Attention is currently required from: Nico Huber, Branden Waldner, Idwer Vollering, Jacob Garber, Mike Banon, Werner Zeh, Iru Cai, Elyes Haouas, Harshit Sharma.
Patch set 2:Code-Review +1
1 comment:
Patchset:
Elyes, thank you for trying, unfortunately GCC 11.3 doesn't work too. Possible outcomes when attempting to use a coreboot+SeaBIOS on AMD Lenovo G505S built with GCC 11.3 :
1) black screen, doesn't show anything - even a SeaBIOS text interface;
2) shows a SeaBIOS interface which hangs on "Press ESC" (pressing ESC doesn't do anything);
3) shows a SeaBIOS interface which works; if I didn't choose anything, then runs a default entry (a KolibriOS ramdisk, which launched successfully the times I tried); if I chose a HDD, see "4)" and "5)"
4) loads GRUB from HDD - it gets a password from keyboard but hangs while decrypting a /boot;
5) successfully decrypts /boot and shows a GRUB menu, but later hangs on "Loading initial ramdisk" ! This booting step precedes a step of entering a key and decrypting a "/" partition of a Full Disk Encryption instal of Artix Linux (user-friendly Arch without SystemD) which I did by this manual - https://wiki.artixlinux.org/Main/InstallationWithFullDiskEncryptionMeanwhile, an equal coreboot+SeaBIOS build with GCC 8.3.0 and the same contents of HDD - works perfectly all the time. I haven't yet tried a mixed build of "coreboot by new GCC, SeaBIOS by old GCC"
It looks like GCC11 aligns packed structs differently which can cause problems when writing to hardware registers. As #pragma pack is used for all the AGESA code I can see why this goes south... Maybe you can see how removing '#pragma pack(1)' from Porting.h fares?
To view, visit change 59636. To unsubscribe, or for help writing mail filters, visit settings.