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>
This reverts commit 38d6b7ebda.
struct global_data contains a pointer to the bd_info structure. This
pointer was populated spl_set_bd() to a pre-allocated bd_info in the
".data" section. The referenced commit replaced this mechanism to one
that uses malloc(). That new mechanism is only used if SPL_ALLOC_BD=y.
which very few boards do.
The result is that (struct global_data)->bd is NULL in SPL on most
platforms. This breaks falcon mode, since arch_fixup_fdt() tries to
access (struct global_data)->bd and set the "/memory" node in the
devicetree. The result is that the "/memory" node contains garbage
values, causing linux to panic() as it sets up the page table.
Instead of trying to fix the mess, potentially causing other issues,
revert to the code that worked, while this change is reworked.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
OP-TEE images are normally packaged with
type = "tee;
os = "tee";
However, fit_image_load() thinks that is somehow invalid. However if
they were declared as type = "kernel", os = "linux", fit_image_load()
would happily accept them and allow the boot to continue. There is no
technical limitation to excluding "tee".
Allowing "tee" images is useful in a boot flow where OP-TEE is
executed before linux.
In fact, I think it's unintuitive for a "load"ing function to also do
parsing and contain a bunch ad-hoc heuristics that only its caller
might know. But I don't make the rules, I just write fixes. In more
polite terms: refactoring the fit_image API is beyond the scope of
this change.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Consider the following FIT:
images {
whipple {};
};
configurations {
conf-1 {
firmware = "whipple";
};
};
Getting the 'firmware' image with fit_image_load() is not possible, as
it doesn't understand 'firmware =' properties. Although one could pass
IH_TYPE_FIRMWARE for 'image_type', this needs to be converted to a
"firmware" string for FDT lookup -- exactly what this change does.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The 'firmware' property of a config node takes precedence over the
'kernel' property. 'standalone' is deprecated. However, give users a
couple of releases where 'standalone' still works, but warns loudly.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
U-Boot expects the FDT to be located right after the _end
linker symbol (see fdtdec.c: board_fdt_blob_setup())
The "basic" LOAD_FIT path is aware of this limitation, and relocates
the FDT at the expected location. Guessing the expected location
probably only works reliably on 32-bit arm, and it feels like a hack.
One proposal would be to pass the FDT address to u-boot
(e.g. using 'r2' on arm platforms).
The variable is named "fdt_hack" to remind future contributors that,
"hey! we should fix the underlying problem". However, that is beyond
the scope of this patch.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
The information on the OS should be contained in the FIT, as the
self-explanatory "os" property of a node under /images. Hard-coding
this to U_BOOT might send us down the wrong path later in the boot
process.
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
The correct FDT to use is described by the "fdt" property of the
configuration node. When the fit_unamep argument to fit_image_load()
is "fdt", we get the "/images/fdt" node. This is incorrect, as it
ignores the "fdt" property of the config node, and in most cases,
the "/images/fdt" node doesn't exist.
Use NULL for the 'fit_unamep' argument. With NULL, fit_image_load()
uses the IH_TYPE_FLATDT value to read the config property "fdt",
which points to the correct FDT node(s).
fit_image_load() should probably be split into a function that reads
an image by name, and one that reads an image by config reference. I
don't make those decisions, I just point out the craziness.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Several architectures had a default board_fit_config_name_match already;
this provides a generic weak version. We default to rejecting all configs.
This will use the FIT's default config, instead of the first config. This
may result in boot failures if there are multiple configurations and the
first config is *not* the default.
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
The "simple" SPL_LOAD_FIT path is the most compliant with the format
documented in doc/uImage.FIT/source_file_format.txt. The other two
paths to load a FIT are SPL_LOAD_FIT_FULL and the "bootm" command.
Since the Kconfig menu is the most likely place for a new user to see
these options, it seems like the most logical candidate to document
the limitations. This documents the _known_ issues, and is not
intended to be a complete list of all follies.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Commit 4afc4f37c7 ("doc: FIT image: Clarify format and simplify
syntax") and delegated FPGA images to be added via the list of
"loadables" in lieu of the "fpga" property. Now actually implement
this in code.
Note that the "compatible" property is ignored for the time being, as
implementing "compatible" loading is beyond the scope of this change.
However, "u-boot,fpga-legacy" is accepted without warning.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Commit 4afc4f37c7 ("doc: FIT image: Clarify format and simplify
syntax") requires that FPGA images be referenced through the
"loadables" in the config node. This means that "fpga" properties in
config nodes are deprecated.
Given that there are likely FIT images which use "fpga", let's not
break those right away. Print a warning message that such use is
deprecated, and give users a couple of releases to update their
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
The FPGA loading code in spl_simple_fit_read() can easily be separated
from the rest of the logic. It is split into two functions instead of
one because spl_fit_upload_fpga() is used in a subsequent patch.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
spl_load_fit_image() will try to load an image at the address given
in the "load" property. Absent such property, it uses
image_info->load_addr
Correct use of this is demonstrated in spl_fit_append_fdt(), which
resets the 'load_addr' before each spl_load_fit_image() call.
