lbmk/script/build/boot/roms_helper

547 lines
16 KiB
Plaintext
Raw Normal View History

#!/usr/bin/env sh
# helper script: create ROM images for a given mainboard
#
# Copyright (C) 2020,2021,2023 Leah Rowe <info@minifree.org>
# Copyright (C) 2021,2022 Ferass El Hafidi
# <vitali64pmemail@protonmail.com>
# Copyright (C) 2022 Caleb La Grange <thonkpeasant@protonmail.com>
# Copyright (C) 2022 Alper Nebi Yasak <alpernebiyasak@gmail.com>
# Copyright (C) 2023 Riku Viitanen <riku.viitanen@protonmail.com>
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
[ "x${DEBUG+set}" = 'xset' ] && set -v
set -u -e
. "include/err.sh"
consolidate u-boot/seabios/coreboot build scripts See file: resources/scripts/build/defconfig/for It is based on: resources/scripts/build/payload/u-boot The u-boot payload script has been deleted, as has the seabios payload script; the build/boot/roms logic has been heavily simplified too, by removing the logic for building of elf files based on defconfig. SeaBIOS, U-Boot and coreboot all use defconfig-type infrastructure for their build systems, and they are fundamentally the *same* in how to compile each codebase, at least in an lbmk context, regardless of actual (and very huge) differences in these codebases. Several hundred sources-lines of code have been eliminated by this change, drastically simplifying everything; U-Boot payload compiling also now errors out when a single build fails, instead of continuing. Also: build/boot/roms no longer re-compiles a coreboot target that was already compiled, which is the same behaviour observed for payloads. (this means you must now manually delete a target, when you wish to re-build it; the build/boot/roms logic now more or less just runs cbfstool; blobutil is handled from build/defconfig/for) ALSO: Since crossgcc is now handled by build/defconfig/for, not build/boot/roms, standalone compiling of u-boot is now possible. This has been tested. You compile it like so: ./build defconfig for u-boot or specific trees, e.g. ./build defconfig for u-boot default One other consequence of this patch is that re-building the same ROM image is now much faster, because the same builds are re-used unless deleted. This could be useful when testing grub.cfg changes, for example, if that's all you change. With things like ccache used (not yet used robustly in lbmk), this could speed things up more, depending on the codebase. This patch demonstrates the raw power of lbmk; it is a very simple and highly efficient build system, and now much more so! Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-17 22:27:30 +00:00
read projectname < projectname
cfgsdir="config/coreboot"
blobs_required=""
consolidate u-boot/seabios/coreboot build scripts See file: resources/scripts/build/defconfig/for It is based on: resources/scripts/build/payload/u-boot The u-boot payload script has been deleted, as has the seabios payload script; the build/boot/roms logic has been heavily simplified too, by removing the logic for building of elf files based on defconfig. SeaBIOS, U-Boot and coreboot all use defconfig-type infrastructure for their build systems, and they are fundamentally the *same* in how to compile each codebase, at least in an lbmk context, regardless of actual (and very huge) differences in these codebases. Several hundred sources-lines of code have been eliminated by this change, drastically simplifying everything; U-Boot payload compiling also now errors out when a single build fails, instead of continuing. Also: build/boot/roms no longer re-compiles a coreboot target that was already compiled, which is the same behaviour observed for payloads. (this means you must now manually delete a target, when you wish to re-build it; the build/boot/roms logic now more or less just runs cbfstool; blobutil is handled from build/defconfig/for) ALSO: Since crossgcc is now handled by build/defconfig/for, not build/boot/roms, standalone compiling of u-boot is now possible. This has been tested. You compile it like so: ./build defconfig for u-boot or specific trees, e.g. ./build defconfig for u-boot default One other consequence of this patch is that re-building the same ROM image is now much faster, because the same builds are re-used unless deleted. This could be useful when testing grub.cfg changes, for example, if that's all you change. With things like ccache used (not yet used robustly in lbmk), this could speed things up more, depending on the codebase. This patch demonstrates the raw power of lbmk; it is a very simple and highly efficient build system, and now much more so! Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-17 22:27:30 +00:00
board=""
ubdir=""
kmapdir="config/grub/keymap"
displaymodes=""
payloads=""
keyboard_layouts=""
2021-11-27 19:06:32 +00:00
grub_timeout=""
grub_scan_disk="undefined"
tree="undefined"
arch="undefined"
