Historically, the loopback driver used the largest packet size of all enabled link layer protocols. This permitted packets to be forward via the loopbak device with no major loss of performance. However, in experimenting with configurations where no other link layer protocols were enabled, this means the loopback packet size was set to the smallest possible size, to the SLIP minimum of 296 bytes. This resulted in terrible loopback performance.
This commit adds an option to increase the loopback packet size with the option CONFIG_NET_LOOPBACK_PACKETSIZE.
The loopback driver packet buffer should be quite large. The larger the loopback packet buffer, the better will be TCP performance of the loopback transfers. The Linux loopback device historically used packet buffers of size 16Kb, but that was increased in recent Linux versions to 64Kb. Those sizes may be excessive for resource constrained MCUs, however.
The network still enforces the lower limit that is the maximum packet size of all enabled link layer protocols. But this new option permits the loopback packet size to be increased from that.
* net/Kconfig: Adds CONFIG_NET_LOOPBACK_PKTSIZE option
* include/nuttx/net/netconfig.h: Assures that the packet size that is used is at least as large as the largest packet size of other link layer protocols.
* drivers/net/loopback.c: Use that larger packet size.
* boards/sim/sim/sim/configs/tcploop/defconfig: Set the loopback packet size to 1500
Trigger SGA and ECHO proactively in character mode otherwise Ubuntu bultin telnet can't enter this mode
Change-Id: I8aa2ab2b31c35007077c701c264b3971152435f0
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Remove g_telnet_common global variable; we can reuse g_clients_sem as the lock guard
Change-Id: Ic3af9f2116f70523a4249b29c65bd1fb83ca4da2
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Telnet driver should return -EAGAIN is O_NONBLOCK is active also should report -EPIPE first
Change-Id: I7ad2df15377c7bec8e22d0f5d1b54f7ce33eb0db
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Refine Ctrl-C handling in telnet driver to avoid issue the kill more than once
Change-Id: I9fcec5d861ea85258170f379d741d2bb8e4d9b9e
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Splict common_ioctl into telnet_ioctl and factoral_ioctl. Rmove the wrong telnet_poll from g_factory_fops
Change-Id: I39f278763ff279d464c5be6728b9936c6cab16eb
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
* S32K add support for Nxp drone boards
* Update arch/arm/src/s32k1xx/hardware/s32k1xx_rtc.h codestyle
Co-Authored-By: David Sidrane <David.Sidrane@Nscdg.com>
Co-authored-by: Jari van Ewijk <jari.vanewijk@nxp.com>
Co-authored-by: David Sidrane <David.Sidrane@Nscdg.com>
* Remove the code duplication in tun_net_receive_tap and remove the unused filep field
* Shouldn't return -EBUSY in tun_write. Let the caller wait until the write buffer free
* Handle that write buffer is ready first correctly in tun_read
* Remove the unused tun_ipv6multicast
Gregory Nutt <gnutt@nuttx.org>
Run all .c and .h files modified in this PR through nxstyle and correct all coding standard problems.
Xiang Xiao <xiaoxiang@xiaomi.com>
Remove TIME_EXTENDED option to more conform C standard
Note: the code/data size increment is small
to fix the follow error:
Refresh skp16c26/ostest
LN: include/arch/board to /workspace/mirtos/nuttx/boards/renesas/m32262f8/skp16c26/include
LN: include/arch/chip to arch/renesas/include/m32262f8
Author: Gregory Nutt <gnutt@nuttx.org>
Run all .c and .h files in last PR through tools/nxstyle and correct all coding standard complaints.
Author: Oleg <ev.mipt@gmail.com>
* Fix CAN driver to work with internal OS interfaces. Store internal file pointer in f_priv and use it then to distinct readers (See issue #111)
* Store reader pointer in f_priv instead of filep
* Don't need in filep in can_reader_s
Author: Alan Carvalho de Assis <acassis@gmail.com>
Run nxstyle on .c and .h files and fix it
Author: Alin Jerpelea <alin.jerpelea@sony.com>
drivers: usbdev: minor fix
drivers: usbdev: usbmsc full speed not available
Change transfer size to be based on maxpacket size.
drivers: usbdev: Fix string ID calculation
For *_STRBASE defines, it already unnecessary because composite device setup
has been changed, it would be calculated by *_composite.c in board sources.
drivers: usbdev: Fix invalid/unsupported command processing
Mass Storage Class shall stall when invalid or unsupported commands
has been recieved.
drivers: usbdev: Remove unnecessary reset logic
drivers: usbdev: Flags comparison fix
drivers: usbdev: Descriptor type mismatch fix when dual speed is enabled
Author: Alan Carvalho de Assis <acassis@gmail.com>
Run nxstyle against .c .h files and fix it
Author: Adam Feuer <adam@starcat.io>
Summary
Adds CDC ECM Ethernet over USB High Speed for SAMA5D36-Xplained
(and maybe other boards) (most of the code was there already,
but didn't work out of the box for the SAMA5D36-Xplained)
Only SAMA5D36-Xplained has been tested so far
Impact
None if you don't use the CDC ECM Ethernet driver
On SAMA5D36-Xplained, this adds high-speed Internet connectivity
over USB 2.0 High Speed. via the USB CDC ECM Gadget driver.
