Fixed parsing hex numbers in chip config files #1169
Conversation
|
@gszy: Please check for build failures, e.g.:
You may also collect several issues in one PR for your convenience in the future - it shall be fine as well. |
process_chipfile() used to improperly parse hex numbers in chip config files (*.chip), because it used atoi(), which read all such numbers as 0. This resulted, among other issues, in chip_id being set to 0 for all read chips. Such chip id could not match any actual MCU. Replace the atoi() calls with sscanf(…, "%i", …), where %i should match integers in base 10, 8 and 16, depending on the number prefix.
|
Sorry for the failures—now it should build cleanly, though I’m not sure if that’s a solution you approve: // may set invalid flash types (not defined in enum stlink_flash_type)
if (sscanf(value, "%i", (int *) &ts->flash_type) < 1) |
|
@gszy This is a good solution to the problem! |
Add a STLINK_FLASH_TYPE_MAX constant that can be used to check if a flash type (an integer) is in the range of valid enum stlink_flash_type values.
I’m trying to make my PRs small and easy to review 🙂
I have added this check in separate commits (feel free to cherry pick what you find useful). There should be some range checks for other parameters as well, but I hope for another PR when I (or someone else) will find more time. |
|
@gszy Thank you! |
process_chipfile()used to improperly parse hex numbers in chip config files (*.chip), because it usedatoi(), which read all such numbers as 0. This resulted, among other issues, inchip_idbeing set to 0 for all read chips. Such chip id could not match any actual MCU.Replace the
atoi()calls withsscanf(…, "%i", …), where%ishould match integers in base 10, 8 and 16, depending on the number prefix.I can rebase this PR once (if) #1166 is merged.