On 2012-10-19 at 10:14 -0700, Jeremy Allison wrote: > On Fri, Oct 19, 2012 at 06:57:03PM +0200, Stefan Metzmacher wrote: > > The branch, master has been updated > > via 7d05ff7 s3:smbd:durable: store stat info for disconnected > > durables and check at reconnect. > > via a6726f9 s3:open_files.idl: add stat-info to > > vfs_default_durable_cookie. > > via 660e90d s3:smbd/durable: remove an extra blank line from > > vfs_default_durable_reconnect() > > from 24d225f s3:smbd/durable: add write_time specific stuff to > > vfs_default_durable_cookie > > > > http://gitweb.samba.org/?p=samba.git;a=shortlog;h=master > > > > > > - Log ----------------------------------------------------------------- > > commit 7d05ff77264cf48d1ee82ca19c8c7e6951e64cf5 > > Author: Michael Adam <[email protected]> > > Date: Fri Oct 19 16:00:05 2012 +0200 > > > > s3:smbd:durable: store stat info for disconnected durables and check at > > reconnect. > > > > At durable reconnect, we check the stat info stored in the cookie > > against the > > current stat data from the file we just opened. If any detail differs, > > we deny > > the durable reconnect, because in that case it is very likely that > > someone > > opened the file while the handle was disconnected, which has to be > > interpreted > > as an oplock break. > > Can you guys split these checks out into a static function > that you call once from vfs_default_durable_reconnect() ? > > Putting these checks inside that function makes it a million > lines long and hard to read :-).
/me goes counting lines and major culprits in various even longer functions like open_file_ntcreate()... ;-) Seriously: Certainly, we can do that. Do you mean a single function of each of the checks, or would you be fine with one function for all the stat detail checks? Cheers - Michael
pgpUPLlVGDpLn.pgp
Description: PGP signature
