Extend the local APIC timer delay so that it can be started,
and waited for, independently.
Add an EOI so that more than one APIC timer interrupt is possible.
Previous to this, because there was no EOI, the first timer
interrupt the CPU took was also the last it would take --
apic_delay would only work one time.
Change-Id: Ib11aeee5b7da81287166ac68fc327e7ae62d1b84
Signed-off-by: Ronald G Minnich <rminnich@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/43323
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: David Hendricks <david.hendricks@gmail.com>
Reviewed-by: Philipp Deppenwiese <zaolin.daisuki@gmail.com>
No other architecture in libpayload outputs anything in the main entry
routine. Let alone an exception test which looks like a real exception
to the normal user and is most likely really misleading. Silence the
startup code.
Change-Id: I6e49f24ad46ce578a4bb111c2d623ca4470a1866
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/43126
Reviewed-by: Julius Werner <jwerner@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
There is no bfd "arm64". The correct bfdname is "aarch64". Fix it. With
this change libpayload will build with the AArch64 GCC.
Change-Id: If7a6b14691107c5d4fc67c3cd3990ecc849d4af1
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/43125
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Stefan thinks they don't add value.
Command used:
sed -i -e '/file is part of /d' $(git grep "file is part of " |egrep ":( */\*.*\*/\$|#|;#|-- | *\* )" | cut -d: -f1 |grep -v crossgcc |grep -v gcov | grep -v /elf.h |grep -v nvramtool)
The exceptions are for:
- crossgcc (patch file)
- gcov (imported from gcc)
- elf.h (imported from GNU's libc)
- nvramtool (more complicated header)
The removed lines are:
- fmt.Fprintln(f, "/* This file is part of the coreboot project. */")
-# This file is part of a set of unofficial pre-commit hooks available
-/* This file is part of coreboot */
-# This file is part of msrtool.
-/* This file is part of msrtool. */
- * This file is part of ncurses, designed to be appended after curses.h.in
-/* This file is part of pgtblgen. */
- * This file is part of the coreboot project.
- /* This file is part of the coreboot project. */
-# This file is part of the coreboot project.
-# This file is part of the coreboot project.
-## This file is part of the coreboot project.
--- This file is part of the coreboot project.
-/* This file is part of the coreboot project */
-/* This file is part of the coreboot project. */
-;## This file is part of the coreboot project.
-# This file is part of the coreboot project. It originated in the
- * This file is part of the coreinfo project.
-## This file is part of the coreinfo project.
- * This file is part of the depthcharge project.
-/* This file is part of the depthcharge project. */
-/* This file is part of the ectool project. */
- * This file is part of the GNU C Library.
- * This file is part of the libpayload project.
-## This file is part of the libpayload project.
-/* This file is part of the Linux kernel. */
-## This file is part of the superiotool project.
-/* This file is part of the superiotool project */
-/* This file is part of uio_usbdebug */
Change-Id: I82d872b3b337388c93d5f5bf704e9ee9e53ab3a9
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/41194
Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This patch makes libpayload enable the instruction cache as the very
first thing, which is similar to how we treat it in coreboot. It also
prevents the icache from being disabled again during mmu_disable() as
part of the two-stage page table setup in post_sysinfo_scan_mmu_setup().
It replaces the existing mmu_disable() implementation with the assembly
version from coreboot which handles certain edge cases better (see
CB:27238 for details).
The SCTLR flag definitions in libpayload seem to have still been
copy&pasted from arm32, so replace with the actual arm64 defintions from
coreboot.
Change-Id: Ifdbec34f0875ecc69fedcbea5c20e943379a3d2d
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/38908
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
According to the POSIX standard, %p is supposed to print a pointer "as
if by %#x", meaning the "0x" prefix should automatically be prepended.
All other implementations out there (glibc, Linux, even libpayload) do
this, so we should make coreboot match. This patch changes vtxprintf()
accordingly and removes any explicit instances of "0x%p" from existing
format strings.
How to handle zero padding is less clear: the official POSIX definition
above technically says there should be no automatic zero padding, but in
practice most other implementations seem to do it and I assume most
programmers would prefer it. The way chosen here is to always zero-pad
to 32 bits, even on a 64-bit system. The rationale for this is that even
on 64-bit systems, coreboot always avoids using any memory above 4GB for
itself, so in practice all pointers should fit in that range and padding
everything to 64 bits would just hurt readability. Padding it this way
also helps pointers that do exceed 4GB (e.g. prints from MMU config on
some arm64 systems) stand out better from the others.
