Constifying global data allows to save lots of RAM. Defining data inside
function as 'static const' allows on the other hand to save stack usage
and reduced code size (because data doesn't have to be copied on stack
in runtime).
This improvement allows to save 640 bytes of RAM and 64 bytes of ROM
when compiled on ARM.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Configure hostname by issuing AT+CWHOSTNAME="<hostname>" command. Do it
just after setting link address, which is used to generate hostname
postfix when CONFIG_NET_HOSTNAME_UNIQUE=y.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
- Add ESP DHCP Support
- Add ESP Static IP Support including
KConfig entries for configuring IP,Gateway and netMask
Signed-off-by: Mohamed ElShahawi <ExtremeGTX@hotmail.com>
So far there was 100ms timeout on allocation of RX net_pkt. This is too
little for cases when lots of data are incoming on pretty
fast (e.g. 1Mbps) UART interface and application layer does not consume
received network packets fast enough. Up to now in such cases all
incoming data was processed from UART and there was no data loss when
utilizing hardware flow control. However there was high chance that data
processed from UART could not be passed further to network stack,
because of the 100ms net_pkt allocation timeout. This happens for
example when low priority application does not have enough time to run.
Increase default RX net_pkt allocation timeout from 100ms to 5s, so
there is much more time to process network packets. Such timeout should
not harm, because only a dedicated RX thread will be suspended for that
time, resulting in suspending UART traffic if hardware flow control is
supported.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
There is now a hardcoded 100ms timeout on allocating new net_pkt for
received data. Move that configuration to Kconfig, so that value can be
tuned according to application needs.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
There is no reason to keep active stream socket when there was some data
loss. Mark such socket for closing and close it when all (so far)
received packets have been processed.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Add uart bus interface to extended esWIFI driver. This enables all
Inventek modules with IWIN AT Commands firmware.
Signed-off-by: Gerson Fernando Budke <gerson.budke@atl-electronics.com>
The WIFI_ESWIFI_NAME config would be uselful when there is no device
tree alternative. The esWIFI driver already is on device tree and the
label property exists. This remove WIFI_ESWIFI_NAME Kconfig variable
and switch to device tree equivalent.
Signed-off-by: Gerson Fernando Budke <gerson.budke@atl-electronics.com>
Remove global access to structure eswifi_spi_data variable. Instead,
add a method to pass access to that structure. This allows better
control to the data bus variable.
Signed-off-by: Gerson Fernando Budke <gerson.budke@atl-electronics.com>
All net_bufs allocated to modem_cmd_handler's data->rx_buf are consumed
synchronously in esp.c in the same thread. This means that allocating
them with timeout makes no sense, because timeout will *always* be hit
when there are no more buffers in net_buf_pool.
Get rid of the unnecessary timeout, as it doesn't help and just slows
down processing of incoming data, increasing possibility of data
overrun.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
So far a dedicated buffer was used for data read from modem
interface. New net_bufs were allocated and filled later, which means
that data was lost when no more net_bufs were available in the pool.
Prevent data loss by allocating net_buf before attempting any read on
modem interface. Process incoming data in a loop as long as reading from
interface results in new data. Also remove dedicated buffer
(data->read_buf) and directly fill net_buf content instead. As a side
effect there are less memory copy operations and RAM usage is reduced.
Pre-allocated net_buf is now always appended to data->rx_buf. When there
was no (more) data read from interface to such net_buf, then this empty
net_buf will be on the end of data->rx_buf fragment list. Update
skipcrlf() and findcrlf() implementations to explicitly check for each
net_buf length, instead of blindly assuming them to have at least single
byte.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Listen func goes through a chain of calls to call winc1500_accept.
This is done to install the accept callback and should not block.
This fixes#28953 where winc1500 driver blocks on listen.
Signed-off-by: Nicolai Glud <nicolai.glud@prevas.dk>
Add power-gpios device-tree binding property to power on module before
communicating with it. This pin is called CHIP_PU in case of ESP32{,-S2}
and CHIP_EN in case of ESP8266. Dedicated reset pin is available only on
the latter, however Espressif recommends (in ESP8266 Hardware Design
Guidelines) to use CHIP_EN instead. Follow those recommendations and use
power-gpios to reset chip if that is provided over device-tree.
Configure power-gpios and reset-gpios as inactive by default, so that
chip becomes ready after executing esp_reset() function, either if one
or both are provided over device-tree.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Most DT bindings use reset-gpios name when there is a pin to reset whole
chip. Rename wifi-reset-gpios to reset-gpios to be more consistent
between various drivers. Additionally this prevents confusion, as
somebody might think that this pin resets only WiFi, which is not true.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
If HW flow control is enabled, then modem_context framework won't drain
UART FIFO blindly, but will stop when there is no more space in RX
ring_buf. This prevents data loss by "pausing" incoming data on hardware
level.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
This API allows to drop use of preallocated isr_buf. Most importantly as
a result RAM usage is reduced for each driver utilizing modem_context
framework. Additionally there is less copying done in ISR context, as
data is direcly read from UART FIFO to ring_buf.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Allow to reconfigure UART baudrate on ESP and on host MCU, so a
non-default baudrate can be used for communication. This option helps
for example to increase network bandwidth without touching ESP chip
firmware.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Add Kconfig options to configure modem command handler buffer size and
count. This will allow to fine tune those based on UART baudrate, system
load and available memory.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Add new Kconfig option to configure modem UART interface handler ring
buffer size. This will allow to lower ring buffer to save some resources
or increase it in case high network bandwidth is utilized (with high
UART baudrate).
