C99 standard: "What constitutes an access to an object that has
volatile-qualified type is implementation-defined".
GCC is not considering the cast to void of volatile structures as an
access and so is not actually issuing reads.
Clang does read those structures by copying them on the stack, which in
this case creates an overflow because of their large size.
This patch removes the cast to void and instead uses the USED attribute
to tell the compiler to retain the static variables.
Change-Id: I952b5056e3f6e91841e7ef9558434352710ab80d
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>
Zelalem Aweke <zelalem.aweke@arm.com>
This patch fixes hangs that happen after soft resetting of rk3399.
Signed-off-by: Piotr Szczepanik <piter75@gmail.com>
Change-Id: If41b12ba1dfcb2ba937361b58eafd50bf5c483d4
This patch adds the Tegra194 SoC information to the nvidia-tegra.rst
file.
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
Change-Id: Id649a5ff1b3f70eeee34b508edb7965e7b7a2454
Tegra194 supports multiple SMMU blocks. This patch adds support to
save register values for SMMU0 and SMMU2, before entering the System
Suspend state.
Change-Id: I3a376cdb606ea057ad7047714717245f9dced5cf
Signed-off-by: Pritesh Raithatha <praithatha@nvidia.com>
This patch updates the memory address space, physical and virtual,
to be 40-bits wide for all Tegra194 platforms.
Change-Id: Ie1bcdec2c4e8e15975048ce1c2a31c2ae0dd494c
Signed-off-by: Steven Kao <skao@nvidia.com>
The per CPU wake times are saved in an array called 't19x_percpu_data'. But,
there is one instance in the code where the name of the variable is misspelt.
This patch fixes this typographical error to fix compilation errors.
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
Change-Id: I52f5f0b150c51d8cc38372675415dec7944a7735
Previously the .init section was created even when the reclaim flag was
manually set to 0.
Change-Id: Ia9e7c7997261f54a4eca725d7ea605192f60bcf8
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>
Zelalem Aweke <zelalem.aweke@arm.com>
Currently tb_fw_cfg_dtb size is fixed to max, which is generally a page
(but depend on the platform). Instead, read the actual size of the dtb
with the libfdt "fdt_totalsize" function.
This avoid flushing extra memory after updating the dtb with mbedtls
heap information when shared heap is used.
Change-Id: Ibec727661116429f486464a0c9f15e9760d7afe2
Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>
Check that entry point information requesting S-EL2
has AArch64 as an execution state during context setup.
Signed-off-by: Artsem Artsemenka <artsem.artsemenka@arm.com>
Change-Id: I447263692fed6e55c1b076913e6eb73b1ea735b7
This patch adds support for enabling S-EL2 if this EL is specified in the entry
point information being used to initialise a secure context. It is the caller's
responsibility to check if S-EL2 is available on the system before requesting
this EL through the entry point information.
Signed-off-by: Achin Gupta <achin.gupta@arm.com>
Change-Id: I2752964f078ab528b2e80de71c7d2f35e60569e1
Consolidate the definition of size_t to one header per AArch, and
the definition of NULL to one header
Signed-off-by: Bence Szépkúti <bence.szepkuti@arm.com>
Change-Id: Iecfbad2cf360cfb705ce7aaa981700fd16219b82
As supporting architectures aside from AArch32 and AArch64 is not a
concern, keeping identical definitions in two places for a large part
of the libc seems counterproductive
The int128 types were left un-unified as __int128 is not supported by
gcc on AArch32
Signed-off-by: Bence Szépkúti <bence.szepkuti@arm.com>
Change-Id: Idf08e6fab7e4680d9da62d3c57266ea2d80472cf
Conceptually, these are supposed to be the largest integers
representable in C, but GCC and Clang define them as long long for
compatibility.
Signed-off-by: Bence Szépkúti <bence.szepkuti@arm.com>
Change-Id: I7c0117f3be167342814d260a371889120dcf6576
Currently, sections within .text/.rodata/.data/.bss are emitted in the
order they are seen by the linker. This leads to wasted space, when a
section with a larger alignment follows one with a smaller alignment.
We can avoid this wasted space by sorting the sections.
To take full advantage of this, we must disable generation of common
symbols, so "common" data can be sorted along with the rest of .bss.
