Skip to contentSkip to conversation

KVM: arm64: pKVM host proxy FF-A fixes

PATCH Applied Sebastian Ene 3 participants 7 messages Jun 13, 2024 – Jun 17, 2024

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v3 cover KVM: arm64: pKVM host proxy FF-A fixes

Hello,

This series contains some small fixes for the host pKVM proxy code. I included some of the patches that I already sent on the list as part of this series to make it easier to keep track of them.

I verified the functionality with OPTEE as a TEE-OS.

Changelog:

v2 -> v3:

* small fixes on the commit messages
* applied the Review-by tags from Sudeep. Thank you Sudeep !

v1 -> v2:

* applied the feedback received from Will on the FFA_VERSION patch:
The spec requires that no other calls to be issued prior to the FFA version negotiation and the current change reflects this. After the version negotiation phase is complete with Trustzone we will just return the hypervisor version.

* corrected some mistakes on the FFA_PARTITION_INFO_GET patch:
- don't trim the number of bytes copied from the hypervisor buffer
- introduce FFA_1_0_PARTITON_INFO_SZ definition
- simplify the logic when the input flag is specified

* collected the Ack from Will and embbeded it in the commit - Thanks Will !

v2:
https://lore.kernel.org/all/[email protected]/

v1:

* previously posted FFA_PARTITION_INFO_GET patch here:
https://lore.kernel.org/kvmarm/[email protected]/
-> minor changes from the previous version, look for the current
ffa_version in the host_buffer structure

* previously posted "Fix the identification range for the FF-A smcs" here:
https://lore.kernel.org/kvmarm/[email protected]/

Thank you,
Sebastian

Sebastian Ene (4):
KVM: arm64: Trap FFA_VERSION host call in pKVM
KVM: arm64: Add support for FFA_PARTITION_INFO_GET
KVM: arm64: Update the identification range for the FF-A smcs
KVM: arm64: Use FF-A 1.1 with pKVM

 arch/arm64/kvm/hyp/include/nvhe/ffa.h |   2 +-
 arch/arm64/kvm/hyp/nvhe/ffa.c         | 180 +++++++++++++++++++++-----
 include/linux/arm_ffa.h               |   3 +
 3 files changed, 152 insertions(+), 33 deletions(-)
Signature · 1 line
2.45.2.505.gda0bf45e8d-goog
#1 lore raw

PATCH v3 1/4 KVM: arm64: Trap FFA_VERSION host call in pKVM

The pKVM hypervisor initializes with FF-A version 1.0. The spec requires that no other FF-A calls to be issued before the version negotiation phase is complete. Split the hypervisor proxy initialization code in two parts so that we can move the later one after the host negotiates its version.

Without trapping the call, the host drivers can negotiate a higher version number with TEE which can result in a different memory layout described during the memory sharing calls.

---
 arch/arm64/kvm/hyp/nvhe/ffa.c | 120 +++++++++++++++++++++++++---------
 1 file changed, 90 insertions(+), 30 deletions(-)
arch/arm64/kvm/hyp/nvhe/ffa.c +90 -30
index 02746f9d0980..c8ab51c331f0 100644
@@ -67,6 +67,9 @@ struct kvm_ffa_buffers {
  */
 static struct kvm_ffa_buffers hyp_buffers;
 static struct kvm_ffa_buffers host_buffers;
+static u32 hyp_ffa_version;
+static bool has_version_negotiated;
+static hyp_spinlock_t version_lock;
 
 static void ffa_to_smccc_error(struct arm_smccc_res *res, u64 ffa_errno)
 {
@@ -639,6 +642,83 @@ static bool do_ffa_features(struct arm_smccc_res *res,
 	return true;
 }
 
