This enum is currently anonymous. Add a name so it can be used in the
code.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Instead of a special function, send an event after driver model is inited
and adjust the boards which use this function.
Signed-off-by: Simon Glass <sjg@chromium.org>
These symbols are incorrect, meaning that binman cannot find the
associated entry. This leads to errors like:
binman: Section '/binman/simple-bin': Symbol '_binman_spl_prop_size'
in entry '/binman/simple-bin/u-boot-spl/u-boot-spl-nodtb':
Entry 'spl' not found in list (mkimage,u-boot-spl-nodtb,
u-boot-spl-bss-pad,u-boot-spl-dtb,u-boot-spl,u-boot-img,main-section)
Fix it.
Signed-off-by: Simon Glass <sjg@chromium.org>
MCFG tables are used on multiple arches. Move to common ACPI lib.
Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Moritz Fischer <moritzf@google.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Use sizeof(*mcfg) instead of sizeof(*header)
Signed-off-by: Simon Glass <sjg@chromium.org>
Commit d953137526 ("x86: Move SSDT table to a writer function")
introduced a bug where the actual MCFG entries are no longer generated.
Cc: Simon Glass <sjg@chromium.org>
Fixes: d953137526 ("x86: Move SSDT table to a writer function")
Signed-off-by: Moritz Fischer <moritzf@google.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
When building for a custom board, it is quite common to maintain a
private branch which include some defconfig and .dts files. But to
hook up those .dts files requires modifying a file "belonging" to
upstream U-Boot, the arch/*/dts/Makefile. Forward-porting that branch
to a newer upstream then often results in a conflict which, while it
is trivial to resolve by hand, makes it harder to have a CI do "try to
build our board against latest upstream".
The .config usually includes information on precisely what .dtb(s) are
needed, so to avoid having to modify the Makefile, simply add the
files in (SPL_)OF_LIST to dtb-y.
A technicality is that (SPL_)OF_LIST is not always defined, so rework
the Kconfig symbols so that (SPL_)OF_LIST is always defined (when
(SPL_)OF_CONTROL), but only prompted for in the cases which used to be
their "depends on".
nios2 and microblaze already have something like this in their
dts/Makefile, and the rationale in commit 41f59f6853 is similar to
the above. So this simply generalizes existing practice. Followup
patches could remove the logic in those two makefiles, just as there's
potential for moving some common boilerplate from all the
arch/*/dts/Makefile files to the new scripts/Makefile.dts.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Reviewed-by: Simon Glass <sjg@chromium.org>
Move this over to use a writer file, moving the code from the x86
implementation.
There is no need to store a separate variable since we can simply access
the ACPI context.
With this, the original monolithic x86 function for writing ACPI tables
is gone.
Note that QEMU has its own implementation.
Signed-off-by: Simon Glass <sjg@chromium.org>
Update this function to the newer style, so we can avoid passing and
returning an address through this function.
Also move this function out of the x86 code so it can be used by other
archs.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Each board has its own way of creating this table. Rather than calling the
acpi_create_fadt() function for each one from a common acpi_write_fadt()
function, just move the writer into the board-specific code.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Move this table over to use a writer function, moving the code from the
x86 implementation.
Add a pointer to the DSDT in struct acpi_ctx so we can reference it later.
Disable this table for sandbox since we don't actually compile real ASL
code.
Signed-off-by: Simon Glass <sjg@chromium.org>
Move this table over to use a writer function, moving the code from the
x86 implementation.
Add a pointer to the DSDT in struct acpi_ctx so we can reference it later.
Signed-off-by: Simon Glass <sjg@chromium.org>
Use the new ACPI writer to write the base tables at the start of the area,
moving this code from the x86 implementation.
Signed-off-by: Simon Glass <sjg@chromium.org>
Use the new ACPI writer to write the ACPI tables. At present this is all
done in one monolithic function. Future work will split this out.
Unfortunately the QFW write_acpi_tables() function conflicts with the
'writer' version, so disable that for sandbox.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present acpi_setup_base_tables() both sets up the ACPI context and
writes out the base tables.
We want to use an ACPI writer to write the base tables, so split this
function into two, with acpi_setup_ctx() doing the context set, and
acpi_setup_base_tables() just doing the base tables.
Disable the writer's write_acpi_tables() function for now, to avoid
build errors. It is enabled in a following patch.
Signed-off-by: Simon Glass <sjg@chromium.org>
Rather than keying everything off ACPIGEN, use the main
GENERATE_ACPI_TABLE option to determine whether the core ACPI code
is included. Make sure these option are not enabled in SPL/TPL since we
never generate tables there.
