Item 1: When running "make flash" I see the following warning: Running make flash... /Users/casner/src/github/esp-idf/components/partition_table/Makefile.projbuild:16: warning: undefined variable `MD5_OPT' Flashing binaries to serial port /dev/cu.SLAB_USBtoUART (app at offset 0x10000)... A grep for MD5_OPT in the esp-idf source tree only finds the two instances in Makefile.projbuild, so it looks like is not defined anywhere if CONFIG_PARTITION_TABLE_MD5=y. Do others see this warning? If not, I must need to update something. Item 2: I don't have an SD card installed, but I see the following errors at boot time (with current Master sources): I (2492) sdcard: SD CARD has been inserted E (2532) sdmmc_req: handle_idle_state_events unhandled: 00000004 00000000 E (2532) sdmmc_cmd: sdmmc_card_init: send_op_cond (1) returned 0x107 -- Steve
I see this too. I’m guessing it is a bug in the ESP IDF version we pulled at last merge. I enabled the option for MD5 checksums on the partition table, for production builds, and it seems to be working (looking at the generated partitions.bin). It also validates correctly. I think this is just a facility to disable the generation of MD5 checksums, and the MD5_OPT variable is not set empty when using MD5 checksums. Should not impact operation. Once the next OTA firmware release goes out, I intend to try another upstream merge for the IDF. Just a bit scary as historically they have broken stuff. Regards, Mark.
On 14 Apr 2018, at 6:38 AM, Stephen Casner <casner@acm.org> wrote:
Item 1: When running "make flash" I see the following warning:
Running make flash... /Users/casner/src/github/esp-idf/components/partition_table/Makefile.projbuild:16: warning: undefined variable `MD5_OPT' Flashing binaries to serial port /dev/cu.SLAB_USBtoUART (app at offset 0x10000)...
A grep for MD5_OPT in the esp-idf source tree only finds the two instances in Makefile.projbuild, so it looks like is not defined anywhere if CONFIG_PARTITION_TABLE_MD5=y.
Do others see this warning? If not, I must need to update something.
Item 2: I don't have an SD card installed, but I see the following errors at boot time (with current Master sources):
I (2492) sdcard: SD CARD has been inserted E (2532) sdmmc_req: handle_idle_state_events unhandled: 00000004 00000000 E (2532) sdmmc_cmd: sdmmc_card_init: send_op_cond (1) returned 0x107
-- Steve _______________________________________________ OvmsDev mailing list OvmsDev@lists.openvehicles.com http://lists.openvehicles.com/mailman/listinfo/ovmsdev
Agree that this warning is harmless. It is not fixed yet in espressif/esp-idf master. I just filed an issue for it. (I could submit a fix, but that review process takes so long...) -- Steve On Sat, 14 Apr 2018, Mark Webb-Johnson wrote:
I see this too. I'm guessing it is a bug in the ESP IDF version we pulled at last merge.
I enabled the option for MD5 checksums on the partition table, for production builds, and it seems to be working (looking at the generated partitions.bin). It also validates correctly.
I think this is just a facility to disable the generation of MD5 checksums, and the MD5_OPT variable is not set empty when using MD5 checksums. Should not impact operation.
Once the next OTA firmware release goes out, I intend to try another upstream merge for the IDF. Just a bit scary as historically they have broken stuff.
Regards, Mark.
On 14 Apr 2018, at 6:38 AM, Stephen Casner <casner@acm.org> wrote:
Item 1: When running "make flash" I see the following warning:
Running make flash... /Users/casner/src/github/esp-idf/components/partition_table/Makefile.projbuild:16: warning: undefined variable `MD5_OPT' Flashing binaries to serial port /dev/cu.SLAB_USBtoUART (app at offset 0x10000)...
A grep for MD5_OPT in the esp-idf source tree only finds the two instances in Makefile.projbuild, so it looks like is not defined anywhere if CONFIG_PARTITION_TABLE_MD5=y.
Do others see this warning? If not, I must need to update something.
