ci: unify zone1 console access and add console severity checks

[dallasxy]

This PR adds two CI reliability improvements:

1.Unified zone1 console handling
Improve how the zone1 console is attached and driven so zone0 and zone1 use the same access path. This reduces flaky CI failures caused by unreliable zone1 serial/console input (for example screen/pty attach and Enter handling).

2.Post-run WARN / ERROR / panic log checks
Persist each BID’s console output to logs/__.log and add a Jenkins stage that scans it after tests. WARN lines are reported; ERROR/CRIT and panic markers fail the job.

Previously CI only checked functional case outcomes and could miss serious hvisor failures. A concrete example on aarch64/qemu-gicv3: zone1 was configured with CPUs 3,4, but QEMU only provides CPUs 0–3. That invalid mapping triggered an hvisor panic at runtime, yet CPU 3 still worked enough for zone1 cases to pass, so CI reported success. With console severity checking, such ERROR/panic lines are surfaced and the pipeline fails.

3.add a script jenkins/run_ci.sh to run ci locally

[agicy]

Merge upstream/dev the dev

I suggest rebasing on upstream/dev instead of merging.

[agicy]

• The Oops: / BUG: alternatives can never match: the trailing \b after : requires a word character to follow, but real messages have a space. Verified no-match for Oops: 0002 [#1] SMP and BUG: unable to handle kernel NULL pointer dereference — i.e. kernel oopses/BUGs, exactly the failures this gate exists for, will not fail the job.
• Conversely, the bare PANIC alternative matches any standalone occurrence of the word "panic" (verified: Kernel command line: ... panic=10, and mode-panic = <0x5242c307> — a DT property present in the rk3568 DT this PR re-enables). A healthy run can be failed by a DT/cmdline echo.