Signed-off-by: Simon Glass <sjg@chromium.org>
Allow this to be used on any arch. Also convert to using macros so that
we can check the CONFIG option in C code.
Signed-off-by: Simon Glass <sjg@chromium.org>
These have sadly found their way to ARM now. Allow any arch to support
generating ACPI tables.
Disable this for the tools build.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present some 32-bit settings are used with the 64-bit app. Fix this by
separating out the two cases.
Be careful not to break the 64-bit payload, which needs to build a 64-bit
EFI stub with a 32-bit U-Boot.
Signed-off-by: Christian Melki <christian.melki@t2data.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
That script is not intended for use with EFI, so update the logic to avoid
using it.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Christian Melki <christian.melki@t2data.com>
Add an empty CPU init function to avoid fiddling with low-level CPU
features in the app. Set up the C runtime correctly for 64-bit use
and avoid clearing BSS, since this is done by EFI when U-Boot is loaded.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present this function requires a pointer to struct efi_entry_memmap
but the only field used in there is the desc_size. We want to be able
to use it from the app, so update it to use desc_size directly.
Signed-off-by: Simon Glass <sjg@chromium.org>
x86 platform uses standard format of Config Address for PCI Configuration
Mechanism #1. So use new U-Boot macro PCI_CONF1_ADDRESS().
Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Show the revision of this table as it can be important.
Also update the 'efi table' entry to show the actual address of the EFI
table rather than our table that points to it. This saves a step and the
intermediate table has nothing else in it.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
At present only 4KB of spare space is left in the DTB when building the
EFI app. Increase this to 32KB so there is plenty of space to insert the
binman definition. This cannot be expanded later (as with OF_SEPARATE)
because the ELF image has already been built.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviwed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
If the 'bootm' command is not enabled then this code is not available and
this causes a link error. Fix it.
Note that for the EFI app, there is no indication of missing code. It just
hangs!
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
At present this is disabled, but it should work so long as the kernel does
not need EFI services. Enable it and add a note about remaining work.
Signed-off-by: Simon Glass <sjg@chromium.org>
The current EFI video driver only works when running in the stub. In that
case the stub calls boot services (before jumping to U-Boot proper) and
copies the graphics info over to the efi table. This is necessary because
the stub exits boot services before jumping to U-Boot.
The app maintains access to boot services throughout its life, so does not
need to do this. Update the driver to support calling boot services
directly.
Enable video output for the app. Note that this uses the
EFI_GRAPHICS_OUTPUT_PROTOCOL protocol, even though it mentions vesa.
A sample qemu command-line for this case is:
qemu-system-x86_64 -bios /usr/share/edk2.git/ovmf-ia32/OVMF-pure-efi.fd
-drive id=disk,file=try.img,if=none,format=raw -nic none
-device ahci,id=ahci -device ide-hd,drive=disk,bus=ahci.0
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This variable is already defined by the EFI code. Drop the duplicate
definition when building a 64-bit EFI app.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Most modern platforms use 64-bit EFI so it is useful to have a U-Boot app
that runs under that. Add a (non-functional) build for this.
Note that --whole-archive causes the gcc 9.2 linker to crash, so disable
this for now. Once this is resolved, things should work.
For now, avoid mentioning the documentation for the 64-bit app, since it
does not work.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Most EFI implementations use 64-bit but U-Boot only supports running as
a 32-bit app at present. While efi-x86_payload64 does boot from 64-bit
UEFI it immediately changes back to 32-bit before starting U-Boot.
In order to support a 64-bit U-Boot app, update the Kconfig to add an
option for 32/64 bit. Update the prompt for the existing option so it is
clear it relates to the stub. Move both up to just under the choice that
controls them, since this looks better and the menu.
Use CONFIG_EFI_APP in the Makefile instead of CONFIG_TARGET_EFI_APP,
since the latter is specific to a single target and we will have two.
Memory size is set to 32MB for now so that it can run on qemu without
increasing the default memory size. We may need to increase the default
later.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Move error message to the caller of mrfld_pinconfig*() in order
to unify them in the future.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Move is_protected assignment closer to its user.
This increases readability and makes maintenance easier.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
There are two PCB designs in the wild which use the opposite
signaling for SD card detection. This makes U-Boot working
in one case and failing in the other. Quirk this out by
disconnecting SD card detection pin from the PCB by switching
it to mode 3. In the disconnected state the read value is always
the same and inverted to what we are expecting in the code.
BugLink: https://github.com/edison-fw/meta-intel-edison/issues/136
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
We would need to quirk out the Card Detect case and for that we allow
configuring the SD/SDIO family of pins.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
When building U-Boot with clang, it notices that the i8254.h include
guard does not work correctly due to a typo. Fix it.