Change-Id: I0171b52f7288abb40e3fc3c8b874aee14b9bdcd6
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37626
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: David Guckian
To avoid trampling over interesting exception artifacts on the real
stack, our arm64 systems switch to a separate exception stack when
entering an exception handler. We don't want that to use up too much
SRAM so we just set it to 512 bytes. I mean it just prints a bunch of
registers, how much stack could it need, right?
Quite a bit it turns out. The whole vtxprintf() call stack goes pretty
deep, and aarch64 generally seems to be very generous with stack space.
Just the varargs handling seems to require 128 bytes for some reason,
and the other stuff adds up too. In the end the current implementation
takes 1008 bytes, so bump the exception stack size to 2K to make sure it
fits.
Change-Id: I910be4c5f6b29fae35eb53929c733a1bd4585377
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37464
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
The MIPS architecture port has been added 5+ years ago in order to
support a Chrome OS project that ended up going nowhere. No other board
has used it since and nobody is still willing or has the expertise and
hardware to maintain it. We have decided that it has become too much of
a mainenance burden and the chance of anyone ever reviving it seems too
slim at this point. This patch eliminates all MIPS code and
MIPS-specific hacks.
Change-Id: I5e49451cd055bbab0a15dcae5f53e0172e6e2ebe
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/34919
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
There's a recurring pattern of reading cbtable entries that point into
cbmem entries. Move that pattern into its own function.
Coccinelle patch used for this:
@@
identifier T, T2;
expression TARGET;
@@
-struct cb_cbmem_tab *const T2 = (struct cb_cbmem_tab *)T;
-TARGET = phys_to_virt(T2->cbmem_tab);
+TARGET = get_cbmem_ptr(T);
Change-Id: I7bd4a7ad8baeeaebf0fa7d4b4de6dbc719bc781f
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/35756
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
One of many steps to compile with -Wconversion, as unsigned int and int
aren't the same thing.
BUG=b:111443775
BRANCH=none
TEST=make junit.xml shows fewer warnings with -Wconversion enabled
Change-Id: I9673ca70da32a1e5117b27fa89167e03379af9c1
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/32183
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
This patch is a raw application of
find payloads/ -type f | \
xargs sed -i -e 's/IS_ENABLED\s*(CONFIG_/CONFIG(/g'
Change-Id: I883b03b189f59b5d998a09a2596b0391a2d5cf33
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/31775
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
This patch adds remote GDB support for the arm64 architecture.
Change-Id: I2fa4dbca6c39f822f489a5e81bd052f53fda98a5
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/29020
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
The arm32 GDB architecture code contains a little hack that allows it to
(sort of) correctly deal with a reentrant exception triggered from
within the GDB stub. The main logic for this isn't really arm32 specific
and could be useful for other architectures as well, so factor it out
into a separate function.
Change-Id: I3c6db8cecf1e86bba23de6fd2ac9fdf0cf69d3c6
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/29019
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch reworks the arm64 exception handling to be more similar to
how it works on arm32. This includes a bunch of features like actually
saving and restoring more exception state in the exception_state
structure and supporting the same sort of partial reentrancy that is
useful for GDB. Since there's no instruction to directly load into or
store out of SP on arm64, we can't do quite the same thing where we use
that to read an exception_state_ptr variable right after exception entry
when no other register is available. But we can do something very
similar by (ab-)using the "high" stack pointer (SP_EL2) as a pointer to
the exception_state struct and providing a function to change it.
Change-Id: Ia16a1124be1824392a309ae1f4cb031547d184c1
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/29018
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch adds the new, faster architectural register accessors to
libpayload that were already added to coreboot in CB:27881. It also
hardcodes the assumption that coreboot payloads run at EL2, which has
already been hardcoded in coreboot with CB:27880 (see rationale there).
This means we can drop all the read_current/write_current stuff which
added a lot of unnecessary helpers to check the current exception level.
This patch breaks payloads that used read_current/write_current
accessors, but it seems unlikely that many payloads deal with this stuff
anyway, and it should be a trivial fix (just replace them with the
respective _el2 versions).
Also add accessors for a couple of more registers that are required to
enable debug mode while I'm here.
Change-Id: Ic9dfa48411f3805747613f03611f8a134a51cc46
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/29017
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Patrick Rudolph <patrick.rudolph@9elements.com>
This will make enabling the APIC safer by ignoring unknown interrupts
and not halting the system. Once all interrupt sources have been found
and handled DIE_ON_UNKNOWN_INTERRUPT can be set if desired.
