/mmio.c b/arch/arm64/kvm/mmio.c
index ab365e839874..bff89d47a4d5 100644
--- a/arch/arm64/kvm/mmio.c
+++ b/arch/arm64/kvm/mmio.c
@@ -6,6 +6,7 @@
#include <linux/kvm_host.h>
#include <asm/kvm_emulate.h>
+#include <asm/rmi_smc.h>
#include <trace/events/kvm.h>
#include "trace.h"
@@ -136,14 +137,21 @@ int kvm_handle_mmio_return(struct kvm_vcpu *vcpu)
trace_kvm_mmio(KVM_TRACE_MMIO_READ, len, run->mmio.phys_addr,
&data);
data = vcpu_data_host_to_guest(vcpu, data, len);
- vcpu_set_reg(vcpu, kvm_vcpu_dabt_get_rd(vcpu), data);
+
+ if (vcpu_is_rec(vcpu))
+ vcpu->arch.rec.run->enter.gprs[0] = data;
+ else
+ vcpu_set_reg(vcpu, kvm_vcpu_dabt_get_rd(vcpu), data);
}
/*
* The MMIO instruction is emulated and should not be re-executed
* in the guest.
*/
- kvm_incr_pc(vcpu);
+ if (vcpu_is_rec(vcpu))
+ vcpu->arch.rec.run->enter.flags |= REC_ENTER_FLAG_EMULATED_MMIO;
+ else
+ kvm_incr_pc(vcpu);
return 1;
}
@@ -162,14 +170,14 @@ int io_mem_abort(struct kvm_vcpu *vcpu, phys_addr_t fault_ipa)
* No valid syndrome? Ask userspace for help if it has
* volunteered to do so, and bail out otherwise.
*
- * In the protected VM case, there isn't much userspace can do
+ * In the protected/realm VM case, there isn't much userspace can do
* though, so directly deliver an exception to the guest.
*/
if (!kvm_vcpu_dabt_isvalid(vcpu)) {
trace_kvm_mmio_nisv(*vcpu_pc(vcpu), kvm_vcpu_get_esr(vcpu),
kvm_vcpu_get_hfar(vcpu), fault_ipa);
- if (vcpu_is_protected(vcpu)) {
+ if (vcpu_is_protected(vcpu) || vcpu_is_rec(vcpu)) {
kvm_inject_dabt(vcpu, kvm_vcpu_get_hfar(vcpu));
return 1;
}
diff --git a/arch/arm64/kvm/rme-exit.c b/arch/arm64/kvm/rme-exit.c
index aae1adefe1a3..c785005f821f 100644
--- a/arch/arm64/kvm/rme-exit.c
+++ b/arch/arm64/kvm/rme-exit.c
@@ -25,6 +25,12 @@ static int rec_exit_reason_notimpl(struct kvm_vcpu *vcpu)
static int rec_exit_sync_dabt(struct kvm_vcpu *vcpu)
{
+ struct realm_rec *rec = &vcpu->arch.rec;
+
+ if (kvm_vcpu_dabt_iswrite(vcpu) && kvm_vcpu_dabt_isvalid(vcpu))
+ vcpu_set_reg(vcpu, kvm_vcpu_dabt_get_rd(vcpu),
+ rec->run->exit.gprs[0]);
+
A comment may be needed to explain why GPR[0] has to be copied over. The contexnt
in GPR[0] isn't needed by all cases, being handled by kvm_handle_guest_abort().
Something like below.
/*
* Copy over GPR[0] to the target GPR, preparing to handle MMIO write
* fault. The content to be written has been saved to GPR[0] by RMM.
* It's overhead to other cases like fault due to MMIO read, shared
* or private space access.
*/
return kvm_handle_guest_abort(vcpu);
}
Thanks,
Gavin
