Get stack analyzer supported to read annotation file and do basic
annotation on the callgraph.
The basic annotation includes:
1. Add missing calls to the callgraph
2. Ignore functions on the callgraph
BUG=chromium:648840
BRANCH=none
TEST=extra/stack_analyzer/stack_analyzer_unittest.py
make BOARD=elm && extra/stack_analyzer/stack_analyzer.py \
--objdump=arm-none-eabi-objdump \
--addr2line=arm-none-eabi-addr2line \
--export_taskinfo=./build/elm/util/export_taskinfo.so \
--section=RW \
--annotation=./extra/stack_analyzer/example_annotation.yaml \
./build/elm/RW/ec.RW.elf
make BOARD=elm SECTION=RW \
ANNOTATION=./extra/stack_analyzer/example_annotation.yaml \
analyzestack
Change-Id: I4cc7c34f422655708a7312db3f6b4416e1af917f
Signed-off-by: Che-yu Wu <cheyuw@google.com>
Reviewed-on: https://chromium-review.googlesource.com/614825
Reviewed-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
Add a shared library to export task information.
Modified the stack analyzer to get information from the shared library.
Show allocated stack sizes of tasks in the stack analyzer.
To get the all task information (including the allocated stack size),
A shared library is added and compiled with the board to export all
configurations of the tasklist.
BUG=chromium:648840
BRANCH=none
TEST=extra/stack_analyzer/stack_analyzer_unittest.py
make BOARD=elm && extra/stack_analyzer/stack_analyzer.py \
--objdump=arm-none-eabi-objdump \
--addr2line=arm-none-eabi-addr2line \
--export_taskinfo=./build/elm/util/export_taskinfo.so \
--section=RW \
./build/elm/RW/ec.RW.elf
make BOARD=${BOARD} SECTION=${SECTION} analyzestack
Change-Id: I72f01424142bb0a99c6776a55735557308e2cab6
Signed-off-by: Che-yu Wu <cheyuw@google.com>
Reviewed-on: https://chromium-review.googlesource.com/611693
Reviewed-by: Nicolas Boichat <drinkcat@chromium.org>
Add a static analysis tool for firmware stack usage.
Add an new Makefile rule to analyze the stack usages of firmwares.
Details about the tool can be found in extra/stack_analyzer/README.md.
BUG=chromium:648840
BRANCH=none
TEST=extra/stack_analyzer/stack_analyzer_unittest.py
make BOARD=elm && make BOARD=elm build/elm/RW/ec.RW.taskinfo && \
extra/stack_analyzer/stack_analyzer.py \
--objdump=arm-none-eabi-objdump \
--addr2line=arm-none-eabi-addr2line \
./build/elm/RW/ec.RW.elf ./build/elm/RW/ec.RW.taskinfo
make BOARD=${BOARD} SECTION=${SECTION} analyzestack
Change-Id: Ifb1b5f5ad6be8f8b125b14d6ee03e25cb385895b
Signed-off-by: Che-yu Wu <cheyuw@google.com>
Reviewed-on: https://chromium-review.googlesource.com/576411
Reviewed-by: Nicolas Boichat <drinkcat@chromium.org>
Instead of having to pass in both the board and the arch used there
(which might be complicated in the light of different ECs with different
arches), allow passing in the board and a set of cross compilers.
The core/*/build.mk then pick the compiler that is responsible for them.
The current method works just the same: If you've already set
CROSS_COMPILE, no override happens. If you set neither CROSS_COMPILE nor
CROSS_COMPILE_$arch, the same default as before this CL is set.
BUG=none
BRANCH=none
TEST=emerge-{samus,kevin} chromeos-ec behave reasonably with adapted
ebuild
Change-Id: Icf1866f296412dd92ecfe134394224c49f7c3df5
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://chromium-review.googlesource.com/549344
Commit-Ready: Patrick Georgi <pgeorgi@chromium.org>
Tested-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Invoking signer with sudo is required only when signing requires a USB
fob. Let's not use it in unless necessary.
BRANCH=cr50
BUG=chromium:728751
TEST=verified that Cr50 build succeeds when both using and not using
the signing fob.
Change-Id: I8f40bd52f1752bfd88ec002f298b991faf7a2512
Reviewed-on: https://chromium-review.googlesource.com/528373
Commit-Ready: Vadim Bendebury <vbendeb@chromium.org>
Tested-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-by: Mary Ruthven <mruthven@chromium.org>
The upcoming "real" signer update will introduce a version which is
not backwards compatible with the existing one wrt the command line
flags: the command line flag '-b' will have to be present.
To keep the default "dummy" signer in sync let's make it accept and
ignore the '-b' command line flag.
BRANCH=none
BUG=none
TEST=verified that the updated signer and the dummy signer both work.
