Optee os update to 3.20+fio (take 2) #1059

ricardosalveti · 2023-03-02T20:33:04Z

No description provided.

Update plug-and-trust to the latest release. Relevant changes: - b4fbe94 scp03 challenge: do not call OP-TEE - 13497ec se050x: policy: do not modify requested policy - 67cac4f ecdsa: op-tee signature handling - 6068e92 ecdsa: support P-521 curve Signed-off-by: Igor Opaniuk <[email protected]>

Operations that require a public key might fallback to a software based implementation. Operations that require a private key might fallback to a software based implementation as long as the private key is not in the secure element. This helps making the software stack compatible with an extended crypto set, allowing xtests to work correctly. Signed-off-by: Ricardo Salveti <[email protected]>

Update OP-TEE to 3.20. Signed-off-by: Igor Opaniuk <[email protected]> Signed-off-by: Ricardo Salveti <[email protected]>

ricardosalveti · 2023-03-02T20:33:31Z

Follow up of #1019 (as igor is off today and I needed to do a bit more testing due some other stuff).

ricardosalveti · 2023-03-02T20:42:29Z

+-----------------------------------------------------
30376 subtests of which 0 failed
128 test cases of which 0 failed
0 test cases were skipped
TEE test application done!
root@imx8mq-evk:~#

ricardosalveti · 2023-03-02T23:13:16Z

+-----------------------------------------------------
30168 subtests of which 0 failed
128 test cases of which 0 failed
0 test cases were skipped
TEE test application done!
root@imx6ulevk:~#

ricardosalveti · 2023-03-03T00:15:58Z

With SE050:

+-----------------------------------------------------
30469 subtests of which 0 failed
127 test cases of which 0 failed
0 test cases were skipped
TEE test application done!
root@imx8mm-lpddr4-evk-sec:~#

ricardosalveti · 2023-03-03T01:02:41Z

+-----------------------------------------------------
31056 subtests of which 0 failed
128 test cases of which 0 failed
0 test cases were skipped
TEE test application done!
root@qemuarm64-secureboot:~#

ricardosalveti · 2023-03-03T01:50:42Z

+-----------------------------------------------------
30342 subtests of which 0 failed
128 test cases of which 0 failed
0 test cases were skipped
TEE test application done!
root@am64xx-evm:~#

ricardosalveti · 2023-03-03T03:02:36Z

+-----------------------------------------------------
30348 subtests of which 0 failed
128 test cases of which 0 failed
0 test cases were skipped
TEE test application done!
root@kv260:~#

MrCry0

UPDATE

Tested with main lmp-manifest +

<project name="meta-lmp" path="layers/meta-lmp" remote="ricardo" revision="optee"/>

30227 subtests of which 2 failed
128 test cases of which 1 failed
0 test cases were skipped
TEE test application done!
fio@imx6ullevk:~$

regression_1039.1 FAILED first error at /usr/src/debug/optee-test/3.20.0-r0/git/host/xtest/regression_1000.c:3206
regression_1039.2 FAILED first error at /usr/src/debug/optee-test/3.20.0-r0/git/host/xtest/regression_1000.c:3213
regression_1039 FAILED

MrCry0 · 2023-03-03T13:40:12Z

+-----------------------------------------------------
30376 subtests of which 2 failed
128 test cases of which 1 failed
0 test cases were skipped
TEE test application done!
fio@apalis-imx8:~$

regression_1039.1 FAILED first error at /usr/src/debug/optee-test/3.20.0-r0/git/host/xtest/regression_1000.c:3206
regression_1039.2 FAILED first error at /usr/src/debug/optee-test/3.20.0-r0/git/host/xtest/regression_1000.c:3213
regression_1039 FAILED

ricardosalveti · 2023-03-03T13:58:39Z

That is weird, both failed the same way. Will try on my imx6ullevk-sec as well.

ricardosalveti · 2023-03-03T13:59:09Z

@MrCry0 can you run just the failed jobs and get the job log?

MrCry0 · 2023-03-03T14:15:12Z

That is weird, both failed the same way. Will try on my imx6ullevk-sec as well.

This is really weird and seems a factory cache issue.

The error is fixed in optee_test commit c7f0b40 ("ta: makefile: add compilation of subkey TAs")

* regression_1039 Test subkey verification
o E/LD:  init_elf:486 sys_open_ta_bin(a720ccbb-51da-417d-b82e-e5445d474a7a)
regression_1039.1 Load TA with two levels of subkeys
/usr/srcE/TC:? 0 ldelf_init_with_ldelf:131 ldelf failed with res: 0xffff000f
/debug/optee-test/3.20.0-r0/git/host/xtest/regression_1000.c:3206: xtest_teec_open_session(&session, &subkey1_ta_uuid, ((void *)0), &ret_orig) has an unexpected value: 0xffff000f = TEEC_ERRS
  regression_1039.1 FAILED
o regression_1039.2 Load TA with identity subkey
/usr/src/debug/optee-test/3.20.0-r0/git/host/xtest/regression_1000.c:3213: xtest_teec_open_session(&session, &subkey2_ta_uuid, ((void *)0), &ret_orig) has an unexpected value: 0xffff000f = S
  regression_1039.2 FAILED
  regression_1039 FAILED

So I think our 3.20 is OK.
The commit c7f0b40 SHOULD have been included in the optee_test for ricardo:optee meta-lmp.

