This moves the management console implementation into its own file so
that BSPs which do not need it do not carry the code around with them.
Closes#5120
This adds generic JFFS2 interworking code that allows JFFS2 to be used
on top of any flashdev backend. It currently only supports NOR flashdev
backends.
Commit 3d782180ea changed the reported Git
hash to short. Change this back to the normal Git hash. Existing
support code may rely on this. The short hash is generated based on the
current repository state. The reported hash should be useful also for
future states of the repository. If Git changes the hash algorithm in
the future, then the hash length may be used as an indicator for the
hash algorithm. With a variable hash length this is more difficult.
Simplify expression.
Update #5037.
Accesses to the L2C registers performed by the L2C driver are now
serialized with spin locks. This avoids concurrent access to the L2C
registers by multiple processors. Proposed by GRLIB-TN-0021.
Update #4925.
This updates the implementation of the LEON3 processor power-down
function.
It now contains the workaround for the GR712RC power-down errata
described in GR712RC-UM, version 2.16, section 1.7.8. The workaround is
compatible with other LEON3/LEON4 components.
Update #4875.
Properly serialize explicit thread priority changes (for example
rtems_task_set_priority()) and thread priority changes carried out by
thread queues (locking protocols such as priority inheritance).
Update #5159.
Test the proper serialization of explicit thread priority changes (for
example rtems_task_set_priority()) and thread priority changes carried
out by thread queues (locking protocols such as priority inheritance).
With RTEMS_DEBUG enabled, the new test case fails with an assertion. It
is fixed in a follow up commit.
Update #5159.
Change the time to convert so that rounding issues do not propergate to
the second next digit. Properly round the result used for the precision
estimate.
The two RTCs supported by that driver (MCP7940M and EOZ9) use slightly
different register order. Add support for a different order to the
i2c-rtc driver.
The RTEMS_UNREACHABLE() is a hint to the compiler that this code
location is unreachable. The compiler may use this information to
optimize the code. If a test case fails and we reach this code
location, then unpredictable things may happen making the test case
debugging more difficult than necessary.
In case of overflows, we may have an RTEMS_RECORD_UPTIME_HIGH event as
the first event. This uptime high event has no associated
RTEMS_RECORD_UPTIME_LOW event.
This code had a uint32_t local variable for seconds from a timespec.
This was assigning a time_t to a uint32_t and ultimately returning it.
This propgated to rtems_timespec_add_to().
Coverity CID 1512527
Closes#5107.
The updated waf version deletes the waf cache directory in the build
tree. This removes also the build specification item cache. Move the
cache directory so that it works across configure commands.
The devfg driver loads the PL with a bitfile image. The driver can also
support scrubbing.
These drivers are from Patrick Gauvin <pggauvin at gmail.com> and a thread
on the devel list: https://lists.rtems.org/pipermail/devel/2017-May/017705.html
When using QEMU configurations that support SMP for Zynq7000 systems,
the second core is started at the same time as the first core instead of
waiting for an event to trigger a check for the value at 0xfffffff0
before jumping into RTEMS code. This makes the erroneously started core
wait as expected and prevents prefetch and data aborts from occurring
before the MMU has been properly configured. This was recently exposed
by cleanup done to the ARM GICv2 driver that removed some delays which
were allowing this to operate normally.
The general optclkbootcpu option is targeted at all ARM QEMU BSPs, but
is not actually linked in to the build tree. This links it in as
expected and removes a redundant and stale copy in the realview PBX BSP.
The RT_TSW_OK field is set to 1 if there was no error. The Message Error
(ME) field indicates an error and must be set to 0 if there was no error
so the result of tsw&RT_TSW_OK must be negated. It is then left-shifted
to the Messsage Error (ME) bit field of the message information word.
Fix issue CID 1399772
This removes the headers imported from the embeddedsw repository in
favor of a much thinner shim. This also removes the complicated build
system configuration necessary to support use of these headers. The
primary reason for removal is that certain external Xilinx libraries
also require use of these headers and this causes version mismatches and
header conflicts that can be avoided.
- Add support for four new ports, UART2-UART5.
- Add build options to allow console device configuration.
- Segregate device-specific definitions from the device family files. X macros
are used to maintain a single source of truth and have the configuration
done at compile-time
- Add raspberrypi_uart_init() to make it convenient for users to install uart
devices
Close#5130
Co-authored-by: Ning Yang <yangn0@qq.com>
* It is not granted that we will always access to natually aligned
access. So, before handling endianess do byte-a-byte load using
the appropriate libfdt function to avoid unaligned access issues
Signed-off-by: Francescodario Cuzzocrea <bosconovic@gmail.com>
This removes the carriage return printed before the shell prompt as it
can cause other output to be overwritten unexpectedly. This was
discovered when using the "cat" shell command on a file whose content
contained no trailing \n\r and was short enough to be entirely swallowed
by the prompt.
The rtime decompression routine does not fully check bounds during the
entirety of the decompression pass and can corrupt memory outside the
decompression buffer if the compressed data is corrupted. This adds the
required check to prevent this failure mode.
Updates #5072
This reverts commit c1aa06dbd2.
Extending the rtems_stack_checker_info like this breaks the existing
visitor API used by rtems_stack_checker_iterate(). After review, using
an information structure contradicts the intention of the reporter to
allow a minimal reporting. Filling up the information structure with
all the details would pull in several dependencies, like
_Thread_Get_name(). The API should be improved to provide the user with
the necessary information from the TCB without having to access members
directly.
This removes the xilinx and lp64 identifiers from the ZynqMP BSPs and
updates those names throughout the build system. The xilinx identifier
is being removed because it is implied by zynqmp and lp64 is being
removed because it is the default for AArch64.
This also renames grp_zu3eg.yml to grp_hardware.yml to more accurately
reflect its purpose.
Updates #5119
Now that minimum.exe is minimum.norun.exe, the tester will not attempt
to run this sample test and will therefore not categorize it as a
failure since it includes no output. BSP-level test configurations for
this test can now be removed since those configurations only existed to
prevent it from being run.
On other targets, the rtems_cache_flush_entire_data() also performs a
clean and invalidate.
This is a workaround for "Answer 68874 - Zynq UltraScale+ MPSoC, APU -
An Eviction Might Overtake A Cache Clean Operation".
Let the caller provide the translation table base and the used page
tables to ease testing.
Simplify the error handling by using early returns.
Return RTEMS_TOO_MANY instead of RTEMS_NO_MEMORY if the page tables are
all used.
There is no need to wait on secondary processors for the GIC distributor
enable since the BSPs for real targets start the secondary processors
in _CPU_SMP_Start_processor().
Whether a floating-point implementation supports trapping of
floating-point exceptions is implementation defined. At least Qemu,
Cortex-A53, and Cortex-A72 do not support them.
In SMP configurations, check that we run on a configured processor. If not,
then there is not much that can be done since we do not have a stack available
for this processor. Just loop forever in this case. Do this in assemlby to
ensure that no stack memory is used.
Make the support for starting in EL2/EL3 customizable. A boot loader or
the Arm Trusted Firmware should start RTEMS in non-secure EL1 mode.
In start.S, use local labels.
For the aarch64/xilinx-zynqmp the support for starting in EL2/EL3 is
disabled by default. For the Qemu xlnx-zcu102 machine, the default is
to start in non-secure EL1 mode. This can be controlled by options, for
example "-machine xlnx-zcu102,secure=on,virtualization=on".
The information from the README.md have been merged into the
documentation.
The necessary tools for the sdcard.sh are quite tricky to build. All
necessary information to create an SD image are in the documentation
already. So the script isn't necessary any more.
Update #5088
Dynamically mapped blocks must be aligned to the MMU page size just like
startup-configured blocks. This was not being enforced and could cause a
hang with bad input.
Rename xilinx_zynqmp_rpu BSP variant to zynqmp_rpu_lock_step to
emphasize that this BSP is for the lock-step mode RPU configuration.
Add BSP variants zynqmp_rpu_split_0 and zynqmp_rpu_split_1 for the split
mode RPU configuration for core 0 and 1 respectively.
Remove copyright from DornerWorks since the files contain not contributions
from this company. Fix the copyright years of the embedded brains
contributions.
Add support for the BSP_ARM_GIC_MULTI_PROCESSOR_SECONDARY build option
for the GICv2 support. This option is useful for multiprocessor systems
without SMP support.
Do not wait in the individual _CPU_SMP_Start_processor() for the
secondary processor. Wait for all of them in
_CPU_SMP_Finalize_initialization() before the L2 cache is enabled.
There is no need to wait for the secondary processor in
_CPU_SMP_Start_processor() since _CPU_SMP_Finalize_initialization() does
nothing.
The caller of _CPU_SMP_Start_processor() ensures that we do not start
the current processor.
Remove copy and paste from the arm SMP support. The shared aarch64
implementation of rtems_cache_enable_data() does not enable a particular
cache, it just enables the C bit in the SCTLR_EL1. This is already done
in aarch64_mmu_enable(). There is no need to wait for secondary
processors in _CPU_SMP_Start_processor().
Calls from Thumb code into ARM code require the use of a BLX instruction
and an optional offset prefix instruction. The BLX instruction encoding
requires that the least significant bit be 0 in all cases while the BL
instruction for Thumb-to-Thumb calls includes that bit as part of the
offset. This ensures that bit 0 of the BLX is set to 0 as required by
the instruction encoding specification.
Fix _CPU_SMP_Get_current_processor(), since read from register MPIDR_EL1
(the least significant byte) is not always right,for example the least
significant byte of MPIDR_EL1 is always zero for cortex-a55.
aarch64 has a thread ID register TPIDR_EL1 available for for OS management
purposes. RTEMS stores per-CPU control in TPIDR_EL1 when the core startup,
so we can use _Per_CPU_Get_index() to get current processor index.
update #5064
- Refactor the pl011 driver to be extensible.
- Add IRQ support and baudrate configuration support for pl011 driver.
- Modify related BSP.
- Add doxygen comments for arm-pl011.
Close#5026
Co-authored-by: Ning Yang <yangn0@qq.com>
It is not clear why this is necessary. For example, the
zynq_uart_initialize() does not issue the '\r' before waiting for an
inactive transmission state.
Similar to the recent commit in tms570-sci.c, the assumption that a UART will
only see printable ASCII characters, instead of any value in the range
0x00-0xFF, is wrong.
A non forgiving binary protocol will be thrown off by this driver sending
"\r\r\r\r" when initializing.
If a user wants to flush the interface, they should explicitely use the
dedicated function `tcflush(fd, TCIOFLUSH);`.
Coverity CID 1619144
In rtems_aio_completed_list_op() memory was getting freed without releasing
a lock.
This commit should also solve coverity issues 1619145, 1619143 and 1619142
since they all seem to be generated by rtems_aio_completed_list_op().
Closes#5127
- change type unsigned to uint64_t, it might overflowed if
the size is greater the 4GB.
- increase buffer size is to eliminate the compiler's warning
that the buffer might be too small
For example, printf("%+i", 1) prints "+1". However, kvprintf() did
print just "1" for this example. According to PRINTF(3):
A sign must always be placed before a number produced by a signed
conversion.
For "%+r" radix conversions, keep the "+" handling as it is, since this
is a non-standard conversion. For "%+p" pointer conversions, continue
to ignore the sign modifier to be in line with libc.
This change allows to support the ' conversion modifier in the future.
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/1310
The struct timespec tv_sec member is of type time_t. Make sure that all
variables related to this member are of the type time_t. This is important for
targets where long is a 32-bit type and time_t a 64-bit type.
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/1373
The IRQ list in this file are specific to PC hardware and should be
part of the BSP. Further, there are paravirtualized environments which
do not follow the PC hardware IRQ list. Moving this avoids collisions.
Resurrect proper file comment. Order copyrights according to template. Remove
white space at the end of line. Do error checks early. Define objects under
optional extern "C".
Update #5047.
Modified the body of lio_listio(), to make sure it return -1 and sets
errno to ENOSYS, when RTEMS_POSIX_API is not defined.
The psxenosys test has been modified to test for lio_listio() only
when RTEMS_POSIX_API is not defined. In the test all the puts() have
been removed. If kept they would have made the output of the test depending
on the build configuration, making the content of psxenosys.scn inconsistent.
Removing the puts() makes the .scn file valid regardless of the build
configuration.
Additionally, rtems_aio_notify() in aio_misc.c has been modified,
simplifying it and reducing the part inside the ifdef.
Closes#5125
This adds support for the 6 SPI interfaces on the STM32H7 series chips
with an initial example for the stm32h750b discovery kit development
board. Configuration is similar to existing peripherals. Chip select
lines are software-controlled since the SPI peripheral only supports a
single hardware-controlled chip select line. This implementation does
not use interrupts.
This reverts commit d81bf04501.
The commit is reverted as this is only partial solution and remaining part
to fix properly is extremely hard and usually done in completely diffetrent
way. Hence let's not enforce compiler but rather enforce RTEMS/H7 users.
This commit adds support for lio_listio().
psxaio05 has been added to test the new method.
To avoid code duplication, the body of aio_read and aio_write has been
moved in two helper methods. In this way i can use it also in lio_listio().
In addition to that, a limit on the number of total aio operation enqueued
has been added.
Updates rtems/programs/gsoc#29
This offset is already accounted for by Stack_check_Usable_stack_start()
called from Stack_check_Visit_stack() and adding it a second time can
cause interaction beyond the bounds of the stack.
The default exception handler uses the Save Program Status Register
(SPSR), however, if _ARMV4_Exception_reserved_default() would get
called, the state of this register is unpredictable. Replace potential
calls to _ARMV4_Exception_reserved_default() with an undefined
instruction.
Add BSP option BSP_ARM_GIC_ENABLE_FIQ_FOR_GROUP_0 to customize the ARM
GIC support. Enable this option for arm/altera-cyclone-v and
arm/xilinx-zynq BSPs by default.
The EOZ9 RTC has a similar register interface like the MCP7940M (and
quite some other I2C RTCs). This commit:
* Extracts the generic parts from MCP7940M and moves it into a generic
i2c-rtc driver.
* Uses the new i2c-rtc for the MCP7940M.
* Uses the new i2c-rtc for the new Abracom EOZ9.
Add support to configure the second region of DDR memory if the
BSP configured RAM size is greater than 2G.
Add the second region's memory to the heap.
This follows current conventions.
Two have been left behind for now until we decide what to do as they are
imports.
* cpukit/compression/xz/COPYING
* cpukit/libmisc/uuid/COPYING
Add filename length checks to prevent the creation of files and folders
that cannot be listed, unlinked or renamed by subsequent calls.
These changes are sufficient to make the jffs2fsrenamelongname and
jffs2nandfsrenamelongname testcases pass.
Updates #5073
Currently, if a BSP is built without MPU alignment enforcement, the MPU
is still enabled and can produce a non-functional binary since code can
be mixed with data within a memory region and memory regions are marked
as read-only or no-execute based on section addresses and sizes leading
to overlapping regions being defined which causes the later-defined
region to take precedence. This change disables the MPU when alignment
is not enforced, allowing the binary to function.
In _Terminate(), everything after invoking the fatal extensions is essentially
dead code. Simplify this code block and provide a reasonable safe fall-back
for badly configured applications. Just disable masked interrupts and execute
an idle loop.
Update #5067.
Previously the first cluster of the file was not updated when
truncating to 0 but was in any case freed.
Now the first cluster of the file is updated during the truncation,
leaving the directory entry in a consistent state.
Closes#2757
This is intended to close any valid FDs that were opened, but 0 is also
a valid FD. This change ensures that a FD of 0 is also closed properly.
Coverity CID 1467408 (original)
Coverity CID 1616151 (followup)
off_by_one: Testing whether handle fd is strictly greater than zero is
suspicious. fd leaks when it is zero.
The patch changes compiler ABI flags option and adds -mno-unaligned-access
there. Usage of this option is also officially recommended way
by the errata document how to work around the issue. For more details
about the errata please see referenced issue number.
Closes#5110.
Fix a race condition in which a file opened with O_TRUNC is made
available to other file functions expecting a file descriptor before
the truncation is performed.
This is only possible if the other thread was using an invalid file
descriptor as the open call has yet to finish.
Closes#5109
- Move the APIC implementation to its on file instead of having it in clock.c;
- Use the MADT for retrieving the Local APIC base address;
- Initialize the APIC during interrupt initialization.
Add a method for initializing the ACPICA table subsystem earlier than
the initilization of the entire ACPICA subsystem. Also add a method to
walk through the subtables of an ACPI table
Coverity Scan spotted the possibility that the
_Objects_Information_table could be over-indexed if the API
value provided was too high. This function is only used during
initialization and the input should be trusted. Adding a debug
_Assert() emphasizes this trust.
Coverity CID 1512508
Closes#5084.
There is certainly no USB 20 TG while there is a known USB 2 OTG tech. The
crash caused by this is well observed from media01.exe from libbsd compiled
for H7 and run on the board.
Coverity CID 1616018
O_SYNC and O_DSYNC have the same value which led Coverity to note that
checking for both values in if's or switches leads to dead code. The
solution is to add a cpp check that they are equal and enough commentary
so if they ever are not the same, there is a hint as to why the cpp
check got tripped.
Closes#5100.
This fixes the ECC info struct naming such that they all use the _Info
suffix. This also adds comments describing which structures are paired
with which event types.
- Implements necessary ACPICA OS Services Layer interfaces;
- Retrieve the RSDP from the FreeBSD Bootloader for amd64 or from multiboot2 for amd64efi;
- Add ACPI initialization routine;
- Use ACPI in bsp_reset.
Now as the default reporter is set to be a quiet reporter, we have stackchk test case for the default quiet reporter, stackchk03 for a custom user-defined reporter and stackchk04 for the detailed reporter function
This patch updates STM32 H7 HAL source files. The files are taken from two
STM projects from their github.com repositories:
(i)
https://github.com/STMicroelectronics/stm32h7xx_hal_driver.git
The project files are still available under BSD-3 license
and the version/commit used is:
fec141ce999da655a48e1a15db83a72d564a1312
which represents Release v1.11.3 exactly.
(ii)
https://github.com/STMicroelectronics/cmsis_device_h7.git
The project files are available under Apache 2.0 license. Fortunately
the project does not contain NOTICE file so no need to do anything special
when used in RTEMS.
The project version/commit imported is:
faccfec37f82f7a1319c21638111b0f7335de7fe
which represents Release v1.10.4 exactly.
GitLab requires explicit exclusions and unfortunately it's one per line while it
does look gross it is clearer to read what's going on.
Also clean up some spacing.
The testcase failed because the rmdir call in that case had two possibilities to fail and the implementation of rmdir checked the other possibility first.
Closes#5071
This file is ancient and there have been many, many contributors to RTEMS that
deserve equal acknowledgement.
I have opened an issue to create a new contributrors section in the
documentation in the docs projects:
* rtems/docs/rtems-docs#21
I have linked this version of the file there but this file should not ship with
the next version of RTEMS given the GitLab migration it will not be converted to
MarkDown
This file is now obsolete now that we have moved to GitLab. A ticket has been
opened in rtems/docs/rtems-docs#21 to discuss adding a contributors section.
GitLab lists who the 'Code Owners' are on top of every file in the web source
browser this is the canonical way to find out who maintainers are of those files
or directories.
Even with knowing who the maintainers are the best way to figure out who is
looking after what is to open an issue or an MR. This way we can keep a group
level of maintainership over the code base and have CODEOWNERS enforce approvals
before merging.
All this does is points to the README. The README points to the docs website as
it should.
INSTALL files are falling out of favour since sites show README.* by default
over an INSTALL file.
This file has not been relevant in at least 8 years and even then only for
texinfo? The file calls itself Experimental and relies on CVS. It was used for
cutting old RTEMS releases.
Merge all licenses into one large file with a TOC. This way we can link to one
file and not have to worry about having many files everywhere.
The original licenses have been copied in without modification though it will
render differently online.
Changed fstest support to allow control on the test result of each fstest, updated all fstest accordingly.
Modified FS_FAIL macro to fail hard and not just print a message which rtems-test runner ignores.
Split failing test cases from the ones that passes; added fsrenameexisting, fsrenamepermexisting, fsrmdirparent, fsrenamelongname and fsrenamemaxlinks test cases.
Marked fsrenameexisting, fsrenamepermexisting as expected to fail due to #2169.
Marked fsrmdirparent as expected to fail, covered by #5071.
Marked fssymlink as expected to fail due to rename problems partially covered by #2169.
Marked rfsfsrenamelongname as expected to fail, covered by #5069.
Marked *fsrenamemaxlinks as expected to fail, covered by #5070.
Marked jffs2fsrenamelongname and jffs2nandfsrenamelongname as expected to fail, as JFFS2 seems to not have a limit on the file name, covered by #5073
rename now actively changes the inode of the file being renamed.
This change has been done to fix an issue when the file being renamed is already opened before the rename, in particular failing to update the inode number (and the two lookup tables) could cause at least two problems:
1) A new file created in the same source folder of the file being renamed may get the same fat file descriptor of the renamed file
2) An open to the renamed file opens a new fat file descriptor, thus putting out of sync the operations on the file (two distinct fat file descriptor working on the same clusters and entries)
Previously the inode was in any case changed once the file was closed.
Another possible solution would have been to separate the inode number from the lookup table, but this solution would have required much more memory because it needed to keep track of all files seen in the file system up to the last operation and not only the opened ones.
Closes#5044
The patch adds the generation of a notification after an Asynchronous I/O
request has been completed. The notification is generated using the aio_sigevent
filed in the aiocb struct, as defined by the posix specifiations.
A new control, ensuring the validity of the field, is added on each call to
aio_read(), aio_write() and aio_fsync().
The patch also adds tests to cover the new code, and fixes formatting errors of
various aio files.
This adds an environment variable that can be used to enable or disable
window size retrieval for the RTEMS shell. This will also disable window
size retrieval upon failure while allowing the user to re-enable it.
On systems that don't provide a proper terminal or have broken window
size detection, the RTEMS shell can take a long time execute commands if
the kernel tick interval is longer than 1ms. This updates the shell
window size retrieval timeout code to account for the kernel tick
interval bringing the experienced timeout from 12.5ms to 1.25ms when
using the default kernel tick interval (10ms).
Import <sys/tree.h> from Newlib as <rtems/score/bsd-tree.h> using Newlib commit:
commit eb03ac17f16f1bd354482148426353fd35cd879d
Author: Sebastian Huber <sebastian.huber@embedded-brains.de>
Date: Tue Oct 5 15:53:47 2021 +0200
sys/tree.h: Red child with black sibling rotations
Add specialized rotations RB_RED_ROTATE_LEFT() and RB_RED_ROTATE_RIGHT() which
may be used if we rotate a red child which has a black sibling. Such a red
node must have at least two child nodes so that the following red-black tree
invariant is fulfilled:
Every path from a given node to any of its descendant NULL nodes goes through
the same number of black nodes.
PARENT
/ \
BLACK RED
/ \
BLACK BLACK
Close#5049.
From the Dhrystone RATIONALE:
o No procedure merging
Although Dhrystone contains some very short procedures where execution would
benefit from procedure merging (inlining, macro expansion of procedures),
procedure merging is not to be used. The reason is that the percentage of
procedure and function calls is part of the "Dhrystone distribution" of
statements contained in [1]. This restriction does not hold for the string
functions of the C version since ANSI C allows an implementation to use
inline code for these functions.
Close#5058.
Closes#5055
## Summary
add rtems_counting_semaphore_wait_timed_ticks and rtems_counting_semaphore_try_wait for self-contained object counting semaphore.
1. rtems_counting_semaphore_wait_timed_ticks(rtems_counting_semaphore \*counting_semaphore,uint32_t ticks)
2. rtems_counting_semaphore_try_wait(rtems_counting_semaphore \*counting_semaphore)
close issue #5055.
<!-- Default settings, if it is a dropdown it will set after submission -->
This switches the XQspiPsu NOR driver layer to use the RTEMS event
system so that waits for interrupt-based data transfers can time out
instead of waiting indefinitely. It is sometimes possible for either the
controller or the NOR chip to fail to issue a transfer completion event
and recovery is preferable to a hang.
Added field to the aio control block to track when a result is retrieved
and modified code accordingly.
Updated documentation for aio_return() and aio_error().
Updated test psxaio01 to make it compatible with specifications.
Closes#5039
Set/clear SCTLR[I] on all online processors. Do not enable/disable the
L2C-310 cache in the instruction cache enable/disable since it is a
unified cache.
On some targets or configurations, the data cache cannot be disabled.
The data cache may be necessary to provide atomic operations. In SMP
configurations, the data cache may be required to ensure data coherency.
Close#5050.
If the RTEMS_INTERRUPT_LOCK_NEEDS_OBJECT define has a non-zero value,
then the interrupt lock operations require an object of type
rtems_interrupt_lock, otherwise no lock object is required.
This indication can be used to avoid the space overhead for lock objects
when they are not needed. In this case, the lock operations will not
use a lock objects parameter.
Update #4957 and #5038.
This adds error reporting for ZynqMP including L1 and L2 cache, on-chip
memory (OCM) error correcting code (ECC), and DDR ECC. OCM ECC supports
fault injection from within RTEMS. DDR ECC technically supports fault
injection as well, but requires that the program injecting faults
operate exclusively outside of DDR. The AArch64 port is not currently
capable of operating exclusively in OCM due to size constraints and
would need to be booted via JTAG or via a non-relocating u-boot to
accomplish this.
When starting multitasking there is a chance an interrupt could be fired
while rsp is still pointing to the ISR stack area if we load rflags
before rsp on "_CPU_Context_restore"
This test was broken by b031cf2b19 since
it expects the macros to evaluate to an empty string for non-SMP builds.
This updates the check for the new content.
Remove ancient SCCS tags from the tree, automated scripting, with two
minor fixup to keep things compiling. All the common forms in the tree
were removed with a perl script.
Sponsored by: Netflix
The interrupt context save is now done in the BSP. This avoids an issue
where a register is modified by the interrupt handler before it is
saved. Specifically, the MSR register was modified by the `addi`
instruction in the interrupt handler before the MSR was saved. This
caused the MSR to be saved with the wrong value.
Closes#4962
## spec/stm32h7: Default nucleo-h755zi BSP to build for CORE_M7
Some of the STM32H7 parts have a secondary Cortex M4 core included and require the user to select which core they are compiling for. The `arm/nucleo-h755zi` BSP was lacking a default core selection, causing the BSP to fail to build with the default options selected. This PR changes the default option to select the larger Cortex M7 core by default.
Use first 1GiB of the DDR RAM with the exception of a potential 64KiB NULL
pointer protection area by default. The upper part of the DDR RAM may be used
by the RPU or for dynamic memory.
The i.MXRT1166 has three iomux controllers. The driver in RTEMS
(originally for i.MX6) doesn't support that. Switching to absolute
addresses for the pins in the FDT works around that.
Compared to changing the IOMUX driver, it has the following advantages:
- If a devicetree is written, the pinctrl-groups don't have to be split
up between different iomux controllers which makes device trees simpler
to read.
- A driver or application doesn't has to handle multiple pinctrl-groups
which simplifies the drivers.
The C files that provide the device trees are generated with
rtems-bin2c. Due to that, the license information in these files is
lost. Therefore this patch adds the header manually. Also adapt the
commands that are used to generate the files to reflect that.
This makes the active level of the FDT available to the user. It doesn't
change the current behaviour. It only provides a function to get the
active or inactive level of a pin.
## Implement a new ST Nucleo STM32H755ZI BSP
Hi all, for a project I recently acquired a Nucleo STM32H755ZI. Thanks to the work of @sebhub, @c-mauderer, and Karel Gardas, it was pretty easy to create a new Nucleo BSP for this chip by copying the Nucleo H743 BSP. I was not able to get the 8 MHz HSE clock enabled, so this is currently clocked off of the PLL using the 64 MHz HSI clock and the RTC clock is disabled. Other features haven't been thoroughly tested, but since they rely on the high quality STM32H7 HAL from ST, I have pretty high confidence in them. I've been able to run most testsuites.
I have tested:
* The CM7 core
* The CM4 core
* Console output on UART3 (the default USB UART) and the shell
* GPIO
* Sample programs like `hello`, `fileio` (and shell), `paranoia`, `unlimited`, etc.
Known Broken:
* Debugging with OpenOCD is wonky. I can set breakpoints, halt the processor and inspect memory and registers and return from functions, but stepping does not work.
* The program begins to run after being flashed but soon (under 1 second) stops running. Pressing the reset button makes the program run continuously afterward. The board resets nearly instantly.
* Shell commands `ls`, `cd`, `cpuuse`, etc. hang for a while before executing. I'm not sure where the time is spent because the built-in `time` command reports that these commands execute nearly instantly.
I haven't tested:
* Pretty much everything else
<!-- Default settings, if it is a dropdown it will set after submission -->
This function may be used to ensure that a count of processors performed a set
of actions and then wait on the barrier performing only memory loads (no memory
stores).
The clock tick rate was off by a factor of two in some configurations.
Use the maximum counter frequency to get the best time resolution. Do
not use the automatic interrupt clear feature.
Update #4982.
The rtems_cache_get_data_cache_size() and
rtems_cache_get_instruction_cache_size() functions shall return the entire
cache size for a level of 0. Levels greater than 0 shall return the size of
the associated level.
Update #4982.
Created an helper method that will be called by rtems_aio_handle().
The method will process the various types of aio requests.
Converted the old style documentation to Doxygen.
Formatted the code to adhere to rtems formatting rules.
Closes#5017
The _ARMV7M_Pendable_service_call() and _ARMV7M_Supervisor_call() work
as a team. The --ef and ++ef is there to preserve the original exception
frame across the jump to and from _ARMV7M_Thread_dispatch().
void _ARMV7M_Pendable_service_call( void )
{
Per_CPU_Control *cpu_self = _Per_CPU_Get();
/*
* We must check here if a thread dispatch is allowed. Right after a
* "msr basepri_max, %[basepri]" instruction an interrupt service may
still
* take place. However, pendable service calls that are activated during
* this interrupt service may be delayed until interrupts are enable
again.