+static int hyp_ffa_post_init(void)
+{
+	size_t min_rxtx_sz;
+	struct arm_smccc_res res;
+
+	arm_smccc_1_1_smc(FFA_ID_GET, 0, 0, 0, 0, 0, 0, 0, &res);
+	if (res.a0 != FFA_SUCCESS)
+		return -EOPNOTSUPP;
+
+	if (res.a2 != HOST_FFA_ID)
+		return -EINVAL;
+
+	arm_smccc_1_1_smc(FFA_FEATURES, FFA_FN64_RXTX_MAP,
+			  0, 0, 0, 0, 0, 0, &res);
+	if (res.a0 != FFA_SUCCESS)
+		return -EOPNOTSUPP;
+
+	switch (res.a2) {
+	case FFA_FEAT_RXTX_MIN_SZ_4K:
+		min_rxtx_sz = SZ_4K;
+		break;
+	case FFA_FEAT_RXTX_MIN_SZ_16K:
+		min_rxtx_sz = SZ_16K;
+		break;
+	case FFA_FEAT_RXTX_MIN_SZ_64K:
+		min_rxtx_sz = SZ_64K;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (min_rxtx_sz > PAGE_SIZE)
+		return -EOPNOTSUPP;
+
+	return 0;
+}
+
+static void do_ffa_version(struct arm_smccc_res *res,
+			   struct kvm_cpu_context *ctxt)
+{
+	DECLARE_REG(u32, ffa_req_version, ctxt, 1);
+
+	if (FFA_MAJOR_VERSION(ffa_req_version) != 1) {
+		res->a0 = FFA_RET_NOT_SUPPORTED;
+		return;
+	}
+
+	hyp_spin_lock(&version_lock);
+	if (has_version_negotiated) {
+		res->a0 = hyp_ffa_version;
+		goto unlock;
+	}
+
+	/*
+	 * If the client driver tries to downgrade the version, we need to ask
+	 * first if TEE supports it.
+	 */
+	if (FFA_MINOR_VERSION(ffa_req_version) < FFA_MINOR_VERSION(hyp_ffa_version)) {
+		arm_smccc_1_1_smc(FFA_VERSION, ffa_req_version, 0,
+				  0, 0, 0, 0, 0,
+				  res);
+		if (res->a0 == FFA_RET_NOT_SUPPORTED)
+			goto unlock;
+
+		hyp_ffa_version = ffa_req_version;
+	}
+
+	if (hyp_ffa_post_init())
+		res->a0 = FFA_RET_NOT_SUPPORTED;
+	else {
+		has_version_negotiated = true;
+		res->a0 = hyp_ffa_version;
+	}
+unlock:
+	hyp_spin_unlock(&version_lock);
+}
+
 bool kvm_host_ffa_handler(struct kvm_cpu_context *host_ctxt, u32 func_id)
 {
 	struct arm_smccc_res res;
@@ -659,6 +739,11 @@ bool kvm_host_ffa_handler(struct kvm_cpu_context *host_ctxt, u32 func_id)
 	if (!is_ffa_call(func_id))
 		return false;
 
+	if (!has_version_negotiated && func_id != FFA_VERSION) {
+		ffa_to_smccc_error(&res, FFA_RET_INVALID_PARAMETERS);
+		goto out_handled;
+	}
+
 	switch (func_id) {
 	case FFA_FEATURES:
 		if (!do_ffa_features(&res, host_ctxt))
@@ -685,6 +770,9 @@ bool kvm_host_ffa_handler(struct kvm_cpu_context *host_ctxt, u32 func_id)
 	case FFA_MEM_FRAG_TX:
 		do_ffa_mem_frag_tx(&res, host_ctxt);
 		goto out_handled;
+	case FFA_VERSION:
+		do_ffa_version(&res, host_ctxt);
+		goto out_handled;
 	}
 
 	if (ffa_call_supported(func_id))
@@ -699,7 +787,6 @@ bool kvm_host_ffa_handler(struct kvm_cpu_context *host_ctxt, u32 func_id)
 int hyp_ffa_init(void *pages)
 {
 	struct arm_smccc_res res;
-	size_t min_rxtx_sz;
 	void *tx, *rx;
 
 	if (kvm_host_psci_config.smccc_version < ARM_SMCCC_VERSION_1_2)
@@ -725,35 +812,7 @@ int hyp_ffa_init(void *pages)
 	if (FFA_MAJOR_VERSION(res.a0) != 1)
 		return -EOPNOTSUPP;
 