BUG=b:116777191
TEST=Booted grunt, halted the kernel, and pushed the power button while
in S5. Verified that depthcharge logged the unknown exception.
APIC Init Started
APIC Configured
Ignoring interrupt vector 39
Change-Id: If4ed566ec284d69786c369f37e4e331d7f892c74
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-on: https://review.coreboot.org/28882
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Not used by x86 code anymore.
BUG=b:116777191
TEST=Validated that depthcharge can be built.
Change-Id: I25ad3903989a5433ce73d657cfdb93dd1f34f7b5
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-on: https://review.coreboot.org/28881
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
We should have a spurious interrupt vector just incase we get one. The
handler doesn't need to do anything.
BUG=b:116777191
TEST=Booted depthcharge on grunt and inspected the register. I can't
generate a spurious interrupt, so I can't validate that the handler gets
called.
Change-Id: I9e49e617f4375eb5eb00d0715c1902f77e2bf284
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-on: https://review.coreboot.org/28880
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Only set end of interrupt (EOI) when the APIC In-Service vector matches
the interrupt vector. This makes it so we don't EOI a non APIC
interrupt.
BUG=b:116777191
TEST=Booted grunt with APIC enabled and verified depthcharge still
works.
Change-Id: I00bd1e7a0fcf2fc004feadc40d22ebfefe68b384
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-on: https://review.coreboot.org/28879
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
As per internal discussion, there's no "ChromiumOS Authors" that's
meaningful outside the Chromium OS project, so change everything to the
contemporary "Google LLC."
While at it, also ensure consistency in the LLC variants (exactly one
trailing period).
"Google Inc" does not need to be touched, so leave them alone.
Change-Id: Ia0780e31cdab879d2aaef62a2f0403e3db0a4ac8
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/28756
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Joel Kitching <kitching@google.com>
This reduces power consumption on grunt by over 3W when sitting at the
depthcharge recovery screen.
BUG=b:109749762
TEST=Booted grunt in the recovery screen and made sure it continued to
work.
Change-Id: Id079c099ee4cf6a07724241af4400063f4551668
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-on: https://review.coreboot.org/28245
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Reviewed-by: Julius Werner <jwerner@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This method has a pause instruction to help the CPU relax a little bit.
Measuring grunt it saves about 80mW.
BUG=b:109749762
TEST=Made sure that grunt boots.
Change-Id: I045a941ed42fcc4f2dbdd65b5cbb42d84813f50c
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-on: https://review.coreboot.org/28244
Reviewed-by: Martin Roth <martinroth@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The apic_delay method will halt the CPU and wait for a timer interrupt
to fire. I went with usec because nsec is too granular to guarantee.
This method will be called from an arch_ndelay() method when the delay
is large enough to justify a sleep.
BUG=b:109749762
TEST=Tested it on grunt by changing the _delay method to call
apic_delay().
Change-Id: I80363f06bdb22d0907f895885e607fde1c4c468d
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-on: https://review.coreboot.org/28242
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
This is just the bare minimum required to initialize the APIC. I only
support xAPIC and chose not to support x2APIC. We can add that
functionality later when it's required.
I also made the exception dispatcher call apic_eoi so that the callbacks
won't forget to call it.
BUG=b:109749762
TEST=Booted grunt and verified that depthcharge continued to function
and that linux booted correctly. Also verified GDB still works.
Change-Id: I420a4eadae84df088525e727b481089ef615183f
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-on: https://review.coreboot.org/28241
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
BUG=b:109749762
TEST=Verified GDB still functions by hitting Ctrl+G on the developer
screen and stepping through some code.
Change-Id: I723a8a95f681c500d9d8e35e49fd1d893cb1f133
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-on: https://review.coreboot.org/28240
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
I need to setup the APIC timer to fire interrupts. I would like to reuse
the existing interrupt table. So I extended it to support user defined
interrupts. I just added all 255 vectors so there wouldn't need to be
any additional build time configuration.
I'm going to deprecate exception_install_hook and remove it in a follow
up. It will be replaced with set_interrupt_handler. This way the
exception lookup does not have to manage a list of callbacks, or have to
worry about the order they are processed.
BUG=b:109749762
TEST=Wrote an interrupt handler and fired an APIC timer interrupt and
verified that vector 32 was returned.