*/
if (
( cpu_self->isr_nest_level |
cpu_self->thread_dispatch_disable_level ) == 0
) {
ARMV7M_Exception_frame *ef;
cpu_self->isr_nest_level = 1;
_ARMV7M_SCB->icsr = ARMV7M_SCB_ICSR_PENDSVCLR;
_ARMV7M_Trigger_lazy_floating_point_context_save();
At this point, the floating point context should be saved on the
exception frame. The FPCCR.LSPACT bit should be 0, to indicate that lazy
state preservation is no longer active.
ef = (ARMV7M_Exception_frame *) _ARMV7M_Get_PSP();
--ef;
_ARMV7M_Set_PSP( (uint32_t) ef );
This new exception frame is just there to jump to
_ARMV7M_Thread_dispatch(). Here was the problem, that FPCCR.LSPACT was
not set to 1. This resulted in a floating-point context from
uninitialized memory which could corrupt the floating-point state. See
also:
https://developer.arm.com/documentation/ddi0403/d/System-Level-Architecture/System-Level-Programmers--Model/ARMv7-M-exception-model/Exception-return-behavior?lang=en
/*
* According to "ARMv7-M Architecture Reference Manual" section B1.5.6
* "Exception entry behavior" the return address is half-word aligned.
*/
ef->register_pc = (void *)
((uintptr_t) _ARMV7M_Thread_dispatch & ~((uintptr_t) 1));
ef->register_xpsr = 0x01000000U;
}
}
Close#4923.
Makes the code in bsp/x86_64/amd64/start/start.S and the stack frame
setup by _CPU_Context_Initialize align the stack properly according to
what is expected by the x86-64 SysV ABI
Interrupt IDs 1020, 1021, 1022, and 1023 are reserved as special INTIDs
and should be ignored for normal RTEMS operation as they signal internal
changes in the interrupt controller that RTEMS itself is directly
causing.
There were some problems in the math calculating RAM length and NOCACHE
origin location. These have been resolved in this patch allowing RTEMS
RAM space to be relocated above 0xfe00000 to better accomodate other
elements in the system.
In e95dabc77 approvers/maintainer was moved to general/maintainer. It was moved
back in the last commit to test permissions though I forgot to note it there.
Re-add since things are working as expected.
This change is to avoid anyone who is in the 'approvers' group for maintenance
showing up as approvers. Access goes only one level up from the current group
so having approval groups one level down was causing issues with me, Chris and
Kinsey showing up.
Add directives to get and set the priority of an interrupt vector.
Implement the directives for the following BSP families:
* arm/lpc24xx
* arm/lpc32xx
* powerpc/mpc55xxevb
* powerpc/qoriq
Implement the directives for the following interrupt controllers:
* GICv2 and GICv3 (arm and aarch64)
* NVIC (arm)
* PLIC (riscv)
Update #5002.
Move the timerfd impelemntation from linux compat code to cpukit/score/src. Use
it to implement the new system calls for timerfd. Add a hook to kern_tc
to allow timerfd to know when the system time has stepped. Add kqueue
support to timerfd. Adjust a few names to be less Linux centric.
RelNotes: YES
Reviewed by: markj (on irc), imp, kib (with reservations), jhb (slack)
Differential Revision: https://reviews.freebsd.org/D38459
L_LINT macro is used with negative numbers [i.e.
L_LINT(time_freq, -MAXFREQ)], it could cause undefined
behavior. It should be similar to the L_RSHIFT(v, n) macro.
MFC after: 2 weeks
Reviewed by: cy
Pull Request: https://github.com/freebsd/freebsd-src/pull/769
Signed-off-by: Dmitriy Alexandrov <d06alexandrov@gmail.com>
The SPDX folks have obsoleted the BSD-2-Clause-FreeBSD identifier. Catch
up to that fact and revert to their recommended match of BSD-2-Clause.
Discussed with: pfg
MFC After: 3 days
Sponsored by: Netflix
Mainly focus on files that use BSD 2-Clause license, however the tool I
was using misidentified many licenses so this was mostly a manual - error
prone - task.
The Software Package Data Exchange (SPDX) group provides a specification
to make it easier for automated tools to detect and summarize well known
opensource licenses. We are gradually adopting the specification, noting
that the tags are considered only advisory and do not, in any way,
superceed or replace the license texts.
The clock from the ARM timer is derived from the system clock. This clock can
change dynamically e.g. if the system goes into reduced power or in low power
mode. Thus the clock speed adapts to the overall system performance
capabilities. For accurate timing it is recommended to use the system timers.
if BSP_CLOCK_USE_SYSTEMTIMER = 1, use the System Timer, otherwise use the ARM
Timer.
In addtion to 1023, the GICC_IAR register may return 1022 as a special value.
Simply check for a valid interrupt vector for the dispatching.
Check the GICC_IAR again after the dispatch to quickly process a next interrupt
without having to go through the interrupt prologue and epiloge.
In addtion to 1023, the GICC_IAR register may return 1022 as a special value.
Simply check for a valid interrupt vector for the dispatching.
Check the GICC_IAR again after the dispatch to quickly process a next interrupt
without having to go through the interrupt prologue and epiloge.
Make the clock driver parameters configurable. Use the maximum counter
frequency to get the best time resolution. Decouple the CPU counter from the
timecounter. Make the tick catch up handling more robust. Add a validation
test for the tick catch up.
The processor mask implementation uses flsl() from <strings.h> which is
only BSD visible. Move the implementation to a separate header file to
hide it from the API level. This fixes build errors with GCC 14.
GitLab has it's own flavour and I've kept that in mind.
This greatly simplifies the document as well GitLab is quite a bit more verbose
than Trac was no need to replicate information.
This is just a first revision we can expand it as necessary.
It's clearly not required not sure why the gitlab-org repo uses it. I think
it's fine because * is a wildcard and ** is typically used for a recursive glob.
While it works it's not necessary.
This also adds stubs in for the rest of the architectures and owners.
Note that any users not currently on our test instance aren't in here we will
have to add them after.
Replace the BSP_CONSOLE_MINOR BSP option for the Xilinx Zynq BSPs with the new
BSP option ZYNQ_UART_KERNEL_IO_BASE_ADDR. Move the kernel I/O support to a
shared file.
Make the initialization and polled functions independent of the Termios
context. This helps to implement the kernel I/O support without a dependency
on the Termios framework.
Move declarations of bsp_interrupt_get_affinity() and
bsp_interrupt_set_affinity() to <bsp/irq-generic.h>. Canonicalize the
<bsp/irq.h> includes.
Implement bsp_interrupt_get_affinity() and bsp_interrupt_set_affinity() only if
needed (usually RTEMS_SMP).
Provide stub implementations for i386 to fix build errors.
The parameters are unused due to API constraints. The functions are
used through function pointers. Alternative implementations may use the
parameters.
Update #4862.
Pass the parameter of the clock interrupt handler to
Clock_driver_support_at_tick() and Clock_driver_timecounter_tick(). This makes
it possible to use the interrupt handler argument in clock drivers.
Use the interrupt handler provided by Clock_driver_support_install_isr() to
avoid local delarations of Clock_isr().
Update #4862.
We observed a strange behavior of the 1Hz timer when running cFS on Zynq
RPU. After some investigation, we reduced the error to the truncation
issue. This patch fixes the issue.
When unmounting a JFFS2 filesystem, any outstanding write buffers must
be flushed to disk. In some circumstances, these write buffers are
instantiated by a garbage collection pass and as such no inode number is
associated with it. Due to the way that JFFS2 processes these garbage
collection passes, a write buffer without any associated inodes will not
be flushed unless it is forced with jffs2_flush_wbuf_pad().
When resetting the QSPI FIFOs, the driver was reading write-only bits of
a register for status information when it was actually in a different
register. This corrects the driver so that it reads the correct status
bits.
Add an opportunistic page cache to the xnandpsu driver since it does not
implement partial page reads and common filesystem access patterns
perform multiple reads from the same page. This has been seen to provide
a 10x speedup to read speeds and a 2x speedup on first initialization
when used with JFFS2.
This fixes a logic inversion that was preventing creation of a Bad Block
Table (BBT) from scratch on devices that lack one. This was discovered
during upstream integration testing. The BBT management layer in this
driver is not designed to be easily testable other than on real hardware.
This adds a test to very that remounting a completely used, but empty
NOR JFFS2 filesystem is possible. Previously, this triggered an edge
condition in the JFFS2 scan code that prevented remount of a correctly
formed and uncorrupted filesystem.
On JFFS2 file systems on NOR flash or dataflash that does not have spare
area for metadata and thus does not invoke delayed writes, it is
possible to put the file system into a state where all blocks have been
written to and all files have been deleted from the filesystem. There is
a bug in the JFFS2 file system scan routine that detects this situation
as a corrupted file system since the scan routine relies on "used" space
to discriminate a valid file system when there are blocks that need to
be erased. The correct fix would require a partial rewrite of the scan
routine, so instead this patch tracks the space marked as obsolete along
with space at the end of each block that is otherwise too small to
contain a JFFS2 node so that it can me compared with the dirty space.
Corrupted data (or otherwise non-JFFS2 blocks) will still cause this
check to fail as corrupted data isn't recognized as obsoleted (deleted)
nodes.
The RAM origin had an unnecessary alignment requirement and the RAM
lenght had a reference that couldn't possibly be used. This removes
those obsolete option specifiers.
Make sure there is no spurious empty line between the gcov info and the
*** END OF GCOV INFO BASE64 ***
line. This helps to recalculate the hash correctly at the consumer
side.
This adds a pthread that runs the test as well to increase test
coverage. The original test would have passed if all threads returned
the address of the Init task's errno since no additional threads or
tasks were checked.
Stacked mode doubles the number of sectors and device size while
parallel mode doubles the sector size and the device size. Make sure
that this is accounted for in the device size accessor.
When the switch to waf occurred, the SMP support in i386 was left out
causing it to accumulate a minor amount of cruft. This enables SMP for
the i386 BSPs that support it and updates them for the API drift that
has occurred since the change.
`tms570_sci_interrupt_handler()` is called when an RX interrupt fires. It checks
in the register `FLR`, the `RXRDY` bit (Receiver ready flag - indicate that the
SCIRD contains new data). If it is set, it calls
`tms570_sci_read_received_chars()`.
`tms570_sci_read_received_chars()` checks the register `RD` against 0. If it is
non zero, it returns 1 to indicate that one byte was read.
In the old behavior, if it is zero, the function returns 0 to indicate that no
data was read.
The new behavior is to not silently drop 0x00 bytes. Ignoring 0x00 bytes is fine
when working with printable text (which, I assume, is how this driver was
tested), but as soon as the UART is used in non canonical (raw) mode, with
potentially 0x00 bytes, these bytes will be silently dropped, causing issues in
the data/protocol layer above.
Update #4982.
This enables the VME support for the MVME2500. Note that the PCIe
support from libbsd is used. So you need the related libbsd patches for
this to work.
If the drivers in libbsd are not enabled, the linker should not pick up
anything from this patch.
Add a function that allows to set the polarity (active-low / negative
edge triggered or active-high / positive edge triggered) and sense
(level or edge sensitive) of the external interrupts.
This alters the API for rtems_cache_coherent_add_area to allow reporting
of failures that can occur during the process of adding a new area to
the coherent cache heap.
Do not set pins to the default function before the actual setting is
applied. If a pin setting needs to be done in a certain order, then
this should be done explicitly through multiple calls to
tms570_bsp_pinmmr_config().
Update #4982.
Remove obsolete tms570ls3137_hdk_with_loader BSP variant. With the new
memory origin/size build options this variant is no longer required.
Update #4982.
Change license to BSD-2-Clause according to file history and contributor
agreements. Add Doxygen file comments.
Update #3053.
Update #3707.
Update #4982.
This cleans up outputUsesInterrupts usage with rtems_termios_device_mode
enum values. The outputUsesInterrupts member was typed as an int, named
as if it were a boolean value, and used as if it were a
rtems_termios_device_mode enum. In this patch, values assigned to
outputUsesInterrupts have been converted to the corresponding
rtems_termios_device_mode enum value, conversions from
deviceOutputUsesInterrupts have been made explicit, and uses of
rtems_termios_device_mode enum values with deviceOutputUsesInterrupts
have been converted to booleans.
When the CPU_CACHE_SUPPORT_PROVIDES_RANGE_FUNCTIONS definition was added
to AArch64 cache management, it obsoleted the *_1_data/instruction_line
functions. These have been removed since they are no longer referenced.
The AArch64_instruction_cache_inner_shareable_invalidate_all function is
only used when RTEMS_SMP is defined, so only define it in that
circumstance.
When encountering an error during filesystem creation, fat_fd must be
cleaned up appropriately once the file is opened. There was an
opportunity for a resource leak due to jumping to the incorrect error
handling label.
Revert JFFS2 to non-granular locking. This makes the superblock
available to the delayed work that it owns so that delayed work
processing can ensure that shared resources are being exclusively
accessed. This change only affects NAND systems. Moving forward with
granular locking will require a significant investment of time in
producing a test harness that doesn't require hardware such that this
can get a reasonable amount of test coverage.
The TM27 support may define TM27_INTERRUPT_VECTOR_ALTERNATIVE to provide
an alternative software generated interrupt request which is raised by
_TM27_Raise_alternative() and cleared by _TM27_Clear_alternative().
Both functions shall return an RTEMS status code. This interrupt vector
may be used to test the interrupt controller support on targets which do
not provide generic software generated interrupts.
Update #3716.
Expose functions to directly manipulate the bad block table (BBT). These
functions are necessary to correct possible BBT corruption caused by
bugs in the BBT management layer.
The XNandPsu_EraseBlock function takes a target device and a block
offset for erasure. Ensure the block offset is within the size of the
target device.
When marking the trailing blocks on a device as reserved for Bad Block
Table usage, ensure that the correct blocks are marked. This resolves an
off-by-one error that was marking one block too low and leaving the last
block in the device unmarked.
The Bad Block Table is a per-device catalog of the dispositions of each
block in the device. Only read enough data to determine the dispositions
of blocks for the device being read.
The xnandpsu driver includes functionality to map back and forth between
the flash-based BBT and the memory-based BBT with the values in each
being a bitwise inversion of each other. This resolves several bugs in
this process and simplifies the inversion from operating on the block
representation to operating on the entire BBT entry (4 blocks, 2 bits
per block, one byte total).
Bugs resolved in XNandPsu_ConvertBbt():
* The calculation of memory BBT entry offset was off by a factor of 4
* The entry offset into the flash BBT has been removed since each flash
BBT directly describes the flash space it is contained within and has
no reference to other devices in the chip
Bugs resolved in XNandPsu_WriteBbt():
* The BBT length calculated was reduced to NumTargetBlocks from
NumBlocks since only the relevant portion of the in-memory BBT should
be written to the flash-based BBT space
* An offset was applied to values retrieved from the in-memory BBT so
that only the relevant portion was converted and written to the
flash-based BBT
The XNandPsu_IsBlockBad() function is insufficient for JFFS2 to
determine whether a block is usable since the function does not report
reserved blocks. JFFS2 is also unable to use the last 4 blocks of each
target attached to the NAND controller since they are reserved for the
Bad Block Table (BBT), but not necessarily marked as such.
Hi all, this is my first email patch submission and my first contribution to RTEMS, so please give any
feedback you have!
This patch enables interrupt driven data reception on USART ports on
STM32F4 series chips. This feature is gated behind the config flag
BSP_CONSOLE_USE_INTERRUPTS. If this flag is not set to True, the older
polling implementation will be used. I tested this feature on STM32F401CE
(blackpill) and STM32 Nucleo F411RE boards, with both capable of keeping
up with a 115200 baud continous data stream. With the older polling
implementation, both would drop bytes at 9600 baud. In addition, I
updated the implementation of usart_set_attributes to support changing
the baud rate of the USART port based on the input speed.
Place the vector table in the start section so that the overlay can be
avoided if we execute from internal flash. The problem is that when the
POM is enabled, the ECC cannot be enabled for the internal flash.
With this, it is possible to use GPIOs as CS pins in the LPSPI. To avoid
additional complexity, the GPIOs will have the same limitations as the
native (hardware) CS pins.
The GPIO CS feature adds a number of extra code when starting SPI
transfers on this controller. Therefore it is possible to disable the
additional code by just setting the IMXRT_LPSPI_MAX_CS option to 0. In
that case only native CS pins are supported.
At the moment, this feature is only enabled on i.MXRT1166 by default
because it is not tested on i.MXRT1050. But it should work there too.
Device trees allow mixing different kinds of GPIOs in one property. For
that it is usefull to only provide a pointer to an arbitrary location in
the property and initialize a GPIO from that.
For uniprocessor configurations, the selection of
RTEMS_MULTIPROCESSOR_RESOURCE_SHARING results in a priority ceiling
mutex.
Build the MrsP validation tests only if RTEMS_SMP is enabled.
Update #3716.
This imports the TTC hardware definitions for the triple timer counters
on various Xilinx platforms. This was imported as specified in the
VERSION file in this commit.
The code coverage build is usually done without compiler optimizations.
This results in long performance test runtimes and useless performance
results.
Enable an individual testsuite only through the corresponding enable.
Use BUILD_TESTS to change the default value of the individual testsuite
enable options. This allows the user to enable all testsuites with
specific exceptions.
For example, this builds all testsuites except the benchmarks:
[arch/bsp]
BUILD_TESTS = True
BUILD_BENCHMARKS = False
Thread_Life_state is used as a bitfield, but is declared as an enum.
This converts the enum typedef to a uint32_t typedef and associated bit
definitions.
The xnandpsu driver conditionally tries to wrap page index to NAND chip
size causing an off-by-one error where the first page of the second chip
is not wrapped correctly. This removes the conditional so that page
index is always wrapped.
On configurations where multiple NAND chips are in use, the erasure
loop in XNandPsu_Erase() can reset the loop counter variable once it
gets to blocks in the second chip causing an infinite loop overwriting
parts of the first chip. This change ensures that the loop counter is
not accidentally reset.
Prefer use of XNandPsu_Erase instead of XNandPsu_EraseBlock since the
XNandPsu driver does not expose the primitives necessary to ensure
device readiness after the operation is complete.
This fixes some issues in the Xilinx support code that are critical to
support the Cortex-R5F cores present in my Xilinx SoCs. The imported
Cortex-R5 xil_cache.c matches the existing information in
bsps/shared/xil/VERSION.
Merge the timecounter and CPU counter support for the leon3 BSP family.
Remove now unused functions from the CPU counter support of the erc32
and leon3 BSPs.
Update #4954.
Share the timecounter instance between the clock and the CPU counter.
This greatly simplifies the clock driver since we have to do the device
selection only in one place, the CPU counter support.
Update #4954.
Using the auto reload counter which generates the clock ticks for the
timecounter or CPU counter is quite difficult and only works in
uniprocessor configurations.
Update #4954.
This improves the standard compatibility of API headers. It fixes
errors like this if RTEMS_MULTIPROCESSING is enabled:
cpukit/include/rtems/score/processormask.h: In function 'uint32_t _Processor_mask_Find_last_set(const Processor_mask*)':
cpukit/include/rtems/score/processormask.h:339:21: error: 'flsl' was not declared in this scope
339 | return (uint32_t) __BIT_FLS( CPU_MAXIMUM_PROCESSORS, a );
| ^~~~~~~~~
Force use of addresses that would be translated by TTBR1 to cause a
translation fault. RTEMS on AArch64 does not use TTBR1 and so attempted
translation of that address range could cause unexpected behavior in the
form of other exception types since TTBR1 is never set.
This provides a non-noop implementation of spinlocks for JFFS2 in terms
of RTEMS mutexes. POSIX spinlocks are not appropriate for the types of
actions that occur during a held JFFS2 spinlock and can cause bad
dispatch level conditions.
Ensure that a single locking session is carried over OOB writes
including the OOB read that is sometimes required. This removes the
possibility of another write occurring between the read and write that
would make the write incorrect.
Explicitly test the 64-bit integer division and modulo operations. They
are essential for the timekeeping services. On most 32-bit targets,
they need a software implementation.
Update #3716.
Measure the runtime of 1000 no-operation instructions. This may help to
investigate runtime measurement variations which do not result from code
generation or source code changes.
The runtime measurement of a function which does nothing and just
returns gives an estimate of the measurement overhead.
Update #3716.
Fix a potential stack corruption in uniprocessor configurations during
start multitasking.
The system initialization uses the interrupt stack. A first level
interrupt shall never interrupt a context which uses the interrupt
stack. Such a use would lead to stack corruption and undefined system
behaviour. Unfortunately, in uniprocessor configurations this was the
case. Multiprocessing is started using _CPU_Context_restore(). The
caller of this function (_Thread_Start_multitasking()) uses the
interrupt stack. Later we have in cpukit/score/cpu/sparc/cpu_asm.S:
mov %g1, %psr ! restore status register and
! **** ENABLE TRAPS ****
ld [%o1 + G5_OFFSET], %g5 ! restore the global registers
ld [%o1 + G7_OFFSET], %g7
! Load thread specific ISR dispatch prevention flag
ld [%o1 + ISR_DISPATCH_DISABLE_STACK_OFFSET], %o2
! Store it to memory later to use the cycles
ldd [%o1 + L0_OFFSET], %l0 ! restore the local registers
ldd [%o1 + L2_OFFSET], %l2
ldd [%o1 + L4_OFFSET], %l4
ldd [%o1 + L6_OFFSET], %l6
! Now restore thread specific ISR dispatch prevention flag
st %o2, [%g6 + PER_CPU_ISR_DISPATCH_DISABLE]
ldd [%o1 + I0_OFFSET], %i0 ! restore the input registers
ldd [%o1 + I2_OFFSET], %i2
ldd [%o1 + I4_OFFSET], %i4
ldd [%o1 + I6_FP_OFFSET], %i6
ldd [%o1 + O6_SP_OFFSET], %o6 ! restore the output registers
Between the ENABLE TRAPS and the restore of the output registers, we
still use the stack of the caller and interrupts may be enabled. If an
interrupt happens in this code block, the interrupt stack is
concurrently used which may lead to a crash.
Fix this by adding a new function _SPARC_Start_multiprocessing() for
uniprocessor configurations. This function first sets the stack pointer
to use the stack of the heir thread.
Close#4955.
Casts from uint64_t to void* must go through a uintptr_t cast to avoid
warnings when building with the ILP32 ABI since this is otherwise an
implicit truncation to 32bits for a pointer.
This reworks the JFFS2 delayed work queue to avoid use of
on-chain/off-chain semantics since they vary in behavior under
RTEMS_DEBUG and are not guaranteed to be safe to use in SMP systems.
This adds all delayed work structs to the chain on FS init and does not
remove them until umount.
Updates #4956
When a buffer is modified by both hardware components such as DMA and by
software components, the buffer cache state must be kept in sync so that
data is not accidentally thrown away during future invalidations.
When polling hardware registers in high performance situations, don't
rely on usleep or other standard sleep functions since they will
necessarily rely on kernel ticks to be woken up. This can easily cause
an immense reduction in throughput.
Make some performance tests compatible to other RTEMS versions which do
not allow changing the priority of a task which owns a priority
inheritance mutex.
Update #3716.
The <rtems/test-info.h> header file is required for every RTEMS test
program. Move the RTEMS test printer support to a dedicated header file
<rtems/test-printer.h>. This removes an unnecessary dependency to the
RTEMS printer support in <rtems/test-info.h>.
Tests using the RTEMS Testing Framework no longer depend on the
<rtems/printer.h>.
Doing the enabled-by processing just for the ldflags and just for the
link custom commands is confusing. Use an option instead which is
intended to be used for such use cases.
Add _CPU_Get_TLS_thread_pointer() to get the thread pointer which is
used to get the address of thread-local storage objects associated with
a thread.
Update #4920.
Fix rtems_configuration_get_interrupt_stack_size() for some code models.
The _ISR_Stack_size symbol has an arbitrary absolute address and may not
be representable in the code model used by the compiler.
Update #4953.
Store symbols with an arbitrary absolute address such as _TLS_Size,
_TLS_Alignment, _TLS_Data_size, and _TLS_BSS_size in an object to avoid issues
with some code models.
Update #4953.
The internal JFFS2 locking does not guarantee that delayed writes will
not step on other reads and writes to the device. This adds locking to
prevent that in the JFFS2 NAND interworking layer.
The changes here ensure correct cache maintenance around DMA operations.
One cache flush was missing and two cache invalidations occurred before
the corresponding read that would make them necessary.
This moves delayed work to a temporary chain to prevent a locking
inversion between the delayed work lock and the alloc_sem lock. Delayed
work is now processed after the delayed work lock is released. Locking
order is any JFFS2 locks before the delayed work lock.
This adds `-u__extendsfdf2` to the `ldflags` for the dl07, dl08, and
dl09 tests to force the inclusion of `__extendsfdf2` in the base image.
This function is part of the GCC software floating point library and is
used in the tests to convert floats to doubles when calling `printf`.
This allows for conditionally setting 'ldflags' in scripts that use
`link_cc` and `link_cxx`. The immediate use case is allowing a linker
flag to be used only for MicroBlaze builds of certain tests.
- Add the section alignment to the size as the allocator may not
provide correctly aligned memory
- Only include symbols in the section when locating symbols. The
powerpc was incorrectly adding SDATA BSS symbols to the BSS offset
overrunning the section
Closes#4950
The MAIR index currently assigned (1) for uncached memory segments is
not configured properly for this purpose. Instead, this switches
uncached memory segment flags to MAIR index 2 which is properly
configured for uncached inner and outer shareable domains.
This change requires an rtems-tools update for symbol generation.
Working architectures:
- aarch64
- arm
- powerpc
- sparc
No newlib TLS support but checked:
- i386
- m69k
Updates #4920
The current ARM support in libdebugger does not cover Cortex-M series
cores since it requires support for CP14 system register accessor
instructions. Cortex-M series cores support debug monitor mode, but its
configuration is accessed by memory mapped registers instead of using
CP14. This omits building libdebugger from BSPs that use a cortex-m ABI
flag.
Updates #4924.
The Regulator is an application support class which is used to
deal with the scenario where there is a bursty input source
which needs to be metered out to a destination sink. The maximum
size of bursts needs to be known and the delivery method must
be configured to deliver messages at a rate that allows the
traffic to not overflow.
Make sure we have enough POSIX key value pairs available. This fixes a test
failure on some targets.
Make objects and functions static. Initialize variable to get rid of warnings.
Fixes the following Coverity warning:
** CID 1539495: Integer handling issues (CONSTANT_EXPRESSION_RESULT)
/bsps/shared/dev/rtc/mcp7940m.c: 317 in mcp7940m_set_time()
Basically coverity warns that (buf[...] & 0x7) can't be bigger than 7.
Just remove the unnecessary comparison.
Define the RTEMS version in the wscript. Optionally use a VERSION file
to change the default values of the wscript. Allow the command line
option --rtems-version to override __RTEMS_MAJOR__. Remove support for
command line configurable options (--rtems-option).
Rename internal define RTEMS_VERSION_VC_KEY to
RTEMS_VERSION_CONTROL_KEY.
The MCP7940M is a I2C RTC chip. The new driver uses the dev/i2c API to
support the RTC. It is written with the intention, that the driver can
be adapted to other RTCs with a similar register layout by just
replacing the initialization function.
Part of the new HAL requires float_t type definition. For that it includes
math.h header file which unfortunately brings a lot of math functions in.
One of them, 'log' conflicts with FreeBSD log function defined in systm.h
Provided hack is really just a crude work around this issue. Never thought
that Motorola 68881 would help me even in 21st century.
This patch updates STM32 H7 HAL source files. The files are taken from two
STM projects from their github.com repositories:
(i)
https://github.com/STMicroelectronics/stm32h7xx_hal_driver.git
The project files are still available under BSD-3 license
and the version/commit used is:
d5fc8d05fc16fa2a2a2f948cf6c6ab39e78358e1
which represents post Release v1.11.1 development tree.
(ii)
https://github.com/STMicroelectronics/cmsis_device_h7.git
The project files were re-licensed from previous BSD-3 to Apache 2.0
license. Fortunately the project does not contain NOTICE file so no need
to do anything special when used in RTEMS.
The project version/commit imported is:
6d5ef249bec5177e0e2a0880ed62df2132874d99
which is code-wise Release v1.10.3 exactly.
The LEON3_IrqCtrl_Adev object is only defined if LEON3_IRQAMP_BASE is
not defined.
The LEON3_Timer_Adev object is only defined if LEON3_GPTIMER_BASE is not
defined.
Close#4850.
The flash configuration is something very board specific. So move the
file to a board specific location. Beneath that, not all controllers and
configurations need the flash config right at the address 0 of the
flash. For example on the i.MXRT11xx, the config has an offset for some
flash types.
The imx-gpio driver used in i.MX and i.MXRT BSPs generates a name based
on a fixed string. The original code only used one digit for the
controller. With the 13 GPIO controllers of the i.MXRT1166, that isn't
enough any more. This patch extends the name to two digits which should
be enough for the next controller generations.
CAVEAT: license change from BSD to Apache2 license!
Explanation:
The imported files come from CMSIS v5 project available on:
https://github.com/ARM-software/CMSIS_5/tree/develop
The files imported are located inside the CMSIS/Core/Include
project sub-directory.
The project does not provide any NOTICE file in its root directory nor
in the directory of the imported files.
The NOTICE file and its usage in the Apache 2 license was/is
so far the only issue mentioned in discussion of RTEMS developers/users
when considering inclusion of the code under Apache 2 license
into the RTEMS project.
Since the CMSIS v5 project is free from this legal hinder, we may freely
use it and update files to the latest version.
Technical: the patch replaces code from 2015 with the latest version
which brings quite a lot of bug fixes and most importantly opens
possibilities to support MCUs based on new ARM cores.
This adds support for the STM32H750B-DK discovery kit. This kit includes
a built-in STLINKv3 debugger which provides a USB serial bridge for
USART3. USART1 is routed to the Arduino header and USART2 is routed to
the STMOD connector. This BSP reuses what would otherwise be duplicated
files from the stm32h747i-disco BSP. Note that system_stm32h7xx.c has
been imported from the STM repository with two minor changes wrapped
with #if __rtems__. This hardware has been tested with hello and ticker.
rtems_task_wake_after takes a parameter in terms of a count of clock
ticks and not a measure in a subunit of seconds. This updates
documentation to reflect that. This also makes obvious the caveat about
the first tick wait not being a whole tick and points the user at a
replacement for better accuracy.
