Simon Glass [Wed, 23 Dec 2020 02:30:29 +0000 (19:30 -0700)]
dm: core: Use access methods for dev/uclass private data
Use these functions in the core code as much as possible. With this, there
are only two places where each priv/plat pointer is accessed, one for read
and one for write.
Simon Glass [Wed, 23 Dec 2020 02:30:22 +0000 (19:30 -0700)]
dm: core: Support dm_dump_all() in SPL
At present the output from this function is hard to read in SPL, due to
(intended) limitations in SPL's printf() function. Add an SPL version so
it is clearer.
Simon Glass [Wed, 23 Dec 2020 02:30:21 +0000 (19:30 -0700)]
spl: Drop bd_info in the data section
This uses up space in the SPL binary but it always starts as zero. Also
some boards cannot support data in TPL (e.g. Intel Apollo Lake).
Use malloc() to allocate this structure instead, by moving the init a
little later, after malloc() is inited. Make this function optional since
it pulls in malloc().
This reduces the TPL binary size on coral by about 64 bytes
Simon Glass [Wed, 23 Dec 2020 02:30:20 +0000 (19:30 -0700)]
sandbox: Drop unnecessary test node
The spl-test4 node deliberately has an invalid compatible string. This
causes a warning from dtoc and the check it does is not really necessary.
Drop it, to avoid the warning and associated confusion.
Simon Glass [Wed, 23 Dec 2020 02:30:15 +0000 (19:30 -0700)]
Makefile: Tidy up SPL dtb production
Use the SPL_NAME variable to simplify the rules. Drop the SPL targets
clean-files since the SPL and TPL dts/ directories are removed by
existing rules. Move the SPL rules into a new spl_dtbs to avoid the
complicated $(if) construct.
Also drop unused pieces from the 'targets' variable.
With this, SPL and TPL have separate dtbs which respect the various
u-boot,dm-spl / u-boot,dm-tpl tags.
Simon Glass [Wed, 23 Dec 2020 02:30:14 +0000 (19:30 -0700)]
Makefile: Build SPL dtbs in the spl/ directory
Rather than putting these in the top-level dts/ directory (which is
intended for U-Boot proper), put them in the correct subdirectory for
SPL (either spl/ or tpl/). This is where other SPL targets are kept,
so this is more consistent.
Simon Glass [Wed, 23 Dec 2020 02:30:13 +0000 (19:30 -0700)]
Makefile: Build a separate .dtb for TPL
At present both SPL and TPL use the same devicetree binary. While there
is logic to run fdtgrep separately on each one, it does not actually
happen.
Add a new TPL rule and use that instead. Make this rule conditional on
there actually being a TPL. Do the same for SPL for consistency.
Note that the SPL and TPL dtbs are build by a Makefule rule used for
U-Boot proper. This is the 'dtbs' target in dts/Makefile. So the check
for CONFIG_TPL_BUILD in cmd_fdtgrep never actually works at present.
We don't support CONFIG_OF_EMBED for TPL at present.
Sughosh Ganu [Wed, 30 Dec 2020 13:57:09 +0000 (19:27 +0530)]
efi: capsule: Add support for uefi capsule authentication
Add support for authenticating uefi capsules. Most of the signature
verification functionality is shared with the uefi secure boot
feature.
The root certificate containing the public key used for the signature
verification is stored as part of the device tree blob. The root
certificate is stored as an efi signature list(esl) file -- this file
contains the x509 certificate which is the root certificate.
Sughosh Ganu [Wed, 30 Dec 2020 13:57:08 +0000 (19:27 +0530)]
efi_loader: Re-factor code to build the signature store from efi signature list
The efi_sigstore_parse_sigdb function reads the uefi authenticated
variable, stored in the signature database format and builds the
signature store structure. Factor out the code for building
the signature store. This can then be used by the capsule
authentication routine to build the signature store even when the
signature database is not stored as an uefi authenticated variable
Sughosh Ganu [Wed, 30 Dec 2020 13:57:07 +0000 (19:27 +0530)]
efi_loader: Make the pkcs7 header parsing function an extern
The pkcs7 header parsing functionality is pretty generic, and can be
used by other features like capsule authentication. Make the function
an extern, also changing it's name to efi_parse_pkcs7_header
Sughosh Ganu [Wed, 30 Dec 2020 13:57:06 +0000 (19:27 +0530)]
dfu_mtd: Add provision to unlock mtd device
Prior to writing to an mtd device, mtd_erase is called. This call
fails in case the sector being erased is locked. Call mtd_unlock to
unlock the region which is to be erased and later written to. Lock the
region once the write to the region has completed.