Change-Id: Id9c2583c7c3d9be4a06a25e546e64399f2b0620c
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-on: https://review.coreboot.org/28100
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Will be used by the APIC.
BUG=b:109749762
TEST=Verified by the other cls in the stack.
Change-Id: Id86f2719d98a90318ac625e09601e5dbb06e3765
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-on: https://review.coreboot.org/28239
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Fixes build with gcc8.1
Change-Id: I042f79ddfb4c249e00b5b259280289b8534f6854
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-on: https://review.coreboot.org/27546
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Add: raw_read_cntfrq_el0() and raw_read_cntpct_el0()
Required to support Arch64 Timer
Change-Id: I86aa97039304b9e9336d0146febfe1811c9e075a
Signed-off-by: T Michael Turney <mturney@codeaurora.org>
Reviewed-on: https://review.coreboot.org/25649
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
The VA space needs to be extended to support 48bit, as on Cavium SoCs
the MMIO starts at 1 << 47.
The following changes were done to coreboot and libpayload:
* Use page table lvl 0
* Increase VA bits to 48
* Enable 256TB in MMU controller
* Add additional asserts
Tested on Cavium SoC and two ARM64 Chromebooks.
Change-Id: I89e6a4809b6b725c3945bad7fce82b0dfee7c262
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Reviewed-on: https://review.coreboot.org/24970
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
The word 'coreboot' should always be written in lowercase, even at the
start of a sentence.
Change-Id: I2ec18ca55e0ea672343a951ab81a24a5630f45fd
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/20028
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Philippe Mathieu-Daudé <philippe.mathieu.daude@gmail.com>
Reviewed-by: Philipp Deppenwiese <zaolin.daisuki@gmail.com>
The coreboot sites support HTTPS, and requests over HTTP with SSL are
also redirected. So use the more secure URLs, which also saves a
request most of the times, as nothing needs to be redirected.
Run the command below to replace all occurences.
```
$ git grep -l -E 'http://(www.|review.|)coreboot.org'
| xargs sed -i 's,http://\(.*\)coreboot.org,https://\1coreboot.org,g'
```
Change-Id: If53f8b66f1ac72fb1a38fa392b26eade9963c369
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-on: https://review.coreboot.org/20034
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
coreboot and libpayload currently use completely different code to
perform a full cache flush on ARM64, with even different function names.
The libpayload code is closely inspired by the ARM32 version, so for the
sake of overall consistency let's sync coreboot to that. Also align a
few other cache management details to work the same way as the
corresponding ARM32 parts (such as only flushing but not invalidating
the data cache after loading a new stage, which may have a small
performance benefit).
Change-Id: I9e05b425eeeaa27a447b37f98c0928fed3f74340
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/19785
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Allows to use SSE and floating point in payloads without digging to
much into x86 assembly code.
Tested on Lenovo T500 (Intel Core2Duo).
Both floating point operation and SSE is properly working.
Change-Id: I4a5fc633f158de421b70435a8bfdc0dcaa504c72
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/18345
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Some simple implementation of the MultiBoot protocol may not pass a
memory map (MULTIBOOT_FLAGS_MMAP missing in the flags) but just the two
values for low and high memory, indicated by the MULTIBOOT_FLAGS_MEMINFO
flag.
Support those kind of boot loaders too, instead of falling back to the
hard-coded values in lib_get_sysinfo().
Tested with a multiboot enhanced version of FILO.
Change-Id: I22cf9e3ec0075aff040390bd177c5cd22d439b81
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Reviewed-on: https://review.coreboot.org/18350
Tested-by: build bot (Jenkins)
Reviewed-by: Nico Huber <nico.h@gmx.de>
Implement the argc/argv passing as described in coreboot’s payload API:
http://www.coreboot.org/Payload_API
While at it, give the code some love by not needlessly trashing register
values.
Change-Id: Ib830f2c67b631b7216843203cefd55d9bb780d83
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Reviewed-on: https://review.coreboot.org/18336
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
Simplify the code by directly using the arguments on the stack as base
pointer relative memory references, instead of loading them into
intermediate registers first.
Make it more robust by preserving all callee saved registers mandated by
the C calling convention (and only those), namely EBP, EBX, ESI and EDI.
Don't assume anything about the register state when the called function
returns -- beside the segment registers and the stack pointer to be
still the same as before the call.
