Skip to content

arm64: dts: exynos: google: gs101-raven: add the ramoops region - #111

Merged
derekross merged 1 commit into
derekross:omamobilefrom
robotosaith:omamobile-gs101-ramoops-pr
Oct 28, 2026
Merged

derekross merged 1 commit into
derekross:omamobilefrom
robotosaith:omamobile-gs101-ramoops-pr

Conversation

@robotosaith

Copy link
Copy Markdown
Contributor

Raven's kernel has CONFIG_PSTORE_RAM, but its device tree has no ramoops region, so a panic or a watchdog reset leaves nothing in /sys/fs/pstore, and an unexplained reset on raven can't be traced.

This describes the region Google's gs101 kernel uses (gs101-rmem.dtsi), with its console and pmsg sizes, as gs201-cheetah does. On raven, 0xfd3ff000–0xfd7fefff is System RAM between the bootloader's debug-kinfo region and its log region, and no reserved-memory node in the bootloader's tree overlaps it.

Tested on raven (7.3.0-rc6): ramoops binds at 0xfd3ff000 with 2 MiB console and 2 MiB pmsg. A pmsg record survived a PSCI warm reset and was gone after the default cold one. The console zone comes back holding the bootloader's own log, including the reset status (RST_STAT) and reboot reason, so the cause of the last reset is recorded either way. The phone has run it since 2026-10-27. tools/kernel-pr-check passes with --build; the 2 checkpatch warnings are the Co-Authored-By: trailer, as in #8.

🤖 Generated with Claude Code

Raven's kernel has CONFIG_PSTORE_RAM, but its device tree has no
ramoops region, so a panic or a watchdog reset leaves nothing in
/sys/fs/pstore: an unexplained reset on raven could not be traced.

Describe the region Google's gs101 kernel uses (gs101-rmem.dtsi), with
its console and pmsg sizes, as gs201-cheetah does. On raven,
0xfd3ff000-0xfd7fefff is System RAM between the bootloader's
debug-kinfo region below and its log region at 0xfd800000 above, and
no reserved-memory node in the bootloader's tree overlaps it.

Tested on raven (as an overlay on 7.3.0-rc6, gb3d6253b0d61): ramoops
binds at 0xfd3ff000 with 2 MiB console and 2 MiB pmsg. A pmsg record
survived a PSCI warm reset (reboot mode warm, SYSTEM_RESET2) and was
gone after the default cold one. The console zone comes back holding
the bootloader's own log, which it writes there at every boot,
including the reset status (RST_STAT) and reboot reason it read, so the
cause of the last reset is recorded either way.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: saiththerobo <the.tom.sharp@gmail.com>
@derekross

Copy link
Copy Markdown
Owner

Thanks. The region is exactly Google's ramoops_mem@fd3ff000 from gs101-rmem.dtsi: address, size, console and pmsg sizes. It's 12-byte reg under gs101's two address cells and one size cell, and it sits between debug_kinfo_reserved and bldr_log_reserved, which gs101.dtsi already reserves. It builds and validates, and the other boards' DTBs are unchanged. Good to see the warm-reset survival tested too.

Merging. Your other raven DT PRs from today all add at the end of gs101-raven.dts, so each will need a quick rebase as the one before it lands. I'll go #111, #112, #113, #114, then #116.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants