Repository navigation
device/tegu SLEEP-PLAN, bringup, AGENTS: the no-wake caught twice with a console, its cause, and #247 - #800
Conversation
…h a console, its cause, and derekross#247 The follow-up derekross#659 left for later. The no-wake on the published kernel came back on 2026-11-12 with the heavy debug switches off, twice, and the console with the device callback events on stops at the start of the noirq resume of an HSI0 USI on CPU 7. A bring-up module showed the cause awake on 11-13: a register access to a USI in the first microseconds after its clock gates are opened never returns. linux-omarchy-mobile#247 keeps the six gates on; on the published tip with only that change 300 of 300 sleeps woke, and touch and GNSS work. - SLEEP-PLAN: the status, the five runs of 11-12 and 11-13 in the run table, "What is left" item 1 rewritten around the cause and the fix, and the dated notes of both catches, the awake test and the fix's runs. - sleep-series.sh: TRACE=2 (the events without the function tracer), PMPRINT=0 (no callback names on the console) and the device PM callback events, the form that caught it. - bringup/hsi0-usi: the two test modules, the clock hold and the stress test, with a README and a header that says the stress test hangs the SoC by design. - AGENTS.md: the 9a sleep entry's open no-wake brought up to the cause and derekross#247.
|
Thank you for this. It's the follow-up I hoped for: the two catches with what each console shows, the awake test that settles the cause, the panic that doesn't reboot and pstore's rejected header, and both modules in the tree with the warning where it belongs. I checked the counts against your reports on linux-omarchy-mobile#247 (1,209 woken and four not, 300 of 300 with #247 alone, the touch and GNSS checks), and they agree. Every row says who was there, the two overnight runs before their attended run included. I'd rather have it written down that plainly than not. The AGENTS change is confined to the 9a's open no-wake, and none of these series had the modem up. CI is green, and it merges cleanly with today's main and with your other open PRs. Two small things before I merge:
Not for this PR: item 3 of "What is left", sleep with the modem up, still describes 11-09. A pointer to the new section in #758 would bring it up to date when that one lands. |
…but a 9a and needs usi=; status words that hold after derekross#247 The stress module now refuses to load unless the machine is google,zumapro-tegu, and usi= has no default, so a bare insmod never switches the touchscreen's USI and no other board runs it by luck of a node path. The status line and item 1's next step no longer depend on derekross#247 being under review.
|
Both are in (3ba8c01):
On item 3: I'll point it at #758's section once that one lands. |
# Conflicts: # AGENTS.md
The previous commit kept the merge's conflict markers in the 9a sleep entry by mistake. This keeps main's derekross#809 note at the head of the entry and this branch's sentence on the no-wake's cause and derekross#247.
|
Thank you, both are in as I asked. The stress module now refuses anything but a 9a, and |
The follow-up #659 left for later, with what you asked for there and on linux-omarchy-mobile#247.
SLEEP-PLAN. The status line and "What is left" item 1 now carry the third cause and its fix. The no-wake on the published kernel came back on 2026-11-12 once the heavy debug switches were off (
PMDEBUG=1 TRACE=2 PMPRINT=0), at the 70th and the 41st sleep of two boots. Both consoles stop on CPU 7 at the start of the noirq resume. The second had the device callback events on, and it names the callback: the resume of the HSI0 USI at 111d00c0. A bring-up module then showed the cause awake (11-13): a register access to one of those USIs in the first microseconds after its gates are opened never returns. With #247, which keeps the six gates on, 300 of 300 woke on the published tip, and touch and GNSS work on that build. The five runs are in the run table, each with the rule it ran under: two of them ran overnight before their attended run, as I wrote on #247. The dated notes of both catches, the awake test and the fix's runs are at the end.The panic that didn't reboot, and the watchdog's blind spot. The stall panic writes the console and the trace, but then its own restart hangs ("Rebooting in 10 seconds.." is the last line), so the palmtop still waits for the buttons. The reset's bit errors then made pstore reject the console's header both times, and the text was read raw from the ramoops region on the next boot. Awake, the cluster watchdog ends a hang while systemd feeds it; in a sleep nothing does. Both are in item 1 and the notes.
sleep-series.shin the form that made the catches:TRACE=2(the events without the function tracer),PMPRINT=0(no callback names on the console) and the device PM callback events.device/tegu/bringup/hsi0-usi/: the two test modules with a README. The clock hold changes nothing on a kernel with #247. The stress test hangs the SoC by design on a kernel without #247, and its header and the README say so, with the conditions it runs under. Neither runs before an attended run with the owner's yes.AGENTS.md: the 9a sleep entry's open no-wake brought up to the cause and #247, with the switches that catch it.
tools/pr-checkandtests/allpass.