It may work on other boards too.
This also fixed full-speed (low-speed) mode for the board too.
Limitations
Hasn't been tested on anything other than SAMA5D36-Xplained board.
TODO
Ideally this would include a composite RNDIS device so it would
also work seamlessly on Windows. That is for a future PR
Ideally this would include software to help configuration via
mDNS/DNS-SD for plug and play compatibility with Linux and macOS.
That is for a future PR.
Detail
Only a few lines of C driver code needed to be changed, since the
capability was there already. The rest is config and documentation.
Changes the SAMA5D3-Xplained board bringup to match the SAMA5D3-EK
board bringup
A helper script to configure Linux routing and iptables NAT is also
provided, along with documentation on how to use it.
Testing
Manual, on a Ubuntu Linux 19.10 system and MacOS 10.14.6 Mojave
MacBook Pro.
How To Verify
Follow the new CDC ECM Ethernet over USB instructions in the board
README.txt file
Commits:
remove non-UTF-8 chars in comment and reformat
removed unneeded comment markers
instructions for using the defconfigs
removed EMAC from config
- to prove this example only needs the CDC ECM Ethernet over USB to work
added CDC-ECM Ethernet over USB info to README
added U-Boot image
added netusb helper script
- this can configure the Linux network interface and routes
so you can ping or access the NuttX system via TCP/IP.
renamed defconfig dirs to be ethernet-over-usb
- was usb-over-ethernet which is not right
added USB DMA to defconfigs
updated readme with autoboot and debugging info
bringing ethernet-over-usb examples into parity
added cdc ecm ethernet over usb with telnetd config
added defconfig
only use phy interrupt if netdevices is ethernet
- because now netdevice could be CDC ECM ethernet over usb
which has no PHY interrupt
add bringup to Makefile
add bringup
app init cleanup
init cdc ecm driver and rndis driver; some cleanup
fixed some typos and odd characters
usb over ethernet working over usb 2.0 hs
Author: Alan Carvalho de Assis <acassis@gmail.com>
Run nxstyle on all .c and .h and fix all issues
Author: Alin Jerpelea <alin.jerpelea@sony.com>
drivers: mtd: smart: Add smartfs fsck feature
Support fsck to check and repair the smartfs file system. If the power
loss occurs during writing into the flash, the dead space are created
in flash after the next power cycle. To avoid this problem, introduce
fsck and keep the consistency of file system in initializing smartfs.
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
commit 60927c5fb6a353994341601f55071c618e97682b
Author: Alin Jerpelea <alin.jerpelea@sony.com>
Date: Thu Dec 28 18:27:21 2017 +0900
drivers: mtd: smart: Fix duplicate sector selection in SmartFS
Add care for 16-bit sequence without CRC.
drivers: mtd: smart: Check CRC of duplicate sectors
In the illegal case by power-loss, when the multiple logical sectors are
duplicated, we compare the sequence number of each sector and select the newer
sector. Just in case, add CRC check for the newer sector. If the newer sector
has CRC error, then we use the older sector.
drivers: mtd: smart: SPI-Flash recovery from the initial error state
The FLASH may be not erased in the initial delivery state.
Just in case for the recovery of this fatal situation,
after once erasing the sector, return the sector as a free sector.
drivers: mtd: smart: Fix error handling in smartfs mtd driver
Add error handling in relocate sector.
drivers: mtd: smart: Fix initialize sector sequence value in smartfs
Fix initialization of sequence value into sector header and also avoid
unaligned memory access when CONFIG_MTD_SMART_ENABLE_CRC=n.
drivers: mtd: smart: Fix handling of duplicate sector in smartfs
In smartfs scan, if duplicate logical sector is found on the device,
then smartfs selects the winner sector by comparing sequence number,
and the loser sector is soon released. Fix a bug this loser sector
is registered into logical-to-physical sector mapping table.
Author: Alan Carvalho de Assis <acassis@gmail.com>
Run nxstyle to check .c and .h files and fix reported issues
Author: liuzhao <happypapa@yeah.net>
Add Quectel EC20 4G LTE Module USB CDC/ACM support
Gregory Nutt