[PATCH] smbd: Fix snapshot query on shares with DFS enabled

[Uri Simchoni]

On 08/19/2016 11:22 AM, Uri Simchoni wrote:
> On 08/19/2016 12:51 AM, Jeremy Allison wrote:
>> On Thu, Aug 18, 2016 at 11:32:35AM -0700, Jeremy Allison wrote:
>>>
>>> We still have the problem that when we create the paths previous
>>> versions path to query the SMB1 server returns NT_STATUS_OBJECT_PATH_NOT_FOUND,
>>> due to the fact we're not setting the flags2 FLAGS2_REPARSE_PATH
>>> bit in the request when we're doing the path query. We don't
>>> currently have capabilities to do that, so I'm going to create
>>> another bug and add create fixes for all these things.
>>
>> OK, here is a patchset that fixes enumerating and accessing
>> shadow copy paths from smbclient and libsmbclient.
>>
>> It looks a bigger patchset than it is :-).
>>
>> Overview:
>>
>> 01	- Add missing FLAGS2_REPARSE_PATH flag definitions.
>> 02 - 04 - Add additional_flags2 arg to internal cli_XXX() functions
>> 	  that need to pass paths that might contain @GMT- tokens.
>> 05	- Add clistr_is_previous_version_path() function that checks
>> 	  for a valid @GMT- path.
>> 06 - 21 - For all cli_XX() functions that take a pathname, call
>> 	  clistr_is_previous_version_path() and add in the
>> 	  FLAGS2_REPARSE_PATH to the additional_flags2 parameter
>> 	  if it maches.
>>
>> 	  Only function not fixed is cli_set_ea_path(), which
>> 	  uses the sync versions of cli_trans() which I didn't
>> 	  change to add additional_flags2 arg. As this is a 'set'
>> 	  function I don't think this works with a @GMT- path.
>> 	  Patch 20 is a reminder to change cli_set_ea_path()
>> 	  to use the async cli_trans_send()/cli_trans_recv()
>> 	  if we ever need to fix this.
>>
>> 22	- Fix wrong size in cli_shadow_copy_data_send() params.
>> 23	- Harden cli_shadow_copy_data_recv() parsing in case
>> 	  of bad server. Only crash bug here, nothing worse.
>> 24	- Fix smbclient to use cli_shadow_copy_data() correctly
>> 	  over SMB1.
>>
>> Using this I can do an 'allinfo' against a Windows server
>> with shadow copies, and see them available to smbclient.
>>
>> Please review and push if happy ! Passes local samba3.blackbox.shadow_copy2
>> tests.
>>
>> Once this is in I'll do the same fix for accessing shadow
>> copies over SMB2.
>>
>> Cheers,
>>
>> 	Jeremy.
>>
> 
> Pushed with a small space-before-tab correction in "libsmb: Add uint16_t
> addtional_flags2 to cli_trans_send()"
> 
> Thanks,
> Uri.
> 
Something strange has happened - it seems to have completed the tests in
autobuild - moved to "master passed", but instead of being committed,
there's another queued job from me without me queuing it. I'll see what
happens with that job.

Uri.