Updates #4772
Remove the BSP_POWER_DOWN_AT_FATAL_HALT BSP option. Applications should
do the customization of the system termination with an initial fatal
extension.
QEMU is known to fail certain tests intermittently due to clock tick
delivery issues. This defines those tests as intermittent for BSPs
intended to run on QEMU alone.
Updates #4922
Updates #4072
Previously, the erase ioctl call would ignore flash regions and erase at
the absolute address passed in. This adds a check for if the region is
set and adds the region offset accordingly.
The NOR driver was not written with SMP systems and caching in mind.
This makes the IsBusy flag volatile for updates across cores and
introduces cache flushing and invalidation where necessary for data
manipulated by the DMA engine in the QSPI peripheral.
The page size ioctl requires an int pointer to retrieve the page size.
The test currently uses a size_t which mostly works fine for systems
where size_t and int are the same size, but can leave junk data in the
upper bits when they differ in size causing an assert in the test to
fail in some cases. This updates the variable to the correct type.
When given the exact bounds of a sector, the current algorithm
calculates that 3 sectors need to be erased. This corrects the
calculation such that only 1 sector needs to be erased for erasures that
exactly match sector boundaries.
This change causes NOR writes to be broken according to page boundaries.
Writes across page boundaries cause the writes beyond the boundary to
fail silently. This also introduces a new function that will explicitly
write pages.
Add <rtems/termiosdevice.h> which does not depend on <rtems/libio.h> to
provide rtems_termios_device_context and rtems_termios_device_handler.
For polled serial device drivers, this removes a header file dependency
to the full file system support.
On most i.MX* the upper bits in SW_PAD_CTL are reserved. On some chips,
like the i.MXRT1166, they are a domain write protection. Setting them to
1 can have unexpected side effects.
The device tree uses these bits for some flags. Make sure that they are
not accidentally written to some value.
The different variants of the i.MXRT have some minimal differences in
the fsl_flexspi_nor_config.h. Make sure that the header supports the
different chips.
The defines for the different clock frequencies in the
fsl_clock_config.h do not represent the clock frequencies that have been
set up in the registers. Remove them to avoid someone trusting in
correct values.
If a different chip variant is used in the i.mxrt BSP, a different
header would have to be included. Make sure that the fsl-edma driver
uses a header that doesn't have to be adapted.
The mcux_sdk has a different interface for getting the clock for
IMXRT11xx than for getting it in IMXRT10xx. Adapt simple drivers to
support that interface.
The number of GPIO devices along with each of their particular
configurations is application-specific. Encoding this information as
build options also introduced a lot of clutter.
The updated files were extracted from:
https://www.zlib.net/zlib-1.2.13.tar.xz
The archive had an SHA-256 hash value of:
d14c38e313afc35a9a8760dadf26042f51ea0f5d154b0630a31da0540107fb98
Close#4902.
The new amd64efi BSP supports:
- multiboot2 boot format. Runs well with GRUB.
- console based on either EFI simple text output or GOP-based framebuffer
- clock based on EFI event/timer API
- early console using either hard-wired PC-AT serial or just memory buffer
- with EFI support disabled the BSP is more or less equivalent to amd64 BSP
with multiboot2 support
The library is imported in minimalist version just to support future
amd64efi BSP.
The FreeBSD tree commit id with imported libefi version is:
ce7b20e5129cf0f269951b313d336a9c7d54d790
There was a cyclic dependency: For RTEMS_STATIC_ANALYSIS we needed
basedefs.h in assert.h. For RTEMS_UNREACHABLE() we needed _Assert() from
assert.h in basedefs.h.
Fix this by introducing _Debug_Unreachable() in basedefs.h.
Add RTEMS_FUNCTION_NAME to basedefs.h and use it in basedefs.h and
assert.h.
Close#4900.
The MMU must be enabled during mapping changes and TLB invalidations.
When this is not the case, TLB updates do not occur correctly in all
cases. This is especially apparent when changing a block entry to a
table entry when remapping small memory ranges in an otherwise
contiguous block.
Do not warn about excessive shifts in imported Xilinx driver code when
building with the ILP32 ABI. The two instances of this warning produce
the desired effect of setting the upper address bits to 0.
This allows any builds targeting an installed RTEMS BSP to override
headers in the installed BSP reliably, including headers previously
installed by that or other builds. This includes applications, network
stacks, libraries, and any other builds.
Closes#4896
This patch includes the spec/build options for the riscv kendrytek210
BSP variant. It includes options to allow the frdme310arty console
UART to be used on multiple BSPS, device tree options, memory
options, and other required options for the variant.
Updates #4876
This patch adds support for the Kendryte K210 RISC-V BSP variant.
The SoC uses the existing Interrupt Controller, Timer, and console UART.
It only needs SoC specific initialization and an embedded device tree binary
similar to the polarfire SoC BSP.
Updates #4876
This patch adds the k210 device tree source and the corresponding
device tree blob encoded in the header which is used for the
embedded device tree blob for the Kendryte K210 BSP variant.
Updates #4876
The existing include path only works from inside the RTEMS build. This
fixes the include path to work both in the RTEMS build and with builds
of external apps since this file gets installed with the BSP.
Use the mtimecmp from the PLIC/CLINT initialization in the clock driver. This
register is defined by the device tree and does not assume a fixed mapping.
In SMP configurations, check that we run on a configured processor. If not,
then there is not much that can be done since we do not have a stack available
for this processor. Just loop forever in this case. Do this in assemlby to
ensure that no stack memory is used.
This adds write buffer and bad block support required for JFFS2
operation on NAND devices. This also adds the minor modifications
necessary for RTEMS support in the Linux header stubs and in wbuf.c.
Memory and NOR backed applications should experience no difference in
operation since they do not expose the callbacks required for write
buffer support.
The kernel space parts of standard system header files were protected
against misuse by a preprocessor error directive. This approach does
not work well with tools such as Doxygen. Provide the kernel space
items only if the required defines are present.
The comment above bintime2timespec() says:
When converting between timestamps on parallel timescales of differing
resolutions it is historical and scientific practice to round down.
However, the delta_nsec value is a time difference and not a timestamp. Also
the rounding errors accumulate in the frequency accumulator, see hardpps().
So, rounding to the closest integer is probably slightly better.
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/604
Let A be the current calculation of the frequency accumulator (pps_fcount)
update in pps_event()
scale = (uint64_t)1 << 63;
scale /= captc->tc_frequency;
scale *= 2;
bt.sec = 0;
bt.frac = 0;
bintime_addx(&bt, scale * tcount);
bintime2timespec(&bt, &ts);
hardpps(tsp, ts.tv_nsec + 1000000000 * ts.tv_sec);
and hardpps(..., delta_nsec):
u_nsec = delta_nsec;
if (u_nsec > (NANOSECOND >> 1))
u_nsec -= NANOSECOND;
else if (u_nsec < -(NANOSECOND >> 1))
u_nsec += NANOSECOND;
pps_fcount += u_nsec;
This change introduces a new calculation which is slightly simpler and more
straight forward. Name it B.
Consider the following sample values with a tcount of 2000000100 and a
tc_frequency of 2000000000 (2GHz).
For A, the scale is 9223372036. Then scale * tcount is 18446744994337203600
which is larger than UINT64_MAX (= 18446744073709551615). The result is
920627651984 == 18446744994337203600 % UINT64_MAX. Since all operands are
unsigned the result is well defined through modulo arithmetic. The result of
bintime2timespec(&bt, &ts) is 49. This is equal to the correct result
1000000049 % NANOSECOND.
In hardpps(), both conditional statements are not executed and pps_fcount is
incremented by 49.
For the new calculation B, we have 1000000000 * tcount is 2000000100000000000
which is less than UINT64_MAX. This yields after the division with tc_frequency
the correct result of 1000000050 for delta_nsec.
In hardpps(), the first conditional statement is executed and pps_fcount is
incremented by 50.
This shows that both methods yield roughly the same results. However, method B
is easier to understand and requires fewer conditional statements.
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/604
Use local variables for the captured timehand and timecounter in pps_event().
This fixes a potential issue in the nsec preparation for hardpps(). Here the
timecounter was accessed through the captured timehand after the generation was
checked.
Make a snapshot of the relevent timehand values early in pps_event(). Check
the timehand generation only once during the capture and event processing. Use
atomic_thread_fence_acq() similar to the other readers.
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/604
Since 32c203577a5e by phk in 1999 (Make even more of the PPSAPI
implementations generic), the "nsec" parameter of hardpps() is a time
difference and no longer a time point. Change the name to "delta_nsec"
and adjust the comment.
Remove comment about a clock tick adjustment which is no longer in the code.
Pull Request: https://github.com/freebsd/freebsd-src/pull/640
Reviewed by: imp
The third argument to this function indicates whether the supplied
ticker is fixed or variable, i.e. requiring calibration. Give this
argument a type and name that better conveys this purpose.
Reviewed by: kib, markj
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35459
(which is used to calculate cputime from cpu ticks) has some imprecision and,
worse, huge timestep (about 20 minutes on 4GHz CPU) near 53.4 days of elapsed
time.
kern_time.c/cputick2timespec() (it is used for clock_gettime() for
querying process or thread consumed cpu time) Uses cputick2usec()
and then needlessly converting usec to nsec, obviously losing
precision even with fixed cputick2usec().
kern_time.c/kern_clock_getres() uses some weird (anyway wrong)
formula for getting cputick resolution.
PR: 262215
Reviewed by: gnn
Differential Revision: https://reviews.freebsd.org/D34558
Interrupts must be disabled during MMU remapping since the majority of
RTEMS including interrupts expects normal memory mapping semantics such
as unaligned accesses.
The whole cache invalidation and flushing functions only ended up
flusing the first N-1 levels of cache due to an off by one error. This
resovles that issue and makes consistent the usage of levels as they
relate to caching.
To ensure data consistency, the cache much be flushed before disabling
the MMU. When the MMU is disabled, all accesses are treated as
non-cachced and thus will bypass the cache.
The application configuration options are documented in
"cpukit/doxygen/appl-config.h". Since the application configuration
option defines are also present in multiple test program sources, the
"#OPTION" references cannot be mapped to a unique definition. Add an
anchor for each option and reference it to avoid the issues with the
multiple definitions.
Update #3994.
Rename the constaints section of application configuration options from
"Value Constraints" in "Constraints. Adjust the constraint wording
accordingly. This is in line with the RTEMS Classic API Guide.
Update #3994.
Each BSP may optionally provide default values for some application
configuration options. Remove the documentation of these items, since
the BSP provided defines are not application configuration options, they
are optional default values. Clarify CONFIGURE_DISABLE_BSP_SETTINGS
accordingly and move it into the "General System Configuration" group.
Update #3994.
The LEON3_Boot_Cpu global object is only used by start.S. Move the definition
of this object to start.S and use a local symbol .Lbootcpuindex for it.
Use a compare-and-swap instruction to assign the boot CPU. This allows a
concurrent initialization.
Close#4845.
This imports the full xil_exception.h instead of an empty stub. This is
required for some Xilinx drivers. The imported files adhere to the
current VERSION file.
This makes xil/sleep.h a stub to prevent multiple differing definitions
of sleep functions from toolchain and local headers. The non-standard
sleep definitions were not in use and can be added later if needed.
There are different cases for _ISR_lock_ISR_disable() and
_ISR_lock_ISR_enable() in the case RTEMS_SMP is defined or
RTEMS_PROFILING is defined, so remove the related code.
../../../cpukit/include/rtems/score/threaddispatch.h: In function '_Thread_Dispatch_disable':
../../../cpukit/include/rtems/score/threaddispatch.h:231:14: warning: 'lock_context' may be used uninitialized [-Wmaybe-uninitialized]
231 | cpu_self = _Thread_Dispatch_disable_critical( &lock_context );
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../../cpukit/include/rtems/score/threaddispatch.h:210:32: note: by argument 1 of type 'const ISR_lock_Context *' to '_Thread_Dispatch_disable_critical' declared here
210 | static inline Per_CPU_Control *_Thread_Dispatch_disable_critical(
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../../cpukit/include/rtems/score/threaddispatch.h:225:21: note: 'lock_context' declared here
225 | ISR_lock_Context lock_context;
| ^~~~~~~~~~~~
Update #4662.
The objxilinxsupport build object was accidentally included twice in
some of the ZynqMP BSPs by two different drivers that required it. This
commit manually deduplicates the inclusions by moving that inclusion to
the BSP. Duplication of object inclusions is considered a bug and can
cause race conditions in the build system.
The QSPI memory is initialized and used only when the BSP configure file
sets QSPI memory size to non-zero value. Currently QSPI is run in memory
mapped mode which allows future RTEMS binary linkage and upload into QSPI
memory.
Sponsored-By: Precidata
Add an assert to _Thread_Initialize_scheduler_and_wait_nodes() which may
help a static analyzer. Use a do/while loop since we have at least one
scheduler.
Update #4832.
This adds helper functions for working with NOR flash connected to the
Xilinx GQSPI controller. The helper functions are based on Xilinx's
QSPIPSU flash interrupt example.
This adds Xilinx's driver for the Xilinx GQSPI controller embedded in
the ZynqMP SoC. Within that device alone, it is possible to access this
peripheral from MicroBlaze, ARMv7, and ARMv8 cores. The imported files
are and should be able to remain unmodified. Import information is kept
in bsps/shared/dev/spi/VERSION.
This adds basic TLS relocation support for AArch64 to libdl. This lets
loadable modules use TLS symbols hosted in the main binary. This does
not allow loadable modules to host their own TLS symbols.
Merge the "default" and "default-by-variant" attributes. Use an
"enabled-by" expression to select the default value based on the enabled
set. This makes it possible to select default values depending on other
options. For example you could choose memory settings based on whether
RTEMS_SMP is enabled or disabled.
The change was tested by comparing the output of
./waf bspdefaults
before and after the change.
Replace the variant patterns in the default-by-variant list with an
explicit list of matching BSPs.
The change was tested by comparing the output of
./waf bspdefaults
before and after the change.
The RTEMS and Linux JFFS2 compression support have a slightly different
interface. This fixes a bug introduced by:
commit 17c3d536ca
Author: Yang Yang <yang.yang29@zte.com.cn>
Date: Thu Jan 28 02:55:35 2021 -0800
jffs2: check the validity of dstlen in jffs2_zlib_compress()
KASAN reports a BUG when download file in jffs2 filesystem.It is
because when dstlen == 1, cpage_out will write array out of bounds.
Actually, data will not be compressed in jffs2_zlib_compress() if
data's length less than 4.
There is a regular need in the kernel to provide a way to declare
having a dynamically sized set of trailing elements in a structure.
Kernel code should always use “flexible array members”[1] for these
cases. The older style of one-element or zero-length arrays should
no longer be used[2].
This code was transformed with the help of Coccinelle:
(linux-5.19-rc2$ spatch --jobs $(getconf _NPROCESSORS_ONLN) --sp-file script.cocci --include-headers --dir . > output.patch)
@@
identifier S, member, array;
type T1, T2;
@@
struct S {
...
T1 member;
T2 array[
- 0
];
};
-fstrict-flex-arrays=3 is coming and we need to land these changes
to prevent issues like these in the short future:
../fs/minix/dir.c:337:3: warning: 'strcpy' will always overflow; destination buffer has size 0,
but the source string has length 2 (including NUL byte) [-Wfortify-source]
strcpy(de3->name, ".");
^
Since these are all [0] to [] changes, the risk to UAPI is nearly zero. If
this breaks anything, we can use a union with a new member name.
[1] https://en.wikipedia.org/wiki/Flexible_array_member
[2] https://www.kernel.org/doc/html/v5.16/process/deprecated.html#zero-length-and-one-element-arrays
Link: https://github.com/KSPP/linux/issues/78
Build-tested-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/lkml/62b675ec.wKX6AOZ6cbE71vtF%25lkp@intel.com/
Acked-by: Dan Williams <dan.j.williams@intel.com> # For ndctl.h
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Use kzalloc rather than duplicating its implementation, which
makes code simple and easy to understand.
Signed-off-by: Haowen Bai <baihaowen@meizu.com>
[rw: Fixed printk string]
Signed-off-by: Richard Weinberger <richard@nod.at>
By making filler_t the same as read_folio, we can use the same function
for both in gfs2. We can push the use of folios down one more level
in jffs2 and nfs. We also increase type safety for future users of the
various read_cache_page() family of functions by forcing the parameter
to be a pointer to struct file (or NULL).
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Andreas Gruenbacher <agruenba@redhat.com>
In preparation for unifying the read_cache_page() and read_folio()
implementations, make jffs2_do_readpage_unlock() get the inode
from the page instead of passing it in from read_cache_page().
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
inode->i_mutex has been replaced with inode->i_rwsem long ago. Fix
comments still mentioning i_mutex.
Signed-off-by: hongnanli <hongnan.li@linux.alibaba.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Add a rcu argument to the ->get_acl() callback to allow
get_cached_acl_rcu() to call the ->get_acl() method in the next patch.
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Building with W=1 shows a few warnings for empty macros:
fs/jffs2/scan.c: In function 'jffs2_scan_xattr_node':
fs/jffs2/scan.c:378:66: error: suggest braces around empty body in an 'if' statement [-Werror=empty-body]
378 | jffs2_sum_add_xattr_mem(s, rx, ofs - jeb->offset);
| ^
fs/jffs2/scan.c: In function 'jffs2_scan_xref_node':
fs/jffs2/scan.c:434:65: error: suggest braces around empty body in an 'if' statement [-Werror=empty-body]
434 | jffs2_sum_add_xref_mem(s, rr, ofs - jeb->offset);
| ^
fs/jffs2/scan.c: In function 'jffs2_scan_eraseblock':
fs/jffs2/scan.c:893:88: error: suggest braces around empty body in an 'if' statement [-Werror=empty-body]
893 | jffs2_sum_add_padding_mem(s, je32_to_cpu(node->totlen));
| ^
Change all these macros to 'do { } while (0)' statements to avoid the
warnings and make the code a little more robust.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
KASAN report a slab-out-of-bounds problem. The logs are listed below.
It is because in function jffs2_scan_dirent_node, we alloc "checkedlen+1"
bytes for fd->name and we check crc with length rd->nsize. If checkedlen
is less than rd->nsize, it will cause the slab-out-of-bounds problem.
jffs2: Dirent at *** has zeroes in name. Truncating to %d char
==================================================================
BUG: KASAN: slab-out-of-bounds in crc32_le+0x1ce/0x260 at addr ffff8800842cf2d1
Read of size 1 by task test_JFFS2/915
=============================================================================
BUG kmalloc-64 (Tainted: G B O ): kasan: bad access detected
-----------------------------------------------------------------------------
INFO: Allocated in jffs2_alloc_full_dirent+0x2a/0x40 age=0 cpu=1 pid=915
___slab_alloc+0x580/0x5f0
__slab_alloc.isra.24+0x4e/0x64
__kmalloc+0x170/0x300
jffs2_alloc_full_dirent+0x2a/0x40
jffs2_scan_eraseblock+0x1ca4/0x3b64
jffs2_scan_medium+0x285/0xfe0
jffs2_do_mount_fs+0x5fb/0x1bbc
jffs2_do_fill_super+0x245/0x6f0
jffs2_fill_super+0x287/0x2e0
mount_mtd_aux.isra.0+0x9a/0x144
mount_mtd+0x222/0x2f0
jffs2_mount+0x41/0x60
mount_fs+0x63/0x230
vfs_kern_mount.part.6+0x6c/0x1f4
do_mount+0xae8/0x1940
SyS_mount+0x105/0x1d0
INFO: Freed in jffs2_free_full_dirent+0x22/0x40 age=27 cpu=1 pid=915
__slab_free+0x372/0x4e4
kfree+0x1d4/0x20c
jffs2_free_full_dirent+0x22/0x40
jffs2_build_remove_unlinked_inode+0x17a/0x1e4
jffs2_do_mount_fs+0x1646/0x1bbc
jffs2_do_fill_super+0x245/0x6f0
jffs2_fill_super+0x287/0x2e0
mount_mtd_aux.isra.0+0x9a/0x144
mount_mtd+0x222/0x2f0
jffs2_mount+0x41/0x60
mount_fs+0x63/0x230
vfs_kern_mount.part.6+0x6c/0x1f4
do_mount+0xae8/0x1940
SyS_mount+0x105/0x1d0
entry_SYSCALL_64_fastpath+0x1e/0x97
Call Trace:
[<ffffffff815befef>] dump_stack+0x59/0x7e
[<ffffffff812d1d65>] print_trailer+0x125/0x1b0
[<ffffffff812d82c8>] object_err+0x34/0x40
[<ffffffff812dadef>] kasan_report.part.1+0x21f/0x534
[<ffffffff81132401>] ? vprintk+0x2d/0x40
[<ffffffff815f1ee2>] ? crc32_le+0x1ce/0x260
[<ffffffff812db41a>] kasan_report+0x26/0x30
[<ffffffff812d9fc1>] __asan_load1+0x3d/0x50
[<ffffffff815f1ee2>] crc32_le+0x1ce/0x260
[<ffffffff814764ae>] ? jffs2_alloc_full_dirent+0x2a/0x40
[<ffffffff81485cec>] jffs2_scan_eraseblock+0x1d0c/0x3b64
[<ffffffff81488813>] ? jffs2_scan_medium+0xccf/0xfe0
[<ffffffff81483fe0>] ? jffs2_scan_make_ino_cache+0x14c/0x14c
[<ffffffff812da3e9>] ? kasan_unpoison_shadow+0x35/0x50
[<ffffffff812da3e9>] ? kasan_unpoison_shadow+0x35/0x50
[<ffffffff812da462>] ? kasan_kmalloc+0x5e/0x70
[<ffffffff812d5d90>] ? kmem_cache_alloc_trace+0x10c/0x2cc
[<ffffffff818169fb>] ? mtd_point+0xf7/0x130
[<ffffffff81487dc9>] jffs2_scan_medium+0x285/0xfe0
[<ffffffff81487b44>] ? jffs2_scan_eraseblock+0x3b64/0x3b64
[<ffffffff812da3e9>] ? kasan_unpoison_shadow+0x35/0x50
[<ffffffff812da3e9>] ? kasan_unpoison_shadow+0x35/0x50
[<ffffffff812da462>] ? kasan_kmalloc+0x5e/0x70
[<ffffffff812d57df>] ? __kmalloc+0x12b/0x300
[<ffffffff812da462>] ? kasan_kmalloc+0x5e/0x70
[<ffffffff814a2753>] ? jffs2_sum_init+0x9f/0x240
[<ffffffff8148b2ff>] jffs2_do_mount_fs+0x5fb/0x1bbc
[<ffffffff8148ad04>] ? jffs2_del_noinode_dirent+0x640/0x640
[<ffffffff812da462>] ? kasan_kmalloc+0x5e/0x70
[<ffffffff81127c5b>] ? __init_rwsem+0x97/0xac
[<ffffffff81492349>] jffs2_do_fill_super+0x245/0x6f0
[<ffffffff81493c5b>] jffs2_fill_super+0x287/0x2e0
[<ffffffff814939d4>] ? jffs2_parse_options+0x594/0x594
[<ffffffff81819bea>] mount_mtd_aux.isra.0+0x9a/0x144
[<ffffffff81819eb6>] mount_mtd+0x222/0x2f0
[<ffffffff814939d4>] ? jffs2_parse_options+0x594/0x594
[<ffffffff81819c94>] ? mount_mtd_aux.isra.0+0x144/0x144
[<ffffffff81258757>] ? free_pages+0x13/0x1c
[<ffffffff814fa0ac>] ? selinux_sb_copy_data+0x278/0x2e0
[<ffffffff81492b35>] jffs2_mount+0x41/0x60
[<ffffffff81302fb7>] mount_fs+0x63/0x230
[<ffffffff8133755f>] ? alloc_vfsmnt+0x32f/0x3b0
[<ffffffff81337f2c>] vfs_kern_mount.part.6+0x6c/0x1f4
[<ffffffff8133ceec>] do_mount+0xae8/0x1940
[<ffffffff811b94e0>] ? audit_filter_rules.constprop.6+0x1d10/0x1d10
[<ffffffff8133c404>] ? copy_mount_string+0x40/0x40
[<ffffffff812cbf78>] ? alloc_pages_current+0xa4/0x1bc
[<ffffffff81253a89>] ? __get_free_pages+0x25/0x50
[<ffffffff81338993>] ? copy_mount_options.part.17+0x183/0x264
[<ffffffff8133e3a9>] SyS_mount+0x105/0x1d0
[<ffffffff8133e2a4>] ? copy_mnt_ns+0x560/0x560
[<ffffffff810e8391>] ? msa_space_switch_handler+0x13d/0x190
[<ffffffff81be184a>] entry_SYSCALL_64_fastpath+0x1e/0x97
[<ffffffff810e9274>] ? msa_space_switch+0xb0/0xe0
Memory state around the buggy address:
ffff8800842cf180: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
ffff8800842cf200: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>ffff8800842cf280: fc fc fc fc fc fc 00 00 00 00 01 fc fc fc fc fc
^
ffff8800842cf300: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
ffff8800842cf380: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
==================================================================
Cc: stable@vger.kernel.org
Reported-by: Kunkun Xu <xukunkun1@huawei.com>
Signed-off-by: lizhe <lizhe67@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Extend some inode methods with an additional user namespace argument. A
filesystem that is aware of idmapped mounts will receive the user
namespace the mount has been marked with. This can be used for
additional permission checking and also to enable filesystems to
translate between uids and gids if they need to. We have implemented all
relevant helpers in earlier patches.
As requested we simply extend the exisiting inode method instead of
introducing new ones. This is a little more code churn but it's mostly
mechanical and doesnt't leave us with additional inode methods.
Link: https://lore.kernel.org/r/20210121131959.646623-25-christian.brauner@ubuntu.com
Cc: Christoph Hellwig <hch@lst.de>
Cc: David Howells <dhowells@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Set rp_size to zero will be ignore during remounting.
The method to identify whether we input a remounting option of
rp_size is to check if the rp_size input is zero. It can not work
well if we pass "rp_size=0".
This patch add a bool variable "set_rp_size" to fix this problem.
Reported-by: Jubin Zhong <zhongjubin@huawei.com>
Signed-off-by: lizhe <lizhe67@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
The log of this problem is:
jffs2: Error garbage collecting node at 0x***!
jffs2: No space for garbage collection. Aborting GC thread
This is because GC believe that it do nothing, so it abort.
After going over the image of jffs2, I find a scene that
can trigger this problem stably.
The scene is: there is a normal dirent node at summary-area,
but abnormal at corresponding not-summary-area with error
name_crc.
The reason that GC exit abnormally is because it find that
abnormal dirent node to GC, but when it goes to function
jffs2_add_fd_to_list, it cannot meet the condition listed
below:
if ((*prev)->nhash == new->nhash && !strcmp((*prev)->name, new->name))
So no node is marked obsolete, statistical information of
erase_block do not change, which cause GC exit abnormally.
The root cause of this problem is: we do not check the
name_crc of the abnormal dirent node with summary is enabled.
Noticed that in function jffs2_scan_dirent_node, we use
function jffs2_scan_dirty_space to deal with the dirent
node with error name_crc. So this patch add a checking
code in function read_direntry to ensure the correctness
of dirent node. If checked failed, the dirent node will
be marked obsolete so GC will pass this node and this
problem will be fixed.
Cc: <stable@vger.kernel.org>
Signed-off-by: Zhe Li <lizhe67@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
When debug (print) macros are not enabled, change them to use the
no_printk() macro instead of <nothing>. This fixes gcc warnings when
-Wextra is used:
../cpukit/libfs/src/jffs2/src/nodelist.c:255:37: warning: suggest braces around empty body in an ‘else’ statement [-Wempty-body]
../cpukit/libfs/src/jffs2/src/nodelist.c:278:38: warning: suggest braces around empty body in an ‘else’ statement [-Wempty-body]
../cpukit/libfs/src/jffs2/src/nodelist.c:558:52: warning: suggest braces around empty body in an ‘else’ statement [-Wempty-body]
../cpukit/libfs/src/jffs2/src/xattr.c:1247:58: warning: suggest braces around empty body in an ‘if’ statement [-Wempty-body]
../cpukit/libfs/src/jffs2/src/xattr.c:1281:65: warning: suggest braces around empty body in an ‘if’ statement [-Wempty-body]
Builds without warnings on all 3 levels of CONFIG_JFFS2_FS_DEBUG.
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Richard Weinberger <richard@nod.at>
Cc: linux-mtd@lists.infradead.org
Signed-off-by: Richard Weinberger <richard@nod.at>
Low-end configurations may want to have the HTIF support removed.
Enable the option by default. Fix formatting. Fix node validity
checks.
Updates #4779.
Xilinx's upstream ILP32 xil_cache.h header is out of date and broken.
This provides a copy of the LP64 header in place of the ILP32 header
since the LP64 header includes all the correct types to work with either
data model.
The address of the nandpsu peripheral is specific to the ZynqMP SoC and
not relevant to other devices that might have one or more instances of
this peripheral.
This adds Xilinx's driver for the Xilinx NAND controller embedded in the
ZynqMP SoC. Within that device alone, it is possible to access this
peripheral from MicroBlaze, ARMv7, and ARMv8 cores. This has been added
to the hardware ZynqMP BSPs since QEMU does not support emulation of
this peripheral. This driver supports polled operation only. The
imported files are and should be able to remain unmodified. Import
information is kept in bsps/shared/dev/nand/VERSION.
This support code is necessary for many Xilinx-provided bare metal device
drivers supported on ARM, AArch64, and MicroBlaze platforms. Support for
all of these architectures is kept under bsps/include due to multiple
architecture variants being supported which requires complex logic in
the build system. The imported files are and should be able to remain
unmodified. Import information is kept in bsps/shared/xil/VERSION.
Currently generic RISC-V BSPs (riscv/riscv) that start with rv64 and not
rv64*_medany will start at 0x70000000. This adds high maintenance overhead
and deviates from almost all other RISC-V-based OSes and baremetal programs
that start at 0x80000000. Further, testing now has to account for an extra
parameter (medany or not) that doubles the number of BSPs need to be
tested.
