On Thu, Sep 17, 2020 at 08:32:39AM -0400, Jeff Layton wrote: > On Wed, 2020-09-16 at 21:11 -0700, Eric Biggers wrote: > > From: Eric Biggers <ebiggers@xxxxxxxxxx> > > > > fscrypt_set_test_dummy_encryption() requires that the optional argument > > to the test_dummy_encryption mount option be specified as a substring_t. > > That doesn't work well with filesystems that use the new mount API, > > since the new way of parsing mount options doesn't use substring_t. > > > > Make it take the argument as a 'const char *' instead. > > > > Instead of moving the match_strdup() into the callers in ext4 and f2fs, > > make them just use arg->from directly. Since the pattern is > > "test_dummy_encryption=%s", the argument will be null-terminated. > > > > Are you sure about that? I thought the point of substring_t was to give > you a token from the string without null terminating it. > > ISTM that when you just pass in ->from, you might end up with trailing > arguments in your string like this. e.g.: > > "v2,foo,bar,baz" > > ...and then that might fail to match properly > in fscrypt_set_test_dummy_encryption. > Yes I'm sure, and I had also tested it. The use of match_token() here is to parse one null-terminated mount option at a time. The reason that match_token() can return multiple substrings is that the pattern might be something like "foo=%d:%d". But here it's just "test_dummy_encryption=%s". "%s" matches until end-of-string. - Eric
