Vincent Stehlé [Wed, 24 Nov 2021 14:54:20 +0000 (15:54 +0100)]
qemu: common: Fix build with update capsule
The common emulation Makefile has a dependency on a non-existent
qemu_capsule.o when building with support for capsule update enabled
(CONFIG_EFI_RUNTIME_UPDATE_CAPSULE=y).
The code which was in qemu_capsule.c has been completely moved to
lib/efi_loader/efi_capsule.c by commit ad59e0b39d7c ("efi_loader: capsule:
add back efi_get_public_key_data()").
Remove the false dependency.
This fixes the following build error:
make[1]: *** No rule to make target 'board/emulation/common/qemu_capsule.o', needed by 'board/emulation/common/built-in.o'. Stop.
Fixes: commit ae6e7a0a65aa ("Revert "efi_capsule: Move signature from DTB to .rodata"") Signed-off-by: Vincent Stehlé <vincent.stehle@arm.com> Cc: Simon Glass <sjg@chromium.org>
Simon Glass [Sat, 23 Oct 2021 23:26:12 +0000 (17:26 -0600)]
test/py: Raise a ValueError if a command fails
At present an Exception is raised if a command fails. This is a very broad
class and makes it difficult for callers to catch the error without also
catching other things, like programming bugs.
Simon Glass [Sat, 23 Oct 2021 23:26:11 +0000 (17:26 -0600)]
test/py: Relax the naming rules for unit tests
At present the collection function used by pytest is quite strict on the
naming of the functions it detects. In particular it requires the name of
the test to be repeated in the function name.
This is not enforced anywhere else, but instead the tests are silently
omitted from the pytest run. This affects a few dozen tests.
The rule does not seem to have any particular purpose. Relax it, so that
all tests that use the UNIT_TEST() macro will run, regardless of the name
of the test function.
Simon Glass [Sat, 23 Oct 2021 23:26:10 +0000 (17:26 -0600)]
sandbox: Enable HEXDUMP for sandbox_flattree
At present the hexdump tests are disabled in sandbox_flattree. This is
good, because they do not pass. Enable the required Kconfig so that they
will, when enabled.
Simon Glass [Sat, 23 Oct 2021 23:26:08 +0000 (17:26 -0600)]
dm: core: Allow finding children / uclasses by partial name
In some cases it is useful to search just by a partial name, such as
when looking for a sibling device that has a common name substring. Add
helper functions to handle these requirements.
Simon Glass [Sat, 23 Oct 2021 23:26:07 +0000 (17:26 -0600)]
dm: core: Add a way to obtain a string list
At present we support reading a string list a string at a time. Apart
from being inefficient, this makes it impossible to separate reading of
the devicetree into the of_to_plat() method where it belongs, since any
code which needs access to the string must read it from the devicetree.
Add a function which returns the string property as an array of pointers
to the strings, which is easily used by clients.
Simon Glass [Sat, 23 Oct 2021 23:26:05 +0000 (17:26 -0600)]
dm: core: Fix handling of uclass pre_unbind method
This method is currently called after the platform data has been freed.
But the pre_unbind() method may wish to access this, e.g. to free some
data structures stored there.
Split the unbinding of devices into two pieces, as is done with removal.
This corrects the problem.
Also tidy a code-style issue in device_remove() while we are here.
Simon Glass [Sat, 23 Oct 2021 23:26:04 +0000 (17:26 -0600)]
dm: core: Add tests for stringlist functions
These functions currently lack tests so add some. The error handling
differs betwee livetree and flattree at present, so only check the error
codes with livetree.
Simon Glass [Sat, 23 Oct 2021 23:26:03 +0000 (17:26 -0600)]
common: Allow a smaller console-recording pre-reloc
Before relocation there is generally not as much available memory and not
that much console output. At present the console-output buffer is the same
side before and after relocation. Add a separate Kconfig option to remove
this limitation.
Simon Glass [Sat, 23 Oct 2021 23:26:02 +0000 (17:26 -0600)]
disk: part_dos: Fix a NULL pointer error
When ext is NULL we cannot dereference it. Update the code flow to avoid
this, so that layout_mbr_partitions() can be used with partition tables
that do not include an extended partition.
Simon Glass [Sat, 23 Oct 2021 23:26:01 +0000 (17:26 -0600)]
mbr: Correct verification check
At present this command considers the partitions to be identical if the
start and size are smaller than expected. It should check that they are
the same. Fix this and tidy up the code style a little.