Signed-off-by: Alistair Delva <adelva@google.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
[bmeng: fixed the other same typo at the end of the same file]
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Fix a warning seen when compiling this dts file.
Signed-off-by: Alistair Delva <adelva@google.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
There is no point to use Method() for the constant.
Replace it with Name() defined object. For the _STA
case it saves 3 bytes per each entry.
Before: 2881
After: 2833
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Allow usage of the bootstage facilities in SPL.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Simon Glass <sjg@chromium.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
It is useful to see some basic EFI info with the command as it forms part
of the information about a board.
Add a hook for this and show the table address as a start.
While here, fix an invalid cast in setup_efi_info(). Note that this
function is using a data structure defined by Linux so we cannot change
it. Also note that ulong is used since this is the standard in U-Boot
(>6k uses), despite there being quite a bit of the more verbose uintptr_t
(930 uses).
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The setup routines are called from zimage but don't really belong in the
zimage header. Add a new EFI header to house these. Add comments so it is
clear what the functions do.
Note that these functions are x86-specific. The zimage business is not
used on other architectures.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
At present this information is stripped when linking. It is useful to keep
it around. Strip it from the .efi files instead.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
-----BEGIN PGP SIGNATURE-----
iQFFBAABCgAvFiEEslwAIq+Gp8wWVbYnfxc6PpAIreYFAmFR2fMRHHNqZ0BjaHJv
bWl1bS5vcmcACgkQfxc6PpAIreaAnAf/ZKPy3CRphbgAmbLhNQhQGpLosLcFp343
cv53gfsv1DakLRrhqMSnOQSVoTg9P22uHbooDDR9WHfZyEL6tvGZ+dgN2/wlDhDj
wbvvsM7FB+uvLVfMF3MURQBvy4dPO+FyCm/uHmry2V2uqWPw4ss+aVhxeQ5s4/oI
LpS6yaiM8UAj2GTUZSqJ1+/W7YG9N6tL73n0PDyCvSrdCkyPWqpww1rtST88SfFI
bBEWc6ksna/29haOwHQ2xopEbVxtHv2byULfqUtnsbtE4Ozw8o0fx66tdwPHTQlM
7eSuRiEbqjlJuUZFIbMat7nb7inZkaWupZ/Sr1HUUiDdKMygrRsQhg==
=x8Uf
-----END PGP SIGNATURE-----
Merge tag 'dm-pull-next-27sep21' of https://source.denx.de/u-boot/custodians/u-boot-dm into next
Various of-platdata improvements, including CONFIG_OF_REAL
Convert some of these occurences to C code, where it is easy to do. This
should help encourage this approach to be used in new code.
Signed-off-by: Simon Glass <sjg@chromium.org>
Now that we have a 'positive' Kconfig option, use this instead of the
negative one, which is harder to understand.
Signed-off-by: Simon Glass <sjg@chromium.org>
The current API is outdated as it requires a devicetree pointer.
Move these functions to use the ofnode API and update this globally. Add
some tests while we are here.
Correct the call in exynos_dsim_config_parse_dt() which is obviously
wrong.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add arch_lmb_reserve() implemented using arch_lmb_reserve_generic().
It is rather likely this architecture also needs to cover U-Boot with
LMB before booting Linux.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Cc: Tom Rini <trini@konsulko.com>
On Intel Tangier the SDHCI #2 provides SD card connection.
Add GPIO card detection for it.
Fixes: 39665beed6 ("x86: tangier: Enable ACPI support for Intel Tangier")
BugLink: https://github.com/edison-fw/meta-intel-edison/issues/135
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Bin Meng <bmeng.cn@gmail.com>
eMMC is non-removable on Intel Edison board. Fix the DTS accordingly.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
U_BOOT_DATE and U_BOOT_TIME are updated on every run of make command.
Therefore mrc.c file is recompiled every time when running make which means
that whole U-Boot binary is recompiled on every run of make command.
Simplify it and do not recompile U-Boot binary on every run of make command
by not depending on macros U_BOOT_DATE and U_BOOT_TIME.
Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
default n/no doesn't need to be specified. It is default option anyway.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
[trini: Rework FSP_USE_UPD portion]
Signed-off-by: Tom Rini <trini@konsulko.com>
We move the SYS_CACHE_SHIFT_N options from arch/arm/Kconfig to
arch/Kconfig, and introduce SYS_CACHE_SHIFT_4 to provide a size of 16.
Introduce select statements for other architectures based on current
usage. For MIPS, we take the existing arch-specific symbol and migrate
to the generic symbol. This lets us remove a little bit of otherwise
unused code.
