Do not use pvconn argument to get the TCP connection pointer because pvconn is
normally NULL for some events like NETDEV_DOWN. Instead, the TCP connection pointer
can be reliably obtained from the corresponding TCP socket.
If a existing file is opened with truncate mode e.g. fopen(file, "w+"),
the file size will be incorrect after writing any data to the file.
Before writing to the first sector, the reading is performed again.
As a result, it makes an invalid file size. When a sector buffer is used,
it must also write to the first sector.
Both the snd_ackcb and snd_datacb callbacks were created and destroyed right after sending every packet.
Whenever TCP_REXMIT event occurred due to TCP send timeout, TCP_REXMIT was ignored because
snd_ackcb callback had been destroyed by the time.
The issue is fixed as follows:
- both the snd_ackcb and snd_datacb callbacks are combined into one snd_cb callback
(the same way as in tcp_send_unbuffered.c).
- the snd_cb callback lives until all requested data (via sendfile) is sent,
including all ACKs and possible retransmissions.
As a positive side effect of the code optimization / fix, sendfile TCP payload throughput is increased.
includes following parts:
add support of sparc in arch/Kconfig
add support of sparc in boards/Kconfig
add sparc dir in arch, add sparc dir in boards
add support of sparc in libs/libc/machine
modify all the coding style problem about saprc
Some SoC like bl602 require the exception entry 8 byte align, it should
be safe for other chips so we can apply it globally.
Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
tcp_sendfile() reads data directly from a file and does not use NET_TCP_WRITE_BUFFERS data flow
even if CONFIG_NET_TCP_WRITE_BUFFERS option is enabled.
Despite this, tcp_sendfile relied on NET_TCP_WRITE_BUFFERS specific flow control variables that
were idle during sendfile operation. Thus it was a total inconsistency.
E.g. because of the issue, TCP socket used by sendfile() operation never issued
FIN packet on close() command, and the TCP connection hung up.
As a result of the fix, simultaneously enabled CONFIG_NET_TCP_WRITE_BUFFERS and
CONFIG_NET_SENDFILE options can coexist.
In case of high TX network traffic, netdriver_loop() that reads data from netdev
was invoked via up_idle() only after high TX network traffic had stopped.
That resulted in massive delay and drop of TCP ACK packets and
any other packets from netdev (tun/tap device).
fw_start and fw_size were miscalculated. What was needed
was the pointed values of the offsets __mpfs_nuttx_start
and __mpfs_nuttx_end, not the values they had in place.
Also add the next_arg1 initialization.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
Add the rest of the OpenSBI code to .text.sbi -section. They belong
to there. This frees up some space in the very limited eNVM.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
RAM is expected to start from 0x08000000, not from
0x80000000 in this case. DDR starts from 0x80000000.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
* fsync doesn't modify the file. It doesn't make sense to require
write-access.
* This matches what ~all other systems do.
At least Linux, macOS, and NetBSD.
