MRC caches in a certain way, that Heads was able to work
around in their build system, for this board.
I've adapted the relevant config differences, from their project
as of heads revision 96440b928acb06de5b925ea12014c9c280b23165
The downside is that CBFS now has to be 8MB in size. The upside
is that the machine also boots much faster
See:
f0792117efhttps://github.com/osresearch/heads/pull/1282#issuecomment-1400634600
I have not adapted their IFD changes, versus Libreboot, because theirs
simply has a different version string, and uses different read/write
permission bits for regions as defined in the IFD.
This affects:
t440p_12mb_mrc
w541_12mb_mrc
S3 suspend/resume still broken on these targets which use the libre
MRC init (replacement code by Angel Pons, recently merged in lbmk):
t440p_12mb
w541_12mb
With clever use of FMAP, the rest of the BIOS region might still be
used. However, for our purposes, 8MB CBFS will do just fine.
Heads's changes configure MRC so that caching is handled properly,
for when the machine returns from sleep. Setting CBFS to be any
higher will result in slower boot times, and broken S3 resume, due
to MRC cache misalignment (this is based on my understanding, reading
through the Heads project looking at their research on this).
At some point in the future, Angel's libre MRC code will probably
be finished, and merged, with more fine tuning possible to allow
bigger CBFS sizes.
libre mrc on haswell is quite buggy for now, but works in
a limited fashion
this patch re-adds the old configs, but as _mrc for example
t440p_12mb_mrc instead of t440p_12mb
and t440p_12mb (without _mrc) still uses the libre mrc code
i found that with libre mrc, usb was broken in grub
however, it worked nicely in seabios
for our purposes, doing seabios-only roms in text mode
is best for now
i'm going to re-add mrc.bin, but for t440p_12mb_mrc
and w541_12mb_mrc, as new config names. the regular
t440p_12mb and w541_12mb will continue to use libre
mrc, but the _mrc ones will use mrc.bin and retain the
grub payload in board.cfg
courtesy of Angel Pons from the coreboot project
this uses the following patch set from gerrit, as yet
unmerged (in coreboot master) on this date:
https://review.coreboot.org/c/coreboot/+/64198/5
logic for downloading mrc blobs has been deleted from
lbmk, as this is now completely obsolete (for haswell
boards)
if other platforms are added later that need mrc.bin,
then logic will be re-added again for that
this fixes the build error:
Error: name not set
Usage: ./download gitmodule [name]
when running:
./download all
running "all" runs all scripts under downloads,
one of which was the gitmodule script itself, therefore
being run without argument
reduce the number of calls to read() by using
bit shifts. when rnum is zero, read again. in
most cases, a nibble will not be zero, so this
will usually result in about 13-15 of of 16
nibbles being used. this is in comparison to
8 nibbles being used before, which means that
the number of calls to read() are roughly
halved. at the same time, the extra amount of
logic is minimal (and probably less) when
compiled, outside of calls to read(), because
shifting is better optimised (on 64-bit machines,
the uint64_t will be shifted with just a single
instruction, if the compiler is decent), whereas
the alternative would be to always precisely use
exactly 16 nibbles by counting up to 16, which
would involve the use of an and mask and still
need a shift, plus...
you get the point. this is probably the most
efficient code ever written, for generating
random numbers between the value of 0 and 15
some checks check for specific utils, which are
then used to indicate the existence of other utils,
which means that building them singularly, as is
currently done, may result in errors later if another
tool doesn't exist compiled yet
this is an obscure bug, fixed by this patch. more of a
workaround really. a dirty hack. when checking for any
of the coreboot utilities required, build all coreboot
utilities that are possibly required
the utilities are small enough that this does not add
much extra time to build, and in most cases, all of them
will be needed anyway
U-Boot can be configured via environment variables which can be saved to
various storage devices. This usually defaults to MMC or SPI depending
on where it boots from, but assumes the device's layout is controlled by
U-Boot.
We should store the environment in SPI flash, but we also need to
configure coreboot FMAPs to reserve the area U-Boot would use as its
environment storage. For now, disable environment storage by setting
ENV_IS_NOWHERE=y to avoid overwriting random regions of SPI or MMC if
someone tries to save the variables.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Set default U-Boot revision to v2023.01 and rebase patches on top of
that. Upstream kconfig status is a bit unstable, so updating configs
with `make oldconfig` would miss important upstream changes.
For each board, run `make savedefconfig` and `diffconfig -m` at the old
version to get a diff from upstream defconfigs. Fix those affected by
upstream changes, like SYS_TEXT_BASE being renamed to TEXT_BASE. Then
append those to the new version's defconfigs and run `make olddefconfig`
to get updated configs.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
the way nvmutil is designed, setWord() is only ever called
under non-error conditions. however, if one part is valid but
the other one isn't, and a command is run that touches both parts,
errno is non-zero write writeGbeFile is called
in situations where one part is valid, but the other isn't, AND the
writes to gbe (in memory) results in a non-change, writeGbeFile is
not called; in this situation, errno is not being reset, despite
non-error condition
this patch fixed the bug, resulting in zero status upon exit under
such conditions
the current code writes part 1 first, and part 0 next,
on the disk, due to the way the swap works.
with this change, swap still swaps the two parts of the file,
on disk, but writes the new file sequentially.
this change might speed up i/o on the file system, on HDDs.
on SSDs, this change likely makes no difference at all.
On many Lenovo GbE regions (in factory firmware), part 0 is
invalid but part 1 is valid.
This change means part 1 is checked first. If part 1 is valid,
part 0 won't be checked at all (due to how most C compilers
optimise).
Most people are just going to extract the factory GbE file,
modify it and re-insert it into the ROM image, so this causes
a nice speedup.
don't constantly open/close the file: /dev/urandom
only read 12 bytes at a time
because of this change, the readFromFile() function now only
handles gbe files
Leah Rowe