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Reimplement getaddrinfo to call SlNetUtil_getAddrInfo from the TI HAL
for a more robust implementation that supports both client and server
modes, and performs better error-checking.
Fixes#11890
Signed-off-by: Vincent Wan <vwan@ti.com>
The value from socket() was directly assigned to offload_context, which
was treated as a unsigned integer when compared. This prevented the
following if statement from catching any errors, leading to random RAM
access. This is fixed by using an intermediate with the same type as the
return value, which is then assigned to offload_context after error
checking.
Signed-off-by: Abram Early <abram.early@gmail.com>
When offloading only ZSOCK_POLLIN and DNS_EAI_MEMORY are available, so
use those instead of the standard ones.
Fixes#28069.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Schedule read work every 100ms for better latency.
Reschedule read work directly in case of a packet is received
(fast dequeue).
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Passing the device's data was sufficient as only the data is being used
by the thread.
Fixes#27399
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Now that device_api attribute is unmodified at runtime, as well as all
the other attributes, it is possible to switch all device driver
instance to be constant.
A coccinelle rule is used for this:
@r_const_dev_1
disable optional_qualifier
@
@@
-struct device *
+const struct device *
@r_const_dev_2
disable optional_qualifier
@
@@
-struct device * const
+const struct device *
Fixes#27399
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Remove non necessary eswifi lock/unlock in read procedure (eswifi
lock support nesting).
Return NULL packet in case socket is closed by remote (tested with
net shell and net recv).
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Variable "sd" is no longer referenced after refactoring in 2ed6b6a8ed.
It caused warning-as-error when building using sanitycheck, e.g. in CI.
This is 2nd fixed of the above refactor, after eb9a18432.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
PR #27485 introduced new implementations for close() that do not build
correctly. This commit fixes the problem by re-implementing them.
Fixes#27587
Signed-off-by: Vincent Wan <vwan@ti.com>
Do not route close() calls via ioctl() as that is error prone
and quite pointless. Instead create a callback for close() in
fdtable and use it directly.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
unsigned byte value is initialized to -1 and so never satisifies
the while condition, leading to broken ipv4 parsing (dead code).
Fix this by making byte value signed.
Reported by coverity #27326.
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
There are possible buffer overflows when parsing the ip address and
SSID. Ensure that we never overwrite the ip and SSID buffers.
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
There are multiple CID errors all based on the test of
'socket > ESWIFI_OFFLOAD_MAX_SOCKETS'. The test needs to
be '>=' in all cases.
Fixes: #27138Fixes: #27139Fixes: #27140Fixes: #27141Fixes: #27142Fixes: #27143
Signed-off-by: David Leach <david.leach@nxp.com>
When CONFIG_POSIX_API is set in lieu of CONFIG_NET_SOCKETS_POSIX_NAMES,
some macros such as POLLIN, POLLOUT and MSG_PEEK are undefined, and
the appropriate posix headers need to be included to access those.
Signed-off-by: Vincent Wan <vwan@ti.com>
Fix compilation warnings when CONFIG_NET_SOCKETS_OFFLOAD enabled
warning: passing argument 5 of '__eswifi_socket_new' makes pointer
from integer without a cast [-Wint-conversion]
395 | idx = __eswifi_socket_new(eswifi, family, type, proto, 0x50CE);
| ^~~~~~
| |
| int
eswifi.h:137:21: note: expected 'void *' but argument is of type 'int'
137 | int proto, void *context);
| ~~~~~~^~~~~~~
In function 'eswifi_socket_bind':
warning: comparison between pointer and integer
461 | if ((addrlen == NULL) || (addr == NULL) ||
|
Signed-off-by: Parthiban Nallathambi <parthiban@linumiz.com>
use common log module across all eswifi sources, which allows
to compile with CONFIG_NET_SOCKETS_OFFLOAD.
Signed-off-by: Parthiban Nallathambi <parthiban@linumiz.com>
It can happen that the command '>' is received between
modem_cmd_send_nolock and modem_cmd_handler_update_cmds. Since the
command handler for '>' is not set, sem_tx_ready will not be given
and _sock_send will timeout. Make sure the command handlers are set
before the send by also passing them to modem_cmd_send_nolock.
Signed-off-by: Tobias Svehagen <tobias.svehagen@gmail.com>