Cc: Alexey Brodkin <alexey.brodkin@synopsys.com>
Cc: Anup Patel <anup.patel@wdc.com>
Cc: Atish Patra <atish.patra@wdc.com>
Cc: Bin Meng <bmeng.cn@gmail.com>
Cc: Leo <ycliang@andestech.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Sean Anderson <seanga2@gmail.com>
Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Acked-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Reviewed-by: Rick Chen <rick@andestech.com>
- Fixed broken ICH SPI driver in software sequencer mode
- Added "m25p,fast-read" to SPI flash node for x86 boards
- Drop ROM_NEEDS_BLOBS and BUILD_ROM for x86 ROM builds
- Define a default TSC timer frequency for all x86 boards
- x86 MTRR MSR programming codes bug fixes
- x86 "hob" command bug fixes
- Don't program MTRR for DRAM for FSP1
- Move INIT_PHASE_END_FIRMWARE to FSP2
- Use external graphics card by default on Intel Crown Bay
- tangier: Fix DMA controller IRQ polarity in CSRT
It is a pain to have to specify the value 16 in each call. Add a new
hextoul() function and update the code to use it.
Add a proper comment to simple_strtoul() while we are here.
Signed-off-by: Simon Glass <sjg@chromium.org>
Initialize 'igd' and 'sdvo' to NULL so that we just need to test
them against NULL later, to be compatible with that case that IGD
and SDVO devices were already in disabled state.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
For FSP1, there is no such INIT_PHASE_END_FIRMWARE.
Move board_final_cleanup() to fsp2 directory.
Fixes: 7c73cea442 ("x86: Notify the FSP of the 'end firmware' event")
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested on chromebook_coral, chromebook_samus, chromebook_link, minnowmax
Tested-by: Simon Glass <sjg@chromium.org>
There are several outstanding issues as to why this does not apply
to FSP1:
* For FSP1, the system memory and reserved memory used by FSP are
already programmed in the MTRR by FSP.
* The 'mtrr_top' mistakenly includes TSEG memory range that has the
same RES_MEM_RESERVED resource type. Its address is programmed
and reported by FSP to be near the top of 4 GiB space, which is
not what we want for SDRAM.
* The call to mtrr_add_request() is not guaranteed to have its size
to be exactly the power of 2. This causes reserved bits of the
IA32_MTRR_PHYSMASK register to be written which generates #GP.
For FSP2, it seems this is necessary as without this, U-Boot boot
process on Chromebook Coral goes very slowly.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested on chromebook_coral, chromebook_samus, chromebook_link, minnowmax
Tested-by: Simon Glass <sjg@chromium.org>
The size parameter of mtrr_add_request() and mtrr_set_next_var()
shall be power of 2, otherwise the logic creates a mask that does
not meet the requirement of IA32_MTRR_PHYSMASK register.
Programming such a mask value to IA32_MTRR_PHYSMASK generates #GP.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested on chromebook_coral, chromebook_samus, chromebook_link, minnowmax
Tested-by: Simon Glass <sjg@chromium.org>
At present mtrr_commit() programs the MTRR MSRs starting from
index 0, which may overwrite MSRs that were already programmed
by previous boot stage or FSP.
Switch to call mtrr_set_next_var() instead.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested on chromebook_coral, chromebook_samus, chromebook_link, minnowmax
Tested-by: Simon Glass <sjg@chromium.org>
Current mtrr_commit() logic assumes that MTRR MSRs are programmed
consecutively from index 0 to its maximum number, and whenever it
detects an unused one, it clears all other MTRRs starting from that
one. However this may not always be the case.
In fact, the clear is not much helpful because these MTRRs come out
of reset as disabled already. Drop the clear codes.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested on chromebook_coral, chromebook_samus, chromebook_link, minnowmax
Tested-by: Simon Glass <sjg@chromium.org>
If for some reason, TSC timer frequency cannot be determined from
hardware, nor is it specified in the device tree, U-Boot will panic
resulting in endless reset during boot.
Let's define a default TSC timer frequency using the Kconfig value
CONFIG_X86_TSC_TIMER_FREQ (note: #include must be used instead of
/include/ otherwise the macro is not pre-processed).
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Currently there are two places to specify the x86 TSC timer frequency
with one in Kconfig used for early timer and the other one in device
tree used when the frequency cannot be determined from hardware.
This may potentially create an inconsistent config where the 2 values
do not match. Let's use the one specified in Kconfig in the device
tree as well.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
These 2 options are no longer needed as now binman is used to build
u-boot.rom.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
IRQ polarity in CSRT has the same definition as by ACPI specification
chapter 19.6.64 "Interrupt (Interrupt Resource Descriptor Macro)", i.e.