This commit defaults all RV64 BSPs to use medany code model to allow starting
all BSPs at 0x80000000. BSPs that require different code models and/or
start addresses are custom and need to add their own entries.
Updates #4775
Since fdt_next_tag() in a public API function all input parameters,
including the fdt blob should not be trusted. It is possible to forge
a blob with invalid property length that will cause integer overflow
during offset calculation. To prevent that, validate the property length
read from the blob before doing calculations.
Signed-off-by: Tadeusz Struk <tadeusz.struk@linaro.org>
Message-Id: <20221005232931.3016047-1-tadeusz.struk@linaro.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
There is no need to check the VALID_DTB repeatedly, and can be combined
into one if statement.
Signed-off-by: Jia Xianhua <jiaxianhua@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Correctly calculate the array index and shift value in
_Processor_mask_To_uint32_t(). The bugs had no impact yet since this
function was always called with a zero value for the index in RTEMS.
With the old build system in RTEMS 5 that was possible by just
overwriting BOARD_Sdram_Config and setting a custom
ATSAM_MEMORY_SDRAM_SIZE during building the BSP. In the new build system
that ATSAM_MEMORY_SDRAM_SIZE is set exclusively by the selected SDRAM
chip.
This patch adds the possibility to specify a "custom-0x100000" or
similar as SDRAM type where the number gives the SDRAM size.
A SPI transfer where the Rx or Tx buffer is set to NULL currently
transfers or overwrites data starting from address 0x00000000 via DMA.
This patch changes the DMA setup so that dummy transfers are done.
Just reading / writing to a single location is simpler than changing the
whole logic of the transfer depending on the passed buffers.
Add ref-clock-num identifiers to the device tree to ensure that
interfaces use the correct clocks even when some are not used due to
unconnected MII busses. This also adjusts the default ZynqMP PHY
attachment to RGMII-ID which was the default before device trees were
introduced.
CID 1506523: Unchecked return value from library (CHECKED_RETURN)
CID 1506522: Unchecked return value from library (CHECKED_RETURN)
CID 1437618: Unchecked return value from library (CHECKED_RETURN)
Close#4718
This enables the tracing of interrupt entry/exit events through an
application configuration option. The interrupt processing can be
viewed with Trace Compass using rtems-record-lttng from the RTEMS Tools.
Update #4769.
This fixes build errors seen when building with console interrupts
enabled. A few places were missing bspopts.h includes, and one of the
UART functions was not defined.
Use of malloc implies errno which adds TLS dependencies and prevents use
of this FDT wrapper library in BSP initialization code. This change
makes use of rtems_malloc and rtems_calloc which avoid TLS dependencies.
This changes the ZynqMP device tree parsing over to direct libfdt calls
to avoid inclusion of malloc() in the base BSP which currently causes
sp01 to fail due to unexpected use of TLS space.
In the original implementation, level -1 was unused and all levels could
have block-like descriptors (level 2 block descriptors are called page
descriptors). When support for level -1 page tables was added the
constraint on level -1 block descriptors was not honored. This prevents
block descriptors from being mapped at level -1 since the hardware will
not map them properly.
This patch changes the license to BSD-2 for all source files where the
copyright is held by Aeroflex Gaisler, Cobham Gaisler, or Gaisler Research.
Updates #3053.
This patch changes the license to BSD-2 for all source files where the
copyright is held by Aeroflex Gaisler, Cobham Gaisler, or Gaisler Research.
Some files also includes copyright right statements from OAR and/or
embedded Brains in addition to Gaisler.
Updates #3053.
This patch changes the license to BSD-2 for all source files where the
copyright is held by Aeroflex Gaisler, Cobham Gaisler, or Gaisler Research.
Some files also includes copyright right statements from OAR and/or
embedded Brains in addition to Gaisler.
Updates #3053.
This patch changes the license to BSD-2 for all source files where the
copyright is held by Aeroflex Gaisler, Cobham Gaisler, or Gaisler Research.
Some files also includes copyright right statements from OAR and/or
embedded Brains in addition to Gaisler.
Updates #3053.
This patch changes the license to BSD-2 for all source files where the
copyright is held by Aeroflex Gaisler, Cobham Gaisler, or Gaisler Research.
Some files also includes copyright right statements from OAR and/or
embedded Brains in addition to Gaisler.
Updates #3053.
This patch changes the license to BSD-2 for all source files where the
copyright is held by Aeroflex Gaisler, Cobham Gaisler, or Gaisler Research.
Some files also includes copyright right statements from OAR and/or
embedded Brains in addition to Gaisler.
Updates #3053.
In SMP configurations, there may be no software interrupt handler
installed when the software interrupt is processed. Add the new
interrupt handler dispatch variant
bsp_interrupt_handler_dispatch_unlikely() for this special case.
Use the tm27 support to test a spurious interrupt. This helps to run the
validation test case on targets which have no software interrupt available for
tests (for example riscv/PLIC/CLINT in the SMP configuration).
Where CPU_ENABLE_ROBUST_THREAD_DISPATCH == TRUE, the only supported interrupt
level allowed to set is 0 (interrupts enabled). This constraint is enforced by
the API level functions which return an error status for other interrupt
levels.
The extra digest copy was introduced by an erroneous merge conflict resolution
for commit "crypto routines: Hint minimum buffer sizes to the compiler".
The AArch64 TRM specifies that when FPCR is set to trap floating point
exceptions, the FPSR exception bits are not set. This ensures that FPSR
is updated as FENV expects even if floating point exception traps are
enabled.
This adds a BSP variant for the ZynqMP BSP family to support the
Innoflight CFC-400X platform. To properly support the CFC-400X, device
trees were added to the ZynqMP platform due to both the optional
management interface as well as alternate physical configuration of the
ethernet interfaces.
- see ARM DUI 0646C Arm Cortex-M7 Devices Generic User Guide
"The RELOAD value is calculated according to its use.
For example, to generate a multi-shot timer with a period
of N processor clock cycles, use a RELOAD value of N-1.
If the SysTick interrupt is required every 100 clock pulses,
set RELOAD to 99."
- see routines used in CMSIS project for reference
Close#4746.
This fixes a problem with parsing the FDT compatible property by
replacing the RISCV_CONSOLE_IS_COMPATIBLE macro with calls to
the fdt_stringlist_contains function. The macro only works when
the compatible FDT entry is a single string and not a list of
strings. The new call will compare each item in the string list.
Close#4728.
By default, allocate the IDLE task storage areas from the RTEMS Workspace.
This avoids having to estimate the thread-local storage size in the default
configuration.
Add the application configuration option CONFIGURE_IDLE_TASK_STORAGE_SIZE to
request a static allocation of the task storage area for IDLE tasks.
Update #3835.
Update #4524.
Add the INTERNAL_ERROR_IDLE_THREAD_CREATE_FAILED fatal error in case the
creation of an idle thread fails. This may happen due to a failing create
extension provided by the application.
Place this member placed directly after the end of the common block so that
the structure offsets are as small as possible. This helps on instruction
set architectures with a very limited range for intermediate values. For
example, see the __aeabi_read_tp() implementation for ARM Thumb-1.
Update #3835.
Allow the IDLE stack allocator to change the stack size. This can be
used by applications with a very dynamic thread-local storage size to
adjust the thread storage area of the IDLE tasks dynamically.
Update #4524.
- Add support to get the parent address and size cells
- Provide support to get a reg prop address map
- Change getting a set of properties to uintptr_t
- Improve the debug mode of the ls command to print all props
Closes#4729
This patch adds new Raspberry pi 4B AArch64 BSP to the RTEMS Family. Currently
only LP64 ABI is supported. ILP32 is not supported. RAM starts from 0x80000 in
64Bit kernel mode and MMU from 0x0. All Raspberrypi Pi 4B models and Raspberry
Pi 400 are supported. All the IRQs are similiar to the older Raspberry pi 2 ARM
BSP.
Raspberry Pi 4B has 2 types of UARTs. Only PL011 serial is supported currently.
Mini-UART is not supported. Mini-UART is default UART on the board so it needs
to be disabled by adding "dtoverlay=disable-bt" to the config.txt. No support
for additional 4 PL011-UARTs on the board.
The raspberrypi.h includes many of the address required for the future
development of the RPi 4B BSP. This includes peripherals, ARM Timer, VideoCore
Timer, Watchdog, Mailbox, AUX, FIQs and IRQs.
The second allocation check was mistakenly rechecking the first
allocation. It now checks the correct allocation and ensures that names
is not NULL.
Updates #4462
If the bsp is integrated and supported a device tree
blob(dtb) then use dtb instead of using it from
the U-Boot (BSP_START_COPY_FDT_FROM_U_BOOT=False).
The Microchip PolarFire SoC support is implemented as a
riscv BSP variant to boot with any individual hart(cpu core)
or SMP based on the boot HARTID configurable and support
components are 4 CPU Cores (U54), Interrupt controller (PLIC),
Timer (CLINT), UART.
Add the basic Microchip PolarFire SoC device tree source and blob
The mpfs-dtb.h is generated by the bin2hex
https://github.com/padmaraob/bin2hex
1.Compile and build the bin2hex.c
$ gcc -o bin2hex bin2hex.c
2.Generate the mpfs.dtb from the mpfs.dts
$ dtc -O dtb -o mpfs.dtb mpfs.dts
3.Generate the mpfs-dtb.h Header file from the mpfs.dtb.
$ ./bin2hex mpfs.dtb
There is not just big-endian on ARM. We have two variants BE32
(obsolete) and BE8. The Cortex-R5F processor supports only BE8,
however, some TMS570 variants are BE32 internally. In GCC 8 and later,
the --be8 option is passed to the linker based on the selected
architecture or CPU. Use BE32 by default for the TMS570 BSP.
In GCC, see:
commit 63d03dcecdafe34715282a5155cfc2162375feca
Author: Richard Earnshaw <rearnsha@arm.com>
Date: Mon Jul 3 13:22:05 2017 +0000
[arm] Clean up generation of BE8 format images.
Note: Resending after learning how to use git send-email, please disregard previous message.
This fixes the riscv fe310 console driver fe310_uart_read function. The function
reads the RX status/data register to check if data is available, but discards
the data and reads it a seconds time.
Also cleared the interrupt enable bit in the first_open function.
Close#4719
All CPU ports used the same _CPU_Counter_difference() implementation. Remove
this CPU port interface and mandate a monotonically increasing CPU counter.
Close#3456.
context after they are finialized after r336539, so do not do it.
Submitted by: David CARLIER <devnexen gmail com>
MFC after: 1 month (after r336539)
Differential Revision: https://reviews.freebsd.org/D16059
copied_key and copied_salt are assigned with NULL and never used
otherwise. Remove the two variables and related code.
Reviewed by: pfg
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D16314
Mainly focus on files that use BSD 2-Clause license, however the tool I
was using mis-identified many licenses so this was mostly a manual - error
prone - task.
The Software Package Data Exchange (SPDX) group provides a specification
to make it easier for automated tools to detect and summarize well known
opensource licenses. We are gradually adopting the specification, noting
that the tags are considered only advisory and do not, in any way,
superceed or replace the license texts.
Mainly focus on files that use BSD 3-Clause license.
The Software Package Data Exchange (SPDX) group provides a specification
to make it easier for automated tools to detect and summarize well known
opensource licenses. We are gradually adopting the specification, noting
that the tags are considered only advisory and do not, in any way,
superceed or replace the license texts.
Special thanks to Wind River for providing access to "The Duke of
Highlander" tool: an older (2014) run over FreeBSD tree was useful as a
starting point.
and when doing it, consistently use explicit_bzero().
Update manual pages to match the behavior.
Reviewed by: pfg, allanjude, jmg
MFC after: 1 month
Differential Revision: https://reviews.freebsd.org/D16316
C99 allows array function parameters to use the static keyword for their
sizes. This tells the compiler that the parameter will have at least the
specified size, and calling code will fail to compile if that guarantee is
not met. However, this syntax is not legal in C++.
This commit reverts r300824, which worked around the problem for
sys/sys/md5.h only, and introduces a new macro: min_size(). min_size(x) can
be used in headers as a static array size, but will still compile in C++
mode.
Reviewed by: cem, ed
MFC after: 4 weeks
Sponsored by: Spectra Logic Corp
Differential Revision: https://reviews.freebsd.org/D8277
with more ILP and a preference for memory load instructions over large code
footprints with embedded immediate variables.
On amd64 CPUs from 2007-2008 there is not a significant change, but
amd64 CPUs from 2009-2010 get roughly 10% more throughput with this
code; amd64 CPUs from 2011-2012 get roughly 15% more throughput; and
AMD64 CPUs from 2013-2015 get 20-25% more throughput. The Raspberry
Pi 2 increases its throughput by 6-8%.
Sponsored by: Tarsnap Backup Inc.
Performance tested by: allanjude
MFC after: 3 weeks
This implements SHA-512/256, which generates a 256 bit hash by
calculating the SHA-512 then truncating the result. A different initial
value is used, making the result different from the first 256 bits of
the SHA-512 of the same input. SHA-512 is ~50% faster than SHA-256 on
64bit platforms, so the result is a faster 256 bit hash.
The main goal of this implementation is to enable support for this
faster hashing algorithm in ZFS. The feature was introduced into ZFS
in r289422, but is disconnected because SHA-512/256 support was missing.
A further commit will enable it in ZFS.
This is the follow on to r292782
Reviewed by: cem
Sponsored by: ScaleEngine Inc.
Differential Revision: https://reviews.freebsd.org/D6061
Use the C99 'static' keyword to hint to the compiler IVs and output digest
sizes. The keyword informs the compiler of the minimum valid size for a given
array. Obviously not every pointer can be validated (i.e., the compiler can
produce false negative but not false positive reports).
No functional change. No ABI change.
Sponsored by: EMC / Isilon Storage Division
cperciva's libmd implementation is 5-30% faster
The same was done for SHA256 previously in r263218
cperciva's implementation was lacking SHA-384 which I implemented, validated against OpenSSL and the NIST documentation
Extend sbin/md5 to create sha384(1)
Chase dependancies on sys/crypto/sha2/sha2.{c,h} and replace them with sha512{c.c,.h}
Reviewed by: cperciva, des, delphij
Approved by: secteam, bapt (mentor)
MFC after: 2 weeks
Sponsored by: ScaleEngine Inc.
Differential Revision: https://reviews.freebsd.org/D3929
The VRSAVE feature of the Altivec unit can be used to reduce the amount of
Altivec registers which need to be saved/restored during interrupt processing
and context switches.
In order to use the VRSAVE optimization a corresponding multilib (-mvrsave) is
required, see GCC configuration. The -mvrsave option must be added to the
ABI_FLAGS of the BSP.
Currently only the -mcpu=e6500 based QorIQ BSP support this optimization.
Update #4712.
Added support for Cobham Gaisler NOEL-V systems. The NOEL-V support
is implemented as a riscv BSP. Both 32-bit and 64-bit processor
systems are supported. Cobham Gaisler's NOEL-V RISC-V processor IP
is described here:
https://www.gaisler.com/NOELV
Compatible with the following NOEL-V FPGA example design ranges
available from Cobham Gaisler. Follow the links for free
bit-streams, DTS/DTB, user's manuals and quick-start guides:
- NOEL-ARTYA7-EX (https://www.gaisler.com/NOEL-ARTYA7)
- NOEL-PF-EX (https://www.gaisler.com/NOEL-PF)
- NOEL-XCKU-EX (https://www.gaisler.com/NOEL-XCKU)
Uses the shared GRLIB APBUART console driver "apbuart_termios.c".
APBUART devices are probed using device tree.
Closes#4225.
Issue an error message if an SMP-specific scheduler is used and
RTEMS_SMP is disabled. This might be a more informative compared to
compiler or linker errors.
The <rtems.h> header file should not unnecessarily include standard C
header files. The <string.h> and <limits.h> header includes were
removed in 2017.
Update #4662.
The ABI flags for the amd64 BSP contain the -Werror=return-type flag. There is no reason for this to be there so it has been removed. The same option has also been removed amd64.cfg file.
- Versal has IO mapped to the upper 64bit address space and
needs full 64bit addresses.
- Add xilinx_versal_aiedge for custom hardware
- Make the hardware settings the defaults and qemu as variants
Closes#4693
Added two pragmas to address, and changed the value of
AARCH64_EXCEPTION_MAKE_ENUM_64_BIT to INT_MAX because the old value was not
in range of an int.
Updates #4662
This adds the configure option CONFIGURE_POSIX_TIMERS_FACE_BEHAVIOR
which allows the application to choose whether to have the POSIX
timer_create() function follow the behavior defined by POSIX or
the FACE Technical Standard.
Updates #4691.
rtl-mdreloc-arm.c was used as the basis for rtl-mdreloc-aarch64.c. This lead
to some code being shared by the two files. The code was consolidated into
rtl-unwind-arm.c.
Closes#4686
A relative CLOCK_REALTIME time out shall not be affected by CLOCK_REALTIME
changes through clock_settime(). Since our CLOCK_REALTIME is basically just
CLOCK_MONOTONIC plus an offset, we can simply use the CLOCK_MONOTONIC watchdog
for relative CLOCK_REALTIME time outs.
Update #4690.
In rtems_rtl_symbol_global_add() the loop that gets to the end of the symbol
table used "unsigned long" to increment the index for the table. For most
architectures this resulted in 4, but with AArch64, it results in 8. This
resulted in the symbols being read in wrong. Changing this to void* along with
changing the RISC-V specific code for 8 byte pointers in rtems-tools to work
independent of the architecture.
Updates #4673Closes#4682
The rtems_task_delete() directive is basically just a combined pthread_cancel()
and pthread_join(). In addition, it removes the PTHREAD_DETACHED state. The
exit value returned by pthread_join() of threads cancelled by
rtems_task_delete() should reflect this by getting a PTHREAD_CANCELED value
instead of NULL which could be a normal exit value.
Close#4680.
Threads may join the thread termination of another thread using the
pthread_join() or rtems_task_delete() directives. The thread cancel operation
used a special case priority boosting mechanism implemented by
_Thread_Raise_real_priority(). The problem was that this approach
* is not transitive,
* does not account for priority adjustments of the calling task
while waiting for the join,
* does not support clustered scheduling, and
* does not detect deadlocks.
All these problems are fixed by using a priority inheritance thread queue for
the join operation.
Close#4679.
- Support DDRMC0 region 0 up to 2G in size
- Support DDRMC0 region 1 with DDR memory greater than 2G
up to the DDRMC0 max amount
- Extend the heap with region 1's memory
Closes#4684
The objects local table must be statically zero-initialized so that
_Objects_Get() and _Objects_Get_no_protection() return NULL if no object is
associated with the identifier.
Update #4678.
The uniprocessor schedulers had some special case logic for the
PRIORITY_PSEUDO_ISR priority. Tasks with a priority of PRIORITY_PSEUDO_ISR
were allowed to preempt a not preemptible task. If other higher priority task
are made ready while a PRIORITY_PSEUDO_ISR task preempts a not preemptible
task, then the other tasks run before the not preemptible task. This made the
RTEMS_NO_PREEMPT mode ineffective.
Remove the PRIORITY_PSEUDO_ISR special case logic. This simplifies the
uniprocessor schedulers. Move the uniprocessor-specific scheduler support to
the new header file <rtems/score/scheduleruniimpl.h>.
Close#2365.
The build option definitions were rearranged such that the option
definitions used in the linker script were not available. This caused
linker errors when building.
This alters the AArch64 page table generation and mapping code and MMU
configuration to use page table level 0 in addition to levels 1, 2, and
3. This allows the mapping of up to 48 bits of memory space and is the
maximum that can be mapped without relying on additional processor
extensions. Mappings are restricted based on the number of physical
address bits that the CPU supports.
In case the Newlib _REENT_THREAD_LOCAL configuration option is enabled, the
struct _reent is not defined (there is only a forward declaration in
<sys/reent.h>). Instead, the usual members of struct _reent are available as
dedicatd thread-local storage objects.
Update #4560.
Sort the .noinit* input sections by name first, then by alignment if two
sections have the same name. This allows the placement of begin/end symbols to
initialize some areas with a special value.
Update #4678.
When fdt_get_string return null, `namep` is not correctly reset.
From the document of `fdt_getprop_by_offset`, the parameter `namep` will
be always overwritten (that is, it will be overwritten without exception
of error occurance).
As for the caller (like
e097c097fe/native/jni/magiskboot/dtb.cpp (L42)),
the code may be like:
```cpp
size_t size;
const char *name;
auto *value = fdt_getprop_by_offset(fdt, prop, &name, &size);
```
and if `value == nullptr`, `size` is also be overwritten correctly but
`name` is not, which is quite inconsistent.
This commit makes sure `name` and `size` behavior consistently (reset to
reasonable value) when error occurs.
Signed-off-by: LoveSy <shana@zju.edu.cn>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Commit 21275b58a5 ("score: Static
Objects_Information initialization") introduced an off-by-one error in the
maintenance of inactive objects.
Close#4677.
Place the statically allocated object control blocks, local tables, and thread
queue heads into the dedicated .noinit intput sections. The output section is
not zero initialized. Placing these elements into the .noinit section reduces
the system initialization time by decreasing the .bss section size.
It may improve the cache efficiency since the mostly read local tables are
placed in a contiguous memory area.
Update #4678.
Sort the .noinit* input sections by name first, then by alignment if two
sections have the same name. This allows the placement of begin/end symbols to
initialize some areas with a special value.
Update #4678.
Exceptions during system initialization work only on targets which do not need
a registration of exception frames during the global construction. In
particular, targets which use the DWARF2 unwinder cannot use exceptions during
system initialization.
Use the existing WRITE_SR() abstraction to access the interrupt group 0 and 1
enable registers. This fixes the build for the AArch32 target.
Add BSP options which define the initial values of CPU Interface registers.
The modification to the hash are already naturally locked by
in_control_sx. Convert the hash lists to CK lists. Remove the
in_ifaddr_rmlock. Assert the network epoch where necessary.
Most cases when the hash lookup is done the epoch is already entered.
Cover a few cases, that need entering the epoch, which mostly is
initial configuration of tunnel interfaces and multicast addresses.
Reviewed by: melifaro
Differential revision: https://reviews.freebsd.org/D32584
Current logic always selects an IFA of the same family from the
outgoing interfaces. In IPv4 over IPv6 setup there can be just
single non-127.0.0.1 ifa, attached to the loopback interface.
Create a separate rt_getifa_family() to handle entire ifa selection
for the IPv4 over IPv6.
Differential Revision: https://reviews.freebsd.org/D31868
MFC after: 1 week
These functions were added in 2001 and are currently unused.
copyinfrom() looks to have never been used. copyinstrfrom() was used
for two weeks before the code was refactored to remove it's sole use.
Reviewed by: brooks, kib
Obtained from: CheriBSD
Sponsored by: DARPA
Differential Revision: https://reviews.freebsd.org/D24928
This particular state during a reconsider help request scheduler operation was
only covered by the existing test suites under some timing conditions.
Update #3716.
There were two bugs with MMU page use that were partially hiding each
other. The linker script page table section was 4x the size it needed to
be and the page table allocation routine was allocating pages PTRSIZE
times larger than it needed to. On ILP32, this resulted in incorrect but
functional allocation. On LP64, this resulted in allocation failures
earlier than expected.
At some point during system initialization, the idle threads are created.
Afterwards, the boot processor basically executes within the context of an idle
thread with thread dispatching disabled. On some architectures, the
thread-local storage area of the associated thread must be set in dedicated
processor registers. Add the new CPU port function to do this:
void _CPU_Use_thread_local_storage( const Context_Control *context )
Close#4672.
The compiler wraps fork(), etc. system calls if coverage generation is enabled.
These functions must be provided by the system. For RTEMS, they just return an
error status.
Update #4670.
Allow separate optimization flags for the BSP, cpukit, and tests. For example,
the BSP and cpukit may be built without optimization if coverage
instrumentation is enabled, however, the tests may still use optimization.
Update #4670.
The waf build system uses lists for tool flags. The build items may use
variable substitution. Add the ability to use the variable substitution in
lists. For example:
MORE_FLAGS = ['-more', '-flags']
flags:
- -some-flag
- ${MORE_FLAGS}
Before this change, the ${MORE_FLAGS} was substituted to "-more -flags". This
would be passed by waf as a single command line argument to the tool.
After this change, the ${MORE_FLAGS} list extends the flags list:
flags = ['-some-flag', '-more', '-flags']
This list extension is performed if a list element consists of exactly one
variable.
Update #4670.
Conditional expressions with inline functions are not optimized away if
optimization is disabled. Avoid such expressions to prevent dead
branches. It helps also during code review to immediately see if a loop
is used or not.
Use CLOCK_REALTIME and CLOCK_MONOTONIC for relative thread queue timeouts
instead of CLOCK_REALTIME_COARSE and CLOCK_MONOTONIC_COARSE. This fixes an
issue with clock_nanosleep() in combination with clock_gettime().
Close#4669.
The original file cpukit/libfs/src/ftpfs/tftpDriver.c
is split into two:
tftpfs.c - This file contains the code from tftpDriver.c
related to file system operations such as mount(),
open(), read(), and so on.
tftpDriver.c - In the original file remains only the code related
to networking. This code implements the Trivial
File Transfer Protocol (TFTP).
Moreover, the code is extended to support
* RFC 2347 TFTP Option Extension
* RFC 2348 TFTP Blocksize Option
* RFC 7440 TFTP Windowsize Option
Update #4666.
Thanks for the advice mentioned in the email.
This is my v3 patch for this problem.
Mounting jffs2 on nand flash will get message "failed: I/O error"
with the steps listed below.
1.umount jffs2
2.erase nand flash
3.mount jffs2 on it (this mounting operation will be successful)
4.do chown or chmod to the mount point directory
5.umount jffs2
6.mount jffs2 on nand flash
After step 6, we will get message "mount ... failed: I/O error".
Typical image of this problem is like:
Empty space found from 0x00000000 to 0x008a0000
Inode node at xx, totlen 0x00000044, #ino 1, version 1, isize 0...
The reason for this mounting failure is that at the end of function
jffs2_scan_medium(), jffs2 will check the used_size and some info
of nr_blocks.If conditions are met, it will return -EIO.
The detail is that, in the steps listed above, step 4 will write
jffs2_raw_inode into flash without jffs2_raw_dirent, which will
cause that there are some jffs2_raw_inode but no jffs2_raw_dirent
on flash. This will meet the condition at the end of function
jffs2_scan_medium() and return -EIO if we umount jffs2 and mount it
again.
We notice that jffs2 add the value of c->unchecked_size if we find
an inode node while mounting. And jffs2 will never add the value of
c->unchecked_size in other situations. So this patch add one more
condition about c->unchecked_size of the judgement to fix this problem.
Signed-off-by: Zhe Li <lizhe67@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
[1] https://en.wikipedia.org/wiki/Flexible_array_member
[2] https://github.com/KSPP/linux/issues/21
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
This reverts commit f2538f999345405f7d2e1194c0c8efa4e11f7b3a. The patch
stopped JFFS2 from being able to mount an existing filesystem with the
following errors:
jffs2: error: (77) jffs2_build_inode_fragtree: Add node to tree failed -22
jffs2: error: (77) jffs2_do_read_inode_internal: Failed to build final fragtree for inode #5377: error -22
Fixes: f2538f999345 ("jffs2: Fix possible null-pointer dereferences...")
Cc: stable@vger.kernel.org
Suggested-by: Hou Tao <houtao1@huawei.com>
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Richard Weinberger <richard@nod.at>
In jffs2_scan_eraseblock(), 'sumptr' is allocated through kmalloc() if
'sumlen' is larger than 'buf_size'. However, it is not deallocated in the
following execution if jffs2_fill_scan_buf() fails, leading to a memory
leak bug. To fix this issue, free 'sumptr' before returning the error.
Signed-off-by: Wenwen Wang <wenwen@cs.uga.edu>
Signed-off-by: Richard Weinberger <richard@nod.at>
Merge these two helpers into the only callers to get rid of some
amazingly bad calling conventions.
Suggested-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
In jffs2_add_frag_to_fragtree(), there is an if statement on line 223 to
check whether "this" is NULL:
if (this)
When "this" is NULL, it is used at several places, such as on line 249:
if (this->node)
and on line 260:
if (newfrag->ofs > this->ofs)
Thus possible null-pointer dereferences may occur.
To fix these bugs, -EINVAL is returned when "this" is NULL.
These bugs are found by a static analysis tool STCheck written by us.
Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Linux kernel tolerates C++ style comments these days. Actually, the
SPDX License tags for .c files start with //.
On the other hand, uapi headers are written in more strict C, where
the C++ comment style is forbidden.
I simply dropped these lines instead of fixing the comment style.
This code has been always commented out since it was added around
Linux 2.4.9 (i.e. commented out for more than 17 years).
'Maybe later...' will never happen.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Richard Weinberger <richard@nod.at>
free the symlink body after the same RCU delay we have for freeing the
struct inode itself, so that traversal during RCU pathwalk wouldn't step
into freed memory.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Need to tell the compiler that the acl entries follow the acl header.
Signed-off-by: Matthew Wilcox <mawilcox@microsoft.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
The false trigger is covered in:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99578
GCC 11 and 12 has been patched for constant pointer casts above
4K. This code casts a constant pointer within the first 4K
page. As a result the patch disables the warning.
Updates #4662
The patch "bsps/atsam: Fix type of options" missed to adapt some parts
of the yml. With that a custom value works well. But if no value is set,
configure doesn't fall back to the default value but instead just causes
an error. This patch fixes that.
Check that the state of rand() and lrand48() is thread-specific,
that they are properly initialized, and return the expected
sequence of pseudo-random numbers for default seed values.
The interrupt enables are used to implement a priority scheme in
bsp_interrupt_dispatch(). Enable interrupts through
bsp_interrupt_vector_enable() only if they are not already enabled to
not interfere with the priority scheme while
bsp_interrupt_vector_enable() is called during interrupt dispatching.
