Make use of the new priv/plat data region if enabled. This is implemented
as a simple offset from the position set up by dtoc to the new position.
So long as all access goes through dm_priv_to_rw() this is safe.
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 the device priv/data data allocated by dtoc is stored in the
data section along with other variables. On some platforms it is better
to allocate space for it separately, e.g. if SPL is running from read-only
memory.
Create a new space with the same size as that allocated by dtoc, ready for
use.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
When of-platdata-inst is active, use the flags in the new udevice_rt
table, dropping them from the main struct udevice. This ensures that the
latter is not updated at runtime.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
At present when driver model needs to change a device it simply updates
the struct udevice structure. But with of-platdata-inst most of the fields
are not modified at runtime. In fact, typically only the flags need to
change.
For systems running SPL from read-only memory it is convenient to separate
out the runtime information, so that the devices don't need to be copied
before being used.
Create a new udevice_rt table, similar to the existing driver_rt. For now
it just holds the flags, although they are not used in this patch.
Add a new Kconfig for the driver_rt data, since this is not needed when
of-platdata-inst is used.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Some systems (e.g. x86 APL) run SPL from read-only memory. The device
instances created by dtoc are therefore not writeable. To make things work
we would need to copy the devices to read/write memory.
To avoid this, add an option to use a separate runtime struct for devices,
just as is done for drivers. This can be used to hold information that
changes at runtime, avoiding the need for a copy.
Also add a Kconfig option for read-only SPL, which selects this feature.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Add this new board to the test plans. Travis-CI is left out, since it is
being removed soon due to lack of capacity.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Move sandbox_spl over to use OF_PLATDATA_INST. Create a new board to
test the case when this is not enabled, since we will be keeping that
code around for several months and want to avoid regressions.
Skip the dm_test_of_plat_dev() test since driver info is not available
for OF_PLATDATA_INST.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
With recent changes this can be supported again. Add it back.
This reverts commit d85f2c4f29.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
At present the i2c emulators require access to the devicetree, which is
not possible (by design) with of-platdata.
Add a way for drivers to record the of-platdata index of their emulator,
so that we can still find the emulator.
This allows i2c emulation to work with of-platdata.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
The test framework reinits driver model tests before running each test.
Since malloc_simple does not support free(), this eventually runs out of
memory.
Fix it for now by increasing the space to 32KB.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This function finds a device by its driver_info index. With
of-platdata-inst we do not use driver_info, but instead instantiate
udevice records at build-time.
However the semantics of using the function are the same in each case:
the caller provides an index and gets back a device.
So rename the function to device_get_by_ofplat_idx(), so that it can be
used for both situations. The caller does not really need to worry about
the details.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
At present the structs used by these drivers are declared in the C files
and so are not accessible to dtoc. Move them to header files, as required.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
At present this function is included in the build but with of-platdata it
only services to produce a confusing link error complaining about a call
to dev_read_u32_default().
Drop it so that any call to uclass_find_device_by_phandle() is flagged as
an error, making it easier to see what is going on.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
This function is now only used in a test. Drop it. Also drop
DM_DRVINFO_GET() which was the only purpose for having the function.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Create a version of this driver for sandbox so that it can use the
of-platdata struct.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
We need to allow SoCs to create their own drivers for this so that they
can use their own of-platdata structs. To minimise code duplication,
export the driver operations and the ofdata_to_plat() setup function.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
At present the structs used by this driver are not accessible outside it,
so cannot be used with OF_PLATDATA_INST. Move them to a header file to
fix this.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
It is not possible to destroy the uclasses when they are created at
build time. Skip this step so that SPL test can complete successfully.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
When dt-structs.h is used, include the dt-decl.h header as well, so that
these declarations are available.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
Which files we generate depends on the setting of OF_PLATDATA_INST in the
build. This might change between builds, but the build directory may be
reused.
Leaving old files around is confusing and switching the OF_PLATDATA_INST
setting does not necessarily regenerate the files, e.g. if the devicetree
has not changed.
Remove all the files before regenerating new ones.
Signed-off-by: Simon Glass <sjg@chromium.org>
There are actually two generated files but only one is currently mentioned
in the Makefile as a dependency. Put them both in a Makefile variable and
use that instead, to avoid inconsistency.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present all possible files are generated, even if some of them just
have a header and an empty body. It is better to generate only the files
that are needed, so that the two types of build (based on the setting of
OF_PLATDATA_INST) can be mutually exclusive.
This is intended to fix a strange problem sometimes found with CI:
Building current source for 1 boards (1 thread, 40 jobs per thread)
sandbox: + sandbox_spl
+drivers/built-in.o: In function `dm_setup_inst':
+drivers/core/root.c:135: undefined reference to
`_u_boot_list_2_udevice_2_root'
+dts/dt-uclass.o:(.u_boot_list_2_uclass_2_serial+0x10): undefined
reference to `_u_boot_list_2_udevice_2_serial'
...
