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#4220
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().
Close#4190.
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#4172
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#4114.
Add a workaround for Cortex-A9 Errata 845369: Under Very Rare Timing
Circumstances Transition into Streaming Mode Might Create Data Corruption.
Update #4114.
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#4096.
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
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#4038.
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#4041.
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#4033.
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#4030
Some objects can be created with a local or global scope in a
multiprocessing network. In non-multiprocessing configurations setting
the scope to local or global had no effect since such a system can be
viewed as a multiprocessing network with just one node. One and all
nodes is the same in such a network. However, if multiprocessing was
configured, creation of a global object in a single node network
resulted in an RTEMS_MP_NOT_CONFIGURED error. Remove this error
condition for symmetry to the non-multiprocessing setup. This is in line
with the task affinity behaviour in SMP systems.
Update #4005.
- Fixes timeout for smpipi01 where:
+ Main thread sends perform jobs to worker cpu while it is already
performing jobs
+ Interrupt on worker cpu performs jobs, but with empty job list
+ Worker cpu continues to execut previous job and adds new job list
to itself, which is never performed, since the interrupt has already
been handled
+ Main thread blocks forever on barrier D
The signal handler of the consumer might start executing
before rtems_signal_send of the producer returns.
Therefore change the state to SIG_1_SENT before sending the signal.
- Do not forward Clock_isr through Clock_driver_support_at_tick as this
will cause every processor to send IPIs with Clock_isr therby creating
an infinie loop
- Instead the processor handling the clock interrupt causes all other
processors to call rtems_timecounter_tick to update their tick count
Create a GS segment in the GDT for each processor for storing TLS.
This makes the GDT in startAP.S obsolete as all processors now share the
same GDT, which is passed to each AP at startup.
The correct segment for each processor is calculated in cpu_asm.S.
Update #3335
start16.S is now only used for SMP configurations to start the
application processors.
This commit removes all unnecessary parts for this job,
i.e. video conssole initalisation, A20 gate activation
and all non-AP related code.
Update #3335
added tests for fesetexeptflag(), fegetexeptflag(),
fegetround(), fesetround().
In the test fegetround() does not return any flag
other then FE_TONEAREST in tests.
This is probably due to soft float.
The test complies successfully and returns assert
at fegetround()
Other tests run without any errors
tested on RISCV/rv32imac
The test prints nothing if runs successfully.
updates #2971
Signed-off-by: Eshan dhawan <eshandhawan51@gmail.com>
Merge the rtems_termios_isig_status_code and
rtems_termios_iproc_status_code enums into a single
rtems_termios_iproc_status_code which is now a part of the API.
Simplify rtems_termios_posix_isig_handler() to avoid unreachable code.
Close#3800.
