HPt630 tests #2

philipandag · 2025-02-05T09:47:47Z

No description provided.

Signed-off-by: Filip Gołaś <[email protected]>

krystian-hebel · 2025-02-05T10:02:04Z

generalhw/hpt630v1/README.md

+Perform installation followed by AEM testing:
+
+```
+openqa-cli api -X POST isos DISTRI=qubesos VERSION=4.2 ARCH=x86_64 BUILD=4.2.0 FLAVOR=install-iso-optiplex


It would be better to use versions we're testing it with, and also, optiplex?

Sorry, it was so long since I copy-pasted this README from optiplex that I forgot that it still has to be changed. I'll look through the whole generalhw/hpt630 again

https://github.com/3mdeb/openqa-tests-qubesos/compare/a39198afd3230fdf0e453e356c1fe72dd65ddd95..c1f63647bdb848a222fd9a6d97cbef1d5a4970c9#diff-effe13d6dacca60a03019f6a5c56dba5883bea6674fdbaa74a19f68713474974L14-R14

krystian-hebel · 2025-02-05T10:09:04Z

generalhw/hpt630v1/README.md

+* `PACKAGES_BASE_URL` - where to look for AEM-related packages.
+* `AEM_VER` - version of `anti-evil-maid` package
+* `GRUB_VER` - version of `grub2-*` packages
+* `XEN_VER` - version of `xen-*` packages


Missing SKL_VER.

https://github.com/3mdeb/openqa-tests-qubesos/compare/a39198afd3230fdf0e453e356c1fe72dd65ddd95..c1f63647bdb848a222fd9a6d97cbef1d5a4970c9#diff-effe13d6dacca60a03019f6a5c56dba5883bea6674fdbaa74a19f68713474974R41

krystian-hebel · 2025-02-05T10:19:47Z

tests/aem_hw.pm

@@ -50,17 +50,20 @@ use Data::Dumper;
 # they are generally machine-specific, so don't assume anything
 my $boot_disk;
 my $boot_part;
-if (check_var('MACHINE', 'optiplex') || check_var('MACHINE', 'supermicro')) {
+if ((check_var('MACHINE', 'optiplex') || check_var('MACHINE', 'supermicro') || check_var('MACHINE', 'hpt630v1')) and (check_var('OS_INSTALL_LEGACY', '1'))) {


I think it should be moved below DRTM and BIOS kind checks, and then we could test just OS_INSTALL_LEGACY value, as all other platforms would be filtered out by BIOS/DRTM checks.

Like that?

https://github.com/3mdeb/openqa-tests-qubesos/compare/c1f63647bdb848a222fd9a6d97cbef1d5a4970c9..2ad98a32caefbfbb51ae8f4ff16a35a8ae849664#diff-5bcc1c5dc564aac714d0933cac3e9cb6d906077998db688ae82132d224af29acR71-R82

tests/aem_hw.pm

krystian-hebel · 2025-02-05T10:32:54Z

generalhw/hpt630v1/power

+# echo "$0: not doing anything"
+# exit 0
+
+function sonoff_pwr() {


We probably should add some delay after power off.

I haven't encountered any issues with powering the platform on/off, but it surely won't hurt.
It's important to not add any delay when powering on, or the screen with the HP logo might not get registered

https://github.com/3mdeb/openqa-tests-qubesos/compare/52d01ba%5E..52d01ba

krystian-hebel · 2025-02-05T10:34:09Z

generalhw/hpt630v1/README.md

+
+These settings can be added to `openqa-cli` posting command to specify which
+packages to use.  Current default uses
+<https://dl.3mdeb.com/open-source-firmware/QubesOS/trenchboot_aem_v0.3/>.


Update according to whatever will be used in template. The same may be needed for other platforms, so maybe this should be made generic and point to template so we won't have to bump it with each new version.

https://github.com/3mdeb/openqa-tests-qubesos/compare/c4107f9%5E..c4107f9

I am not sure about MSI. The README and the instructions to start the tests are different than for other platforms.
I think it should probably still be fine to override the parameters like in the case of the other platforms.

krystian-hebel · 2025-02-05T10:40:58Z

generalhw/hpt630v1/ks.cfg

+
+# allow all USB devices, this setting seems to appear on first boot, hence
+# the update is performed here
+sed -i -e 's/authorized_default=0/authorized_default=1/' /boot/grub2/grub.cfg /etc/default/grub


For this platform, we should also add spec-ctrl=no-ibpb-entry to Xen cmdline, otherwise VMs may not be able to start before timeout under some conditions. It will also make it much faster, and for testing we care more about time than about security implications this option has.

Does it look good?
https://github.com/3mdeb/openqa-tests-qubesos/compare/bba37a8%5E..bba37a8

I think you need to escape $ so it won't be parsed by shell executing echo. Maybe putting the whole string in single quotes and unescaping inner double quotes instead would be cleaner.

