Building of LLVM tests and examples is enabled by default, but they are
not necessary. Thus disable them.
Change-Id: I58b09e276967e97856da65e5876b27f0bae3f0cc
Signed-off-by: Felix Singer <felixsinger@posteo.net>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69942
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
To improve the readability and visibility of the configure options, move
each of them to a separate line.
Change-Id: Ifc39e4d0849d220d85e1d9ce92fc008fec610694
Signed-off-by: Felix Singer <felixsinger@posteo.net>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69941
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
The patches for kconfig need to be in a format compatible with the quilt
tool, and usually also contain a header with some additional info like
the git commit. This header is in the same format as patches produced by
`git format-patch`, but the diff style git uses is incompatible with
quilt and there does not seem to be a straightforward way to format the
diff section to work.
Add some documentation for a method I found to go from a git commit to a
quilt compatible patch with git headers.
Change-Id: I7a8bbe41e0864be1d28116742b6b8b3fc440cc31
Signed-off-by: Nicholas Chin <nic.c3.14@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69458
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
CPU nodes are now declared in a common chipset.cb.
TESTED: generates a proper devicetree for x220 based on logs.
Change-Id: Ic1f2d3d611aa3979b846706b6f743f79a3c4e54d
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69501
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Right now cbmem uses a fix path to reach coreboot src path (../../).
This makes it impossible to compile cbmem out of the coreboot tree (e.g.
copy just the cbmem directory elsewhere and compile).
This patch adapts the technique from cbfstool and adds a variable called
'TOP' which points to coreboot root directory and which can be
overridden at build time by providing it to make as an argument. This
will enable a stand-alone build of cbmem.
Change-Id: I2732f75310e10716e5aa74e094e0bf628ad22f0b
Signed-off-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69686
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Singer <felixsinger@posteo.net>
The download links from acpica.org [1] are not stable, and for some
reason they named the release tarballs with .tar_0.gz. Thus, use the
tarballs from their GitHub repository generated out of the release
tags [2].
Tested locally and also IASL patch applies.
[1] https://www.acpica.org/downloads
[2] https://github.com/acpica/acpica/tags
Signed-off-by: Felix Singer <felixsinger@posteo.net>
Change-Id: I7b10dd1db4299aaef96bc29023bed874b660aba0
Reviewed-on: https://review.coreboot.org/c/coreboot/+/70021
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
GCC 12 incorrectly warns about an array out of bounds issue:
```
$ make V=1 # emulation/qemu-i440fx
[…]
CC ramstage/arch/x86/ebda.o
x86_64-linux-gnu-gcc-12 -MMD -Isrc -Isrc/include -Isrc/commonlib/include -Isrc/commonlib/bsd/include -Ibuild -I3rdparty/vboot/firmware/include -include src/include/kconfig.h -include src/include/rules.h -include src/commonlib/bsd/include/commonlib/bsd/compiler.h -I3rdparty -D__BUILD_DIR__=\"build\" -Isrc/arch/x86/include -D__ARCH_x86_32__ -pipe -g -nostdinc -std=gnu11 -nostdlib -Wall -Wundef -Wstrict-prototypes -Wmissing-prototypes -Wwrite-strings -Wredundant-decls -Wno-trigraphs -Wimplicit-fallthrough -Wshadow -Wdate-time -Wtype-limits -Wvla -Wdangling-else -fno-common -ffreestanding -fno-builtin -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -fno-pie -Wno-packed-not-aligned -fconserve-stack -Wnull-dereference -Wreturn-type -Wlogical-op -Wduplicated-cond -Wno-unused-but-set-variable -Werror -Os -Wno-address-of-packed-member -m32 -Wl,-b,elf32-i386 -Wl,-melf_i386 -m32 -fuse-ld=bfd -fno-stack-protector -Wl,--build-id=none -fno-delete-null-pointer-checks -Wlogical-op -march=i686 -mno-mmx -MT build/ramstage/arch/x86/ebda.o -D__RAMSTAGE__ -c -o build/ramstage/arch/x86/ebda.o src/arch/x86/ebda.c
In file included from src/arch/x86/ebda.c:6:
In function 'write_ble8',
inlined from 'write_le8' at src/commonlib/include/commonlib/endian.h:155:2,
inlined from 'write_le16' at src/commonlib/include/commonlib/endian.h:178:2,
inlined from 'setup_ebda' at src/arch/x86/ebda.c:35:2,
inlined from 'setup_default_ebda' at src/arch/x86/ebda.c:48:2:
src/commonlib/include/commonlib/endian.h:27:26: error: array subscript 0 is outside array bounds of 'void[0]' [-Werror=array-bounds]
27 | *(uint8_t *)dest = val;
| ~~~~~~~~~~~~~~~~~^~~~~
[…]
```
[In GCC 12 the new parameter `min-pagesize` is added and defaults 4 kB.][1]
It treats INTEGER_CST addresses smaller than that as assumed results of
pointer arithmetics from NULL while addresses equal or larger than that
as expected user constant addresses. For GCC 13 we can represent results
from pointer arithmetics on NULL using &MEM[(void*)0 + offset] instead
of (void*)offset INTEGER_CSTs.
