Andreas, Could you send me the network capture for the test below ? Also, if I can run your test for a repro, it will be easier for me the debug the behavior. Is your test a new test in smbtorture ?
Thanks! Hongwei -----Original Message----- From: Andreas Schneider [mailto:m...@cynapses.org] Sent: Tuesday, May 25, 2010 9:45 AM To: Hongwei Sun Cc: p...@tridgell.net; cifs-proto...@samba.org; MSSolve Case Email Subject: Re: [REG:110052070326491] [cifs-protocol] MS_RRP: Question on Symbolic Links On Monday 24 May 2010 23:58:34 Hongwei Sun wrote: > Andreas, > > When you open the key with REG_OPTION_LINK flag set, the server will > return the handle to the source key. With a valid handle, client should > be able to update the target of the symbolic link by changing the value of > SymbolicLinkValue and also delete the key that is referenced by the > handle. As explicitly pointed out in 3.1.1.11, the SymbolicLinkValue for > target link should contain Fully Qualified Name(3.1.1.1.1), which is > something like HKEY_LOCAL_MACHINE\SYSTEM\MountedDevices. It is not in > the kernel mode string such as \registry\machine\system\MountedDevice. I've done some testing with FQN as link value against Windows 2008 here. If I set the 'SymbolicLinkValue' to HKEY_LOCAL_MACHINE\Software\Samba\torture_test\symlink_destination and then do a normal OpenKey (follow the symlink) it fails with WERR_BADFILE. If I set the 'SymbolicLinkValue' to '\Registry\MACHINE\Software\Samba\torture_test\symlink_destination' I can follow the symlink. I think the documentation is wrong here. At least for Windows 2008. I haven't tested it against other Windows versions yet. Best regards, -- andreas _______________________________________________ cifs-protocol mailing list cifs-protocol@cifs.org https://lists.samba.org/mailman/listinfo/cifs-protocol