Change-Id: Ia8ab6d7ae01d249046f267608b5971a7a7c95e29
Signed-off-by: Vadim Bendebury <vbendeb@google.com>
Reviewed-on: https://chromium-review.googlesource.com/517678
Commit-Ready: Vadim Bendebury <vbendeb@chromium.org>
Tested-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-by: Marius Schilder <mschilder@chromium.org>
Reviewed-by: Mary Ruthven <mruthven@chromium.org>
Create an app to extract relevant information
from the EC code base that's used to create Vendor
Information Files (VIFs) needed for USB Type-C
compliance testing.
BUG=chromium:701852
BRANCH=none
TEST=make -j buildall
Compared generated VIFs to expected values
Change-Id: I600ca78b9fb5d2de78aa65a58264c6f79b36ea17
Reviewed-on: https://chromium-review.googlesource.com/455280
Commit-Ready: Sam Hurst <shurst@google.com>
Tested-by: Sam Hurst <shurst@google.com>
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
usbpd1 futility image type is deprecated and should not be used for
new designs. This adds proper support for rwsig image type.
Key and signatures are added at linker stage step (futility cannot
directly create such signed images). Thanks to VB21 header, rwsig.c
can now tell how many bytes of the RW image need to be
cryptographically verified, and ensure that the rest is blank (0xff).
BRANCH=none
BUG=chromium:690773
TEST=make BOARD=hammer; flash, RW image is verified correctly.
TEST=make runtests -j
TEST=For the rest of the tests:
Change config option to CONFIG_RWSIG_TYPE_RWSIG
TEST=make BOARD=hammer; flash, hammer still verifies correctly.
TEST=cp build/hammer/ec.RW.bin build/hammer/ec.RW.bin.orig;
futility sign --type rwsig --prikey build/hammer/key.vbprik2 \
build/hammer/ec.RW.bin
diff build/hammer/ec.RW.bin build/hammer/ec.RW.bin.orig
=> Same file
TEST=Add CONFIG_CMD_FLASH, flashwrite 0x1e000, reboot, EC does
not verify anymore.
TEST=dump_fmap build/hammer/ec.bin shows KEY_RO and SIG_RW at
correct locations.
Change-Id: I50ec828284c2d1eca67fa8cbddaf6f3b06606c82
Reviewed-on: https://chromium-review.googlesource.com/441546
Commit-Ready: Nicolas Boichat <drinkcat@chromium.org>
Tested-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
In the chroot, we use ccache to speed up compilation. However, it's
possible for this cache to become corrupt which could lead to slower
builds. In order to keep things fresh, 'make clobber' should clean up
that ccache cache as well. This commit makes it so.
Note, this is NOT *clearing* the cache entirely.
BUG=chromium:681961
BRANCH=None
TEST=make clobber; verify that ccache is cleaned.
Change-Id: I8ccd7df4858c445485ed740843f3126af0abd33a
Signed-off-by: Aseda Aboagye <aaboagye@google.com>
Reviewed-on: https://chromium-review.googlesource.com/433108
Commit-Ready: Aseda Aboagye <aaboagye@chromium.org>
Tested-by: Aseda Aboagye <aaboagye@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Shawn N <shawnn@chromium.org>
Add a 'echo' command which can be called to generate output. If V=0 then
it is suppressed. This drops the unwanted output when V=0.
Signed-off-by: Simon Glass <sjg@chromium.org>
BUG=chromium:680243
BRANCH=none
TEST=V=0 emerge-reef chromeos-ec; See that the BUILD lines are gone
Change-Id: Ie4474024c84345d427f4d95a9ff194dc740f860f
Reviewed-on: https://chromium-review.googlesource.com/427364
Commit-Ready: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
At present the EC Makefile supports two levels of verbosity:
V unset: Show an abbreviated build log (operation and file only)
V=1: Show the full build log, including all commands
However, even the abbreviated build log includes a lot of output. It is
basically a long list of filenames and is of little use during
development. It is more useful to show just warnings and errors.
Add a new setting, V=0, which provides this.
BUG=chromium:680243
BRANCH=none
TEST=emerge-reef chromeos-ec; test each of V=, V=0, V=1 and see that the
Makefile does the correct thing.
Change-Id: I85c0423c5299fa3ab624ed9f7f7b6b7f73236611
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/427363
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Add HOST_LDFLAGS to cmd_c_to_host to allow for cross-compilation of
utils such as 'ectool'.
BRANCH=master
BUG=none
TEST=Cross compile using arm-oe-linux-gnueabi with corresponding
linkerflags and run resulting binary on armv7 device
Change-Id: Iebdfa92239eb5f245d4ad3623cbdceafd82d1675
Signed-off-by: Moritz Fischer <moritz.fischer@ettus.com>
Reviewed-on: https://chromium-review.googlesource.com/424855
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Martin Roth <martinroth@chromium.org>
In order to reduce build times during development, this change allows an
engineer to build a subset of all boards, writing the results for each
board into the ".failedboards" directory.
