If ns_msg->flags == RPMSG_NS_CREATE_ACK, means already know peer's address
so direclty use usr_ept->dest_addr
Signed-off-by: yintao <yintao@xiaomi.com>
fist free when rpmsg_port_unregister at rpmsg spi thread
<free+12>
<rpmsg_router_hub_ept_release+6>
<rpmsg_ept_decref+28>
<rpmsg_unregister_endpoint+120>
<rpmsg_destroy_ept+40>
<rpmsg_router_hub_unbind+22>
<rpmsg_device_destory+172>
<rpmsg_port_unregister+26>
this thread will free r:droid and r:audio at last;
But in this process,
rpmsg_destroy_ept "r:droid" will send NS_destroy to audio;
audio will response NS_destroy to ap; if r:droid has not been
removed from the ept list yet, the rptun_audio thread on ap will
occur as follows:
3 0x103016fa in kasan_check_report
4 0x103018b6 in __asan_store4_noabort
5 0x106ca352 in metal_list_del
6 in rpmsg_unregister_endpoint
7 0x106ca77a in rpmsg_destroy_ept
8 0x102db232 in rpmsg_router_hub_unbind
9 0x106cabb2 in rpmsg_virtio_ns_callback
10 0x106cad5c in rpmsg_virtio_rx_callback
11 0x106cc190 in virtqueue_notification
12 0x106ca064 in rproc_virtio_notified
13 0x106c9ad0 in remoteproc_get_notification
14 0x102dd4ba in rptun_worker at rptun/rptun.c:334
15 rptun_worker (arg=<optimized out>) at rptun/rptun.c:328
16 0x102dd974 in rptun_thread rptun/rptun.c:353
17 0x102cd558 in nxtask_start () at task/task_start.c:1
this will lead to r:audio be freed again.
Signed-off-by: yintao <yintao@xiaomi.com>
Use rpmsg_send_offchannel_raw to fix stuck issue avoid getting stuck
when remote destroy ept.
When send messages from ap to android in Rptun thread, Android may
have sent NS_DESTROY through rpmsg_port and changed ept->dst to
RPMSG_ADDR_ANY in rpmsg_port_ns_callback; So rptun thread may be
stuck at this time because rpmsg_send detected dst_ept's dst_addr
is RPMSG_ADDR_ANY.
rpmsg_virtio_thread(rptun audio):
rpmsg_virtio_rx_callback -> rpmsg_router_cb -> rpmsg_send(dst_ept)
send to android
rpmsg_port_thread (android send NS_DESTROY):
rpmsg_port_ns_callback -> dst_ept->dest_addr = RPMSG_ADDR_ANY
Signed-off-by: yintao <yintao@xiaomi.com>
when edge_ept received NS_DESTROY message to sync the behavior with
other rpmsg transport
if (ns_msg.flags == RPMSG_NS_DESTROY) {
if (_ept)
_ept->dest_addr = RPMSG_ADDR_ANY;
...
Signed-off-by: yintao <yintao@xiaomi.com>
implement the pm feature in rpmsg virtio common part, now the rpmsg
virtio can use in the low power case.
Signed-off-by: Bowen Wang <wangbowen6@xiaomi.com>
Already copy the rpmsg_internal.h from openamp to the nuttx/include,
so do not need add the include patch for .c files.
Signed-off-by: yangao1 <yangao1@xiaomi.com>
Signed-off-by: wangshaoxin <wangshaoxin@xiaomi.com>
This is the virtio-rpmsg and vhost-rpmsg common implementation for
VirtIO-Rpmsg: drivers/virtio/virtio-rpmsg.c
Vhost-Rpmsg: drivers/vhost/vhost-rpmsg.c
Later we will seperate the rpmsg virtio from the rptun framework to
make the rpmsg virtio can work with all the virtio transport layer.
Signed-off-by: Bowen Wang <wangbowen6@xiaomi.com>
Now the rpmsg router transport also support get the signals, and
this signals value is modified by other physical rpmsg transport
(port, virtio) other routed by the rpmsg router.
