M2354: add TRNG to device_has list

[ghost]

Summary of changes

The NU_M2354 target does have a TRNG but it is missing from the device_has list
in targets.json. This commit updates this file such that DEVICE_TRNG is defined
for NU_M2354. No further additions are required to port the M2345's true random
number generator to Mbed OS.

Impact of changes

Migration actions required

Documentation

None

---

Pull request type

[x] Patch update (Bug fix / Target update / Docs update / Test update / Refactor)
[ ] Feature update (New feature / Functionality change / New API)
[ ] Major update (Breaking change E.g. Return code change / API behaviour change)

---

Test results

[ ] No Tests required for this change (E.g docs only update)
[x] Covered by existing mbed-os tests (Greentea or Unittest)
[ ] Tests / results supplied as part of this PR

---

Reviewers

@MarceloSalazar
@ccli8

---

[LDong-Arm]

LGTM.

The TRNG implementation is based on PSA where psa_generate_random is provided by TF-M.

Maybe in the future we can consider globally enabling TRNG on TF-M targets in Mbed OS - are TF-M targets required/guaranteed to have it? @noonfom

[0xc0170]

Was this tested on hw or if we don't have it currently, can @ARMmbed/team-nuvoton help us?

[MarceloSalazar]

@noonfom as you have HW, can you please run the corresponding Greentea test and share the logs?

mbed test -t <toolchain> -m <target> -v -n *mbed-os-tests-mbed_hal-trng*

[ghost]

@LDong-Arm TF-M targets are not required/guaranteed to have a TRNG, e.g., AN521 uses a PRNG.

@0xc0170 I built and ran mbed-os-example-sockets with a TLS socket. That's about all the testing I've done so far.

@MarceloSalazar Sure, I'll run the Greentea test this afternoon.

[ccli8]

M2354 has TRNG implementation on TF-M side. And it gets available through PSA TRNG on Mbed.

Actually, this PR isn't necessary. Looking into mbed-os/targets/targets.json, NU_M2354 target has had TRNG through inheriting PSA_V8_M target:

mbed-os/targets/targets.json

Lines 141 to 152 in 3319beb

	"PSA_V8_M": {
	"inherits": [
	"PSA_Target"
	],
	"extra_labels": [
	"TFM",
	"TFM_LATEST",
	"TFM_V8M"
	],
	"device_has": [
	"TRNG"
	],

[ccli8]

Actually, this PR isn't necessary...

Sorry, missing one point. DEVICE_TRNG isn't defined without change. The suggested fix should be: change device_has with device_has_add. Then TRNG will get in through inheriting PSA_V8_M target.

mbed-os/targets/targets.json

Lines 6997 to 7002 in 3319beb

	"device_has": [
	"USTICKER",
	"LPTICKER",
	"RTC",
	"ANALOGIN",
	"ANALOGOUT",

[ccli8]

@noonfom How about changedevice_has with device_has_add for adding TRNG in?

[LDong-Arm]

@noonfom As @ccli8 noted, the base target PSA_V8_M declares TRNG via device_has but it doesn't work as it gets replaced by each target's own device_has.

@LDong-Arm TF-M targets are not required/guaranteed to have a TRNG, e.g., AN521 uses a PRNG.

If TRNG is not guaranteed by TF-M, maybe we should remove it from PSA_V8_M and add it to individual PSA_V8_M targets. This would also mean device_has_add -> device_has in each target, as there would be not be a base to "add" to anymore.

[ghost]

Thanks @ccli8 and @LDong-Arm. Closing in favour of #14807.