-	arm_smccc_1_1_smc(FFA_ID_GET, 0, 0, 0, 0, 0, 0, 0, &res);
-	if (res.a0 != FFA_SUCCESS)
-		return -EOPNOTSUPP;
-
-	if (res.a2 != HOST_FFA_ID)
-		return -EINVAL;
-
-	arm_smccc_1_1_smc(FFA_FEATURES, FFA_FN64_RXTX_MAP,
-			  0, 0, 0, 0, 0, 0, &res);
-	if (res.a0 != FFA_SUCCESS)
-		return -EOPNOTSUPP;
-
-	switch (res.a2) {
-	case FFA_FEAT_RXTX_MIN_SZ_4K:
-		min_rxtx_sz = SZ_4K;
-		break;
-	case FFA_FEAT_RXTX_MIN_SZ_16K:
-		min_rxtx_sz = SZ_16K;
-		break;
-	case FFA_FEAT_RXTX_MIN_SZ_64K:
-		min_rxtx_sz = SZ_64K;
-		break;
-	default:
-		return -EINVAL;
-	}
-
-	if (min_rxtx_sz > PAGE_SIZE)
-		return -EOPNOTSUPP;
-
+	hyp_ffa_version = FFA_VERSION_1_0;
 	tx = pages;
 	pages += KVM_FFA_MBOX_NR_PAGES * PAGE_SIZE;
 	rx = pages;
@@ -775,5 +834,6 @@ int hyp_ffa_init(void *pages)
 		.lock	= __HYP_SPIN_LOCK_UNLOCKED,
 	};
 
+	version_lock = __HYP_SPIN_LOCK_UNLOCKED;
 	return 0;
 }
Signature · 1 line
2.45.2.505.gda0bf45e8d-goog
#2 lore raw

PATCH v3 2/4 KVM: arm64: Add support for FFA_PARTITION_INFO_GET

In reply to Sebastian Ene

Handle the FFA_PARTITION_INFO_GET host call inside the pKVM hypervisor and copy the response message back to the host buffers.

---
 arch/arm64/kvm/hyp/nvhe/ffa.c | 52 +++++++++++++++++++++++++++++++++++
 include/linux/arm_ffa.h       |  3 ++
 2 files changed, 55 insertions(+)
arch/arm64/kvm/hyp/nvhe/ffa.c +52 -0
index c8ab51c331f0..4eaef673e98d 100644
@@ -719,6 +719,55 @@ static void do_ffa_version(struct arm_smccc_res *res,
 	hyp_spin_unlock(&version_lock);
 }
 
+static void do_ffa_part_get(struct arm_smccc_res *res,
+			    struct kvm_cpu_context *ctxt)
+{
+	DECLARE_REG(u32, uuid0, ctxt, 1);
+	DECLARE_REG(u32, uuid1, ctxt, 2);
+	DECLARE_REG(u32, uuid2, ctxt, 3);
+	DECLARE_REG(u32, uuid3, ctxt, 4);
+	DECLARE_REG(u32, flags, ctxt, 5);
+	u32 count, partition_sz, copy_sz;
+
+	hyp_spin_lock(&host_buffers.lock);
+	if (!host_buffers.rx) {
+		ffa_to_smccc_res(res, FFA_RET_BUSY);
+		goto out_unlock;
+	}
+
+	arm_smccc_1_1_smc(FFA_PARTITION_INFO_GET, uuid0, uuid1,
+			  uuid2, uuid3, flags, 0, 0,
+			  res);
+
+	if (res->a0 != FFA_SUCCESS)
+		goto out_unlock;
+
+	count = res->a2;
+	if (!count)
+		goto out_unlock;
+
+	if (hyp_ffa_version > FFA_VERSION_1_0) {
+		/* Get the number of partitions deployed in the system */
+		if (flags & 0x1)
+			goto out_unlock;
+
+		partition_sz  = res->a3;
+	} else {
+		/* FFA_VERSION_1_0 lacks the size in the response */
+		partition_sz = FFA_1_0_PARTITON_INFO_SZ;
+	}
+
+	copy_sz = partition_sz * count;
+	if (copy_sz > KVM_FFA_MBOX_NR_PAGES * PAGE_SIZE) {
+		ffa_to_smccc_res(res, FFA_RET_ABORTED);
+		goto out_unlock;
+	}
+
+	memcpy(host_buffers.rx, hyp_buffers.rx, copy_sz);
+out_unlock:
+	hyp_spin_unlock(&host_buffers.lock);
+}
+
 bool kvm_host_ffa_handler(struct kvm_cpu_context *host_ctxt, u32 func_id)
 {
 	struct arm_smccc_res res;
@@ -773,6 +822,9 @@ bool kvm_host_ffa_handler(struct kvm_cpu_context *host_ctxt, u32 func_id)
 	case FFA_VERSION:
 		do_ffa_version(&res, host_ctxt);
 		goto out_handled;
+	case FFA_PARTITION_INFO_GET:
+		do_ffa_part_get(&res, host_ctxt);
+		goto out_handled;
 	}
 