Sections with identical attributes may be contiguous with a respective
begin and end address which is not on a minimum region boundary. The
begin address is aligned down to the region base address. The end
address is aligned up to the region end address. Account for this in
the check for contiguous sections.
Update #4202.
A section may span up to the end of the address range. In this case the
end address is zero. Use the base address to check if a region should
be before another region.
Update #4202.
ATSAM_CONSOLE_DEVICE_INDEX and ATSAM_CONSOLE_DEVICE_TYPE have to be
integers like suggested by their description. Otherwise it's not
possible to select (for example) USART2 as console device.
This is the default configuration of the board out of the box.
Any other possible/supported configuration requires soldering,
so definitely not out of the box experience.
Sponsored-By: Precidata
This patch adds stm32h747i-disco-m4 BSP variant and puts it in sync
with the stm32h747i-disco BSP variant hardware support. That means,
only USART 1, 2 and UART 8 are enabled. Also SDRAM 2 is set to 32MB,
SDRAM 1 size is set to 0.
Sponsored-By: Precidata
This patch disables all U(S)ARTs which are not supported by the board
itself and its provided connectors. That means only USART1 and 2
and UART8 are enabled.
Sponsored-By: Precidata
Besides C files for the BSP variant the patch also provides license
clarification on system_stm32h7xx.c file which is provided
in boards/stm/stm32h747i-disco directory.
The files comes from STM32CubeH7 project and references "root directory"
in its license comment and it's not clear where this points out.
Let's add clarification comment about it and also based on it
and resulting license let's add SPDX license identifier.
Sponsored-By: Precidata
This is minimalist configuration for the stm32h757i-eval-m4 BSP provided
here. The only general enhancement worth mention is a flash origin address
configuration which is needed for simplification as M4 core boots
from second flash bank which starts at 0x8100000 by default. The boot
address of the core may be changed by using STM32CubeProgrammer. If done
so then also BSP configuration needs to be changed accordingly.
As the BSP variant is running on M4 core, there is also more configuration
changes required here. E.g. boot core and ABI (compilation flags)
in comparison with stm32h757i-eval BSP. On the other hand, C code is shared
completely with this BSP variant.
Sponsored-By: Precidata
This is minimalist configuration for the stm32h757i-eval BSP provided
here. The only general enhancement worth mention is boot core
configuration which is needed here as this is the first dual-core board
supported by stm32h7 BSP family and we need to choose boot core in order
to get C files compiling well.
Sponsored-By: Precidata
Besides C files for the BSP variant the patch also provides license
clarification on system_stm32h7xx.c file which is provided
in boards/stm/stm32h757i-eval directory.
The files comes from STM32CubeH7 project and references "root directory"
in its license comment and it's not clear where this points out.
Let's add clarification comment about it and also based on it
and resulting license let's add SPDX license identifier.
Sponsored-By: Precidata
REGION_WORK may be backed by external RAM which may not be initialized
in a time we need stack to work well. E.g. code loaded in flash,
stack allocated on in-cpu SRAM and data (REGION_WORK) on external SDRAM.
Sponsored-By: Precidata
Start code for most platforms requires hand-coded ASM but some can be
bootstrapped entirely in C, especially for paravirtualized platforms.
This change allows start code to be written in C where possible instead
of requiring architecture-specific ASM to bridge to C.
This expands the ability to substitute variables outside the current
limitation of values in options to asflags, cflags, cppflags, cxxflags,
ldflags, and includes. It is possible for all of these flags to utilize
user-defined information in config.ini, especially for paths to external
resources.
In the SMP_FATAL_MULTITASKING_START_ON_NOT_ONLINE_PROCESSOR test case,
the boot processor ends up in the idle body. Do not issue a fatal error
here and instead enter the idle loop. The secondary processor will
terminate the system.
Currently the following sequence causes a endless loop when extending an
IMFS file:
- Create a file with zero length and close it.
- Make sure nearly no allocatable memory is left.
- Open the file and write enough data into it that more than the
remaining memory will be used.
In that case when extending the IMFS file, the file currently need zero
blocks. If allocating enough new blocks fails, the already allocated new
blocks will be freed again.
The comparison of block>=old_blocks that has been used prior to this
patch compared two unsigned numbers. If old_blocks was zero, the
comparison of these two numbers always evaluated to true.
This patch frees the last block in a separate step to avoid this
problem.
Fixes#4639
The patch merges differences in FMC configuration between system_stm32h7xx.c
file generated by STM32CubeIDE for 743i-eval2 board and the current RTEMS
ext-mem-ctl.c file.
Sponsored-By: Precidata
In contrast to rtems_shell_main_loop(), this new function does not
perform all sorts of initialization based on environment settings. For
example, due to the use of isatty() in rtems_shell_main_loop() it is
impossible to run an interactive shell through a socket connection.
Separate the Interrupt Manager implementation from the generic Arm GICv3
support. Move parts of the Arm GICv3 support into a new header file. This
helps to support systems with a clustered structure in which multiple GICv3
instances are present. For example, two clusters of two Cortex-R52 cores where
each cluster has a dedicated GICv3 instance.
The changes provided here are a result of a merge from various examples
system_stm32h7xx.c files provided by STMicroelectronics for the STM32H7B3I-DK
board with the original RTEMS file provided for the STM32H743I-EVAL2 board.
Added tests for exit procedures to ensure proper resource
cleanup. The test now checks cleanup for files assigned
to stdio streams and non-stdio streams.
This code did not have any copyrights or licenses and a bit
of archeology was needed to determine authorship.
This code was in the initial import into the RTEMS CVS repository when
it was established in May 1995. There was very little, if any, code not
written by OAR Corporation in that initial import. After discussion
with Chris Johns, it was determined that this code was from OAR
Corporation and that he had added a few features later. Both
Chris Johns and OAR Corporation have given permission to relicense.
Updates #3053.
Add SMP-specifc SMP_FATAL_MULTITASKING_START_ON_NOT_ONLINE_PROCESSOR
fatal error. This fatal error helps to diagnose a broken SMP startup
sequence. Without this error a context switch using the NULL pointer
for the thread control block happens which may be difficult to debug.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
The test source code is generated from specification items
by the "./spec2modules.py" script contained in the
git://git.rtems.org/rtems-central.git Git repository.
Please read the "How-To" section in the "Software Requirements Engineering"
chapter of the RTEMS Software Engineering manual to get more information about
the process.
Update #3716.
Add support to record scheduler operations. This support is especially
important for tests in SMP configurations since the thread switch extension is
quite difficult to use due to the asynchronous nature of thread dispatching.
In contrast, the scheduler operations occur normally in a deterministic order.
Update #3716.
This function may be used to burn a couple of processor cycles with
minimum impact on the system bus. It may be used in busy wait loops.
Since it is a global function, it is possible to wrap it in device
driver test code.
The rtems_shell_init() passed the address of a stack variable
(exit_code) to rtems_shell_run(). If wait == false, then the stack
variable goes out of scope but may be accessed by the created shell
thread.
The rtems_shell_script() was affected by the same problem.
Close#4629.
Restrict the affinity set to the set of online processors. Make sure
the affinity set for an interrupt vector contains at least one online
processor.
Update #3269.
Add an architecture-specific implementation for
_CPU_Get_current_per_CPU_control() to reduce overhead for getting the
current CPU's Per_CPU_Control structure.
Quite some files in the bsps/m68k/genmcf548x mention a
Freescale_license.txt file. The file has been accidentally removed
during the source reorganization in 2018. This commit restores it and
moves it to the right location for licenses.
Update #4625.
There is not really anything left from the referenced cpuinit.c file
from MPC8260ads CPU initialization. So ignore that part of the header
during rework.
Update #4625.
The previous SMP multitasking start assumed that the initial heir thread of a
processor starts execution in _Thread_Handler(). The _Thread_Handler() sets
the interrupt state explicitly by _ISR_Set_level() before it calls the thread
entry. Under certain timing conditions, processors may perform an initial
context switch to a thread which already executes its thread body (see
smptests/smpstart01). In this case, interrupts are disabled after the context
switch on targets which do not save/restore the interrupt state during a
context switch (aarch64, arm, and riscv).
Close#4627.
Disable thread dispatching earlier on secondary processors. This ensures that
fatal error and per-CPU job handlers are called with thread dispatching
disabled. On the boot processor, the thread dispatching is already disabled by
_Thread_Dispatch_initialization().
This is required for ISA 2.0 support, see chapter
"Zicsr", Control and Status Register (CSR) Instructions, Version 2.0
in
RISC-V Instruction Set Manual, Volume I: RISC-V User-Level ISA
When committed, the MicroBlaze RAM size was hard-coded to 16MB. This
changes the default to 256MB and sets the KCU105 BSPs to 2GB since that
is what the board has on it.
Add MicroBlaze support for libdebugger. This uses only software break
type instructions to provide self-hosted GDB debugging support for
applications since internal control of debug hardware is not possible.
Also of note, this implementation for MicroBlaze would typically use the
brki instruction for software break, but instead uses an illegal opcode
to manage software breaks as exceptions. This is due to poor interaction
with the debug hardware where the debug hardware will intercept software
breaks instead of allowing the software break vector to execute.
This updates behavior of libdebugger to handle debug exceptions in
interrupt context by temporarily removing a software breakpoint,
stepping, and then resuming afterward.
Exception handling should be enabled at all times during execution to
ensure that exceptions are not ignored which would cause further
problems. This separates use of the exception enable bit from use of the
interrupt enable bit in the machine status register so that they can be
manipulated independently.
Add a capability that allows for implementations that operate purely
using software breaks. Due to this implementation method, software
breaks must not be restored until just before returning control to the
thread itself and will be handled by the implementation through thread
switch and interrupt hooks.
It is possible to remove software breaks without actually restoring the
original instruction to memory. When this happens, the original
instruction is lost. This ensures that the original instruction is
restored when a software break is removed.
The file was imported from this repository:
https://github.com/freebsd/freebsd.git
This commit was used:
commit 3ec0dc367bff27c345ad83240625b2057af391b9
Author: Sebastian Huber <sebastian.huber@embedded-brains.de>
Date: Mon Feb 7 14:16:16 2022 -0700
kern_ntptime.c: Remove ntp_init()
The ntp_init() function did set a couple of global objects to zero. These
objects are in the .bss section and already initialized to zero during kernel
or module loading.
Update #2348.
There are two places where we convert from a timecounter delta to
a bintime delta: tc_windup and bintime_off.
Both functions use the same calculations when the timecounter delta is
small. But for a large delta (greater than approximately an equivalent
of 1 second) the calculations were different. Both functions use
approximate calculations based on th_scale that avoid division. Both
produce values slightly greater than a true value, calculated with
division by tc_frequency, would be. tc_windup is slightly more
accurate, so its result is closer to the true value and, thus, smaller
than bintime_off result.
As a consequence there can be a jump back in time when time hands are
switched after a long period of time (a large delta). Just before the
switch the time would be calculated with a large delta from
th_offset_count in bintime_off. tc_windup does the switch using its own
calculations of a new th_offset using the large delta. As explained
earlier, the new th_offset may end up being less than the previously
produced binuptime. So, for a period of time new binuptime values may
be "back in time" comparing to values just before the switch.
Such a jump must never happen. All the code assumes that the uptime is
monotonically nondecreasing and some code works incorrectly when that
assumption is broken. For example, we have observed sleepq_timeout()
ignoring a timeout when the sbinuptime value obtained by the callout
code was greater than the expiration value, but the sbinuptime obtained
in sleepq_timeout() was less than it. In that case the target thread
would never get woken up.
The unified calculations should ensure the monotonic property of the
uptime.
The problem is quite rare as normally tc_windup should be called HZ
times per second (typically 1000 or 100). But it may happen in VMs on
very busy hypervisors where a VM's virtual CPU may not get an execution
time slot for a second or more.
Reviewed by: kib
MFC after: 2 weeks
Sponsored by: Panzura LLC
Hyper-V wants to register its MSR-based timecounter during
SI_SUB_HYPERVISOR, before SI_SUB_LOCK, since an emulated 8254 may not be
available for DELAY(). So we cannot use MTX_SYSINIT to initialize the
timecounter lock.
PR: 259878
Reviewed by: kib
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D33014
Do not initialize Thread_Control::libc_reent if
CONFIGURE_DISABLE_NEWLIB_REENTRANCY is defined. This helps to catch errors
with a NULL pointer exception rather than a corruption of the thread control
block.
A device tree binary has to be 8 byte aligned in memory. This is checked
since RTEMS commit 34052ef78c "libfdt: Add
FDT alignment check to fdt_check_header()".
Edit assert statements in worker thread to check initialization against the
__cleanup member of struct _reent instead of __sdidinit. This will allow the
removal of sdidinit in a follow up Newlib patch.
Put the next character into the send buffer if the buffer is empty and
not when the last character has been sent out to the line. This improves
the performance slightly.
Before that patch, the receive path was faster than the transmit path.
Therefore a simple echo could drop characters on a busy connection.
With this patch sending and receiving has about the same performance so
that no characters are lost.
Fixes#4610
The driver didn't return with an error on (for example) a NACK on the
bus. This adds the expected error return. Due to the new case that a
transfer can be interrupted on an error, there were some new edge cases.
This patch therefore also fixes these edge cases by removing the
transfer_state that more or less duplicated the interrupt states.
Fixes#4592
Do some clean ups. Remove superfluous variables. Eliminate some overly
complex logic (information about transfer and remaining bytes has been
tracked redundantly in multiple variables).
This patch doesn't change the behavior of the driver.
Update #4592
At the moment the line discipline start function (l_start) has no
possibility to get feedback about the number of characters that have
been sent. This patch passes that information via an additional
parameter.
The change might trigger a warning on existing code because of a pointer
mismatch but it shouldn't break it. An existing function with the old
API will just ignore the additional parameter.
Update #4493
The dispatch code was unnecessarily saving and restoring an extra
interrupt frame. This avoids the extra frame and folds the dispatch call
into a fallthrough to the interrupt frame restoration code.
This patch adds a vector for debug events along with a hook similar to
the exception framework. The debug vector generates an exception frame
for use by libdebugger.
This patch updates the CPU_Exception_frame to include all necessary
registers, combines hardware snd software exception handlers into a
shared vector, provides an architecture-specific hook for taking
control of exception handling, and moves exception handling over to
actually using the CPU_Exception_frame instead of a minimal interrupt
stack frame. As the significant contents of _exception_handler.S have
been entirely rewritten, the copyright information on this file has been
updated to reflect that.
Avoid a dependency on errno which might be a thread-local object. The tests
sp01, spstkalloc02, and sptls03 assume that no thread-local storage object is
present.
Update #4560.
When the stack checker is not enabled, the stack checker reporting
function can still be called. This prevents that call from performing a
null memory access in trying to find the high water mark if the stack
checker was never initialized.
This also introduces a test to ensure this call does not cause a crash.
Closes#4588
GCC versions prior to 6.1 used a RTEMS thread model based on
rtems_gxx_*() functions. GCC version 6.1 or later uses the
self-contained synchronization objects of Newlib <sys/lock.h> for the
RTEMS thread model.
Remove the obsolete implementation.
Close#3143.
If the system is busy with other interrupts and the UART is set to a
fast baud rate, it's possible to loose UART interrupts and therefore
characters. This allows to optionally enable a DMA for the UARTs so that
a number of lost interrupts can be tolerated.
The number of DMAs on this chip is limited and not not all applications
need that feature. Therefore the DMA is disabled by default.
Close#4578
If no extended features of the USART are used and if the comparison
feature of the UART is not used, the two modules are compatible. The
drivers were nearly identical except for some names of the defines.
This patch merges the two drivers into one.
Update #4578
Use volatile register r0 for the per-CPU control of the current
processor instead of the non-volatile register r7. This enables the use
of r7 in a follow up patch. Do the interrupt handling in ARM mode.
Update #4579.
Most BSPs which used the stubbed benachmark timer provide a CPU counter.
All BSPs provide at least a stub CPU counter. Simply use the benchmark
timer implementation using the CPU counter.
Always start the executive in Exception Level 1, Non-Secure mode.
If we boot in EL3 Secure with GICv3 then we have to initialize
the distributor and redistributor to set up G1NS interrupts
early in the boot sequence before stepping down from EL3S to EL1NS.
Now there is no need to distinguish between secure and non-secure
world execution after the primary core boots, so get rid of the
AARCH64_IS_NONSECURE configuration option.
This is done to get the target path for the overlay nodes which is very useful
in many cases. For example, Xen hypervisor needs it when applying overlays
because Xen needs to do further processing of the overlay nodes, e.g. mapping of
resources(IRQs and IOMMUs) to other VMs, creation of SMMU pagetables, etc.
Signed-off-by: Vikram Garhwal <fnu.vikram@xilinx.com>
Message-Id: <1637204036-382159-2-git-send-email-fnu.vikram@xilinx.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
UINT32_MAX is an integer of type unsigned int. UINT32_MAX + 1 overflows
unless explicitly computed as unsigned long long. This led to some
invalid addresses being treated as valid.
Cast UINT32_MAX to uint64_t explicitly.
Signed-off-by: Elvira Khabirova <e.khabirova@omp.ru>
With -Wsign-compare, compilers warn about a mismatching signedness in
the different legs of the conditional operator, in fdtget.c.
In the questionable expression, we are constructing a 16-bit value out of
two unsigned 8-bit values, however are relying on the compiler's
automatic expansion of the uint8_t to a larger type, to survive the left
shift. This larger type happens to be an "int", so this part of the
expression becomes signed.
Fix this by explicitly blowing up the uint8_t to a larger *unsigned* type,
before doing the left shift. And while we are at it, convert the hardly
readable conditional operator usage into a sane switch/case expression.
This fixes "make fdtget", when compiled with -Wsign-compare.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20210618172030.9684-3-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Coverity gets a bit confused by loading fdt_size_dt_strings() and
using it in a memmove(). In fact this is safe because the callers
have verified this information (via FDT_RW_PROBE() in fdt_pack() or
construction in fdt_open_into()).
Passing in strings_size like we already do struct_size seems to get
Coverity to follow what's going on here.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Apparently the unchecked return value of the first fdt_next_tag() call in
fdt_add_subnode_namelen() is tripping Coverity Scan in some circumstances,
although it appears not to for the scan on our project itself.
This fdt_next_tag() should always return FDT_BEGIN_NODE, since otherwise
the fdt_subnode_offset_namelen() above would have returned BADOFFSET or
BADSTRUCTURE.
Still, add a check to shut Coverity up, gated by a can_assume() to avoid
bloat in small builds.
Reported-by: Ryan Long <ryan.long@oarcorp.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Changes in v3:
- Remove noop version sets
- Set version correctly on loaded fdt in fdt_open_into
Fixes: f1879e1a50eb ("Add limited read-only support for older (V2 and V3) device tree to libfdt.")
Signed-off-by: Justin Covell <jujugoboom@gmail.com>
Message-Id: <20201229041749.2187-1-jujugoboom@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
There's a small inaccuracy in the comment describing these new helpers.
This corrects it, and reformats while we're there.
Fixes: f98f28ab ("libfdt: Internally perform potentially unaligned loads")
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Commits 6dcb8ba4 "libfdt: Add helpers for accessing unaligned words"
introduced changes to support unaligned reads for ARM platforms and
11738cf01f15 "libfdt: Don't use memcpy to handle unaligned reads on ARM"
improved the performance of these helpers.
On further discussion, while there are potential cases where we could be
used on platforms that do not fixup unaligned reads for us, making this
choice the default is very expensive in terms of binary size and access
time. To address this, introduce and use new fdt{32,64}_ld_ functions
that call fdt{32,64}_to_cpu() as was done prior to the above mentioned
commits. Leave the existing load functions as unaligned-safe and
include comments in both cases.
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Message-Id: <20201211022736.31657-1-trini@konsulko.com>
Tested-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
With -Wsign-compare, compilers warn about a mismatching signedness
in a comparison in fdt_strerror().
Force FDT_ERRTABSIZE to be signed (it's surely small enough to fit), so
that the types match. Also move the minus sign to errval, as this is
actually what we use in the next line.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20201001164630.4980-7-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
With -Wsign-compare, compilers warn about a mismatching signedness in
comparisons in fdt_get_string().
Introduce a new usigned variable, which holds the actual (negated)
stroffset value, so we avoid negating all the other variables and have
proper types everywhere.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20201001164630.4980-6-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
With -Wsign-compare, compilers warn about a mismatching signedness
in a comparison in fdt_setprop_inplace_namelen_partial().
fdt_getprop_namelen() will only return negative error values in "proplen"
if the return value is NULL. So we can rely on "proplen" being positive
in our case and can safely cast it to an unsigned type.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20201001164630.4980-5-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
With -Wsign-compare, compilers warn about a mismatching signedness
in a comparison in fdt_create_with_flags().
By making hdrsize a signed integer (we are sure it's a very small
number), we avoid all the casts and have matching types.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20201001164630.4980-4-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
With -Wsign-compare, compilers warn about a mismatching signedness
in comparisons in fdt_move().
This stems from "bufsize" being passed in as a signed integer, even
though we would expect a buffer size to be positive.
Short of changing the prototype, check that bufsize is not negative, and
cast it to an unsigned type in the comparison.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20201001164630.4980-3-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
With -Wsign-compare, compilers warn about a mismatching signedness
in a comparison in fdt_add_string_().
Make all variables unsigned, and express the negative offset trick via
subtractions in the code.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20201001164630.4980-2-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
With -Wsign-compare, compilers warn about a mismatching signedness
in a comparison in fdt_node_offset_by_phandle().
Uses a better suited bitwise NOT operator to denote the special value of
-1, which automatically results in an unsigned type.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20200921165303.9115-14-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
With -Wsign-compare, compilers warn about a mismatching signedness
in a comparison in fdt_resize().
A negative buffer size will surely do us no good, so let's rule this
case out first.
In the actual comparison we then know that a cast to an unsigned type is
safe.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20200921165303.9115-10-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
With -Wsign-compare, compilers warn about a mismatching signedness in
comparisons in fdt_get_string().
In the first two cases, we have just established that the signed values
are not negative, so it's safe to cast the values to an unsigned type.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20200921165303.9115-7-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
With -Wsign-compare, compilers warn about a mismatching signedness
in a comparison in fdt_grab_space_().
All the involved values cannot be negative, so let's switch the types of
the local variables to unsigned to make the compiler happy.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20200921165303.9115-4-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
With -Wsign-compare, compilers warn about mismatching signedness in
comparisons in fdt_offset_ptr().
This mostly stems from "offset" being passed in as a signed integer,
even though the function would not really tolerate negative values.
Short of changing the prototype, check that offset is not negative, and
use an unsigned type internally.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Message-Id: <20200921165303.9115-2-andre.przywara@arm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
fdt_check_node_offset_() checks for a valid offset but also changes the
offset by calling fdt_next_tag(). Hence, do not skip this function if
ASSUME_VALID_INPUT is set but only omit the initial offset check in that
case.
As this function works very similar to fdt_check_prop_offset_(), do the
offset check there as well depending on ASSUME_VALID_INPUT.
Message-Id: <1913141.TlUzK5foHS@noys4>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
libfdt: fix undefined behaviour in fdt_splice_()
Along the lines of commit d0b3ab0a0f46 ("libfdt: Fix undefined behaviour
in fdt_offset_ptr()"), fdt_splice_() similarly may not use pointer
arithmetic to do overflow checks. (The left side of the checks added by
d4c7c25c9ed1 ["libfdt: check for potential overrun in _fdt_splice()"]
doesn't really lend itself to similar replacement though.)
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Message-Id: <f2d09e81-7cb8-c5cc-9699-1ac05b0626ff@suse.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Using rtems_interrupt_entry_install() instead of
rtems_interrupt_handler_install() avoids a dependency on the dynamic memory
allocation.
Use Interrupt Manager directives instead of a BSP-specific API. Use inline
functions. In SMP configurations, set an affinity to all online processors and
raise the interrupt on the current processor.
The AArch64 cache implementation does not define
rtems_cache_disable_data(), but declares that it does via
CPU_CACHE_SUPPORT_PROVIDES_DISABLE_DATA. The existing implementation of
_CPU_cache_disable_data() is sufficient to enable this functionality
without the erroneous cache feature flag.
Closes#4569
When the cadence I2C code was moved to a shared directory, the
references were updated but the install locations weren't. This updates
the install locations to match what out-of-tree applications expect.
RTEMS untar implementation had problems with overwriting or integrating
archives into existing directory structures. This patch adapts the
behavior to mimic that of a GNU tar or BSD tar and extends the tar01
test to check for the behavior. That is:
* If a directory structure exists, the files from the archive will be
integrated. Existing files are overwritten.
* If a file exists and the archive contains a directory with the same
name, the file is removed and a directory is created. In the above
example: if l1/l2 is a file it will be overwritten with a new
directory.
* If a directory exists and the archive contains a file with the same
name, the directory will be replaced if it is empty. If it contains
files, the result is an error.
* An archive also can contain only a file without the parent
directories. If in that case one of the parent directories exists as a
file extracting the archive results in an error. In the example: if
l1/l2 is a file and the archive doesn't contain the directories but
only the file l1/l2/x.txt that would be an error.
* In case of an error, it is possible that the archive has been
partially extracted.
Closes#4568
In _Workspace_Initialize_for_one_area(), properly check if there is enough
free memory available for the configured workspace size.
The bug was introduced by commit 3d0620b607.
Since the previous Interrupt Manager Extension API moved to the
Interrupt Manager, we can move the interfaces to <rtems/rtems/intr.h> so
that they are available via #include <rtems.h>.
The BSPs provide memory for the separate C Program Heap initialization
via _Memory_Get(). Most BSPs provide exactly one memory area. Only two
BSPs provide more than one memory area (arm/altera-cyclone-v and
bsps/powerpc/mpc55xxevb). Only if more than one memory area is
provided, there is a need to use _Heap_Extend(). Provide two
implementations to initialize the separate C Program Heap and let the
BSP select one of the implementations based on the number of provided
memory areas. This gets rid of a dependency on _Heap_Extend(). It
also avoids dead code sections for most BSPs.
Change licence to BSD-2-Clause according to file history.
Update #3053.
The BSPs provide memory for the workspace initialization via
_Memory_Get(). Most BSPs provide exactly one memory area. Only two
BSPs provide more than one memory area (arm/altera-cyclone-v and
bsps/powerpc/mpc55xxevb). Only if more than one memory area is
provided, there is a need to use _Heap_Extend(). Provide two
implementations to initialize the workspace handler and let the BSP
select one of the implementations based on the number of provided memory
areas. This gets rid of a dependency on _Heap_Extend(). It also avoids
dead code sections for most BSPs.
Affected components are the GR712RC, UT699, UT699E, UT700, and
LEON3FT-RTAX. Strictly, the workaround is only necessary if the MMU is
enabled. Use __FIX_LEON3FT_B2BST to enable the workaround. This is not
100% appropriate, but the best thing we can use to enable the
workaround.
Close#4551.
The _Thread_queue_Extract() does not deal with potential priority
updates and the SMP locking protocol handling. Use
_Thread_queue_Continue(). For the POSIX signals processing this is
currently probably unnecessary, however, the use case is similar to the
restart so use the same appoach.
Close#4546.
Remove the THREAD_WAIT_STATE_READY_AGAIN and simply use the initial value to
indicate that a thread does not wait on something. Rename
THREAD_WAIT_FLAGS_INITIAL to THREAD_WAIT_STATE_READY. This change is necessary
so that _Thread_Continue() can be called for threads which never waited on
something (for example dormant threads).
Update #4546.
Remove _Thread_queue_Extract_with_proxy() and move the proxy extraction
to _Thread_MP_Extract_proxy(). Move similar code blocks of the previous
caller of _Thread_queue_Extract_with_proxy() to helper functions.
Update #4546.
The SMP EDF scheduler supports a one-to-one and one-to-all thread to
processor affinity. It accepted affinity sets which are a proper
subset of the online processor containing at least two processors owned by
the scheduler. In this case it used a one-to-one thread to processor
affinity. This leads to undefined behaviour if a processor is removed
since the higher level check in rtems_scheduler_remove_processor() does
not account for this implementation detail.
Restrict the affinity set accepted by the SMP EDF scheduler to
1. all online processors, or
2. exactly one processor owned by the scheduler.
Close#4545.
Return an error status for the following error condition in
rtems_scheduler_remove_processor():
While an attempt is made to remove a processor from a scheduler, while
the processor is the only processor owned by the scheduler, if a thread
exists which uses the scheduler as a helping scheduler, then the
processor shall not be removed.
The reason is that ask for help requests and withdraw node requests are
processed asynchronously in any order. An ask for help request carried
out on a scheduler without a processor is undefined behaviour.
Update error status description.
Update #4544.
Process ask for help requests on the current processor. This avoids
using inter-processor interrupts to make the system behaviour a bit more
predictable.
Update #4531.
Simplify callers of _Scheduler_SMP_Schedule_highest_ready(). Move the node
state change and the extraction from scheduled into
_Scheduler_SMP_Schedule_highest_ready(). Move the idle thread release to the
caller which have more information about the presence of an idle thread.
Update #4531.
The SMP EDF scheduler supports one-to-one and one-to-all thread to
processor affinities. The one-to-one thread to processor affinity
introduces a constraint on the ordering of threads. The implementation
uses one ready queue for threads which have a one-to-all affinity and
one for each one-to-one affinity group. To order threads across the
ready queues, a generation number is used. However, the approach to
update the generation number each time a thread is inserted into a ready
queue was wrong. The generation number needs to be updated only in the
enqueue and enqueue scheduled operations where an insert priority is
available. The scheduled chain needs to take the generation number into
account.
An example scenario which shows the bug is this. Let T be a high
priority task affine to processor X. Let A be a lower priority task
affine to processor X. Let B be a lower priority task with no affinity
to a particular processor which executes on processor Y. Let B be in
the same priority group than A and after A. Let T set the affinity to
all processors. Now A (higher priority relative to B) should execute on
X and T (high priority) should execute on Y.
Close#4534.
