We needed to add support for the RV32M1_LPTMR_TIMER to the test so its
knows what the IRQ of the timer is.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Apply a similar fix to commit 0289a410ba
to the openisa_rv32m1 linker script for handling .srodata sections.
Otherwise we get oprhan read-only data section. Fix this by adding
the .srodata section to the RISC-V linker script.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
This turns on the linker flag to sort sections by alignment
in decreasing size of symbols. This helps to minimize
padding between symbols.
This also adds the linker options to sort common symbols by
alignment in descending to minimize the need for padding.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
During testing with sorting section by alignment with qemu_nios2,
if rodata section is not aligned on 4-byte boundary and its size
not of multiple of 4, it would never boot correctly. So align
the rodata here. This is in preparation to enable the linker
option to sort sections by alignment.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The shared page is inherently used in multiprocessor contexts where
the compiler optimizer can trip us up (specifically, a spin on
num_active_pus was being hoisted out of the loop on some gcc's). Put
the volatile declartion into the struct pointer itself instead of
relying on the code to get it right.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
We can turn this off in application code, but the default should be
SMP.
Oddball note: the SMP setting is in the soc layer and not arch, even
though I think it would be better there. We don't actually have a
per-arch defconfig file included by kconfig.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The various tests would all do a "wait for threads to exit" step
before checking the results, but this was implemented with a simple
busy wait that turns out to need careful tuning (because there was
busy waiting in the threads).
Rather than try to synchronize this, white box the issue (it's a low
level SMP test, after all) by spinning on the thread states directly
watching for the kernel to flag them dead. The downside here is that
if the process fails for some reason we'll get a hang and a timeout
reported from sanitycheck and not a synchronous ztest assertion. But
in return, successful tests run much faster and I don't need to worry
about how to tune them for IPI latency on different platforms.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This case was predicated on a mistake. The behavior of k_wakeup() has
always been NOT to wake up threads that are "pending" on a wait queue,
only ones blocked on a timeout in k_sleep(). As written, this test
case could never pass.
(Really there's no good reason for that. It seems reasonable to me to
expect wakeup to work symmetrically, and the docs are sort of
ambiguous on the subject. But the code in k_wakeup() is clear:
threads flagged pending get an early exit and the call becomes a
noop.)
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This was always doing a remove/add of the _current thread to the run
queue, which is wrong because in SMP _current isn't in the queue to
remove. But it went undetected until the recent dlist changes.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
In SMP, we are setting the _current pointer while holding the
scheduler spinlock locally, which means that when we try to release it
the validation layer (not the spinlock per se) will scream at us
because the thread that took the lock doesn't match the one releasing
it.
Special case this when validation is enabled.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The k_spin_lock() validation was setting the new owner of the spinlock
BEFORE the actual lock was taken, so it could race against other
processors trying the same thing. Split the modification step out
into a separate function that can be called after we affirmatively
have the lock.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
There was a test-created thread that wasn't including this. It's a
huge stack and doesn't overflow (though I thought briefly that it
was), but it's a rule that we need to have that buffer and I'm trying
to fix these as I find them.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The tracing fixes in commit e87193896a ("subsys: debug: tracing: Fix
thread tracing") were... not a readability win. The point appears to
have been to put a tracing hook immediately before and after the
assignment to the _current pointer. So do that in an abstracted
function and clean up _get_next_switch_handle() (which is a subtle and
important function already polluted with some unavoidable preprocessor
testing!)
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Currently thread abort doesn't work if a thread is currently scheduled
on a different CPU, because we have no way of delivering an interrupt
to the other CPU to force the issue. This patch adds a simple
framework for an architecture to provide such an IPI, implements it
for x86_64, and uses it to implement a spin loop in abort for the case
where a thread is currently scheduled elsewhere.
On SMP architectures (xtensa) where no such IPI is implemented, we
fall back to waiting on an arbitrary interrupt to occur. This "works"
for typical code (and all current tests), but of course it cannot be
guaranteed on such an architecture that k_thread_abort() will return
in finite time (e.g. the other thread on the other CPU might have
taken a spinlock and entered an infinite loop, so it will never
receive an interrupt to terminate itself)!
On non-SMP architectures this patch changes no code paths at all.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Tickless timers mean that k_busy_wait() won't work until after the
timer driver is initialized, which is very early but not as early as
SMP. No need for it, just spin.
