* Fix filesystem mutex handling
This fixes the re-initialization of the filesystem mutex when the dyn storage
module is reinitialized. Using xSemaphoreCreateMutex would cause a memory leak
when the begin function is called multiple times and a semaphore leakage could
be caused by the re-initialization of the global fsMutex variable while the
semaphore is taken.
The fsMutex should not be taken while the logLine function is called as this
would cause a nested take of the fsMutex, which causes a deadlock.
Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
* Fix hard coded littlefs in json upload
Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
* Add new AP hardware support OpenEPaperLink_ESP32-PoE-ISO_AP
This is based on Olimex ESP32-PoE-ISO
https://www.olimex.com/Products/IoT/ESP32/ESP32-POE-ISO/open-source-hardware
It has a SD Card slot that is used to store all filesystem data on SD.
Use the prepare_sdcard.sh script to copy all needed data to an empty
SD card that is formatted with FAT filesystem. The AP firmware will format
the SD if an unformatted or from formatted card is used. This can be used
to intially prepare an empty SD card for usage.
For tag communication a ESP32-C6-WROOM-1(U) is used with the following
connection scheme:
ESP32-PoE-ISO | ESP32-C6-WROOM-1
---------------+------------------
GPIO5 | EN
GPIO13 | GPIO9
GPIO36 | GPIO3
GPIO4 | GPIO2
GPIO33 | GPIO24
GPIO32 | GPIO25
| GPIO8 pullup 5.1k
Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
* Avoid error message log print when parsers.json is missing
Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
* Workaround for IEEE802.15.4 modem stuck issue
The ESP32-C6 esp-idf based modem firmware can run into a case where it
does not receive data anymore from the tags. This can be detected when
it starts to print
"receive buffer full, drop the current frame"
and does not recover from that. In such a case a modem reset is triggered.
Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
* Add OpenEPaperLink_ESP32-PoE-ISO_AP to release build
Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
* Add OpenEPaperLink_ESP32-PoE-ISO_AP to condidional build
Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
* Add Ethernet support
The ethernet support allows to make the network/internet connection
via LAN cable instead of WiFi. LAN is preferred, if a LAN cable is
connected and a valid IP configuration via DHCP can be obtained, WiFi
is switched off. If the LAN cable is disconnected, a fall back to
WiFi is done.
Use those defines in platform.ini for PHY settings:
ETHERNET_PHY_POWER: IO pin where the PHY can be switched of/on, can be
-1 if not used.
ETHERNET_CLK_MODE: PHY clock mode, see eth_clock_mode_t in ETH.h
ETHERNET_PHY_MDC: PHY MDC pin
ETHERNET_PHY_MDIO: PHY MDIO pin
ETHERNET_PHY_TYPE: PHY type, see eth_phy_type_t in ETH.h
Limitations:
- only DHCP is supported, no static IP configuration for LAN so far.
- If GPIO0 is used for one of the ETHERNET_CLK_MODE modes, then GPIO0
cannot be used to clear the WiFi configuration.
Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
---------
Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
Co-authored-by: Frank Kunz <mailinglists@kunz-im-inter.net>
- turn off Buienradar and Google Calendar content on ESP32-S2 based AP's because the firmware doesn't fit the flash memory anymore. If you need this and you use e.g. the mini AP v2, don't update. ESP32-S3 based AP's are not affected.
- On the update tab, the list of available versions now only shows the versions that are actually available for the build environment in use, so when we leave out S2-based builds in future releases, they won't show the new version.
So far the serial output is a mix of \r\n for println() and \n for
sprintf() and such.
This commit consolidates this to a consistent \r\n.
There should be no noticeable difference for users of terminals that
auto-substitute a sole \n to \r\n.
This will however make the serial output friendly to default behavior of
various terminal emulators like 'screen' and 'minicom'.
This commit doesn't touch anything that is not serial output and also
leaves 'ets_printf' alone just in case.
Co-authored-by: Thomas B. Rücker <thomas.ruecker@relexsolutions.com>
- adds tag serial debug in webflasher and power up/down buttons
- fixes bug where esp32-s3 connected via usb cdc slows down when no terminal connection is present
- fix: ghost tags reporting a different channel id than the AP are now ignored
- webflasher now ramps up/down power to prevent spikes
- bugfix: C6 doesn't get reset anoymore by the watchdog during flashing
- bugfix: C6 could go in flash mode unintended
- a failed connection during swd write blocks is now reported back instead of silently ignored
- added 4 areas of Sweden to day ahead price content
- new context menu option 'Delete all inactive tags' (only when right clicking on an inactive tag).
- adjusted some timings
- added webinterface for tag flasher
- added tcp transport for communicating with tag flasher (OTG USB also still works)
- added content 'timestamp', makes use of preloaded images and buttons on the 2.9" M3
- webinterface is now aware of C6 and flasher capabilities
- AP can run without ieee801.15.4 radio (i.e. flasher only) by shorting FLASHER_AP_TXD and FLASHER_AP_RXD
- added tcp transport option to OEPL-Flasher.py (serial also still works)
- added new environment OpenEPaperLink_Mini_AP_v4
- lots of finetuning and bug fixes
Pretty complex change here: pending images/commands are now queued.