This patch fixes an issue with the idle thread handling in the SMP
scheduler framework used for the MrsP locking protocol. The approach to
use a simple chain of unused idle threads is broken for schedulers which
support thread to processor affinity. The reason is that the thread to
processor affinity introduces another ordering indicator which may under
certain conditions lead to a reordering of idle threads in the scheduled
chain. This reordering is not propagated to the chain of unused idle
threads. This could lead to use an idle thread for a sticky scheduler
node which is already in use. This locks up the system in infinite
loops in the thread context switch procedure.
To fix this, the SMP scheduler implementations must now provide
callbacks to get and release an unused idle thread.
Update #4531.
This patch fixes the following broken behaviour:
While a thread is scheduled on a helping scheduler, while it does not
own a MrsP semaphore, if it obtains a MrsP semaphore, then no
scheduler node using an idle thread and the ceiling priority of the
semaphore is unblocked for the home scheduler.
This could lead to priority inversion issues and is not in line
with the MrsP protocol.
Introduce two new scheduler operations which are only enabled if
RTEMS_SMP is defined. The operations are used to make the scheduler
node of the home scheduler sticky and to clean the sticky property.
This helps to keep the sticky handing out of the frequently used
priority update operation.
Close#4532.
If a node is moved from the scheduled chain to the ready queue, then we
know that it is the highest priority ready node. So, it can be
prepended to the ready queue without doing any comparisons.
Update #4531.
Rework the handling of the affine ready queue for the EDF SMP scheduler.
Do the queue handling in the node insert, move, and extract operations.
Remove the queue handling from _Scheduler_EDF_SMP_Allocate_processor().
Update #4531.
Insert nodes after moving the second node to reduce the items in the
data structure for the insert operation. This also avoids having two
nodes for the same processor inserted into the scheduled chain.
Update #4531.
Do not set the CPU of the idle thread in _Scheduler_Use_idle_thread(). This
helps to use _Scheduler_Try_to_schedule_node() under more general conditions in
the future, for example in case the owner and user of a node are not the same.
Update #4531.
If we block the executing thread and it is not the heir thread, then there is
no need to run the schedule operation. The scheduler already selected a new
heir.
These two OAR copyright headers are the only two in the codebase with
a format that differs from the typical OAR copyright header. This makes
all of the OAR copyright headers consistent.
Ensure when both RTEMS_DEBUG is specified and pointers are large that
enough space is allocated to accomodate the Per_CPU_Control structure.
This changes the calculation to be more compositional instead of trying
to list out every permutation of options possible.
These external functions rtems_scheduler_get_processor() and
rtems_scheduler_get_processor_maximum() which may be used by bindings
for languages other than C/C++.
Check all IO driver table configuration options which are used to initialize
_IO_Driver_address_table[]. Checks for the following settings were missing:
* CONFIGURE_BSP_PREREQUISITE_DRIVERS
* CONFIGURE_APPLICATION_PREREQUISITE_DRIVERS
* CONFIGURE_APPLICATION_NEEDS_WATCHDOG_DRIVER
* CONFIGURE_APPLICATION_EXTRA_DRIVERS
Update #3875.
Evaluate CONFIGURE_DISABLE_BSP_SETTINGS for each affected application
configuration option. This makes the code easier to review since the influence
of CONFIGURE_DISABLE_BSP_SETTINGS is locally visible in the code.
Update #3875.
Priority ceiling semaphores (in contrast to MrsP semaphores) have only
one priority ceiling which is defined for the home scheduler of the task
which created the semaphore. Always check during a semaphore seize that
the home scheduler of the calling task is equal to this scheduler.
This avoids a scheduler mismatch if the semphore is acquired after the
calling thread blocked on the semaphore.
This change is a slight performance optimization for systems with a slow
64-bit division.
The th->th_scale and th->th_large_delta values only depend on the
timecounter frequency and the th->th_adjustment. The timecounter
frequency of a timehand only changes when a new timecounter is activated
for the timehand. The th->th_adjustment is only changed by the NTP
second update. The NTP second update is not done for every call of
tc_windup().
Move the code block to recalculate the scaling factor and
the large delta of a timehand to the new helper function
recalculate_scaling_factor_and_large_delta().
Call recalculate_scaling_factor_and_large_delta() when a new timecounter
is activated and a NTP second update occurred.
MFC after: 1 week
Timecounter registration is dynamic, i.e., there is no requirement that
timecounters must be registered during single-threaded boot. Loadable
drivers may in principle register timecounters (which can be switched to
automatically). Timecounters cannot be unregistered, though this could
be implemented.
Registered timecounters belong to a global linked list. Add a mutex to
synchronize insertions and the traversals done by (mpsafe) sysctl
handlers. No functional change intended.
Reviewed by: imp, kib
MFC after: 2 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D32511
FreeBSD Foundation sys/ copyrights
These ones were unambiguous cases where the Foundation was the only
listed copyright holder (in the associated license block).
Sponsored by: The FreeBSD Foundation
In FreeBSD, the current time is computed from uptime + boottime. Uptime
is a continuous, smooth function that's monotonically increasing. To
effect changes to the current time, boottime is adjusted. boottime is
mutable and shouldn't be cached against future need. Document the
current implementation, with the caveat that we may stop stepping
boottime on resume in the future and will step uptime instead (noted in
the commit message, but not in the code).
Sponsored by: Netflix
Reviewed by: phk, rpokala
Differential Revision: https://reviews.freebsd.org/D30116
on freebsd/arm64:
- Implement a dtrace_getnanouptime(), matching the existing
dtrace_getnanotime(), to avoid DTrace calling out to a potentially
instrumentable function.
(These should probably both be under KDTRACE_HOOKS. Also, it's not clear
to me that they are correct implementations for the DTrace thread time
functions they are used in .. fixes for another commit.)
- Don't allow FBT to instrument functions involved in EL1 exception handling
that are involved in FBT trap processing: handle_el1h_sync() and
do_el1h_sync().
- Don't allow FBT to instrument DDB and KDB functions, as that makes it
rather harder to debug FBT problems.
Prior to these changes, use of FBT on FreeBSD/arm64 rapidly led to kernel
panics due to recursion in DTrace.
Reliable FBT on FreeBSD/arm64 is reliant on another change from @andrew to
have the aarch64 instrumentor more carefully check that instructions it
replaces are against the stack pointer, which can otherwise lead to memory
corruption. That change remains under review.
MFC after: 2 weeks
Reviewed by: andrew, kp, markj (earlier version), jrtc27 (earlier version)
Differential revision: https://reviews.freebsd.org/D27766
to warm timecounters.
It seems that second call does not add any useful state change for all
implemented timecounters.
Discussed with: bde
Sponsored by: The FreeBSD Foundation
MFC after: 3 weeks
or CTLFLAG_NEEDGIANT (17 of many)
r357614 added CTLFLAG_NEEDGIANT to make it easier to find nodes that are
still not MPSAFE (or already are but aren’t properly marked).
Use it in preparation for a general review of all nodes.
This is non-functional change that adds annotations to SYSCTL_NODE and
SYSCTL_PROC nodes using one of the soon-to-be-required flags.
Mark all obvious cases as MPSAFE. All entries that haven't been marked
as MPSAFE before are by default marked as NEEDGIANT
Approved by: kib (mentor, blanket)
Commented by: kib, gallatin, melifaro
Differential Revision: https://reviews.freebsd.org/D23718
and fix possible overflow in bintime()/binuptime().
The algorithm to read the consistent snapshot of current timehand is
repeated in each accessor, including the details proper rollup
detection and synchronization with the writer. In fact there are only
two different kind of readers: one for bintime()/binuptime() which has
to do the in-place calculation, and another kind which fetches some
member from struct timehand.
Extract the logic into type-checked macros, GETTHBINTIME() for bintime
calculation, and GETTHMEMBER() for safe read of a structure' member.
This way, the synchronization is only written in bintime_off() and
getthmember().
In bintime_off(), use overflow-safe calculation of th_scale *
delta(timecounter). In tc_windup, pre-calculate the min delta value
which overflows and require slow algorithm, into the new timehands
th_large_delta member.
This part with overflow fix was written by Bruce Evans.
Reported by: Mark Millard <marklmi@yahoo.com> (the overflow issue)
Tested by: pho
Discussed with: emaste
Sponsored by: The FreeBSD Foundation (kib)
MFC after: 3 weeks
On arm64 (and possible other architectures) we are unable to use static
DPCPU data in kernel modules. This is because the compiler will generate
PC-relative accesses, however the runtime-linker expects to be able to
relocate these.
In preparation to fix this create two macros depending on if the data is
global or static.
Reviewed by: bz, emaste, markj
Sponsored by: ABT Systems Ltd
Differential Revision: https://reviews.freebsd.org/D16140
to opt_global.h.
opt_compat.h is mentioned in nearly 180 files. In-progress network
driver compabibility improvements may add over 100 more so this is
closer to "just about everywhere" than "only some files" per the
guidance in sys/conf/options.
Keep COMPAT_LINUX32 in opt_compat.h as it is confined to a subset of
sys/compat/linux/*.c. A fake _COMPAT_LINUX option ensure opt_compat.h
is created on all architectures.
Move COMPAT_LINUXKPI to opt_dontuse.h as it is only used to control the
set of compiled files.
Reviewed by: kib, cem, jhb, jtl
Sponsored by: DARPA, AFRL
Differential Revision: https://reviews.freebsd.org/D14941
In this case volatile qualifiers enusre that a compiler does not
optimize the accesses out.
Reviewed by: alc, jhb
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D13534
This patch set replaces the CPU budget algorithm enumeration with a set of CPU
budget operations which implement a particular CPU budget algorithm. This
helps to hide the CPU budget algorithm implementation details from the general
thread handling. The CPU budget callouts are turned into CPU budget
operations. This slightly reduces the size of the thread control block.
All schedulers used the default scheduler tick implementation. The tick
scheduler operation is removed and the CPU budget operations are directly used
in _Watchdog_Tick() if the executing thread uses a CPU budget algorithm. This
is performance improvement for all threads which do not use a CPU budget
algorithm (default behaviour).
This fix came from Code Inspector, flagged by one of the analysis that
was done.
This test code is compiled only but better to address the issues so
future static analysis reports will not include it.
Debug events should be masked at least until after the first context
switch and should usually be masked until a debugger is attached for
application debugging.
The existing code is functional but inccorrect and blindly modifies the
other masking bits. It is important to preserve those other bits since
they control masking of important system events.
Certain input parameters for MMU mapping operations could cause an
infinite recursion if block end boundaries didn't align to 4k. This
ensures that recursion descent does not exceed 2 levels and instead
rounds up to the nearest 4k block if necessary.
This moves the AArch64 MMU memory type definitions into cpukit for use
by libdebugger since remapping of memory is required to insert software
breakpoints.
The N used in the breakpoint and watchpoint register names is intended
to be an integer between 0 and 15 (inclusive) and will not compile when
used as is. This adds the accessors necessary to access all of these
breakpoint and watchpoint registers.
This adds a confdef option allowing an application to request mapping
machine exceptions to POSIX signals. This is required for some languages
such as Ada.
This adds the set of functions necessary to allow more generic handling
of machine exceptions. This initial patch offers the ability to
manipulate a CPU_Exception_frame and resume execution using that
exception information with or without thread dispatch. These functions
are gated behind the RTEMS_EXCEPTION_EXTENSIONS configuration option.
Using 32bit types like uint32_t for pointers creates issues on 64 bit
architectures like AArch64. Replaced occurrences of these with
uintptr_t, which will work for both 32 and 64 bit architectures. Added
hex_decode_addr function to rtems-debugger.
Mention that resetting the processor usage time of tasks has no impact
on the period status and statistics.
Remove no longer relevant RTEMS_NOT_DEFINED error status.
Update #4528.
The rate monotonic period statistics were affected by
rtems_cpu_usage_reset(). The logic to detect and work around a CPU
usage reset was broken.
The Thread_Contol::cpu_time_used is changed to contain the processor
time used throughout the entire lifetime of the thread. The new member
Thread_Contol::cpu_time_used_at_last_reset is added to contain the
processor time used at the time of the last reset through
rtems_cpu_usage_reset(). This decouples the resets of the CPU usage and
the rate monotonic period statistics.
Update #4528.
The spconfig01 and spmisc01 tests were disabled for all AArch64 BSPs due
to a toolchain issue preventing them from compiling correctly. The
binutils version that contains the fix has been released and integrated
into RSB such that these two tests now build and operate correctly.
Add a stack allocator hook specifically for allocation of IDLE thread stacks.
This allows the user to decide if IDLE thread stacks are statically allocated
or handled by the same custom allocator mechanism as other thread stacks.
Closes#4524.
Do not use a direct thread dispatch in
_Thread_queue_Surrender_priority_ceiling() since it may be used in condition
variables using POSIX mutexes.
Close#4526.
Set the postponed jobs count to zero in rtems_rate_monotonic_cancel() so that
rtems_rate_monotonic_get_status() returns a consistent status for inactive
periods.
Update #4511.
The schedule operation is only called by rtems_task_mode(). It is
called if preempt mode of the executing thread changes from disabled to
enabled. Since the EDF SMP scheduler does not support the disabled
preemption mode, the schedule operation is never called.
The utimbuf structs in fsimfsconfig01 and fsimfsconfig02 were
being passed to utime uninitialized. This did not cause problems until
utime was changed to use utimensat behind the scenes. Now that
utimensat is called, the values of the utimbuf struct are checked, and
EINVAL is set for invalid values. The utimebuf structs in these tests
could contain invalid values since they are uninitialized.
By zero-initializing the utimbuf structs, we ensure that they pass the
checks in utimensat and that errno is set to the expected ENOTSUP.
Context validation for AArch64 was ported from the ARM implementation
without a reinterpretation of the actual requirements. The spcontext01
test just happened to pass because the set of scratch registers in ARM
is a subset of the scratch registers in AArch64.
Replace the boolen return value with the new enum
Thread_queue_Deadlock_status. This improves the code readability.
Improve documentation. Shorten function names.
For uniprocessor configurations, this patch removes dead code in the
_Thread_queue_Surrender() and _Thread_queue_Surrender_priority_ceiling()
functions.
Dead code is removed from _Thread_queue_Surrender_sticky().
This function was unused. It was a relict of the thread queue rework done
during the SMP support development. In an early stage, the extract operation
was called with a NULL thread queue. However, this is no longer the case. The
extract operation is only called if we have a non-NULL thread queue.
ARM's GICv2 is configurable and its attributes vary between
implementations including omission of specific interrupts. This allows
BSPs to accomodate those varying implementations with customized
attribute sets.
The order in which step checks for 1 and 2 are not rigidly defined and
may actually occur in either order depending on how the threads execute.
This waits for the job to complete to enforce the existing ordering.
This adds the SMP function that supports spinup of additional CPU cores
using the ARM standard PSCI inteface. This interface is provided by QEMU
as well as ARM Trusted Firmware running in monitor mode (EL3) on ARMv7 and
AArch64 CPUs. This supports activation va SMC or HVC instructions
depending on BSP configuration.
The T_now_tick() is a fall back time measurement using the CPU counter
in case no Clock Driver is configured. Some CPU counter may overflow
during the test execution. Accumulate the elapsed time to reduce the
chance of CPU counter overflows.
If SMP support is enabled and the system has exactly one processor, then
it may use an uniprocessor scheduler. The ask for help, reconsider help
request, and withdraw node operations can be NULL in this case, since
they are only used if a thread has at least one helping scheduler node.
At least two schedulers are required to get a helping node and each
scheduler involved must own at least one processor. This is not
possible on a system with exactly one processor.
Issue the new fatal SMP error
SMP_FATAL_SCHEDULER_REQUIRES_EXACTLY_ONE_PROCESSOR if the system starts
with not exactly one processor and an uniprocessor scheduler is
configured.
The uniprocessor schedulers do not support systems with more than more
processors. So they rivially support thread pinning and thus the
SMP_FATAL_SCHEDULER_PIN_OR_UNPIN_NOT_SUPPORTED cannot happen.
Add a second default implementation for SMP schedulers which do not
support thread pinning.
Change license to BSD-2-Clause according to file history and re-licensing
agreement.
Update #3053.
psxdevctl is supposed to return the value in errno. Before, it was
returning -1 and setting errno. Changed the tests to reflect these
changes. Added code from RRADE's posix_devctl.c.
Closes#4506
Remove the duplicated description which is an out dated copy and paste
from the Classic API Guide. Instead, thoroughly document the individual
extensions.
Change license to BSD-2-Clause according to file histories and
documentation re-licensing agreement.
Update #3899.
Update #3993.
Certain files related to the Zynq BSP's I2C driver are useable by the ZynqMP BSP as well.
Moved these files to shared directory in anticipation of I2C support for ZynqMP.
The rtems_rate_monotonic_get_status() directive returns an arbitrary
number for the count of postponed jobs if it is called for a newly
created period object. Set the count of postponed jobs to zero during
object creation.
Close#4511.
Split up rbtreenext.c since only _RBTree_Minimum() is used by the operating
system core services (thread queues and the EDF scheduler).
Change license to BSD-2-Clause according to file history and re-licensing
agreement.
Update #3053.
This simplifies the implementation a bit. Declare _TOD_Days_to_date[] in
<rtems/score/todimpl.h>. Make _TOD_Days_per_month[] and
_TOD_Days_since_last_leap_year[] static.
Update #4338.
Move the TOD validation to the callers of _TOD_Set(). This avoids dead code in
case only rtems_clock_set() is used in an application because rtems_clock_set()
always calls _TOD_Set() with a valid time of day.
The _Thread_Dispatch() function was customized over time and now the
work is done by _Thread_Do_dispatch() and specialized wrappers. The
plain _Thread_Dispatch() is now only used in some CPU ports. Move it to
a separate file to avoid dead code in the general.
Change license to BSD-2-Clause according to file history and
re-licensing agreement.
Update #3053.
Rename LEON3_FATAL_INVALID_CACHE_CONFIG_MAIN_PROCESSOR in
LEON3_FATAL_INVALID_CACHE_CONFIG_BOOT_PROCESSOR since the term
"boot processor" is used elsewhere in the code base.
Move the transition map members of the test context to a dedicated
structure. Move the transition variant pre-condition prepare, action,
and post-condition checks to a separate function to reduce the
indentation level and allow skipping of transition variants.
It wasn't possible to keep the CS line low between multiple message
descriptors in one transfer. This patch reworks the driver so that it is
possible.
Update #4180
In _Thread_queue_Flush_critical(), update the priority of the thread
queue owner only if necessary. The scheduler update priority operation
could be expensive.
In order to ensure FIFO fairness across schedulers, the thread queue
surrender operation must be used to dequeue a thread from the thread
queue. The thread queue extract operation is intended for timeouts.
Add _Thread_queue_Resume() which may be used to make extracted or
surrendered threads ready again.
Remove the now unused _Thread_queue_Extract_critical() function.
Close#4509.
The priority queues in clustered scheduling configurations use a per
scheduler priority queue rotation to ensure FIFO fairness across
schedulers. This mechanism is implemented in the thread queue surrender
operation. Unfortunately some semaphore and message queue directives
used wrongly the thread queue extract operation. Fix this through the
use of _Thread_queue_Surrender().
Update #4358.
The behaviour of the futex operations is defined by Linux:
https://man7.org/linux/man-pages/man2/futex.2.html
Use EAGIN instead of EWOULDBLOCK to be in line with the Linux man page.
These error numbers have the same value in Newlib. Using the same error
numbers helps to avoid confusion.
When you look at the history of the Linux man page you see that they
replaced EWOULDBLOCK with EAGAIN over time. At the time of the RTEMS
futex implementation they used EWOULDBLOCK.
- Support all possible descriptors in a select call. Borrowed
from Christain and his mDNS change in LibBSD
- If select (or poll) fails pause for a bit rather than
locking up in a hard loop
The Pi firmware added a wfe(wait for event), the cores 1-3 wait
for the start address being written to the mailbox register, followed
by a SEV poke to the mailbox that acts as a wfe wake-up event.
The NULL pointer check for the executing thread was introduced by
commit:
commit be3c257286
Author: Sebastian Huber <sebastian.huber@embedded-brains.de>
Date: Thu Jun 5 11:17:26 2014 +0200
score: Avoid NULL pointer access
Check that the executing thread is not NULL in _Scheduler_Tick(). It
may be NULL in case the processor has an optional scheduler assigned and
the system was not able to start the processor.
However, it is no longer necessary since now the clock interrupt is
distributed to the online processors.
In SMP configurations, on 64-bit architectures use plain atomic
operations to set/get the priority value of a scheduler node. On 32-bit
architectures use an ISR lock. Using a sequence lock has no real
benefit since it uses atomic read-modify-write operations for both the
read and the write lock. Simply use a ticket lock instead so that only
one SMP synchronization primitive is used for everything.
Use the new Priority_Group_order enum instead of a boolean to indicated if a
priority should be inserted as the first or last node into its priority group.
This makes the code more expressive. It is also a bit more efficient since a
branch in _Scheduler_Node_set_priority() is avoided and a simple bitwise or
operation can be used.
The rtems_partition_return_buffer() wrongly accepted which were exactly
at the buffer area end. Use the buffer area limit address for the range
checking.
Close#4490.
the timer_create() method can use CLOCK_MONOTONIC
but there was no test for this.
Also it implements the functionality to
create a CLOCK_MONOTONIC timer and gettime() .
Closes#3888
Cloning under Cygwin turned off executable permission on these
files. This shows them as modified even though they have not
explicitly been touched. Executable permission should not have
been on for these files so this is just a minor clean up.
Change license to BSD-2-Clause according to file histories and
documentation re-licensing agreement.
Place the group into the I/O Manager group. Add all source files to the
group.
Update #3899.
Update #3993.
Update #4482.
Move _CPU_Fatal_halt() declaration to <rtems/score/cpuimpl.h> and make sure it
is a proper declaration of a function which does not return. Fix the type of
the error code. If necessary, add the implementation to cpu.c. Implementing
_CPU_Fatal_halt() as a function makes it possible to wrap this function for
example to fully test _Terminate().
The per-CPU states which control the SMP system initialization were added quite
early during the SMP support development. Replace this initial implementation
with a simplified one. There is no longer a global SMP lock required which
serialized the state changes of all processors. The new implementation better
integrates with the per-CPU jobs.
Remove _CPU_SMP_Processor_event_broadcast() and
_CPU_SMP_Processor_event_receive(). These functions are hard to use since they
are subject to the lost wake up problem.
Use the same function to output the '\r\n' combination produced by
rtems_putc(). Fix the format.
Change licence according to file history.
Update #3053.
Users have access to the fatal error source and code though the fatal error
extension. The user-specific fatal error handling should be done in statically
initialized fatal error handlers. The _Internal_errors_What_happened was
updated after the fatal error extension. In addition, there was no API to get
the information stored in _Internal_errors_What_happened. In SMP
configurations, this object could contain a mix of different fatal errors.
Remove this object to save some bytes of storage.
Do not return a status code in bsp_interrupt_facility_initialize() since this
leads to unreachable code in bsp_interrupt_initialize(). Use RTEMS_DEBUG
assertions in bsp_interrupt_facility_initialize() if necessary.
Add default implementations for bsp_interrupt_get_affinity() and
bsp_interrupt_set_affinity() which are required to link all tests in SMP
configurations.
Update #3269.
Add rtems_interrupt_entry_remove(). Split up irq-generic.c into several files.
In particular, place all functions which use dynamic memory into their own
file.
Add optional macros to let the BSP customize the vector installation after
installing the first entry and the vector removal before removing the last
entry:
* bsp_interrupt_vector_install()
* bsp_interrupt_vector_remove()
Use these new customization options in the m68k/genmcf548x BSP so re-use the
generic interrupt controller support.
Update #3269.
Return RTEMS_INCORRECT_STATE instead of RTEMS_INTERNAL_ERROR in case the
interrupt support is not initialized. This is similar to
rtems_timer_server_fire_after() for example.
Update #3269.
Add RTEMS_INTERRUPT_ENTRY_INITIALIZER(),
rtems_interrupt_entry_initialize(), and
rtems_interrupt_entry_remove(). This allows to install interrupt
handlers using user-provided storage as an alternative to
rtems_interrupt_handler_install() which has to allocate memory.
Update #3269.
Add rtems_interrupt_raise_on() and rtems_interrupt_clear().
Add a default implementation which just returns RTEMS_UNSATISFIED for
valid parameters.
Update #3269.
Add RTEMS_FATAL_SOURCE_SPURIOUS_INTERRUPT as the fatal source for
spurious interrupts. Use the interrupt vector number of the spurious
interrupt for the fatal code.
Update #3269.
Use <rtems/score/chain.h> which just provides the data types and avoid a
dependency on <rtems/chain.h> which contains the full chain
implementation.
Change license to BSD-2-Clause according to file histories and
documentation re-licensing agreement.
Update #3269.
Update #3899.
Update #3993.
Add rtems_interrupt_raise_on(). Document the currently not implemented
rtems_interrupt_clear(). Remove the not implemented and badly named
rtems_interrupt_cause() directive.
Update #3269.
When compiling the lwIP port for the TMS570, there
were issues with the BSP. Headers are expected in a folder
named ti_herc which did not exist. This fixes the issue.
Furthermore, there were multiple warnings about define redefinitions.
This was fixed as well.
These patches were submitted a few months ago, but it was found out
that the default-by-family: [] were missing in the GPIO .yml lines.
This was fixed in this patch.
This patch accounts for different pins for the ETH peripheral
on STM32H7 devices. For example, the Nucleo H743ZI has slightly
different pins than other STM32H7 boards.
Prefer RTEMS_FATAL_SOURCE_EXCEPTION over
INTERNAL_ERROR_ILLEGAL_USE_OF_FLOATING_POINT_UNIT since the fatal code
(rtems_exception_frame) provides more context.
Fully support the interrupt extension API to set/get the interrupt affinity.
Remove LEON3_irq_to_cpu which defined the interrupt to processor mapping in a
BSP-specific way.
Update #3269.
Skip the data cache initialization if we are a secondary processor.
The bug was introduced by e164df5e33 and
did not show up in tests using Qemu since the data cache behaviour is
not emulated.
Calling the memory FLASH and EXTRAM instead of FLEXSPI and SDRAM makes
it simpler to support other types of external RAM. This patch also
removes some of the calculations and improves names and documentation to
avoid pitfalls. It removes a unnecessary memory definition.
Update #4180
This reverts commit 10041a4cfc.
This type of configuration does not belong in RTEMS and is better
constrained to libbsd where the defines are actually being used.
The fix to address CID 1399742 (NO_EFFECT) in commit
f8b6359415 introduced a bug since
LEON3_IrqCtrl_EIrq == -1 in case no extended interrupts are supported by
the interrupt controller. Fix this by checking for
LEON3_IrqCtrl_EIrq > 0.
In addition, interrupt number 0 is reserved and should not be used.
The GICv3 support is shared between AArch32 and AArch64. For AArch32,
the new AARCH64_IS_NONSECURE is never defined. Use ARM_MULTILIB_ARCH_V4
instead.
This issue was introduced by 76c6caad52.
There is still a change in bsp_interrupt_vector_enable() for AArch32
compared to the version before 76c6caad52.
The existing fix for the ZynqMP UART hardware bug only caught the vast
majority of instances where it could occur. To fully fix the data
corruption, this fix must be applied after every baud rate change. This
makes the logic reset and kick apply in any locations where the baud
rate could be changed.
The Cortex-R52 does not support cache coherency and the shareable memory
attribute. If a region is configured to be shareable, then it falls
back to use non-cacheable memory.
Update #4202.
Provide the options necessary to enable any combination of CGEM ethernet
interfaces in LibBSD. The default is still CGEM3, so this should
continue to operate as expected on typical Zynq Ultrascale+ MPSoC
development hardware.
This change allows for the migration of higher priority tasks on the
arrival of a lower priority task limited by affinity constraints.
Change license to BSD-2-Clause according to file history and
re-licensing agreement.
Update #3053.
Move _ISR_Handler() to a separate file since it is now only used if a handler
is installed by _CPU_ISR_install_raw_handler().
Statically initialize the traps for external interrupts to use the new
_SPARC_Interrupt_trap() which directly dispatches the interrupt handlers
installed by rtems_interrupt_handler_install() via the BSP-provided
_SPARC_Interrupt_dispatch().
Since the trap table is now fully statically initialized, there is no longer a
dependency on the Cache Manager in the default configuration.
Update #4458.
Add bsp_interrupt_handler_dispatch_unchecked() as an alternative to
bsp_interrupt_handler_dispatch(). It may be used if the caller can ensure that
the vector number is valid.
Statically initialize the trap table in start.S to jump to _SPARC_Bad_trap()
for all unexpected traps. This enables a proper RTEMS fatal error handling
right from the start. Do not rely on the stack and register settings which
caused an unexpected trap. Use the ISR stack of the processor to do the fatal
error handling. Save the full context which caused the trap. Fatal error
handler may use it for error logging.
Unify the _CPU_Exception_frame_print() implementations and move it to cpukit.
Update #4459.
Move _CPU_ISR_install_raw_handler() and _CPU_ISR_install_vector() to separate
files. The goal is to make their use optional.
Update #4458.
Update #4459.
This makes it easier to review start.o and set break points to trap table
entries. This change was checked by inspecting the trap table in start.o with
objdump.
Update #4458.
The _ISR_Handler_initialization() does not touch the _ISR_Vector_table[]. Move
the definition of _ISR_Vector_table[] to a separate file.
Change license to BSD-2-Clause according to file history and re-licensing
agreement.
Update #3053.
This CPU port macro was not used. Since the _ISR_Vector_table[] is statically
allocated, CPU ports could initialize this table in _CPU_Initialize() if
necessary. Remove _CPU_Initialize_vectors() to simplify the CPU port
interface.
Change the default value of BSP_INTERRUPT_VECTOR_COUNT so that no interrupt
vectors are supported and all related directives return RTEMS_INVALID_ID.
Update #3269.
Assert BSP_INTERRUPT_VECTOR_MAX + 1 == BSP_INTERRUPT_VECTOR_COUNT.
After building all BSPs with this patch, BSP_INTERRUPT_VECTOR_MAX can be
removed and replaced by BSP_INTERRUPT_VECTOR_COUNT. The
BSP_INTERRUPT_VECTOR_COUNT allows a default implementation which supports no
interrupt vector at all. Using COUNT instead of MAX may avoid some
interpretation issues, for example is the maximum value a valid vector number
or not.
