[PATCH] Fix the new async copy-chunk for streams

Ralph Böhme slow at samba.org
Mon May 15 18:29:00 UTC 2017


On Sun, May 14, 2017 at 08:25:44PM +0200, David Disseldorp via samba-technical wrote:
> > Please review & push if ok. Thanks!
> 
> The VFS changes themselves look fine:
> Reviewed-by: David Disseldorp <ddiss at samba.org>

thanks!

> Given that named stream copy-chunk requests aren't macOS specific,

not from a protocol perspective, but once we go into the VFS stack the streams
stuff is very VFS specific. vfs_fruit implements special handling for two
specific streams and it's those two cases can only be tested in the vfs.fruit
testsuite.

In the testcases I added I test with three different streams: one is just some
arbitrary named stream where none of our VFS modules is expected to do any
special handling, the other two are the two subtle macOS streams, so

file:foo
file:AFP_AfpInfo
file:AFP_Resource

Generally these days the vfs.fruit tests test all relevant permutations of
streams modules (xattr vs depot), with and without vfs_fruit intercepting the
macOS streams. So by putting the testcases in vfs.fruit we effectively test all
cases.

Nevertheless I agree that it makes sense to add a basic test to smb2.ioctl
itself. I will post an updated patchset.

> please add the generic tests to the smb2.ioctl suite - Is there any
> reason why the suite doesn't currently run against the vfs_fruit
> shares?

That wouldn't increase test coverage because smb2.ioctl doesn't use macOS
specific streams.

-slow



More information about the samba-technical mailing list