>I’d use imperative mood in the commit message, and be more explicit. > >> mdadm: Use PATH_MAX over MAX_SYSFS_PATH_LEN > >> GCC8 make more strict checks of possible truncation during snprintf > >s/make/makes/ > >or: GCC 8 checks possible truncation during snprintf more strictly than >GCC 7. > >> calls than GCC7 which cause compilation errors. This patch >> fixes compilation of mdadm on GCC8 compiler. > >So you increase the buffer size from 120 to PATH_MAX (4096)? What is >the logic behind that besides just being bigger? > >PATH_MAX seems to be tricky: https://eklitzke.org/path-max-is-tricky I agree with you regarding solution using PATH_MAX. MAX_SYSFS_PATH_LEN defines max path and this value should be used. I will upload second version of patch with check against errors and truncation. Best regards, Krzysztof