Masked ROMs are the silent killers of boot speed on devices without
memory-mapped SPI flash. They often contain awfully slow SPI drivers
(presumably bit-banged) that take hundreds of milliseconds to load our
bootblock, and every extra kilobyte of bootblock size has a hugely
disproportionate impact on boot speed. The coreboot timestamps can never
show that component, but it impacts our users all the same.
This patch tries to alleviate that issue a bit by allowing us to
compress the bootblock with LZ4, which can cut its size down to nearly
half. Of course, masked ROMs usually don't come with decompression
algorithms built in, so we need to introduce a little decompression stub
that can decompress the rest of the bootblock. This is done by creating
a new "decompressor" stage which runs before the bootblock, but includes
the compressed bootblock code in its data section. It needs to be as
small as possible to get a real benefit from this approach, which means
no device drivers, no console output, no exception handling, etc.
Besides the decompression algorithm itself we only include the timer
driver so that we can measure the boot speed impact of decompression. On
ARM and ARM64 systems, we also need to give SoC code a chance to
initialize the MMU, since running decompression without MMU is
prohibitively slow on these architectures.
This feature is implemented for ARM and ARM64 architectures for now,
although most of it is architecture-independent and it should be
relatively simple to port to other platforms where a masked ROM loads
the bootblock into SRAM. It is also supposed to be a clean starting
point from which later optimizations can hopefully cut down the
decompression stub size (currently ~4K on RK3399) a bit more.
NOTE: Bootblock compression is not for everyone. Possible side effects
include trying to run LZ4 on CPUs that come out of reset extremely
underclocked or enabling this too early in SoC bring-up and getting
frustrated trying to find issues in an undebuggable environment. Ask
your SoC vendor if bootblock compression is right for you.
Change-Id: I0dc1cad9ae7508892e477739e743cd1afb5945e8
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/26340
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch attempts to finish the separation between CONFIG_VBOOT and
CONFIG_CHROMEOS by moving the remaining options and code (including
image generation code for things like FWID and GBB flags, which are
intrinsic to vboot itself) from src/vendorcode/google/chromeos to
src/vboot. Also taking this opportunity to namespace all VBOOT Kconfig
options, and clean up menuconfig visibility for them (i.e. some options
were visible even though they were tied to the hardware while others
were invisible even though it might make sense to change them).
CQ-DEPEND=CL:459088
Change-Id: I3e2e31150ebf5a96b6fe507ebeb53a41ecf88122
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/18984
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
No board has ever tried to combine CONFIG_SEPARATE_VERSTAGE with
CONFIG_VBOOT_STARTS_IN_ROMSTAGE. There are probably many reasons why
this wouldn't work (e.g. x86 CAR migration logic currently always
assumes verstage code to run pre-migration). It would also not really
make sense: the reason we use separate verstages is to decrease
bootblock size (mitigating the boot speed cost of slow boot ROM SPI
drivers) and to allow the SRAM-saving RETURN_FROM_VERSTAGE trick,
neither of which would apply to the after-romstage case. It is better to
just forbid that case explicitly and give programmers more guarantees
about what the verstage is (e.g. now the assumption that it runs pre-RAM
is always valid).
Since Kconfig dependencies aren't always guaranteed in the face of
'select' statements, also add some explicit compile-time assertions to
the vboot code. We can simplify some of the loader logic which now no
longer needs to provide for the forbidden case. In addition, also try to
make some of the loader logic more readable by writing it in a more
functional style that allows us to put more assertions about which cases
should be unreachable in there, which will hopefully make it more robust
and fail-fast with future changes (e.g. addition of new stages).
Change-Id: Iaf60040af4eff711d9b80ee0e5950ce05958b3aa
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/18983
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
The timestamp code asserts that the _timestamp region (allocated in
memlayout for pre-RAM stages) is large enough for the assumptions it
makes. This is good, except that we often initialize timestamps
extremely early in the bootblock, even before console output. Debugging
a BUG() that hits before console_init() is no fun.
This patch adds a link-time assertion for the size of the _timestamp
region in memlayout to prevent people from accidentally running into
this issue.
Change-Id: Ibe4301fb89c47fde28e883fd11647d6b62a66fb0
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/16270
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Certain chipsets don't have a memory-mapped boot media
so their code execution for stages prior to DRAM initialization
is backed by SRAM or cache-as-ram. The postcar stage/phase
handles the cache-as-ram situation where in order to tear down
cache-as-ram one needs to be executing out of a backing
store that isn't transient. By current definition, cache-as-ram
is volatile and tearing it down leads to its contents disappearing.
