Masahiro Yamada [Mon, 3 Feb 2020 10:46:26 +0000 (19:46 +0900)]
uniphier: make on-chip SRAM region configurable
The on-chip SRAM region will be changed in the next SoC. Make it
configurable. Also, split the mmap code into a new helper function
so that it can be re-used for another boot mode.
Masahiro Yamada [Mon, 3 Feb 2020 09:40:37 +0000 (18:40 +0900)]
uniphier: change block_addressing flag to bool
The flag, uniphier_emmc_block_addressing, is boolean logic, so
"bool' is more suitable.
uniphier_emmc_is_over_2gb() is not boolean - it returns 1 / 0
depending on the card density, or a negative value on failure.
Rename it to make it less confusing.
Paul Beesley [Thu, 16 May 2019 12:33:18 +0000 (13:33 +0100)]
doc: Split and expand coding style documentation
This patch expands the coding style documentation, splitting it
into two documents: the core style rules and extended guidelines.
Note that it does not redefine or change the coding style (aside
from section 4.6.2) - generally, it is only documenting the
existing style in more detail.
The aim is for the coding style to be more readable and, in turn,
for it to be followed by more people. We can use this as a more
concrete reference when discussing the accepted style with external
contributors.
Change-Id: I87405ace9a879d7f81e6b0b91b93ca69535e50ff Signed-off-by: Paul Beesley <paul.beesley@arm.com> Signed-off-by: Petre-Ionut Tudor <petre-ionut.tudor@arm.com>
The checks on size and offset_address in get_entry always resolve to
false provided those fields are long long int and cannot be greater
than LONG_MAX.
When Trusted Boot is enabled, images are loaded and authenticated
following up the root of trust. This means that between the initial
console message saying that an image is being loaded, and the final one
where it says that it failed to load it, BL2 may print several messages
about other images on the chain of trust being loaded, thus it is not
always clear which image we failed loading at the end of the day.
Imre Kis [Mon, 3 Feb 2020 13:48:21 +0000 (14:48 +0100)]
doc: Remove backquotes from external hyperlinks
Since Sphinx 2.3.0 backquotes are replaced to \textasciigrave{} during
building latexpdf. Using this element in a \sphinxhref{} breaks the
build. In order to avoid this error backquotes must not be used in
external hyperlinks.
Signed-off-by: Imre Kis <imre.kis@arm.com>
Change-Id: Ie3cf454427e3d5a7b7f9829b42be45aebda7f0dd
Alexei Fedorov [Wed, 29 Jan 2020 16:21:28 +0000 (16:21 +0000)]
FDT wrappers: add functions for read/write bytes
This patch adds 'fdtw_read_bytes' and 'fdtw_write_inplace_bytes'
functions for read/write array of bytes from/to a given property.
It also adds 'fdt_setprop_inplace_namelen_partial' to jmptbl.i
files for builds with USE_ROMLIB=1 option.
Masahiro Yamada [Thu, 26 Dec 2019 04:26:49 +0000 (13:26 +0900)]
doc: qemu: fix and update documentation
The current URL for QEMU_EFI.fd is not found. Update the link to
point to the new one.
If you run the shell command as instructed, you will see this error:
qemu-system-aarch64: keep_bootcon: Could not open 'keep_bootcon': No such file or directory
The part "console=ttyAMA0,38400 keep_bootcon root=/dev/vda2" is the
kernel parameter, so it must be quoted.
As of writing, QEMU v4.2.0 is the latest, but it does not work for
TF-A (It has been fixed in the mainline.) QEMU v4.1.0 works fine.
With those issues addressed, I succeeded in booting the latest kernel.
Tested with QEMU v4.1.0 and Linux 5.5 (defconfig with no modification).
Update the tested versions.
