mirror of
https://codeberg.org/canoeboot/cbmk.git
synced 2025-01-10 17:19:57 +00:00
ce9030b74f
This commit realises an intellectual dream: what if I made a Canoeboot release, but using upstream revisions from GNU Boot 0.1 RC3? Canoeboot uses much newer revisions, but I've recently been sending patches to the GNU Boot project, making their 0.1 series build on modern distros, and I also added support for U-Boot building (gru_bob and gru_kevin boards added), and Dell Latitude E6400. This commit *reverts* Canoeboot back to the older upstream revisions used in GNU Boot, based on the Libreboot 20220710 revisions, but with their changes and with my fixes. I've also included my improvements to the grub.cfg file, such as EFI System Partition support, which was also sent upstream to GNU Boot for review. This commit *removes* argon2 support from GRUB, because GNU Boot 0.1 doesn't have it! A few other things from Canoeboot have been retained, such as the ability to build serprog firmware. The memtest86plus revision was even downgraded back to the 5.x one used by GNU Boot (Canoeboot uses 6.2). This means that memtest86plus is 32-bit again, not 64-bit, and re-enabled on X60/T60, but it is only provided in configurations, on each board, where text mode startup is used (on Canoeboot, only 64-bit builds are provided, so unavailable for X60/T60, but memtest 6.2 works on corebootfb and txtmode startup). TL;DR this commit/branch/release of Canoeboot, dubbed Canoeboot v0.1, is more or less precisely in sync with GNU Boot 0.1 RC3, but with my build fixes, and the additional boards (gru_bob, gru_kevin, dell e6400). I did this so that I could have a more reliable benchmark, comparing the build system performance of GNU Boot 0.1 RC and Canoeboot. The idea is that if exact revisions are used across both projects, that are the same, then the test is more likely to be reliable, in showing how fast or slow each build system design is. Certain design changes from Canoeboot have been retained here, such as use of pre-generated ICH9M IFDs (GNU Boot re-creates them using my ich9gen, but Canoeboot now includes pre-made ones generated from ich9gen), and declaring PIKE2008 fake roms in coreboot menuconfig instead, defining them as paths to /dev/null, so as to still insert empty ROMs (while not calling cbfstool from cbmk, instead relying on the coreboot build system). Also merged these GRUB fixes from Libreboot: d44c9551c5e7456c2caa4a2815d33ff978dc55ef build/roms: regression fix: uninitialised variable df007d22ec801679b5e8f43ee861b78515518ce2 build/roms: err if -k layout doesn't exist 37817e6bcb7c7272d7c70c3afe89a5b3b2604824 GRUB: insert only 1 keymap per board, in cbfs (instead of memdisk, and use compressed .gkb files) This essentially syncs with lbmk up to commit: 3e7e0c7d4881a187f82404beb34a2cd014a409f8 Signed-off-by: Leah Rowe <info@minifree.org>
22 lines
683 B
Diff
22 lines
683 B
Diff
From 1ee893f9fc5251968500695824ab9fd39461d318 Mon Sep 17 00:00:00 2001
|
|
From: Leah Rowe <leah@libreboot.org>
|
|
Date: Fri, 14 May 2021 13:21:39 +0100
|
|
Subject: [PATCH 08/11] acer/g43t-am3: set VRAM to 352MiB by default
|
|
|
|
---
|
|
src/mainboard/acer/g43t-am3/cmos.default | 2 +-
|
|
1 file changed, 1 insertion(+), 1 deletion(-)
|
|
|
|
diff --git a/src/mainboard/acer/g43t-am3/cmos.default b/src/mainboard/acer/g43t-am3/cmos.default
|
|
index 706f5dd551..98899e8bf5 100644
|
|
--- a/src/mainboard/acer/g43t-am3/cmos.default
|
|
+++ b/src/mainboard/acer/g43t-am3/cmos.default
|
|
@@ -3,4 +3,4 @@ debug_level=Debug
|
|
power_on_after_fail=Disable
|
|
nmi=Enable
|
|
sata_mode=AHCI
|
|
-gfx_uma_size=64M
|
|
+gfx_uma_size=352M
|
|
--
|
|
2.25.1
|
|
|