## The problem
WiFiClient's connect method variant where no timeout is passed can block esp32 MCU and may then cause watchdog to kick in and reset the device. This behavior is different from that, what is in arduino-esp8266 core.
## Summary
Some cross-esp libraries (working both on esp32 and 8266), like PubSubClient simply call connect method on WiFiClient, to get connected to remote server. However, connect behavior varies betwen esp arduino 8266 and esp arduino 32 cores. This pull request tries introduce same behavior - to make connect method non-blocking on esp32, like it is with 8266 arduino core.
## Proposed solution
Introduce default fixed timeout that can be changed by #define - by default set to 3 seconds.
### Affected components:
WiFiClient
### Affected methods:
```c++
int connect(IPAddress ip, uint16_t port);
int connect(const char *host, uint16_t port);
```
### Impact
May impact projects or libraries using connect method variant without specified timeout, where:
- remote is located far away or
- connection is heavily limited, or
- remote is slow, when it comes to accept the connection
* Update toolchain
* Update package_esp32_index.template.json
* add optional component dependencies after Kconfig options are known (#5404)
Until this commit, Kconfig options (e.g. CONFIG_TINYUSB_ENABLED) were
used in conditions preceding idf_component_register to determine which
components need to be added to `arduino` component requirements.
However the Kconfig options aren't known at the early expansion stage,
when the component CMakeLists.txt files are expanded the first time
and requirements are evaluated. So all the conditions evaluated as if
the options were not set.
This commit changes the logic to only add these components as
dependencies when the Kconfig options are known. Dependencies become
"weak", which means that if one of the components isn't included into
the build for some reason, it is not added as a dependency.
This may happen, for example, if the component is not present in the
`components` directory or is excluded by setting `COMPONENTS` variable
in the project CMakeLists.txt file.
This also ensures that if the component is not present, it will not be
added as a dependency, and this will allow the build to proceed.
Follow-up to https://github.com/espressif/arduino-esp32/pull/5391.
Closes https://github.com/espressif/arduino-esp32/issues/5319.
* IDF master d93887f9f
* PlatformIO updates for CI (#5387)
* Update PlatformIO CI build script
- Switch to the latest toolchains 8.4.0 for ESP32, ESP32S2, ESP32C3
- Use PlatformIO from master branch for better robustness
* Update package.json for PlatformIO
Co-authored-by: Ivan Grokhotkov <ivan@espressif.com>
Co-authored-by: Valerii Koval <valeros@users.noreply.github.com>
Commit d15e1b0e [1] in the 2.0.0 pull request #4996 introduced a filter
to connect only to WiFi networks with a RSSI of -75 or better. This
results in strage behaviour, as a scan still shows networks, that can't
be connected to, even though with older versions, connecting to those
networks was possible and the connection (albeit slow) was stable.
Remove the RSSI filter for now by setting the threshold to -127, i.e.
the lowest possible value. Maybe in the future the filter threshold
could be exposed to users, to allow filtering out nearly unreachable
networks.
[1] github.com/espressif/arduino-esp32/pull/4996/commits/d15e1b0
When iterating through APlist from first to the last element, a corrected password for already known SSID is not used. Therefore, I propose to iterate from the back of the list. With iterating from the back through the vector, an already known SSID with a corrected password is used instead.
Thi may happen if read() gets called repeatedly (such as in HttpClient to parse response headers) and the connection is closed unexpectedly or the remote peer may have unexpected behavior that causes the underlying socket to report an error. In that case read() itself calls stop(), which invalidates the receive buffer object. Then when read() is called again without checking, such as inside readStringUntil(), the _rxBuffer is null and ESP32 crashes.
This adds support for the KSZ8081 ethernet phy.
Only the IDF 4+ specific code is modified, as the phy support was only
added recently:
espressif/esp-idf@aecfbf96
This is very much still work in progress and much more will change before the final 2.0.0
Some APIs have changed. New libraries have been added. LittleFS included.
