> > This patch check DataLength in stream > > entry only if the type is ALLOC_NO_FAT_CHAIN and add the check ensure > > that dentry offset does not exceed max dentries size(256 MB) to avoid > > the circular FAT chain issue. > > I think it's preferable to add a 256MB check during dir-scan.(as I said in the previous) If you want > to solve "the circular FAT chain issue", you should add it to other dir-scan loops. > (exfat_search_empty_slot, exfat_check_dir_empty, exfat_count_dir_entries, etc ...). > > Also, the dir-scan loop may not terminate when TYPE_UNUSED is detected. > Even if TYPE_UNUSED is detected, just break the inner for-loop will continue the outer while-loop, so > the next cluster will be accessed. > If we can't use DataLength for checking, we should check the contents more strictly instead. > > The current implementation has several similar dir-scan loops. > These are similar logics and should be abstracted, if possible. Can you send me the patch instead of just talking? > > BR > Kohada Tetsuhiro <Kohada.Tetsuhiro@xxxxxxxxxxxxxxxxxxxxxxxxxxx>