Sughosh Ganu [Wed, 30 Dec 2020 13:57:05 +0000 (19:27 +0530)]
efi_loader: Add logic to parse EDKII specific fmp payload header
When building the capsule using scripts in edk2, a fmp header is
added on top of the binary payload. Add logic to detect presence of
the header. When present, the pointer to the image needs to be
adjusted as per the size of the header to point to the actual binary
payload.
Sughosh Ganu [Wed, 30 Dec 2020 13:57:04 +0000 (19:27 +0530)]
fsp: Move and rename fsp_types.h file
The fsp_types.h header file contains macros for building signatures of
different widths. These signature macros are architecture agnostic,
and can be used in all places which use signatures in a data
structure. Move and rename the fsp_types.h under the common include
header.
Signed-off-by: Sughosh Ganu <sughosh.ganu@linaro.org> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Sughosh Ganu [Wed, 30 Dec 2020 13:57:03 +0000 (19:27 +0530)]
qemu: common: Set dfu_alt_info variable for the platform
The dfu framework uses the dfu_alt_info environment variable to get
information that is needed for performing the firmware update. Add
logic to set the dfu_alt_info for the qemu arm64 platform to reflect
the two mtd partitions created for the u-boot env and the firmware
image. This can be subsequently extended for other qemu architectures
which need this variable set.
Sughosh Ganu [Wed, 30 Dec 2020 13:57:02 +0000 (19:27 +0530)]
qemu: common: Add support for dynamic mtdparts for the platform
Add support for setting the default values for mtd partitions on the
platform. This would be used for updating the firmware image using
uefi capsule update with the dfu mtd backend driver.
Currently, values have been defined for the qemu arm64 platform, with
default values defined for the mtd partitions based on the NOR
flash. This can be subsequently extended for other qemu architectures
which need mtdparts set.
Sughosh Ganu [Wed, 30 Dec 2020 13:57:01 +0000 (19:27 +0530)]
crypto: Fix the logic to calculate hash with authattributes set
RFC 2315 Section 9.3 describes the message digesting process. The
digest calculated depends on whether the authenticated attributes are
present. In case of a scenario where the authenticated attributes are
present, the message digest that gets signed and is part of the pkcs7
message is computed from the auth attributes rather than the contents
field.
Check if the auth attributes are present, and if set, use the auth
attributes to compute the hash that would be compared with the
encrypted hash on the pkcs7 message.
Sughosh Ganu [Wed, 30 Dec 2020 13:57:00 +0000 (19:27 +0530)]
qemu: arm: Initialise virtio devices in board_late_init
On the qemu arm platform, the virtio devices are initialised in the
board_init function, which gets called before the initr_pci. With
this sequence, the virtio block devices on the pci bus are not
initialised. Move the initialisation of the virtio devices to
board_late_init which gets called after the call to initr_pci.
Sughosh Ganu [Wed, 30 Dec 2020 13:56:59 +0000 (19:26 +0530)]
mkeficapsule: Add support for embedding public key in a dtb
Add options for embedding the public key esl(efi signature list) file
to the platform's dtb. The esl file is then retrieved and used for
authenticating the capsule to be used for updating firmare components
on the platform.
The esl file can now be embedded in the dtb by invoking the following
command
mkeficapsule -K <pub_key.esl> -D <dtb>
In the scenario where the esl file is to be embedded in an overlay,
this can be done through the following command
mkeficapsule -O -K <pub_key.esl> -D <dtb>
This will create a node named 'signature' in the dtb, and the esl file
will be stored as 'capsule-key'
When a new event is queued we have to process the event queue by calling
efi_process_event_queue(). But there is not reason to call the function
when the event is not queueable.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Ilias Apalodimas [Thu, 31 Dec 2020 10:26:46 +0000 (12:26 +0200)]
efi_loader: Add size checks to efi_create_indexed_name()
Although the function description states the caller must provide a
sufficient buffer, it's better to have in function checks that the
destination buffer can hold the intended value.
So let's add an extra argument with the buffer size and check that
before doing any copying.
