Justin Stitt <justinstitt@xxxxxxxxxx> wrote: > Prefer `strscpy` to `strncpy` for use on NUL-terminated destination > buffers. > > This fixes a potential bug due to the fact that both `t->u.user.name` > and `name` share the same size. This replacement seems fine. > Signed-off-by: Justin Stitt <justinstitt@xxxxxxxxxx> > > --- > Here's an example of what happens when dest and src share same size: > | #define MAXLEN 5 > | char dest[MAXLEN]; > | const char *src = "hello"; > | strncpy(dest, src, MAXLEN); // -> should use strscpy() > | // dest is now not NUL-terminated This can't happen here, the source string is coming from the kernel (xt target and matchinfo struct). But, even if it would it should be fine, this function prepares the translated 64bit blob which gets passed to translate_table(), and that function has to check for '\0' presence. Normally it handles the native (non-compat) data originating from userspace, so m-->user.name can not be assumed to contain a \0.
