Doc: Fix internal links in interrupt framework design guide
Some of the sections numbers have been incorrect since the document
structure has changed. Using references to sections rather than
hard-coded section numbers fixes the issue while making it more
future-proof to updates.
While converting the interrupt framework design document from Markdown
to reStructuredText format, see commit 6f6257476754 ("Convert
documentation to reStructuredText"), the structure of the document was
changed and one of the titles got lost. Fix that.
Louis Mayencourt [Wed, 27 Feb 2019 14:24:16 +0000 (14:24 +0000)]
Cortex-A73: Implement workaround for errata 852427
In AArch32, execution of 2 instructions with opposite condition code
might lead to either a data corruption or a CPU deadlock. Set the bit
12 of the Diagnostic Register to prevent this.
Change-Id: I22b4f25fe933e2942fd785e411e7c0aa39d5c1f4 Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>
Louis Mayencourt [Mon, 11 Feb 2019 11:25:50 +0000 (11:25 +0000)]
MPAM: enable MPAM EL2 traps
Complete the MPAM enablement in TF-A for lower ELs by enabling the EL2
traps in MPAMHCR_EL2 and MPAM2_EL2.This prevents an
MPAM-unaware-hypervisor to be restricted by an MPAM-aware-guest.
Change-Id: I47bf3f833fa22baa590f83d49cc0e3f2974e698d Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>
Varun Wadekar [Fri, 11 Jan 2019 22:47:48 +0000 (14:47 -0800)]
Tegra: Support for scatterfile for the BL31 image
This patch provides support for using the scatterfile format as
the linker script with the 'armlink' linker for Tegra platforms.
In order to enable the scatterfile usage the following changes
have been made:
* provide mapping for ld.S symbols in bl_common.h
* include bl_common.h from all the affected files
* update the makefile rules to use the scatterfile and armlink
to compile BL31
* update pubsub.h to add sections to the scatterfile
NOTE: THIS CHANGE HAS BEEN VERIFIED WITH TEGRA PLATFORMS ONLY.
The previous commit added the infrastructure to load and save
ARMv8.3-PAuth registers during Non-secure <-> Secure world switches, but
didn't actually enable pointer authentication in the firmware.
This patch adds the functionality needed for platforms to provide
authentication keys for the firmware, and a new option (ENABLE_PAUTH) to
enable pointer authentication in the firmware itself. This option is
disabled by default, and it requires CTX_INCLUDE_PAUTH_REGS to be
enabled.
Change-Id: I35127ec271e1198d43209044de39fa712ef202a5 Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
ARMv8.3-PAuth adds functionality that supports address authentication of
the contents of a register before that register is used as the target of
an indirect branch, or as a load.
This feature is supported only in AArch64 state.
This feature is mandatory in ARMv8.3 implementations.
This feature adds several registers to EL1. A new option called
CTX_INCLUDE_PAUTH_REGS has been added to select if the TF needs to save
them during Non-secure <-> Secure world switches. This option must be
enabled if the hardware has the registers or the values will be leaked
during world switches.
To prevent leaks, this patch also disables pointer authentication in the
Secure world if CTX_INCLUDE_PAUTH_REGS is 0. Any attempt to use it will
be trapped in EL3.
Change-Id: I27beba9907b9a86c6df1d0c5bf6180c972830855 Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
This commit improves the SDHost driver for RPi3 as following:
* Unblock MMC_CMD(17). Using MMC_CMD(17) is more efficient on
block reading.
* In some low probability that SEND_OP_COND might results CRC7
error. We can consider that the command runs correctly. We don't
need to retry this command so removing the code for retry.
* Using MMC_BUS_WIDTH_1 as MMC default value to improve the stability.
* Increase the clock to 50Mhz in data mode to speed up the io.
* Change the pull resistors configuration to gain more stability.
Signed-off-by: Ying-Chun Liu (PaulLiu) <paulliu@debian.org>
Streaming store under specific conditions might cause deadlock or data
corruption. Set bit 25:24 of CPUECTLR_EL1, which disables write
streaming to the L2 to prevent this.
Change-Id: Ib5cabb997b35ada78b27e75787afd610ea606dcf Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>
TLBI VAAE1 or TLBI VAALE1 targeting a page within hardware page
aggregated address translation data in the L2 TLB might cause
corruption of address translation data. Set bit 59 of CPUACTLR2_EL1 to
prevent this.
Change-Id: I59f3edea54e87d264e0794f5ca2a8c68a636e586 Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>
Louis Mayencourt [Wed, 20 Feb 2019 12:11:41 +0000 (12:11 +0000)]
Add workaround for errata 764081 of Cortex-A75
Implicit Error Synchronization Barrier (IESB) might not be correctly
generated in Cortex-A75 r0p0. To prevent this, IESB are enabled at all
expection levels.
