Juan Castillo [Fri, 7 Nov 2014 09:44:58 +0000 (09:44 +0000)]
Juno: Add support for image overlaying in Trusted SRAM
This patch allows the BL3-1 NOBITS section to overlap the BL1 R/W
section since the former will always be used after the latter.
Similarly, the BL3-2 NOBITS section can overlay the BL2 image
when BL3-2 is loaded in Trusted SRAM.
Due to the current size of the images, there is no actual overlap.
Nevertheless, this reorganization may help to optimise the Trusted
SRAM usage when the images size grows.
Note that because BL3-1 NOBITS section is allowed to overlap the
BL1 R/W section, BL1 global variables will remain valid only until
execution reaches the BL3-1 entry point during a cold boot.
User Guide: Enable secure memory on Foundation FVP
Previously, the User Guide recommended launching the Foundation
FVP with the parameter --no-secure-memory, which disabled security
control of the address map. This was due to missing support for
secure memory regions in v1 of the Foundation FVP. This is no longer
needed as secure memory is now supported on the Foundation FVP.
This patch updates the User Guide to recommend enabling secure
memory instead.
Juan Castillo [Thu, 13 Nov 2014 17:04:33 +0000 (17:04 +0000)]
Bug fix: initialize bl30_image_info fields before use
This patch initializes the version field in the bl30_image_info
structure when loading BL30. This initialization must be done before
calling load_image().
Calls to system register read accessors functions may be optimised
out by the compiler if called twice in a row for the same register.
This is because the compiler is not aware that the result from
the instruction may be modified by external agents. Therefore, if
nothing modifies the register between the 2 reads as far as the
compiler knows then it might consider that it is useless to read
it twice and emit only 1 call.
This behaviour is faulty for registers that may not have the same
value if read twice in succession. E.g.: counters, timer
control/countdown registers, GICv3 interrupt status registers and
so on.
The same problem happens for calls to system register write
accessors functions. The compiler might optimise out some calls
if it considers that it will produce the same result. Again, this
behaviour is faulty for cases where intermediate writes to these
registers make a difference in the system.
This patch fixes the problem by making these assembly register
accesses volatile.
Juan Castillo [Mon, 24 Nov 2014 11:45:12 +0000 (11:45 +0000)]
Specify FIP filename at build time
This patch allows to define the name of the FIP at build time by
defining the FIP_NAME variable. If FIP_NAME is not defined, default
name 'fip.bin' is used.
Soby Mathew [Thu, 2 Oct 2014 16:24:19 +0000 (17:24 +0100)]
Fix CPU_SUSPEND when invoked with affinity level higher than get_max_afflvl()
This patch fixes the assertion failure when CPU_SUSPEND is invoked with
an affinity level higher than supported by the platform by adding suitable
checks for affinity level within `psci_cpu_suspend`. Also added suitable
bound checks within `psci_aff_map_get_idx` to prevent indexing beyond array
limits.
Kévin Petit [Mon, 8 Dec 2014 13:23:09 +0000 (13:23 +0000)]
fip_create: don't succeed if one of the passed files doesn't exist
If one of the files passed to fip_create on the command line doesn't
exist, it will print an error message but produce an incomplete
fip.bin file and report success. This behaviour could potentially
hide errors made in the command line arguments.
This patch addresses the issue by having the tool bail out if one of
the supplied files can't be processed.
Signed-off-by: Kévin Petit <kevin.petit@arm.com>
Fixes ARM-software/tf-issues#279
In both FVP and Juno ports, IRQ #16, which is a PPI, is incorrectly
identified as secure SGI #8 through the constant IRQ_SEC_SGI_8.
This patch removes it.
Fix problem of dependencies on the fiptool makefile target
The 'fiptool' target doesn't depend on fip_create's source files,
neither directly nor indirectly. As a result, the FIP tool is not
rebuilt whenever its source files change.
This patch makes the ${FIPTOOL} target into a phony target so that the
FIP tool's sub-makefile is always called. The sub-makefile correctly
handles the dependencies. It also moves the completion message into
the sub-makefile so that it is only displayed when the tool is
actually recompiled.