Therefore provide a shim layer, postcar, that's loaded into
memory and executed which does 2 things:
1. Tears down cache-as-ram with a chipset helper function.
2. Loads and runs ramstage.
Because those 2 things are executed out of ram there's no issue
of the code's backing store while executing the code that
tears down cache-as-ram. The current implementation makes no
assumption regarding cacheability of the DRAM itself. If the
chipset code wishes to cache DRAM for loading of the postcar
stage/phase then it's also up to the chipset to handle any
coherency issues pertaining to cache-as-ram destruction.
Change-Id: Ia58efdadd0b48f20cfe7de2f49ab462306c3a19b
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/14140
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Stages are inconsistent with other memlayout regions in that they don't
have _<name> and _e<name> symbols defined. We have _program and
_eprogram, but that always only refers to the current stage and
_eprogram marks the actual end of the executable's memory footprint, not
the end of the area allocated in memlayout. Both of these are sometimes
useful to know, so let's add another set of symbols that allow the stage
areas to be treated more similarly to other regions.
Change-Id: I9e8cff46bb15b51c71a87bd11affb37610aa7df9
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/13737
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Instead of tagging object files with .<class>, move them to a <class>
directory below $(obj)/. This way we can keep a 1:1 mapping between
source- and object-file names.
The 1:1 mapping is a prerequisite for Ada, where the compiler refuses
any other object-file name.
Tested by verifying that the resulting coreboot.rom files didn't change
for all of Jenkins' abuild configurations.
Change-Id: Idb7a8abec4ea0a37021d9fc24cc8583c4d3bf67c
Signed-off-by: Nico Huber <nico.h@gmx.de>
Reviewed-on: https://review.coreboot.org/13181
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
commit a8aef3ac (cbfs_spi: Initialize spi_flash when initializing
cbfs_cache) introduced a bug that makes the rarely-used unified
CBFS_CACHE() memlayout macro break when used in conjunction with
cbfs_spi.c (since that macro does not define a separate
postram_cbfs_cache region). This patch fixes the problem by making all
three region names always available for both the unified and split
macros in every stage (and adds code to ensure we don't reinitialize
the same buffer again in romstage, which might be a bad idea if
previous mappings are still in use).
BRANCH=None
BUG=None
TEST=Compiled for both kinds of macros, manually checked symbols in
disassembled stages.
Change-Id: I114933e93080c8eceab04bfdba3aabf0f75f8ef9
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 0f270f88e54b42afb8b5057b0773644c4ef357ef
Original-Change-Id: If172d9fa3d1fe587aa449bd4de7b5ca87d0f4915
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/318834
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/12933
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Most devices do not use SPI before they initialize CBMEM. This change
initializes spi_flash in the CBMEM_INIT_HOOK to initialize the postram
cbfs cache so it is not overwritten when boot_device_init is called
later.
BUG=chromium:210230
BRANCH=none
TEST=confirm that the first cbfs access can occur before RAM initialized
and after on panther and jerry.
Change-Id: If3b6efc04082190e81c3773c0d3ce116bb12421f
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 0ab242786a16eba7fb423694f6b266e27d7660ec
Original-Change-Id: I5f884b473e51e6813fdd726bba06b56baf3841b0
Original-Signed-off-by: Mary Ruthven <mruthven@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/314311
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/12601
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Currently the CBFS mmap cannot be accessed at the beginning of romstage
because it waits until DRAM is initialized. This change first loads CBFS
into SRAM and then switches to using DRAM as the backing once it is
initialized.
BUG=chromium:210230
BRANCH=none
TEST=confirm that the cbfs can be access at the beginning and end of
romstage on different boards.
Change-Id: I9fdaef392349c27ba1c19d4cd07e8ee0ac92dddc
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: ccaaba266386c7d5cc62de63bdca81a0cc7c4d83
Original-Change-Id: Idabfab99765b52069755e1d1aa61bbee39501796
Original-Signed-off-by: Mary Ruthven <mruthven@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/312577
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/12586
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
It encourages users from writing to the FSF without giving an address.
Linux also prefers to drop that and their checkpatch.pl (that we
imported) looks out for that.
This is the result of util/scripts/no-fsf-addresses.sh with no further
editing.
Change-Id: Ie96faea295fe001911d77dbc51e9a6789558fbd6
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: http://review.coreboot.org/11888
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Bring rmodule linking into the common linking method.