Change-Id: I2a1a568668a31e4f3f38d0fba1d632ad9939e5ad Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>
Tien Hock, Loh [Tue, 26 Feb 2019 01:25:14 +0000 (09:25 +0800)]
plat: intel: Add BL31 support to Intel Stratix10 SoCFPGA platform
This adds BL31 support to Intel Stratix10 SoCFPGA platform. BL31 in TF-A
supports:
- PSCI calls to enable 4 CPU cores
- PSCI mailbox calls for FPGA reconfiguration
Signed-off-by: Loh Tien Hock <tien.hock.loh@intel.com>
Yann Gautier [Fri, 15 Feb 2019 16:33:27 +0000 (17:33 +0100)]
stm32mp1: add minimal support for co-processor Cortex-M4
STM32MP1 chip embeds a dual Cortex-A7 and a Cortex-M4.
The support for Cortex-M4 clocks is added when configuring the clock tree.
Some minimal security features to allow communications between A7 and M4
are also added.
Marek Vasut [Mon, 11 Feb 2019 23:09:46 +0000 (00:09 +0100)]
rcar_gen3: plat: Prevent PCIe hang during L1X config access
In case the PCIe controller receives a L1_Enter_PM DLLP, it will
disable the internal PLLs. The system software cannot predict it
and can attempt to perform device config space access across the
PCIe link while the controller is in this transitional state. If
such condition happens, the PCIe controller register access will
trigger ARM64 SError exception.
This patch adds checks for which PCIe controller is enabled,
checks whether the PCIe controller is in such a transitional
state and if so, first completes the transition and then restarts
the instruction which caused the SError.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Usama Arif [Wed, 12 Dec 2018 17:14:29 +0000 (17:14 +0000)]
plat/arm: Support for Cortex A5 in FVP Versatile Express platform
Cortex A5 doesnt support VFP, Large Page addressing and generic timer
which are addressed in this patch. The device tree for Cortex a5
is also included.
Change-Id: I0722345721b145dfcc80bebd36a1afbdc44bb678 Signed-off-by: Usama Arif <usama.arif@arm.com>
Usama Arif [Wed, 12 Dec 2018 17:08:33 +0000 (17:08 +0000)]
Division functionality for cores that dont have divide hardware.
Cortex a5 doesnt support hardware division such as sdiv and udiv commands.
This commit adds a software division function in assembly as well as include
appropriate files for software divison.
The software division algorithm is a modified version obtained from:
http://www.keil.com/support/man/docs/armasm/armasm_dom1359731155623.htm
Change-Id: Ib405a330da5f1cea1e68e07e7b520edeef9e2652 Signed-off-by: Usama Arif <usama.arif@arm.com>
This patch adds support for Versatile express FVP (Fast models).
Versatile express is a family of platforms that are based on ARM v7.
Currently this port has only been tested on Cortex A7, although it
should work with other ARM V7 cores that support LPAE, generic timers,
VFP and hardware divide. Future patches will support other
cores like Cortex A5 that dont support features like LPAE
and hardware divide. This platform is tested on and only expected to
work on single core models.
Change-Id: I10893af65b8bb64da7b3bd851cab8231718e61dd Signed-off-by: Usama Arif <usama.arif@arm.com>
Sathees Balya [Wed, 30 Jan 2019 15:56:44 +0000 (15:56 +0000)]
docs: Document romlib design
Change-Id: I2b75be16f452a8ab7c2445ccd519fb057a135812 Co-authored-by: John Tsichritzis <john.tsichritzis@arm.com> Signed-off-by: John Tsichritzis <john.tsichritzis@arm.com>
Commit 2559b2c8256f ("xlat v2: Dynamically detect need for CnP bit")
modified the code to convert the compile-time check for ARMv8.2-TTCNP to
a runtime check, but forgot to update the documentation associated to it.
Change-Id: I6d33a4de389d976dbdcce65d8fdf138959530669 Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
- If EL3 is using AArch32, setting MDCR_EL3.SCCD to 1 disables counting
in Secure state in PMCCNTR.
- If EL3 is using AArch64, setting SDCR.SCCD to 1 disables counting in
Secure state in PMCCNTR_EL0.
So far this effect has been achieved by setting PMCR_EL0.DP (in AArch64)
or PMCR.DP (in AArch32) to 1 instead, but this isn't considered secure
as any EL can change that value.
Change-Id: I82cbb3e48f2e5a55c44d9c4445683c5881ef1f6f Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
Samuel Holland [Sun, 17 Feb 2019 21:33:33 +0000 (15:33 -0600)]
allwinner: Clean up CPU ops functions
Convert them to take an mpidr instead of a (cluster, core) pair. This
simplifies all of the call sites, and actually makes the functions a bit
smaller.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Since commit 01fc1c24b9a0 ("BL31: Use helper function to save registers
in SMC handler") all the general-purpose registers are saved when
entering EL3. It isn't needed to save them here.