Achin Gupta [Mon, 10 Nov 2014 11:50:30 +0000 (11:50 +0000)]
Precede a 'sev' with a 'dsb' in bakery lock code
This patch fixes a bug in the bakery lock implementation where a data
synchronisation barrier instruction is not issued before sending an event as
mandated by the ARMv8 ARM. This can cause a event to be signalled before the
related memory accesses have completed resulting in erroneous execution.
Use the BL3-1 runtime console as the crash console.
This patch reassigns the crash console on Juno and FVP to use the runtime
BL3-1 console. The crash console is changed to SoC UART0 (UART2) from the
previous FPGA UART0 (UART0) on Juno. In FVP, it is changed from UART0 to
UART1.
Juan Castillo [Mon, 20 Oct 2014 11:29:58 +0000 (12:29 +0100)]
Juno: Use the generic ARM GIC driver
This patch replaces the usage of the GIC private driver in Juno with
the generic ARM GIC driver. The private driver is no longer necessary
and has been removed from the Juno port.
Juan Castillo [Mon, 20 Oct 2014 11:27:28 +0000 (12:27 +0100)]
Improvements to ARM GIC driver
This patch introduces several improvements to the ARM GIC driver:
* In function gicd_set_itargetsr(), target CPU is specified using
the same bit mask detailed in the GICD_ITARGETSRn register instead
of the CPU linear ID, removing the dependency between bit position
and linear ID in the platform porting. The current CPU bit mask
may be obtained by reading GICD_ITARGETSR0.
* PPIs and SGIs are initialized in arm_gic_pcpu_distif_setup().
SPIs are initialized in arm_gic_distif_setup().
* By default, non secure interrupts are assigned the maximum
priority allowed to a non secure interrupt (defined by
GIC_HIGHEST_NS_PRIORITY).
* GICR base address is allowed to be NULL for GICv1 and GICv2.
Optimize Cortex-A57 cluster power down sequence on Juno
This patch optimizes the Cortex-A57 cluster power down sequence by not
flushing the Level1 data cache. The L1 data cache and the L2 unified
cache are inclusive. A flush of the L2 by set/way flushes any dirty
lines from the L1 as well. This is a known safe deviation from the
Cortex-A57 TRM defined power down sequence. This optimization can be
enabled by the platform through the 'SKIP_A57_L1_FLUSH_PWR_DWN' build
flag. Each Cortex-A57 based platform must make its own decision on
whether to use the optimization.
This patch also renames the cpu-errata-workarounds.md to
cpu-specific-build-macros.md as this facilitates documentation
of both CPU Specific errata and CPU Specific Optimization
build macros.
Apply errata workarounds only when major/minor revisions match.
Prior to this patch, the errata workarounds were applied for any version
of the CPU in the release build and in the debug build an assert
failure resulted when the revision did not match. This patch applies
errata workarounds in the Cortex-A57 reset handler only if the 'variant'
and 'revision' fields read from the MIDR_EL1 match. In the debug build,
a warning message is printed for each errata workaround which is not
applied.
The patch modifies the register usage in 'reset_handler` so
as to adhere to ARM procedure calling standards.
Add support for level specific cache maintenance operations
This patch adds level specific cache maintenance functions
to cache_helpers.S. The new functions 'dcsw_op_levelx',
where '1 <= x <= 3', allow to perform cache maintenance by
set/way for that particular level of cache. With this patch,
functions to support cache maintenance upto level 3 have
been implemented since it is the highest cache level for
most ARM SoCs.
These functions are now utilized in CPU specific power down
sequences to implement them as mandated by processor specific
technical reference manual.
Juan Castillo [Wed, 24 Sep 2014 09:00:06 +0000 (10:00 +0100)]
FVP: keep shared data in Trusted SRAM
This patch deprecates the build option to relocate the shared data
into Trusted DRAM in FVP. After this change, shared data is always
located at the base of Trusted SRAM. This reduces the complexity
of the memory map and the number of combinations in the build
options.