Signed-off-by: Simon Glass <sjg@chromium.org> Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Simon Glass [Sat, 23 Oct 2021 23:25:59 +0000 (17:25 -0600)]
sandbox: mmc: Support a backing file
Provide a way for sandbox MMC to present data from a backing file. This
allows a filesystem to be created on the host and easily served via an
emulated mmc device.
Simon Glass [Sun, 19 Sep 2021 21:49:35 +0000 (15:49 -0600)]
sf: doc: Add documentation for the 'sf' command
This command is fairly complicated so documentation is useful.
Unfortunately I an not sure how the MTD side of things works and cannot
find information about that.
Signed-off-by: Simon Glass <sjg@chromium.org> Acked-by: Pratyush Yadav <p.yadav@ti.com> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Acked-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Simon Glass [Sun, 19 Sep 2021 21:49:33 +0000 (15:49 -0600)]
sf: Use const for the stage name
This is not updated at runtime so should be marked const. Update the code
accordingly.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Simon Glass [Sun, 19 Sep 2021 21:49:32 +0000 (15:49 -0600)]
command: Use a constant pointer for the help
This text should never change during execution, so it makes sense to
use a const char * so that it can be declared as const in the code.
Update struct cmd_tbl with a const char * pointer for 'help'.
We cannot make usage const because of the bmode command, used on mx53ppd
for example.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com> Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Tom Rini [Fri, 26 Nov 2021 22:10:53 +0000 (17:10 -0500)]
Merge tag 'efi-2022-01-rc3-2' of https://source.denx.de/u-boot/custodians/u-boot-efi
Pull request for efi-2022-01-rc3-2
Test:
* fix pylint warnings
UEFI:
* disable interrupts before removing devices in ExitBootServices()
* implement poweroff in efi_system_reset() on sandbox
* allow booting via EFI even if some block device fails
Here efi_resister_disk() failed due to an error in the BTRFS implementation.
Move the logic to clear EFI_OS_INDICATIONS_FILE_CAPSULE_DELIVERY_SUPPORTED
to the rest of the capsule code.
If CONFIG_EFI_IGNORE_OSINDICATIONS=y, we should still clear the flag.
If OsIndications does not exist, we should not create it as it is owned by
the operating system.
Our algorithm for creating USB device paths may lead to duplicate device
paths which result in efi_disk_register() failing. Instead we should just
skip devices that cannot be registered as EFI block devices.
Fix a memory leak in efi_disk_add_dev() caused by the duplicate device
path.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Tom Rini [Fri, 19 Nov 2021 21:33:04 +0000 (16:33 -0500)]
efi: Call bootm_disable_interrupts earlier in efi_exit_boot_services
If we look at the path that bootm/booti take when preparing to boot the
OS, we see that as part of (or prior to calling do_bootm_states,
explicitly) the process, bootm_disable_interrupts() is called prior to
announce_and_cleanup() which is where udc_disconnect() /
board_quiesce_devices() / dm_remove_devices_flags() are called from. In
the EFI path, these are called afterwards. In efi_exit_boot_services()
however we have been calling bootm_disable_interrupts() after the above
functions, as part of ensuring that we disable interrupts as required
by the spec. However, bootm_disable_interrupts() is also where we go
and call usb_stop(). While this has been fine before, on the TI J721E
platform this leads us to an exception. This exception seems likely to
be the case that we're trying to stop devices that we have already
disabled clocks for. The most direct way to handle this particular
problem is to make EFI behave like the do_bootm_states() process and
ensure we call bootm_disable_interrupts() prior to ending up in
usb_stop().
Cc: Ilias Apalodimas <ilias.apalodimas@linaro.org> Cc: Heinrich Schuchardt <xypron.glpk@gmx.de> Cc: Simon Glass <sjg@chromium.org> Suggested-by: Ilias Apalodimas <ilias.apalodimas@linaro.org> Signed-off-by: Tom Rini <trini@konsulko.com> Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
This change implements an SCMI transport for agent interfacing the
OP-TEE SCMI service. OP-TEE provides an SCMI PTA (Pseudo-TA) for
non-secure world to send SCMI messages over an identified channel.
The driver implemented here uses a SMT shared memory for passing
messages between client and server.