Change-Id: Ic540a5441b89b70888da587ab8fc3b2508cef8cc Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
Update macro to check need for CVE-2017-5715 mitigation
Armv8.5 introduces the field CSV2 to register ID_AA64PFR0_EL1. It can
have the following 3 values:
- 0: Branch targets trained in one hardware described context may affect
speculative execution in a different hardware described context. In
some CPUs it may be needed to apply mitigations.
- 1: Branch targets trained in one hardware described context can only
affect speculative execution in a different hardware described
context in a hard-to-determine way. No mitigation required.
- 2: Same as 1, but the device is also aware of SCXTNUM_ELx register
contexts. The TF doesn't use the registers, so there is no
difference with 1.
The field CSV2 was originally introduced in the TRM of the Cortex-A76
before the release of the Armv8.5 architecture. That TRM only mentions
the meaning of values 0 and 1. Because of this, the code only checks if
the field has value 1 to know whether to enable or disable the
mitigations.
This patch makes it aware of value 2 as well. Both values 1 and 2
disable the mitigation, and 0 enables it.
Change-Id: I5af33de25a0197c98173f52c6c8c77b51a51429f Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
Yann Gautier [Thu, 14 Feb 2019 08:17:55 +0000 (09:17 +0100)]
stm32mp1: add timeout detection in reset driver
This change makes the platform to panic in case of peripheral reset
resource malfunction.
Change-Id: I17eb9cb045b78a4e5142a8c33b744e84992d732a Signed-off-by: Yann Gautier <yann.gautier@st.com> Signed-off-by: Etienne Carriere <etienne.carriere@st.com> Signed-off-by: Nicolas LE BAYON <nicolas.le.bayon@st.com>
Yann Gautier [Thu, 14 Feb 2019 10:01:20 +0000 (11:01 +0100)]
stm32mp1: use functions to retrieve some peripheral addresses
PWR, RCC, DDRPHYC & DDRCTRL addresses can be retrieved from device tree.
Platform asserts the value read from the DT are the SoC addresses.
Change-Id: I43f0890b51918a30c87ac067d3780ab27a0f59de Signed-off-by: Yann Gautier <yann.gautier@st.com> Signed-off-by: Etienne Carriere <etienne.carriere@st.com> Signed-off-by: Nicolas LE BAYON <nicolas.le.bayon@st.com>
Yann Gautier [Thu, 14 Feb 2019 10:15:03 +0000 (11:15 +0100)]
stm32mp1: update I2C and PMIC drivers
Regulator configuration at boot takes more information from DT.
I2C configuration from DT is done in I2C driver.
I2C driver manages more transfer modes.
The min voltage of buck1 should also be increased to 1.2V,
else the platform does not boot.
Heavily modifies stm32_i2c.c since many functions move inside the source
file to remove redundant declarations.
Change-Id: I0bee5d776cf3ff15e687427cd6abc06ab237d025 Signed-off-by: Yann Gautier <yann.gautier@st.com> Signed-off-by: Etienne Carriere <etienne.carriere@st.com> Signed-off-by: Lionel Debieve <lionel.debieve@st.com> Signed-off-by: Nicolas LE BAYON <nicolas.le.bayon@st.com>
Yann Gautier [Thu, 14 Feb 2019 10:14:39 +0000 (11:14 +0100)]
stm32mp1: use new functions to manage timeouts
Remove the previously use function: get_timer, and use new functions
timeout_init_us and timeout_elapsed.
Change-Id: I4e95b123648bff7ca91e40462a2a3ae24cfe1697 Signed-off-by: Yann Gautier <yann.gautier@st.com> Signed-off-by: Etienne Carriere <etienne.carriere@st.com> Signed-off-by: Nicolas LE BAYON <nicolas.le.bayon@st.com>
Yann Gautier [Thu, 14 Feb 2019 10:13:50 +0000 (11:13 +0100)]
stm32mp1: remove some dependencies on clocks and reset in drivers
Include all RCC, clocks and reset headers from stm32mp1_def.h
which if exported to the firmware through platform_def.h.
The same dependency removal is done in common code as well.
Some useless includes are also removed in stm32_sdmmc2 driver.
Loh Tien Hock [Wed, 13 Feb 2019 06:39:31 +0000 (14:39 +0800)]
plat: intel: Fix faulty DDR calibration value
A DDR calibration value is missing write mask, causing ECC DDR calibration
to fail. This patch addresses the issue. ECC should also be scrubbed before
MMU initializes, thus the scrubbing is moved to ddr intialization phase.
Signed-off-by: Loh Tien Hock <tien.hock.loh@intel.com>
Loh Tien Hock [Mon, 11 Feb 2019 02:56:28 +0000 (10:56 +0800)]
drivers: partition: Add simple MBR partition entries support
This is to add simple MBR partition entry support. This will read all four
MBR partition into the partition list, and the partition type will be saved
in the list.name[0] entry.
Signed-off-by: Loh Tien Hock <tien.hock.loh@intel.com>
projects / arm-tf.git / log