Juan Castillo [Fri, 5 Sep 2014 16:29:38 +0000 (17:29 +0100)]
Juno: Reserve some DDR-DRAM for secure use
This patch configures the TrustZone Controller in Juno to split
the 2GB DDR-DRAM memory at 0x80000000 into Secure and Non-Secure
regions:
- Secure DDR-DRAM: top 16 MB, except for the last 2 MB which are
used by the SCP for DDR retraining
- Non-Secure DDR-DRAM: remaining DRAM starting at base address
Build option PLAT_TSP_LOCATION selects the location of the secure
payload (BL3-2):
- 'tsram' : Trusted SRAM (default option)
- 'dram' : Secure region in the DDR-DRAM (set by the TrustZone
controller)
The MMU memory map has been updated to give BL2 permission to load
BL3-2 into the DDR-DRAM secure region.
Juan Castillo [Thu, 4 Sep 2014 13:43:09 +0000 (14:43 +0100)]
Juno: Use TZC-400 driver calls
This patch replaces direct accesses to the TZC-400 registers by the
appropiate calls to the generic driver available in the Trusted
Firmware in order to initialize the TrustZone Controller.
Functions related to the initialization of the secure memory,
like the TZC-400 configuration, have been moved to a new file
'plat_security.c'. This reorganization makes easier to set up
the secure memory from any BL stage.
TZC-400 initialization has been moved from BL1 to BL2 because BL1
does not access the non-secure memory. It is BL2's responsibility
to enable and configure the TZC-400 before loading the next BL
images.
In Juno, BL3-0 initializes some of the platform peripherals, like
the DDR controller. Thus, BL3-0 must be loaded before configuring
the TrustZone Controller. As a consequence, the IO layer
initialization has been moved to early platform initialization.
This patch uses the IMAGE_BL<x> constants to create translation tables specific
to a boot loader stage. This allows each stage to create mappings only for areas
in the memory map that it needs.
Remove BSS section access by 'plat_print_gic' during crash reporting
This patch avoids the problem of crash reporting mechanism accessing
global data in BSS by 'plat_print_gic_regs' for FVP platforms. Earlier
it depended on the global 'plat_config' object for the GIC Base address
in FVP platforms which would have caused exception if it were accessed
before the BSS was initialized. It is now fixed by dynamically
querying the V2M_SYS_ID to find the FVP model type and accordingly
selecting the appropriate GIC Base address.
This patch also fixes the 'plat_print_gic_regs' to use the correct GIC
Distributor base address for printing GICD_IS_PENDR register values
for both Juno and FVP platforms.
Add support for specifying pre-built BL binaries in Makefile
This patch adds support for supplying pre-built BL binaries for BL2,
BL3-1 and BL3-2 during trusted firmware build. Specifying BLx = <path_to_BLx>
in the build command line, where 'x' is any one of BL2, BL3-1 or BL3-2, will
skip building that BL stage from source and include the specified binary in
final fip image.
This patch also makes BL3-3 binary for FIP optional depending on the
value of 'NEED_BL33' flag which is defined by the platform.
Soby Mathew [Fri, 29 Aug 2014 13:41:58 +0000 (14:41 +0100)]
Reset CNTVOFF_EL2 register before exit into EL1 on warm boot
This patch resets the value of CNTVOFF_EL2 before exit to EL1 on
warm boot. This needs to be done if only the Trusted Firmware exits
to EL1 instead of EL2, otherwise the hypervisor would be responsible
for this.
Dan Handley [Thu, 28 Aug 2014 13:04:06 +0000 (14:04 +0100)]
Fix minor issues in user guide
* Fix broken link to SCP download.
* Remove requirement to install `ia32-libs`. This package is no
longer available in current versions of Ubuntu and is no
longer required when using the Linaro toolchain.
Dan Handley [Tue, 26 Aug 2014 15:02:19 +0000 (16:02 +0100)]
Fix reset to BL3-1 instructions in user guide, part 2
Fix the instructions for resetting to the BL3-1 entrypoint in the
user guide. The BL3-1 and BL3-2 image locations changed in the fix
to ARM-software/tf-issues#100 (commit 186c1d4). This is distinct
from the similar issue fixed in commit bfb1dd5.
Also clarify the dependence on the FVP_SHARED_DATA_LOCATION and
FVP_TSP_RAM_LOCATION build options, and tidy up the "Notes
regarding Base FVP configuration options" section.
Soby Mathew [Thu, 21 Aug 2014 14:20:27 +0000 (15:20 +0100)]
Rationalize UART usage among different BL stages
This patch changes the UART port assignment for various BL stages
so as to make it consistent on the platform ports. The BL1, BL2 and
BL3-1 now uses UART0 on the FVP port and SoC UART0 on the Juno port.