# Disable all payloads by default.
# target.cfg files have to specifically enable [a] payload(s)
payload_grub="n"
payload_grub_withseabios="n" # seabios chainloaded from grub
payload_seabios="n"
payload_seabios_withgrub="n" # i386-coreboot grub from SeaBIOS boot menu
payload_memtest="n"
payload_uboot="n"
uboot_config="undefined"
romdir=""
cbdir=""
cbfstool=""
corebootrom=""
consolidate u-boot/seabios/coreboot build scripts See file: resources/scripts/build/defconfig/for It is based on: resources/scripts/build/payload/u-boot The u-boot payload script has been deleted, as has the seabios payload script; the build/boot/roms logic has been heavily simplified too, by removing the logic for building of elf files based on defconfig. SeaBIOS, U-Boot and coreboot all use defconfig-type infrastructure for their build systems, and they are fundamentally the *same* in how to compile each codebase, at least in an lbmk context, regardless of actual (and very huge) differences in these codebases. Several hundred sources-lines of code have been eliminated by this change, drastically simplifying everything; U-Boot payload compiling also now errors out when a single build fails, instead of continuing. Also: build/boot/roms no longer re-compiles a coreboot target that was already compiled, which is the same behaviour observed for payloads. (this means you must now manually delete a target, when you wish to re-build it; the build/boot/roms logic now more or less just runs cbfstool; blobutil is handled from build/defconfig/for) ALSO: Since crossgcc is now handled by build/defconfig/for, not build/boot/roms, standalone compiling of u-boot is now possible. This has been tested. You compile it like so: ./build defconfig for u-boot or specific trees, e.g. ./build defconfig for u-boot default One other consequence of this patch is that re-building the same ROM image is now much faster, because the same builds are re-used unless deleted. This could be useful when testing grub.cfg changes, for example, if that's all you change. With things like ccache used (not yet used robustly in lbmk), this could speed things up more, depending on the codebase. This patch demonstrates the raw power of lbmk; it is a very simple and highly efficient build system, and now much more so! Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-17 22:27:30 +00:00
seavgabiosrom="elf/seabios/default/libgfxinit/vgabios.bin"
grub_background="background1280x800.png"
initmode=""
displaymode=""
cbcfg=""
targetdir=""
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
grubelf="elf/grub/grub.elf"
main()
{
while [ $# -gt 0 ]; do
case ${1} in
-d)
displaymodes="${displaymodes}${2}"
shift ;;
-p)
payloads="${payloads}${2}"
shift ;;
-k)
keyboard_layouts="${keyboard_layouts}${2}"
shift ;;
*)
board=${1} ;;
esac
shift
done
printf "\n\nboard %s, kb %s, displaymode %s, payloads %s\n" \
"${board}" "${keyboard_layouts}" "${displaymodes}" "${payloads}"
configure_target
build_dependencies
build_target
}
configure_target()
{
targetdir="${cfgsdir}/${board}"
[ -d "${targetdir}" ] || \
err "Target not defined: ${board}"
[ -f "${targetdir}/target.cfg" ] || \
err "Missing target.cfg for target: ${board}"
# Override the above defaults using target.cfg
. "${targetdir}/target.cfg"
[ "${grub_scan_disk}" = "undefined" ] && \
grub_scan_disk="both"
[ "${grub_scan_disk}" != "both" ] && \
[ "${grub_scan_disk}" != "ata" ] && \
[ "${grub_scan_disk}" != "ahci" ] && \
grub_scan_disk="both"
[ "${tree}" = "undefined" ] && \
err "Target '${board}' defines no tree. Skipping build."
[ "${arch}" = "undefined" ] && \
err "Target '${board}' defines no arch. Skipping build."
[ "${payload_memtest}" != "y" ] && \
payload_memtest="n"
[ "${payload_grub_withseabios}" = "y" ] && \
payload_grub="y"
if [ "${payload_grub_withseabios}" = "y" ]; then
payload_seabios="y"
payload_seabios_withgrub="y"
fi
[ "${payload_seabios_withgrub}" = "y" ] && \
payload_seabios="y"
# The reverse logic must not be applied. If SeaBIOS-with-GRUB works,
# that doesn't mean GRUB-withSeaBIOS will. For example, the board
# might have a graphics card whose vga rom coreboot doesn't execute
if [ "${payload_grub}" != "y" ] && \
[ "${payload_seabios}" != "y" ] && \
[ "${payload_uboot}" != "y" ]; then
for configfile in "${targetdir}/config/"*; do
[ -e "${configfile}" ] || continue
err "target '${board}' defines no payload"
done
fi
[ "${payload_uboot}" != "n" ] && [ "${payload_uboot}" != "y" ] && \
payload_uboot="n"
[ "${payload_uboot}" = "y" ] && [ "${uboot_config}" = "undefined" ] && \
uboot_config="default"
[ "${blobs_required}" != "n" ] && [ "${blobs_required}" != "y" ] && \
blobs_required="y"
# Override all payload directives with cmdline args
if [ ! -z ${payloads} ]; then
printf "setting payloads %s\n" "$payloads"
