the temporary rom per build was not being deleted after
finishing the current target. this adds up in /tmp during
large builds, when building for many targets. fix this!
Signed-off-by: Leah Rowe <leah@libreboot.org>
release="n" is set in target.cfg on haswell build targets
that use mrc.bin
script/update/release exports LBMK_RELEASE="y"
script/build/roms skips building a given target if release="n"
in target.cfg *and* LBMK_RELEASE="y"
you could also do the export yourself before running ./build roms,
for example:
export LBMK_RELEASE="y"
./build roms all
This would skip these ROM images. The native haswell raminit is
now stable enough in my testing, that I wish to delete the MRC-based
targets. This is in line with Libreboot's Binary Blob Reduction Policy,
which states: if a blob can be avoided, it should be avoided.
The problem is that users often run the inject script in *lbmk* from
Git, instead of from the src release archive. I forsee some users
running this on modern lbmk with older release images. If the mrc-based
target isn't there, the user may use an NRI-based target name, and
think it works; they will insert without MRC. I foresaw this ages
ago, which is why Caleb and I ensured that the script checks hashes,
and hashes are included in releases.
Therefore: for the time being, keep the MRC-based configs in lbmk
but do not include images for them in releases. This can be done
indefinitely, but I'll probably remove those configs entirely at
some point.
On the following boards, Libreboot now will *only* provide NRI-based
ROM images for the following machines:
* Dell OptiPlex 9020 SFF
* Dell OptiPlex 9020 MT
* Lenovo ThinkPad T440p
* Lenovo ThinkPad W541/W540
I now recommend exclusive use of NRI-based images, on Haswell
hardware. It's stable enough in my testing, and now supports S3.
Signed-off-by: Leah Rowe <leah@libreboot.org>
in shell scripts, a function named the same as a program included in
the $PATH will override that program. for example, you could make a
function called ls() and this would override the standand "ls".
in lbmk, a part of it was first trying to run the "fail" command,
deferring to "err", because some scripts call fail() which does
some minor cleanup before calling err.
in most cases, fail() is not defined, and it's possible that the user
could have a program called "fail" in their $PATH, the behaviour of
which we could not determine, and it could have disastrous effects.
lbmk error handling has been re-engineered in such a way that the
err function is defined in a variable, which defaults to err_ which
calls err_, so defined under include/err.sh.
in functions that require cleanup prior to error handling, a fail()
function is still defined, and err is overridden, thus:
err="fail"
this change has made xx_() obsolete, so now only x_ is used. the x_
function is a wrapper that can be used to run a command and exit with
non-zero status (from lbmk) if the command fails. the xx_ command
did the same thing, but called fail() which would have called err();
now everything is $err
example:
rm -f "$filename" || err "could not delete file"
this would now be:
rm -f "$filename" || $err "could not delete file"
overriding of err= must be done *after* including err.sh. for
example:
err="fail"
. "include/err.sh"
^ this is wrong. instead, one must do:
. "include/err.sh"
err="fail"
this is because err is set as a global variable under err.sh
the new error handling is much cleaner, and safer. it also reduces
the chance of mistakes such as: calling err when you meant to
call fail. this is because the standard way is now to call $err,
so you set err="fail" at the top of the script and all is well.
Signed-off-by: Leah Rowe <leah@libreboot.org>
There is no need to add multiple keymap files, because
GRUB can load keymaps from CBFS. The current build logic
is designed to avoid building multiple GRUB binaries,
which are expensive computationally because each one
would then have to be compressed for each board.
This patch provides the best of both worlds: less space
used in flash like in the old lbmk design (1 keymap per
board), but retaining the current build speeds and therefore
not re-introducing the slowness of lbmk's previous GRUB
build logic.
The grub.cfg file has been modified, accordingly. It now
only loads a keymap.gkb file from CBFS, by default. It does
this, only if that file exists; if not, GRUB already defaults
to US Qwerty layout anyway.
ALSO: compress all keymap gkb files with xz -6
GRUB automatically decompresses files when accessed.
This results in about 2KB of flash space saved in CBFS.
Here is real-world data, showing the increased flash space:
< fallback/payload 0x3eb80 simple elf 548821 none
< keymap.cfg 0xc4bc0 raw 16 none
< (empty) 0xc4c00 null 11633316 none
---
> fallback/payload 0x3eb80 simple elf 546787 none
> keymap.gkb 0xc43c0 raw 344 none
> (empty) 0xc4540 null 11635044 none
This was taken by diffing the cbfstool "print" output,
both before and after. The *after* result is with this change.
11633316. In this example, 1728 bytes have been saved. Therefore,
with compression taken into account, this patch saves about 1.7KB
of space in CBFS.
