Skip to contentSkip to conversation

Xen FF-A mediator

PATCH Superseded Jens Wiklander 2 participants 7 messages Jun 7, 2022 – Jun 8, 2022

v1 of 11 Superseded Latest is v11 →

lore
Share & tools
lore
#0 lore raw

cover Xen FF-A mediator

Hi,

This patch sets add a FF-A [1] mediator modeled after the TEE mediator already present in Xen. The FF-A mediator implements the subset of the FF-A 1.1 specification needed to communicate with OP-TEE using FF-A as transport mechanism instead of SMC/HVC as with the TEE mediator. It allows a similar design in OP-TEE as with the TEE mediator where OP-TEE presents one virtual partition of itself to each guest in Xen.

The FF-A mediator is generic in the sense it has nothing OP-TEE specific except that only the subset needed for OP-TEE is implemented so far. The hooks needed to inform OP-TEE that a guest is created or destroyed is part of the FF-A specification.

It should be possible to extend the FF-A mediator to implement a larger portion of the FF-A 1.1 specification without breaking with the way OP-TEE is communicated with here. So it should be possible to support any TEE or Secure Partition using FF-A as transport with this mediator.

[1] https://developer.arm.com/documentation/den0077/latest

Thanks,
Jens

Jens Wiklander (2):
xen/arm: smccc: add support for SMCCCv1.2 extended input/output
registers
xen/arm: add FF-A mediator

 xen/arch/arm/Kconfig         |   11 +
 xen/arch/arm/Makefile        |    1 +
 xen/arch/arm/arm64/smc.S     |   43 +
 xen/arch/arm/domain.c        |   10 +
 xen/arch/arm/ffa.c           | 1624 ++++++++++++++++++++++++++++++++++
 xen/arch/arm/vsmc.c          |   19 +-
 xen/include/asm-arm/domain.h |    4 +
 xen/include/asm-arm/ffa.h    |   71 ++
 xen/include/asm-arm/smccc.h  |   42 +
 9 files changed, 1821 insertions(+), 4 deletions(-)
 create mode 100644 xen/arch/arm/ffa.c
 create mode 100644 xen/include/asm-arm/ffa.h
Signature · 1 line
2.31.1
#1 lore raw

PATCH 1/2 xen/arm: smccc: add support for SMCCCv1.2 extended input/output registers

SMCCC v1.2 AArch64 allows x0-x17 to be used as both parameter registers and result registers for the SMC and HVC instructions.

Arm Firmware Framework for Armv8-A specification makes use of x0-x7 as parameter and result registers.

Let us add new interface to support this extended set of input/output registers.

This is based on 3fdc0cb59d97 ("arm64: smccc: Add support for SMCCCv1.2 extended input/output registers") by Sudeep Holla from the Linux kernel

---
 xen/arch/arm/arm64/smc.S    | 43 +++++++++++++++++++++++++++++++++++++
 xen/arch/arm/vsmc.c         |  2 +-
 xen/include/asm-arm/smccc.h | 42 ++++++++++++++++++++++++++++++++++++
 3 files changed, 86 insertions(+), 1 deletion(-)
xen/arch/arm/arm64/smc.S +43 -0
index b0752be57e8f..e024fa4d44d1 100644
@@ -30,3 +30,46 @@ ENTRY(__arm_smccc_1_0_smc)
         stp     x2, x3, [x4, #SMCCC_RES_a2]
 1:
         ret
+
+
+/*
+ * void arm_smccc_1_2_smc(const struct arm_smccc_1_2_regs *args,
+ *                        struct arm_smccc_1_2_regs *res)
+ */
+ENTRY(arm_smccc_1_2_smc)
+    /* Save `res` and free a GPR that won't be clobbered */
+    stp     x1, x19, [sp, #-16]!
+
+    /* Ensure `args` won't be clobbered while loading regs in next step */
+    mov	x19, x0
+
+    /* Load the registers x0 - x17 from the struct arm_smccc_1_2_regs */
+    ldp	x0, x1, [x19, #0]
+    ldp	x2, x3, [x19, #16]
+    ldp	x4, x5, [x19, #32]
+    ldp	x6, x7, [x19, #48]
+    ldp	x8, x9, [x19, #64]
+    ldp	x10, x11, [x19, #80]
+    ldp	x12, x13, [x19, #96]
+    ldp	x14, x15, [x19, #112]
+    ldp	x16, x17, [x19, #128]
+
+    smc #0
+
+    /* Load the `res` from the stack */
+    ldr	x19, [sp]
+
+    /* Store the registers x0 - x17 into the result structure */
+    stp	x0, x1, [x19, #0]
+    stp	x2, x3, [x19, #16]
+    stp	x4, x5, [x19, #32]
+    stp	x6, x7, [x19, #48]
+    stp	x8, x9, [x19, #64]
+    stp	x10, x11, [x19, #80]
+    stp	x12, x13, [x19, #96]
+    stp	x14, x15, [x19, #112]
+    stp	x16, x17, [x19, #128]
+
+    /* Restore original x19 */
+    ldp     xzr, x19, [sp], #16
+    ret
xen/arch/arm/vsmc.c +1 -1
index a36db15fffc0..33b0ddc634da 100644
@@ -93,7 +93,7 @@ static bool handle_arch(struct cpu_user_regs *regs)
     switch ( fid )
     {
     case ARM_SMCCC_VERSION_FID:
-        set_user_reg(regs, 0, ARM_SMCCC_VERSION_1_1);
+        set_user_reg(regs, 0, ARM_SMCCC_VERSION_1_2);
         return true;
 