The implementation opens and releases channel resources for each
passed SCMI message so that resources allocated (sessions) or
registered (shared memory areas) in OP-TEE firmware are released for
example before relocation as the driver will likely allocate/register
them back when probed after relocation.
The integration of the driver using dedicated config switch
CONFIG_SCMI_AGENT_OPTEE is designed on the model posted to the
U-Boot ML by Patrick Delaunay [1].
Link: [1] https://lore.kernel.org/all/20211028191222.v3.4.Ib2e58ee67f4d023823d8b5404332dc4d7e847277@changeid/ Cc: Patrick Delaunay <patrick.delaunay@foss.st.com> Cc: Wolfgang Denk <wd@denx.de> Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org> Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
dt-bindings: arm: scmi: OP-TEE as transport channel for SCMI messages
Introduce compatible "linaro,scmi-optee" for SCMI transport channel
based on an OP-TEE service invocation.
Define "linaro,optee-channel-id" property to identify the OP-TEE SCMI
channel used by the protocol(s). OP-TEE SCMI transport can either use
shared memory or a static shared memory buffer identified by the DT.
These bindings were posted to the Linux kernel DT bindings mailing list
and acked by maintainer [1].
tee: optee: define TEE error code TEE_ERROR_SHORT_BUFFER
Adds TEE_ERROR_SHORT_BUFFER as TEE error code. This error code is
commonly used by TEEs to inform caller that the buffer(s) it provided
is too small for the desired operation.
Define identifiers for clnt_login field in struct tee_open_session_arg
based in GlobalPlatform Device TEE IDs and on the REE_KERNEL identifier
extension from OP-TEE OS.
firmware: scmi: smccc transport: use plat data, not priv data
Change SCMI smccc transport drivers to use platform data rather
than private data for channel reference since it only stores platform
data retrieved from the DT. Consequently the probe handler is replaced
with a of_to_plat handler.
Cc: Simon Glass <sjg@chromium.org> Cc: Patrice Chotard <patrice.chotard@foss.st.com> Cc: Patrick Delaunay <patrick.delaunay@foss.st.com> Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org> Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
firmware: scmi: mailbox transport: use plat data, not priv data
Change SCMI mailbox transport drivers to use platform data rather
than private data for channel reference since it only stores platform
data retrieved from the DT. Consequently the probe handler is replaced
with a of_to_plat handler.
Cc: Simon Glass <sjg@chromium.org> Cc: Patrice Chotard <patrice.chotard@foss.st.com> Cc: Patrick Delaunay <patrick.delaunay@foss.st.com> Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org> Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Samuel Holland [Wed, 13 Oct 2021 02:07:32 +0000 (21:07 -0500)]
net: phy: realtek: Add tx/rx delay config for 8211e
Some boards need to change the tx/rx delay config in order for
gigabit Ethernet to work.
In Linux commit bbc4d71d6354 ("net: phy: realtek: fix rtl8211e rx/tx
delay config"), Realtek documented the bits for overriding the delays
from the hardware straps.
Copy the logic from linux, so the delay config is set from the PHY's
interface type (the phy-mode property in the device tree).
This removes the need for a one-off workaround for the Pine A64+ board.
Signed-off-by: Samuel Holland <samuel@sholland.org> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Peter Hoyes [Thu, 14 Oct 2021 08:40:04 +0000 (09:40 +0100)]
cmd: pxe_utils: Check fdtcontroladdr in label_boot
If using OF_CONTROL, fdtcontroladdr is set to the fdt used to configure
U-Boot. When using PXE, if no fdt is defined in the menu file, and
there is no fdt at fdt_addr, add fall back on fdtcontroladdr too.
We are developing board support for the Armv8r64 FVP using
config_distro_bootcmd. We are also using OF_BOARD and would like the
PXE boot option to default to the fdt provided by board_fdt_blob_setup.
Signed-off-by: Peter Hoyes <Peter.Hoyes@arm.com> Reviewed-by: Simon Glass <sjg@chromium.org> Signed-off-by: Ramon Fried <rfried.dev@gmail.com>
Walter Stoll [Tue, 12 Oct 2021 11:01:59 +0000 (11:01 +0000)]
net: bootp: Correct VCI string transmission
The VCI string sent during bootp of U-Boot-SPL is corrupt. This is
because the byte counter is not adjusted within the bootp_extended()
function when the VCI string is added. We fix this.