A command (like LED flasher) will not overwrite a pending image anymore. Also, sending multiple preloaded images is possible.
Also works (at least, as far as I could test) in combination with Multi AP and mirroring tags ('display a copy' content type).
It you want to test this: don't forget to upload the changed files in /www (the pending icon is now displaying the amount of queued messages). Timing improvements will follow later (only one message can be transmitted every checkin. If multiple messages are queued, at this moment, you have to wait until the next checkin which takes 40-60 sec).
This comes also with the advantage of better stability if you upload multiple images to the same tag in succession. Before queuing, if was possible to replace the image between sending the pending message and the image transfer to the tag, causing md5 mismatches and instability.
Solves #47
- changed ESPAsyncWebServer and AsyncTCP (both not properly maintained for more than 5 years) to more recent fork with bugfixes
- fixed concurrency issues in /imgupload. Hurts the upload performance a lot, but probably that's not a real issue. Parallel uploads are now possible (still within the max 5 TCP connection limit)
- added short date format definition in language.json
- changed english date format to mm/dd
- added lightgrey (4)/darkgrey (5)/pink (6) colors in jsontemplate (uses pattern dither)
- fixed ntp init timing
- prevent using html file for tag firmware update
- removed excessive logging
- fallback to .bak on tagDB load error
- scheduled reboot once at night around 4:00
- neopixel patterns optimized. The 'breathing' led state now is green colored if everything is okay, and blue if there are one or more tags timed out.
- time zone is now set before wifi connect to show correct time zone in the logs during startup
- concurrent image upload POST is now blocked. If an upload is in progress while you do a second http POST, http status 409 Conflict is returned.
- small synchronisation bug fix in web interface on loading tag type
- dialog window close bugfix in painter
- image upload is now logged in /log.txt
* Add formatString convenience function
* Use String& for wsLog, wsErr and wsSerial
* Add tag data parser and parse tag data
* Make logLine use String&
* Fix issue with formatString
* Reuse payloadLength in processTagReturnData
* Fix parsing of unsigned/signed inetegers and cleanup
* Use c++17 standard
* Cleanup logging
- clean up webproxy getExtUrl
- bugfix init multicast on wifi reconnect
- add ip address of remote AP that's taken over a tag
- renamed button 'edit contentFS' to 'file system'
- apconfig: configure night time, where the system is not updating tags, and the tags are sleeping
- fixed bug calculating expected checkin
- throttle down apinfo update to once per minute
- fixed too many concurrent requests getting tagtypes
- resend current image when a tag reboots and the content type is 'static image' (all other content types were already regenerating the content)
- fixed timing of main loop
- AP info screen content card. Can run on any tag.
- now, you can use {variables} in the 'text' entries in a json template. You can set variables via a http call. If you update a variable, all tags with a json template containing that variable get updated.
- font name in json template is more flexible. You can just use 'filename.ttf' or 'filename.vlw'. A full path is still possible.
- colors in the json template can now be set using #rrggbb color values, and using 'black', 'white' or 'red'.
- added direct output for the TFT display for the yellow esp32-s3. No file writes needed.
- added POST variable 'ttl' to json template upload and image upload, to set the next checkin time
- added /variables-demo.html to demonstrate the variables.
- json templates received from jsonupload are now saved in /current, and reused.
- known issue: 'backup db' button doesn't work due to some browser policy change. Fixing.
thanks to @steinwedel for the inspiration on the variables and some other fixes.
This one saves a lot of flash space.
Now, wifimanager is replaced with our own method for managing the wifi connections.
Also, the whole thing will run now even if there is no wifi. If wifi disappears, it automatically tries to reconnect in the background every few minutes.
Wifi settings now available at all times at /setup
If it cannot connect, an access point 'OpenEPaperLink' is created where you can set the wifi credentials (http://192.168.4.1/setup).
To be able to store more images and never worry about running
out of storage again we are adding support of SDCards.
We are faced with two options:
1) Remove the LittleFS code
2) Allow for both FSs but increase the app partition as we run
out of space.
We have opted for 2 for two reasons:
1) Unformated SDCards can be initialized from the littleFS
2) If the SDCard and the flasher share the same SPI pins
the SDCard needs to be completely disconnected so there needs
to be a FS in that rare case.
Performance wise the SDCard is not slowing the system down.
Writing the AP db with 70 APs takes 2.5 seconds, on the SD
only 800ms. This however is SD and setup depending.
By default the SDCard is expected on pins 18,19,23 however this
is configurable via preprocessor defines.
Globally we replace LittleFS with a new pointer to the currently
used FS allowing for hotswapping where the FS pointer is not saved.
One case which is therefore not working is the SPIFSEditor which
copies the FS interally.
Signed-off-by: Mimoja <git@mimoja.de>
- added Swedish å Å to Bahnschrift20 and 30
- option to turn dither on/off in 'static image'
- major overhaul of contentmanager.cpp, content is now generated from json template.
- fallback to 1bpp for 4.2" when no psram
- added 4.2" content layouts
- fixed small apconfig bug
- pause content generation for 60 seconds after crash, to prevent uncontrollable boot loop