On the other hand loading "loadables" loop in spl_load_simple_fit()
completely ignores this. It re-uses the same structure, but doesn't
reset load_addr. If loadable [i] does not have a "load" property, its
load address defaults to load_addr, which still contains the address
of loadable [i - 1].
A simple solution is to treat NULL as an invalid load address. The
caller can set load_addr = 0 to request an abort if the "load"
property is absent.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
mkimage supports rsa2048, and rsa4096 signatures. With newer silicon
now supporting hardware-accelerated ECDSA, it makes sense to expand
signing support to elliptic curves.
Implement host-side ECDSA signing and verification with libcrypto.
Device-side implementation of signature verification is beyond the
scope of this patch.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
rsa-checksum.c sontains the hash_calculate() implementations. Despite
the "rsa-" file prefix, this function is useful for other algorithms.
To prevent confusion, move this file to lib/, and rename it to
hash-checksum.c, to give it a more "generic" feel.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
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>
If there were no variable substitutions in a command, then initial
assignments would be misinterpreted as commands, instead of being skipped
over. This is demonstrated by the following example:
=> foo=bar echo baz
Unknown command 'foo=bar' - try 'help'
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Similar to support for SHA1 and SHA256, allow the use of hardware hashing
engine by enabling the algorithm and setting CONFIG_SHA_HW_ACCEL /
CONFIG_SHA_PROG_HW_ACCEL.
Signed-off-by: Joel Stanley <joel@jms.id.au>
The optee_copy_fdt_nodes is only used to copy op-tee nodes
of U-Boot device tree (from gd->fdt_blob when OF_LIVE is not activated)
to external device tree but it is not compatible with OF_LIVE.
This patch migrates all used function fdt_ functions to read node on
old_blob to ofnode functions, compatible with OF_LIVE and remove this
parameter "old_blob".
The generated "device tree" is checked on stm32mp platform with OF_LIVE
activated.
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Update the debug trace for the reserved video memory to KB as indicated
in the message with "%luk"; before the patch the computed size
gd->relocaddr - addr is in bytes.
This patch aligns the debug trace in reserve_video() with others
functions, for example on stm32mp157c-dk2:
- Reserving 3080192k for video at: dfd00000
+ Reserving 3008k for video at: dfd00000
Reserving 873k for U-Boot at: dfc25000
Reserving 32776k for malloc() at: ddc23000
Reserving 72 Bytes for Board Info at: ddc22fb0
Reserving 280 Bytes for Global Data at: ddc22e90
Reserving 119072 Bytes for FDT at: ddc05d70
Reserving 0x278 Bytes for bootstage at: ddc05af0
Fixes: 5630d2fbc5 ("board: Show memory for frame buffers")
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Add a weak nand_get_mtd function for nand drivers to provide mtd info
and use this to set pagesize such that reading of non page-aligned
elements can succeed.
The spl_load_simple_fit already handles block block access so all we
need to do is provide the nand writesize as the block length.
Further cleanup of the drivers which use nand_spl_loaders.c such as
am335x_spl_bch.c, atmel_nand.c, and nand_spl_simple.c could be done
using info from mtd_info instead of statically defined details.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
commit 9f6a14c47f ("spl: fit: nand: fix fit loading in case of bad blocks")
added support for adjusting the image offset to account for bad blocks.
However this requires nand_spl_adjust_offset() which requires fully defined
specifics of the NAND chip being used may not be avialable.
Allow skipping this support for drivers or configs which don't specify
the NAND chip details statically with defines.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
In preparation to add SiFive Unmatched board support, let's rename
the existing fu540 board to unleashed.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
Add missing endianness conversions to usb_get_port_status(). This
(amongst others) is necessary to enable the use of USB 3 hubs on
big-endian platforms like MIPS Octeon.
Signed-off-by: Aaron Williams <awilliams@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Chandrakala Chavva <cchavva@marvell.com>
Cc: Bin Meng <bmeng.cn@gmail.com>
Cc: Marek Vasut <marex@denx.de>
Some boards may want to show the SKU ID or other information obtained at
runtime. Allow this to come from sysinfo. The board can then provide a
sysinfo driver to provide it.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-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>
At present bootstage silently ignores new records if it runs out of
space. It is sometimes obvious by looking at the report, but the IDs are
not contiguous, so it is easy to miss.
Aad a message so that action can be taken.
Signed-off-by: Simon Glass <sjg@chromium.org>
We have two separate places that need to figure out the bootstage ID to
use. Put this code in a function so that the logic is in one place.
Signed-off-by: Simon Glass <sjg@chromium.org>
Sandbox is special in that it is used for testing and it does not match
any particular target architecture. Allow it to load an image from any
architecture, so that 'bootm' can be used as needed.
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>
part_get_info_by_name will return -1 on error, and >0 on success.