An example of the improvement, from `make DEBUG=1 PLAT=sun50i_a64 bl31`:
.text => no change
.rodata => 16 bytes saved
.data => 11 bytes saved
.bss => 576 bytes saved
As a side effect, the addition of `-fno-common` in TF_CFLAGS makes it
easier to spot bugs in header files.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Change-Id: I073630a9b0b84e7302a7a500d4bb4b547be01d51
This is not conforming C and does not compile with -fno-common.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Change-Id: I6535954cc567d6efa06919069b91e3f50975b073
This was found by compiling with -fno-common:
./build/picopi/release/bl2/imx_snvs.o:(.bss.__packed+0x0): multiple definition of `__packed';
./build/picopi/release/bl2/imx_caam.o:(.bss.__packed+0x0): first defined here
__packed was intended to be the attribute macro from cdefs.h, not an
object of the structure type.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Change-Id: Id02fac3f098be2d71c35c6b4a18012515532f32a
The current range check for the offset is wrong: it is counting bytes,
while indexing an array of uint32_t. Since the offset is always zero,
the parameter is unnecessary. Instead of adding more code to fix the
check, remove the parameter to avoid the problem entirely.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Change-Id: Iadfc7d027155adc754e017b3462233ce9a1d64f6
-Wunused-const-variable=1 is already included by -Wunused-variable,
which is part of -Wall. -Wunused-const-variable=2, which is what we have
been using as part of W=1, warns for unused static const variables in
headers, which will likely produce a lot of false positives that will
take a large effort to fix.
Additionally, some of these issues may be caused by different builds of
TF-A where some features are used in some builds and ignored in others.
Change-Id: Ifa0b16a75344cc1f6240e8d5745005f8f2046d34
Signed-off-by: Justin Chadwell <justin.chadwell@arm.com>
* changes:
Tegra194: add support to reset GPU
Tegra194: memctrl: fix logic to check TZDRAM config register access
Tegra: introduce plat_enable_console()
Tegra: include: drivers: introduce spe.h
Tegra194: update nvg header to v6.4
Tegra194: mce: enable strict checking
Tegra194: CC6 state from last offline CPU in the cluster
Tegra194: console driver compilation from platform makefiles
Tegra194: memctrl: platform handler for TZDRAM setup
Tegra194: memctrl: override SE client as coherent
Tegra194: save system suspend entry marker to TZDRAM
Tegra194: helper functions for CPU rst handler and SMMU ctx offset
Tegra194: cleanup references to Tegra186
Tegra194: mce: display NVG header version during boot
Tegra194: mce: fix cg_cstate encoding format
Tegra194: drivers: SE and RNG1/PKA1 context save support
Tegra194: rename secure scratch register macros
Tegra194: SiP: Fix Rule 8.4 and Rule 10.4 violation
Tegra194: mce: remove unsupported functionality
Tegra194: sanity check target cluster during core power on
Tegra194: fix defects flagged by MISRA scan
Tegra194: mce: fix defects flagged by MISRA scan
Tegra194: remove the GPU reset register macro
Tegra194: MC registers to allow CPU accesses to TZRAM
Tegra194: increase MAX_MMAP_REGIONS macro value
Tegra194: update nvg header to v6.1
Tegra194: update cache operations supported by the ROC
Tegra194: memctrl: platform handlers to reprogram MSS
Tegra194: core and cluster count values
Tegra194: correct the TEGRA_CAR_RESET_BASE macro value
Tegra194: add MC_SECURITY mask defines
Tegra194: Update wake mask, wake time for cpu offlining
Tegra194: program stream ids for XUSB
Tegra194: Update checks for c-state stats
Tegra194: smmu: fix mask for board revision id
Tegra194: smmu: ISO support
Tegra194: Initialize smmu on system suspend exit
Tegra194: Update cpu core-id calculation
Tegra194: read-modify-write ACTLR_ELx registers
Tegra194: Enable fake system suspend
Tegra194: convert 'target_cpu' and 'target_cluster' to 32-bits
Tegra194: platform support for memctrl/smmu drivers
Tegra194: Support for cpu suspend
This patch enables LTO for TF-A when compiled with GCC.
LTO is disabled by default and is enabled by
ENABLE_LTO=1 build option.
LTO is enabled only for aarch64 as there seem to be
a bug in the aarch32 compiler when LTO is enabled.
The changes in the makefiles include:
- Adding -flto and associated flags to enable LTO.
- Using gcc as a wrapper at link time instead of ld.
This is recommended when using LTO as gcc internally
takes care of invoking the necessary plugins for LTO.
- Adding switches to pass options to ld.
- Adding a flag to disable fix for erratum cortex-a53-843419
unless explicitly enabled. This is needed because GCC
seem to automatically add the erratum fix when used
as a wrapper for LD.
Additionally, this patch updates the TF-A user guide with
the new build option.