payload_grub="n"
payload_grub_withseabios="n" # seabios chainloaded from grub
payload_seabios="n"
payload_seabios_withgrub="n" # i386-coreboot grub from SeaBIOS
payload_uboot="n"
payload_memtest="n"
for payload in ${payloads} ; do
eval "payload_${payload}=y"
done
fi
}
build_dependencies()
{
romdir="bin/${board}"
cbdir="coreboot/${board}"
[ "${board}" = "${tree}" ] || \
cbdir="coreboot/${tree}"
cbfstool="cbutils/${tree}/cbfstool"
corebootrom="${cbdir}/build/coreboot.rom"
./build coreboot utils ${tree} || err "cannot build cbutils/${tree}"
build_dependency_seabios
memtest_bin="memtest86plus/build${arch#*_}/memtest.bin"
[ "${payload_memtest}" != "y" ] || [ -f "${memtest_bin}" ] || \
./handle make file -b ${memtest_bin%/*} || \
err "cannot build memtest86+"
rm -f "${romdir}"/* || err "cannot: rm -f \"${romdir}\"/*"
build_dependency_grub
build_dependency_uboot
}
build_dependency_seabios()
{
[ "${payload_seabios}" = "y" ] || return 0
if [ ! -f "${seavgabiosrom}" ] \
|| [ ! -f elf/seabios/default/libgfxinit/bios.bin.elf ] \
|| [ ! -f elf/seabios/default/vgarom/bios.bin.elf ] \
|| [ ! -f elf/seabios/default/normal/bios.bin.elf ]; then
./handle make config -b seabios || \
much, much stricter, more verbose error handling lbmk is much more likely to crash now, in error conditions, which is a boon for further auditing. also: in "fetch", remove the downloaded program if fail() was called. this would also be done for gnulib, when downloading grub, but done in such a way that gnulib goes first. where calls to err write "ERROR" in the string, they no longer say "ERROR" because the "err" function itself now does that automatically. also: listmodes/listoptions (in "lbmk") now reports an error if no scripts and/or directories are found. also: where a warning is given, but not an error, i've gone through in some places and redirected the output to stderr, not stdout as part of error checks: running anything as root, except for the "./build dependencies *" commands, is no longer permitted and lbmk will throw an error mrc downloads: debugfs output no longer redirected to /dev/null, and stderr no longer redirected to stdout. everything is verbose. certain non-error states are also more verbose. for example, patch_rom in blobs/inject will now state when injection succeeds certain actual errors(bugs) were fixed: for example, build/release/roms now correctly prepares the blobs hash files for a given target, containing only the files and checksums in the list. Previously, a printf message was included. Now, with this new code: blobutil/inject rightly verifies hashes. doing all of this in one giant patch is cleaner than 100 patches changing each file. even this is yet part of a much larger audit going on in the Libreboot project. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-24 19:19:41 +00:00
err "cannot build seabios"
fi
}
build_dependency_grub()
{
[ "${payload_grub}" != "y" ] && \
[ "${payload_seabios_withgrub}" != "y" ] && return 0
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
rebuild_grub="n"
[ ! -f "${grubelf}" ] && rebuild_grub="y"
for keymapfile in "${kmapdir}"/*.gkb; do
[ "${rebuild_grub}" = "y" ] || break
[ -f "${keymapfile}" ] || continue
keymap="${keymapfile##*/}"
keymap="${keymap%.gkb}"
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
[ ! -f "elf/grub/keymap_${keymap}.cfg" ] && \
rebuild_grub="y" && break
done
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
if [ "${rebuild_grub}" = "y" ]; then
./build grub payload || \
err "build_dependency_grub: cannot build grub payload"
fi
}
build_dependency_uboot()
{
[ "${payload_uboot}" = "y" ] || return 0
./handle make config -b u-boot ${board} || \
much, much stricter, more verbose error handling lbmk is much more likely to crash now, in error conditions, which is a boon for further auditing. also: in "fetch", remove the downloaded program if fail() was called. this would also be done for gnulib, when downloading grub, but done in such a way that gnulib goes first. where calls to err write "ERROR" in the string, they no longer say "ERROR" because the "err" function itself now does that automatically. also: listmodes/listoptions (in "lbmk") now reports an error if no scripts and/or directories are found. also: where a warning is given, but not an error, i've gone through in some places and redirected the output to stderr, not stdout as part of error checks: running anything as root, except for the "./build dependencies *" commands, is no longer permitted and lbmk will throw an error mrc downloads: debugfs output no longer redirected to /dev/null, and stderr no longer redirected to stdout. everything is verbose. certain non-error states are also more verbose. for example, patch_rom in blobs/inject will now state when injection succeeds certain actual errors(bugs) were fixed: for example, build/release/roms now correctly prepares the blobs hash files for a given target, containing only the files and checksums in the list. Previously, a printf message was included. Now, with this new code: blobutil/inject rightly verifies hashes. doing all of this in one giant patch is cleaner than 100 patches changing each file. even this is yet part of a much larger audit going on in the Libreboot project. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-24 19:19:41 +00:00