Signed-off-by: yintao <yintao@xiaomi.com>
1. move drop packets function from port spi to port;
2. add drop packets logic when reconnect to peer;
Signed-off-by: Bowen Wang <wangbowen6@xiaomi.com>
Signed-off-by: liaoao <liaoao@xiaomi.com>
rpmsg_test is a rpmsg api test service inside the kernel, and user
can use ioctl to start this test.
Signed-off-by: mazhuang <mazhuang@xiaomi.com>
Signed-off-by: Bowen Wang <wangbowen6@xiaomi.com>
Signed-off-by: wangshaoxin <wangshaoxin@xiaomi.com>
return -EAGAIN will cause rpmsg_virtio_get_tx_payload_buffer() in
OpenAMP break immediately, but should return to the sleep() logic
to wait buffer is return.
Signed-off-by: Bowen Wang <wangbowen6@xiaomi.com>
Signed-off-by: Yongrong Wang <wangyongrong@xiaomi.com>
If a ns_bind message come at between traverse g_rpmsg_cb and add g_rpmsg_cb list,
the cb->node has not been added to the g_rpmsg_cb list yet.
So rpmsg_ns_bind will not find this callback and not call the cb->ns_bind()
to notify the rpmsg services.
Signed-off-by: Yongrong Wang <wangyongrong@xiaomi.com>
The reconnect will fail because the state changes to connecting when
shutdown cmd is still in process.
And after the shutdown cmd processed, it will reset the state to
unconnected, even though the connect cmd is in ready list,
it will still not be processed.
This patch solve this issue.
Signed-off-by: liaoao <liaoao@xiaomi.com>
Signed-off-by: Bowen Wang <wangbowen6@xiaomi.com>
If connect cmd is dropped during the shutdown cmd process, it will
not be able to reconnect because current side will not try to send
connect cmd after shutdown cmd processed.
So change state to connecting when received connect cmd during the
shutdown cmd process.
Signed-off-by: liaoao <liaoao@xiaomi.com>
Signed-off-by: Bowen Wang <wangbowen6@xiaomi.com>
Now rpmsg port spi can receive peer's suspend and resume command,
suspend command means peer enter into the lowpower state and
resume command means peer waked up
These two commands can let us know peer's running status, so we
will modify the rpmsg signals to let the rpmsg service can get
peer's running state too.
Signed-off-by: liaoao <liaoao@xiaomi.com>
Shutdown mean peer's want to disconnect the connection untial
peer want connect with local side again.
After receive shutdown command, local will clear the state until
reconnecting and clear the gpio status to avoid the leak current.
Signed-off-by: liaoao <liaoao@xiaomi.com>
Because rxq may have received reconnection data, when connect state
changed to unconnected but the reconnect data has not been processed,
and during that time there is a tx buffer to be sent, the
drop_packets may drop the reconnect data and two sides can not
reconnected anymore.
Signed-off-by: liaoao <liaoao@xiaomi.com>
1. add error log when transferring keeps rasing because of no spi
exchange complete callback;
2. this log can also be used to check if there is an interrupt has
been caught when peer side keeps trying to send data;
Signed-off-by: liaoao <liaoao@xiaomi.com>
0x106b9ad2 in metal_list_del (node=<optimized out>) at /home/yintao/O61/nuttx/include/metal/list.h:85
rpmsg_unregister_endpoint (ept=0x1857ba30) at open-amp/lib/rpmsg/rpmsg.c:296
rpmsg_destroy_ept (ept=0x1857ba30) at open-amp/lib/rpmsg/rpmsg.c:389
0x102b871a in rpmsg_router_hub_unbind (ept=0x1857b330) at rpmsg/rpmsg_router_hub.c:139
0x102b7fec in rpmsg_device_destory (rpmsg=rpmsg@entry=0x1859e4e8) at rpmsg/rpmsg.c:486
0x102b9928 in rpmsg_port_unregister (port=port@entry=0x1859e4e8) at rpmsg/rpmsg_port.c:749
0x102b9bca in rpmsg_port_spi_process_packet (rxhdr=0x185ba718, rpspi=<optimized out>) at rpmsg/rpmsg_port_spi.c:408
rpmsg_port_spi_thread (argc=<optimized out>, argv=<optimized out>) at rpmsg/rpmsg_port_spi.c:452
0x102aa96e in nxtask_start () at task/task_start.c:111
Signed-off-by: yintao <yintao@xiaomi.com>
And set signal_running when rpmsg devices registered
Now rpmsg port transport support to get the remote core's state
by api rpmsg_get_signals()
Signed-off-by: liaoao <liaoao@xiaomi.com>
Signed-off-by: Bowen Wang <wangbowen6@xiaomi.com>
provide get_signals ops to the transport layer implementation.