Item 2: I don't have an SD card installed, but I see the following errors at boot time (with current Master sources):
I (2492) sdcard: SD CARD has been inserted E (2532) sdmmc_req: handle_idle_state_events unhandled: 00000004 00000000 E (2532) sdmmc_cmd: sdmmc_card_init: send_op_cond (1) returned 0x107
Mea culpa. A bug introduced a few commits ago. Michael fixed it earlier today. Regards, Mark.
On 14 Apr 2018, at 6:38 AM, Stephen Casner <casner@acm.org> wrote:
Item 1: When running "make flash" I see the following warning:
Running make flash... /Users/casner/src/github/esp-idf/components/partition_table/Makefile.projbuild:16: warning: undefined variable `MD5_OPT' Flashing binaries to serial port /dev/cu.SLAB_USBtoUART (app at offset 0x10000)...
A grep for MD5_OPT in the esp-idf source tree only finds the two instances in Makefile.projbuild, so it looks like is not defined anywhere if CONFIG_PARTITION_TABLE_MD5=y.
Do others see this warning? If not, I must need to update something.
Item 2: I don't have an SD card installed, but I see the following errors at boot time (with current Master sources):
I (2492) sdcard: SD CARD has been inserted E (2532) sdmmc_req: handle_idle_state_events unhandled: 00000004 00000000 E (2532) sdmmc_cmd: sdmmc_card_init: send_op_cond (1) returned 0x107
-- Steve _______________________________________________ OvmsDev mailing list OvmsDev@lists.openvehicles.com http://lists.openvehicles.com/mailman/listinfo/ovmsdev
The reason I encountered this error message was that I had to remove the SD card in order to test ssh because otherwise there is not enough free RAM for the connection to open when using v3.0 hardware with the default configuration. What the client sees in this case is just that the connection immediately closes without becoming open first. Unfortunately, the error message that appears on the async console is "E (95787) ssh: Error in reception: -29" which is not very clear. I could provide the text corresponding to that error number, but that is "crypto action failed" which might be misleading. The problem is that the memory allocation failure occurs several layers down, so it gets hidden by higher-level error returns. I might be able to supply my own malloc function that logs an error message if the actual malloc fails. I did go back and reduce the configuration to remove (for now) the MAX7317, EXT12V, OBD2ECU, RE_TOOLS, EDITOR, CANOPN and the V3 server and all the vehicles other than null and Roadster. That allows me to open an SSH connection with the SD card installed. My question for you guys is what items that are not part of OVMS could we safely remove from the default configuration? Do we need SPIFFS, MBED_TLS, OPENSSL, PTHREADS? I'm not sure what libraries are needed for https. -- Steve On Sat, 14 Apr 2018, Mark Webb-Johnson wrote:
Item 2: I don't have an SD card installed, but I see the following errors at boot time (with current Master sources):
I (2492) sdcard: SD CARD has been inserted E (2532) sdmmc_req: handle_idle_state_events unhandled: 00000004 00000000 E (2532) sdmmc_cmd: sdmmc_card_init: send_op_cond (1) returned 0x107
Mea culpa. A bug introduced a few commits ago. Michael fixed it earlier today.
Regards, Mark.
On 14 Apr 2018, at 6:38 AM, Stephen Casner <casner@acm.org> wrote:
Item 1: When running "make flash" I see the following warning:
Running make flash... /Users/casner/src/github/esp-idf/components/partition_table/Makefile.projbuild:16: warning: undefined variable `MD5_OPT' Flashing binaries to serial port /dev/cu.SLAB_USBtoUART (app at offset 0x10000)...
A grep for MD5_OPT in the esp-idf source tree only finds the two instances in Makefile.projbuild, so it looks like is not defined anywhere if CONFIG_PARTITION_TABLE_MD5=y.
Do others see this warning? If not, I must need to update something.
Item 2: I don't have an SD card installed, but I see the following errors at boot time (with current Master sources):
I (2492) sdcard: SD CARD has been inserted E (2532) sdmmc_req: handle_idle_state_events unhandled: 00000004 00000000 E (2532) sdmmc_cmd: sdmmc_card_init: send_op_cond (1) returned 0x107
-- Steve _______________________________________________ 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
participants (2)
-
Mark Webb-Johnson -
Stephen Casner