Craig, there's a new idf 3.3 config against the starvation, did you set that? Your sdkconfig should have this: CONFIG_SPI_FLASH_YIELD_DURING_ERASE=y CONFIG_SPI_FLASH_ERASE_YIELD_DURATION_MS=20 CONFIG_SPI_FLASH_ERASE_YIELD_TICKS=1 Those are the defaults, working for me. Regards, Michael Am 29.07.19 um 10:01 schrieb Mark Webb-Johnson:
Really no idea. I’ve seen cases of the wifi getting messed up during flash erase (bug in esp idf as it starves the wifi stack during the long erase operation, and supposedly fixed in the 3.3 we are now using).
Can you copy off the elf before doing the build/update, so you have that to decode the backtrace?
Regards, Mark
On 29 Jul 2019, at 12:11 PM, Craig Leres <leres@xse.com> wrote:
I think there might be a new problem with the code. The most common way for me to load new firmware is via the "Flash from web" from my system at home. Several times this weekend I've gotten "Request error, please retry" when starting the flash. Just now I got one half way through the flash. I haven't been able to get a useful stack trace because my build/ovms3.elf has already been overwritten with a new image when this happens and xtensa-esp32-elf-addr2line just prints a bunch of "??:0" lines. I've had this multiple times on two different modules today and so far 3 times in a row. I finally resorted to booting from the factory partition (which I try to keep loaded with the current public release). That resulted in it completely losing its config and firing up the OVMS ssid instead of joining my network.
I've attached some frame grabs.
Craig <flash.png><module.png>_______________________________________________ OvmsDev mailing list OvmsDev@lists.openvehicles.com http://lists.openvehicles.com/mailman/listinfo/ovmsdev
OvmsDev mailing list OvmsDev@lists.openvehicles.com http://lists.openvehicles.com/mailman/listinfo/ovmsdev
-- Michael Balzer * Helkenberger Weg 9 * D-58256 Ennepetal Fon 02333 / 833 5735 * Handy 0176 / 206 989 26