Also the original code used a stack variable for the start flag, which
racily presumed that _arch_start_cpu() would comes back synchronously
with the other CPU fired up and running right now. The cleaned up smp
bringup API on x86_64 isn't so perky, so it exposed the bug. The flag
just needs to be static.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Before we're initialized and can use proper synchronization, the CPU
initialization path spins on the thread entry function to be non-null.
But the data wasn't tagged volatile, and with gcc 8.2.1 (but not
6.2.0) the optimizer was hoisting the reads to SMP init would spin
forever.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
In SMP, _current is not "queued". (The run queue only stores
unscheduled threads because we can't rely on the head of the list
being _current). We weren't updating the cache choice, which would
flag swap_ok, so calling k_thread_abort(_current) (for example, when a
thread exits from its entry function) would try to switch back into
the thread and then run off the end of the function.
Amusingly this was more benign than you'd think. Stumbled on it by
accident.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
When CONFIG_SMP is enabled but CONFIG_MP_NUM_CPUS is 1 (which is a
legal configuration, though a weird one) this static function ends up
being defined but unused, producing a compiler warning.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The previous scheme where the xuk layer would call out to the to
"fetch" the stack for a SMP CPU at startup was sorta weird, and an
impedance mismatch with Zephyr which has a "start this CPU" call
instead. It also got broken when x86_64 started launching CPUs
(correctly) on their interrupt stacks instead of a temporary area;
they weren't ready yet when xuk initialization was happening and the
system would deadlock waiting for code that can't run yet to provide a
stack.
Note that this preserves the somewhat quirky behavior that Zephyr's
CPU numbering is just the order in which the SMP CPUs emerge from
initialization and not a hardware ID.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
No need to do this in arch code, all qemu variants speak the same
command line and the kconfig values we need are easily available.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This is intended to initialize CPU-local timer devices, but HPET is
global so we have nothing to do.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
When this code was written, there was no "stack frame" struct defined.
There is now, so use that for clarity and concision. Also remove an
obvious comment (I mean, duh, we can put any segment selectors in
those fields we want).
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
If CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT is enabled,
the app shared memory partition may cause waste of memory
due to the need for padding.
For example, tests/subsys/jwt and board mps2_an385:
z_test_mem_partition: addr 0x20000000, size 52
z_libc_partition : addr 0x20000040, size 4
k_mbedtls_partition : addr 0x20008000, size 32736
ending at 0x2000ffff, taking up 65536 bytes
With power-of-two size and alignment requirement,
k_mbedtls_partition takes up 32KB memory and needs to be
aligned on 32KB boundary. If the above partitions are
ordered as shown, there needs to be a lot of padding
after z_libc_partition before k_mbedtls_partition can
start. In order to minimize padding, these partitions
need to be sort by size in descending order.
After the changes here, the partitions are:
k_mbedtls_partition : addr 0x20000000, size 32736
z_test_mem_partition: addr 0x20008000, size 52
z_libc_partition : addr 0x20008040, size 4
ending at 0x2000805f, taking up 32864 bytes
With the above example, sorting results in a saving
of 32672 bytes of saving.
Fixes#14121
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The app_smem.ld is also being used by architectures other than ARM.
So move the linker script out of include/arch/arm and into
include/linker.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
mqtt_client_tls_connect() calls setsockopt() to set the TLS_PEER_VERIFY
option to require verfication. To enable mqtt, we need to return
success at a minimum when this option is set to 2.
Signed-off-by: Vincent Wan <vincent.wan@linaro.org>
The driver is not returning correct error codes when error occurs.
The error handling function is expecting negative input values, but
that is not true for BSD error codes. So I am taking an approach
where I use a function to convert SimpleLink error codes to BSD
error codes, and call slcb_setErrno() to set the errno independently.
Fixes#12745
Signed-off-by: Vincent Wan <vincent.wan@linaro.org>
Appears within an 'if MBEDTLS'.
'if FOO' is just shorthand for adding 'depends on FOO' to each item
within the 'if'.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Add CONFIG_SYS_CLOCK_TICKS_PER_SEC=1000 to nucleo_l053r8_defconfig.
All STM32 boards seem to have this defined.
Makes samples with ksleep work properly.
Signed-off-by: Anthony Kreft <anthony.kreft@gmail.com>
Fix an out of bound access in the native_posix uart driver,
when generating the string to autoattach a terminal to the UART.