err "cannot build u-boot target: ${board}"
consolidate u-boot/seabios/coreboot build scripts See file: resources/scripts/build/defconfig/for It is based on: resources/scripts/build/payload/u-boot The u-boot payload script has been deleted, as has the seabios payload script; the build/boot/roms logic has been heavily simplified too, by removing the logic for building of elf files based on defconfig. SeaBIOS, U-Boot and coreboot all use defconfig-type infrastructure for their build systems, and they are fundamentally the *same* in how to compile each codebase, at least in an lbmk context, regardless of actual (and very huge) differences in these codebases. Several hundred sources-lines of code have been eliminated by this change, drastically simplifying everything; U-Boot payload compiling also now errors out when a single build fails, instead of continuing. Also: build/boot/roms no longer re-compiles a coreboot target that was already compiled, which is the same behaviour observed for payloads. (this means you must now manually delete a target, when you wish to re-build it; the build/boot/roms logic now more or less just runs cbfstool; blobutil is handled from build/defconfig/for) ALSO: Since crossgcc is now handled by build/defconfig/for, not build/boot/roms, standalone compiling of u-boot is now possible. This has been tested. You compile it like so: ./build defconfig for u-boot or specific trees, e.g. ./build defconfig for u-boot default One other consequence of this patch is that re-building the same ROM image is now much faster, because the same builds are re-used unless deleted. This could be useful when testing grub.cfg changes, for example, if that's all you change. With things like ccache used (not yet used robustly in lbmk), this could speed things up more, depending on the codebase. This patch demonstrates the raw power of lbmk; it is a very simple and highly efficient build system, and now much more so! Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-17 22:27:30 +00:00
ubdir="elf/u-boot/${board}/${uboot_config}"
ubootelf="${ubdir}/u-boot.elf"
[ ! -f "${ubootelf}" ] && [ -f "${ubdir}/u-boot.bin" ] && \
ubootelf="${ubdir}/u-boot.bin"
if [ ! -f "${ubootelf}" ]; then
consolidate u-boot/seabios/coreboot build scripts See file: resources/scripts/build/defconfig/for It is based on: resources/scripts/build/payload/u-boot The u-boot payload script has been deleted, as has the seabios payload script; the build/boot/roms logic has been heavily simplified too, by removing the logic for building of elf files based on defconfig. SeaBIOS, U-Boot and coreboot all use defconfig-type infrastructure for their build systems, and they are fundamentally the *same* in how to compile each codebase, at least in an lbmk context, regardless of actual (and very huge) differences in these codebases. Several hundred sources-lines of code have been eliminated by this change, drastically simplifying everything; U-Boot payload compiling also now errors out when a single build fails, instead of continuing. Also: build/boot/roms no longer re-compiles a coreboot target that was already compiled, which is the same behaviour observed for payloads. (this means you must now manually delete a target, when you wish to re-build it; the build/boot/roms logic now more or less just runs cbfstool; blobutil is handled from build/defconfig/for) ALSO: Since crossgcc is now handled by build/defconfig/for, not build/boot/roms, standalone compiling of u-boot is now possible. This has been tested. You compile it like so: ./build defconfig for u-boot or specific trees, e.g. ./build defconfig for u-boot default One other consequence of this patch is that re-building the same ROM image is now much faster, because the same builds are re-used unless deleted. This could be useful when testing grub.cfg changes, for example, if that's all you change. With things like ccache used (not yet used robustly in lbmk), this could speed things up more, depending on the codebase. This patch demonstrates the raw power of lbmk; it is a very simple and highly efficient build system, and now much more so! Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-17 22:27:30 +00:00