Varun Wadekar [Fri, 25 May 2018 23:17:53 +0000 (16:17 -0700)]
Tegra194: mce: fix multiple MISRA issues
This patch fixes violations of the following MISRA rules
* Rule 8.5 "An external object or function shall be declared once in
one and only one file"
* Rule 10.3 "The value of an expression shall not be assigned to an
object with a narrower essential type or of a different
esential type category"
Varun Wadekar [Fri, 25 May 2018 21:34:53 +0000 (14:34 -0700)]
Tegra: bpmp: fix multiple MISRA issues
This patch fixes violations for the following MISRA rules
* Rule 5.7 "A tag name shall be a unique identifier"
* Rule 10.1 "Operands shall not be of an inappropriate essential type"
* Rule 10.3 "The value of an expression shall not be assigned to an object
with a narrower essential type or of a different essential type
category"
* Rule 10.4 "Both operands of an operator in which the usual arithmetic
conversions are performed shall have the same essential type
category"
* Rule 20.7 "Expressions resulting from the expansion of macro parameters
shall be enclosed in parentheses"
* Rule 21.1 "#define and #undef shall not be used on a reserved identifier
or reserved macro name"
Varun Wadekar [Fri, 25 May 2018 22:22:58 +0000 (15:22 -0700)]
Tegra194: se: fix multiple MISRA issues
This patch fixes violations for the following MISRA rules
* Rule 8.4 "A compatible declaration shall be visible when an object or
function with external linkage is defined"
* Rule 10.1 "Operands shall not be of an inappropriate essential type"
* Rule 10.6 "Both operands of an operator in which the usual arithmetic
conversions are perdormed shall have the same essential type
category"
* Rule 17.7 "The value returned by a function having non-void return
type shall be used"
Varun Wadekar [Thu, 17 May 2018 18:10:13 +0000 (11:10 -0700)]
Tegra: compile PMC driver for Tegra132/Tegra210 platforms
The PMC driver is used only by Tegra210 and Tegra132 platforms. This
patch removes pmc.c from the common makefile and moves it to the
platform specific makefiles.
As a result, the PMC code from common code has been moved to Tegra132
and Tegra210 platform ports.
Varun Wadekar [Thu, 17 May 2018 16:36:38 +0000 (09:36 -0700)]
Tegra: remove weakly defined platform setup handlers
This patch converts the weakly defined platform setup handlers into
actual platform specific handlers to improve code coverage numbers
and some MISRA defects.
The weakly defined handlers never get executed thus resulting in
lower coverage - function, function calls, statements, branches
and pairs.
Varun Wadekar [Tue, 15 May 2018 18:24:59 +0000 (11:24 -0700)]
Tegra: per-SoC DRAM base values
Tegra194 supports upto 64GB of DRAM, whereas the previous SoCs support
upto 32GB DRAM. This patch moves the common DRAM base/end macros to
individual Tegra SoC headers to fix this anomaly.
Move all TBBR-specific stuff out of the tool's makefile into a
sub-makefile. This will make it easier to define and select an alternate
chain of trust in the future.
This flag warns if anything is declared more than once in the same
scope, even in cases where multiple declaration is valid and changes
nothing.
Consequently, this patch also fixes the issues reported by this
flag. Consider the following two lines of code from two different source
files(bl_common.h and bl31_plat_setup.c):
The IMPORT_SYM macro which actually imports a linker symbol as a C expression.
The macro defines the __RO_START__ as an extern variable twice, one for each
instance. __RO_START__ symbol is defined by the linker script to mark the start
of the Read-Only area of the memory map.
Essentially, the platform code redefines the linker symbol with a different
(relevant) name rather than using the standard symbol. A simple solution to
fix this issue in the platform code for redundant declarations warning is
to remove the second IMPORT_SYM and replace it with following assignment
static const unsigned long BL2_RO_BASE = BL_CODE_BASE;
Soby Mathew [Tue, 28 Jan 2020 10:43:36 +0000 (10:43 +0000)]
Merge changes from topic "ti-cluster-power" into integration
* changes:
ti: k3: drivers: ti_sci: Put sequence number in coherent memory
ti: k3: drivers: ti_sci: Remove indirect structure of const data
ti: k3: common: Enable ARM cluster power down
ti: k3: common: Rename device IDs to be more consistent
In order to support SEPARATE_NOBITS_REGION for Arm platforms, we need to load
BL31 PROGBITS into secure DRAM space and BL31 NOBITS into SRAM. Hence mandate
the build to require that ARM_BL31_IN_DRAM is enabled as well.
Naturally with SEPARATE_NOBITS_REGION enabled, the BL31 initialization code
cannot be reclaimed to be used for runtime data such as secondary cpu stacks.
Memory map for BL31 NOBITS region also has to be created.
Changes necessary to support SEPARATE_NOBITS_REGION feature
Since BL31 PROGBITS and BL31 NOBITS sections are going to be
in non-adjacent memory regions, potentially far from each other,
some fixes are needed to support it completely.
1. adr instruction only allows computing the effective address
of a location only within 1MB range of the PC. However, adrp
instruction together with an add permits position independent
address of any location with 4GB range of PC.