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org> Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Ilias Apalodimas [Wed, 30 Dec 2020 15:07:14 +0000 (17:07 +0200)]
efi_loader: Remove unconditional installation of file2 protocol for initrd
Up to now we install the 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 start the installation.
A following patch introduces a different logic where we search for an
initrd path defined in an EFI variable named 'Initrd####'.
If the bootmgr is used to launch the EFI payload, we'll will try to match
the BootCurrent value and find the corresponding initrd
(i.e Boot0000 -> Initrd0000 etc). If the file is found, we'll install
the required protocol which the kernel's efi-stub can use and load our
initrd.
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org> Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The FAT filesystem implementation uses several marcros referring to a magic
variable name mydata which renders the code less readable. Eliminate one of
them which is only used for a debug() statement.
Use log_debug() instead of debug().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Reviewed-by: Simon Glass <sjg@chromium.org>
Ilias Apalodimas [Wed, 23 Dec 2020 11:25:00 +0000 (13:25 +0200)]
efi_loader: Extra checks while opening an OPTEE session
When opening an OP-TEE session we need to check the internal return
value of OP-TEE call arguments as well the return code of the
function itself.
The code was also ignoring to close the OP-TEE session in case the
shared memory registration failed.
Fixes: f042e47e8fb43 ("efi_loader: Implement EFI variable handling via OP-TEE") Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org> Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Simon Glass [Thu, 17 Dec 2020 00:25:06 +0000 (17:25 -0700)]
dm: core: Inline a few ofnode functions in SPL
A recent change to unify the flattree/livetree code introduced a small
size increase in SPL on some boards. For example SPL code size for
px30-core-ctouch2-px30 increased by 40 bytes.
To address this we can take advantage of the fact that some of the ofnode
functions are only called a few times in SPL, so it is worth inlining
them.
Add new Kconfig options to control this. These functions are not inlined
for U-Boot proper, since this increases code size.
Fixes: 2ebea5eaebf ("dm: core: Combine the flattree and livetree binding code") Signed-off-by: Simon Glass <sjg@chromium.org>
Simon Glass [Thu, 17 Dec 2020 00:24:17 +0000 (17:24 -0700)]
buildman: Remove output binaries before building
Buildman reuses build directories from previous builds to avoid the cost
of 'make mrproper' for every build. If the previous build produced an SPL
image but the current one does not, the SPL image will remain and buildman
will think it is a result of building the current board.
Remove these files before building, to avoid this problem.
Niel Fourie [Wed, 16 Dec 2020 11:11:52 +0000 (12:11 +0100)]
dm: spi: Fix spi_free_slave() freed memory write
Remove setting slave->dev to NULL after the device_remove() call.
The slave pointer points to dev->parent_priv, which has already
been freed by device_free(), called from device_remove() in the
preceding line. Writing to slave->dev may cause corruption of the
dlmalloc free chunk forward pointer of the previously freed chunk.
Signed-off-by: Niel Fourie <lusus@denx.de> Cc: Simon Glass <sjg@chromium.org> Reviewed-by: Simon Glass <sjg@chromium.org>
Ovidiu Panait [Mon, 14 Dec 2020 17:06:51 +0000 (19:06 +0200)]
test: dm: spi: Add testcase for spi_claim_bus()
Add testcase for spi_claim_bus(), which checks that sandbox spi bus
speed/mode settings are updated correctly when multiple slaves use
the bus consecutively. The following configurations are used for the
two spi slaves involved:
* different max_hz / different modes
* different max_hz / same modes
* different modes / same max_hz
asm/test.h header is added in order to be able to retrieve the current
speed/mode of the sandbox spi bus, via sandbox_spi_get_{speed, mode}.
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Currently, when different spi slaves claim the bus consecutively using
spi_claim_bus(), spi_set_speed_mode() will only be executed on the first
two calls, leaving the bus in a bad state starting with the third call.
This patch drops spi_slave->speed member and adds caching of bus
speed/mode in dm_spi_bus struct. It also updates spi_claim_bus() to call
spi_set_speed_mode() if either speed or mode is different from what the
bus is currently configured for. Current behavior is to only take into
account the speed, but not the mode, which seems wrong.
Fixes: 60e2809a848 ("dm: spi: Avoid setting the speed with every transfer") Reviewed-by: Simon Glass <sjg@chromium.org> Reported-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk> Reported-by: Moshe, Yaniv <yanivmo@amazon.com> Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Introduce sandbox_spi_get_{speed, mode} public interface to retrieve the
sandbox spi bus internal state. They are meant to be used in sandbox spi
testcases.