 	if (ffa_call_supported(func_id))
include/linux/arm_ffa.h +3 -0
index c82d56768101..c6d18f50f671 100644
@@ -212,6 +212,9 @@ bool ffa_device_is_valid(struct ffa_device *ffa_dev) { return false; }
 
 extern const struct bus_type ffa_bus_type;
 
+/* The FF-A 1.0 partition structure lacks the uuid[4] */
+#define FFA_1_0_PARTITON_INFO_SZ	(8)
+
 /* FFA transport related */
 struct ffa_partition_info {
 	u16 id;
Signature · 1 line
2.45.2.505.gda0bf45e8d-goog
#3 lore raw

PATCH v3 3/4 KVM: arm64: Update the identification range for the FF-A smcs

In reply to Sebastian Ene

The FF-A spec 1.2 reserves the following ranges for identifying FF-A calls:
0x84000060-0x840000FF: FF-A 32-bit calls 0xC4000060-0xC40000FF: FF-A 64-bit calls.

Use the range identification according to the spec and allow calls that are currently out of the range(eg. FFA_MSG_SEND_DIRECT_REQ2) to be identified correctly.

---
 arch/arm64/kvm/hyp/include/nvhe/ffa.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm64/kvm/hyp/include/nvhe/ffa.h +1 -1
index d9fd5e6c7d3c..146e0aebfa1c 100644
@@ -9,7 +9,7 @@
 #include <asm/kvm_host.h>
 
 #define FFA_MIN_FUNC_NUM 0x60
-#define FFA_MAX_FUNC_NUM 0x7F
+#define FFA_MAX_FUNC_NUM 0xFF
 
 int hyp_ffa_init(void *pages);
 bool kvm_host_ffa_handler(struct kvm_cpu_context *host_ctxt, u32 func_id);
Signature · 1 line
2.45.2.505.gda0bf45e8d-goog
#4 lore raw

PATCH v3 4/4 KVM: arm64: Use FF-A 1.1 with pKVM

In reply to Sebastian Ene

Now that the layout of the structures is compatible with 1.1 it is time to probe the 1.1 version of the FF-A protocol inside the hypervisor. If the TEE doesn't support it, it should return the minimum supported version.

---
 arch/arm64/kvm/hyp/nvhe/ffa.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)
arch/arm64/kvm/hyp/nvhe/ffa.c +8 -4
index 4eaef673e98d..fdb63b7857ec 100644
@@ -457,7 +457,7 @@ static __always_inline void do_ffa_mem_xfer(const u64 func_id,
 	memcpy(buf, host_buffers.tx, fraglen);
 
 	ep_mem_access = (void *)buf +
-			ffa_mem_desc_offset(buf, 0, FFA_VERSION_1_0);
+			ffa_mem_desc_offset(buf, 0, hyp_ffa_version);
 	offset = ep_mem_access->composite_off;
 	if (!offset || buf->ep_count != 1 || buf->sender_id != HOST_FFA_ID) {
 		ret = FFA_RET_INVALID_PARAMETERS;
@@ -536,7 +536,7 @@ static void do_ffa_mem_reclaim(struct arm_smccc_res *res,
 	fraglen = res->a2;
 
 	ep_mem_access = (void *)buf +
-			ffa_mem_desc_offset(buf, 0, FFA_VERSION_1_0);
+			ffa_mem_desc_offset(buf, 0, hyp_ffa_version);
 	offset = ep_mem_access->composite_off;
 	/*
 	 * We can trust the SPMD to get this right, but let's at least
@@ -844,7 +844,7 @@ int hyp_ffa_init(void *pages)
 	if (kvm_host_psci_config.smccc_version < ARM_SMCCC_VERSION_1_2)
 		return 0;
 
-	arm_smccc_1_1_smc(FFA_VERSION, FFA_VERSION_1_0, 0, 0, 0, 0, 0, 0, &res);
+	arm_smccc_1_1_smc(FFA_VERSION, FFA_VERSION_1_1, 0, 0, 0, 0, 0, 0, &res);
 	if (res.a0 == FFA_RET_NOT_SUPPORTED)
 		return 0;
 