This change means that lbmk can now scale to support hundreds
of keymaps, without increasing the amount of flash space used,
in each given image. Since the keymap files are compressed in
lbmk.git, in advance, we spend no additional time on compression
at build time. The resulting change in build speed in negligible.
Adding your own keymap.gkb file was already possible, for changing
the keymap in libreboot images, if you didn't want to change the
memdisk (and thus re-compile grub.elf). Now, this is the default
behaviour, and the only way to do it. It's much more efficient.
The original keymap files can be restored, by running unxz.
Signed-off-by: Leah Rowe <leah@libreboot.org>
the "kmapdir" variable was removed in an earlier audit,
but was overlooked for -k because that option was untested.
rather than initialise the variable, re-use grubcfgsdir.
this fix enables e.g. "-k usdvorak" to work again.
Signed-off-by: Leah Rowe <leah@libreboot.org>
the update/trees script checks this binary itself, before
deciding whether to recompile/compile, so we don't need
to do such checks here.
Signed-off-by: Leah Rowe <leah@libreboot.org>
the script can now also handle autoconf build systems,
whereas this could previously only be done for grub.
with this change, the overall sloccount is also lower
Signed-off-by: Leah Rowe <leah@libreboot.org>
arch no longer needs to be set, on multi-tree projects,
and it has been renamed to xarch
the new behaviour is: if xarch is set, treat it as a
list of crossgcc targets and go through the list. set
the first one as the target, for what lbmk builds, but
build all of the defined crossgccc targets
crossgcc_ada is now xlang, and defines which languages
to build, rather than whether to build gcc-gnat
Signed-off-by: Leah Rowe <leah@libreboot.org>
while seemingly pedantic, this does actually make code
easier to read. mostly just switching to shorthand for
variable names, where no expansions or patterns are used
Signed-off-by: Leah Rowe <leah@libreboot.org>
the rom functions print a path to the rom they built,
which is then used, but these are called inside what
are essentially subshells, and we had no error handling
Signed-off-by: Leah Rowe <leah@libreboot.org>
cros roms are always using libgfxinit, with a coreboot
framebuffer, so the "normal" initmode is never used.
Signed-off-by: Leah Rowe <leah@libreboot.org>
lbmk used to set version/versiondate directly in
err.sh, but now it's handled there by a function,
which is called by the main script.
script/update/release hadn't yet been adapted. the
only change necessary is to call check_project()
script/update/trees also makes use of it
script/build/roms is using "projectname"
Signed-off-by: Leah Rowe <leah@libreboot.org>
at present, lbmk can remove microcode updates on images for
a given target, if the target specifies
microcode_required="n" in target.cfg
lbmk then provides images with microcode, and images without,
in a given release. although the user can also remove them
manually, this just makes it a bit more convenient, for those
users who do wish to run without the updates. this functionality
is provided only on those platforms where no-microcode is tested.
well, this behaviour implements a compromise on libreboot policy,
which is to always include microcode updates by default. see:
Binary Blob Reduction Policy
the *canoeboot* project now exists, developed in parallel with
libreboot, and it ships without microcode updates, on the same
targets where lbmk also handled this.
running without microcode updates is foolish, and should not
be encouraged. clean up lbmk by not providing this kludge.
the libreboot documentation will be updated, telling such users
to try canoeboot instead, or to remove the update from a given
libreboot rom - this is still possible, and mitigations such as
PECI disablement on GM45 are still in place (and will be kept),
so that this continues to work well.
Signed-off-by: Leah Rowe <leah@libreboot.org>
don't handle "romtype" at all, in board target.cfg files
add /dev/null as pike2008 rom on amd boards. this serves
the same purpose, adding them as empty vga roms, to add
an empty rom in cbfs. pike2008 cards cause seabios to hang,
when their oproms are executed, so we insert a fake rom
on i945 thinkpads, use the coreboot config option:
CONFIG_INTEL_ADD_TOP_SWAP_BOOTBLOCK
when set, this enables the same bootblock copy, for use
with bucts. these two cases, namely pike2008 roms and
i945 bootblock copies, no longer need to be handled in code
Signed-off-by: Leah Rowe <leah@libreboot.org>
all it did was set -v in the shell, which doesn't yield
very useful results. this is a relic of very old design
in the libreboot build system, that is no longer needed.
Signed-off-by: Leah Rowe <leah@libreboot.org>
lbmk didn't quote certain arguments in commands, or
used ! -z instead of -n, things like that. simple fixes.
Signed-off-by: Leah Rowe <leah@libreboot.org>
most of these are probably redundant, and will never
be called, but lbmk needs to be as safe as possible
under fault conditions. fail early, fail hard.
