[sival,triag] Some SiVal ROM_EXT are broken #21706

pamaury · 2024-02-27T09:39:36Z

Description

The following tests have been found to be broken in //hw/top_earlgrey:fpga_cw310_sival_rom_ext execution environment. See #21580 for details. This issue tracks the lists for further triaging.

edn_kat_fpga_cw310_sival_rom_ext -> fixed
flash_ctrl_mem_protection_test_fpga_cw310_sival_rom_ext
flash_ctrl_test_fpga_cw310_sival_rom_ext
pmp_smoketest_tor_fpga_cw310_sival_rom_ext
rv_core_ibex_address_translation_test_fpga_cw310_sival_rom_ext
sram_ctrl_execution_test_fpga_cw310_sival_rom_ext
uart_tx_rx_test_fpga_cw310_sival_rom_ext
keymgr_test_fpga_cw310_sival_rom_ext
boot_data_functest_fpga_cw310_sival_rom_ext
irq_asm_functest_fpga_cw310_sival_rom_ext
alert_functest_fpga_cw310_sival_rom_ext
keymgr_functest_fpga_cw310_sival_rom_ext
rnd_functest_fpga_cw310_sival_rom_ext
rstmgr_functest_fpga_cw310_sival_rom_ext
watchdog_functest_fpga_cw310_sival_rom_ext
mod_exp_otbn_functest_hardcoded_fpga_cw310_sival_rom_ext
rsa_verify_functest_fpga_cw310_sival_rom_ext
sigverify_dynamic_functest_hardcoded_fpga_cw310_sival_rom_ext
chip_power_idle_load_fpga_cw310_sival_rom_ext
chip_power_sleep_load_fpga_cw310_sival_rom_ext
csrng_edn_concurrency_test_fpga_cw310_sival_rom_ext -> fixed
csrng_kat_test_fpga_cw310_sival_rom_ext -> fixed
entropy_src_csrng_test_fpga_cw310_sival_rom_ext -> fixed
entropy_src_edn_reqs_test_fpga_cw310_sival_rom_ext -> fixed
flash_ctrl_clock_freqs_test_fpga_cw310_sival_rom_ext
flash_ctrl_ops_test_fpga_cw310_sival_rom_ext
keymgr_key_derivation_test_fpga_cw310_sival_rom_ext
keymgr_sideload_aes_test_fpga_cw310_sival_rom_ext
keymgr_sideload_kmac_test_fpga_cw310_sival_rom_ext
lc_ctrl_otp_hw_cfg0_test_fpga_cw310_sival_rom_ext
otp_ctrl_smoketest_fpga_cw310_sival_rom_ext
pmp_smoketest_napot_fpga_cw310_sival_rom_ext
sram_ctrl_sleep_sram_ret_contents_scramble_test_fpga_cw310_sival_rom_ext
aes_kwp_sideload_functest_fpga_cw310_sival_rom_ext
aes_sideload_functest_fpga_cw310_sival_rom_ext
ecdh_p256_sideload_functest_fpga_cw310_sival_rom_ext
ecdsa_p256_sideload_functest_fpga_cw310_sival_rom_ext
kdf_kmac_sideload_functest_hardcoded_fpga_cw310_sival_rom_ext
kmac_sideload_functest_hardcoded_fpga_cw310_sival_rom_ext
pwrmgr_random_sleep_all_reset_reqs_test_fpga_cw310_sival_rom_ext
pwrmgr_all_reset_reqs_test_fpga_cw310_sival_rom_ext

There are some common patterns but each test needs to be investigated:

some are enabled in ROM_EXT but cannot actually run in ROM_EXT because some registers are locked: those should disable the //hw/top_earlgrey:fpga_cw310_sival_rom_ext and enable the //hw/top_earlgrey:fpga_cw310_sival exec env instead
some make assumptions about the OTP values in //hw/top_earlgrey:fpga_rom_with_fake_keys but the values in the sival ROM_EXT are different

The corresponding sival issues should be reopened.

The text was updated successfully, but these errors were encountered:

pamaury · 2024-02-27T09:39:51Z

Tag @johngt for visibility

…ival ROM_EXT This test also seems to be part of issue lowRISC#21706, so this commit marks it as broken in sival ROM_EXT. This should fix the currently failing CI checks on `master`. Signed-off-by: Andreas Kurth <[email protected]>

The following tests also seem to be part of issue lowRISC#21706: - //sw/device/tests:pwrmgr_all_reset_reqs_test - //sw/device/tests:pwrmgr_random_sleep_all_reset_reqs_test Thus this commit marks them as broken in SiVal ROM_EXT. This should contribute to fixing the currently failing CI checks on `master`. Signed-off-by: Andreas Kurth <[email protected]>

The following tests also seem to be part of issue #21706: - //sw/device/tests:pwrmgr_all_reset_reqs_test - //sw/device/tests:pwrmgr_random_sleep_all_reset_reqs_test Thus this commit marks them as broken in SiVal ROM_EXT. This should contribute to fixing the currently failing CI checks on `master`. Signed-off-by: Andreas Kurth <[email protected]>