The BL3-2 uses UART2 on the FVP port and FPGA UART0 on the Juno
port.
This provides an interim fix to ARM-software/tf-issues#220 until
support is added for changing the UART port for a BL image between
cold boot and runtime.
Dan Handley [Thu, 14 Aug 2014 08:14:52 +0000 (09:14 +0100)]
Move up dependency versions in user guide
Move up the version numbers of the following Trusted Firmware
dependencies in the user guide:
* Foundation and Base FVPs (latest publically available
versions).
* EDK2 implementation. The guide now uses the latest version from
https://github.com/ARM-software/edk2.git. This requires the
`iasl` package to also be installed.
* Linux kernel. The guide now uses the latest version from
https://github.com/ARM-software/linux.git.
Juan Castillo [Tue, 12 Aug 2014 16:24:30 +0000 (17:24 +0100)]
Juno: Read primary CPU MPID from SCC GPR_1
This patch removes the PRIMARY_CPU definition hardcoded in the
Juno port. Instead, the primary CPU is obtained at runtime by
reading the SCC General Purpose Register 1 (GPR_1), whose value
is copied by the SCP into shared memory during the boot process.
This patch implements the TSP on Juno. It executes from on-chip Trusted
SRAM.
Also, the other bootloader images (i.e. BL1 R/W, BL2 and BL3-1) have
been moved around. The reason is, although there was enough space
overall to store the TSP in SRAM, there was no contiguous free chunk
of SRAM big enough to hold it.
This patch keeps the overall memory layout (i.e. keeping BL1 R/W at
the bottom, BL2 at the top and BL3-1 in between) but moves the base
addresses of all the bootloader images in such a way that:
- memory fragmentation is reduced enough to fit BL3-2 in;
- new base addresses are suitable for release builds as well as debug
ones;
- each image has a few extra kilobytes for future growth.
BL3-1 and BL3-2 are the images which received the biggest allocations
since they will most probably grow the most.
This patch also adds instruction synchronization barriers around the code which
handles the timer interrupt in the TSP. This ensures that the interrupt is not
acknowledged after or EOIed before it is deactivated at the peripheral.
Achin Gupta [Mon, 18 Aug 2014 13:40:27 +0000 (14:40 +0100)]
Juno: Implement PSCI CPU_OFF and CPU_SUSPEND APIs
This patch adds support for PSCI CPU_OFF and CPU_SUSPEND APIs to the Juno port
of the ARM Trusted Firmware. The maximum affinity level that can be suspended is
the cluster level (affinity level 1). Support for suspending the system level is
not present.
This patch adds the initial port of the ARM Trusted Firmware on the Juno
development platform. This port does not support a BL3-2 image or any PSCI APIs
apart from PSCI_VERSION and PSCI_CPU_ON. It enables workarounds for selected
Cortex-A57 (#806969 & #813420) errata and implements the workaround for a Juno
platform errata (Defect id 831273).
Achin Gupta [Wed, 20 Aug 2014 16:33:09 +0000 (17:33 +0100)]
FVP: Update device trees to match cpuidle driver
This patch updates the representation of idle tables and cpu/cluster topology in
the device tree source files for the FVP to what the latest cpuidle driver in
Linux expects. The device tree binaries have also been updated.
Soby Mathew [Mon, 18 Aug 2014 15:57:56 +0000 (16:57 +0100)]
Add documentation for CPU specific abstraction and Errata workarounds
This patch adds documentation for CPU specific abstraction in the firmware-
design.md and adds a new document cpu-errata-workarounds.md to describe
the cpu errata workaround build flags.
Soby Mathew [Thu, 14 Aug 2014 15:19:29 +0000 (16:19 +0100)]
Add support for selected Cortex-A57 errata workarounds
This patch adds workarounds for selected errata which affect the Cortex-A57 r0p0
part. Each workaround has a build time flag which should be used by the platform
port to enable or disable the corresponding workaround. The workarounds are
disabled by default. An assertion is raised if the platform enables a workaround
which does not match the CPU revision at runtime.
