our IoT product usually sends a steady stream of data to our server. If we lose connectivity to the server, we buffer the messages in NVS. To be able to bridge longer offline periods without any data loss, we want to use a large partition for NVS, which is why we upgraded our flash memory to 16 MB.
Unfortunately, I noticed that when I increase the NVS partition size, the chip won't start any more due to an out-of-memory error. Looking into the nvs_pagemanager.cpp file, it seams that the NVS system tries to allocate RAM for each page, which exhausts the heap. It works for a 1MB NVS partition, but fails for 2MB or larger partitions.
The backtrace (for a 6MB NVS partition):
Code: Select all
0x400d0b4f: __cxa_end_catch at /home/david/esp/esp-idf/components/cxx/./cxx_exception_stubs.cpp:13
0x40089273: invoke_abort at /home/david/esp/esp-idf/components/esp32/./panic.c:572
0x4008929f: abort at /home/david/esp/esp-idf/components/esp32/./panic.c:572
0x400d0b4f: __cxa_end_catch at /home/david/esp/esp-idf/components/cxx/./cxx_exception_stubs.cpp:13
0x4012f998: operator new(unsigned int) at /builds/idf/crosstool-NG/.build/src/gcc-5.2.0/libstdc++-v3/libsupc++/new_op.cc:54
0x4012f95d: operator new[](unsigned int) at /builds/idf/crosstool-NG/.build/src/gcc-5.2.0/libstdc++-v3/libsupc++/new_opv.cc:32
0x40125710: nvs::PageManager::load(unsigned int, unsigned int) at /home/david/esp/esp-idf/components/nvs_flash/src/nvs_pagemanager.cpp:24 (discriminator 4)
0x401244b6: nvs::Storage::init(unsigned int, unsigned int) at /home/david/esp/esp-idf/components/nvs_flash/src/nvs_storage.cpp:41
0x40123d23: nvs_flash_init_custom at /home/david/esp/esp-idf/components/nvs_flash/src/nvs_api.cpp:454
0x40123d8d: nvs_flash_init_partition at /home/david/esp/esp-idf/components/nvs_flash/src/nvs_api.cpp:454
If there is no way, then I probably need create a SPIFFS partition...
Thank you for your help, you guys rock!