Summary:
When the user sets an illegal signo, the error will only be known when timer_signotify is triggered during timeout.
By intercepting timer_create, we can directly locate the problem
Signed-off-by: chenrun1 <chenrun1@xiaomi.com>
reason:
Replace "inline" with "inline_function" for "up_interrupt_context" to ensure consistency with other arch
Signed-off-by: hujun5 <hujun5@xiaomi.com>
we will drop packet when tcp_close_eventhandler
is register and invoke by tcp_input. then we will always early return and
never stop, the peer will only close the connection if we send reset packet.
precondition:
close -> register tcp_close_eventhandler;
tcp_input -> tcp_callback(TCP_NEWDATA) -> devif_conn_event -> tcp_close_eventhandler
-> flags &= ~TCP_NEWDATA -> NOT entry tcp_data_event -> conn->recvreq NOT increase
old flow:
tcp_input -> tcp->seqno greater than conn->rcvseq -> tcp_send(TCP_ACK)
with this patch:
tcp_input -> tcp->seqno greater than conn->rcvseq -> !TCP_ESTABLISHED
-> case TCP_FIN_WAIT_1 -> dev->d_len greater than 0 -> tcp_reset
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
This primarily fixes allocated memory overrun due to invalidly
calculated limit of the chunk. The function here allocates chunk of size
that includes required header. The error was that size of the chunk was
invalidly again added when limit was being calculated. This was causing
memory overrun and issues especially with object growing (reallocation).
The secondary fix here is to the algorithm that rounds the required size
to the multiple of chunk size. In short chunk size must be reduced by
one to get the correct mask. The condition that was generating the mask
was also invalid because it must perform shift with at most one less
than number of bits (not bytes).
Summary:
add check !defined(__STDC_NO_ATOMICS__)
If the macro constant __STDC_NO_ATOMICS__(C11) is defined by the compiler, the header <stdatomic.h>, the keyword _Atomic, and all of the names listed here are not provided.
refer to:https://en.cppreference.com/w/c/atomic
Signed-off-by: chenrun1 <chenrun1@xiaomi.com>
1. Since the implementation of gcov has changed since clang17, versions before clang17 need to use the libunwind.a file
Signed-off-by: wangmingrong1 <wangmingrong1@xiaomi.com>
currently, nuttx implements readv/writev on the top of read/write.
while it might work for the simplest cases, it's broken by design.
for example, it's impossible to make it work correctly for files
which need to preserve data boundaries without allocating a single
contiguous buffer. (udp socket, some character devices, etc)
this change is a start of the migration to a better design.
that is, implement read/write on the top of readv/writev.
to avoid a single huge change, following things will NOT be done in
this commit:
* fix actual bugs caused by the original readv-based-on-read design.
(cf. https://github.com/apache/nuttx/pull/12674)
* adapt filesystems/drivers to actually benefit from the new interface.
(except a few trivial examples)
* eventually retire the old interface.
* retire read/write syscalls. implement them in libc instead.
* pread/pwrite/preadv/pwritev (except the introduction of struct uio,
which is a preparation to back these variations with the new
interface.)
1. CONFIG_ARCH_COVERAGE has been replaced by CONFIG_SCHED_GCOV
2. Delete the SIM-specific GCOV_ALL configuration and change it to a universal configuration for all architectures
Signed-off-by: wangmingrong1 <wangmingrong1@xiaomi.com>
tmpfile.rename(args.output)
^^^^^^^^^^^^^^
AttributeError: '_io.BufferedRandom' object has no attribute 'rename'. Did you mean: 'name'?
Signed-off-by: anjiahao <anjiahao@xiaomi.com>
Add sem_wait fast operations, use atomic to ensure
atomicity of semcount operations, and do not depend
on critical section.
Test with robot:
before modify:
nxmutex_lock cost: 78 ns
nxmutex_unlock cost: 82 ns
after modify:
nxmutex_lock cost: 28 ns
nxmutex_unlock cost: 14 ns
Signed-off-by: zhangyuan29 <zhangyuan29@xiaomi.com>
If device encryption is not enabled by eFuse, and partiton mark as
encrypted flag, then encrypted MTD is used.
That is no problem in write and read operation, but failed while
using spi_flash_mmap(...) since de-encrypt is not processed.
So, back to use non-encrypted MTD following API Guide:
If flash encryption is not enabled, the flag "encrypted" has no effect
- i.MX93 LPUARTs have 16-byte RX and TX FIFOs. Take those into use and correct some related register definitions
- There is no reason to loop inside interrupt handler, remove the looping
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
- Remove GPIO (SW) based flow control. It didn't work, and pure HW flow control seems to work fine
- Remove some unneeded ifdefs and change bit-field flags to booleans to clean up the code
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
1. cmake uses clang++ as a connector, and does not currently support:
clang++: error: unknown argument: '-wl,--print-memory-usage'
clang++: error: no input files
Signed-off-by: wangmingrong1 <wangmingrong1@xiaomi.com>
adapts to third-party code compilation. in the process of porting EtherCAT,
we encounter some situations where the structure is not defined, or the
returned data types do not match the expectations. Refer to the EtherCAT
implementation of other systems and add relevant definitions.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
supports the usrsock client to subscribe netlink events,
so that it can detect events such as link connection changes.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
- When opening a NFS file in append mode, its file pointer was at offset
0 instead of the end of file.
- When creating a NFS file, the response read pointer wasn't advanced
after reading the attributes_follows bool, which caused the attributes
to be off by 4 bytes. For example, the file size read the GID.
Signed-off-by: Alejandro Aguirre <alejandro.aguirre@midokura.com>