The expectation is that, while working on code that affects multiple
boards, the engineer will invoke "make build_boards BOARDS=<board-list>"
periodically to confirm that all affected boards still build.
Before pushing any code for review, however, the engineer is still expected
to run "make buildall" to ensure all boards still build and tests pass.
* Define a new make variable, BOARDS, whose default value is $(boards).
This variable can be overridden on the make command line or via the
environment. Its value must be non-empty.
* Use the BOARDS make variable to control which boards will be built by the
new "build_boards" and "try_build_boards" targets.
* Refactor the "buildall" target so that it depends on "build_boards". This
indirectly allows the boards built to be controlled via the BOARDS make
variable.
* Here are some make command lines that were used to test the behavior of
the change:
make -j buildall # Builds all boards, runs tests.
make -j build_boards # Builds all boards, doesn't run tests.
make -j build_boards BOARDS="cr50 reef" # Builds cr50 and reef boards
make -j build_boards BOARDS= # Fails due to empty BOARDS.
BUG=none
BRANCH=none
TEST=make -j buildall
Change-Id: I833277842e5e20d86eaa90df6457264a2f86a3fa
Reviewed-on: https://chromium-review.googlesource.com/420110
Commit-Ready: Carl Hamilton <carlh@chromium.org>
Tested-by: Carl Hamilton <carlh@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Martin Roth <martinroth@chromium.org>
This prints all the boards that are part of the build. This will be
used on coreboot's build system for running coverity. This allows
the builds to keep in sync as platforms are added, and as anything is
added to the "skip_boards" list.
for board in $(make showboards); do
make proj-$board CROSS_COMPILE="arm-eabi-"
done
BUG=none
BRANCH=none
TEST=make showboards shows boards, and excludes skip_boards list
Change-Id: I34d8bc95cc5c7d05b4f2ec38b16b75f9f0c72c6f
Signed-off-by: Martin Roth <martinroth@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/403498
Reviewed-by: Aseda Aboagye <aaboagye@chromium.org>
Since the ec binaries are so tight on space, it would be nice to be
able to see just how much changes are affecting the size. This allows
users to easily do before/after comparisons.
Sample output:
build/sweetberry/RW/ec.RW.flat shrank by 44 bytes: (43828 to 43784)
build/twinkie/RO/ec.RO.flat shrank by 64 bytes: (46312 to 46248)
build/twinkie/RW/ec.RW.flat shrank by 40 bytes: (45900 to 45860)
build/wheatley/RW/ec.RW.flat shrank by 40 bytes: (102692 to 102652)
Compared 156 of 156 files.
81 files changed.
Total size change: -3100 bytes.
Average size change: -38 bytes.
BRANCH=none
BUG=none
TEST=make clobber buildall -j; make savesizes
[change some code that changes file sizes]
make clobber buildall -j ; make newsizes
[Shows size differences]
Change-Id: I48b440063eb6eb6c00900af3d0dfa075be6f9ec7
Signed-off-by: Martin Roth <martinroth@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/410860
Reviewed-by: Patrick Georgi <pgeorgi@chromium.org>
Currently eCTS suites share the same directory (e.g. build/stm32l476g
-eval) to put build artifacts even though some files (e.g. board.c)
compile differently suite to suite. So, if cts-i2c-stm32l476g-eval is
built, followed by cts-gpio-stm32l476g-eval, build fails or produces
incorrect binary.
This patch makes eCTS create different directories for each suite.
As a bonus, we can now builds eCTS suites in parallel.
BUG=chromium:654549
BRANCH=none
TEST=make buildall -j (with uncommitted change)
Change-Id: I4abedc917787be5f79b97e0e50d0d08e01bd5f9d
Signed-off-by: Daisuke Nojiri <dnojiri@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/398281
We don't really have the ability to build the latest signer yet, but
this should not stop us from being able to build a properly signed
image using the ec makefiles.
As a stopgap measure the suggestion is to keep the latest signer
binary in ~/bin/codesigner in chroot, then with this patch applied
invoking make with CR50_DEV=1 will cause the proper sighner used and
proper signing procedure followed.
The signed targets need to be built in series to avoid concurrent use
of the signer fob, an addition dependency is being added to enforce
that.
BRANCH=none
BUG=chrome-os-partner:55557
TEST=ran make as follows:
CR50_DEV=1 make BOARD=cr50
touched the fob when requested, uploaded the generated
build/cr50/ec.bin on a kevin-tpm2 using usb_updater, and observed
it boot properly with the new version.
