[SCM] Samba Shared Repository - branch master updated

[Jeremy Allison]

On Tue, Mar 23, 2010 at 07:45:27PM +0100, Stefan (metze) Metzmacher wrote:
> Jeremy Allison schrieb:
> > On Mon, Mar 22, 2010 at 11:26:50AM -0500, Stefan Metzmacher wrote:
> >> The branch, master has been updated
> >>        via  cad0c00... s3: Implement an asynchronous echo responder process
> >>        via  0c77e3a... s3:smbd: don't allow SMB2 if the async echo handler is active
> >>        via  5a069f7... s3:smbd: disable SMB encryption when the echo handler is active
> >>        via  d663b4c... s3:smbd: disallow readbraw and writebraw if the echo handler is active
> >>        via  fbf112b... s3:smbd: disable sendfile if the echo handler is active
> >>        via  453e6af... s3:smbd: don't use recvfile if the echo handler is active
> >>        via  79e5e3d... s3:smbd: setup a shared memory area for the signing state
> >>        via  44d655b... s3:smbd: add echo handler information to struct smbd_server_connection
> >>        via  752240c... s3:param: add "async smb echo handler" option
> >>        via  b2c107f... s3:smbd: pass down trusted_channel via receive_smb_talloc()
> >>        via  1e7086e... s3:smbd: let reply_readbraw_error use the locked socket
> >>        via  c1653e3... s3:smbd: send keepalive packets under the socket lock
> >>        via  977aa66... s3:smbd: smbd_[un]lock_socket() while accessing the socket to the client
> >>        via  8de8554... s3:smbd: add smbd_[un]lock_socket() dummies
> >>        via  0b7da43... s3:smbd: add an option to skip signings checks srv_check_sign_mac for trusted channels
> >>        via  048c919... s3:libsmb: add a smb_signing_init_ex() function
> >>        via  01f2c02... lib/util: add allocate_anonymous_shared()
> > 
> > Metze, this is really nice code ! I'm still digesting this....
> > 
> > One point. If we make the smbd_[un]lock socket calls reference
> > counted so they're recursive I think we can still allow sendfile().
> > 
> > Thoughts ?
> 
> sendfile() wouldn't be the hardest problem.
> It's isolated in smbd/reply.c
> 
> But recvfile() is the problem that would be very, very hard to fix
> and we can create really ugly bugs if we allow it.

Yep, I agree, that's why I didn't suggest that one :-).

Jeremy