Signed-off-by: zelalem-aweke <zelalem.aweke@arm.com>
Change-Id: I1188c11974da98434b7dc9344e058cd1eacf5468
This patch adds macros, to define registers required to support GPU
reset, for Tegra194 SoCs.
Change-Id: Ifa7e0161b9e8de695a33856193f500b847a03526
Signed-off-by: Jeetesh Burman <jburman@nvidia.com>
This patch fixes the logic to check if the previous bootloader has
disabled access to the TZDRAM configuration registers. The polarity
for the bit was incorrect in the previous check.
Change-Id: I7a0ba4f7b1714997508ece904c0261ca2c901a03
Signed-off-by: Steven Kao <skao@nvidia.com>
This patch introduces the 'plat_enable_console' handler to allow
the platform to enable the right console. Tegra194 platform supports
multiple console, while all the previous platforms support only one
console.
For Tegra194 platforms, the previous bootloader checks the platform
config and sets the uart-id boot parameter, to 0xFE. On seeing this
boot parameter, the platform port uses the proper memory aperture
base address to communicate with the SPE. This functionality is
currently protected by a platform macro, ENABLE_CONSOLE_SPE.
Change-Id: I3972aa376d66bd10d868495f561dc08fe32fcb10
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
This patch introduces a header file for the spe-console driver. This
file currently provides a device struct and a registration function
call for clients.
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
Change-Id: Ic65c056f5bd60871d8a3f44f2c1210035f878799
This patch updates the header, t194_nvg.h, to v6.4. This
gets it in synch with MTS pre-release 2 - cl39748439.
Change-Id: I1093c9f5dea7b7f230b3267c90b54b7f3005ecd7
Signed-off-by: Steven Kao <skao@nvidia.com>
"Strict checking" is a mode where secure world can access
secure-only areas unlike legacy mode where secure world could
access non-secure spaces as well. Secure-only areas are defined
as the TZ-DRAM carveout and any GSC with the CPU_SECURE bit set.
This mode not only helps prevent issues with IO-Coherency but aids
with security as well.
This patch implements the programming sequence required to enable
strict checking mode for Tegra194 SoCs.
Change-Id: Ic2e594f79ec7c5bc1339b509e67c4c62efb9d0c0
Signed-off-by: Dilan Lee <dilee@nvidia.com>
This patch enables the CC6 cluster state for the cluster, if the
current CPU being offlined is the last CPU in the cluster.
Change-Id: I3380a969b534fcd14f9c46433471cc1c2adf6011
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
This patch includes the console driver from individual platform
makefile, to allow future platforms to include consoles of their
choice.
Change-Id: I4c92199717da410c8b5e8d45af67f4345f743dbd
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
This patch provides the platform with flexibility to perform custom
steps during TZDRAM setup. Tegra194 platforms checks if the config
registers are locked and TZDRAM setup has already been done by the
previous bootloaders, before setting up the fence.
Change-Id: Ifee7077d4b46a7031c4568934c63e361c53a12e3
Signed-off-by: Steven Kao <skao@nvidia.com>
This patch enables IO coherency for SE clients, SEWR and SERD,
by overriding their platform settings to "normal_coherent".
This setting also converts read/write requests from these SE
clients to Normal type.
Change-Id: I31ad195ad30ecc9ee785e5e84184cda2eea5c45a
Signed-off-by: Puneet Saxena <puneets@nvidia.com>
Signed-off-by: Shravani Dingari <shravanid@nvidia.com>
Signed-off-by: Jeff Tsai <jefft@nvidia.com>
This patch adds support to save the system suspend entry and exit
markers to TZDRAM to help the trampoline code decide if the current
warmboot is actually an exit from System Suspend.
The Tegra194 platform handler sets the system suspend entry marker
before entering SC7 state and the trampoline flips the state back to
system resume, on exiting SC7.
Change-Id: I29d73f1693c89ebc8d19d7abb1df1e460eb5558e
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
This patch adds a helper function to get the SMMU context's offset
and uses another helper function to get the CPU trampoline offset.
These helper functions are used by the System Suspend entry sequence
to save the SMMU context and CPU reset handler to TZDRAM.
Change-Id: I95e2862fe37ccad00fa48ec165c6e4024df01147
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
Ambroise Vincent
Piotr Szczepanik
Varun Wadekar
Pritesh Raithatha
Steven Kao
Manish Pandey
Alexei Fedorov
Sandrine Bailleux
Louis Mayencourt
Artsem Artsemenka
Achin Gupta
Soby Mathew
Bence Szépkúti
Mark Dykes
Samuel Holland
Justin Chadwell
zelalem-aweke
Jeetesh Burman
Dilan Lee
Puneet Saxena