Update #3269.
Remove BSP_INTERRUPT_VECTOR_MIN and unconditionally let interrupt vector
numbers start with zero.
The BSP_INTERRUPT_VECTOR_MIN == 0 invariant was tested by the previous commit
and building all BSPs.
Update #3269.
This BSP uses a customized implementation of the interrupt extension API. It
was the only BSP which defined BSP_INTERRUPT_VECTOR_MIN to a value other than
zero. Define it to zero and use a custom bsp_interrupt_is_valid_vector()
function instead.
Update #3269.
This adds the possibility to open an I2C bus with O_NONBLOCK (or set it
later via fcntl) to get non-blocking transmissions. This means that if
the bus is busy, a read, write or transfer ioctl will return with a
EAGAIN errno.
The write to RBAR didn't have the valid flag set. Therefore the write to
RASR had an influence on the previously set region. That means for
example that if Region 0 had been enabled but 1 should be disabled due
to a size of 0, the previous code would have disabled region 0 instead.
This patch fixes that behaviour.
Close#4450
Get rid of the hack to output into the idle stack during the early system
initialization. This fixes also a couple of test program failures which fail
due to missing output.
Make the initialization of the per-CPU data optional.
Change license to BSD-2-Clause according to file history and
re-licensing agreement.
Update #3053.
Directly initialize the memory in the start sequence defined by start.S
instead of using a system initialization handler. This avoids using the
global variable rdb_start which used a memory location which was shared
with _ERC32_MEC_Timer_Control_Mirror. This change makes it possible to
use _Memory_Allocate() even before the system initialization is started.
Change license to BSD-2-Clause according to file history and
re-licensing agreement.
Update #3053.
Initialize the stacks for all processors in one place. Do not rely on
Per_CPU_Control::interrupt_stack_high and directly use the statically
allocated interrupt stack area.
Remove the support to load the data section and rely on the boot loader. The
code is an artifact from the old erc32 days, when we would boot and execute
from ROM and the .data had to be copied over to RAM. With leon1/2/3, this is
not used anymore as a boot loader is made from the RAM image using a custom
tool (mkprom).
In SMP configurations, this support was also broken since LEON3_Boot_Cpu
(in the data section due to the -1 initialization value) was used quite
early in the start sequence.
If the data copy is really necessary, then an application can still add this
step as a very early system initialization step, since boot_card() and the
system initialization loop does not use initialized read-write data (only
read-only and BSS data). However, the SMP startup would still not work in this
case. A boot loader is a better place to load the sections.
Set Thread_queue_Context::timeout_absolute in
_Thread_queue_Context_set_timeout_argument() to avoid using it uninitialized.
The bug was introduced by a89ecaa1a9.
A warning was present when building RTEMS that stated that the argument
for malloc() exceeded the maximum object size. To get rid of this, I
changed many places where 'int' was being used to 'size_t'.
The addition of the entire *utime*() family of functions resulted
in this call returning ENOENT not ENXIO. This is better aligned
with the POSIX definition of the methods.
Created futimens.c and utimensat.c to add support for the POSIX
methods futimens() and utimensat().
utime() and utimes() are considered obsolote by POSIX, but RTEMS
will continue to support them.
Closes#4396
Add the BSPs for running on the ZU3EG Ultrascale+ Zynq MPSoC and alter
the option defaults necessary for them to run properly using the
standard BOOT.BIN configured for PetaLinux that comes in the Out-of-Box
package.
Currently, the AArch64 BSPs have a hard time running on real hardware
without building the toolchain and the bsps with -mstrict-align in
multiple places. Configuring the MMU on these chips allows for unaligned
memory accesses for non-device memory which avoids requiring strict
alignment in the toolchain and in the BSPs themselves.
In writing this driver, it was found that the synchronous exception
handling code needed to be rewritten since it relied on clearing SCTLR_EL1 to
avoid thread stack misalignments in RTEMS_DEBUG mode. This is now
avoided by exactly preserving thread mode stack and flags and the new
implementation is compatible with the draft information provided on the
mailing list covering the Exception Management API.
The default available RAM on the A53 BSP is quite small at 8MB. This
bumps that to 128MB to avoid allocation failures in tmcontext01 caused
by large allocations on a cache size of 16MB reported by the system
registers in QEMU.
This fixes a bug where addresses were not being aligned correctly.
Addresses used in cache functions are now aligned consistently using
RTEMS_ALIGN_DOWN.
Ensure the stack remains aligned by keeping the context frame at a
multiple of 16 bytes. This avoids stack alignment exceptions which occur
when the stack pointer is not 16 byte aligned.
Commit 73ebf9a27e accidentally removed the
direct thread dispatch in a self thread restart. In case of a self
restart (always in task context) the directive shall not return. If
this is not possible due to a bad thread dispatch disable level, then a
fatal error shall occur.
Update #4412.
The _Thread_Cancel() (in contrast to _Thread_Restart() which used a
similar code block) may have produced ready threads with an active timer
in case the thread to cancel had its thread life protection enabled. The
problem was this code block:
Priority_Control priority;
_Thread_Add_life_change_request( the_thread );
if ( _Thread_Is_life_change_allowed( previous ) ) {
_Thread_State_release( the_thread, &lock_context );
_Thread_queue_Extract_with_proxy( the_thread );
_Thread_Timer_remove( the_thread );
} else {
_Thread_Clear_state_locked( the_thread, STATES_SUSPENDED );
_Thread_State_release( the_thread, &lock_context );
}
priority = _Thread_Get_priority( executing );
_Thread_Raise_real_priority( the_thread, priority );
_Thread_Remove_life_change_request( the_thread );
The life change request should only be added/removed if a life change is
allowed (see _Thread_Restart()). Add _Thread_Try_life_change_request()
and use it in _Thread_Cancel() and _Thread_Restart().
Close#4435.
This change eliminates a system initialization dependentcy which resulted
in an application without a file system or console referencing errno being
forced to include the code to open(/dev/console), close(), atexit(),
and the unmount infrastructure.
Closes#4439.
Add Thread_queue_Context::timeout_absolute to specify an absolute or
relative timeout. This avoid having to get the current time twice for
timeouts relative to the current time. It moves also functionality to
common code.
The __builtin_unreachable() cannot be used with current GCC versions to
tell the compiler that a function does not return to the caller, see:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99151
Add a no return variant of _CPU_Context_switch() to avoid generation of
dead code in _Thread_Start_multitasking() if RTEMS was built with SMP
support enabled.
Move this diagnostic function to a separate file since it does not
provide a core function of the system.
Change license to BSD-2-Clause according to file history and
re-licensing agreement.
Update #3053.
Move clock_nanosleep() to a separate file to avoid a dependency on errno
which pulls in the Newlib reentrancy support. This is an issue since
most parts which are pulled in cannot be garbage collected by the linker
due to the system initialization linker set.
The sizes are configurable via fuses or per software via some registers.
At the moment the registers are not changed. Changing the registers
destroys data stored in the RAM areas (like application code or data).
So either the fuses or some bootloader should be used to set them before
the application starts.
This also adds an OCRAM only linker command file.
Update #4180
Add addresses and interrupts for most internal peripherals to the dts.
The additional aliases make it possible for an application to easily
access these informations.
Update #4180
rtems_task_restart() may be called from within interrupt context. So
checking only that the thread to restart is equal to the executing
thread is insufficient to determine a self restart. We have to also
check that no ISR is in progress. Merge _Thread_Restart_other() and
_Thread_Restart_self() into one _Thread_Restart() since they share a lot
of common code.
Close#4412.
Add a parameter to _TOD_Validate() to disable the validation of the
ticks member.
There are two reasons for this change. Firstly, in
rtems_task_wake_when() was a double check for time_buffer == NULL (one
in rtems_task_wake_when() and one in _TOD_Validate()). Secondly, the
ticks member is ignored by rtems_task_wake_when(). This was done with a
write of zero to the ticks member and thus a modification of the
user-provided structure. Now the structure is no longer modified.
Using a mask parameter is quite efficient. You just have to load an
immediate value and there are no additional branches in _TOD_Validate().
Close#4406.
This patch fixes bug #4403. Directives
* rtems_timer_fire_when()
* rtems_timer_server_fire_when()
* rtems_task_wake_when()
are documented to return RTEMS_INVALID_ADDRESS when their time-of-day
argument is NULL. But actually they return RTEMS_INVALID_CLOCK. To fix
the issue this patch changes _TOD_Validate() to return a
status code instead of just true/false.
Close#4403
Use _Objects_Free_nothing() for rtems_message_queue_construct() to avoid
unreachable code in _CORE_message_queue_Close() in case only
user-provided message buffers are used.
Update #4007.
In the action requirements, some pre-conditions are not applicable
depending on the state of other pre-conditions. There is no particular
order in the pre-conditions, so we have to check the N/A status
individually.
This patch adds asserts to fix coverity defects
1) CID 1474437 (Out-of-bounds access)
2) CID 1474436 (Out-of-bounds access)
From manual inspection, out of bounds access cannot occur due to
bounds checking but coverity fails to detect the checks.
We are adding asserts as a secondary check.
Move all error checks into posix_memalign() so that the returned memory
pointer is set to NULL under all error conditions except
memptr == NULL.
Use parameter names of POSIX documentation.
In case the processor set is not large enough to contain the processor
set owned by the scheduler return RTEMS_INVALID_SIZE instead of
RTEMS_INVALID_NUMBER. This is more in line with other directives since
the issue is related to the size of an object.
Close#4401.
Using 32bit types like uint32_t for pointers creates issues on 64 bit
architectures like AArch64. Replaced occurrences of these with uintptr_t,
which will work for both 32 and 64 bit architectures.
Using 32bit types like uint32_t for pointers creates issues on 64 bit
architectures like AArch64. Replaced occurrences of these with uintptr_t,
which will work for both 32 and 64 bit architectures.
The allocator mutex is recursive and already used for the task creation
in rtems_timer_initiate_server(). Just use this mutex instead of the
once mutex to serialize the initialization. This avoids a dependency on
condition variables which are not used here.
The zero size allocations had no consistent behaviour in RTEMS. For
example, malloc( 0 ) returned NULL and posix_memalign( &p, align, 0 )
returned in p a unique pointer (or NULL if no memory is available). In
POSIX, zero size memory allocations are implementation-defined
behaviour. The implementation has two options:
https://pubs.opengroup.org/onlinepubs/9699919799/functions/malloc.htmlhttps://pubs.opengroup.org/onlinepubs/9699919799/functions/posix_memalign.html
Linux and FreeBSD return a unique pointer for zero size memory
allocations. Use this approach for RTEMS as well throughout the memory
allocation directives
Close#4390.
The __builtin_unreachable() cannot be used with current GCC versions to
tell the compiler that a function does not return to the caller, see:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99151
Add a no return variant of _Thread_Dispatch_direct() to avoid generation
of dead code.
If only one constraint is present, then use a single sentence, otherwise
use a list for the constraints. Format the constraints so that the line
length limit is maintained.
The 'sc' variable was originally storing the return value of mkdir().
This was causing an issue, so it was changed to make it to where we
ignored the return value with (void). The 'sc' variable was left in
by mistake.
The "powerpc/shared/console" code has the start-up console value fixed
at 9600 baud. This changes the hard-wired constant "9600" in the code
to the configuration setting "BSP_CONSOLE_BAUD" and adds configuration
support in both the "waf" and the legacy configuration systems.
Note that the VME BSPs beatnik, mvme3100, and mve5100 can be improved
by adding a "mvmexxxx" BSP family. This configuration change, as well
as future configuration changes, could then be made in a "grp.yml" file.
In case the processor set is not large enough to contain the processor
affinity set of the task return RTEMS_INVALID_SIZE instead of
RTEMS_INVALID_NUMBER. This is more in line with other directives since
the issue is related to the size of an object.
Close#4393.
Refactored the i2c driver to parse register values from the device
tree rather than hardcoding them. But still the clocks have to
initialized manually.
This patch fixes issue #4338 by changing _TOD_Validate()
to only accept years till 2105. This requires another patch
to change the documentation of rtems_clock_set() and other
affected API functions (indicating the end date is 2105 not 2514).
I tried to support till year 2514 but it turned out that
this needs changing the Timer Manager too. That in turn
would mean to change _TOD_Seconds_since_epoch( void )
from 32 to 64 bit. Sebastian pointed out that a naive extension
leads to trouble with 32 bit processors. He deemed a safe
re-implementation too costly performance wise considering
that year 2106 is far away and current binaries using RTEMS
Classic API are unlikely to be in use by 2106.
The constant TOD_SECONDS_AT_2100_03_01_00_00 in
cpukit/rtems/src/clocktodtoseconds.c happens to be wrong by
1 hour. When setting the date 2100-Feb-28 23:59:59 and then
reading the date again you will find yourself in 2100-Feb-27.
Update #4338
Set the Main Stack Pointer (MSP) to the ISR stack area end just in case
we run using the Process Stack Pointer (PSP). This helps if
applications are started by a boot loader.
Move rtems_calloc() since it only depends on rtems_malloc(). This may
make it easier to customize the heap allocator.
Change licence to BSD-2-Clause according to file history.
Update #3053.
Last use was removed by:
commit 54550e048d
Author: Sebastian Huber <sebastian.huber@embedded-brains.de>
Date: Fri May 13 08:16:30 2016 +0200
posix: Rework pthread_join()
Rework pthread_join() to use _Thread_Join().
The _CPU_ISR_Is_enabled() function operates on ISR cookies and so must
mask off the appropriate status bits. This also fixes the naming of the
parameters of the _CPU_ISR_* functions to indicate use of ISR cookies
instead of interrupt enable/disable levels.
Constantly reinitializing the Cadence UART on every character output
causes data corruption/loss on some ZynqMP hardware. Only initialize
the UART once for early output and give it a kick on startup.
Do not adjust the stack area begin address since this may confuse the
stack allocator and result in failed stack frees.
Account for the alignment overhead in the stack space size estimate.
Check that the stack size is in the expected interval.
At this point in time, /etc can be created in multiple ways. There
is a discussion (#4354) that would define a mechanism for
instantiating a base file system with some flexibility for the
set of directories included. For now, this particular mkdir()
call can fail because /etc could already have been created by
at least an initial filesystem image, the shell, or libbsd.
closes#4382.
Adds a peek function that allows (for example) a file system to suggest
the next blocks that should be used for read ahead. This can increase
the read speed of fragmented files.
Update #3689
The patch was:
gen_uuid.c: Fix two Unchecked return value from library errors
CID 1049146: Unchecked return value from library in get_clock().
CID 1049147: Unchecked return value from library in get_random_fd().
Reopen#4280
Removed the _Assert_unused_variable_equals macro due to /etc
having already been created by the network stack initialization
or an initial filesystem image.
Closes#4282
This patch imports the beagle pinmux driver from FreeBSD into
RTEMS. Previously this driver was placed in RTEMS-libBSD but
this caused double initialization of few pins once during RTEMS
initialization and second time during libBSD initialization.
The following patches port the driver from FreeBSD to RTEMS and
will remove the driver from RTEMS-libBSD.
RTEMS Ticket Update: #3784
FreeBSD head: ca12b7e9534b10af022fed60c043984dbaf4003d
The following files have been imported from FreeBSD:
1) freebsd/sys/arm/ti/am335x/am335x_scm_padconf.h
2) freebsd/sys/arm/ti/am335x/am335x_scm_padconf.c
3) freebsd/sys/arm/ti/ti_pinmux.h
4) freebsd/sys/arm/ti/ti_pinmux.c
Check only for invalid modes if a valid handler is presented to
rtems_signal_catch(). When NULL is used for the handler, ASR processing
is disabled and the mode is not used.
Update #4244.
Do not mix checks for different error indicators (RTEMS_NOT_DEFINED and
RTEMS_INVALID_NUMBER). This simplifies the specification of
rtems_semaphore_create().
MrsP semaphores are a generalization of the priority ceiling semaphores
for SMP configurations. Priority ceiling semaphores are required to use
the priority task wait queue discipline. Require this discipline also
for MrsP semaphores.
Close#4347.
Overview
========
The errata is worked around in the kernel without requiring toolchain
modifications. It is triggered the JMPL/RETT return from trap instruction
sequence never generated by the compiler and. There are also other
conditions that must must be true to trigger the errata, for example the
instruction that the trap returns to has to be a JMPL instruction. The
errata can only be triggered if certain data is corrected by ECC
(inflicted by radiation), thus it can not be triggered under normal
operation. For more information see:
www.gaisler.com/notes
Affected RTEMS target BSPs:
* GR712RC
* UT699
* UT700/699E
The work around is enabled by defining __FIX_LEON3_TN0018 at build time.
After applying the following GCC patch, GCC will set the define when
compiling for an affected multilib:
* GR712RC (-mcpu=leon3 -mfix-gr712rc)
* UT700/UT699E (-mpcu=leon3 -mfix-ut700)
* UT699 (-mcpu=leon -mfix-ut699)
When building for another multilib and TN0018 is still required, it
is possible to enable it on the RTEMS kernel configure line using the
TARGET_CFLAGS (-D__FIX_LEON3FT_TN0018) or other by other means.
The following GCC patch sets __FIX_LEON3FT_TN0018 for the affected RTEMS
multilibs:
---------
diff --git a/gcc/config/sparc/rtemself.h b/gcc/config/sparc/rtemself.h
index 6570590..ddec98c 100644
--- a/gcc/config/sparc/rtemself.h
+++ b/gcc/config/sparc/rtemself.h
@@ -33,6 +33,8 @@
builtin_assert ("system=rtems"); \
if (sparc_fix_b2bst) \
builtin_define ("__FIX_LEON3FT_B2BST"); \
+ if (sparc_fix_gr712rc || sparc_fix_ut700 || sparc_fix_ut699) \
+ builtin_define ("__FIX_LEON3FT_TN0018"); \
} \
while (0)
---------
Workaround Implementation
=========================
In general there are two approaches that the workaround uses:
A) avoid ECC restarting the RETT instruction
B) avoid returning from trap to a JMPL instruction
Where A) comes at a higher performance cost than B), so B) is used
where posssible. B) can be achived for certain returns from trap
handlers if trap entry is controlled by assembly, such as system calls.
A)
A special JMPL/RETT sequence where instruction cache is disabled
temporarily to avoid RETT containing ECC errors, and reading of RETT
source registers to "clean" them from incorrect ECC just before RETT
is executed.
B)
The work around prevents JMPL after system calls (TA instruction) and
modifies assembly code on return from traps jumping back to application
code. Note that for some traps the trapped instruction is always
re-executed and can therefore not trigger the errata, for example the
SAVE instruction causing window overflow or an float instruction causing
FPU disabled trap.
RTEMS SPARC traps workaround implementation:
NAME NOTE TRAP COMMENT
* window overflow 1 - 0x05 always returns to a SAVE
* window underflow 1 - 0x06 always returns to a RESTORE
* interrupt traps 2 - 0x10..1f special rett sequence workaround
* syscall 3 - 0x80 shutdown system - never returns
* ABI flush windows 2 - 0x83 special rett sequence workaround
* syscall_irqdis 4 - 0x89
* syscall_irqen 4 - 0x8A
* syscall_irqdis_fp 1 - 0x8B always jumps back to FP instruction
* syscall_lazy_fp_switch 5 - 0x04 A) jumps back to FP instruction, or to
B) _Internal_error() starting with SAVE
Notes:
1) no workaround needed because trap always returns to non-JMPL instruction
2) workaround implemented by special rett sequence
3) no workaround needed because system call never returns
4) workaround implemented by inserting NOP in system call generation. Thus
fall into 1) when workaround is enabled and no trap handler fix needed.
5) trap handler branches into both 1) and returning to _Internal_error()
which starts with a SAVE and besides since it shuts down the system that
RETT should never be in cache (only executed once) so fix not necessary
in this case.
Any custom trap handlers may also have to be updated. To simplify that,
helper work around assembly code in macros are available in a separate
include file <libcpu/grlib-tn-0018.h>.
Close#4155.
Without this smp05 and smpthreadlife01 tests may fail
depending on how the boot loader initialized the GPTIMER.
Before the time counter stopped counting when reaching
zero, but tests could work since it could take 2^32 us
before stopping.
The timer driver will potentially overwrite this, but it
happens later due to the initialization order having
RTEMS_SYSINIT_CPU_COUNTER very early.
Moves drvmgr_drivers[] from amba.c to a separate file in order
to avoid the dependecy on APBUART/GPTIMER drivers. This has
an effect when user configured not to use timer or uart
in their project.
The descriptor table size is equal to its alignment and set when
configuring the HW IP through VHDL generics. This SW patch simply
probes the HW how large the RX/TX descriptor tables are and adjusts
accordingly.
The number of descriptors actual used are controlled by other
settings (rxDescs and txDescs) controlled by the user.
The new GRCAN_FD IP supports CAN FD standard and is mostly backwards
compatible with GRCAN SW interface. The GRCAN driver have been extended
to support the GRCANFD IP using the same driver.
Additional functions have been added that uses a new CAN FD frame
format and read/write/baud-rate functions that supports both GRCANFD
and GRCAN. To keep the SW API fully backwards compatible with GRCAN,
the old functions remain.
Update #4324.
When the DMA table has been allocated dynamically, the IOCTL_SET_PACKETSIZE
will trigger an issue where pDev->rx and pDev->tx are not updated with
the new DMA tables base address. Instead the old pointers are used.
There is no point in reallocting the DMA tables because there is no
configuration option to it. Therefore the DMA tables allocation is
moved to a separate function never called from SET_PACKETSIZE.
CID 1399726: Missing break in switch in task_usage().
CID 1399728: Missing break in switch in task_usage().
CID 1399742: Missing break in switch in task_usage().
Closes#4278
CID 1399761: Missing break in switch in process_dma().
CID 1399765: Missing break in switch in process_dma().
CID 1399766: Missing break in switch in process_dma().
CID 1399777: Missing break in switch in process_dma().
Closes#4277
The current ZynqMP BSPs don't have _qemu in their name as do all other
RTEMS BSPs that are specifically made to run on QEMU. This fixes the
naming for those ZynqMP BSP variants for easier identification.
Run with stack alignment faults enabled under RTEMS_DEBUG to catch any
stack misalignments early. This makes it easier to track them down
should they ever occur.
AArch64 on hardware is often started at EL2 instead of EL1 from either
u-boot or a first stage bootloader. This allows RTEMS to drop from EL2
execution to EL1 to operate as normal.
According to commentary on GCC bug
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99143, the alignment
behavior of linker sections on which RTEMS has relied was never
guaranteed to be consistent across platforms and any alignment
requirements for linker sections needs to be enforced explicitly.
This adds those explicit alignment requirements.
Closes#4255.
ZynqMP hardware appears to have an odd hard-wired SGI implementation in
which the SGIs are permanently set as enabled or disabled. Allow the
TM27 IRQs to be overridden as necessary.
Remove usage of SUBALIGN() in aarch64 linkcmds which works around a
difference in behavior on AArch64 platforms. This is no longer necessary
since alignment is now enforced explicitly.
Closes#4178.
Make sure that a user-provided stack size is the minimum size allocated
for the stack.
Make sure we meet the stack alignment requirement also for CPU ports
with CPU_STACK_ALIGNMENT > CPU_HEAP_ALIGNMENT.
- The change to building all code with code and data sections means
we have a section per function. Make sure all functions are
placed in the text section.
Closes#4266
Close the thread object if a thread create extension fails. Also call
the delete extension to avoid resource leaks in early extensions if a
late extension fails.
Close#4270.
Use the right system initialization step for the stack allocator
initialization. Do the stack allocator initialization before the
workspace initialization so that _Memory_Allocate() can be used to get
memory for the allocator.
This type is not only used by Task Manager directives. This is a
preparation to move the rtems_scheduler_*() directives from
<rtems/rtems/tasks.h> to a new <rtems/rtems/scheduler.h>.
Do the mode changes necessary for the ASR processing directly under
protection of the thread state lock to avoid the recursive calls to
thread dispatching done in rtems_task_mode().
Close#4244.
Move _Signal_Action_handler() and signal MP initialization to
rtems_signal_send(). The goal is to make _Signal_Action_handler() a
static function in a follow up patch.
Update #4244.
Require that the corresponding lock is acquired before the action
handler returns. This helps to avoid recursion in the signal
processing.
Update #4244.
- These calls clash with the Linux IO header in LibBSD. Making these
conditional here means BSPs build and the imported Linux header is
untouched.
Updates #4245
- Add support to the BSP to enable irq-generic management
- Update the powerpc shared irq code to support irq-generic. This
is an opt in option for existing powerpc bsps. This change
should be simpler now
- Fix a number of issues in ISA IRQ controller handling by porting
fixes from the i386 (PC) BSP
Closes#4238Closes#4239
CID 1468681: Dereference before null check in rtems_debugger_thread_continue().
CID 1468690: Dereference before null check in rtems_debugger_thread_system_resume().
CID 1468694: Dereference before null check in rtems_debugger_thread_find_index().
Closes#4241.
CID 1437625: Unchecked return value from call to rtems_task_create().
CID 1472765: Unchecked return value from call to rtems_task_start().
Closes#4237.
Change RTEMS_STRING() and RTEMS_XSTRING() to accept a variable number of
arguments which is stringified. This can be used for example to create
register lists for inline assembler statements.
- Add the bool flag no_regular_file_mknod to the mount table so a file
system can indicate creating regular files is not done by
use the mknod handler. The file system will handle creating a
file node in the open handler.
- Note, the mount option is an enum which means there is only one
exclusive option supported. As a result no encapsulation is
provided and file systems need to set no_regular_file_mknod directly.
Closes#4222
- The call to enable the openpic irq for the ISA bridge fails
because the IRQ used is offset by the ISA bus signals and
the openpic call expects an IRQ relative to its signals.
- Add the MVME 2600/2700 to the list of boards with an ISA bridge.
Closes#4231
Add a "Constraints" paragraph to the documentation.
Provide prototypes for programming language bindings. Use the macro
implementation to implement the corresponding functions.
Update #3993.
The goal is to reduce false positive NULL_RETURNS,
PW.NORETURN_FUNCTION_DOES_RETURN, PW.SET_BUT_NOT_USED,
UNUSED_VALUE, etc. issues produced by Coverity.
Define RTEMS_STATIC_ANALYSIS if a static analysis run by Coverity is
performed.
Use the idle stack to buffer early uses of printk(). Print the buffered
characters during initialization when the UART is available and before
the idle stack is used normally.
This fix relates to a Coverity issue (PW.SET_BUT_NOT_USED).
Do not shadow the queue_context parameter with a local variable. Reuse
the queue context to reduce the required stack space.
This fix relates to a Coverity issue (PW.PARAMETER_HIDDEN).
We do not need all the checks if we have a valid indentifier to a thread
class object.
Using the new _Thread_Get_objects_information() instead of the inline
function _Thread_Get_objects_information_by_id() avoids dead code since
the identifier in a thread control is always valid and the return NULL
path in _Thread_Get_objects_information_by_id() would be dead code. The
_Thread_Get_objects_information_by_id() should be an inline function
since it is used by _Thread_Get() and thus performance critical. Static
analyzers which cannot derive that the identifier in a thread control is
always valid, may find a potential NULL pointer access (or otherwise
find dead code).
The identifier in an object control is always valid, see
_Objects_Initialize_information() and _Objects_Extend_information().
Move _RTEMS_tasks_Free() to the only source file which calls this
function.
Rename _Thread_Get_objects_information() in
_Thread_Get_objects_information_by_id() to emphasize that this thread
method uses an object identifier and not a thread control.
There is not need to check that vector >= BSP_INTERRUPT_VECTOR_MIN since
BSP_INTERRUPT_VECTOR_MIN is zero and vector is unsigned.
This fix relates to CID 1399742 (NO_EFFECT).
If CPU_DATA_CACHE_ALIGNMENT == CPU_INSTRUCTION_CACHE_ALIGNMENT we had
dead code with the previous implementation.
This fix relates to CID 1399776 (DEADCODE).
The maximum frequency is UINT32_MAX. Converted to a uint64_t variable
it can be shifted by 32. The addition does not overflow since bin_per_s
- 1 is UINT32_MAX.
GCC 11 produced warnings like this:
items.c:21:1: warning: ignoring attribute
'section (".rtemsrwset.test_rw.content.0.1")' because it conflicts with
previous 'section (".rtemsrwset.test_rw.content.1")' [-Wattributes]
items.c:23:1: warning: ignoring attribute
'section (".rtemsroset.test_ro.content.0.OC")' because it conflicts with
previous 'section (".rtemsroset.test_ro.content.1")' [-Wattributes]
For i.MX7 U-Boot initializes the system counter. On i.MX6 Barebox is
often used which doesn't initialize the counter. With this patch, we try
to auto-detect whether the counter is initialized or not and do the
initialization ourself if necessary.
Closes#3869
This allows an application to get the registers of the LPSPI. That is
usefull for applications that want to use DMA for a very specialized and
highly optimized communication.
Update #4180
Note: The changes have been done with portability in mind. The driver
should (in theory) be able to replace the original one in the MPC BSPs
too. For full compatibility an adaption layer and especially a test
would be necessary. Because both are missing, don't integrate it into
the MPC BSP now.
Update #4180
Return STATUS_DEADLOCK (RTEMS_INCORRECT_STATE) to indicate a nested
seize since this is a kind of deadlock. This status code is also used
for other deadlocks.
Update #4217.
RTEMS OFW is a FDT only implementation of the OpenFirmWare
interface. This API is created to be compatible with FreeBSD
OpenFirmWare interface. The main intention is to make
porting of FreeBSD drivers to RTEMS easier.
Most functions implemented have an direct one-one mapping
with the original OFW API and some extra auxiliary functions
were implemented to make working with device trees easier in
RTEMS.
Update #3784
This BSP supports the Arm Fixed Virtual Platform. Only the Cortex-R52
processor configuration is supported by the BSP. It should be easy to
add support for other variants if needed.
Update #4202.
Set the VBAR to the vector table in the start section before
bsp_start_hook_0() is called to earlier handle exceptions in RTEMS.
Set the VBAR to the normal vector table in start.S for the main
processor. Secondary processors set it in bsp_start_hook_0().