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Ovidiu Panait [Mon, 14 Dec 2020 17:06:48 +0000 (19:06 +0200)]
spi: sandbox_spi: Implement speed/mode setup
Implement sandbox_spi_set_{speed, mode} routines, to be able to keep track
of the current bus speed/mode. This will help determine whether the values
passed from dm_spi_claim_bus() are valid.
Ovidiu Panait [Mon, 14 Dec 2020 17:06:47 +0000 (19:06 +0200)]
sandbox: test: Add a second SPI slave on sandbox_spi bus
Place a second spi slave on the sandbox_spi bus, to be used by the
spi_claim_bus() testcase we are about to introduce. We need to make sure
that jumping between slaves calling spi_claim_bus() sets the bus speed and
mode appropriately. Use different max-hz and mode properties for this new
slave.
Also, update sandbox_spi cs_info call to allow activity on CS0/CS1 and
adapt dm_test_spi_find() testcase for this new setup.
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Ovidiu Panait [Mon, 14 Dec 2020 17:06:46 +0000 (19:06 +0200)]
sandbox: spi: Drop unused sandbox_spi_parse_spec function
Commit 1289e96797bf ("sandbox: spi: Drop command-line SPI option") dropped
support for specifying SPI devices on the command line, removing the only
user of sandbox_spi_parse_spec(). Remove the function too.
Fixes: 1289e96797bf ("sandbox: spi: Drop command-line SPI option") Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com> Reviewed-by: Simon Glass <sjg@chromium.org>
fdt: Use phandle to distinguish DT nodes with same name
While assigning the sequence number to subsystem instances by reading the
aliases property, only DT nodes names are compared and not the complete
path. This causes a problem when there are two DT nodes with same name but
have different paths.
In arch/arm/dts/k3-am65-main.dtsi there are two USB controllers with the
same device tree node name but different path. When aliases are defined for
these USB controllers then fdtdec_get_alias_seq() fails to pick the correct
instance for a given index.
fdt_path_offset() function is slow and this would effect the U-Boot
startup. To avert the time penalty on all boards, apply this extra check
only when required by using a config option.
Fix it by comparing the phandles of DT nodes after the node names match,
under a config option.
Signed-off-by: Aswath Govindraju <a-govindraju@ti.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Fix whitespace error in Kconfig: Signed-off-by: Simon Glass <sjg@chromium.org>
Philipp Tomsich [Tue, 24 Nov 2020 17:14:52 +0000 (18:14 +0100)]
patman: Add --no-signoff to suppress adding signoffs
To enable use of patman with FSF/GNU projects, such as GCC or
Binutils, no Signed-off-by may be added. This adds a command
line flag '--no-signoff' to suppress adding signoffs in patman
when processing commits.
Signed-off-by: Philipp Tomsich <philipp.tomsich@vrull.eu> Reviewed-by: Simon Glass <sjg@chromium.org>
Fix patman testBranch() test: Signed-off-by: Simon Glass <sjg@chromium.org>
Simon Glass [Thu, 17 Dec 2020 04:20:30 +0000 (21:20 -0700)]
dm: core: Update uclass_find_next_free_req_seq() for new scheme
This function current deals with req_seq which is deprecated. Update it to
use the new sequence numbers, putting them above existing aliases. Rename
the function to make this clear.
Simon Glass [Thu, 17 Dec 2020 04:20:29 +0000 (21:20 -0700)]
dm: Drop the unused arg in uclass_find_device_by_seq()
Now that there is only one sequence number (rather than both requested and
assigned ones) we can simplify this function. Also update its caller to
simplify the logic.
Simon Glass [Thu, 17 Dec 2020 04:20:22 +0000 (21:20 -0700)]
x86: Drop unnecessary mp_init logic
Now that sequence numbers are set up when devices are bound, this code is
not needed. Also, we should use dev_seq() instead of req_seq. Update the
whole file accordingly.
Simon Glass [Thu, 17 Dec 2020 04:20:20 +0000 (21:20 -0700)]
usb: ehci-mx6: Drop assignment of sequence number
This hack cannot work in the new sequence-numbering scheme. Remove it
while we wait for the maintainer to complete DM conversion as noted in
the existing comment.