Well, and this is not a sstate cache issue:

Level 21: Sstate summary: Wanted 3548 Local 0 Mirrors 0 Missed 3548 Current 0 (0% match, 0% complete)

Building locally.

ricardosalveti · 2023-03-03T15:17:51Z

+-----------------------------------------------------
30109 subtests of which 2 failed
128 test cases of which 1 failed
0 test cases were skipped
TEE test application done!
root@stm32mp15-disco:~#

CKR_HOST_MEMORY, and I don't get this failure on every run.

* pkcs11_1007 PKCS11: Check operations release at session closure
/usr/src/debug/optee-test/3.20.0-r0/git/host/xtest/pkcs11_1000.c:1491: rv has an unexpected value: 0x2 = CKR_HOST_MEMORY, expected 0x0 = CKR_OK
/usr/src/debug/optee-test/3.20.0-r0/git/host/xtest/pkcs11_1000.c:1539: rv has an unexpected value: 0x2 = CKR_HOST_MEMORY, expected 0x0 = CKR_OK
  pkcs11_1007 FAILED

ricardosalveti · 2023-03-03T15:24:19Z

+-----------------------------------------------------
3717 subtests of which 0 failed
27 test cases of which 0 failed
0 test cases were skipped
TEE test application done!
root@stm32mp15-disco:~# xtest -t pkcs11

Just the pkcs11 tests work ok.

ricardosalveti · 2023-03-03T22:13:51Z

CKR_HOST_MEMORY, and I don't get this failure on every run.

It fails when trying to allocate SHM memory for the test case, which is just a normal shm alloc operation, and why it works OK when retrying or when just running the pkcs11 tests.

@igoropaniuk is this something you faced as well? I'm ok merging even with this failure but we should investigate this one a bit more (not an issue for the normal scenario as we never really stress optee this way).

ricardosalveti · 2023-03-03T22:35:08Z

+-----------------------------------------------------
30227 subtests of which 0 failed
128 test cases of which 0 failed
0 test cases were skipped
TEE test application done!
root@imx6ullevk-sec:~#

@MrCry0 working fine here.

ricardosalveti · 2023-03-03T22:35:57Z

Will merge so we get a wider testing on this (we can fix as needed).

MrCry0 · 2023-03-04T07:13:14Z

+-----------------------------------------------------
30227 subtests of which 0 failed
128 test cases of which 0 failed
0 test cases were skipped
TEE test application done!
root@imx6ullevk-sec:~#

@MrCry0 working fine here.

Yeah, I said - that is not an op-tee issue, but building one. Somehow factory built an older version of optee-test.

igoropaniuk and others added 3 commits March 2, 2023 15:15

base/bsp: update optee to 3.20

9c3f413

Update OP-TEE to 3.20. Signed-off-by: Igor Opaniuk <[email protected]> Signed-off-by: Ricardo Salveti <[email protected]>

ricardosalveti requested a review from a team March 2, 2023 20:33

ricardosalveti mentioned this pull request Mar 2, 2023

Optee os 3.20+fio testing #1019

Closed

quaresmajose approved these changes Mar 3, 2023

View reviewed changes

MrCry0 approved these changes Mar 3, 2023

View reviewed changes

MrCry0 self-requested a review March 3, 2023 13:38

angolini approved these changes Mar 3, 2023

View reviewed changes

ricardosalveti merged commit 292595e into foundriesio:main Mar 3, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Optee os update to 3.20+fio (take 2) #1059

Optee os update to 3.20+fio (take 2) #1059

ricardosalveti commented Mar 2, 2023

ricardosalveti commented Mar 2, 2023

ricardosalveti commented Mar 2, 2023

ricardosalveti commented Mar 2, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

MrCry0 left a comment •

edited

Loading

MrCry0 commented Mar 3, 2023 •

edited

Loading

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

MrCry0 commented Mar 3, 2023 •

edited

Loading

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

MrCry0 commented Mar 4, 2023

Optee os update to 3.20+fio (take 2) #1059

Optee os update to 3.20+fio (take 2) #1059

Conversation

ricardosalveti commented Mar 2, 2023

ricardosalveti commented Mar 2, 2023

ricardosalveti commented Mar 2, 2023

ricardosalveti commented Mar 2, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

MrCry0 left a comment • edited Loading

Choose a reason for hiding this comment

MrCry0 commented Mar 3, 2023 • edited Loading

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

MrCry0 commented Mar 3, 2023 • edited Loading

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

ricardosalveti commented Mar 3, 2023

MrCry0 commented Mar 4, 2023

MrCry0 left a comment •

edited

Loading

MrCry0 commented Mar 3, 2023 •

edited

Loading

MrCry0 commented Mar 3, 2023 •

edited

Loading