err "Could not find u-boot build for board, ${board}"
fi
}
build_target()
{
for x in "normal" "vgarom" "libgfxinit"; do
initmode="${x}"
hmode="vesafb"
[ "${initmode}" = "vgarom" ] || hmode="corebootfb"
modes="${hmode} txtmode"
[ -z ${displaymodes} ] || modes="${displaymodes}"
for y in ${modes}; do
displaymode="${y}"
[ "${initmode}" = "normal" ] && \
[ "$displaymode" != "txtmode" ] && continue
cbcfg="${targetdir}/config/${initmode}_${displaymode}"
[ "${initmode}" = "normal" ] && cbcfg="${cbcfg%_*}"
build_roms "${cbcfg}" "${displaymode}" "${initmode}"
done
done
}
# Main ROM building function. This calls all other functions below
build_roms()
{
cbcfg="${1}"
displaymode="${2}"
initmode="${3}"
[ ! -f "${cbcfg}" ] && \
printf "'%s' does not exist. Skipping build for %s %s %s\n" \
"${cbcfg}" "${board}" "${displaymode}" "${initmode}" \
1>&2 && return 0
./handle make config -b coreboot ${board} || \
err "build_roms: cannot build coreboot for target: ${board}"
_corebootrom="elf/coreboot/${board}/${initmode}_${displaymode}"
[ "${initmode}" = "normal" ] && \
_corebootrom="${_corebootrom%_${displaymode}}"
_corebootrom="${_corebootrom}/coreboot.rom"
corebootrom="$(mktemp -t coreboot_rom.XXXXXXXXXX)"
cp "${_corebootrom}" "${corebootrom}" || \
err "build_roms: cannot copy rom"
coreboot: re-add asus kgpe-d16/kcma-d8/kfsn4-dre Libreboot 20220710 was the last release to support these boards. I plan to eventually port code differences between D8/D16 to Dasharo, for KCMA-D8 support in Dasharo, to then use in Libreboot for both KCMA-D8 and KGPE-D16, but I have no plans to update the KFSN4-DRE code, at least for now. Libreboot 20220710 used coreboot 4.11, whereas this patch makes use of coreboot 4.11_branch; the crossgcc toolchains no longer compile on modern distros, so I spent time patching those (tested in Debian Sid, will also work on Arch Linux and so on). The acpica downloads now fail, in 4.11_branch, because Intel made some changes upstream for these tarball downloads. Newer coreboot works around this by grabbing tarballs from github, itself a non-ideal solution, but I digress; this patch changes coreboot crossgcc (in 4.11_branch) to download the acpica tarball from libreboot rsync, where I've added it. This patch also re-introduces the PIKE2008 fix, where empty option ROMs for these are inserted into CBFS. This prevents SeaBIOS from loading the real option ROMs, which would cause SeaBIOS to hang. This means that SAS drives are not supported in SeaBIOS, for these boards in Libreboot. I previously said, in the Censored Libreboot c20230710 announcement, that I would *only* merge D8/D16 when I've added Dasharo support to Libreboot, and use that, but the work to make coreboot 4.11_branch compile is something I'm quite proud of and I see no reason to exclude from lbmk master branch. Honestly, there's not much different than 4.11, code-wise. I *probably* won't use 4.11_branch for the next Libreboot release, on D8/D16. By then, I might have Dasharo integrated in lbmk instead. We shall see. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-07-16 02:18:58 +00:00
[ "${payload_memtest}" != "y" ] || \
"${cbfstool}" "${corebootrom}" add-payload \
-f "${memtest_bin}" -n img/memtest -c lzma || \
err "build_roms: cannot add img/memtest to coreboot rom"
[ "${payload_seabios}" = "y" ] && \
build_seabios_roms
[ "${payload_grub}" != "y" ] || \
build_grub_roms "${corebootrom}" "${initmode}" \
"${displaymode}" "grub" || \
err "build_roms: build_grub_roms failed"
if [ "${payload_uboot}" = "y" ]; then
build_uboot_roms
fi
}
build_seabios_roms()
{
if [ "${payload_seabios_withgrub}" = "y" ]; then
tmprom=$(mktemp -t coreboot_rom.XXXXXXXXXX)
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
cp "${corebootrom}" "${tmprom}" || \
err "build_seabios_roms: cannot copy to tmprom"
build_grub_roms "${tmprom}" "${initmode}" \
"${displaymode}" "seabios_withgrub" || \
err "build_roms: cannot build grub roms, seabios w/grub"
rm -f "${tmprom}" || err "build_roms: can't remove tmprom"
else
x=${corebootrom}
y=${initmode}
t=$(mkSeabiosRom "$x" "fallback/payload" "$y") || \
err "build_seabios_roms: cannot build tmprom"
if [ "${initmode}" = "normal" ]; then
newrompath="${romdir}/seabios_${board}_"
newrompath="${newrompath}${initmode}.rom"
else
newrompath="${romdir}/seabios_${board}_"
newrompath="${newrompath}${initmode}_"
newrompath="${newrompath}${displaymode}.rom"
fi
moverom "${t}" "${newrompath}" || \
err "build_roms: cannot copy rom"
rm -f "${t}" || err "cannot rm ${t}"
fi
}
# Make separate ROM images with GRUB payload, for each supported keymap
build_grub_roms() {
tmprompath="${1}"
initmode="${2}"
displaymode="${3}"
firstpayloadname="${4}" # allow values: grub, seabios, seabios_withgrub
