This adds a basic BSP for RockPro64, which includes a UART,
an ARM timer, and GICv3 support, and utilizes all available
memory as defined by the upstream U-Boot configuration.
Closes#5414
JFFS2 uses the high values of the unsigned flash offset in a raw node
reference as special identifiers and specifies them using negative
integer literals. This disables those errors for non-RTEMS source and
constrains such comparisons in RTEMS code to avoid larger reworks of the
JFFS2 upstream source.
This removes the adapters that provide glue code directly from JFFS2 to
backend storage drivers since JFFS2, the NAND driver, and the QSPI NOR
driver all support the flashdev interface.
This moves the JFFS2 sources imported from Linux into the contrib/
subdirectory since that is where 3rd party sources belong. The import
revision is currently "v6.1" with some alterations.
This is to workaround a difficult bug in newlib related to
configure-time generation of this flag in newlib.h. Since there is no
multi-lib mechanism for headers, _LDBL_EQ_DBL ends up in newlib.h, which
is installed architecture wide. The default -mcpu= setting for the 68k
toolchain does support long double, so this flag is not defined in
newlib.h.
However, Coldfire targets lack support for long double, leading to some
undesirable behavior in the C/C++ standard headers. ieeefp.h has
numerous sanity checks that will #error and <cmath> from C++ will fail
to compile.
It doesn't look like this issue will ever be fixed in newlib, so we'll
just define this manually as a workaround.
Implements the required bsp directives for the 'new' style rtems
interrupt system. Previously, uC5282 only supported interrupts through
the BSP_installVME_isr directive.
BSP_installVME_isr still behaves the same, but is instead implemented
using the new IRQ directives. To ensure compatibility with existing
EPICS related code without causing subtle bugs, BSP_installVME_isr will
also poke at the EPORT interrupt settings to enable IRQ1 when installing
a "fake VME" interrupt (vector >= 192).
This also fixes a regression from e9cb088995 that broke clock interrupt
handlers, leading to the BSP crashing shortly after boot.
Added API header OK test in psxtests/psxhdrs
as a part of my work to GSoC 25
note that functions ppoll(), sem_clockwait() and timespec_get()
don't have a prototype yet so the tests will fail. I won't
add them to the libpsxhdrs.yml file until my PR is approved in newlib.
Add build options for start address, vector base, and BRAM length,
wire them into the microblaze_fpga linker script and BSP group,
and add the dependencies to the kcu105_qemu BSP spec.
- Extend IMFS and libio headers to define default allocator/deallocator ops
and logic for configuring memfile_ops.
- Update imfs_memfile to use custom allocators for memory blocks.
- Fix imfs_statvfs.f_bfree calculation to respect user-defined ops.
- Add psximfs03 test case to validate custom allocator behavior.
- Refactor psximfs01 to share helper functions with psximfs03.
Add AXI timer definitions and a BSP btimer implementation and
install them in the microblaze_fpga BSP.
Co-authored-by: Alex White <alex.white@oarcorp.com>
Added the implementation of timespec_get() with tests.
The implementation used clock_gettime() as a reference.
The tests used psxtimes01 as a reference.
Updates
rtems/programs/gsoc#69
In order for the linker to pick the correct soft-float libraries when
building for a target without an FPU the -msoft-float flag needs to be
present.
This is already the case in the pkg-config file where ABI_FLAGS are
included in Ldflags:
Ldflags: ${ABI_FLAGS} -B${libdir} -qrtems -Wl,--gc-sections
But this is not present in the custom make includes:
LDFLAGS = -B$(exec_prefix)/$(RTEMS_BSP)/lib -qrtems -Wl,--gc-sections
The RTEMS template make rule for linking includes all of CFLAGS:
define bsp-link-c
$(LINK.c) $(CPU_CFLAGS) $(AM_CFLAGS) $(AM_LDFLAGS) \
-o $(basename $@)$(EXEEXT) $(LINK_OBJS) $(LINK_LIBS)
endef
The standard (GNU) built-in make rule for linking does not include
CFLAGS when linking:
$(CC) $(LDFLAGS) $* $(LOADLIBES) $(LDLIBS) -o $@
Add ABI_FLAGS to LDFLAGS in the custom make includes such that building
soft-float applications works with the default built-in make rules as
well.
The sizeof(ino_t) can potentially vary based on the target. This
resulted in -Wformat warnings. Add PRIdino_t in <rtems/inttypes.h>
to have a portable format specifier.
Closes#5393.
This adds a backend for flashdev that simulates a flash device with
adjustable delays to simulate real flash devices and offers
configurable flash parameters. This also migrates the JFFS2
filesystem tests to the flashdev framework and eliminates the
purpose-built simulated flash in the testsuite in favor of a more
generic and widely usable solution.
This adds summary support to the build with a global option to enable
this support and necessary changes to the imported summary support
source to integrate it into the build system and resolve errors.
This makes the number of active chipselects a BSP configuration
parameter. The name of the macro in upstream source is misleading and
suggests that the driver can support up to the selected value when in
actuality it requires that number of chips to be present. If this number
of chips is not present, the driver will hang on initialization waiting
for a chip to respond that does not exist.
Move JFFS2 from the Cygwin implementation of CRC32 to the zlib
implementation which can take advantage of hardware acceleration and
reduces code duplication. Synthetic benchmarks on AArch64 hardware show
approximately a 17x improvement in CRC32 calculation speed and
real-world testing with JFFS2 shows a 1% improvement in mount times, a
5% improvement in data write times, and a 9% improvement in garbage
collection times.
A test was written to compare hash generation between the two algorithms
and a partition was created with one and then remounted with the other
with no errors.
Adds the registers that expose the second microblaze timer interrupt.
Moves the system clock to a shared interrupt so users may add a new
interrupt off the second interrupt.
Uses the CONFIGURE_MAXIMUM_PRIORITY to set the priority range for
EDF and CBS schedulers under the control of applications.
The default priority behavior is now the same as other schedulers.
The user API to get a task's current priority returns 0 for deadline
driven tasks.
Fixes#5390
All but ~50 warnings have been fixed if you account for unmerged
MRs. These settings should give a warning free build given the
current state.
Warnings by Type
=================================
2 -Wshift-negative-value
3 -Wenum-conversion
7 -Wcast-function-type
7 -Wmissing-field-initializers
8 -Wclobbered
32 -Wtype-limits
This adds the Xilinx compatibility layer for use with various Xilinx
libraries. The first consumer of this will be lwIP support for the
Zynq7000 platform.
Added pthread_cond_clockwait(), pthread_mutex_clocklock()
pthread_rwlock_clockrdlock() and pthread_rwlock_clockwrlock()
that are new in POSIX Issue 8. Also added tests.
The implementation used the timed versions of these functions
as a reference.
Updates
rtems/programs/gsoc#69
Newer C versions require that the storage-class specifier like
static or _Thread_Local be the first thing in a declaration. Adding
this warning to the default set to let us eliminate violations.