Right, it makes more sense if Grub interprets the variable.

https://github.com/3mdeb/openqa-tests-qubesos/compare/455291aad9cc99c2752ef8824ab92650daab4e20..d759221a46937501909ed0ad25f260e9530a568c

krystian-hebel · 2025-02-05T10:51:49Z

tests/aem_hw.pm

@@ -350,18 +452,31 @@ sub setup_aem {
    # cleanup in case AEM was previously initialized (useful for debugging this test)


(comment to code few lines above, but GH won't allow me to add it there)

Currently, install_packages fails on dnf reinstall step, because there is nothing to reinstall. Ideally we shouldn't use dnf directly, instead relying on qubes-dom0-update, as was done in blog posts. There still may be some install vs reinstall issues, and I don't know if assert_script_run is enough to catch all problems, like missing packages or version conflicts.

assert_script_run returns the command output, so there is no reason not to check it and use it to decide on fail/pass.
I'll see if using qubes-dom0-update helps and if there are some characteristic output meaning a success/fail.

@krystian-hebel
In the blog post we first update dom0 before installing anything. The tests are not doing that. Should a general update to the whole dom0 be added?

This is worth considering, but I don't think we should do it here and now. Let's focus on getting it to work in the first place.

krystian-hebel

Do we need v1 in the platform name? Ideally we should reuse the same setup for both platforms.

krystian-hebel · 2025-02-05T11:08:29Z

generalhw/hpt630v1/power

+    cmd=$(cut -d "p" -f2 <<< "$cmd" )
+    post_params="cmnd=Power $cmd"
+    echo "Sending sonoff POST request: $post_params\n"
+    curl -X POST http://192.168.10.157/cm -d "$post_params" && echo "\n"


Hardcoded IP, perhaps we should pass it as $1 in place of $rte_ip, AFAICT we're not using it for anything.

Right. Now I see that, for example, supermicro also has a different meaning of the first parameter. It's a good idea then.

https://github.com/3mdeb/openqa-tests-qubesos/compare/75fb8a7%5E..75fb8a7

Signed-off-by: Filip Gołaś <[email protected]>

(It's really slow...) Signed-off-by: Filip Gołaś <[email protected]>

Signed-off-by: Filip Gołaś <[email protected]>

philipandag · 2025-02-05T13:17:27Z

Do we need v1 in the platform name? Ideally we should reuse the same setup for both platforms.

The two platforms (v1 and v2) have different versions of TPM chips. Other than that I think they are identical, and the tests don't need to take that into account. If it is not important to clearly differentiate them when scheduling / presenting the test results, then I think this suffix can be removed without problems.

Signed-off-by: Filip Gołaś <[email protected]>

to make sure the motherboard discharges completely Signed-off-by: Filip Gołaś <[email protected]>

It's defined in the templates. It is enough to point there Signed-off-by: Filip Gołaś <[email protected]>

Signed-off-by: Filip Gołaś <[email protected]>

Co-authored-by: Krystian Hebel <[email protected]> Signed-off-by: Filip Gołaś <[email protected]>

Co-authored-by: Krystian Hebel <[email protected]>

philipandag added 4 commits January 27, 2025 13:09

3mdeb-templates.json: Update products to qubes R4.2.3

e04a46b

Signed-off-by: Filip Gołaś <[email protected]>

3mdeb-templates.json: Update AEM packages

e91f059

Signed-off-by: Filip Gołaś <[email protected]>

3mdeb-templates.json: Add new producsts & machines

11375e0

Signed-off-by: Filip Gołaś <[email protected]>

3mdeb-templates: Update jobgroups

d4cba72

Signed-off-by: Filip Gołaś <[email protected]>

krystian-hebel reviewed Feb 5, 2025

View reviewed changes

philipandag added 13 commits February 5, 2025 13:43

generalhw: hpt630v1: Add

c263ca0

Signed-off-by: Filip Gołaś <[email protected]>

generalhw/hpt630v1: Add +x to scripts

de34bea

Signed-off-by: Filip Gołaś <[email protected]>

generalhw/hpt630v1/power: Fix parsing action to sonoff api

4fd460d

Signed-off-by: Filip Gołaś <[email protected]>

3mdeb-templates: Update MAX_ISO_SIZE

23af81e

Signed-off-by: Filip Gołaś <[email protected]>

3mdeb-templates: Add "legacy_boot" key to jobs with legacy boot

75b3ec4

Signed-off-by: Filip Gołaś <[email protected]>

needles: HP GRUB bootmenu needles

3ee1206

Signed-off-by: Filip Gołaś <[email protected]>

install_partitioning_default: check for install over existing

298ab73

Signed-off-by: Filip Gołaś <[email protected]>

firstboot & 3mdeb-templates: Workaround for missing video in lukspass

6bc8db4

Signed-off-by: Filip Gołaś <[email protected]>

3mdeb-templates.json: Set TIMEOUT_SCALE for hpt630v1 machine

e42c64c

(It's really slow...) Signed-off-by: Filip Gołaś <[email protected]>

tests/aem_hw.pm: Support HPt630v1

ff4d340

Signed-off-by: Filip Gołaś <[email protected]>

3mdeb-templates: Add SKL_VER to aem-setup test suite

7311cd4

Signed-off-by: Filip Gołaś <[email protected]>

tests/aem_hw.pm: Running Qubes: update grub traversal

9d015ba

Signed-off-by: Filip Gołaś <[email protected]>

aem_hw.pm: Workaround for missing signal on lukspass

a39000e

Signed-off-by: Filip Gołaś <[email protected]>

philipandag force-pushed the qubes-r4.2.3-update branch from a39198a to c1f6364 Compare February 5, 2025 12:44