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
grubelf_cbfs="fallback/payload"
x=${tmprompath}
y=${initmode}
if [ "${payload_grub_withseabios}" = "y" ] && \
[ "${firstpayloadname}" = "grub" ]; then
mv "$(mkSeabiosRom "${x}" "seabios.elf" "${y}")" \
"${tmprompath}" || \
err "build_grub_roms: cannot move tmp rom (seabios.elf)"
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
elif [ "${payload_seabios_withgrub}" = "y" ] && \
[ "${firstpayloadname}" != "grub" ]; then
mv "$(mkSeabiosRom "${x}" "fallback/payload" "$y")" \
"${tmprompath}" || \
err "build_grub_roms: cannot move SeaBIOS ROM to tmprom"
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
grubelf_cbfs="img/grub2"
fi
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
# we only need insert grub.elf once, for each coreboot config:
"${cbfstool}" "${tmprompath}" add-payload -f "${grubelf}" \
-n ${grubelf_cbfs} -c lzma || \
err "build_grub_roms: cannot add grub payload to tmprom"
# we only need insert background.png once, for each coreboot config:
if [ "${displaymode}" = "vesafb" ] || \
[ "${displaymode}" = "corebootfb" ]; then
backgroundfile="config/grub/background/${grub_background}"
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
"${cbfstool}" "${tmprompath}" add -f ${backgroundfile} \
-n background.png -t raw || \
err "build_grub_roms: cannot add background.png to tmprom"
fi
tmpcfg=$(mktemp -t coreboot_rom.XXXXXXXXXX) ||
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
err "mkGrubRom: cannot create temporary scan.cfg"
printf "set grub_scan_disk=\"%s\"\n" "${grub_scan_disk}" > \
"${tmpcfg}" || \
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
err "mkGrubRom: cannot insert into temporary scan.cfg"
[ "${grub_scan_disk}" = "both" ] || \
"${cbfstool}" "${tmprompath}" add -f "${tmpcfg}" -n scan.cfg \
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
-t raw || err "mkGrubROM: cannot insert scan.cfg into CBFS"
printf "set timeout=%s\n" "${grub_timeout}" > "${tmpcfg}" || \
err "mkGrubRom: cannot insert into temporary timeout.cfg"
[ -z "${grub_timeout}" ] || \
"${cbfstool}" "${tmprompath}" add -f "${tmpcfg}" -n timeout.cfg \
-t raw || err "mkGrubROM: cannot insert timeout.cfg into CBFS"
rm -f "${tmpcfg}" || err "cannot delete tmpcfg"
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
keymaps=""
if [ -z ${keyboard_layouts} ]; then
for kmapfile in "${kmapdir}"/*; do
keymaps="${keymaps} ${kmapfile}"
done
else
for keymapname in ${keyboard_layouts}; do
keymaps="${keymaps} ${kmapdir}/${keymapname}.gkb"
done
fi
for keymapfile in ${keymaps}; do
printf "keymaps is %s, keymapfile is %s\n" \
"${keymaps}" "${keymapfile}"
[ -f "${keymapfile}" ] || continue
keymap="${keymapfile##*/}"
keymap="${keymap%.gkb}"
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
tmpgrubrom="$(mkGrubRom "${keymap}" "${tmprompath}")"
if [ "${initmode}" = "normal" ]; then
newrompath="${romdir}/${firstpayloadname}_${board}_"
newrompath="${newrompath}${initmode}_${keymap}.rom"
else
newrompath="${romdir}/${firstpayloadname}_${board}_"
newrompath="${newrompath}${initmode}_${displaymode}_"
newrompath="${newrompath}${keymap}.rom"
fi
moverom "${tmpgrubrom}" "${newrompath}" || \
err "build_grub_roms, moverom"
rm -f "${tmpgrubrom}" || err "rm tmpgrubrom, build_grub_roms"
done
}
# make a rom in /tmp/ and then print the path of that ROM
mkGrubRom() {
target_keymap="${1}"
target_cbrom="${2}"
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
keymapcfg="elf/grub/keymap_${target_keymap}.cfg"
tmprom=$(mktemp -t coreboot_rom.XXXXXXXXXX) || \
err "mkGrubRom: cannot create tmprom"
cp "${target_cbrom}" "${tmprom}" || \
err "mkGrubRom: cannot copy to tmprom"
grub: all one grub.elf containing keymaps and cfg new behaviour: * grub.cfg and grubtest.cfg no longer inserted to cbfs * grub.cfg in memdisk instead * grub.cfg in memdisk defers to cbfs/grub.cfg if added (not added by default, anymore) * does not defer to grubtest.cfg even if available * only shows link to grubtest.cfg if available, as a menuentry item keymaps: if /keymap.gkb exists in cbfs, it uses that by default, but by default this isn't added. instead, it looks for a file named keymap.cfg and sources that, which then sets the keymap to one that is located under memdisk. this file is inserted for each rom, per layout. if keymap.gkb and keymap.cfg both absent, grub.cfg in memdisk shall defer to usqwerty as the default keymap grub_scan_disk: grub.cfg looks for cbfs file "scan.cfg" and sources that if found, which will be inserted with the string: set grub_scandisk=setting_goes_here (based on target.cfg, generated by build/boot/roms automatically). If no scan.cfg is found, it defaults to "both" The "background.png" file remains unchanged, and present in CBFS, used by grub.cfg