ActiveHigh is 0, and ActiveLow is 1. On Intel Tangier the DMA controller
IRQ polarity is ActiveHigh.
Note, in DSDT (see southcluster.asl) it's described correctly.
Fixes: 5e99fde34a ("x86: tangier: Populate CSRT for shared DMA controller")
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Except ICH7 SPI, all SPI flashes connected to ICH9 / Fast SPI should
have "m25p,fast-read" property present in their DT nodes.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
SMBIOS tables only support 32bit addresses. If we don't have memory here
handle the error gracefully:
* on x86_64 fail to start U-Boot
* during UEFI booting ignore the missing table
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
This is a revert of a recent logic change in setup_zimage(). We do
actually need to install this information always. Change it to install
from the Coreboot tables if available, else the normal source.
Fixes: e7bae8283f ("x86: Allow installing an e820 when booting from coreboot")
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
All the x86 devicetree files are built at once, whichever board is
actually being built. If coreboot is the target build, CONFIG_ROM_SIZE
is not defined and samus cannot build Chromium OS verified boot. Add
this condition to avoid errors about CONFIG_ROM_SIZE being missing.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Use VENDOR_COREBOOT instead of TARGET_COREBOOT so we can have multiple
coreboot boards, sharing options. Only SYS_CONFIG_NAME needs to be
defined TARGET_COREBOOT.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Add a function comment for get_coreboot_info() and a declaration for
cb_get_sysinfo(), since this may be called from elsewhere.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
A recent change to disable cache setup when booting from coreboot
assumed that this has been done by SPL. The result is that for the
coreboot board, the cache is disabled (in start.S) and never
re-enabled.
If the cache was turned off, as it is on boards without SPL, we should
turn it back on. Add this new condition.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
These constants conflict with error codes returned by the MP
implementation when something is wrong. In particular, mp_first_cpu()
returns MP_SELECT_BSP when running without multiprocessing enabled.
Since this is -2, it is interpreted as an error by callers, which
expect a positive CPU number for the first CPU.
Correct this by using a different range for the pre-defined CPU
numbers, above zero and out of the range of possible CPU values. For
now it is safe to assume there are no more than 64K CPUs.
This fixes the 'mtrr' command when CONFIG_SMP is not enabled.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
When starting U-Boot from a previous-stage bootloader we presumably don't
need to set up the variable MTRRs. In fact this could be harmful if the
existing settings are not what U-Boot uses.
Skip that step in this case.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
With x86 we can execute an option ROM either natively or using the
x86 emulator (if enabled with CONFIG_BIOSEMU). Both of these share
the _X86EMU_env variable, with the native code using it to hold
register state during interrupt processing.
At present, in 32-bit U-Boot, the variable is declared twice, once
in common code and once in code only compiled with CONFIG_BIOSEMU.
With GCC 11 this causes a 'multiple definitions' error on boards
with CONFIG_BIOSEMU.
Drop the emulator definition when CONFIG_BIOSEMU is used.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
When we do not have bootstage enabled, rather than include an empty
dummy function, we just don't reference it. This saves us space in some
tight builds. This also shows a few cases where show_boot_progress was
incorrectly guarded before.
Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
When switching to kernel.org x86_64 gcc 11.1.0 toolchain, u-boot.rom
built from qemu-x86_defconfig no longer boots anymore. Investigation
shows that U-Boot fails at a very early stage during the boot process,
in fdtdec_prepare_fdt() where fdt_check_header() complains that there
is not a valid device tree found at gd->fdt_blob which points to _end.
Now _end points to an allocated section .note.gnu.property which of
course is wrong.
This issue is however not seen when using the default Ubuntu 20.04 gnu
toolchain (gcc 9.3.0 with binutils 2.34). Further investigation shows
that it is caused by a behavior change of binutils v2.36 which is part
of the kernel.org gcc 11.1.0 toolchain, via the following commit:
939b95c77bf2 ("Linux/x86: Configure gas with --enable-x86-used-note by default")
In fact, there was already a regression bug report [1] for binutils two
months ago, but the binutils folks did not think it is a bug :(
To resolve this, there are several options:
* pass -Wa,-mx86-used-note=no to gas
* pass -R .note.gnu.property to objcopy
* discard the section in the linker script
Linux kernel uses the discard way [2], so let's do the same for U-Boot.
[1] https://sourceware.org/bugzilla/show_bug.cgi?id=27753
[2] commit 4caffe6a28d3 ("x86/vdso: Discard .note.gnu.property sections in vDSO")
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
This commit does the same thing as Linux commit 33def8498fdd.
Use a more generic form for __section that requires quotes to avoid
complications with clang and gcc differences.
Remove the quote operator # from compiler_attributes.h __section macro.