Change-Id: I383d6ccefc5b3d5cca37a1c9b638c231bbc48aa8
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Reviewed-on: https://review.coreboot.org/18335
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
According to coreboot’s payload API [1], the called payload should be
able to return a value via %eax. Support this by changing the prototype
of start_main() and pass on the return value of main() to the caller
instead of discarding it.
[1] https://www.coreboot.org/Payload_API
Change-Id: I8442faea19cc8e04487092f8e61aa4e5cba3ba76
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Reviewed-on: https://review.coreboot.org/18334
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
According to coreboot’s payload API [1] the argc value should be passed
at stack offset 0x10, so we need to push a dummy value to comply to the
API.
[1] https://www.coreboot.org/Payload_API
Change-Id: Id20424185a5bf7e4d94de1886a2cece3f3968371
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Reviewed-on: https://review.coreboot.org/18333
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
The pointer to write the return value to is in %ecx, not %eax. Writing
to (%eax) leads to memory corruptions as %eax holds the return value,
e.g. would write zero to address zero for a "successful" returning
payload.
Change-Id: I82df27ae89a9e3d25f479ebdda2b50ea57565459
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Reviewed-on: https://review.coreboot.org/18332
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
According to coreboot’s payload API [1] the magic value passed to the
payload should be 0x12345678, not 12345678. Fix that.
[1] https://www.coreboot.org/Payload_API
Change-Id: I10a7f7b1a4aec100416c5e7e4ba7f8add10ef5c5
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Reviewed-on: https://review.coreboot.org/18331
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
3 out of 4 architectures currently zero out the payload BSS in early
assembly code, which is pointless since the code loading the payload has
already done that (with a more efficient memset). ARM64 has never had
any code like this and can run just fine without it. This also defeats
the new optimization of moving the heap out of the BSS, since all three
implementations assume that everything between _edata and _end is BSS.
We should just take this out.
Change-Id: I45cd2dabd94da43ff0f77e990f11c877cee6cda1
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/16091
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The ARM64 MMU code maintains a list of used ranges, to avoid mapping the
DMA buffer over the coreboot tables and things like that. Unfortunately,
the overlap with ranges in that list is checked with
(start1 >= start2 && start1 <= end2) || (end1 >= start2 && end1 <= end2)
which is not a full overlap check and misses the case where the second
region is completely contained within the first. This patch replaces
that code with a properly vetted primitive from Stack Overflow.
BRANCH=none
BUG=chrome-os-partner:54416
TEST=Observe how Kevin recovery screen now gets drawn at 10x the speed.
Change-Id: I7e2706426762794e160d743bbfc40da1e26eee12
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/16075
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Some exceptions (like from calling a NULL function pointer) are easier
to narrow down with a dump of the call stack. Let's take a page out of
ARM32's book and add that feature to ARM64 as well. Also change the
output format to two register columns, to make it easier to fit a whole
exception dump on one screen.
Applying to both coreboot and libpayload and syncing the output format
between both back up.
Change-Id: I19768d13d8fa8adb84f0edda2af12f20508eb2db
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/14931
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
If a framebuffer is already configured by coreboot, libpayload's
MMU tables didn't mark its memory DMAable (unlike when libpayload
set up its own framebuffer memory).
BRANCH=none
BUG=chrome-os-partner:52826
TEST=depthcharge's recovery screen is not corrupted anymore on kevin
Change-Id: I228a861b3fdcf1298a3cfa0a054214c78ed55e70
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 889e8358a0f2f504abd9910549aa68f3992bb4e8
Original-Change-Id: I7ba79151ccc1eb605f82e1869a74b539a6be5e99
Original-Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Original-Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/341092
Reviewed-on: https://review.coreboot.org/14685
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
This allows to accommodate different platforms' default
configurations, memory configuration is fine tuned later during boot
process.
BUG=chrome-os-partner:51537
BRANCH=none
TEST=none yet, the full stack of patches boots fine on EVB
Change-Id: I39da4ce247422f67451711ac0ed5a5e1119ed836
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 97a9a71ade4df8a501043f9ae58463a3135e2a4f
Original-Change-Id: I39da4ce247422f67451711ac0ed5a5e1119ed836
Original-Signed-off-by: huang lin <hl@rock-chips.com>
Original-Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/332384
Original-Reviewed-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/13914
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This will allow more payloads to use the standard linker script
instead of implementing their own.
Change-Id: Ie60120769829f427ceb722109d85859b61dbde31
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: https://review.coreboot.org/14074
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Ronald G Minnich