[1]: https://web.archive.org/web/20220711061810/https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99578
TEST=No compile error with gcc (Debian 12.2.0-3) 12.2.0
Change-Id: I6e36633f42cb4dc5af53212c10c919a86e451ee0
Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/62830
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
We don't currently use the artifacts from the Clang or CrOS GCC
builds, so don't bother saving them.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I69fe803e4b4213a199d0b76089da443aa769aa92
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69954
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Singer <felixsinger@posteo.net>
This reduce the difference with linux v6.0-rc3.
Signed-off-by: Elyes Haouas <ehaouas@noos.fr>
Change-Id: Ica20264d744ea8f77b56c63d29e1fafc2e68a869
Reviewed-on: https://review.coreboot.org/c/coreboot/+/67338
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin L Roth <gaumless@gmail.com>
This patch extends the available device paths with a new device 'mdio'.
MDIO is the 'Management Data Input/Output' called interface which is
used to access an Ethernet PHY behind a MAC to change settings. The real
payload data path is not handled by this interface.
To address the PHY correctly on the MDIO bus, there is a 5 bit address
needed, which often can be configured via pins on the mainboard.
Therefore, the new introduced device has an 'addr' field to define its
address. If one wants to use a MDIO device in devicetree, the syntax is
straight forward (example):
device mdio 0x2 on end
As the MDIO interface is driven by the MAC, most likely this MDIO device
will be hooked in as a child device of the (PCI attached) MAC device.
With the new introduced ops_mdio a new interface is added to provide an
API for read and write access over MDIO.
Change-Id: I6691f92c4233bc30afc9029840b06f74bb1eb4b2
Signed-off-by: Mario Scheithauer <mario.scheithauer@siemens.com>
Signed-off-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69382
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
No real functional changes, just cleaning up shellcheck issues, putting
braces around variables, add comments and the like.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I6e79afc8d725e86ddbf7f4eb4685bed190c20738
Reviewed-on: https://review.coreboot.org/c/coreboot/+/67319
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
"../cmake" introduced on Change-Id: I3144a83
Remove "../cmake" when the build is done.
Change-Id: I289bfaca1fd8d3f004455babd99849ca8aa2d6db
Signed-off-by: Elyes Haouas <ehaouas@noos.fr>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69869
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin L Roth <gaumless@gmail.com>
coreboot only supports a small subset of the targets that LLVM supports.
It's not needed to enable all possible targets. Thus limit the targets
to the following ones:
* X86
* RISC-V
* AArch32
* AArch64
* PowerPC
Change-Id: I9938bf176b5fe2b0a631c3b1ae858f988898a196
Signed-off-by: Felix Singer <felixsinger@posteo.net>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69841
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur.heymans@9elements.com>
With the addition of the clang tests, the jenkins builds are taking a
really long time to run the tests. This change allows the
"what-jenkins-does" build to be split into separate builds on jenkins.
Additionally, some jenkins builds like coverity don't need (or want)
to build clang or even the linters.
Update help with the variables.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I0f8ac68c1bc8f8ff9be62d80db850355e742ee74
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69495
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
This tests building a single target with scanbuild so to make sure that
option hasn't been broken. Since it's a different type of build, it
hasn't previously been tested with what-jenkins-does.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I8a74dac203f4d38c0cb30a0b64724e6f9095b9dd
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69861
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
This gets rid of the duplicated directory and xml filename and uses the
--name argument to abuild instead, which also updates the test name in
the junit xml file.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: Ibe538da42280696190b0a7a0c63fd86a63e40214
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69860
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Previously, the testclass variable was only updated with the chromeos
or Kconfig option values, and the output directory and xml file names
were updated independently.
With the --name option, all of these can be set simultaneously. This
also prevents jenkins from seeing clang and gcc tests as the same
because the testclass variable wasn't updated.
If --name is not set, all behavior is as it was previously.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I8f52779b92d213386a3eb371d1f30ee32ed48b85
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69859
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Instead of having duplicate lines in the what-jenkins-does target and
the test-tools target, make test-tools from what-jenkins-does.
Now there's only one place to update when changing the call.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: Id62d6bb1e729892ec123ea970ca8a31e03a812d0
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69838
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
It's hard to tell what is what with the short option names, so use the
long options here.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I1371e098bba1077dedfaffa56287a28656197b40
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69837
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Instead of having duplicate lines in the what-jenkins-does target and
the test-abuild target, make test-abuild from what-jenkins-does.
The test-abuild target had not been updated to use the ABUILD_OPTIONS
variable, so update it with the commands from what-jenkins-does.
Now there's only one place to update when changing the call.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I4552193894c16301defb851eb3db4bdfbfa49803
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69836
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Instead of having duplicate lines in the what-jenkins-does target and
the test-lint target, make test-lint with the --junit argument from
what-jenkins-does.