if present (and it is, by default) This change actually *saves* space in CBFS, due to compression, and means that the grub.cfg is now compressed heavily. This is also safer, because now the user overrides grub.cfg by adding it, and they can still add grubtest.cfg for testing first. If they accidentally delete both configs from cbfs, Libreboot will fall back to the one in memdisk which would presumably not be deleted. This also means that lbmk can now more easily be used by other build systems, that just want the GRUB part to re-use in their own project. For example, people who want to build custom coreboot images without using Libreboot's build system. This change also *speeds* up the build process considerably, on the parts where ROM images are copied. It's less than half a second now, whereas previously it took about 30-45 seconds for ROM images to copy, because of grub.elf being re-added in each ROM via cbfstool, where compression is used; I believe the compression part is what caused slowness. Much, much faster, more versatile builds. Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-28 17:10:35 +00:00
"${cbfstool}" "${tmprom}" add -f "${keymapcfg}" -n keymap.cfg -t raw \
|| err "mkGrubRom: cannot add keymap.cfg to tmprom"
printf "%s\n" "${tmprom}"
}
# make a rom in /tmp/ and then print the path of that ROM
mkSeabiosRom() {
target_cbrom="${1}" # rom to insert seabios in. will not be touched
# (a tmpfile will be made instead)
target_seabios_cbfs_path="${2}" # e.g. fallback/payload
target_initmode="${3}" # e.g. libgfxinit
target_seabioself="elf/seabios/default/${target_initmode}/bios.bin.elf"
tmprom=$(mktemp -t coreboot_rom.XXXXXXXXXX)
cp "${target_cbrom}" "${tmprom}" || \
err "mkSeabiosRom: cannot copy to tmprom"
"${cbfstool}" "${tmprom}" add-payload -f "${target_seabioself}" \
-n ${target_seabios_cbfs_path} -c lzma || \
err "mkSeabiosRom: can't add payload, ${target_seabioself}, to rom"
"${cbfstool}" "${tmprom}" add-int -i 3000 -n etc/ps2-keyboard-spinup \
|| err "mkSeabiosRom: cbfs add-int etc/ps2-keyboard-spinup 3000"
if [ "${target_initmode}" = "normal" ] || \
[ "${target_initmode}" = "libgfxinit" ]; then
"${cbfstool}" "${tmprom}" add-int -i 2 \
-n etc/pci-optionrom-exec || \
err "mkSeabiosRom: cbfs add-int etc/pci-optionrom-exec 2"
elif [ "${target_initmode}" = "vgarom" ]; then # coreboot executes it
"${cbfstool}" "${tmprom}" add-int -i 0 \
-n etc/pci-optionrom-exec || \
err "mkSeabiosRom: cbfs add-int etc/pci-optionrom-exec 0"
fi # for undefined modes, don't add this integer. use SeaBIOS defaults
"${cbfstool}" "${tmprom}" add-int -i 0 -n etc/optionroms-checksum || \
err "mkSeabiosRom: cbfs add-int etc/optionroms-checksum 0"
[ "${target_initmode}" != "libgfxinit" ] || \
"${cbfstool}" "${tmprom}" add -f "${seavgabiosrom}" \
-n vgaroms/seavgabios.bin -t raw || \
err "mkSeabiosRom: cbfs add-raw vgaroms/seavgabios.bin"
printf "%s\n" "${tmprom}"
}
build_uboot_roms()
consolidate u-boot/seabios/coreboot build scripts See file: resources/scripts/build/defconfig/for It is based on: resources/scripts/build/payload/u-boot The u-boot payload script has been deleted, as has the seabios payload script; the build/boot/roms logic has been heavily simplified too, by removing the logic for building of elf files based on defconfig. SeaBIOS, U-Boot and coreboot all use defconfig-type infrastructure for their build systems, and they are fundamentally the *same* in how to compile each codebase, at least in an lbmk context, regardless of actual (and very huge) differences in these codebases. Several hundred sources-lines of code have been eliminated by this change, drastically simplifying everything; U-Boot payload compiling also now errors out when a single build fails, instead of continuing. Also: build/boot/roms no longer re-compiles a coreboot target that was already compiled, which is the same behaviour observed for payloads. (this means you must now manually delete a target, when you wish to re-build it; the build/boot/roms logic now more or less just runs cbfstool; blobutil is handled from build/defconfig/for) ALSO: Since crossgcc is now handled by build/defconfig/for, not build/boot/roms, standalone compiling of u-boot is now possible. This has been tested. You compile it like so: ./build defconfig for u-boot or specific trees, e.g. ./build defconfig for u-boot default One other consequence of this patch is that re-building the same ROM image is now much faster, because the same builds are re-used unless deleted. This could be useful when testing grub.cfg changes, for example, if that's all you change. With things like ccache used (not yet used robustly in lbmk), this could speed things up more, depending on the codebase. This patch demonstrates the raw power of lbmk; it is a very simple and highly efficient build system, and now much more so! Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-08-17 22:27:30 +00:00