The __rmodule_entry symbol was removed while using
a more common _start symbol. The rmodtool will honor
the entry point found within the ELF header. Add
ENV_RMODULE so that one can distinguish the environment
when generating linker scripts for rmodules. Lastly,
directly use program.ld for the rmodule.ld linker script.
BUG=chrome-os-partner:44827
BRANCH=None
TEST=Built rambi and analyzed the relocatable ramstage,
sipi_vector, and smm rmodules.
Change-Id: Iaa499eb229d8171272add9ee6d27cff75e7534ac
Signed-off-by: Aaron Durbin <adubin@chromium.org>
Reviewed-on: http://review.coreboot.org/11517
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
There's no reason to have a separate verstage.ld now
that there is a unified stage linking strategy. Moreover
verstage support is throughout the code base as it is
so bring in those link script macros into the common
memlayout.h as that removes one more specific thing a
board/chipset needs to do in order to turn on verstage.
BUG=chrome-os-partner:44827
BRANCH=None
TEST=None
Change-Id: I1195e06e06c1f81a758f68a026167689c19589dd
Signed-off-by: Aaron Durbin <adubin@chromium.org>
Reviewed-on: http://review.coreboot.org/11516
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Instead of having separate <stage>.ld files in src/lib
one file can be used: program.ld. There's now only one
touch point for stage layout.
BUG=chrome-os-partner:44827
BRANCH=None
TEST=Built a myriad of boards. Analyzed readelf output.
Change-Id: I4c3e3671d696caa2c7601065a85fab803e86f971
Signed-off-by: Aaron Durbin <adubin@chromium.org>
Reviewed-on: http://review.coreboot.org/11509
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Though coreboot started as x86 only, the current approach to x86
linking is out of the norm with respect to other architectures.
To start alleviating that the way ramstage is linked is partially
unified. A new file, program.ld, was added to provide a common way
to link stages by deferring to per-stage architectural overrides.
The previous ramstage.ld is no longer required.
Note that this change doesn't handle RELOCATABLE_RAMSTAGE
because that is handled by rmodule.ld. Future convergence
can be achieved, but for the time being that's being left out.
BUG=chrome-os-partner:44827
BRANCH=None
TEST=Built a myriad of boards.
Change-Id: I5d689bfa7e0e9aff3a148178515ef241b5f70661
Signed-off-by: Aaron Durbin <adubin@chromium.org>
Reviewed-on: http://review.coreboot.org/11507
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Julius Werner <jwerner@chromium.org>
In order to accommodate tracking timestamps in all the
__PRE_RAM__ stages (bootblock, verstage, romstage, etc)
of a platform one needs to provide a way to specify
a persistent region of SRAM or cache-as-ram to store
the timestamps until cbmem comes online. Provide that
infrastructure.
Based on original patches from chromium.org:
Original-Change-Id: I4d78653c0595523eeeb02115423e7fecceea5e1e
Original-Signed-off-by: Furquan Shaikh <furquan@google.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/223348
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Tested-by: Furquan Shaikh <furquan@chromium.org>
Original-Commit-Queue: Furquan Shaikh <furquan@chromium.org>
Original-Change-Id: Ie5ffda3112d626068bd1904afcc5a09bc4916d16
Original-Signed-off-by: Furquan Shaikh <furquan@google.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/224024
Original-Reviewed-by: Furquan Shaikh <furquan@chromium.org>
Original-Commit-Queue: Furquan Shaikh <furquan@chromium.org>
Original-Tested-by: Furquan Shaikh <furquan@chromium.org>
Change-Id: I8779526136e89ae61a6f177ce5c74a6530469ae1
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/10790
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
A new CBFS API is introduced to allow making CBFS access
easier for providing multiple CBFS sources. That is achieved
by decoupling the cbfs source from a CBFS file. A CBFS
source is described by a descriptor. It contains the necessary
properties for walking a CBFS to locate a file. The CBFS
file is then decoupled from the CBFS descriptor in that it's
no longer needed to access the contents of the file.
All of this is accomplished using the regions infrastructure
by repsenting CBFS sources and files as region_devices. Because
region_devices can be chained together forming subregions this
allows one to decouple a CBFS source from a file. This also allows
one to provide CBFS files that came from other sources for
payload and/or stage loading.
The program loading takes advantage of those very properties
by allowing multiple sources for locating a program. Because of
this we can reduce the overhead of loading programs because
it's all done in the common code paths. Only locating the
program is per source.