(The space for the null termination was missing)
Fixes: #14401
Coverity issue CID: 195855
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
fcntl() may (very unlikely) fail when setting the new pseudoterminal
to non-blocking.
Let's check for this condition, handle it, and in the process
silence a coverity issue.
Fixes#14396
Coverity CID: 195872
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
When user disconnects the USB cable, peripheral should be
immediately disabled, howewer a delay may occur when driver
events are processed from a workqueue or higher-priority
thread/ISR is active. This may lead to a fake resume/reset
event (peripheral-specific behavior). This fix drops such
events when cable is detached.
Fixes#13822
Signed-off-by: Paweł Zadrożniak <pawel.zadrozniak@nordicsemi.no>
This adds BT_GATT_DISCOVER_ATTRIBUTE which can be used to discover any
type of attribute in a given range.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Those are not considered Descriptors and shall be discovered with
use of BT_GATT_DISCOVER_CHARACTERISTIC.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
When discovering descriptor Find Information procedure is used which
does not allow any filtering by the server so it will return all
attributes in the given range including services and characteristics.
Fixes#14265
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Ensure that no I2C start condition is inserted between messages of
a same transfer, except if explicitely requested (restart flag).
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Ensure that no I2C start condition is inserted between messages of
a same transfer, except if explicitely requested (restart flag).
This fix issue with I2C register write functions.
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Out-of-tree Kconfig frontends need to know which environment variables
are exported by Zephyr to Kconfig.
To support this we allow additional Kconfig python frontend targets to
be specified from out-of-tree.
This support was added on request by a third-party with a custom
Kconfig frontend.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Since NET_L2_BT no longer "select"s BT and the rest of required BT
features, enable them within the defconfig file.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Maybe this is some "just in case" thing that got copied around. There's
no need to have a blank line at the beginning or end of Kconfig files.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
When we build with newlib we don't set -nostdinc. In that case make
sure that we leave it to the toolchain to set the system include paths.
The one exception to leaving to the toolchain to set the system include
paths is the path to the newlib headers. Since we build
with -ffreestanding we need to make sure the newlib header path is the
before the toolchain headers. Otherwise the toolchain's 'freestanding'
headers get picked up and that causes issues (for example getting PRI*64
defined properly from inttypes.h due to __STDC_HOSTED__ being '0').
For newlib we accomplish this by having the only system header specified
by zephyr_system_include_directories() being just the newlib headers.
Note: for minlibc we leave things alone as things just happen to work as
the -I include of the libc headers takes precedence over -isystem so we
get the libc headers over the toolchain ones. For the newlib case it
appears that setting both -I and -isystem for the same dir causes the
-I to be ignored.
Fixes#14310
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Configure lvgl defaults for imx rt boards in their respective board
defconfigs rather than the lvgl sample application.
Signed-off-by: Maureen Helm <maureen.helm@nxp.com>
Removing the TX timeout handling in the GMAC driver (commit 18b07e09e0)
revealed some issues with the way hardware priority queues work.
For cases with both hardware priority queues enabled, with the default
recommended delta bandwidths (0% - 75%), the lower priority queue (0%)
is hardly able to send any packets. This became visible, because without
the timeout mechanism, we quickly ran out of available TX buffers if
there were multiple packets being queued to the queue.
Here is an excerpt from 802.1Q, chapter 34.3.1 which describes how Qav
bandwidth sharing SHOULD work:
The deltaBandwidth(N) for a given N, plus the deltaBandwidth(N) values
for any higher priority queues (larger values of N) defines the total
percentage of the Port’s bandwidth that can be reserved for that queue
and all higher priority queues. For the highest priority queue, this
means that the maximum value of operIdleSlope(N) is deltaBandwidth(N)%
of portTransmitRate. However, if operIdleSlope(N) is actually less
than this maximum value, any lower priority queue that supports the
credit-based shaper algorithm can make use of the reservable bandwidth
that is unused by the higher priority queue. So, for queue N-1, the
maximum value of (operIdleSlope(N) + operIdleSlope(N-1)) is
(deltaBandwidth(N) + deltaBandwidth(N1))% of portTransmitRate.
However in reality, the lower priority queues (N-1) on the SAM GMAC
hardware DO NOT use the bandwidth available from the higher priority
queues (N).
This commits fixes the issue by changing the defaults. These are still
set to the recommended 75% (total), but this percentage is split between
the priority queues manually.
Signed-off-by: Tomasz Gorochowik <tgorochowik@antmicro.com>