Change-Id: Ia9494bdc60b4bd3b8e5e09cbcbd8b27409c739d2
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/376885
Tested-by: Bill Richardson <wfrichar@chromium.org>
Reviewed-by: Bill Richardson <wfrichar@chromium.org>
We need to (at least) build CTS for all boards & suites supported by CTS.
This will prevent our investment from accidentally being broken.
BUG=chromium:627252
BRANCH=none
TEST=make buildall builds CTS suites
Change-Id: Ib7bceaafd5e27ce9285b9d1bf35339bf6b04ba72
Signed-off-by: Daisuke Nojiri <dnojiri@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/359947
Add an explicit dependency on utils-build (includes ecst) to prevent
ec_elf_to_flat from running before ecst is built during parallel make.
BUG=chrome-os-partner:52777
BRANCH=None
TEST=Modify Makefile.rules to add 5sec sleep in $(build-utils) rule,
verify that `make BOARD=kevin clean && make BOARD=kevin -j` still
succeeds.
Signed-off-by: Shawn Nematbakhsh <shawnn@chromium.org>
Change-Id: Ia443f8b880cbd4451d41302817e176201f7870dc
Reviewed-on: https://chromium-review.googlesource.com/350162
Commit-Ready: Shawn N <shawnn@chromium.org>
Tested-by: Shawn N <shawnn@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
It is quite annoying to see an error reported by buildall, but not
knowing what board(s) actually failed to build.
Create file for each board being built, put build progress information
in it, and remove it if the board build succeeded. Then, once the
build is completed see the failed boards in .failedboards/, if any,
and the states they failed at.
BRANCH=none
BUG=none
TEST=verified that the directory is empty after a successful buildall
run, and has files named after failed boards in case of build
failures.
Change-Id: I67eb1671cadf58d9f8feccebfcc860524f33c2a0
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/338883
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Bill Richardson <wfrichar@chromium.org>
The new signer command line option should be used by some boards, not
all of them. Add it for cr50 for now.
BRANCH=none
BUG=none
TEST=verified that --cros is added to the command line when cr50 blobs
are signed, and that the signed blob boots fine on b1.
Change-Id: Ibc38cc6233b57838129ae7d554921c1bae0e6b8c
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/329595
Reviewed-by: Bill Richardson <wfrichar@chromium.org>
Creating bootloader for the chip involves signing the image with an
'air gap' - some physical presence action is required. We don't want
this to be required when the builder is building cr50 for test
purposes.
The solution is to keep using the dummy private key when building by
default, and invoking make differently when building an image which
would be accepted by the hardware.
Setting CR50_RO_KEY variable in the environment or in the make command
line will cause the signer use the value of this variable as the name
of the file containing the key to use for signing the RO image.
Should this file be a public key, the signer will stop and look for a
fob containing the matching private key, and will stream the RO image
through the fob for signing.
Using the fob requires that the signer runs under sudo, but we do not
want the generated files to belong to root, some more code is added to
change the generated files' ownership to user running the make.
BRANCH=none
BUG=chrome-os-partner:49950
TEST=ran the following tests:
- verified that the build still succeeds by default.
- invoked make as follows:
CR50_RO_KEY=cr50_rom0-dev-blsign.pem.pub make BOARD=cr50
observed the signer stop to wait for the user to interact with the
USB fob and proceed. Made sure that the generated image runs
successfully on the evaluation board.
- verified that 'make BOARD=cr50 clean' still works (i.e. none of the
generated files is owned by root).
Change-Id: I733ec6386c1dfd838d83d22fb589fa64e5eeaced
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/326484
Reviewed-by: Bill Richardson <wfrichar@chromium.org>
A typical EC image includes two similar in their functionality
subsections, RO and RW. CR50 has a small RO subsection, all it does -
detects a proper RW image to run and starts it up. To provide for
reliable firmware updates, the CR50 image needs to include two RW
sections, while the code is running from one RW subsection, the other
one can be upgraded.
This patch adds the ability to generate two identical RW sections,
mapped half flash size apart, and include them into the resulting EC
image.
To keep things simple the previously existing RW section's name is not
being changed, while the new (identical) RW section is named RW_B.
Two configuration options need to be defined to enable building of the
new image type: CONFIG_RW_B to enable the feature and
CONFIG_RW_B_MEM_OFF to define where RW_B should be mapped into the
flash.
A new rule added to Makefile.rules allows to generate a different lds
file from the same source (core/cortex-m/ec.lds.S) by defining a
compile time variable to pick a different base address for the
rewritable section, when RW_B is built.