Simon Glass [Thu, 17 Dec 2020 04:20:19 +0000 (21:20 -0700)]
spi: Update for new sequence numbers
Use the new sequence number in all cases. Drop the rockchip case because
the sequence number should be 0 anyway, and assigning to the sequence
number is not permitted.
Simon Glass [Thu, 17 Dec 2020 04:20:17 +0000 (21:20 -0700)]
dm: core: Allow manual sequence numbering
Some buses have their own rules which require assigning sequence numbers
with a bus-specific algorithm. For example, PCI requires that sub-buses
are numbered higher than their parent buses, meaning effectively that
parent buses must be numbered only after all of their child buses have
been numbered.
Add a uclass flag to indicate that driver model should not assign sequence
numbers. In this case, the uclass must do it.
Simon Glass [Thu, 17 Dec 2020 04:20:16 +0000 (21:20 -0700)]
net: Update to use new sequence numbers
Checking for seq == -1 is effectively checking that the device is
activated. The new sequence numbers are never -1 for a bound device, so
update the check.
Also drop the note about valid sequence numbers so it is accurate with the
new approach.
Simon Glass [Thu, 17 Dec 2020 04:20:14 +0000 (21:20 -0700)]
octeon: Don't attempt to set the sequence number
Several Octeon drivers operate by setting the sequence number of their
device. This should not be needed with the new sequence number setup. Also
it is not permitted. Drop it.
Simon Glass [Thu, 17 Dec 2020 04:20:09 +0000 (21:20 -0700)]
dm: core: Add a new sequence number for devices
At present each device has two sequence numbers, with 'req_seq' being
set up at bind time and 'seq' at probe time. The idea is that devices
can 'request' a sequence number and then the conflicts are resolved when
the device is probed.
This makes things complicated in a few cases, since we don't really know
what the sequence number will end up being. We want to honour the
bind-time requests if at all possible, but in fact the only source of
these at present is the devicetree aliases. Since we have the devicetree
available at bind time, we may as well just use it, in the hope that the
required processing will turn out to be useful later (i.e. the device
actually gets used).
Add a new 'sqq' member, the bind-time sequence number. It operates in
parallel to the old values for now. All devices get a valid sqq value,
i.e. it is never -1.
At present this is passed a uclass ID and it has to do a lookup. The
callers all have the uclass pointer, except for the I2C uclass where the
code will soon be deleted.
Update the argument to a uclass * instead of an ID since it is more
efficient.
Simon Glass [Thu, 17 Dec 2020 04:20:07 +0000 (21:20 -0700)]
dm: Avoid accessing seq directly
At present various drivers etc. access the device's 'seq' member directly.
This makes it harder to change the meaning of that member. Change access
to go through a function instead.
The drivers/i2c/lpc32xx_i2c.c file is left unchanged for now.
Simon Glass [Thu, 17 Dec 2020 04:20:06 +0000 (21:20 -0700)]
linker_lists: Fix alignment issue
The linker script uses alphabetic sorting to group the different linker
lists together. Each group has its own struct and potentially its own
alignment. But when the linker packs the structs together it cannot ensure
that a linker list starts on the expected alignment boundary.
For example, if the first list has a struct size of 8 and we place 3 of
them in the image, that means that the next struct will start at offset
0x18 from the start of the linker_list section. If the next struct has
a size of 16 then it will start at an 8-byte aligned offset, but not a
16-byte aligned offset.
With sandbox on x86_64, a reference to a linker list item using
ll_entry_get() can force alignment of that particular linker_list item,
if it is in the same file as the linker_list item is declared.
Consider this example, where struct driver is 0x80 bytes:
If these two lines of code are in the same file, then the entry is forced
to be aligned at the 'struct driver' alignment, which is 16 bytes. If the
second line of code is in a different file, then no action is taken, since
the compiler cannot update the alignment of the linker_list item.
In the first case, an 8-byte 'fill' region is added:
With this, the linker_list no-longer works since items after testfdt1_drv
are not at the expected address.
Ideally we would have a way to tell gcc not to align structs in this way.
It is not clear how we could do this, and in any case it would require us
to adjust every struct used by the linker_list feature.
One possible fix is to force each separate linker_list to start on the
largest possible boundary that can be required by the compiler. However
that does not seem to work on x86_64, which uses 16-byte alignment in this
case but needs 32-byte alignment.
So add a Kconfig option to handle this. Set the default value to 4 so
as to avoid changing platforms that don't need it.