Hi, I am investigating two autoport issues:
https://ticket.coreboot.org/issues/652 https://ticket.coreboot.org/issues/586 For #652, I reproduced the failure on current main (3f89aa30bb) using the attached MacBookPro11,4 logs. The missing 00:1f.2 SATA controller causes an unchecked ConfigDump[0x92] access. After bypassing it, the missing 00:1d.0 EHCI controller causes another unchecked ConfigDump[0x64] access. Temporarily skipping both assumptions allows autoport to complete with the existing 00:14.0 xHCI controller. Would the preferred behavior be: 1. Treat SATA and EHCI controllers as optional and continue with explicit warnings/FIXMEs when configuration cannot be inferred; or 2. Stop cleanly with an actionable error instead of generating incomplete USB2 configuration? For #586, autoport only consumes the FACP and DSDT tables. Would using: acpidump -n FACP -n DSDT as the safe default be preferred over dumping all tables or using `acpidump -b` and deleting SLIC/MSDM afterward? Should there also be an explicit unredacted opt-in mode? I would appreciate feedback on the expected behavior before preparing implementation patches and regression tests. Best regards, Zax
_______________________________________________ coreboot mailing list -- [email protected] To unsubscribe send an email to [email protected]