@@ -864,7 +864,11 @@ int hyp_ffa_init(void *pages)
 	if (FFA_MAJOR_VERSION(res.a0) != 1)
 		return -EOPNOTSUPP;
 
-	hyp_ffa_version = FFA_VERSION_1_0;
+	if (FFA_MINOR_VERSION(res.a0) < FFA_MINOR_VERSION(FFA_VERSION_1_1))
+		hyp_ffa_version = res.a0;
+	else
+		hyp_ffa_version = FFA_VERSION_1_1;
+
 	tx = pages;
 	pages += KVM_FFA_MBOX_NR_PAGES * PAGE_SIZE;
 	rx = pages;
Signature · 1 line
2.45.2.505.gda0bf45e8d-goog
maintainer Acked-by on the v3 series
#5 lore raw
In reply to Sebastian Ene
Quoted 55 lines · Sebastian Ene

Hello,

This series contains some small fixes for the host pKVM proxy code. I included some of the patches that I already sent on the list as part of this series to make it easier to keep track of them.

I verified the functionality with OPTEE as a TEE-OS.

Changelog:

v2 -> v3:

* small fixes on the commit messages
* applied the Review-by tags from Sudeep. Thank you Sudeep !

v1 -> v2:

* applied the feedback received from Will on the FFA_VERSION patch:
The spec requires that no other calls to be issued prior to the FFA version negotiation and the current change reflects this. After the version negotiation phase is complete with Trustzone we will just return the hypervisor version.

* corrected some mistakes on the FFA_PARTITION_INFO_GET patch:
- don't trim the number of bytes copied from the hypervisor buffer
- introduce FFA_1_0_PARTITON_INFO_SZ definition
- simplify the logic when the input flag is specified

* collected the Ack from Will and embbeded it in the commit - Thanks Will !

v2:
https://lore.kernel.org/all/[email protected]/

v1:

* previously posted FFA_PARTITION_INFO_GET patch here:
https://lore.kernel.org/kvmarm/[email protected]/
-> minor changes from the previous version, look for the current
ffa_version in the host_buffer structure

* previously posted "Fix the identification range for the FF-A smcs" here:
https://lore.kernel.org/kvmarm/[email protected]/

Thank you,
Sebastian

Sebastian Ene (4):
KVM: arm64: Trap FFA_VERSION host call in pKVM
KVM: arm64: Add support for FFA_PARTITION_INFO_GET
KVM: arm64: Update the identification range for the FF-A smcs
KVM: arm64: Use FF-A 1.1 with pKVM

Quoted 4 lines
 arch/arm64/kvm/hyp/include/nvhe/ffa.h |   2 +-
 arch/arm64/kvm/hyp/nvhe/ffa.c         | 180 +++++++++++++++++++++-----
 include/linux/arm_ffa.h               |   3 +
 3 files changed, 152 insertions(+), 33 deletions(-)

Although these are labelled as fixes, I don't think they warrant being taken into 6.10. So assuming Oliver will take them into 6.11:

	M.
Signature · 1 line
Without deviation from the norm, progress is not possible.
Applied
#6 lore raw
In reply to Sebastian Ene
Quoted 9 lines · Sebastian Ene

This series contains some small fixes for the host pKVM proxy code. I included some of the patches that I already sent on the list as part of this series to make it easier to keep track of them.

I verified the functionality with OPTEE as a TEE-OS.

Changelog:

[...]

Applied to kvmarm/next, thanks!

[1/4] KVM: arm64: Trap FFA_VERSION host call in pKVM

      https://git.kernel.org/kvmarm/kvmarm/c/c9c012625e12
[2/4] KVM: arm64: Add support for FFA_PARTITION_INFO_GET
      https://git.kernel.org/kvmarm/kvmarm/c/894376385a2d
[3/4] KVM: arm64: Update the identification range for the FF-A smcs
      https://git.kernel.org/kvmarm/kvmarm/c/0dd60c4632a1
[4/4] KVM: arm64: Use FF-A 1.1 with pKVM
      https://git.kernel.org/kvmarm/kvmarm/c/42fb33dde42b

--
Best,
Oliver