This likely happens when switching from !OF_PLATDATA_INST to
OF_PLATDATA_INST since running 'make xxx_defconfig" does not currently
cause any change in which files are generated. With !OF_PLATDATA_INST
the dt-device.c file has no declarations and this is assumed to be the
starting state. The error above seems to indicate that, after changing
to OF_PLATDATA_INST, the dt-uclass.c file is regenerated but the
dt-device.c files is not. This does not seem possible from the relevant
Makefile.spl rule:
u-boot-spl-platdata := $(obj)/dts/dt-plat.o $(obj)/dts/dt-uclass.o
$(obj)/dts/dt-device.o
cmd_dtoc = $(DTOC_ARGS) -c $(obj)/dts -C include/generated all
include/generated/dt-structs-gen.h $(u-boot-spl-platdata_c) &: \
$(obj)/$(SPL_BIN).dtb
@[ -d $(obj)/dts ] || mkdir -p $(obj)/dts
$(call if_changed,dtoc)
It seems that this cannot regenerate dt-uclass.c without dt-device.c since
'dtoc all' is used. So here the trail ends for now.
In any case it seems better to generate files that are uses and not bother
with those that serve no purpose. So update dtoc to do this automatically.
Signed-off-by: Simon Glass <sjg@chromium.org>
When OF_PLATDATA_INST is enabled, we need dt-uclass.c and dt-device.c for
the build to work. When OF_PLATDATA_INST is not enabled, we only need
dt-plat.c
Update the Makefile rules to indicate this.
At present all files are generated and compiled regardless of which are
actually needed.
Signed-off-by: Simon Glass <sjg@chromium.org>
Since the dst_dir rule always runs, it causes a rebuild of the of-platdata
files even if not needed.
Create the directory inside the rule instead.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present the build rule for pylibfdt depends on _libfdt.so but modern
Python versions add a different suffix to the output file, resulting in
something like _libfdt.cpython-38-x86_64-linux-gnu.so
The result is that pylibfdt is rebuilt every time.
Rename the file the standard name so that the rule works correctly. Also
add libfdt.py to the dependencies, so that file is always created if
missing.
Signed-off-by: Simon Glass <sjg@chromium.org>
We recently added a common function for calculating file size,
instead of copy pasting the code around. Switch one of the
occurences over to the common function
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reviwed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Setup will always occur before ExitBootServices(). So eliminate
EFI_SETUP_AFTER_BOOTTIME_EXIT. Put the SetVirtualAddressMap() test into a
separate class so that we can execute it last.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
TCG PC Client Platform Firmware Profile Spec mandates that an S-CRTM
event for the version identifier using the event type EV_S_CRTM_VERSION
must be measured.
So since we are trying to add more conformance into U-Boot, let's add
the event using U_BOOT_VERSION_STRING, extend PCR[0] accordingly and log
it in the EventLog
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Current code doesn't free the efi allocated memory in case the protocol
failed to install
Fixes: c8d0fd5825 ("efi_loader: Introduce eventlog support for TCG2_PROTOCOL")
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
This commit slightly extends test_efi_capsule_fw3.
In order to run the test the following must be added to
sandbox_defconfig:
+CONFIG_CMD_SF=y
+CONFIG_CMD_MEMORY=y
+CONFIG_CMD_FAT=y
+CONFIG_DFU=y
The ESRT is printed in the u-boot shell by calling efidebug esrt.
The test ensures that, after the capsule is installed, the ESRT
contains entries with the GUIDs:
- EFI_FIRMWARE_IMAGE_TYPE_UBOOT_FIT_GUID;
- EFI_FIRMWARE_IMAGE_TYPE_UBOOT_RAW_GUID;
test invocation:
sudo ./test/py/test.py --bd sandbox -k capsule_fw3 -l --build
CC: Heinrich Schuchardt <xypron.glpk@gmx.de>
CC: Sughosh Ganu <sughosh.ganu@linaro.org>
CC: AKASHI Takahiro <takahiro.akashi@linaro.org>
CC: Ilias Apalodimas <ilias.apalodimas@linaro.org>
CC: Andre Przywara <andre.przywara@arm.com>
CC: Alexander Graf <agraf@csgraf.de>
CC: nd@arm.com
Signed-off-by: Jose Marinho <jose.marinho@arm.com>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This commmit exercises the ESRT creation in a EFI selftest.
A fake FMP, with TEST_ESRT_NUM_ENTRIES FW images, is installed in the
system leading to the corresponding ESRT entries being populated.