BRANCH=none
BUG=chromium:43025
TEST=as follows:
- make buildall -j still succeeds
- verified that regular CR50 image starts successfully
- modified chip/g/loader/main.c to launch RW_B first, re-built and
re-run the image, observed on the console:
vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
cr50 bootloader, 20151118_11218@80881, no USB, full crypto
Valid image found at 0x00084000, jumping
--- UART initialized after reboot ---
[Reset cause: power-on]
[Image: unknown, cr50_v1.1.4160-4c8a789-dirty 2015-12-07 18:54:27 vbendeb@eskimo.mtv.corp.google.com]
[0.001148 Inits done]
This FPGA image has no USB support
Console is enabled; type HELP for help.
> [0.002212 task 2 waiting for events...]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(note that the image base address is 0x840000, which is RW_B).
Change-Id: Ia2f90d5e5b7a9f252ea3ecf3ff5babfad8a97444
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/316703
Commit-Ready: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-by: Aseda Aboagye <aaboagye@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
This commit introduces a new CONFIG_* option. To utilise the
experimental console, simply define the following config option in the
board's board.h file.
#define CONFIG_EXPERIMENTAL_CONSOLE
This is a temporary option which allows the EC console to be work with
the EC-3PO interactive console interface. When this option is enabled,
the EC expects commands to packed in a particular format. This is for
command integrity and allows the interpreter to perform automatic
command retrying in the event that a character is dropped from the sent
command.
It also removes a lot of the console editing methods since they are now
being served by EC-3PO.
Once the EC-3PO interpreter is pulled into servod, we can enable this
feature by default and the config option can go away.
BUG=chrome-os-partner:46054
BRANCH=None
TEST=make -j buildall tests
TEST=Enable CONFIG_EXPERIMENTAL_CONSOLE on GLaDOS; Flash EC and verify
that console works via the EC-3PO interactive console interface.
TEST=Build and flash on GLaDOS and verify normal console operation on
standard EC UART.
CQ-DEPEND=CL:308615
Change-Id: I5e66eb94e31299b27ce029b7f7ce6ba0a7fb6816
Signed-off-by: Aseda Aboagye <aaboagye@google.com>
Reviewed-on: https://chromium-review.googlesource.com/309991
Commit-Ready: Aseda Aboagye <aaboagye@chromium.org>
Tested-by: Aseda Aboagye <aaboagye@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
This patch completes introduction of building of proper RO and RW
images for cr50.
A few small mods were required:
- both RO and RW images have to be signed, using the same dedicated
signer, but with different keys, dev_key.pem is not needed any more.
- the RW image offset is not at the half of available flash, a chip
specific value of 16K is used instead.
The suggested new image layout is as follows:
+----------------------------------------+
| 1KB RO signature header. |
+----------------------------------------+
- -
| 15KB RO image. |
- -
+========================================+
| 1KB RW-A signature header. |
+----------------------------------------+
- -
- -
| 239K RW-A image. |
- -
- -
+========================================+
- -
| 16 KB NVRAM, shared |
- -
+========================================+
| 1KB RW-B signature header. |
+----------------------------------------+
- -
- -
| 239K RW-B image. |
- -
- -
+========================================+
BRANCH=none
BUG=chrome-os-partner:43025
TEST=The combined image (build/cr50/ec.hex) is successfully loaded and
started by the spiflash utility from the latest FPGA tarball.
Corrupting a byte in the generated image in the RW section causes
failure to verify.
Change-Id: I41a05168b0d4e9f88efa1003f261b6dd03972a24
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/311422
Reviewed-by: Bill Richardson <wfrichar@chromium.org>
When building with -j, it's easy to miss errors. If you don't
have your shell configured to warn you about nonzero exit, you
may not even notice that "make buildall -j" failed. To make it
more obvious, we'll do one level of recursion for that target.
That will ensure that the last line of output is always a
pass/fail message.
BUG=none
BRANCH=none
TEST=make buildall -j (with and without errors)
Before this CL, a successful "make buildall -j" ends with this:
MV ec.bin
OBJCOPY ec.hex
LD RW/ec.RW.elf
NM RO/ec.RO.smap
OBJCOPY RO/ec.RO.flat
NM RW/ec.RW.smap
OBJCOPY RW/ec.RW.flat
CAT ec.obj
OBJCOPY ec.bin
COPY_RW ec.bin
MV ec.bin
buildall completed successfully!
(cr) ((fa7baa6...)) ~/trunk/src/platform/ec $
while a failing one looks like this:
MV ec.bin
OBJCOPY ec.hex
LD RW/ec.RW.elf
NM RO/ec.RO.smap
OBJCOPY RO/ec.RO.flat
NM RW/ec.RW.smap
OBJCOPY RW/ec.RW.flat
CAT ec.obj
OBJCOPY ec.bin
COPY_RW ec.bin
MV ec.bin
(cr) ((fa7baa6...)) ~/trunk/src/platform/ec $
Did you see the difference? I suspect some people miss it.