Signed-off-by: Walter Stoll <walter.stoll@duagon.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Vladimir Oltean [Wed, 29 Sep 2021 15:04:42 +0000 (18:04 +0300)]
net: dsa: sja1105: add support for SGMII
The list of ports which support SGMII depending on switch generation is
available here:
https://www.kernel.org/doc/html/latest/networking/dsa/sja1105.html#port-compatibility-matrix
SGMII can either be used to connect to an external PHY or to the host
port. In the first case, the use of in-band autoneg is expected, in the
last, in-band autoneg is expected to be turned off (fixed-link). So the
driver supports both cases.
SGMII support means configuring the PCS and PMA. The PCS is a Synopsys
Designware XPCS, in Linux this has a separate driver but here it is
embedded within the sja1105 driver. If needed it can be taken out later,
although we would need a UCLASS_PCS for it, which we don't have atm.
Nonetheless, I did go all the way to export an internal MDIO bus for PCS
access, because it is nice to be able to debug the PCS through commands
such as:
=> mdio read ethernet-switch@1-pcs 4 1f.0
Reading from bus ethernet-switch@1-pcs
PHY at address 4:
31.0 - 0x1140
The internal MDIO bus is not registered with DM because there is no
udevice on it, as mentioned. But the XPCS code can still be ripped out,
as needed.
I did not add support for 2500base-x because I do not expect this
interface type to be used as a boot source for anybody, it would just
add unnecessary bloat.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Vladimir Oltean [Wed, 29 Sep 2021 15:04:41 +0000 (18:04 +0300)]
net: dsa: add driver for NXP SJA1105 L2 switch
The SJA1105 driver is largely reused from Linux. Its programming model
is that it is blank out of reset, and it waits for a static
configuration stream over SPI, which contains all runtime parameters (it
has no notion of "default values").
Keeping a binary array for the configuration stream would have meant
that aspects such as the CPU port and the MAC speeds could have not been
configured easily, and would have been static and board-dependent.
Live-patching the binary array means recalculating the static config
table CRCs, which is not a fun process.
So we create an abstraction over the static config tables, using the
packing API, same as in Linux. The tables are kept as C structures, and
the binary configuration stream is constructed on-the-go, with CRC and
all.
All static config tables instantiated in this driver are mandatory.
The hardware reference manual can be found at:
https://www.nxp.com/docs/en/user-guide/UM10944.pdf
For tagging, a simplified version of tag_8021q from Linux is used. The
VLAN EtherType is the same (0xdadb) but since we don't want switching in
U-Boot, there is no reason to have a TX VLAN and an RX VLAN for each
port. We just need the RX VLANs to act as the unique pvid of each
front-panel port, to decode the switch port number. The RX VLAN is used
for both RX and TX.
The device tree bindings are the same as in Linux.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Vladimir Oltean [Wed, 29 Sep 2021 15:04:40 +0000 (18:04 +0300)]
net: dsa: felix: configure the in-band autoneg property based on OF node info
Instead of trying to guess which operating modes need in-band
negotiation to be active and which ones don't, parse the available
information from the device tree. That will be correct in the cases we
can already guess, and more.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Vladimir Oltean [Wed, 29 Sep 2021 15:04:39 +0000 (18:04 +0300)]
net: introduce a helper to determine whether to use in-band autoneg
Certain serial SERDES protocols like 1000base-x, 2500base-x, SGMII,
USXGMII can operate either in a mode where the PHY (be it on-board or
inside an SFP module) passes the link parameters (speed, duplex, pause)
to the MAC through in-band through control words standardized by IEEE
802.3 clause 37, or in a mode where the MAC must configure (force) its
link parameters based on information obtained out-of-band (MDIO reads,
guesswork etc).
In Linux, the OF node property named "managed" is parsed by the phylink
framework, and the convention is that if a driver uses phylink, then the
presence of this property means that in-band autoneg should be enabled,
otherwise it shouldn't.
To be compatible with the OF node bindings of drivers that use phylink
in Linux, introduce parsing support for this property in U-Boot too.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com> Reviewed-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Vladimir Oltean [Wed, 29 Sep 2021 15:04:38 +0000 (18:04 +0300)]
net: dsa: allow drivers to get the port OF node
In the current DSA switch driver API, only the udevice of the switch
(belonging to UCLASS_DSA) is exposed, as well as an "int port" argument.
So drivers do not have access to the udevice of individual ports
(belonging to UCLASS_ETH), one of the reasons being that not all ports
have an associated UCLASS_ETH udevice.