Also fix minor typos in other pwrmgr top-level test code. Addresses lowRISC#21706 Signed-off-by: Guillermo Maturana <[email protected]>

Also fix minor typos in other pwrmgr top-level test code. Addresses #21706 Signed-off-by: Guillermo Maturana <[email protected]>

matutem · 2024-03-01T17:15:18Z

Also, the test pwrmgr_all_reset_reqs_test_fpga_cw310_sival_rom_ext is passing for me.

nasahlpa · 2024-04-02T17:40:47Z

Can confirm that edn_kat_fpga_cw310_sival_rom_ext fails for me. Is there a plan of fixing this test or any idea what is going wrong such that I can fix it?

jwnrt · 2024-04-16T08:49:29Z

We found that sram_ctrl_sleep_sram_ret_contents_scramble_test passes in the test_rom and rom_with_fake_keys environments on the FPGA, but not in the sival or sival_rom_ext environments. The no_scramble variant of the test always passes.

@matute do you know of any OTP configuration that might cause the scramble version to fail on in the sival configurations?

engdoreis · 2024-04-19T12:41:22Z

@matutem I'm looking into pwrmgr_all_reset_reqs_test_fpga_cw310_sival_rom_ext and it is actually flaky. Sometimes an unexpected reset (0x40 = kDifRstmgrResetInfoEscalation) occurs and mess up with the test.
This PR should fix it.

The `otbn_boot_services_functest` must run in the ROM_EXT slot since it manipulates the keygmr state, which normally is done by the ROM_EXT. This partially addresses lowRISC#21706. Signed-off-by: Tim Trippel <[email protected]>

The `otbn_boot_services_functest` must run in the ROM_EXT slot since it manipulates the keygmr state, which normally is done by the ROM_EXT. This partially addresses #21706. Signed-off-by: Tim Trippel <[email protected]>

The `otbn_boot_services_functest` must run in the ROM_EXT slot since it manipulates the keygmr state, which normally is done by the ROM_EXT. This partially addresses lowRISC#21706. Signed-off-by: Tim Trippel <[email protected]> (cherry picked from commit 698b5a6)

The `otbn_boot_services_functest` must run in the ROM_EXT slot since it manipulates the keygmr state, which normally is done by the ROM_EXT. This partially addresses #21706. Signed-off-by: Tim Trippel <[email protected]> (cherry picked from commit 698b5a6)

These tests are all working in ROM_EXT on master and earlgrey_es_sival now that we've fixed the underlying keymgr issue. For details, see lowRISC#22819 (comment) This is related to lowRISC#21706 and lowRISC#22140. Signed-off-by: Pirmin Vogel <[email protected]>

These tests are all working in ROM_EXT on master and earlgrey_es_sival now that we've fixed the underlying keymgr issue. For details, see #22819 (comment) This is related to #21706 and #22140. Signed-off-by: Pirmin Vogel <[email protected]>

These tests are all working in ROM_EXT on master and earlgrey_es_sival now that we've fixed the underlying keymgr issue. For details, see lowRISC#22819 (comment) This is related to lowRISC#21706 and lowRISC#22140. Signed-off-by: Pirmin Vogel <[email protected]>

These tests are all working in ROM_EXT on master and earlgrey_es_sival now that we've fixed the underlying keymgr issue. For details, see #22819 (comment) This is related to #21706 and #22140. Signed-off-by: Pirmin Vogel <[email protected]>

These tests are all working in ROM_EXT on master and earlgrey_es_sival now that we've fixed the underlying keymgr issue. For details, see lowRISC#22819 (comment) This is related to lowRISC#21706 and lowRISC#22140. Signed-off-by: Pirmin Vogel <[email protected]>

keymgr_key_derivation_test_fpga_cw310_sival_rom_ext test was marked as broken in lowRISC#21706. This commit fixes this and addresses lowRISC#21500. The source of the bug is due to the nature of the ROM_EXT that runs through a CDI attestation flow advancing the key manager state to "OwnerRootKey" in the process. This was not accounted for in the test, which assumed an active CreatorRootKey" state (this works for sim_dv tests but not with ROM_EXT). The solution to this issue lies in rendering the test state-agnostic, meaning it will execute its verifications independently of the current key manager state. This approach thus mimics the keymgr_sideload_{aes,kmac,otbn} tests which follow a similar strategy. Signed-off-by: Andrea Caforio <[email protected]>

keymgr_key_derivation_test_fpga_cw310_sival_rom_ext test was marked as broken in lowRISC#21706. This commit fixes this and addresses lowRISC#21500. The source of the bug is due to the nature of the ROM_EXT that runs through a CDI attestation flow advancing the key manager state to "OwnerRootKey" in the process. This was not accounted for in the test, which assumed an active "CreatorRootKey" state (this works for sim_dv tests but not with ROM_EXT). The solution to this issue lies in rendering the test state-agnostic, meaning it will execute its verifications independently of the current key manager state. This approach thus mimics the keymgr_sideload_{aes,kmac,otbn} tests which follow a similar strategy. Signed-off-by: Andrea Caforio <[email protected]>

