While I was adding the JavaDocs I have realized that the attribute name string is of unspecified encoding, which, I guess, means some native encoding will be used. But the attribute name can be non-ASCII as well which requires known encoding to be used. I can use attribute name with Japanese:

~/jna/jna> setfattr -n 'user.日本語' -v test TODO 
~/jna/jna> getfattr -d TODO                      
# file: TODO
user.日本語="test"

So it seems to me the utility functions should take name encoding parameter as well and all other variants should forward to it. What do you think?

According to the manpage the signature for setxattr is:

int setxattr (const char *path, const char *name, const void *value, size_t size, int flags);

For the datatypes it is said:

The VFS imposes limitations that an attribute names is limited to 255 bytes and an attribute value is limited to 64 kB. The list of attribute names that can be returned is also limited to 64 kB (see BUGS in listxattr(2)).

There are multiple mappings possible. The existing mappings map C char[] to java String. This can result in unaccessible files (where the encoded filename can't be represented in a String)/attributes, but I think for normal bindings this is a valid tradeoff. If you need to be able to access all files and attributes, a C char[] should be mapped as java byte[].

So I would bind the above as:

int setxattr (String path, String name, byte[] value, size_t size, int flags);

@wilx could you please have a look at this branch:
https://github.com/matthiasblaesing/jna/commits/upstream-pr-1021

I squashed your commits and added two new ones:

matthiasblaesing@62cd54e - adds a CHANGES.md entry for the new function
matthiasblaesing@bd184d5 - adds bindings based on byte[]

The latter commit is intended to circumvent Memory usage. Memory uses finalize to clean up native resources and a ByteBuffer created from a Memory object retains another reference to the Memory object, so adds another step in GC. Using a byte[] reduces this to two memcpy (to and from native).

@wilx I am OK with that.

Merged via 4e88709