Co-authored-by: Seon Rozenblum <seonr@3sprockets.com>
Co-authored-by: Me No Dev <me-no-dev@users.noreply.github.com>
Co-authored-by: geeksville <kevinh@geeksville.com>
Co-authored-by: Mike Dunston <m_dunston@comcast.net>
Co-authored-by: Unexpected Maker <seon@unexpectedmaker.com>
Co-authored-by: Seon Rozenblum <seonr@3sprockets.com>
Co-authored-by: microDev <70126934+microDev1@users.noreply.github.com>
Co-authored-by: tobozo <tobozo@users.noreply.github.com>
Co-authored-by: bobobo1618 <bobobo1618@users.noreply.github.com>
Co-authored-by: lorol <lorolouis@gmail.com>
Co-authored-by: geeksville <kevinh@geeksville.com>
Co-authored-by: Limor "Ladyada" Fried <limor@ladyada.net>
Co-authored-by: Sweety <switi.mhaiske@espressif.com>
Co-authored-by: Loick MAHIEUX <loick111@gmail.com>
Co-authored-by: Larry Bernstone <lbernstone@gmail.com>
Co-authored-by: Valerii Koval <valeros@users.noreply.github.com>
Co-authored-by: 快乐的我531 <2302004040@qq.com>
Co-authored-by: chegewara <imperiaonline4@gmail.com>
Co-authored-by: Clemens Kirchgatterer <clemens@1541.org>
Co-authored-by: Aron Rubin <aronrubin@gmail.com>
Co-authored-by: Pete Lewis <601236+lewispg228@users.noreply.github.com>
In case you have multiple APs with the same SSID/password and WiFi.begin(ssid, pwd) is called, ESP32 defaults to connect to THE FIRST AP FOUND matching ssid - according to:
https://docs.espressif.com/projects/esp-idf/en/latest/esp32/api-reference/network/esp_wifi.html#_CPPv418wifi_scan_method_t
This can cause situations that ESP32 is trying to connect to AP which is far away from it (weak signal) even there is AP close to it, just as in my house - I have AP on channel 6 which is in the 1st floor (quite far from room where I do the programming) and AP on channel 13 in the same room I do the programming (which is in the 2nd floor) - result: ESP32 is trying to connect to the AP on channel 6 because it finds it first and never try the AP on channel 13 in the same room, result of this is very unreliable WiFi connection.
When scan_method is set to WIFI_ALL_CHANNEL_SCAN, ESP32 scans all channels and choose the nearest / strongest AP (matching the ssid of course) as expected - result is no connection problems at all.
Therefore I propose adding this parameter into WiFi.begin, connection problems as described above are quite confusing (especially for beginners), I can imagine that for example Schools there are usually using mutliple APs and this can cause intermittent connection problems without obvious reason.
See #3960 for more details of the problem and the solution. I only implemented what was proposed in this ticket, as it solves my problem, which was the same as in this ticket. Credits for the code going to @etrinh ;-)
This also is a more consistence behaviour compared to esp8266, where it also is possible to restart the wifiserver immediately on the same port.
Real DNS resolving timeout used by lwip library is 14[s] (7[s] for DNS1 + 7[s] for DNS2). Function WiFiGenericClass::hostByName() has timeout set to lower value (only 4[s]), so callback function may be called after this low timeout and it may overlappe stack memory used now by other function.
Fixes#3722
1. WiFiProv.ino sketch is added that allows arduino users to do provisioning via SoftAP or BLE. WiFi.beginProvision( ) API is designed for provisioning in Arduino.
2. In WiFiProv.h provisioning class is defined.
3. WiFiProv.cpp contains implementation for provisioning class.
4. README.md file is added which contains detail information for working.
* std::shared_ptr Memory Leak
clientSocketHande and _rxBuffer are std::shared_ptr, the stop() call was not correctly releasing them and the operator= had similar problems fix for #3679
* operator= second attempt
* operator= third time
WiFiClient.connected() was hanging thinking there was still a connection when the remote had already closed. The one-liner in this patch addresses recv() returning 0 and errno==128. I couldn't find the corresponding errno for 128 but its caught by the case statement which includes EPIPE, ENOTCONN, ECONNRESET and ECONNABORTED so I assume its one of those. Broken pipe maybe?
```c
[D][WiFiClient.cpp:511] connected(): Disconnected: RES: 0, ERR: 128
```
EDIT: added comment to reflect that recv() can set errno when it returns 0.
fixed the connected() function so that it only checks errno if recv returns a value of -1.
"in the even of an error, errno is set to indicate the error" --manpage
This fixes the ESP32 Webserver when dealing with a modern webserver with a slow SD card.
This event name was missing in the list:
d5e2bb12ca/tools/sdk/include/esp32/esp_event_legacy.h (L43)
E.g., it was giving
[D][WiFiGeneric.cpp:337] _eventCallback(): Event: 24 - ETH_GOT_IP
When it should have been
[D][WiFiGeneric.cpp:337] _eventCallback(): Event: 24 - ETH_DISCONNECTED