keymgr_key_derivation_test_fpga_cw310_sival_rom_ext test was marked as broken in #21706. This commit fixes this and addresses #21500. The source of the bug is due to the nature of the ROM_EXT that runs through a CDI attestation flow advancing the key manager state to "OwnerRootKey" in the process. This was not accounted for in the test, which assumed an active "CreatorRootKey" state (this works for sim_dv tests but not with ROM_EXT). The solution to this issue lies in rendering the test state-agnostic, meaning it will execute its verifications independently of the current key manager state. This approach thus mimics the keymgr_sideload_{aes,kmac,otbn} tests which follow a similar strategy. Signed-off-by: Andrea Caforio <[email protected]>

keymgr_key_derivation_test_fpga_cw310_sival_rom_ext test was marked as broken in #21706. This commit fixes this and addresses #21500. The source of the bug is due to the nature of the ROM_EXT that runs through a CDI attestation flow advancing the key manager state to "OwnerRootKey" in the process. This was not accounted for in the test, which assumed an active "CreatorRootKey" state (this works for sim_dv tests but not with ROM_EXT). The solution to this issue lies in rendering the test state-agnostic, meaning it will execute its verifications independently of the current key manager state. This approach thus mimics the keymgr_sideload_{aes,kmac,otbn} tests which follow a similar strategy. Signed-off-by: Andrea Caforio <[email protected]> (cherry picked from commit aa4031d)

pamaury mentioned this issue Feb 27, 2024

[ci] Avoid running duplicate hyperdebug-based tests #21580

Merged

This was referenced Feb 29, 2024

[sw/tests] Mark more tests as broken in SiVal ROM_EXT #21751

Merged

[vendor/pulp_riscv_dbg] Update to pulp-platform/riscv-dbg@358f901 #21719

Merged

matutem added a commit to matutem/opentitan that referenced this issue Mar 1, 2024

[sival,pwrmgr] Fix pwrmgr_random_sleep_all_reset_reqs_test

e84f290

Also fix minor typos in other pwrmgr top-level test code. Addresses lowRISC#21706 Signed-off-by: Guillermo Maturana <[email protected]>

matutem mentioned this issue Mar 1, 2024

[sival,pwrmgr] Fix pwrmgr_random_sleep_all_reset_reqs_test #21782

Merged

jadephilipoom mentioned this issue Mar 1, 2024

[boot_services] Add functional tests for the OTBN boot services library. #21717

Merged

matutem added a commit that referenced this issue Mar 1, 2024

[sival,pwrmgr] Fix pwrmgr_random_sleep_all_reset_reqs_test

bf0bb58

Also fix minor typos in other pwrmgr top-level test code. Addresses #21706 Signed-off-by: Guillermo Maturana <[email protected]>

This was referenced Mar 4, 2024

[pwrmgr] D2(S) Signoff #20990

Closed

[pwrmgr] V2(S) Signoff #21025

Closed

andreaskurth mentioned this issue Mar 28, 2024

[keymgr] D2(S) Signoff #20981

Closed

This was referenced Mar 29, 2024

[edn] D2S Signoff #20982

Closed

[csrng] D2S Signoff #20976

Closed

vogelpi mentioned this issue May 9, 2024

[sival] Mark fixed entropy complex tests as unbroken in ROM_EXT #23035

Merged

vogelpi mentioned this issue May 10, 2024

[sival] Mark fixed entropy complex tests as unbroken in ROM_EXT #23048

Merged

vogelpi mentioned this issue Jun 7, 2024

[edn] V2S Signoff #22468

Closed

vogelpi mentioned this issue Jun 25, 2024

[csrng] V2S Signoff #22467

Closed

andrea-caforio mentioned this issue Oct 9, 2024

[sival, keymgr] State-agnostic chip_sw_key_derivation_test #24755

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[sival,triag] Some SiVal ROM_EXT are broken #21706

[sival,triag] Some SiVal ROM_EXT are broken #21706

pamaury commented Feb 27, 2024 •

edited by vogelpi

Loading

pamaury commented Feb 27, 2024

matutem commented Mar 1, 2024

nasahlpa commented Apr 2, 2024

jwnrt commented Apr 16, 2024

engdoreis commented Apr 19, 2024 •

edited

Loading

[sival,triag] Some SiVal ROM_EXT are broken #21706

[sival,triag] Some SiVal ROM_EXT are broken #21706

Comments

pamaury commented Feb 27, 2024 • edited by vogelpi Loading

Description

pamaury commented Feb 27, 2024

matutem commented Mar 1, 2024

nasahlpa commented Apr 2, 2024

jwnrt commented Apr 16, 2024

engdoreis commented Apr 19, 2024 • edited Loading

pamaury commented Feb 27, 2024 •

edited by vogelpi

Loading

engdoreis commented Apr 19, 2024 •

edited

Loading