add rpmsg_get_signals API for rpmsg service to get remote core's
status.
Signed-off-by: Yongrong Wang <wangyongrong@xiaomi.com>
Signed-off-by: cuiziwei <cuiziwei@xiaomi.com>
Signed-off-by: Bowen Wang <wangbowen6@xiaomi.com>
Before this patch, the rpmsg services will not be destroied after
call the rpmsg_device_destory(). And will cause error after
call rpmsg_device_created() at the secound connecting witt peer.
Signed-off-by: liaoao <liaoao@xiaomi.com>
Signed-off-by: Bowen Wang <wangbowen6@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>
If SPI dependency is not set, the following warning will be generated during compilation:
[109/1450] Building C object drivers/CMakeFiles/drivers.dir/rpmsg/rpmsg_port_spi.c.o
/data/code/nuttxspace/nuttx/drivers/rpmsg/rpmsg_port_spi.c: In function ‘rpmsg_port_spi_exchange’:
/data/code/nuttxspace/nuttx/drivers/rpmsg/rpmsg_port_spi.c:233:3: warning: implicit declaration of function ‘SPI_EXCHANGE’ [-Wimplicit-function-declaration]
233 | SPI_EXCHANGE(rpspi->spi, txhdr, rpspi->rxhdr,
| ^~~~~~~~~~~~
[1450/1450] Pac SIM with dynamic libs in nuttx.tgz
Signed-off-by: lijing.ly <lijing.ly@bytedance.com>
In file included from nuttx/drivers/rpmsg/rpmsg_port_uart.c:27:
nuttx/drivers/rpmsg/rpmsg_port_uart.c: In function ‘rpmsg_port_uart_send_connect_req’:
nuttx/drivers/rpmsg/rpmsg_port_uart.c:219:16: warning: format ‘%d’ expects argument of type ‘int’, but argument 3 has type ‘ssize_t’ {aka ‘long int’} [-Wformat=]
219 | rpmsgerr("Send connect request failed, ret=%d\n", ret);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~
| |
| ssize_t {aka long int}
nuttx/drivers/rpmsg/rpmsg_port_uart.c:219:51: note: format string is defined here
219 | rpmsgerr("Send connect request failed, ret=%d\n", ret);
| ~^
| |
| int
| %ld
nuttx/drivers/rpmsg/rpmsg_port_uart.c: In function ‘rpmsg_port_uart_send_connect_ack’:
nuttx/drivers/rpmsg/rpmsg_port_uart.c:235:16: warning: format ‘%d’ expects argument of type ‘int’, but argument 3 has type ‘ssize_t’ {aka ‘long int’} [-Wformat=]
235 | rpmsgerr("Send connect ack failed, ret=%d\n", ret);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~
| |
| ssize_t {aka long int}
nuttx/drivers/rpmsg/rpmsg_port_uart.c:235:47: note: format string is defined here
235 | rpmsgerr("Send connect ack failed, ret=%d\n", ret);
| ~^
| |
| int
| %ld
Signed-off-by: chao an <anchao.archer@bytedance.com>
Because rpmsg_virtio_lite is a better name to inform user that the
rpmsg virtio (original name) is a lite implementation of rpmsg virtio.
Signed-off-by: Bowen Wang <wangbowen6@xiaomi.com>
Modify the kernel to use only atomic_xx and atomic64_xx interfaces,
avoiding the use of sizeof or typeof to determine the type of
atomic operations, thereby simplifying the kernel's atomic
interface operations.
Signed-off-by: zhangyuan29 <zhangyuan29@xiaomi.com>
