etwen/Blanton_TTL_Script
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 18 Wiki Activity

18 Releases 18 Tags

RSS Feed
  • V1.1.5 215fed6e95
    Compare

    v1.1.5 — The stress jobs now last the whole soak Stable

    etwen released this 2026-09-01 11:04:55 +08:00 | 19 commits to main since this release

    🐛 Bug fixes

    USB stress ran for sixty seconds of an overnight soak

    • qfx5252-stress-usb was started with --runtime 60. It finished a minute in and the USB path sat idle for the rest of the night. Nothing failed, nothing said so, and the run looked complete.
    • SSD stress had the same shape at --runtime 3600 — one hour of a soak that usually runs far longer.
    • Both are now --runtime 86400, and the BMC DDR memtester lost its loop count (64M 1 → 64M) so it runs until it is stopped rather than finishing one pass.
    • 86400 is 24 hours, not "forever". A soak longer than a day leaves SSD and USB dead for the remainder. Script B prints bgctl list every round (since v1.0.10), so the shorter list is visible there — but nothing raises an alarm.

    ✨ New features

    Multiphase VRM margin

    • New blanton_multiphase_margin.sh drives the multiphase controllers on both switch boards over the native i2c buses — SWB0 on /dev/i2c-13 and -15, SWB1 on -14 and -16. 19 rails per board, 38 in all.

    • high3 / low3 / comboA / comboB set the voltages and report one line per rail:

      SWB0_0V9_ANLG0 - 0.930V Configure Done
SWB0_0V75_ANLG0 - 0.775V Configure Done
...
[ OK ] high3: 38 rail(s) configured

    • save commits them. It has no per-rail mapping, so it reports as one operation: Multiphase Save All.

    • -u 0 / -u 1 for a single board; rails belonging to the other board are skipped and counted separately, not failed.

    • save writes NVM permanently (PMBus STORE_USER_ALL, and 0x17 on the TH6 core rail). A power cycle does not undo it.

    • These are not the LTC2980s that margin.sh drives. The two tools reach the same rails from different sides and each assumes it is the only writer. There is no interlock — do not run both.

    Traffic bursts carry a random pattern

    • blanton_traffic_linespeed v0.7.0 appends PatternRandom=yes to every tx, ce0 included. A fixed pattern can sit on a benign bit sequence for an entire run; a changing one is what makes a weak lane show itself.
    • -r no forces the fixed pattern; -r "" leaves the argument off entirely, which is the pre-v0.7.0 wire format if a future SDK rejects it.
    • Defaults change from tx 100 length=512 to tx 200 length=324.
    • Counter totals are not comparable across this build: 64800 bytes per burst instead of 51200, about 27% more. The PASS/FAIL cross-check compares a pair against each other and is unaffected — only absolute figures moved.

    📦 Downloads

    File Contents
    Script_ABC_Blanton_V1.1.5.zip The full Tera Term working directory, including Blanton_Script/ to copy onto the DUT

    After copying to the DUT:

    chmod +x ~/Blanton_Script/*.sh      # see below
grep -rlU $'\r' ~/Blanton_Script    # expect no output

    blanton_ber.sh, nfc_polling.sh, blanton_tr518.sh, blanton_multiphase_margin.sh and the two TH6_SWB*_power_readback.sh are invoked through ./ or ~/, so they need the execute bit along with bmc_monitor.sh and usb_target.sh — neither git nor a Windows/USB copy carries it reliably.

    ⚠️ Before you run

    • A soak longer than 24 hours needs the runtimes raised. --runtime 86400 is a ceiling; past it, SSD and USB are idle and only a shorter bgctl list hints at it.
    • blanton_multiphase_margin.sh save is permanent. Read the rails back before committing, and never run it during a test.
    • Do not mix blanton_multiphase_margin.sh and margin.sh — same rails, two different access paths, no interlock.
    • Traffic counters changed scale. Any absolute thresholds taken from earlier logs need redoing.
    • Script 4 (thermal/safety) has no teardown — finish with blanton_tr518.sh stop and kill $(jobs -p).
    • These three cannot share the switch: blanton_ber.sh, blanton_tr518.sh, blanton_traffic_linespeed.
    • Margin is only sampled at the start and end of a soak (since v1.1.4), not during it.
    • FAN_SPEED is 100 — the fans run flat out unless you change config.ttl.
    • The three items listed as unverified in v1.1.0 are still unverified: the ce0 ingress mirror, the 120s settle in blanton_tr518.sh, and test mode nr=no on stop.

    🔗 Links

    Full changelog: V1.1.4...V1.1.5

    Downloads