Now there's only one place to update when changing the call.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I2f90df76126f453fbcd91f4c4af5d784ac2dbe88
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69835
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Instead of having the what-jenkins-does target clean up before building,
have it call the test_cleanup target.
Clean the tegra targets.
Remove distclean from test_cleanup target - I don't think that's
expected, and people might be upset by having their .config deleted.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: Ia9d585df05343365c89e49b1c01dba9ba865003f
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69834
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
If no boards are tested by abuild, an error is currently shown because
no boards failed, but no boards passed either. Account for this
possibility.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I097d3c728ca1acc652d5a1b7b49e57d01b0e513b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69520
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
The change being reverted [1] caused all the Kconfig dependency files
to be generated at the top level of coreboot's build directory.
This reverts that behavior and puts the dependencies back where we're
used to them being.
[1] 1b9e740a81
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: Ic4b48831705c3206e7c2e09f01d072d1cde9c9c4
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69535
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
revert commit 1b9e740a8 (kconfig: fix failing to generate auto.conf) [1]
The above change caused all of the enabled kconfig options to be written
into the top level build directory. We don't want that, so go back to
the old behavior for the coreboot tree.
[1] 1b9e740a81
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I2904f69a5d85337ad0a6b48590ccd4b4a6e38b70
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69536
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Testing for the presence of intel-sec-tools doesn't need to happen
inside the what-jenkins-does target.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I6faa5bd5292ac5cceba9a64fe81939c0e25b9f3e
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69519
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Also comment out the SMSC FDC37M602 which has a conflicting
ID and has never had the LDN/register layout anyway.
Tested on a Dell Latitude E6400
Change-Id: I5b1900e6ef599c422a1d6eca7a2ac4691d56d874
Signed-off-by: Nicholas Chin <nic.c3.14@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69481
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
There doesn't seem to be a datasheet available for the NCT6685D, but
there is one for the NCT6686D. The 85D seems to return the same ID as
the 86D, and the registers do seem to be returning valid data other than
LDN 0xf which returns all 1s. The LDN and register layout appears to be
identical to the NCT6687D-W.
Tested on a Lenovo ThinkCentre M900 with a NCT6685D.
Change-Id: I4de0e7b86422a14ab9ccb15b7571597611d755d5
Signed-off-by: Nicholas Chin <nic.c3.14@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69480
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
-malign-abi does not exist on clang (v15.0.0) and the -ccc-gcc-name
variable is not needed anymore.
TESTED: This also boots on qemu q35
Change-Id: I7f99ebea18d5c09fdc7ced5c793d57d6fedd2e47
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/69232
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin L Roth <gaumless@gmail.com>
Some platforms correctly build and boot with clang. Add this to our CI.
Change-Id: I82d756e071a0e575db73fbd91167d27cae3ddc18
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/62173
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin L Roth <gaumless@gmail.com>
These are used on newer platforms.
Change-Id: I20dc77fb6f83dc813e3da5fe30f8f52068fc4662
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/68119
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Sean Rhodes <sean@starlabs.systems>
This finds all the boards using a specified Kconfig option and runs both
CrOS and non-CrOS abuilds on them to make sure they're working.
Nobody wants to run the full what-jenkins-does build on their host
machine. Hopefully this can help get some tests run locally before
pushing to coreboot.org.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: Ifc71c28bf64a805f203a815a9468ff9fe882aad3
Reviewed-on: https://review.coreboot.org/c/coreboot/+/68956
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
The functions create_bpdt_hdr and create_cse_layout
in bpdt_1_6.c are defined to return pointers but
not integers as was previouly implemented.
Reported-by: Coverity(CID:1469323)
Reported-by: Coverity(CID:1469353)
Signed-off-by: Solomon Alan-Dei <alandei.solomon@gmail.com>
Change-Id: Idb78d94be7a75a25ad954f062e9e52b1f0b921dc
Reviewed-on: https://review.coreboot.org/c/coreboot/+/68986
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin L Roth <gaumless@gmail.com>
Fix readouts from the hardware monitor on Fintek F71808A Super I/O.
The HWM port is +5 to the base address stored in LDN 0x4 at index 0x60/0x61.
Referred to util/superiotool/winbond.c and the Linux kernel driver f71882fg.
Tested on a HP 500-319na (Memphis-S / IPM87-MP).
Signed-off-by: Ravi Mistry <rvstry@protonmail.com>
Change-Id: I2b2b98c62f9305c6f4885c2ce3b1444801dcb9d5
Reviewed-on: https://review.coreboot.org/c/coreboot/+/62060
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Some of the cbmem area names have gotten longer, and were making the
output of cbmem -l look bad, so expand the name area to 20 characters.
Instead of printing a blank area if the name isn't recognized, call it
unknown.
Change the method of printing the title to match the way the actual text
of the table is printed.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I9d91d21c6ad418d9fee9880550fb6cb9e41e93f0
Reviewed-on: https://review.coreboot.org/c/coreboot/+/68644
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-by: Eric Lai <eric_lai@quanta.corp-partner.google.com>
Felix Singer