Convert all unquoted __section(foo) uses to quoted __section("foo").
Also convert __attribute__((section("foo"))) uses to __section("foo")
even if the __attribute__ has multiple list entry forms.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Add support for stack protector for UBOOT, SPL, and TPL
as well as new pytest for stackprotector
Signed-off-by: Joel Peshkin <joel.peshkin@broadcom.com>
Adjust UEFI build flags.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The current usage of the variable CFLAGS_NON_EFI on the x86 architecture
deviates from other architectures.
Variable CFLAGS_NON_EFI is the list of compiler flags to be removed when
building UEFI applications. It is not a list of flags to be added anywhere.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Fix copy/paste errors in the descriptions of mtrr_close () and mtrr_set().
Signed-off-by: Wolfgang Wallner <wolfgang.wallner@br-automation.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
We move qfw into its own uclass and split the PIO functions into a
specific driver for that uclass. The PIO driver is selected in the
qemu-x86 board config (this covers x86 and x86_64).
include/qfw.h is cleaned up and documentation added.
Signed-off-by: Asherah Connor <ashe@kivikakk.ee>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Provide the model information through sysinfo so that it shows up on
boot. For memconfig 4 pins are provided, for 16 combinations. For SKU
ID there are two options:
- two pins provided in a ternary arrangement, for 9 combinations.
- reading from the EC
Add a binding doc and drop the unused #defines as well.
Example:
U-Boot 2021.01-rc5
CPU: Intel(R) Celeron(R) CPU N3450 @ 1.10GHz
DRAM: 3.9 GiB
MMC: sdmmc@1b,0: 1, emmc@1c,0: 2
Video: 1024x768x32 @ b0000000
Model: Google Coral (memconfig 5, SKU 3)
This depends on the GPIO series:
http://patchwork.ozlabs.org/project/uboot/list/?series=228126
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Bin Meng <bmeng.cn@gmail.com>
The _SUPPORT suffix is from an earlier time and interferes with use of
the CONFIG_IS_ENABLED() macro. Rename the option to drop the suffix.
Tidy up the TODO that prompted this.
Signed-off-by: Simon Glass <sjg@chromium.org>
When booting from coreboot there is no need to notify the FSP of anything,
since coreboot has already done it. Nor it is possible, since the FSP
details are not provided by coreboot.
Skip it in this case.
Signed-off-by: Simon Glass <sjg@chromium.org>
When booting from coreboot the FSP video information is no-longer
available. Enable the coreboot driver so that we can get some sort of
display in this case.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present the debug UART is only set up in SPL, on the assumption that
the boot flow will always pass through there. When booting from coreboot,
SPL is not used, so the debug UART is not available.
Move the code into a common place so that it can be used in U-Boot proper
also. Add the required init to start_from_spl.S as well.
Signed-off-by: Simon Glass <sjg@chromium.org>
This driver cannot work when booted from coreboot, since the FSP
information is not available. Disable it in that case, so that the
coreboot video driver can be used instead.
Signed-off-by: Simon Glass <sjg@chromium.org>
Move this code into a generic location so that it can be used by other x86
boards which want to boot from coreboot. Also ensure that this is called
if booting from coreboot.
Signed-off-by: Simon Glass <sjg@chromium.org>
Quite a few new tag types have been added over the years. Bring these into
U-Boot so that all required tags can be parsed.
Add a proper comment to struct sysinfo_t while we are here, since many of
the meanings are not obvious.
Signed-off-by: Simon Glass <sjg@chromium.org>
It is useful to be able to parse coreboot tables on any x86 build which is
booted from coreboot. Add a new Kconfig option to enable this feature and
move the code so it can be used on any board, if enabled.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add new timestamp codes that are present in coreboot, so that we can decode
these in U-Boot.
At present TS_U_BOOT_START_KERNEL is used twice. It should only be used
just before jumping to Linux, so update the other call site to use
TS_START_KERNEL.
Signed-off-by: Simon Glass <sjg@chromium.org>
This all relates to the sysinfo structure provided by coreboot. Put the
timestamp definitions into the same file as the others. Tidy up a few
comments at the same time.
Signed-off-by: Simon Glass <sjg@chromium.org>
It is possible to boot U-Boot for chromebook_coral either 'bare metal' or
from coreboot. In the latter case we want to provide access to the coreboot
sysinfo tables. Move the definitions into a file available to any x86
board.
Signed-off-by: Simon Glass <sjg@chromium.org>
In some cases CBFS does not start with a header but is just a collection
of files. It is possible to support this so long as the size of the CBFS
is provided.
Update the cbfs_init_mem() function to support this.
