Commit 7d6a15e63c95 ("Convert toolchain file to TOML syntax") switched
from the bare toolchain file to the TOML-based one for better management
of the toolchain and components used.
Commit cf7fc2c540d5 ("scripts: Update rustup or inform user of env
vars") added an explicit `rustup self update` because there were still
cases, a year later, of people not having a rustup new enough to support
the TOML-based toolchain file.
Now 2 years after that, it should be safe to drop the explicit self
update. The TOML format has widespread adoption and rustup now self
updates by default. This should allow distro-provided rustup, which
disables the self update feature, to work if it is already installed in
place of the one downloaded from https://rustup.rs.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
- Do cold reboot on success instead of shutting down
- Reboot on success after flashing all components
- Do not boot to firmware setup after MeSet
- Flash full EC ROM size instead of hard-coding to 128K
- Add addw4, lemp13, oryp12
Signed-off-by: Tim Crawford <tcrawford@system76.com>
CSME should be enabled when S0ix is used for power savings during
suspend. TGL-U is just one case where we default as our TGL-U models
do not support S3. Other models may use S0ix as well, such as darp8 and
darp9 due to new batches just not working with S3 anymore [1].
[1]: https://github.com/system76/firmware-open/issues/507
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Cherry-pick the following upstream commits:
- coreboot/coreboot@fb401e74da: soc/intel/alderlake: Sync UPD Usb4CmMode with Kconfig
- coreboot/coreboot@961ed9fe27: soc/intel/adl: Set slp-s0 counter frequency
- coreboot/coreboot@377845a9d4: soc/intel/tigerlake: Fix processor hang while plug unplug of TBT device
- coreboot/coreboot@354a54ac84: soc/intel/tigerlake: Remove IOM Mctp command from TCSS ASL
- coreboot/coreboot@adf042f6c6: lib/rtc: Fix off-by-one error in February day count in leap year
This also brings in the WIP addw4, but that will documented in the PR
that adds full support for the board.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Commit 77581d11fcda ("models: Remove generated C files") removed the HDA
and GPIO files, but bonw14 had an odd named file that was not removed
with the `find` command.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Since 4.19-based releases took so long, we are already at the next
coreboot release. Update to this release for the slew of MTL fixes.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
These files are generated while running proprietary firmware, but once
they are added to coreboot they are no longer needed. They also quickly
become stale as the files are always changed in coreboot.
find models/ -name gpio.h -exec rm {} \;
find models/ -name hda_verb.c -exec rm {} \;
Signed-off-by: Tim Crawford <tcrawford@system76.com>
I always add HDA reset to the coreboot file when upstreaming, so just
have the script do it.
- Proprietary firmware does it
- It fixed audio init on oryp5
Signed-off-by: Tim Crawford <tcrawford@system76.com>
The whole coreboot generated `.config` has not been used since
ff0a27ad9c60 ("Use defconfig to generate .config"). The required
`coreboot.config` is much simpler now and can usually just be copied
from another model.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
This script is stale. It hasn't worked since TGL and only produced a few
useful values. The format of devicetree.cb also continues to change with
each SoC generation.
The removal of devmem2 should also fix setup on Debian, which does not
package it.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Always sync submodules before the SCM step to handle coreboot adding
new submodules. This is needed because coreboot uses relative paths for
submodules, so git tries to clone it from the System76 GitHub
organization. We are not mirroring submodules and want to clone from
gerrit.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
As part of this, the RPL boards now use the client FSP release from
Intel's public repo [1].
[1]: https://github.com/intel/fsp
Signed-off-by: Tim Crawford <tcrawford@system76.com>
coreboot expects paths to be relative to it, and vboot breaks include
paths that are absolute. Fixes compilation failing due to the addition
of `-Wmissing-include-dirs`.
Ref: https://review.coreboot.org/c/coreboot/+/70251
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Fixes suspend with the following drives:
- Kingston KC3000 (SKC3000D/4096G)
- Kingston HyperX (SHPM2280P2H/240G)
- Solidigm P44 Pro (SSDPFKKW010X7)
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Avoid the issue on new boards where the PCH will not de-assert the
`PLTRST#` virtual wire on S3 resume, causing the unit to hang.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Secure Boot support is enabled. Make it clear in the doc that it was
enabled so Windows could be installed, and not as a means for securing
the system.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Report issues by shell files with:
git ls-files '*.sh' | xargs shellcheck --exclude=SC2162
Address the following:
- SC1087: Use braces when expanding arrays
- SC1091: Not following
- SC2004: `$`/`${}` is unnecessary on arithmetic variables
- SC2024: `sudo` doesn't affect redirects
- SC2034: foo appears unused. Verify it or export it
- SC2086: Double quote to prevent globbing and word splitting
- SC2087: Quote `EOF`
- SC2115: Use `"${var:?}"` to ensure this never expands to `/*`
- SC2148: Add a shebang
Addresses (at least partially) some POSIX/dash issues:
- SC2113: `function` keyword is non-standard
- SC3010: In POSIX sh, `[[` `]]` is undefined
- SC3014: In POSIX sh, `==` in place of `=` is undefined
- SC3020: In POSIX sh, `&>` is undefined
- SC3046: In POSIX sh, `source` in place of `.` is undefined
Does not address:
- SC2162: `read` without `-r` will mangle backslashes
- Any other POSIX/dash-specific issues
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Specify dependencies specifically for building coreboot toolchains in
the file, so they are not conflated with the dependencies required for
building firmware-open.
Remove building the toolchain when building firmware, so that the new
script is the single source for building coreboot toolchains.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Allow overriding the coreboot toolchain location with `XGCCPATH` so
that xgcc can be installed to a separate location, such as `/opt`.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Building EC was broken as the build output from the previous board was
not being cleaned. `BUILD` is now always set, defaulting to "build",
instead of only when `BOARD` is specified. For good measure, add it
to the clean command in case a custom path is used.
Fixes: 569321f9ac79 ("scripts: Set EC build dir")
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Fix a potential lock up during S0ix opportunistic suspend, caused by
waiting indefinitely for a PECI command to complete when it is not
available.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Specify the EC build directory so that it does not require the model or
version to be determined in the script.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
- Use `GIT_COMMIT` instead of stashing the commit
- Do not shadow built-in `GIT_BRANCH` variable with a parameter
- Update `scmGit` config based on docs
Signed-off-by: Tim Crawford <tcrawford@system76.com>
I had forgotten that CSME reporting an error was normal when disabled.
Document what is should report when enabled/disabled.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
The scripted projects using the default Slack notifications settings
report the time the build took. Add it to the declarative jobs as well.
Signed-off-by: Tim Crawford <tcrawford@system76.com>