Signed-off-by: schspa <schspa@gmail.com>
Cc: Igor Opaniuk <igor.opaniuk@gmail.com>
Reviewed-by: Igor Opaniuk <igor.opaniuk@gmail.com>
Unfortunately -ENODATA is not available in OpenBSD. Use -EBADMSG
instead, to indicate a missing timestamp.
Fixes: c5819701a3 image: Adjust the workings of fit_check_format()
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Mark Kettenis <kettenis@openbsd.org>
This Trusted Application allows enabling SCP03 as well as provisioning
the keys on TEE controlled secure element (ie, NXP SE050).
All the information flowing on buses (ie I2C) between the processor
and the secure element must be encrypted. Secure elements are
pre-provisioned with a set of keys known to the user so that the
secure channel protocol (encryption) can be enforced on the first
boot. This situation is however unsafe since the keys are publically
available.
For example, in the case of the NXP SE050, these keys would be
available in the OP-TEE source tree [2] and of course in the
documentation corresponding to the part.
To address that, users are required to rotate/provision those keys
(ie, generate new keys and write them in the secure element's
persistent memory).
For information on SCP03, check the Global Platform HomePage and
google for that term [1]
[1] globalplatform.org
[2] https://github.com/OP-TEE/optee_os/
check:
core/drivers/crypto/se050/adaptors/utils/scp_config.c
Signed-off-by: Jorge Ramirez-Ortiz <jorge@foundries.io>
Reviewed-by: Simon Glass <sjg@chromium.org>
When multiple log() calls are used which don't end in newline, the
log prefix is prepended multiple times in the same line. This makes the
output look strange.
Fix this by detecting when the previous log record did not end in newline.
In that case, setting a flag.
Drop the unused BUFFSIZE in the test while we are here.
As an example implementation, update log_console to check the flag and
produce the expected output.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present only a single flag (force_debug) is used in log records. Before
adding more, convert this into a bitfield, so more can be added without
using more space.
To avoid expanding the log_record struct itself (which some drivers may
wish to store in memory) reduce the line-number field to 16 bits. This
provides for up to 64K lines which should be enough for anyone.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present this casts addresses to pointers so cannot work with sandbox.
Update the code to use map_sysmem() instead.
As part of this change, the existing load_ptr is renamed to src_ptr since
it is not a pointer to load_addr. It is confusing to use a similar name
for something that is not actually related. For the alignment code,
ALIGN() is used instead of open-coded alignment. Add a comment to the line
that casts away a const.
Use a (new) load_ptr variable to access memory at address load_addr.
Signed-off-by: Simon Glass <sjg@chromium.org>
Vendor Authorized Boot is a security feature for authenticating
the images such as U-Boot, ARM trusted Firmware, Linux kernel,
device tree blob and etc loaded from FIT. After those images are
loaded from FIT, the VAB certificate and signature block appended
at the end of each image are sent to Secure Device Manager (SDM)
for authentication. U-Boot will validate the SHA384 of the image
against the SHA384 hash stored in the VAB certificate before
sending the image to SDM for authentication.
Signed-off-by: Siew Chin Lim <elly.siew.chin.lim@intel.com>
Reviewed-by: Ley Foon Tan <ley.foon.tan@intel.com>
At present initr_addr_map() is put at a late stage in the
init_sequence_r[] calls. This won't work because lot of
device driver initialization (e.g.: serial port) happens
before it but is lack of the address translation support.
This moves the call to a bit earlier, right after the DM
initialization.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
At present fdt_read_prop() can only handle 1 or 2 cells. It is
called by fdt_read_range() which may be used to read PCI address
from <ranges> for a PCI bus node where the number of PCI address
cell is 3. The <ranges> property is an array of:
{ <child address> <parent address> <size in child address space> }
When trying to read <child address> from a PCI bus node using
fdt_read_prop(), as the codes below:
/* Read <child address> */
if (child_addr) {
r = fdt_read_prop(ranges, ranges_len, cell, child_addr,
acells);
if (r)
return r;
}
it will fail, because the PCI child address is made up of 3 cells
but fdt_read_prop() cannot handle it. We advance the cell offset
by 1 so that the <child address> can be correctly read.
This adds the special handling of such case.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
At present these three Kconfigs exist even when bootstage is not enabled.
This is not necessary since bootstage.c is only built if BOOTSTAGE is
enabled.
Make them conditional. Also fix up the overflow message to mention TPL.
Signed-off-by: Simon Glass <sjg@chromium.org>
This is printed twice but we only need one message, since there is very
little processing in between them. Drop the second one, since all branches
of the switch() already have output. Update the U-Boot message to include
the phase being jumped to.
Signed-off-by: Simon Glass <sjg@chromium.org>
Some compression algorithms currently can be enabled for SPL and
U-Boot proper separately. Therefore take into account USE_HOSTCC
is well as CONFIG_IS_ENABLED() in these cases to prevent compiling these
functions in case of a host tool build.
Signed-off-by: Frieder Schrempf <frieder.schrempf@kontron.de>
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
