Add WLAN_STA_STATE_DISCONNECTED state to prevent race conditions or invalid state handling during disconnect.
Signed-off-by: liqinhui <liqinhui@xiaomi.com>
modify the code of the adapted protocol stack to avoid deadlocks and the
logic that cannot be protected by locks after modification.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
avoid packet processing delays caused by task switching,
to support those applications that are extremely time-sensitive.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
The loopback network driver sending task is placed in the LPWORK queue, which may result in delayed scheduling when the CPU usage reaches 100%, leading to the timely sending of cached messages from the loopback network driver and causing issues with post waitsem being too late.we decied to set loopback sending task to the HPWORK queue.
Signed-off-by: wangchen <wangchen41@xiaomi.com>
avoid packet processing delays caused by task switching,
to support those applications that are extremely time-sensitive.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
Upperhalf supports multiple working modes at the same time,
which is specified by NIC when register
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
Remove the previously supported polling mode. If necessary in the
future, switch to a new solution for implementation.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
avoid illegal address access caused by removing vlan
network cards during the traversal of vlan network cards
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
1. We add default PCP because some of our apps may not want to set PCP
manually (e.g. Our user may just ping with pre-set PCP)
2. The `vlan_qos` is used as PCP when setting Linux's priority mapping:
https://github.com/torvalds/linux/blob/v6.12/net/8021q/vlan.c#L590
Although `vlan_qos` is not used when creating VLAN on Linux, we can
use it as PCP on creating VLAN (without changing its meaning), and
keep compatible with Linux's creating (Exactly the same when
`vlan_qos` is 0).
Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
Inspired by Linux's way, we also create VLAN devices for managing VLAN,
which will become interfaces like `eth0.58`.
QinQ is also supported, we can create VLAN devices above another VLAN
devices, like `eth0.100.101` (or even `eth0.1.2.3.4`, also supported on
Linux).
Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
after the network card is turned off, read and write operations need
to return -ENETDOWN, and it is necessary to wake up the waiting read
and write operations.
Signed-off-by: wangchen <wangchen41@xiaomi.com>
Nuttx currently has 2 types of sleep interfaces:
1. Signal-scheduled sleep: nxsig_sleep() / nxsig_usleep() / nxsig_nanosleep()
Weaknesses:
a. Signal-dependent: The signal-scheduled sleep method is bound to the signal framework, while some driver sleep operations do not depend on signals.
b. Timespec conversion: Signal-scheduled sleep involves timespec conversion, which has a significant impact on performance.
2. Busy sleep: up_mdelay() / up_udelay()
Weaknesses:
a. Does not actively trigger scheduling, occupy the CPU loading.
3. New interfaces: Scheduled sleep: nxsched_sleep() / nxsched_usleep() / nxsched_msleep() / nxsched_ticksleep()
Strengths:
a. Does not depend on the signal framework.
b. Tick-based, without additional computational overhead.
Currently, the Nuttx driver framework extensively uses nxsig_* interfaces. However, the driver does not need to rely on signals or timespec conversion.
Therefore, a new set of APIs is added to reduce dependencies on other modules.
(This PR also aims to make signals optional, further reducing the code size of Nuttx.)
Signed-off-by: chao an <anchao.archer@bytedance.com>
Currently the mdio communication is part of the monolithic 'full netdevs'.
This commit serves as an way to add modularity for the netdevs drivers.
A new upperhalf/lowerhalf mdio device comes with this commit that manages the data transfer ofer mdio interface.
Signed-off-by: Luchian Mihai <luchiann.mihai@gmail.com>
Add driver for the LAN865x 10BASE-T1S SPI MAC-PHY.
The driver is a lower-half driver to the OA-TC6 base driver.
Signed-off-by: michal matias <mich4l.matias@gmail.com>
Add driver for the Onsemi NCV7410 10BASE-T1S SPI MAC-PHY.
The driver also works for the NCN26010, which is an identical chip.
The driver is a lower-half driver to the OA-TC6 base driver.
Signed-off-by: michal matias <mich4l.matias@gmail.com>
Add base driver common for OPEN Alliance 10BASE-T1x MAC-PHY Serial Interface (OA-TC6)
protocol SPI MAC-PHYs.
Signed-off-by: michal matias <mich4l.matias@gmail.com>
Instead of verifying MMD indirect access by checking
EEE ADVERTISEMENT register has value 0x6 (which may be changed),
the verification is done by write and check LED MODE register.
1. Backup current LED_MODE register value
2. Write value Single-LED mode (0x10) to LED_MODE
3. Read LED_MODE register and verify it returns value 0x10
4. Write original (backup) value to LED_MODE
From the ksz9477 errata, setup the configuration which is tested to work.
- Improve PHY receive performance
- Disable EEE for ports 1-5
- Fix supply current values
Signed-off-by: Jani Paalijarvi <jani.paalijarvi@unikie.com>
This commit removes existing support for the NCV7410 10BASE-T1S MAC-PHY.
The driver will be replaced by its generalized version with different organization.
The new driver has naming clashes with the old, therefore the old has to be removed first.
Signed-off-by: michal matias <mich4l.matias@gmail.com>
A descriptor ring length must be aligned to 128, one descriptor is 16B length,
so we can provide this condition with the appropriate number of descriptors
Signed-off-by: p-szafonimateusz <p-szafonimateusz@xiaomi.com>
Descriptor head is managed by HW and should not be modified by SW unless:
1. device is after a reset
2. device is before enabling TX or RX
Also set correct tail for RX which should point at the end of ring
(descriptor ring is zero-indexed).
Signed-off-by: p-szafonimateusz <p-szafonimateusz@xiaomi.com>
Do not reset RX/TX rings when link status change.
This can break internal card state which is impossible to recovery.
Signed-off-by: p-szafonimateusz <p-szafonimateusz@xiaomi.com>
reset RX/TX rings **after** RX/TX are disabled to make sure the rings are
not used by hardware when software modify rings state.
Signed-off-by: p-szafonimateusz <p-szafonimateusz@xiaomi.com>