     case ARM_SMCCC_ARCH_FEATURES_FID:
xen/include/asm-arm/smccc.h +42 -0

62 lines not on this page · raw message

Signature · 1 line
2.31.1
#2 lore raw

PATCH 2/2 xen/arm: add FF-A mediator

In reply to Jens Wiklander

Adds a FF-A version 1.1 [1] mediator to communicate with a Secure Partition in secure world.

The implementation is the bare minimum to be able to communicate with OP-TEE running as an SPMC at S-EL1.

This is loosely based on the TEE mediator framework and the OP-TEE mediator.

[1] https://developer.arm.com/documentation/den0077/latest

---
 xen/arch/arm/Kconfig         |   11 +
 xen/arch/arm/Makefile        |    1 +
 xen/arch/arm/domain.c        |   10 +
 xen/arch/arm/ffa.c           | 1624 ++++++++++++++++++++++++++++++++++
 xen/arch/arm/vsmc.c          |   17 +-
 xen/include/asm-arm/domain.h |    4 +
 xen/include/asm-arm/ffa.h    |   71 ++
 7 files changed, 1735 insertions(+), 3 deletions(-)
 create mode 100644 xen/arch/arm/ffa.c
 create mode 100644 xen/include/asm-arm/ffa.h
xen/arch/arm/Kconfig +11 -0
index 277738826581..281edb85cbff 100644
@@ -114,6 +114,17 @@ config TEE
 
 source "arch/arm/tee/Kconfig"
 
+config FFA
+	bool "Enable FF-A mediator support" if EXPERT
+	default n
+	depends on ARM_64
+	help
+	  This option enables a minamal FF-A mediator. The mediator is
+	  generic as it follows the FF-A specification [1], but it only
+	  implements a small substet of the specification.
+
+	  [1] https://developer.arm.com/documentation/den0077/latest
+
 endmenu
 
 menu "ARM errata workaround via the alternative framework"
xen/arch/arm/Makefile +1 -0
index 1ded44d20047..90c43a4a9edf 100644
@@ -17,6 +17,7 @@ obj-y += domain.o
 obj-y += domain_build.init.o
 obj-y += domctl.o
 obj-$(CONFIG_EARLY_PRINTK) += early_printk.o
+obj-$(CONFIG_FFA) += ffa.o
 obj-y += gic.o
 obj-y += gic-v2.o
 obj-$(CONFIG_GICV3) += gic-v3.o
xen/arch/arm/domain.c +10 -0
index ac6a419f55c9..451f3eed550f 100644
@@ -25,6 +25,7 @@
 #include <asm/cpufeature.h>
 #include <asm/current.h>
 #include <asm/event.h>
+#include <asm/ffa.h>
 #include <asm/gic.h>
 #include <asm/guest_access.h>
 #include <asm/guest_atomics.h>
@@ -737,6 +738,9 @@ int arch_domain_create(struct domain *d,
     if ( (rc = tee_domain_init(d, config->arch.tee_type)) != 0 )
         goto fail;
 
