On 11/18/21 13:32, Michael Balzer wrote:
there was no change to the RTC or IRAM sections lately. Maybe somehow your linker definition got messed up?
"git status" doesn't show any modified files in my copy of the tree. And I don't see any changes to my toolchain, looks like the last time I updated it was in July. The only shared libraries the toolchain ld uses are libz and libc from the base os (not updated in more than a year). The amount of overflow seems unreasonably large? /usr/local/xtensa-esp32-elf/bin/../lib/gcc/xtensa-esp32-elf/5.2.0/../../../../xtensa-esp32-elf/bin/ld: RTC_SLOW segment data does not fit. /usr/local/xtensa-esp32-elf/bin/../lib/gcc/xtensa-esp32-elf/5.2.0/../../../../xtensa-esp32-elf/bin/ld: region `iram0_0_seg' overflowed by 2172504 bytes /usr/local/xtensa-esp32-elf/bin/../lib/gcc/xtensa-esp32-elf/5.2.0/../../../../xtensa-esp32-elf/bin/ld: region `rtc_slow_seg' overflowed by 33681 bytes 2172504 bytes is over 2MB?!?! I tried reverting to ~Oct 29th but I get the same errors so clearly something is recently wrong with my build environment. I found this: https://github.com/espressif/esp-idf/issues/6914 And tried running idf_size.py: ice 34 % python $IDF_PATH/tools/idf_size.py build/ovms3.map load: 0.09 cmd: python3.8 85307 [running] 3.59r 3.57u 0.00s 29% 13616k load: 0.45 cmd: python3.8 85307 [running] 12.20r 12.18u 0.00s 73% 19824k Total sizes: DRAM .data size: 0 bytes DRAM .bss size: 0 bytes Used static DRAM: 0 bytes ( 180736 available, 0.0% used) Used static IRAM: 16775 bytes ( 114297 available, 12.8% used) Flash code: 0 bytes Flash rodata: 64388 bytes Total image size:~ 81163 bytes (.bin may be padded larger) The end of that github issue points to changes in idf_size, could this be a python issue? Maybe python packages have been updated on this box since my last build at the end of October. Another seemingly related thread: https://githubmemory.com/repo/aws/amazon-freertos/issues/3356 "I check the components size. And you can see in the log, program uses the 0.8% of IRAM then how should the IRAM memory region overflow?" This feels like broken math somewhere. Craig