Update #4202.
Somehow the table index has been missing in the latest patch version.
With that, the configuration for the first region has been applied
multiple times.
Update #4180
This adds some commands that are usefull for debugging simple serial
interfaces.
Even if they are a complete re-implementation, the i2c* commands use a
simmilar call like the Linux i2c tools.
Expressions in static assertions must be integral constant expressions. In
integral constant expressions the use of address constant expressions is not
allowed.
In static initializers the address constant expressions are allowed. Introduce
a new macro _CONFIGURE_ASSERT_NOT_NULL() which leads to a compile time error if
the second parameter is NULL. It generates error messages like this if for
example
#define CONFIGURE_INIT_TASK_ENTRY_POINT NULL
is provided by the application:
cpukit/include/rtems/confdefs/inittask.h:51:26: error: size of unnamed array is negative
51 | ( _type ) sizeof( int[ ( _value ) != NULL ? 1 : -1 ] ) )
| ^
cpukit/include/rtems/confdefs/inittask.h:170:3: note: in expansion of macro '_CONFIGURE_ASSERT_NOT_NULL'
170 | _CONFIGURE_ASSERT_NOT_NULL(
| ^~~~~~~~~~~~~~~~~~~~~~~~~~
This fix relates to CID 1470570 (PARSE_ERROR).
Update #4181.
The arm_cp15 function for accessing the current CPU index is specific
to ARMv7 while this header is used for ARMv8 as well. Instead, use a
generic accessor that is part of the standard CPU API.
Use the targets parameter to determine the targets of the SGI. Change
targets parameter type to 32-bit to ease the parameter passing. GICv3
supports up to 16 targets.
Update #4202.
Currently, zynq-uart code is always built and has some requirements for
BSPs that use it. Instead of making all BSPs satisfy that requirement or
working around it by setting defaults, this moves the zynq-uart code
into its own spec build object so it can be included if needed.
This adds a BSP family that runs on the Xilinx Ultrascale+ MPSOC
(ZynqMP) family of chips. It is configured to be usable on the Qemu
ZCU102 machine definition and should be almost trivially portable to
ZynqMP development boards and custom hardware. It is also configured to
be usable with libbsd.
The zynq-uart set_attributes implementation was configured to always
return false which causes spconsole01 to fail. This restores the
disabled implementation which sets the baud rate registers
appropriately and allows spconsole01 to pass. This also expands the
set_attributes functionality to allow setting of the stop bits,
character width, and parity.
For architectures with relatively large minimum stack sizes, this test
breaks during compilation due to a static check. The init task stack
size should not need to be set for this test.
The libnetworking is excluded from the Doxygen output. Exclude also the
nfsclient source code, otherwise there are some orphaned Doxygen groups.
Update #3706.
Rename Classic API top-level group from Classic to RTEMSImplClassic.
Use RTEMSImplClassic as a prefix for the subgroups. Change the group
names to be in line with the API group names. Use common phrases for
the group brief descriptions.
Update #3706.
This directive has no implementation. It is documented in the RTEMS
Classic API Guide.
It was added by
commit 6c06288f64
Author: Joel Sherrill <joel.sherrill@OARcorp.com>
Date: Tue Jan 29 21:52:21 2008 +0000
without an implementation. The later change
commit fdc57ca4b6
Author: Joel Sherrill <joel.sherrill@OARcorp.com>
Date: Mon Nov 23 14:53:04 2009 +0000
renamed the similar rtems_object_id_api_minimum_class() in
rtems_object_api_minimum_class().
The rtems_object_api_maximum_class() is documented and implemented. It
can be assumed that the rtems_object_id_api_maximum_class() is a
fragement left over from development.
In order to better support applications which use the new
rtems_task_construct() directive add the
CONFIGURE_INIT_TASK_CONSTURCT_STORAGE_SIZE configuration option. If
this option is specified, then the Classic API initialization task is
constructed with rtems_task_construct().
Update #4181.
Change the _Per_CPU_Information declaration so that Doxygen can parse
it. This attribute placement is also in line with the _Alignas()
specifier of C11.
The spintrcritical03-05 and psx12 tests are known to fail on Qemu when
the host system is heavily loaded. A single run of Qemu per core
during a testsuite run tends to yield positive results, but any
additional load on a system will result in test failures.
This patch also applies the correct expected test state for intermittent
failures so that those tests will still build.
The variety of expected test states are not currently applied to tests
with names containing '-' correctly due to a failure to replace '-' with
'_' before adding the CPPFLAGS to the environment for that test. This
ensures that all additions of CPPFLAGS have that replacement performed
so that the CPPFLAGS are applied properly during compilation.
Also start interrupt server tasks on processors which do not have a
scheduler. Applications may dynamically manage processors using
rtems_scheduler_remove_processor() and rtems_scheduler_add_processor().
The SUBALIGN(4) required on rtemsroset and rtemsrwset for ILP32 builds
was previously present on LP64 builds and causes no issues within RTEMS,
but causes relocation/alignment issues when building libbsd. This
restricts those alignment changes to ILP32 builds.
This primitive test case memory allocator uses memory from the low-level
memory information provided by the BSP. At the beginning of each test
case, the memory available to the test case is reinitialized. This
allows the use of a simple allocate only allocator.
In "bsp/atsam: Simplify XDMAD_Handler()" (5f813694f6) the interrupt
callback has been made unconditional. That allowed to avoid some special
deadlock situations in error cases. But it removed part of the XDMAD
status handling.
This patch adds the ability to update the XDMAD status from the
callback if that is necessary for the driver.
Fixes#4173
This improves support for the --out option. Previously, the cache file
was placed in the source directory under "build/...". Now, it is placed
in the output directory, which is "build" by default. So, if you don't
use the --out option nothing changes. However, if you use the --out
option, then the cache file is placed under the specified directory.
The compiler warning was:
../../../cpukit/libmisc/rtems-fdt/rtems-fdt.c:267:5: warning:
'strncpy' specified bound depends on the length of the source argument
267 | strncpy(path, name, namelen);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
It turns out that the `strncpy()` nor the buffer `path` is needed when
one uses `strncmp()` instead of `strcmp()`. This needs some change to
the algorithm but has the advantage that `name` is never truncated
to the size of the buffer `path`.
Prevent a call to _SMP_lock_Stats_register_or_max_section_time(). This fixes a
thread stack corruption in case RTEMS_PROFILING and RTEMS_SMP is enabled.
Close#4157.
Add a workaround for Cortex-A9 Errata 845369: A short loop including a DMB
instruction might cause a denial of service on another which executes a CP15
broadcast operation.
Close#4115.
Add a workaround for Cortex-A9 Errata 845369: Under Very Rare Timing
Circumstances Transition into Streaming Mode Might Create Data Corruption.
Update #4115.
The compiler warning was:
../../../cpukit/libmisc/rtems-fdt/rtems-fdt.c:267:5: warning:
'strncpy' specified bound depends on the length of the source argument
267 | strncpy(path, name, namelen);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
It turns out that the `strncpy()` nor the buffer `path` is needed when
one uses `strncmp()` instead of `strcmp()`. This needs some change to
the algorithm but has the advantage that `name` is never truncated
to the size of the buffer `path`.
Note:
rtems-fdt.c, rtems-fdt-shell.c and cpukit/include/rtems/rtems-fdt.h
seem to be dead code. They implement a shell command `fdt` but that
command is not part of the shell nor of any macro in
cpukit/include/rtems/shellconfig.h.
Using strlcpy() instead of strncpy():
1) Prevents the compiler warnings
2) Ensures, the string is NUL terminated.
3) Avoids that strncpy() unnecessary fills the unused part of the buffer with
0 bytes.
(Note that realpath() also returns NULL if the file does not exist - that
happens always if someone creates a new file with the editor of the shell.)
This is an illegal use of strcpy() because one is not allowed to
use this function with overlapping source and destination buffers;
whereas memmove() is explicitly designed to handle such cases.
The copiler warning was:
../../../cpukit/libmisc/shell/shell.c:626:13: warning:
'strcpy' accessing between 1 and 2147483645 bytes at offsets
0 and [1, 2147483647] may overlap up to 2147483644 bytes at
offset [1, 2147483644] [-Wrestrict]
This patch fixes a compiler warning:
../../../testsuites/fstests/fsdosfsname01/init.c:430:19: warning:
'%s' directive output may be truncated writing up to 6424 bytes into
a region of size 257 [-Wformat-truncation=]
The buffer 'dirname' is exactly large enough so that no truncation
can ever occur. Using the return value of snprintf() is an official
supported way to suppress the warning.
I considered the comment of Joel Sherrill about not replacing snprintf():
https://lists.rtems.org/pipermail/devel/2020-September/062113.html
This fixes warnings like this:
warning: implicit declaration of function 'rtems_interrupt_disable'
warning: implicit declaration of function 'rtems_interrupt_enable'
warning: implicit declaration of function 'rtems_interrupt_flash'
When I use the 'shell' from the fileio sample with the command below:
env QEMU_AUDIO_DRV="none" \
qemu-system-arm -no-reboot -net none -nographic -M realview-pbx-a9 \
-m 256M \
-kernel build/arm/realview_pbx_a9_qemu/testsuites/samples/fileio.exe
The executable crashes with an "BLOWN STACK!!!" as soon as I try to login
as 'root' with password. (The logins without password work fine.)
Increasing the stack size of the affected thread a bit solves the issue.
Hence, I suggest this patch.
My config.ini was
[arm/realview_pbx_a9_qemu]
RTEMS_DEBUG = True
RTEMS_NETWORKING = True
RTEMS_POSIX_API = True
RTEMS_SMP = True
BUILD_TESTS = True
RTEMS origin.master at a479686c11Close#4143
The shell has an 'fdisk' command which has sub-commands 'mount' and 'unmount'.
These two sub-commands have a bug which causes them to be not able
to mount anything. This proposed patch removes the buggy file
cpukit/libblock/src/bdpart-mount.c and the mount/unmount commands
from 'fdisk' as bug fix. The 'fdisk' command itself is not removed.
The reasons for removing the sub-commands (instead of fixing the issue) are:
1) The bug has been introduced on 2010-May-31 with commit
29e92b090c. Since ten years no one
can use this feature, nor has anybody complained about it.
2) Besides of the 'fdisk' 'mount' sub-command, the shell has the
usual 'mount' and 'unmount' commands which can serve as
substitutes.
3) There are additional minor issues (see further down) which needed to
be addressed when the file will be kept.
What follows below is the precise bug description.
The bug is in function rtems_bdpart_mount() which is only be used
by the 'fdisk' shell command to mount all partitions of a disk with a
single command:
> fdisk DISK_NAME mount
> mounts the file system of each partition of the disk
>
> fdisk DISK_NAME unmount
> unmounts the file system of each partition of the disk
The whole command does not work because in file
cpukit/libblock/src/bdpart-mount.c line 103 specifies the file system type
of each partition to be "msdos". Yet, "msdos" does not exist. The name
must be "dosfs".
Beside of this fundamental problem, there are more issues with the code
in bdpart-mount.c:
1) The function returns RTEMS_SUCCESSFUL despite the mount always fails.
2) The reason for errors is not written to the terminal.
3) The directory '/mnt' is created but not deleted later on (failure or not).
3) There is no documentation about this special 'fdisk' feature in the
RTEMS Shell Guide ('fdisk' is mentioned but its documentation is a
bit short):
https://docs.rtems.org/branches/master/shell/
file_and_directory.html#fdisk-format-disk
4) Only "msdos" formatted partitions can be mounted and all partitions
are mounted read-only. This is hard coded and cannot be changed by
options. Moreover, there is no information about this to the user of
the shell (i.e. using 'fdisk' mount requires insider knowledge).
How to reproduce:
1) For testing, I use the 'testsuites/samples/fileio.exe' sample with qemu:
> cd rtems
> env QEMU_AUDIO_DRV="none" qemu-system-arm -net none -nographic \
> -M realview-pbx-a9 -m 256M -kernel \
> build/arm/realview_pbx_a9_qemu/testsuites/samples/fileio.exe
2) Type any key to stop the timer and enter the sample tool.
Type 's' to enter the shell, login as 'root' with the password
shown in the terminal.
3) Type the following shell commands (they create a RAM disk,
partition it, register it, format it and try to mount it):
> mkrd
> fdisk /dev/rda fat32 16 write mbr
> fdisk /dev/rda register
> mkdos /dev/rda1
> fdisk /dev/rda mount
4) The last line above is the command which fails - without an error
message. There exists a '/mnt' directory but no '/mnt/rda1' directory
as it should be:
> ls -la /mnt
5) If you change line 103 of 'cpukit/libblock/src/bdpart-mount.c'
from "msdos" to "dosfs", compile and build the executable and
re-run the above test, '/mnt/rda1' exists (but the file system
is mounted read-only).
Close#4131
This is actually an illegal use of strcpy() because one is not allowed to
use this function with overlapping source and destination buffers; whereas
memmove() is explicitly designed to handle such cases.
The compiler warning was:
../../../cpukit/libmisc/monitor/mon-editor.c:342:15: warning:
'strcpy' accessing 1 byte at offsets [0, 75] and [0, 75] overlaps
1 byte at offset [0, 74] [-Wrestrict]
This fixes the following compiler warning:
testsuites/psxtests/psxndbm01/init.c:221:3: warning: 'strncpy' output truncated
before terminating nul copying 5 bytes from a string of the same length
221 | strncpy( test_strings, "Hello", 5 );
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In addition, the comments from Sebastian Huber on an old version of
such a patch have been taken into account:
1) The use of `sizeof()` in `key.dsize = sizeof( test_strings );` is wrong.
2) There is no need to allocate the string. One can simply use a string
constant.
(See https://lists.rtems.org/pipermail/devel/2020-August/061418.html)
The manager documentation is a consolidation of the comments in Doxygen
markup and the documentation sources in Sphinx markup. The
documentation was transfered to interface specification items. This
header file was generated from the items by a script.
Change license to BSD-2-Clause according to file histories and
documentation re-licensing agreement.
Update #3899.
Update #3993.
This fixes the compiler warning below.
../../../cpukit/mghttpd/mongoose.c:1919:45: warning:
'.gz' directive output may be truncated writing 3 bytes into a region
of size between 1 and 255 [-Wformat-truncation=]
This patch does not only fix the compiler warning below. memcpy() is the
better function at this place as the terminating NUL character is never
copied here. Instead more characters will be appended to the
'logical_disk_name' later on.
../../../cpukit/libblock/src/bdpart-register.c:41:5:
warning: 'strncpy' output truncated before terminating nul copying
as many bytes from a string as its length [-Wstringop-truncation]
The strlcpy() function ensures there is always a NUL character
at the end of the string. Hence it is safer as strncpy() and it avoids the
compiler warning
Fix a compiler warning about a missing prototype.
If you wonder why the function name is not `static`:
The code in the file only serves a test where the dynamic object loader
(aka dlopen()) should handle duplicated symbols in an archive.
testsuites/libtests/dl10/dl10-o6.c:14:5: warning: no previous prototype
for 'rtems_main_o5' [-Wmissing-prototypes]
14 | int rtems_main_o5 (void)
| ^~~~~~~~~~~~~
rtems_name is a four byte integer.
Giving an rtems_name as value instead of a pointer to ctrace_task_name_add()
fixes not only the compiler warning but it is also a bit more safe
For those who have asked for the warning:
../../../cpukit/libmisc/capture/capture_support.c:352:49: warning:
taking address of packed member of 'struct rtems_capture_task_record'
may result in an unaligned pointer value [-Waddress-of-packed-member]
352 | ctrace_task_name_add (rec_out->task_id, &task_rec.name);
| ^~~~~~~~~~~~~~
Provide RTEMS_NO_RETURN also in case RTEMS_DEBUG is defined to prevent errors
like this:
error: no return statement in function returning non-void [-Werror=return-type]
Use C11 and C++11 standard means to declare a no-return function.
Close#4122.
This breaks AArch32-specific portions of the ARM GPT driver into their
own file so that the generic code can be moved for reuse by other
architectures.
Before this patch RTEMS_Malloc_Initialize() had a fixed dependency on
_Workspace_Area. Introduce _Workspace_Malloc_initializer to have this
dependency only if CONFIGURE_UNIFIED_WORK_AREAS is defined by the
application configuration.
The Autoconf/Automake build system did something similar. This fixes
the build of sigprogmask.c which uses RTEMS_POSIX_API and expects to get
it defined via "config.h".
Update #3818.
This patch fixes a tiny bug in the command line editing of the RTEMS shell.
Typing CTRL-U in the shell should remove all characters left of the cursor.
After pressing CTRL-U, the current implementation does wrongly place the cursor
at the end of the line instead at its beginning.
To reproduce the bug, start the shell and type 'abc123' (no <RETURN>):
> ~/src/rtems $ qemu-system-arm -net none -nographic -M realview-pbx-a9 \
-m 256M -kernel build/arm/realview_pbx_a9_qemu/testsuites/libtests/dl10.exe
> *** BEGIN OF TEST libdl (RTL) 10 ***
> *** TEST VERSION: 6.0.0.d9bdf166644f612dd628fe4951c12c6f8e94ba5f
> *** TEST STATE: USER_INPUT
> *** TEST BUILD: RTEMS_DEBUG RTEMS_NETWORKING RTEMS_POSIX_API RTEMS_SMP
> *** TEST TOOLS: 10.2.1 20200904 \
(RTEMS 6, RSB 31f936a7b74d60bda609a9960c6e1a705ba54974, Newlib a0d7982)
> RTL (libdl) commands: dl, rtl
>
> RTEMS Shell on /dev/foobar. Use 'help' to list commands.
> SHLL [/] # abc123
Then move the cursor onto the '1' by hitting three times the <ARROW-LEFT> key.
Next type <CTRL>-U:
> SHLL [/] # 123
Note that the cursor is at the end of the line (after '3') instead of correctly
at the beginning (on the '1'), now.
Continuing typing 'echo ' incorrectly results in the output:
> SHLL [/] # 123echo 123
The patch changes this behavior so that the cursor in the second last step will
be on the '1' and typing 'echo ' will then correctly reflected as:
> SHLL [/] # echo 123
Close#4097.
In contrast to message queues created by rtems_message_queue_create(), the
message queues constructed by this directive use a user-provided message buffer
storage area.
Add RTEMS_MESSAGE_QUEUE_BUFFER() to define a message buffer type for message
buffer storage areas.
Update #4007.
Add Message_queue_Control::is_global if RTEMS_MULTIPROCESSING is defined. This
reduces the Message_queue_Control size in standard RTEMS configurations.
Update #4007.
Move the CORE_message_queue_Buffer definition to a separate header file to be
able to use it independent of the remaining Message Queue Handler API.
Change license to BSD-2-Clause according to file history.
Update #3053.
Update #4007.
Merge CORE_message_queue_Buffer structure into
CORE_message_queue_Buffer_control.
Use a zero-length array for the actual message buffer. This reduces the
structure size on all targets.
Update #4007.
The previous multiplication error check is broken on 64-bit machines. Use the
recommended check from SEI CERT C Coding Standard, "INT30-C. Ensure that
unsigned integer operations do not wrap".
Make sure the message size computation does not overflow.
Update #4007.
commit 20c89ab7c5091ee48535392cae2177aa1a1c43eb
Author: Segher Boessenkool <segher@kernel.crashing.org>
Date: Fri Jan 12 21:50:52 2018 +0100
rs6000: Remove -mstring
-mstring is only enabled by default on 601, and with -Os on some
configurations. It is almost always slower (than not using it) and
does not very often lead to smaller code.
This patch disables it. If a user uses -mstring he gets a warning
(but not with -mno-string). I left the target attribute in place, it
just doesn't do anything anymore.
The patch also deletes a whole bunch of code. The 'N' and 'O' output
modifiers are now unused, but now is not the time to delete them.
Use -B for the linker since some architectures use extra start files defined by
the GCC specification and provided by the BSP, for example powerpc. Start
files are not found by -L.
In the long run, the GCC defined start files should be provided by GCC. This
is a GCC 11 topic.
Update #3818
Add a general purpose test suite for validation tests.
This is the first test suite generated from a specification item in the
rtems-central repository.
Update #3959.
In contrast to rtems_task_create() this function constructs a task with
a user-provided task storage area. The new directive uses a
configuration structure instead of individual parameters.
Add RTEMS_TASK_STORAGE_SIZE() to calculate the recommended size of a
task storage area based on the task attributes and the size dedicated to
the task stack and thread-local storage. This macro may allow future
extensions without breaking the API.
Add application configuration option
CONFIGURE_MINIMUM_TASKS_WITH_USER_PROVIDED_STORAGE to adjust RTEMS
Workspace size estimate.
Update #3959.
Add this application configuration option. This configuration option can be
used to reserve space for the dynamic linking of modules with thread-local
storage objects.
Add RTEMS_TASK_STORAGE_ALIGNMENT to define the minium alignment of a
thread-local storage size.
Update #4074.
QEMU is now stricter with MMIO sizes and accesses. uintptr_t on RV64
is 8 bytes and generates an sd instruction that Store/AMO faults
because sifive_test MMIO expects 4 bytes accesses.
Add a feature to enforce an explicit target file for assembler sources.
Add a build start file node list and use it as a test program
dependency.
The fix for #3846 and #4080 needs to be combined, because the fix
for #3846 requires the removal of 'before=["cstlib"]'. This patch fixes
two issues:
1. The tracking of start file dependencies.
2. Reflect that executables depend on the start files.
We need a start.o file in the right path so that the linker can find is
as specified by the linker script, and not for example a start.S.17.o
file in some path. This part is addressed by the "explicit_asm_target"
feature.
This build process extension
@after("apply_link")
@feature("cprogram", "cxxprogram")
def process_start_files(self):
if getattr(self, "start_files", False):
self.link_task.dep_nodes.extend(self.bld.start_files)
addresses 2.
Close#3846.
Close#4080.
In a multi-processor system we must broadcast the TLB maintenance operation to
the Inner Shareable domain to ensure that the other processors update their TLB
caches accordingly.
Close#4068.
Do the stack allocator initialization and sanity check only if a
user-provided stack allocator was configured. This avoids a dependency
of _Thread_Handler_initialization() on the stack allocator.
Update #3959.
Only provide the file descriptor array if
CONFIGURE_MAXIMUM_FILE_DESCRIPTORS > 0. If someone configured
CONFIGURE_MAXIMUM_FILE_DESCRIPTORS == 0 and the appplication uses
something which requires a file descriptor, then a linker error will
show up. An alternative would be to add a zero-length array
rtems_libio_iops[ 0 ] to librtemscpu.a which would be used in this
case.
Conditionally include some header files.
Remove the superfluous invalid name check since the object creation
directives ensure that objects with such a name cannot exist. Also
finding an object with such a name would be no catastrophy if it really
exists.
Simplify object name to identifier directives. Using
_RTEMS_Name_to_id() to implement the directives enables a tail call
optimization.
Change license to BSD-2-Clause according to file history.
Update #3053.
The argument to the ctype functions must be an int and the value of the
character must be representable as an unsigned char or equal to the
value of the macro EOF. If the argument has any other value, the
behavior is undefined.
The RTEMS_BARRIER_AUTOMATIC_RELEASE and RTEMS_BINARY_SEMAPHORE options
had the same value. In order to better detect a misuse of option values
(for example using RTEMS_BINARY_SEMAPHORE for rtems_barrier_create()),
the options should have unique values.
Close#4054.
Normally, the expected test step must be a compile time constant. Allow
variable expected test steps for the T_step() and T_assert_step(). This
can be used for parameterized test loops with individual fixtures.
Remove the ability to use custom failure messages due to some
implementation constraints.
Update #3199.
In the nominal case checked by spsysinit01, obj_err is unmodified if
_POSIX_Shm_Get_by_name returns non-NULL. In the case of shm_unlink, this means
an uninitialized value is passed into the switch and it appears tests using it
were passing by virtue of the stack having the right value on it in most cases.
This now checks obj_err only if _POSIX_Shm_Get_by_name returns NULL.
Close#4016
Add support to record thread switch events. This can be used to check
that a blocking operation results in the expected sequence of thread
switches.
Update #3199.
Add a new directory for Doxygen-specific documentation content. Add a
Doxygen only header file containing documentation of the application
configuration options. The header file is generated from specification
items.
Close#3994.
Since we need to test the strncpy function, using a character array with
a fixed array size in this case in place of character pointer can avoid
the string turncation warning.
The external UART over SPI device SC16IS752 uses the interrupt server
for interrupt processing. The interrupt server is also heavily used by
libbsd. The interrupt processing for the SC16IS752 is time critical and
doesn't work if network traffic is processed at the same priority.
With #4033 custom interrupt servers are available. Change
atsam_sc16is752_spi_create() to support user-defined interrupt servers.
Introduced atsam_sc16is752_spi_config to cut down the argument count of
this function.
Close#4039.
This addresses compiler warnings like this:
warning: array subscript 0 is outside the bounds of an interior
zero-length array 'abc[0]' [-Wzero-length-bounds]
Currently if mount fails, a converter isn't destroyed. We have to take
care of two cases:
1. The user doesn't provide a converter.
In this case mounting a dosfs creates a default converter. This patch
makes sure that the converter is destroyed again if mount failes for
this case.
2. The user provides a converter.
In this case it's not sure that the dosfs specific routines are reached
because mount can fail before that. Therefore the user has to destroy
the converter himself again. This patch adds a documentation for that
and implements it in the media server.
Closes#4042.
Add rtems_interrupt_server_destroy().
Before this patch, the only way to create interrupt servers was
rtems_interrupt_server_initialize(). This function creates the default
interrupt server and in SMP configurations additional interrupt servers
for the additional processors. The interrupt server is heavily used by
libbsd. This includes the epoch based reclamation which performs time
consuming resource and memory deallocation work. This does not work well
with time critical services, for example an UART over SPI or I2C. One
approach to address this problem is to allow the application to create
custom interrupt servers with the right priority and task properties.
The interrupt server API accounted for this, however, it was not
implemented before this patch.
Close#4034.
The chip select lines of the iMX SPI module doesn't work well for a
generic API like the one RTEMS uses. The existing solution only worked
in some special cases and had odd bugs when trying transfers of
different sizes (like deselecting between each byte for lengths that are
not dividable by 4).
With this patch the same approach like on FreeBSD or Linux is used:
Treat the CS lines as GPIOs.
Update 3869
Having a duplicate header with the shared includes is a build system
inconsistency. You may use one header file to build the libraries and
another one is installed (overwriting the other).
Update #3269.
Make sure that the esp is restored before the eflags register.
When the init task is initially restored, system interrupts are activated when the
eflags register is loaded.
If the esp register still points to an address in the interrupt stack
area (from early system initlization) the ISR might overwrite its own
stack.
Closes#4031
Only include the deferred free support if free() is actually used by the
application.
The free() implementation in RTEMS supports that allocated memory is
freed in interrupt context. Since the heap is protected by a mutex, the
frees issued in interrupt context cannot immediately be freed to the
heap, instead they are placed on a deferred free list. This list is
emptied by the core allocation function
rtems_heap_allocate_aligned_with_boundary(). This adds a dependency to
free() in rtems_heap_allocate_aligned_with_boundary(). In order to
better support applications which only allocate memory and never free
it, this dependency should be avoided. Provide a weak default
implementation of _Malloc_Process_deferred_frees() for
rtems_heap_allocate_aligned_with_boundary(). In the free() module
provide the strong implementation.
Close#4032.
These two tests check conditions which no longer exist in the thread
queue implementation. The are obsolete since the change to use
red-black trees for the priority queues.
Rename internal function T_check_true() to T_check() and use the new
flag T_CHECK_FMT to indicate if a format string is present. This is a
preparation step to make the format string optional.
Make the check context the first parameter.
The API remains the same.
Update #3199.
Rename this header file to later move <t.h> to <rtems/test.h>. The main
feature provided by <rtems/test-info.h> is the output of standard test
information which is consumed by the RTEMS Tester.
Update #3199.
If we interrupt a thread dispatch critical section (thread dispatch
disable level != ISR nest level), then we should not do the fast idle
mode since this may delay an ongoing system call forever.
This adds the possibility to request a retry in the media-listener if an
operation failed. Usefull for example if you want to automatically
reformat a disk if it wasn't possible to mount it.
Remove superflous __atexit stuff in start.o. It is no longer required
by GCC and may cause linker problems.
See GCC commit:
commit 362c63a5e8b5aacfff3e5af0911e42ba7c775042
Author: Geoff Keating <geoffk@cygnus.com>
Date: Fri Apr 14 23:16:25 2000 +0000
GCC 10 no longer passes -many to the assembler. This enables more
checks in the assembler.
The 0 in the tlbie instruction is the L operand which selects a 4KiB
page size.
This introduces the CPU_USE_LIBC_INIT_FINI_ARRAY define for use by CPU
ports to determine which global constructor and destructor methods are
used instead of placing architecture defines where they shouldn't be.
Close#4018
cfg_bsp=$(echo ${bsp_path} | sed -e "s/.*\///" -e 's/\.cfg//')
if test x$bsp = x$cfg_bsp; then
cfg_arch=$(echo ${bsp_path} | sed -e "s/${libbsp_e}*\///" -e 's/\/.*//')
case ${target_arch} in
${cfg_arch}* )
;;
* )
AC_MSG_ERROR([BSP '$bsp' architecture does not match the --target architecture, run 'rtems-bsps' (in the top of the source tree) for a valid BSP list])
;;
esac
found=yes
break
fi
done
if test $found = no; then
AC_MSG_ERROR([BSP '$bsp' not found, run 'rtems-bsps' (in the top of the source tree) for a valid BSP list])
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
#include<bsp.h>
#include<bsp/irq-generic.h>
#include<bsp/start.h>
#ifdef BSP_START_ENABLE_EL3_START_SUPPORT
BSP_START_TEXT_SECTIONvoidbsp_start_hook_0(void)
{
bsp_interrupt_facility_initialize();
}
#endif
BSP_START_TEXT_SECTIONvoidbsp_start_hook_1(void)
{
AArch64_start_set_vector_base();
bsp_start_copy_sections();
versal_setup_mmu_and_cache();
bsp_start_clear_bss();
}
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