philipandag added 5 commits February 5, 2025 13:50

aem_hw: Set corect boot partition depending on legacy/efi boot

fd2a597

Signed-off-by: Filip Gołaś <[email protected]>

aem_hw: Do lukspass no signal workaround in loop to save time

c951bfb

Signed-off-by: Filip Gołaś <[email protected]>

aem_hw: Don't do grub2-install on efi systems

77a5104

Signed-off-by: Filip Gołaś <[email protected]>

needles: Add needles for aem_hw clearing TPM on HPt630v1

3d94db9

Signed-off-by: Filip Gołaś <[email protected]>

aem_hw.pm: Change poweroff to reboot on the end of run

2a79832

Signed-off-by: Filip Gołaś <[email protected]>

philipandag force-pushed the qubes-r4.2.3-update branch from c1f6364 to 2ad98a3 Compare February 5, 2025 12:50

philipandag force-pushed the qubes-r4.2.3-update branch from fcbbb2e to bba37a8 Compare February 5, 2025 14:15

philipandag added 2 commits February 5, 2025 15:36

aem_hw: clear_tpm_hp: Add verifying TPM is still enabled after clear

8cdd2a3

Signed-off-by: Filip Gołaś <[email protected]>

aem_hw: Adjust installed packages depending on Legacy/EFI boot

cdc14e6

Signed-off-by: Filip Gołaś <[email protected]>

philipandag added 7 commits February 5, 2025 15:36

needles: Add needles to verify TPM enabled on HP

540243e

Signed-off-by: Filip Gołaś <[email protected]>

3mdeb-templates: Fix XEN_VER for installation test suite

b1331d2

Signed-off-by: Filip Gołaś <[email protected]>

3mdeb-templates: Seperate test suites into Legacy/UEFI versions

ea4c6c8

Signed-off-by: Filip Gołaś <[email protected]>

3mdeb-templates: Update START_AFTER for EFI/legacy test variants

eebfab1

Signed-off-by: Filip Gołaś <[email protected]>

hpt630v1/power: Add delay after power off

620c76e

to make sure the motherboard discharges completely Signed-off-by: Filip Gołaś <[email protected]>

generalhw: Generalize hardcoded AEM packages version

5adeb88

It's defined in the templates. It is enough to point there Signed-off-by: Filip Gołaś <[email protected]>

hpt630v1/power: Use $1 for sonoff ip instead of hardcoding it

d0776fa

Signed-off-by: Filip Gołaś <[email protected]>

philipandag force-pushed the qubes-r4.2.3-update branch from bba37a8 to 455291a Compare February 5, 2025 14:36

hpt630v1/ks.cfg: Add spec-ctrl=no-ibpb-entry for xen

d759221

Co-authored-by: Krystian Hebel <[email protected]> Signed-off-by: Filip Gołaś <[email protected]>

philipandag force-pushed the qubes-r4.2.3-update branch from 455291a to d759221 Compare February 6, 2025 06:16

aem_hw.pm: Fix comment on optional grub2-install call

fa5d277

Co-authored-by: Krystian Hebel <[email protected]>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

HPt630 tests #2

HPt630 tests #2

philipandag commented Feb 5, 2025

krystian-hebel Feb 5, 2025

philipandag Feb 5, 2025

philipandag Feb 5, 2025

krystian-hebel Feb 5, 2025

philipandag Feb 5, 2025

krystian-hebel Feb 5, 2025

philipandag Feb 5, 2025

krystian-hebel Feb 5, 2025

philipandag Feb 5, 2025

krystian-hebel Feb 5, 2025

philipandag Feb 5, 2025

krystian-hebel Feb 5, 2025

philipandag Feb 5, 2025

krystian-hebel Feb 5, 2025 •

edited

Loading

philipandag Feb 6, 2025

krystian-hebel Feb 5, 2025

philipandag Feb 5, 2025

philipandag Feb 5, 2025

krystian-hebel Feb 5, 2025

krystian-hebel left a comment

krystian-hebel Feb 5, 2025

philipandag Feb 5, 2025

philipandag Feb 5, 2025

philipandag commented Feb 5, 2025

		@@ -350,18 +452,31 @@ sub setup_aem {
		# cleanup in case AEM was previously initialized (useful for debugging this test)

HPt630 tests #2

Are you sure you want to change the base?

HPt630 tests #2

Conversation

philipandag commented Feb 5, 2025

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

krystian-hebel Feb 5, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

krystian-hebel left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

philipandag commented Feb 5, 2025

krystian-hebel Feb 5, 2025 •

edited

Loading