Attention is currently required from: Hung-Te Lin, Paul Menzel, Yu-Ping Wu, Felix Held. Rex-BC Chen has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/57270 )
Change subject: soc/mediatek: preserve WDT reset reason for debugging ......................................................................
Patch Set 3: Code-Review+1
(4 comments)
Commit Message:
https://review.coreboot.org/c/coreboot/+/57270/comment/0e18a9ff_f05c606c PS2, Line 7: reserve
preserve
Done
https://review.coreboot.org/c/coreboot/+/57270/comment/e1dfbd60_a62eb3ab PS2, Line 10: reserve
Yeah "preserve" is the correct term here.
Done
https://review.coreboot.org/c/coreboot/+/57270/comment/7bca084a_68592704 PS2, Line 11:
Why is `mark_watchdog_tombstone()` removed?
We think it does not impact for our modification, so we recover is back.
File src/soc/mediatek/common/wdt.c:
https://review.coreboot.org/c/coreboot/+/57270/comment/3f4b7609_dff90f14 PS2, Line 22: write32(&mtk_wdt->wdt_swrst, MTK_WDT_SWRST_KEY);
i'd replace this line with a do_board_reset() call which makes it much clearer what this is doing
original wdt reset flow: when wdt reset occurs: 1. wdt hw send signal to info ec, and do fully reset from ec. (this setting is in kernel driver.)
new wdt reset flow in this patch: when wdt reset occcurs: 1. wdt hw don't send signal to info ec in kernel. 2. after resetting, we will check register status and record the status and wdt hw will send signal to info ec to do fully reset.
So I think do_board_reset() is not part of this flow.
-
Rex-BC Chen (Code Review)