search

raburton / rboot

An open source bootloader for the ESP8266
https://richard.burtons.org/tag/rboot/?order=ASC
MIT License
300 stars 72 forks source link

Error in checksum calculation? #53

Open gpreviato opened 5 years ago

gpreviato commented 5 years ago

Hi, I had worked a lot you your great code (but also I've made some mess in...), and I was looking to what seem be a miscalculation in the checksum. Actually I have a mismatch between the checksum calculated by rboot and the one calculated from esptool.py. Seems the latter doesn't consider the first segment in checksum calculation. Here the results from a modified esptool (for log): ./esptool.py --chip esp8266 image_info ../../firmware/firmware.bin Entry point: 4010023c 4 segments

Segment 1: len 0xc1530 load 0x00000000 file_offs 0x00000008 Segment 2: len 0x077b4 load 0x40100000 file_offs 0x000c1548 Segment 3: len 0x00cf4 load 0x3ffe8000 file_offs 0x000c8d04 Segment 4: len 0x019bc load 0x3ffe8cf8 file_offs 0x000c9a00 Seg: 0 -> seg.include_in_checksum: 0. DataLen: 0xc1530 Checksum: 0xef Seg: 1 -> seg.include_in_checksum: 1. DataLen: 0x77b4 Checksum: 0x56 Seg: 2 -> seg.include_in_checksum: 1. DataLen: 0xcf4 Checksum: 0x66 Seg: 3 -> seg.include_in_checksum: 1. DataLen: 0x19bc Checksum: 0x76 Checksum: 76 (valid)

As you can see the first segment is excluded from calculation. Doing the same with your code, i.e. adding: if (sectcurrent!=0) { for (loop = 0; loop < readlen; loop++) { chksum ^= buffer[loop]; } } seems set everything ok.

What do you think?

Rgds,

--gra

gpreviato commented 5 years ago

Hi, I had worked a lot you your great code (but also I've made some mess in...), and I was looking to what seem be a miscalculation in the checksum. Actually I have a mismatch between the checksum calculated by rboot and the one calculated from esptool.py. Seems the latter doesn't consider the first segment in checksum calculation. Here the results from a modified esptool (for log): ./esptool.py --chip esp8266 image_info ../../firmware/firmware.bin Entry point: 4010023c 4 segments

Segment 1: len 0xc1530 load 0x00000000 file_offs 0x00000008 Segment 2: len 0x077b4 load 0x40100000 file_offs 0x000c1548 Segment 3: len 0x00cf4 load 0x3ffe8000 file_offs 0x000c8d04 Segment 4: len 0x019bc load 0x3ffe8cf8 file_offs 0x000c9a00 Seg: 0 -> seg.include_in_checksum: 0. DataLen: 0xc1530 Checksum: 0xef Seg: 1 -> seg.include_in_checksum: 1. DataLen: 0x77b4 Checksum: 0x56 Seg: 2 -> seg.include_in_checksum: 1. DataLen: 0xcf4 Checksum: 0x66 Seg: 3 -> seg.include_in_checksum: 1. DataLen: 0x19bc Checksum: 0x76 Checksum: 76 (valid)

As you can see the first segment is excluded from calculation. Doing the same with your code, i.e. adding: if (sectcurrent!=0) { for (loop = 0; loop < readlen; loop++) { chksum ^= buffer[loop]; } } seems set everything ok.

What do you think?

Rgds,

--gra

Do not mind about this... I've to go digger into....

raburton commented 5 years ago

Sorry I haven't had chance to look at this, I'm not sure which checksum you're referring too. If it's the standard one then it must be right or the esp wouldn't run the code. if it's the extra rBoot checksum, as far as I know this is not calculated by esptool.py.