+    if ( (rc = ffa_domain_init(d)) != 0 )
+        goto fail;
+
     update_domain_wallclock_time(d);
 
     /*
@@ -974,6 +978,7 @@ static int relinquish_memory(struct domain *d, struct page_list_head *list)
  */
 enum {
     PROG_tee = 1,
+    PROG_ffa,
     PROG_xen,
     PROG_page,
     PROG_mapping,
@@ -1012,6 +1017,11 @@ int domain_relinquish_resources(struct domain *d)
         if (ret )
             return ret;
 
+    PROGRESS(ffa):
+        ret = ffa_relinquish_resources(d);
+        if (ret )
+            return ret;
+
     PROGRESS(xen):
         ret = relinquish_memory(d, &d->xenpage_list);
         if ( ret )
xen/arch/arm/ffa.c new file +1624 -0

1,624 lines not on this page · raw message

xen/arch/arm/vsmc.c +14 -3
index 33b0ddc634da..cfc56715191b 100644
@@ -20,6 +20,7 @@
 #include <public/arch-arm/smccc.h>
 #include <asm/cpuerrata.h>
 #include <asm/cpufeature.h>
+#include <asm/ffa.h>
 #include <asm/monitor.h>
 #include <asm/regs.h>
 #include <asm/smccc.h>
@@ -32,7 +33,7 @@
 #define XEN_SMCCC_FUNCTION_COUNT 3
 
 /* Number of functions currently supported by Standard Service Service Calls. */
-#define SSSC_SMCCC_FUNCTION_COUNT (3 + VPSCI_NR_FUNCS)
+#define SSSC_SMCCC_FUNCTION_COUNT (3 + VPSCI_NR_FUNCS + FFA_NR_FUNCS)
 
 static bool fill_uid(struct cpu_user_regs *regs, xen_uuid_t uuid)
 {
@@ -186,13 +187,23 @@ static bool handle_existing_apis(struct cpu_user_regs *regs)
     return do_vpsci_0_1_call(regs, fid);
 }
 
+static bool is_psci_fid(uint32_t fid)
+{
+    uint32_t fn = fid & ARM_SMCCC_FUNC_MASK;
+
+    return fn >= 0 && fn <= 0x1fU;
+}
+
 /* PSCI 0.2 interface and other Standard Secure Calls */
 static bool handle_sssc(struct cpu_user_regs *regs)
 {
     uint32_t fid = (uint32_t)get_user_reg(regs, 0);
 
-    if ( do_vpsci_0_2_call(regs, fid) )
-        return true;
+    if ( is_psci_fid(fid) )
+        return do_vpsci_0_2_call(regs, fid);
+
+    if ( is_ffa_fid(fid) )
+        return ffa_handle_call(regs, fid);
 
     switch ( fid )
     {
xen/include/asm-arm/domain.h +4 -0
index 4e2f582006a5..e8ac49e29899 100644
@@ -91,6 +91,10 @@ struct arch_domain
 #ifdef CONFIG_TEE
     void *tee;
 #endif
+
+#ifdef CONFIG_FFA
+    void *ffa;
+#endif
 }  __cacheline_aligned;
 
 struct arch_vcpu
xen/include/asm-arm/ffa.h new file +71 -0

71 lines not on this page · raw message

Signature · 1 line
2.31.1
#3 lore raw
In reply to Jens Wiklander
Quoted 19 lines · Jens Wiklander

Hi,

This patch sets add a FF-A [1] mediator modeled after the TEE mediator already present in Xen. The FF-A mediator implements the subset of the FF-A 1.1 specification needed to communicate with OP-TEE using FF-A as transport mechanism instead of SMC/HVC as with the TEE mediator. It allows a similar design in OP-TEE as with the TEE mediator where OP-TEE presents one virtual partition of itself to each guest in Xen.

The FF-A mediator is generic in the sense it has nothing OP-TEE specific except that only the subset needed for OP-TEE is implemented so far. The hooks needed to inform OP-TEE that a guest is created or destroyed is part of the FF-A specification.

It should be possible to extend the FF-A mediator to implement a larger portion of the FF-A 1.1 specification without breaking with the way OP-TEE is communicated with here. So it should be possible to support any TEE or Secure Partition using FF-A as transport with this mediator.