Signed-off-by: Leah Rowe <leah@libreboot.org>
in cases where lbmk must always return from a function,
there are some cases where it relies on non-zero exit
status, which in practise is always the case, but may
change in the future if the relevant part is modified
e.g. do_something && return 0
the proper form is:
do_something
return 0
also do this for unconditional exits
Signed-off-by: Leah Rowe <leah@libreboot.org>
it didn't work in the past, but it does work nowadays;
specifically, it only worked with libgfxinit in the past,
but not on VGA ROMs.
now it does work on VGA ROMs, tested on e6400 and t1650 so
it was enabled there.
in this setup, a special image is provided where SeaBIOS is
the main payload, but it only loads GRUB; nothing else, every.
this is called SeaGRUB. this setup is useful in cases where
the user only has a GPU that lacks libgfxinit support.
Signed-off-by: Leah Rowe <leah@libreboot.org>
when printing the name of the rom being created, it's
done before the check to rename based on vendorfiles
in target.cfg. this patch fixes that bug.
Signed-off-by: Leah Rowe <leah@libreboot.org>
it wasn't being copied right
the roms under elf/ were being copied, but not the ones
under bin/ - i need to audit it further
for now, i run modify_coreboot_roms from build/roms
instead of update/trees
so, the ones under elf/ no longer have bootblocks copied.
it's only done in bin/
Signed-off-by: Leah Rowe <leah@libreboot.org>
when building only for u-boot, the current script
works just fine. however, when building for other
payloads in additional to u-boot, the final u-boot
stage fails because other payloads are already
inserted via cbfs.
when we build u-boot, we do that last because we want
u-boot setups to only be u-boot, nothing else.
this patch enables qemu x86 to build properly with
u-boot.
Signed-off-by: Leah Rowe <leah@libreboot.org>
keymaps weren't being set in keymay.cfg of cbfs, due
to use of x_ in the rom script, and x_ doesn't handle
quotes or spaces in arguments well.
i'm going to remove use of x_ and xx_ (it's in my todo),
for next release.
for now, hot patch the release. i've gone through and
replaced use of x_ with || err, in some places.
not just the keymap.cfg command, but others too. in case
there are more issues we missed.
this commit is being tagged "20231021fix" and i'm using
this tag to re-build the 20231021 release. i'll just
replace the tarballs in rsync and add errata to the news
page announcing the release. all i did was break peoples
umlauts, i didn't brick their machines fortunately!
very minor bug. anyway, x_/xx_ is a great idea, but sh
isn't really designed for that style of programming. i'll
go back to using just || err in the next release.
Signed-off-by: Leah Rowe <leah@libreboot.org>
as opposed to the current 3-level structure.
recent build system simplifications have enabled
this change, thus:
./build fw coreboot -> ./build roms
./build fw grub -> ./build grub
./build fw serprog -> ./build serprog
./update project release -> ./update release
./update project trees -> ./update trees
./update vendor download -> ./vendor download
./update vendor inject -> ./vendor inject
alper criticised that the commands were too long,
so i made them shorter!
Signed-off-by: Leah Rowe <leah@libreboot.org>
move it all to other files where items are used, and not
used anywhere else. this reduces the size of vendor.sh.
also remove a few redundant variables, or variables that
are not meaningfully used.
a few items have been moved to include/option.sh
Signed-off-by: Leah Rowe <leah@libreboot.org>
The "u-boot.bin" file generated by U-Boot builds is a raw binary. When
adding payloads to a CBFS, we need to use ELF files with add-payload
or manually pass the entry point and load address of the payload binary
with add-flat-binary.
We primarily use the "u-boot.elf" which gets build with the REMAKE_ELF
option, as it also has the necessary device-tree binary that U-Boot
usually needs to work. When the option is not set (e.g. for QEMU), we
need to use the "u-boot" file which is an ELF.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
it's sometimes done unconditionally. this change
ensures that it is not repeated needlessly.
i observed otherwise that cbfstool would be
re-built from time to time, even if it was built.
Signed-off-by: Leah Rowe <leah@libreboot.org>
also rename elf/coreboot to something scary
some users were flashing roms built under elf/, which
lack payloads. lbmk builds no-payload roms (and payloads)
under elf/ then inserts them, creating full (flashable)
images under bin/
Signed-off-by: Leah Rowe <leah@libreboot.org>
some users reported build errors. technically, there's
nothing wrong with lbmk but it relies on hostcc, and
hostcc is hit or miss when it comes to cross compiling
32-bit, depending on the build system of whatever project.
lbmk needs to handle cross compilation. for now, i'm just
disabling memtest86plus on non-64-bit hosts.
Signed-off-by: Leah Rowe <leah@libreboot.org>
the script used to be called once per target, now it
handles every target. the grub background image wasn't
being set, so if it changed at build time, it would
stay changed.
keep the default in place for each run, while still
allowing target.cfg files to change it per target.
Signed-off-by: Leah Rowe <leah@libreboot.org>
Leah Rowe