The ESRT entries are checked against the datastructure used to
initialize the FMP.
Invocation from the sandbox platform:
add to sandbox_defconfig:
+CONFIG_CMD_BOOTEFI_SELFTEST=y
make sandbox_capsule_defconfig all
./u-boot -d arch/sandbox/dts/test.dtb
bootefi selftest
CC: Heinrich Schuchardt <xypron.glpk@gmx.de>
CC: Sughosh Ganu <sughosh.ganu@linaro.org>
CC: AKASHI Takahiro <takahiro.akashi@linaro.org>
CC: Ilias Apalodimas <ilias.apalodimas@linaro.org>
CC: Andre Przywara <andre.przywara@arm.com>
CC: Alexander Graf <agraf@csgraf.de>
CC: nd@arm.com
Signed-off-by: Jose Marinho <jose.marinho@arm.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This commit enables the ESRT printing from the u-boot shell by invoking:
- efidebug capsule esrt
CC: Heinrich Schuchardt <xypron.glpk@gmx.de>
CC: Sughosh Ganu <sughosh.ganu@linaro.org>
CC: AKASHI Takahiro <takahiro.akashi@linaro.org>
CC: Ilias Apalodimas <ilias.apalodimas@linaro.org>
CC: Andre Przywara <andre.przywara@arm.com>
CC: Alexander Graf <agraf@csgraf.de>
CC: nd@arm.com
Signed-off-by: Jose Marinho <jose.marinho@arm.com>
Reviwed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The ESRT is initialised during efi_init_objlist after
efi_initialize_system_table().
The ESRT is recreated from scratch at the following events:
- successful UpdateCapsule;
- FMP instance install.
The code ensures that every ESRT entry has a unique fw_class value.
Limitations:
- The ESRT is not updated if an FMP instance is uninstalled;
- the fields image_type and flags are in the current implementation left
undefined. Setting these values will require a per-platform function
that returns the image_type/flags as a function of the image fw_class.
CC: Heinrich Schuchardt <xypron.glpk@gmx.de>
CC: Sughosh Ganu <sughosh.ganu@linaro.org>
CC: AKASHI Takahiro <takahiro.akashi@linaro.org>
CC: Ilias Apalodimas <ilias.apalodimas@linaro.org>
CC: Andre Przywara <andre.przywara@arm.com>
CC: Alexander Graf <agraf@csgraf.de>
CC: nd@arm.com
Signed-off-by: Jose Marinho <jose.marinho@arm.com>
Remove two EFI_CALL() indirections.
Move ESRT GUID in efidebug's list of GUIDs.
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Multiple DFU entities may share the same SPI device. We must make sure that
the SPI device is only freed once.
When using the driver model it is not necessary to free the SPI device.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Lukasz Majewski <lukma@denx.de>
The UEFI spec allows a packed array of UEFI device paths in the
FilePathList[] of an EFI_LOAD_OPTION. The first file path must
describe the loaded image but the rest are OS specific.
Previous patches parse the device path and try to use the second
member of the array as an initrd. So let's modify efidebug slightly
and install the second file described in the command line as the
initrd device path.
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Up to now we install EFI_LOAD_FILE2_PROTOCOL to load an initrd
unconditionally. Although we correctly return various EFI exit codes
depending on the file status (i.e EFI_NO_MEDIA, EFI_NOT_FOUND etc), the
kernel loader, only falls back to the cmdline interpreted initrd if the
protocol is not installed.
This creates a problem for EFI installers, since they won't be able to
load their own initrd and continue the installation. It also makes the
feature hard to use, since we can either have a single initrd or we have
to recompile u-boot if the filename changes.
So let's introduce a different logic that will decouple the initrd
path from the config option we currently have.
When defining a UEFI BootXXXX we can use the filepathlist and store
a file path pointing to our initrd. Specifically the EFI spec describes:
"The first element of the array is a device path that describes the device
and location of the Image for this load option. Other device paths may
optionally exist in the FilePathList, but their usage is OSV specific"
When the EFI application is launched through the bootmgr, we'll try to
interpret the extra device path. If that points to a file that exists on
our disk, we'll now install the load_file2 and the efi-stub will be able
to use it.
This opens up another path using U-Boot and defines a new boot flow.
A user will be able to control the kernel/initrd pairs without explicit
cmdline args or GRUB.
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
A following patch introduces a different logic for loading initrd's
based on the EFI_LOAD_FILE2_PROTOCOL.
Since similar logic can be applied in the future for other system files
(i.e DTBs), let's add some helper functions which will retrieve and
parse file paths stored in EFI variables.