{
x=${corebootrom}
y=${uboot_config}
z=${cbfstool}
tmprom="$(mkUbootRom "$x" "fallback/payload" "$y" "$z")"
newrompath="${romdir}/uboot_payload_${board}_${initmode}_${displaymode}.rom"
[ "${initmode}" = "normal" ] && \
newrompath="${romdir}/uboot_payload_${board}_${initmode}.rom"
moverom "${tmprom}" "${newrompath}" || \
err "build_roms: moverom fail (u-boot)"
rm -f "${tmprom}" || \
err "build_roms: cannot rm u-boot rom"
}
# make a rom in /tmp/ and then print the path of that ROM
mkUbootRom() {
target_cbrom="${1}" # rom to insert u-boot in. it won't be touched
# (a tmpfile will be made instead)
target_uboot_cbfs_path="${2}" # e.g. fallback/payload
target_uboot_config="${3}"
cbfstool_path="${4}"
build/boot/roms: Support removing microcode From now on, the following rules are available for all mainboards, in resources/coreboot/boardname/board.cfg: * blobs_required="n" or "y" * microcode_required="n" or "y" The blobs setting, if set to "n", simply renames filename.rom to filename_noblobs.rom. The microcode setting, if set to "n", copies the ROM (with or without _noblobs) to filename_nomicrocode.rom (if blobs="n", it would be filename_noblobs_nomicrocode.rom). Where "nomicrocode" is set, ROMs with microcode will still be provided by lbmk and in relesase, but ROMs will also be provided alongside it that lacks any microcode updates. If the *original* ROM already lacks microcode updates, then the original ROM will be *renamed* to include "nomicrocode" in the name. This is done on images for ARM platforms, for instance, where microcode is never used whatsoever. Example filenames now generated: seabios_e6400_4mb_libgfxinit_corebootfb_noblobs_nomicrocode.rom seabios_e6400_4mb_libgfxinit_corebootfb_noblobs.rom seabios_withgrub_hp8300usdt_16mb_libgfxinit_corebootfb_colemak_nomicrocode.rom seabios_withgrub_hp8300usdt_16mb_libgfxinit_corebootfb_colemak.rom uboot_payload_gru_kevin_libgfxinit_corebootfb_noblobs_nomicrocode.rom A vocal minority of people were not happy with some of the changes made in Libreboot last year, including on existing supported hardware from before those changes were made. I did this before the last release, out of respect: https://libreboot.org/news/gm45microcode.html (re-add mitigations for no-microcode setup on GM45) This new change is done as an further, extended courtesy. Tested and works fine. (testing using cbfstool-print) Actual Libreboot policy about binary blobs is nuanced. See: https://libreboot.org/news/policy.html (reduction policy) and: https://libreboot.org/freedom-status.html (implementation) Well, the status page talks about descriptor vs non-descriptor on Intel platforms, and where me_cleaner is used (on platforms that need Intel ME firmware), it regards the descriptored setups to be blob-free if coreboot does not require binary blobs. In this paradigm, microcode updates are not considered to be binary blobs, because they aren't technically software, they're more like config files that just turn certain features on or off within the CPU. However, for lbmk purposes, "noblobs" means that, after the ROM is fully ready to flash on the chip, there will be no blobs in it (except microcode). So for example, an X200 that does not require ME firmware is considered blob-free under this paradigm, even though Libreboot policy regards X230 as equally libre when me_cleaner is used; in this setup, ROMs will not contain "blobfree" in the filename, for X230 (as one example). Signed-off-by: Leah Rowe <leah@libreboot.org>
2023-06-18 13:12:31 +00:00
target_ubdir="elf/u-boot/${board}/${target_uboot_config}"
target_ubootelf="${target_ubdir}/u-boot.elf"
[ -f "${target_ubootelf}" ] || \
target_ubootelf="${target_ubdir}/u-boot.bin"
[ -f "${target_ubootelf}" ] || \
err "mkUbootRom: ${board}: cant find u-boot build"
tmprom=$(mktemp -t coreboot_rom.XXXXXXXXXX)
cp "${target_cbrom}" "${tmprom}" || \
err "mkUbootRom: cannot copy to tmprom"
"${cbfstool}" "${tmprom}" add-payload -f "${target_ubootelf}" \
-n ${target_uboot_cbfs_path} -c lzma || \
err "mkUbootRom: cannot add u-boot to tmprom"
printf "%s\n" "${tmprom}"
}
# it is assumed that no other work will be done on the ROM
# after calling this function. therefore this function is "final"
moverom() {
rompath="${1}"
newrompath="${2}"
printf "\nCreating new ROM image: %s\n" "${newrompath}"
[ -d "${newrompath%/*}" ] || mkdir -p "${newrompath%/*}/" || \
err "moverom: cannot create rom directory: \"${newrompath%/*}\""
[ "${blobs_required}" = "n" ] && \
newrompath="${newrompath%.rom}_noblobs.rom"
cp "${rompath}" "${newrompath}" || err "moverom: can't copy rom"
}
main $@