Change-Id: I339b84fce95f03d1dbb63a0f54a26be5eb07f7c8
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/9134
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
As per discussion with lawyers[tm], it's not a good idea to
shorten the license header too much - not for legal reasons
but because there are tools that look for them, and giving
them a standard pattern simplifies things.
However, we got confirmation that we don't have to update
every file ever added to coreboot whenever the FSF gets a
new lease, but can drop the address instead.
util/kconfig is excluded because that's imported code that
we may want to synchronize every now and then.
$ find * -type f -exec sed -i "s:Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, *MA[, ]*02110-1301[, ]*USA:Foundation, Inc.:" {} +
$ find * -type f -exec sed -i "s:Foundation, Inc., 51 Franklin Street, Suite 500, Boston, MA 02110-1335, USA:Foundation, Inc.:" {} +
$ find * -type f -exec sed -i "s:Foundation, Inc., 59 Temple Place[-, ]*Suite 330, Boston, MA *02111-1307[, ]*USA:Foundation, Inc.:" {} +
$ find * -type f -exec sed -i "s:Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.:Foundation, Inc.:" {} +
$ find * -type f
-a \! -name \*.patch \
-a \! -name \*_shipped \
-a \! -name LICENSE_GPL \
-a \! -name LGPL.txt \
-a \! -name COPYING \
-a \! -name DISCLAIMER \
-exec sed -i "/Foundation, Inc./ N;s:Foundation, Inc.* USA\.* *:Foundation, Inc. :;s:Foundation, Inc. $:Foundation, Inc.:" {} +
Change-Id: Icc968a5a5f3a5df8d32b940f9cdb35350654bef9
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: http://review.coreboot.org/9233
Tested-by: build bot (Jenkins)
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
This patch introduces a new option (CONFIG_MULTIPLE_CBFS_INSTANCES) to
allow multiple CBFS instances in the bootrom.
When the new option is enabled, the code running on the target
controls which CBFS instance is used. Since all other then header CBFS
structures use relative addressing, the only value which needs
explicit setting is the offset of the CBFS header in the bootrom.
This patch adds a facility to set the CBFS header offset. The offset
value of zero means default. i.e. the CBFS initialization code still
discovers the offset through the value saved at the top of the ROM.
BRANCH=storm
BUG=chrome-os-partner:34161, chromium:445938
TEST=with the rest patches in, storm target successfully boots from RW
section A.
Change-Id: Id8333c9373e61597f0c653c727dcee4ef6a58cd2
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: e57a3a15bba7cdcca4a5d684ed78f8ac6dbbc95e
Original-Change-Id: I4c026389ec4fbaa19bd11b2160202282d2f9283c
Original-Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/237569
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/9747
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Edward O'Callaghan <edward.ocallaghan@koparo.com>
This patch creates a new mechanism to define the static memory layout
(primarily in SRAM) for a given board, superseding the brittle mass of
Kconfigs that we were using before. The core part is a memlayout.ld file
in the mainboard directory (although boards are expected to just include
the SoC default in most cases), which is the primary linker script for
all stages (though not rmodules for now). It uses preprocessor macros
from <memlayout.h> to form a different valid linker script for all
stages while looking like a declarative, boilerplate-free map of memory
addresses to the programmer. Linker asserts will automatically guarantee
that the defined regions cannot overlap. Stages are defined with a
maximum size that will be enforced by the linker. The file serves to
both define and document the memory layout, so that the documentation
cannot go missing or out of date.
The mechanism is implemented for all boards in the ARM, ARM64 and MIPS
architectures, and should be extended onto all systems using SRAM in the
future. The CAR/XIP environment on x86 has very different requirements
and the layout is generally not as static, so it will stay like it is
and be unaffected by this patch (save for aligning some symbol names for
consistency and sharing the new common ramstage linker script include).
BUG=None
TEST=Booted normally and in recovery mode, checked suspend/resume and
the CBMEM console on Falco, Blaze (both normal and vboot2), Pinky and
Pit. Compiled Ryu, Storm and Urara, manually compared the disassemblies
with ToT and looked for red flags.
Change-Id: Ifd2276417f2036cbe9c056f17e42f051bcd20e81
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: f1e2028e7ebceeb2d71ff366150a37564595e614
Original-Change-Id: I005506add4e8fcdb74db6d5e6cb2d4cb1bd3cda5
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/213370
Reviewed-on: http://review.coreboot.org/9283
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Tauner <stefan.tauner@gmx.at>
Reviewed-by: Aaron Durbin <adurbin@google.com>
Kyösti Mälkki