With this CL, a failing "make buildall -j" looks like this:
NM RW/ec.RW.smap
OBJCOPY RW/ec.RW.flat
CAT ec.obj
OBJCOPY ec.bin
NM RW/chip/mec1322/lfw/ec_lfw-lfw.smap
OBJCOPY RW/chip/mec1322/lfw/ec_lfw-lfw.flat
COPY_RW ec.bin
MV ec.bin
COPY_RW ec.bin
MV ec.bin
OBJCOPY ec.bin
COPY_RW ec.bin
MV ec.bin
make[1]: Leaving directory '/mnt/host/source/src/platform/ec'
Makefile.rules:93: recipe for target 'buildall' failed
make: *** [buildall] Error 2
(cr) (stopit) ~/trunk/src/platform/ec $
Change-Id: Id9b47d2869f61e8e3e44b3c618399ca9223f0a71
Signed-off-by: Bill Richardson <wfrichar@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/303811
Reviewed-by: Aseda Aboagye <aaboagye@chromium.org>
This patch brings in the enhanced signer utility published along with
FPGA version 20150925_21715, and the image.cc file updated to fix the
bug where it was not picking up the initialized data segment from the
elf file.
The new signed image header format, among other things, describes
memory areas as read-only and read-execute, which allows the bootrom
to configure the MMU appropriately.
Makefiles had to be modified to reflect the fact that the signed image
now depends on .elf, not on .raw, and that building the signer
requires more source files. Note that some signer features are not yet
being utilized (like processing xml files describing fuses or
retrieving keys from gnubby), the source are kept for completeness.
BRANCH=none
BUG=chrome-os-partner:43791
TEST=build the cr50 image and boot in on the FPGA board using the
spiflash utility outside chroot. Observe the target starting the
console session.
Change-Id: Ib59b8ebbeb98a8146d4d997e1f78178c4fbc031a
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/303070
Commit-Ready: Marius Schilder <mschilder@chromium.org>
Reviewed-by: Marius Schilder <mschilder@chromium.org>
For signed EC RW images (CONFIG_RWSIG), there's no point in
embedding the public key or signature into the image itself since
it will just be replaced by the signer (either as the next step
in the build process, or after the fact for MP releases). This
takes that out and just points to where the pubkey and signature
will be placed.
BUG=none
BRANCH=none
TEST=make buildall
I also checked the signatures with
futility show -t build/*/ec.bin
They still look good, and the one signed image I booted (Cr50)
works as before.
Change-Id: Ib39b7c508914851f81a1bebb2450e08ef0def76c
Signed-off-by: Bill Richardson <wfrichar@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/302630
Reviewed-by: Randall Spangler <rspangler@chromium.org>
We have been using the time of compilation to determine the
version string. This CL will keep doing that if the git repo has
uncommitted changes, but if the repo is clean we'll just use the
author date of the last commit. This ensures that the same source
will produce bitwise-identical builds (assuming no toolchain
changes, of course).
BUG=chrome-os-partner:45616
BRANCH=none
TEST=manual
cd src/platform/ec
make buildall
mv build build.one
make buildall
md5sum build{,.one}/*/ec.bin | sort
Observe that successive builds produce identical binaries.
Change-Id: Ie2ef44b216586097589c9c15f12e05c87a53f991
Signed-off-by: Bill Richardson <wfrichar@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/302140
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Shawn N <shawnn@chromium.org>
The linker scripts are passed through the preprocessor so that
they can distinguish between RO and RW images. This change makes
the same macros available when compiling, so that code can
contain directives like
#ifdef SECTION_IS_RO
or
#define FW_SIZE_(section) CONFIG_##section##_SIZE
#define FW_SIZE(section) FW_SIZE_(section)
int size = FW_SIZE(SECTION);
BUG=none
BRANCH=none
TEST=make buildall
Nothing uses this yet, so there's no change to the images.
Change-Id: I6e03cd07c134f4b86aeddd9d516b74bbdb95b8a8
Signed-off-by: Bill Richardson <wfrichar@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/301255
Reviewed-by: Randall Spangler <rspangler@chromium.org>
This removes an unused vboot recipe left over from long, long ago.
It was added with commit 8101b71 (May 2012) to enable vboot on
EC. It should have been removed with commit 45cd846 (Aug 2012),
but wasn't. It hasn't been useful since then.
BUG=none
BRANCH=none
TEST=make buildall
Change-Id: I1176898064c4e714639dc63f041fb0fd85d73bab
Signed-off-by: Bill Richardson <wfrichar@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/301254
Reviewed-by: Randall Spangler <rspangler@chromium.org>
The signer utility needs to be built and the flat image needs to be
signed. The signer utility is written in C++, supporting this required
adding a new make command to Makefile.rules and a build file for the
utility.