2. Since BL31 _RW_END_ marks the end of BL31 image, care must be
taken that it is aligned to page size since we map this memory
region in BL31 using xlat_v2 lib utils which mandate alignment of
image size to page granularity.
Andrew F. Davis [Thu, 16 Jan 2020 21:58:34 +0000 (15:58 -0600)]
ti: k3: drivers: ti_sci: Put sequence number in coherent memory
The current message sequence number is accessed both with caches on and
off so put this memory in the un-cached coherent section so accesses
are consistent and coherency is maintained.
Signed-off-by: Andrew F. Davis <afd@ti.com>
Change-Id: Ieeefefeaffc691e4e4c4de7c74490d50ff9de807
Andrew F. Davis [Mon, 11 Feb 2019 22:12:31 +0000 (16:12 -0600)]
ti: k3: common: Enable ARM cluster power down
When all cores in a cluster are powered down the parent cluster can
be also powered down. When the last core has requested powering down
follow by sending the cluster power down sequence to the system
power controller firmware.
Signed-off-by: Andrew F. Davis <afd@ti.com>
Change-Id: I0ffeb339852c66ef62743aecd3e17ca20bad6216
T589: Fix insufficient ordering guarantees in bakery lock
bakery_lock_get() uses DMB LD after lock acquisition and
bakery_lock_release() uses DMB ST before releasing the lock. This is
insufficient in both cases. With just DMB LD, stores in the critical
section can be reordered before the DMB LD which could mean writes in
the critical section completing before the lock has been acquired
successfully. Similarly, with just DMB ST, a load in the critical section
could be reordered after the the DMB ST. DMB is the least expensive
barrier that can provide the required ordering.
Soby Mathew [Mon, 27 Jan 2020 17:01:07 +0000 (17:01 +0000)]
Merge changes from topic "pie" into integration
* changes:
uniphier: make all BL images completely position-independent
uniphier: make uniphier_mmap_setup() work with PIE
uniphier: pass SCP base address as a function parameter
uniphier: set buffer offset and length for io_block dynamically
uniphier: use more mmap_add_dynamic_region() for loading images
bl_common: add BL_END macro
uniphier: turn on ENABLE_PIE
TSP: add PIE support
BL2_AT_EL3: add PIE support
BL31: discard .dynsym .dynstr .hash sections to make ENABLE_PIE work
PIE: pass PIE options only to BL31
Build: support per-BL LDFLAGS
Manish Pandey [Fri, 24 Jan 2020 11:54:44 +0000 (11:54 +0000)]
Neovers N1: added support to update presence of External LLC
CPUECTLR_EL1.EXTLLC bit indicates the presense of internal or external
last level cache(LLC) in the system, the reset value is internal LLC.
To cater for the platforms(like N1SDP) which has external LLC present
introduce a new build option 'NEOVERSE_N1_EXTERNAL_LLC' which can be
enabled by platform port.
plat/arm/sgi: move topology information to board folder
The platform topology description of the upcoming Arm's RD platforms
have different topology than those listed in the sgi_topology.c file. So
instead of adding platform specific topology into existing
sgi_topology.c file, those can be added to respective board files. In
order to maintain consistency with the upcoming platforms, move the
existing platform topology description to respective board files.
Deepika Bhavnani [Fri, 13 Dec 2019 16:53:56 +0000 (10:53 -0600)]
xilinx: Unify Platform specific defines for PSCI module
PLATFORM_CORE_COUNT - Unsigned int
PLATFORM_CLUSTER_COUNT - Unsigned int
PLATFORM_MAX_CPUS_PER_CLUSTER - Unsigned int
PLATFORM_CORE_COUNT_PER_CLUSTER - Unsigned int
Masahiro Yamada [Fri, 17 Jan 2020 04:46:48 +0000 (13:46 +0900)]
uniphier: make all BL images completely position-independent
This platform supports multiple SoCs. The next SoC will still keep
quite similar architecture, but the memory base will be changed.
The ENABLE_PIE improves the maintainability and usability. You can reuse
a single set of BL images for other SoC/board without re-compiling TF-A
at all. This will also keep the code cleaner because it avoids #ifdef
around various base addresses.
By defining ENABLE_PIE, BL2_AT_EL3, BL31, and BL32 (TSP) are really
position-independent now. You can load them anywhere irrespective of
their link address.
Masahiro Yamada [Fri, 17 Jan 2020 04:46:38 +0000 (13:46 +0900)]
uniphier: make uniphier_mmap_setup() work with PIE
BL2_BASE, BL31_BASE, and BL32_BASE are defined in platform_def.h,
that is, determined at link-time.