Soby Mathew [Thu, 14 Aug 2014 12:36:41 +0000 (13:36 +0100)]
Add CPU specific crash reporting handlers
This patch adds handlers for dumping Cortex-A57 and Cortex-A53 specific register
state to the CPU specific operations framework. The contents of CPUECTLR_EL1 are
dumped currently.
Soby Mathew [Thu, 14 Aug 2014 11:49:05 +0000 (12:49 +0100)]
Add CPU specific power management operations
This patch adds CPU core and cluster power down sequences to the CPU specific
operations framework introduced in a earlier patch. Cortex-A53, Cortex-A57 and
generic AEM sequences have been added. The latter is suitable for the
Foundation and Base AEM FVPs. A pointer to each CPU's operations structure is
saved in the per-cpu data so that it can be easily accessed during power down
seqeunces.
An optional platform API has been introduced to allow a platform to disable the
Accelerator Coherency Port (ACP) during a cluster power down sequence. The weak
definition of this function (plat_disable_acp()) does not take any action. It
should be overriden with a strong definition if the ACP is present on a
platform.
Soby Mathew [Thu, 14 Aug 2014 11:22:32 +0000 (12:22 +0100)]
Add platform API for reset handling
This patch adds an optional platform API (plat_reset_handler) which allows the
platform to perform any actions immediately after a cold or warm reset
e.g. implement errata workarounds. The function is called with MMU and caches
turned off. This API is weakly defined and does nothing by default but can be
overriden by a platform with a strong definition.
Soby Mathew [Thu, 14 Aug 2014 10:33:56 +0000 (11:33 +0100)]
Introduce framework for CPU specific operations
This patch introduces a framework which will allow CPUs to perform
implementation defined actions after a CPU reset, during a CPU or cluster power
down, and when a crash occurs. CPU specific reset handlers have been implemented
in this patch. Other handlers will be implemented in subsequent patches.
Also moved cpu_helpers.S to the new directory lib/cpus/aarch64/.
Soby Mathew [Tue, 19 Aug 2014 10:26:00 +0000 (11:26 +0100)]
Rework use of labels in assembly macros.
This patch provides a workaround for the ASM_ASSERT label issue
and also reworks the use of labels in assembly macros.
If the caller of the ASM_ASSERT macro happened to use the
label '1' to jump past the ASM_ASSERT macro, it would not have
worked since the ASM_ASSERT macro internally used the same label.
Hence, as a workaround, this patch makes the label a high
number in the expectation that the caller will never use it.
Also updated the other assembly macros using numerical labels to
named lables.
This patch reworks FVP port's power management implementation to perform
platform actions only when the platform exported hook is invoked for the highest
affinity level to enter/exit the OFF state.
For example, during a CPU_OFF operation, fvp_affinst_off() is called twice: for
affinity level 0 and affinity level 1 (in that order). CPU specific operations
are deferred until the next invocation if it is determined through a call to
psci_get_max_phys_off_afflvl() that this is CPU is the last in the cluster.
Similarly, during power up if the CPU is the first in the cluster, both CPU and
cluster specific operations are performed when fvp_affinst_on_finish() is
invoked for affinity level 1. Earlier, they were done across the two invocations
of the handler.
Add APIs to preserve highest affinity level in OFF state
This patch adds APIs to find, save and retrieve the highest affinity level which
will enter or exit from the physical OFF state during a PSCI power management
operation. The level is stored in per-cpu data.
It then reworks the PSCI implementation to perform cache maintenance only
when the handler for the highest affinity level to enter/exit the OFF state is
called.
For example. during a CPU_SUSPEND operation, state management is done prior to
calling the affinity level specific handlers. The highest affinity level which
will be turned off is determined using the psci_find_max_phys_off_afflvl()
API. This level is saved using the psci_set_max_phys_off_afflvl() API. In the
code that does generic handling for each level, prior to performing cache
maintenance it is first determined if the current affinity level matches the
value returned by psci_get_max_phys_off_afflvl(). Cache maintenance is done if
the values match.
This change allows the last CPU in a cluster to perform cache maintenance
independently. Earlier, cache maintenance was started in the level 0 handler and
finished in the level 1 handler. This change in approach will facilitate
implementation of tf-issues#98.