Signed-off-by: Simon Glass <sjg@chromium.org>
Since the recent bug fix, it doesn't matter which GPIO phandle is used so
long as the GPIO number is right. Still, we may as well use the correct
one to avoid confusion.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
At present the eMMC device does not have an alias so it appears after
the SD card which is device 1. There is no device 0 which is odd.
Make the eMMC device be the first one. Update the boot script to use the
new device.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
We don't need to spell out the separate pieces of U-Boot phase binaries
anymore. Revert to using the simple entry and let binman do the expansion
itself as needed.
Signed-off-by: Simon Glass <sjg@chromium.org>
With Apollo Lake, SPL is placed in read-only memory. Set this new option
so that OF_PLATDATA_INST can be used.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
These devices are not actually built in TPL but are currently active in
the TPL devicetree. For of-platdata-inst this means that we will try to
generate devices for them, which fails.
Update them to be active only in U-Boot proper.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
We don't normally need this driver in TPL/SPL, so drop it for now. It can
be enabled by individual boards if needed.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
With TPL we don't need full PCI support and it adds to code size. Instead,
a simple_bus driver is good enough to be able to read and write the PCI
config and do a little basic setup.
So at present there are two drivers in U-Boot called pci_x86. One is in
UCLASS_PCI, used in SPL and U-Boot proper. The other is in
UCLASS_SIMPLE_BUS and used only in TPL.
Add a tag to tell dtoc about this, so it knows which one to use when
generating the devices and uclasses.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
With the standard of-platdata we must fix up driver_data manually. With
of-platadata-inst this is not necessary, since it is added to the device
by dtoc.
Update the code to handle this.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
This enum is needed to generate build-time devices. Tell dtoc where to
find the header, to avoid compile errors in the generated code.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
The dm.h header should come first. In fact it needs to, since otherwise
the driver model definitions are not available to dt-structs.h
Fix this, since it causes problems with OF_PLATDATA_INST.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Collect this together in one place, so driver model can access set it up
in a new place if needed.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
At present most of the Intel-specific code is built on all devices, even
those which don't have software support for the features provided there.
This means that any board can enable CONFIG_INTEL_ACPIGEN even if it does
not have the required features.
Add a new INTEL_SOC option to control this access. This must be selected
by SoCs that can support the required features.
Reported-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
[bmeng: fixed a typo in arch/x86/Kconfig]
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
This option is better placed in the x86 code since it is not generic
enough to be in the core code. Move it.
Reported-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
[bmeng: fixed a typo in arch/x86/Kconfig]
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
When CONFIG_CHROMEOS is not enabled this currently does not build. Fix it.
Reported-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Building qemu-x86_64_defconfig with GCC 11.0 fails with:
arch/x86/cpu/intel_common/lpc.c:
In function ‘lpc_common_early_init’:
arch/x86/cpu/intel_common/lpc.c:56:40:
error: expression does not compute the number of elements in this array;
element type is ‘struct reg_info’, not ‘u32’ {aka ‘unsigned int’}
[-Werror=sizeof-array-div]
56 | sizeof(values) / sizeof(u32));
| ^
arch/x86/cpu/intel_common/lpc.c:56:40: note: add parentheses around the
second ‘sizeof’ to silence this warning
arch/x86/cpu/intel_common/lpc.c:50:11: note: array ‘values’ declared here
50 | } values[4], *ptr;
| ^~~~~~
Add parentheses to silence warning.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
We don't need to implement direction_input() and direction_output()
anymore. Drop them and use update_flags() instead.
Signed-off-by: Simon Glass <sjg@chromium.org>
Historically, the reset_cpu() function had an `addr` parameter which was
meant to pass in an address of the reset vector location, where the CPU
should reset to. This feature is no longer used anywhere in U-Boot as
all reset_cpu() implementations now ignore the passed value. Generic
code has been added which always calls reset_cpu() with `0` which means
this feature can no longer be used easily anyway.
Over time, many implementations seem to have "misunderstood" the
existence of this parameter as a way to customize/parameterize the reset
(e.g. COLD vs WARM resets). As this is not properly supported, the
code will almost always not do what it is intended to (because all
call-sites just call reset_cpu() with 0).
To avoid confusion and to clean up the codebase from unused left-overs
of the past, remove the `addr` parameter entirely. Code which intends
to support different kinds of resets should be rewritten as a sysreset
driver instead.
This transformation was done with the following coccinelle patch:
@@
expression argvalue;
@@
- reset_cpu(argvalue)
+ reset_cpu()
@@
identifier argname;
type argtype;
@@
- reset_cpu(argtype argname)
+ reset_cpu(void)
{ ... }
Signed-off-by: Harald Seiler <hws@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Unfortunately the multi-core boot for QEMU x86 has been broken since
commit 77a5e2d3bc ("x86: mp_init: Set up the CPU numbers at the start").