On the other hand, BL2_END, BL31_END, and BL32_END are derived from
the symbols produced by the linker scripts. So, they are fixed-up
at run-time if ENABLE_PIE is enabled.
To make it work in a position-indepenent manner, use BL_CODE_BASE and
BL_END, both of which are relocatable.
Masahiro Yamada [Fri, 17 Jan 2020 04:45:47 +0000 (13:45 +0900)]
bl_common: add BL_END macro
Currently, the end address macros are defined per BL, like BL2_END,
BL31_END, BL32_END. They are not handy in the common code shared
between multiple BL stages.
This commit introduces BL_END, which is equivalent to BL{2,31,32}_END,
and will be useful for the BL-common code.
Masahiro Yamada [Fri, 17 Jan 2020 04:44:50 +0000 (13:44 +0900)]
BL31: discard .dynsym .dynstr .hash sections to make ENABLE_PIE work
When I tried ENABLE_PIE for my PLAT=uniphier platform, BL31 crashed
at its entry. When it is built with ENABLE_PIE=1, some sections are
inserted before the executable code.
$ make PLAT=uniphier CROSS_COMPILE=aarch64-linux-gnu- ENABLE_PIE=1 bl31
$ aarch64-linux-gnu-objdump -h build/uniphier/release/bl31/bl31.elf | head -n 13
build/uniphier/release/bl31/bl31.elf: file format elf64-littleaarch64
The previous stage loader generally jumps over to the base address of
BL31, where no valid instruction exists.
I checked the linker script of Linux (arch/arm64/kernel/vmlinux.lds.S)
and U-Boot (arch/arm/cpu/armv8/u-boot.lds), both of which support
relocation. They simply discard those sections.
Masahiro Yamada [Fri, 17 Jan 2020 04:44:37 +0000 (13:44 +0900)]
PIE: pass PIE options only to BL31
docs/getting_started/build-options.rst clearly says ENABLE_PIE is
currently only supported in BL31, but in fact, it has a stronger
limitation:
Defining ENABLE_PIE may corrupt BL1 and BL2. So, ENABLE_PIE is
supported only for platforms where BL31 is the only image built
in the TF-A tree.
Currently, ENABLE_PIE is enabled by two platforms,
plat/arm/common/arm_common.mk and ti/k3/common/plat_common.mk,
both of which enable ENABLE_PIE together with RESET_TO_BL31.
For platforms with the full boot sequence, ENABLE_PIE may break earlier
BL stages. For example, if I build PLAT=qemu with ENABLE_PIE=1, it
fails in BL1.
When ENABLE_PIE is enabled, PIE options are added to TF_CFLAGS and
TF_LDFLAGS, so all BL images are affected. It is problematic because
currently only the BL31 linker script handles it. Even if BL1/BL2
works, the image size would increase needlessly, at least.
Pass the PIE options only to BL images that support it.
Masahiro Yamada [Fri, 17 Jan 2020 04:44:20 +0000 (13:44 +0900)]
Build: support per-BL LDFLAGS
make_helpers/build_macros.mk supports per-BL CFLAGS. For example,
you can pass compiler flags only to BL31 by using BL31_CFLAGS.
This commit adds per-BL LDFLAGS support, which is useful as well.
My main motivation of this addition is to use it for ENABLE_PIE.
When ENABLE_PIE is enabled, some linker flags are added to TF_LDFLAGS,
which affects all the TF images. It will make more sense to pass the
relevant options only to BL images that support it.
Deepika Bhavnani [Fri, 13 Dec 2019 16:53:34 +0000 (10:53 -0600)]
ti: Unify Platform specific defines for PSCI module
PLATFORM_CORE_COUNT - Unsigned int
PLATFORM_CLUSTER_COUNT - Unsigned int
PLATFORM_MAX_CPUS_PER_CLUSTER - Unsigned int
PLATFORM_CORE_COUNT_PER_CLUSTER - Unsigned int
Deepika Bhavnani [Fri, 13 Dec 2019 16:53:12 +0000 (10:53 -0600)]
st: Unify Platform specific defines for PSCI module
PLATFORM_CORE_COUNT - Unsigned int
PLATFORM_CLUSTER_COUNT - Unsigned int
PLATFORM_MAX_CPUS_PER_CLUSTER - Unsigned int
PLATFORM_CORE_COUNT_PER_CLUSTER - Unsigned int
projects / arm-tf.git / log