However, all DSA ports have an OF node, and in some cases the driver
needs a handle to it, for all ports including the CPU port. Example: the
following Linux per-port device tree property:
managed = "in-band-status";
states whether a port should operate with clause 37 in-band autoneg
enabled or not.
This patch exposes a function which can be called by individual drivers
as needed.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Vladimir Oltean [Wed, 29 Sep 2021 15:04:36 +0000 (18:04 +0300)]
net: tsec: add support for promiscuous mode
The Freescale TSEC can be a DSA master, and the ports of the attached
DSA switch can have different MAC addresses compared to the TSEC.
Nonetheless, the TSEC must receive the packets on behalf of those switch
ports. Therefore, implement the promiscuous mode method to allow DSA to
set this.
Note that the init_registers() function called from eth_ops :: start
overwrites this setting. There is no reason why the RCTRL register
should be zero-initialized, so just stop clearing it so that the setting
we applied in eth_ops :: set_promisc sticks.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com> Reviewed-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Vladimir Oltean [Mon, 27 Sep 2021 23:13:42 +0000 (02:13 +0300)]
net: phy: mscc: add support for VSC8502 in dual RGMII mode
The VSC8502 is a Microchip (formerly Microsemi, formerly Vitesse)
dual port, gigabit Ethernet copper PHY which supports the MII, GMII and
RGMII MAC-side interfaces.
Of these, I could only test RGMII, and my board needed RGMII delays to
be applied by software, so I am able to confirm that this patch handles
that properly.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Tom Rini [Sat, 20 Nov 2021 14:36:37 +0000 (09:36 -0500)]
Merge branch 'efi-2022-01' of https://source.denx.de/u-boot/custodians/u-boot-efi
Scripts:
* Update spelling.txt
LMB:
* remove extern keyword in lmb.h
* drop unused lmb_size_bytes()
Test:
* test truncation in snprintf()
Documentation:
* add include/lmb.h to HTML documentation
UEFI:
* reduce non-debug logging output for measured boot
* fix use after free in measured boot
* startup the tpm device when installing the protocol
* implement EFI_EVENT_GROUP_BEFORE_EXIT_BOOT_SERVICES
* record capsule result only if capsule is read
Ilias Apalodimas [Thu, 18 Nov 2021 08:13:42 +0000 (10:13 +0200)]
efi_loader: startup the tpm device when installing the protocol
Due to U-Boot's lazy binding mentality the TPM is probed but not properly
initialized. The user can startup the device from the command line
e.g 'tpm2 startup TPM2_SU_CLEAR'. However we can initialize the TPM during
the TCG protocol installation, which is easier to use overall.
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org> Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Ilias Apalodimas [Thu, 18 Nov 2021 07:03:39 +0000 (09:03 +0200)]
efi_loader: fix FinalEvents table if an EFI app invoked GetEventLog
As described in the TCG spec [1] in sections 7.1.1 and 7.1.2 the FinalEvent
table should include events after GetEventLog has been called. This
currently works for us as long as the kernel is the only EFI application
calling that. Specifically we only implement what's described in 7.1.1.
So refactor the code a bit and support EFI application(s) calling
GetEventLog. Events will now be logged in both the EventLog and FinalEvent
table as long as ExitBootServices haven't been invoked.
Masami Hiramatsu [Fri, 12 Nov 2021 13:05:15 +0000 (22:05 +0900)]
efi_loader: capsule: Record capsule result only if capsule is read
Record capsule update result only if the capsule file is
successfully read, because the capsule GUID is not sure when
the file can not be read or the file is not a capsule.
Without this fix, if user puts a dummy (non-capsule) file
under (ESP)EFI/UpdateCapsule, U-Boot causes a synchronous
abort.
This also fixes use-after-free bug of the 'capsule' variable.
Masahisa Kojima [Tue, 9 Nov 2021 11:35:53 +0000 (20:35 +0900)]
efi_loader: Reduce efi_tcg2 logging statement
log_info() is used for the debug level logging statement
which should use log_debug() instead. Convert it to reduce the
log output.
Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org> Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Masahisa Kojima [Fri, 12 Nov 2021 07:24:27 +0000 (16:24 +0900)]
efi: add comment for efi_system_table and efi_configuration_table
This commit adds the comment for efi_system_table and
efi_configuration_table structure.
Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>