Quoted 4 lines

[1] https://developer.arm.com/documentation/den0077/latest

Thanks,
Jens

Hi Jens,

Many thanks for the patches! I tried to apply them to the master branch but unfortunately they don't apply any longer. Could you please rebase them on master (or even better rebase them on staging) and resend?

Thank you!

Quoted 17 lines

Jens Wiklander (2):
xen/arm: smccc: add support for SMCCCv1.2 extended input/output
registers
xen/arm: add FF-A mediator

 xen/arch/arm/Kconfig         |   11 +
 xen/arch/arm/Makefile        |    1 +
 xen/arch/arm/arm64/smc.S     |   43 +
 xen/arch/arm/domain.c        |   10 +
 xen/arch/arm/ffa.c           | 1624 ++++++++++++++++++++++++++++++++++
 xen/arch/arm/vsmc.c          |   19 +-
 xen/include/asm-arm/domain.h |    4 +
 xen/include/asm-arm/ffa.h    |   71 ++
 xen/include/asm-arm/smccc.h  |   42 +
 9 files changed, 1821 insertions(+), 4 deletions(-)
 create mode 100644 xen/arch/arm/ffa.c
 create mode 100644 xen/include/asm-arm/ffa.h
#4 lore raw
Quoted 25 lines · Stefano Stabellini
Quoted 19 lines · Jens Wiklander

Hi,

This patch sets add a FF-A [1] mediator modeled after the TEE mediator already present in Xen. The FF-A mediator implements the subset of the FF-A 1.1 specification needed to communicate with OP-TEE using FF-A as transport mechanism instead of SMC/HVC as with the TEE mediator. It allows a similar design in OP-TEE as with the TEE mediator where OP-TEE presents one virtual partition of itself to each guest in Xen.

The FF-A mediator is generic in the sense it has nothing OP-TEE specific except that only the subset needed for OP-TEE is implemented so far. The hooks needed to inform OP-TEE that a guest is created or destroyed is part of the FF-A specification.

It should be possible to extend the FF-A mediator to implement a larger portion of the FF-A 1.1 specification without breaking with the way OP-TEE is communicated with here. So it should be possible to support any TEE or Secure Partition using FF-A as transport with this mediator.

Quoted 4 lines

[1] https://developer.arm.com/documentation/den0077/latest

Thanks,
Jens

Quoted 7 lines

Hi Jens,

Many thanks for the patches! I tried to apply them to the master branch but unfortunately they don't apply any longer. Could you please rebase them on master (or even better rebase them on staging) and resend?

Thank you!

One question without having looked at the patches in details. These patches are necessary to mediate OS (e.g. Linux) interactions with OPTEE. The difference between xen/arch/arm/ffa.c and xen/arch/arm/tee/optee.c is the transport mechanism: shared mem vs. SMC. Is that right?

If only the transport is different, would it make sense to place ffa.c under xen/arch/arm/tee?

Without having looked at the details of the transport or the FF-A protocol let me ask you a question. Do you think it would be possible to share part of the implementation with xen/arch/arm/tee/optee.c? I am asking because intuitively, if only the transport is differenti I would have thought that some things could be common. But it doesn't look like the current patches are reusing anything from xen/arch/arm/tee/optee.c. Are the two protocols too different?

#5 lore raw
Quoted 35 lines · Stefano Stabellini
Quoted 25 lines · Stefano Stabellini
Quoted 19 lines · Jens Wiklander

Hi,

This patch sets add a FF-A [1] mediator modeled after the TEE mediator already present in Xen. The FF-A mediator implements the subset of the FF-A 1.1 specification needed to communicate with OP-TEE using FF-A as transport mechanism instead of SMC/HVC as with the TEE mediator. It allows a similar design in OP-TEE as with the TEE mediator where OP-TEE presents one virtual partition of itself to each guest in Xen.

The FF-A mediator is generic in the sense it has nothing OP-TEE specific except that only the subset needed for OP-TEE is implemented so far. The hooks needed to inform OP-TEE that a guest is created or destroyed is part of the FF-A specification.

It should be possible to extend the FF-A mediator to implement a larger portion of the FF-A 1.1 specification without breaking with the way OP-TEE is communicated with here. So it should be possible to support any TEE or Secure Partition using FF-A as transport with this mediator.