In order to support QEMU x86 multi-core boot, the /cpus node must be
bound before any actual fix up in qemu_cpu_fixup(). This adds the
uclass_get() call to ensure this, just like what was done before.
Fixes: 77a5e2d3bc ("x86: mp_init: Set up the CPU numbers at the start")
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Move this out of the common header and include it only where needed. In
a number of cases this requires adding "struct udevice;" to avoid adding
another large header or in other cases replacing / adding missing header
files that had been pulled in, very indirectly. Finally, we have a few
cases where we did not need to include <asm/global_data.h> at all, so
remove that include.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Enhance the debugging to show the next stage being booted as well as a
dump of the start of the image.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
This should be done even if not using TPL, since BSS may be in use or
boards that only use SPL. Fix it.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
At present if the command line is very long it is truncated by the
printf() statement, which works within a limited buffer. Use puts()
instead. Also show better debugging with the command-line setup
fails.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
At present it is possible to dump an image within the zimage command, but
it is also useful to be able to dump it from elsewhere, for example in a
loader that has special handling for the different zimage stages.
Export this feature as a new function.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
The Global Non-Volatile Storage struct has some fields with particular
meanings. Rename these to make things easier to follow. Also add a few
more boot flags.
GNVS should not be confused with GNVQ (Going Nowhere Very Quickly).
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
At present BSS is always placed in SDRAM. If a separate BSS is not in use
this means that BSS doesn't work as expected. Make the setting conditional
on the SEPARATE_BSS option.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
The part of U-Boot that actually ends up in u-boot-nodtb.bin is not built
with any particular alignment. It ends at the start of the BSS section.
The BSS section selects its own alignment, which may larger.
This means that there can be a gap of a few bytes between the image
ending and BSS starting.
Since u-boot.bin is build by joining u-boot-nodtb.bin and u-boot.dtb (with
perhaps some padding for BSS), the expected result is not obtained. U-Boot
uses the end of BSS to find the devicetree, so this means that it cannot
be found.
Add 32-byte alignment of BSS so that the image size is correct and
appending the devicetree will place it at the end of BSS.
Example SPL output without this patch:
Sections:
Idx Name Size VMA LMA File off Algn
0 .text 000142a1 fef40000 fef40000 00001000 2**4
CONTENTS, ALLOC, LOAD, RELOC, READONLY, CODE
1 .u_boot_list 000014a4 fef542a8 fef542a8 000152a8 2**3
CONTENTS, ALLOC, LOAD, RELOC, DATA
2 .rodata 0000599c fef55760 fef55760 00016760 2**5
CONTENTS, ALLOC, LOAD, RELOC, READONLY, DATA
3 .data 00000970 fef5b100 fef5b100 0001c100 2**5
CONTENTS, ALLOC, LOAD, RELOC, DATA
4 .binman_sym_table 00000020 fef5ba70 fef5ba70 0001ca70 2**2
CONTENTS, ALLOC, LOAD, DATA
5 .bss 00000060 fef5baa0 fef5baa0 00000000 2**5
ALLOC
You can see that .bss is aligned to 2**5 (32 bytes). This is because of
the mallinfo struct in dlmalloc.c:
17 .bss.current_mallinfo 00000028 00000000 00000000 000004c0 2**5
ALLOC
In this case the size of u-boot-spl-nodtb.bin is 0x1ba90. This matches up
with the _image_binary_end symbol:
fef5ba90 g .binman_sym_table 00000000 _image_binary_end
But BSS starts 16 bytes later, at 0xfef5baa0, due to the 32-byte
alignment. So we must align _image_binary_end to a 32-byte boundary. This
forces the binary size to be 0x1baa0, i.e. ending at the start of bss, as
expected.
Note that gcc reports __BIGGEST_ALIGNMENT__ of 16 on this build, even
though it generates an object file with a member that requests 32-byte
alignment.
The current_mallinfo struct is 40 bytes in size. Increasing the struct to
68 bytes (i.e. just above a 64-byte boundary) does not cause the alignment
to go above 32 bytes. So it seems that 32 bytes is the maximum alignment
at present.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
[bmeng: add more details in the commit message to help people understand]
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Move to log_debug() and make use of the new SPL function to find the
text base.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Use a driver name in line with the compatible string so that of-platdata
can use this driver.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Add a node for this so we can indicate that it is does not require any
ACPI code.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Update the devicetree so that the TPM and RTC can be used in SPL. Also
enable the pins used for getting the memory configuration settings while
we are here.
Signed-off-by: Simon Glass <sjg@chromium.org>