The signing now needs to be a part of generating the .flat file. To
achieve this an alternative set of rules is defined in Makfile.rules
for targets where RO image needs to be signed.
Rules for converting elf to hex have been consolidated as there is no
need to omit the --set-section-flags when it does not apply.
BRANCH=none
BUG=chrome-os-partner:43025
TEST=as follows:
- ran 'rm build/cr50; make BOARD=cr50'
- observed that both build/cr50/ec.bin and
build/cr50/RO/ec.RO.flat have the required signature header in
the first 1024 bytes.
- verified that the cr50 board can be booted over SPI using the
image in build/cr50/RO/ec.RO.flat
Change-Id: Iacc22561de67fadfaf8e049bf9578cbd08cfad86
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/295291
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Aseda Aboagye <aaboagye@chromium.org>
The public sources look for board configurations in directories
named board/$BOARD/
Sometimes it's necessary to keep sensitive projects out of the
public view for a bit.
This CL allows board configurations to also appear in directories
named private*/board/$BOARD/
BUG=none
BRANCH=none
TEST=manual
First, ebuilds and "make buildall" seem to work just as before.
Second, I copied 24 of the existing boards (those without
board-specific #ifdefs in the code) into a private*/board/
directory, renamed them to something unique, and ran "make
buildall" again.
Both public and private boards compiled and passed their tests.
Change-Id: I977c23cb8e73e40677c8f329abca8bbc51fd53df
Signed-off-by: Bill Richardson <wfrichar@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/292428
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
The ec_date.h file had incorrect dependencies. $(objs) had no meaning
outside of the building the object files as it gets privately overidden
with the corresponding target objects (RO, RW, libsharedobjs). This
caused ec_date.h to only be generated once from a clean. This commit
fixes that by adding all of the RO and RW objects as dependencies (with
the exception of version.o).
BUG=chrome-os-partner:43373
BRANCH=None
TEST=Built ryu, checked build timestamp in build_info. Touched a file,
rebuilt, verified that build timestamp was updated.
TEST=make -j buildall tests
Change-Id: I0ab107efc1a504b4f871ebcf595754db1d414c7a
Signed-off-by: Aseda Aboagye <aaboagye@google.com>
Reviewed-on: https://chromium-review.googlesource.com/289338
Tested-by: Aseda Aboagye <aaboagye@chromium.org>
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
Commit-Queue: Aseda Aboagye <aaboagye@chromium.org>
Trybot-Ready: Aseda Aboagye <aaboagye@chromium.org>
Add ECST tool to modify the header used by npcx booter.
Modified drivers:
1. i2c.c: Modify for i2c_port design.
2. i2c.c: Fixed bugs when mutil-tasks use the same i2c port and pull-up issue.
3. hwtimer.c: Fixed bug whcih event expired time is behide current timer.
4. lpc.c: Add intializing host settings after pltrst is deasserted.
5. uart.c/clock.c/register.h: Fixed bug which cannot enter deep-idle
when gpio is any-edge trigger mode.
6. task.c: Add workaround method for hard fault issue.
7. keyboard_raw.c: Modified for support CONFIG_KEYBOARD_KSO_BASE
8. lpc.c: Modified for support CONFIG_KEYBOARD_IRQ_GPIO
9. lpc.c: fixed obe interrupt bug during 8042 initialization
10.Adjust path of flat files for new Makefile rules
11.Fixed build error on lpc.c without CONFIG_KEYBOARD_IRQ_GPIO
BUG=chrome-os-partner:34346
TEST=make buildall -j; test nuvoton IC specific drivers
BRANCH=none
Change-Id: Icf9494174b245b4026e396be877d578f36b6f6a5
Signed-off-by: Ian Chao <mlchao@nuvoton.com>
Reviewed-on: https://chromium-review.googlesource.com/284036
Reviewed-by: Shawn N <shawnn@chromium.org>
Tested-by: Shawn N <shawnn@chromium.org>
Commit-Queue: Shawn N <shawnn@chromium.org>
This commit introduces the build infrastructure changes needed for
creating a shared RO library. (libsharedobjs). The end goal is for the
library to contain various objects that can be shared with both the RO
and RW EC images.
Now, there are 3 make goals: ro, rw, and libsharedobjs.
In order for changes that are only specific to a single image (ie: RW
only) to be applied correctly, the object files are now built separately
for the RO, RW, shared objects library targets.
NOTE: Certain EC targets are incompatible with this model due to the
fact that only one image is present within flash at a time.
BRANCH=none
BUG=None
TEST=make -j buildall tests
TEST=make -j BOARD=cr50 xrefs
TEST=make BOARD=samus dis
TEST=Built samus EC image and compared that the final EC image was
identical to the upstream version (except for the git SHAs & version
strings).