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
On the following patches we allow for an initrd path to be stored in
Boot#### variables. Specifically we encode in the FIlePathList[] of
the EFI_LOAD_OPTIONS for each Boot#### variable.
The FilePathList[] array looks like this:
kernel - 0xff - VenMedia(initrd GUID) - initrd1 - 0x01 initrd2 - 0xff
So let's add the relevant functions to concatenate and retrieve a device
path based on a Vendor GUID.
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reformat function descriptions.
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
When uploading an EFI binary via the UART we have to call
efi_set_bootdev() or we won't be able to execute it.
Put the includes into alphabetic order.
Fixes: 5f59518a7b ("efi_loader: setting boot device")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
When uploading an EFI binary via the UART we need to assign a device path.
* Provide devicepath node to text conversion for Uart() node.
* Provide function to create Uart() device path.
* Add UART support to efi_dp_from_name().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Even if CONFIG_DM_VIDEO=y and stdout="vidconsole", a video device may not
be available. Check the return values of the relevant functions.
If no video output device is available, assume that the serial console is
in use.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This occur when building on Raspberry Pi 400 (32-bit ARM). Fix them.
Examples:
cmd/efidebug.c: In function ‘do_efi_capsule_update’:
cmd/efidebug.c:75:49: warning: cast from pointer to integer of different
size [-Wpointer-to-int-cast]
ret = EFI_CALL(RT->update_capsule(&capsule, 1, (u64)NULL));
^
include/efi_loader.h:104:9: note: in definition of macro ‘EFI_CALL’
typeof(exp) _r = exp; \
^~~
cmd/efidebug.c:75:49: warning: cast from pointer to integer of different
size [-Wpointer-to-int-cast]
ret = EFI_CALL(RT->update_capsule(&capsule, 1, (u64)NULL));
^
include/efi_loader.h:104:19: note: in definition of macro ‘EFI_CALL’
typeof(exp) _r = exp; \
^~~
In file included from include/common.h:20,
from lib/efi_loader/efi_capsule.c:9:
lib/efi_loader/efi_capsule.c: In function ‘efi_update_capsule’:
include/efi_loader.h:83:8: warning: format ‘%lu’ expects argument of type
‘long unsigned int’, but argument 10 has type ‘size_t’
{aka ‘unsigned int’} [-Wformat=]
debug("%sEFI: Entry %s(" format ")\n", __efi_nesting_inc(), \
^~~~~~~~~~~~~~~~~~
include/linux/printk.h:37:21: note: in definition of macro ‘pr_fmt’
#define pr_fmt(fmt) fmt
^~~
include/log.h:229:2: note: in expansion of macro ‘log’
log(LOG_CATEGORY, LOGL_DEBUG, fmt, ##args); \
^~~
include/log.h:249:2: note: in expansion of macro ‘debug_cond’
debug_cond(_DEBUG, fmt, ##args)
^~~~~~~~~~
include/efi_loader.h:83:2: note: in expansion of macro ‘debug’
debug("%sEFI: Entry %s(" format ")\n", __efi_nesting_inc(), \
^~~~~
lib/efi_loader/efi_capsule.c:444:2: note: in expansion of macro ‘EFI_ENTRY’
EFI_ENTRY("%p, %lu, %llu\n", capsule_header_array, capsule_count,
^~~~~~~~~
lib/efi_loader/efi_capsule.c:444:19: note: format string is defined here
EFI_ENTRY("%p, %lu, %llu\n", capsule_header_array, capsule_count,
~~^
%u
Signed-off-by: Simon Glass <sjg@chromium.org>
Replace (uintptr_t)NULL by 0.
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
We are redefining how u-boot locates the initrd to load via the kernel
LoadFile2 protocol. This selftest is not relevant any more, so remove
it. A new one will be added later
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The (yet unreleased version of the) devicetree specification clearly
states that:
As with the /reserved-memory node, when booting via UEFI
entries in the Memory Reservation Block must also be listed
in the system memory map obtained via the GetMemoryMap() toi
protect against allocations by UEFI applications. The memory
reservation block entries should be listed with type
EfiReservedMemoryType.
This restores the behaviour that was changed by commit 4cbb2930bd
("efi_loader: consider no-map property of reserved memory").
Fixes: 4cbb2930bd ("efi_loader: consider no-map property of reserved memory")
Signed-off-by: Mark Kettenis <kettenis@openbsd.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
GRUB_ARM32_WORKAROUND can be disabled on ARCH_SUNXI as the Allwinner SoCs
only have a level 2 cache controlled via CP15 and not an architecturally
defined cache. Having the cache available speeds up booting Linux.
On ARCH_BCM283X it is already disabled via rpi_2_defconfig. But let's move
this setting to Kconfig.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