Rework state management in the PSCI implementation
This patch pulls out state management from the affinity level specific handlers
into the top level functions specific to the operation
i.e. psci_afflvl_suspend(), psci_afflvl_on() etc.
In the power down path this patch will allow an affinity instance at level X to
determine the state that an affinity instance at level X+1 will enter before the
level specific handlers are called. This will be useful to determine whether a
CPU is the last in the cluster during a suspend/off request and so on.
Similarly, in the power up path this patch will allow an affinity instance at
level X to determine the state that an affinity instance at level X+1 has
emerged from, even after the level specific handlers have been called. This will
be useful in determining whether a CPU is the first in the cluster during a
on/resume request and so on.
As before, while powering down, state is updated before the level specific
handlers are invoked so that they can perform actions based upon their target
state. While powering up, state is updated after the level specific handlers have
been invoked so that they can perform actions based upon the state they emerged
from.
This patch adds a structure defined by the PSCI service to the per-CPU data
array. The structure is used to save the 'power_state' parameter specified
during a 'cpu_suspend' call on the current CPU. This parameter was being saved
in the cpu node in the PSCI topology tree earlier.
The existing API to return the state id specified during a PSCI CPU_SUSPEND call
i.e. psci_get_suspend_stateid(mpidr) has been renamed to
psci_get_suspend_stateid_by_mpidr(mpidr). The new psci_get_suspend_stateid() API
returns the state id of the current cpu.
The psci_get_suspend_afflvl() API has been changed to return the target affinity
level of the current CPU. This was specified using the 'mpidr' parameter in the
old implementation.
The behaviour of the get_power_on_target_afflvl() has been tweaked such that
traversal of the PSCI topology tree to locate the affinity instance node for the
current CPU is done only in the debug build as it is an expensive operation.
This patch adds a macro which will flush the contents of the specified member of
the per-CPU data structure to the PoC. This is required to enable an update of a
per-CPU data member to be visible to all observers.
Juan Castillo [Tue, 12 Aug 2014 10:17:06 +0000 (11:17 +0100)]
Add support for PSCI SYSTEM_OFF and SYSTEM_RESET APIs
This patch adds support for SYSTEM_OFF and SYSTEM_RESET PSCI
operations. A platform should export handlers to complete the
requested operation. The FVP port exports fvp_system_off() and
fvp_system_reset() as an example.
If the SPD provides a power management hook for system off and
system reset, then the SPD is notified about the corresponding
operation so it can do some bookkeeping. The TSPD exports
tspd_system_off() and tspd_system_reset() for that purpose.
Versatile Express shutdown and reset methods have been removed
from the FDT as new PSCI sys_poweroff and sys_reset services
have been added. For those kernels that do not support yet these
PSCI services (i.e. GICv3 kernel), the original dtsi files have
been renamed to *-no_psci.dtsi.
Dan Handley [Mon, 4 Aug 2014 10:41:20 +0000 (11:41 +0100)]
Clarify platform porting interface to TSP
* Move TSP platform porting functions to new file:
include/bl32/tsp/platform_tsp.h.
* Create new TSP_IRQ_SEC_PHY_TIMER definition for use by the generic
TSP interrupt handling code, instead of depending on the FVP
specific definition IRQ_SEC_PHY_TIMER.
* Rename TSP platform porting functions from bl32_* to tsp_*, and
definitions from BL32_* to TSP_*.
* Update generic TSP code to use new platform porting function names
and definitions.
* Update FVP port accordingly and move all TSP source files to:
plat/fvp/tsp/.
* Update porting guide with above changes.
Note: THIS CHANGE REQUIRES ALL PLATFORM PORTS OF THE TSP TO
BE UPDATED
Achin Gupta [Mon, 4 Aug 2014 22:13:10 +0000 (23:13 +0100)]
Unmask SError interrupt and clear SCR_EL3.EA bit
This patch disables routing of external aborts from lower exception levels to
EL3 and ensures that a SError interrupt generated as a result of execution in
EL3 is taken locally instead of a lower exception level.
The SError interrupt is enabled in the TSP code only when the operation has not
been directly initiated by the normal world. This is to prevent the possibility
of an asynchronous external abort which originated in normal world from being
taken when execution is in S-EL1.
projects / arm-tf.git / log