Quoted 4 lines

[1] https://developer.arm.com/documentation/den0077/latest

Thanks,
Jens

Quoted 7 lines

Hi Jens,

Many thanks for the patches! I tried to apply them to the master branch but unfortunately they don't apply any longer. Could you please rebase them on master (or even better rebase them on staging) and resend?

Thank you!

Quoted 8 lines

One question without having looked at the patches in details. These patches are necessary to mediate OS (e.g. Linux) interactions with OPTEE. The difference between xen/arch/arm/ffa.c and xen/arch/arm/tee/optee.c is the transport mechanism: shared mem vs. SMC. Is that right?

If only the transport is different, would it make sense to place ffa.c under xen/arch/arm/tee?

FF-A is a standard Arm service with a different range of SMC Function Identifiers. FF-A is used to communicate with SPs, Secure Partitions. An SP can be a TEE but also something different.

There are similarities between xen/arch/arm/ffa.c and xen/arch/arm/tee/optee.c, but I believe it's more in broader terms on the surface.

FF-A is a generic transport protocol that is generic enough that Xen doesn't even need to know anything about the entity it's facilitating communication with except what's provided by FF-A. The idea is that it's only the end points that need to be aware of details of the protocol run on top of FF-A. This means that EL2 (Xen), EL3 (SPMD/TF-A) and S-EL2 (SPMC/Hafnium) in principle can be kept unchanged and agnostic to Trusted OSes and what not.

Quoted 8 lines

Without having looked at the details of the transport or the FF-A protocol let me ask you a question. Do you think it would be possible to share part of the implementation with xen/arch/arm/tee/optee.c? I am asking because intuitively, if only the transport is differenti I would have thought that some things could be common. But it doesn't look like the current patches are reusing anything from xen/arch/arm/tee/optee.c. Are the two protocols too different?

The two protocols are quite different. So far I haven't been able to identify suitable common helper functions.

Thanks,
Jens

#6 lore raw
In reply to Stefano Stabellini

Hi Stefano,

Quoted 23 lines · Stefano Stabellini
Quoted 21 lines · Jens Wiklander

Hi,

This patch sets add a FF-A [1] mediator modeled after the TEE mediator already present in Xen. The FF-A mediator implements the subset of the FF-A 1.1 specification needed to communicate with OP-TEE using FF-A as transport mechanism instead of SMC/HVC as with the TEE mediator. It allows a similar design in OP-TEE as with the TEE mediator where OP-TEE presents one virtual partition of itself to each guest in Xen.

The FF-A mediator is generic in the sense it has nothing OP-TEE specific except that only the subset needed for OP-TEE is implemented so far. The hooks needed to inform OP-TEE that a guest is created or destroyed is part of the FF-A specification.

It should be possible to extend the FF-A mediator to implement a larger portion of the FF-A 1.1 specification without breaking with the way OP-TEE is communicated with here. So it should be possible to support any TEE or Secure Partition using FF-A as transport with this mediator.

[1] https://developer.arm.com/documentation/den0077/latest

Quoted 8 lines
Quoted 2 lines

Thanks,
Jens

Hi Jens,

Many thanks for the patches! I tried to apply them to the master branch but unfortunately they don't apply any longer. Could you please rebase them on master (or even better rebase them on staging) and resend?

No problem, I'll rebase and send out a v2.

Thanks,
Jens

Quoted 23 lines

Thank you!

Quoted 17 lines

Jens Wiklander (2):
xen/arm: smccc: add support for SMCCCv1.2 extended input/output
registers
xen/arm: add FF-A mediator

 xen/arch/arm/Kconfig         |   11 +
 xen/arch/arm/Makefile        |    1 +
 xen/arch/arm/arm64/smc.S     |   43 +
 xen/arch/arm/domain.c        |   10 +
 xen/arch/arm/ffa.c           | 1624 ++++++++++++++++++++++++++++++++++
 xen/arch/arm/vsmc.c          |   19 +-
 xen/include/asm-arm/domain.h |    4 +
 xen/include/asm-arm/ffa.h    |   71 ++
 xen/include/asm-arm/smccc.h  |   42 +
 9 files changed, 1821 insertions(+), 4 deletions(-)
 create mode 100644 xen/arch/arm/ffa.c
 create mode 100644 xen/include/asm-arm/ffa.h