gwillcox-r7
commented
2 years ago Before Patch:
~/git/ruby_smb │ master sudo ruby examples/file_server.rb --path /var/public --share public --username MSFLAB\\tester --password letsTest123!
D, [2022-04-11T14:13:37.275411 #17237] DEBUG -- : Adding disk share: public
server is running
received connection
#<Thread:0x00007f52d41ccfe0 /home/gwillcox/git/ruby_smb/lib/ruby_smb/server.rb:56 run> terminated with exception (report_on_exception is true):
/root/.gem/gems/bindata-2.4.10/lib/bindata/choice.rb:115:in `instantiate_choice': selection '8' does not exist in :choices for obj.negotiate_context_list[3].data (IndexError)
from /root/.gem/gems/bindata-2.4.10/lib/bindata/choice.rb:109:in `current_choice'
from (eval):2:in `do_read'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/struct.rb:139:in `block in do_read'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/struct.rb:139:in `each'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/struct.rb:139:in `do_read'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/array.rb:318:in `block in do_read'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/array.rb:315:in `loop'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/array.rb:315:in `do_read'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/struct.rb:139:in `block in do_read'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/struct.rb:139:in `each'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/struct.rb:139:in `do_read'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/base.rb:147:in `block in read'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/base.rb:253:in `start_read'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/base.rb:145:in `read'
from /root/.gem/gems/bindata-2.4.10/lib/bindata/base.rb:21:in `read'
from /home/gwillcox/git/ruby_smb/lib/ruby_smb/generic_packet.rb:44:in `read'
from /home/gwillcox/git/ruby_smb/lib/ruby_smb/server/server_client/negotiation.rb:19:in `handle_negotiate'
from /home/gwillcox/git/ruby_smb/lib/ruby_smb/server/server_client.rb:161:in `block in run'
from /home/gwillcox/git/ruby_smb/lib/ruby_smb/server/server_client.rb:153:in `loop'
from /home/gwillcox/git/ruby_smb/lib/ruby_smb/server/server_client.rb:153:in `run'
from /home/gwillcox/git/ruby_smb/lib/ruby_smb/server.rb:56:in `block (2 levels) in run'And on server we would just hang before getting this message:
C:\Users\*****>type \\172.31.117.32\public\cirt-default-usernames.txt
The semaphore timeout period has expired.
C:\Users\****>
This fixes a couple of small bugs that were preventing Windows 11 clients from reading files on the SMB server.
The main issue was that the new negotiate contexts could not be parsed. This meant the server couldn't negotiate the connection. The contexts now have a default string field that just consumes the data to keep things aligned but doesn't process it. This allows the negotiation to be complete without needing to be aware of all the negotiate contexts which should make it more future proof.
Additionally, the logic used to negotiate the encryption was improved to select the first mutually supported algorithm instead of only selecting the first algorithm if it's supported. This should also help make things more future-proof should additional encryption algorithms be added.
Finally, AES-256 support was fixed for the server. The key generation needed to be updated to account for the encryption algorithm so the bit length would be set correctly.
Testing
Use the example file server and a Windows 11 client.
ruby examples/file_server.rb --path /var/public --share public --username MSFLAB\\smcintyre --password Password1!/var/public, shares them using the share name "public", and adds an account with which to access themtype \\192.168.159.128\public\test.txttest.txtfile just fine.Fixes #205