CQ-DEPEND=CL:285934
Change-Id: I8e67f089710be9c6d7017718109262394bdad2f5
Signed-off-by: Aseda Aboagye <aaboagye@google.com>
Reviewed-on: https://chromium-review.googlesource.com/274079
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Aseda Aboagye <aaboagye@chromium.org>
Tested-by: Aseda Aboagye <aaboagye@chromium.org>
Commit-Queue: Aseda Aboagye <aaboagye@chromium.org>
Add npcx_evb_arm board-level driver for arm-based platform.
Add header.c: for booting from NPCX5M5G A3 Booter.
Remove lfw folder due to those functionalitie have been replaced with Booter
Modified drivers for
Patch Set 1:
1. flash.c: Implement UMA lock, tri-state and selection register lock functionalities
2. hwtimer.c: Add ITIM32 for hwtimer
3. lpc.c: Add checking for LRESET
4. system.c: Modified CODERAM_ARCH functions for NPCX5M5G A3 Booter.
5. uart.c: Add support for module 2
Patch Set 2:
6. lpc.c: Modified lpc_get_pltrst_asserted() func
Patch Set 3:
7. minimize the changes for CONFIG_CODERAM_ARCH in common layer
8. comments of Patch Set1/2
Patch Set 4:
9. Modified CONFIG_RO_MEM_OFF point to ro image and keep header as a part of ec.RO.flat.
10. Fixed RO_FRID and RW_FRID issues which caused by CONFIG_CODERAM_ARCH.
Patch Set 5:
11. Modified system.c in common folder for supporting *_STORAGE_OFF.
12. Use *_STORAGE_OFF in firmware_image.lds.S to indicate flat file layout in flash.
Patch Set 6:
13. rebase to newest version
14. system.c: Modified for the newest include/system.h
Patch Set 7:
15. Merge from version 0625
BUG=chrome-os-partner:34346
TEST=make buildall -j; test nuvoton IC specific drivers
BRANCH=none
Change-Id: Ifd7c10b81b5781ccd75bb2558dc236486976e8ed
Signed-off-by: Ian Chao <mlchao@nuvoton.com>
Reviewed-on: https://chromium-review.googlesource.com/272034
Reviewed-by: Shawn N <shawnn@chromium.org>
Tested-by: Shawn N <shawnn@chromium.org>
Commit-Queue: Shawn N <shawnn@chromium.org>
The flash_ec rule calls the flash_ec tool, passing the inferred board
and image based on where make was run. This means that from a board
directory you can run "make flash_ec" and an up to date image will be
flashed using flash_ec.
Signed-off-by: Anton Staaf <robotboy@chromium.org>
BRANCH=None
BUG=None
TEST=make buildall -j
cd board/ryu_p4p5
make flash_ec
Change-Id: I51e2a62f4d0de427f8d36e0848941aef742e0d3d
Reviewed-on: https://chromium-review.googlesource.com/272264
Tested-by: Anton Staaf <robotboy@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Commit-Queue: Anton Staaf <robotboy@chromium.org>
Trybot-Ready: Anton Staaf <robotboy@chromium.org>
Sorted symbol maps are useful when one wants to look up an address to
associate it with a function or a variable, or to estimate how much
room certain components of the image take.
This patch adds a rule and creates a dependency to make sure that
sorted maps (assigned extension .smap) are generated during builds.
BRANCH=none
BUG=none
TEST=make buildall -j
observed 91 .elf files and 89 .smap files generated.
the two images not triggering .smap file generation are
build/npcx_evb/chip/npcx/lfw/ec_lfw
build/npcx_evb/chip/npcx/spiflashfw/ec_npcxflash
they could be added in a follow up patch if deemed necessary.
Change-Id: I196a2ffe059a83481b7a313617d516634762dc60
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/270117
Reviewed-by: Bill Richardson <wfrichar@chromium.org>
Previously for the mec1322 chip an ec.bin file was created in the normal way
and then it was "packed" in a post-processing stage to produce ec.spi.bin.
This change allows a chip or board build.mk file to specify the rules used to
produce ec.bin, and uses this for the mec1322 to do the packing. This means
that we can use the standard "ec.bin" name, and do not need to alter other
scripts, such as the script which creates chromeos-firmwareupdate.
BUG=None
TEST=buildall -j, flash on strago and see it still works.
BRANCH=NONE
Change-Id: I3f880d64e60d14f82cb1d21c8b3f2d4ae5e0dfef
Signed-off-by: Icarus Sparry <icarus.w.sparry@intel.com>
Reviewed-on: https://chromium-review.googlesource.com/265544
Tested-by: Kevin K Wong <kevin.k.wong@intel.com>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Commit-Queue: Kevin K Wong <kevin.k.wong@intel.com>
