Bug 31565

Summary: USB devices cripple SMBFS
Product: Base System Reporter: brad <brad>
Component: kernAssignee: freebsd-bugs (Nobody) <bugs>
Status: Closed FIXED    
Severity: Affects Only Me    
Priority: Normal    
Version: 4.4-STABLE   
Hardware: Any   
OS: Any   

Description brad 2001-10-28 20:20:03 UTC
Compile the following into your kernel:

options         NETSMB                  #SMB/CIFS requester
options         NETSMBCRYPTO            #encrypted password support for
SMB

# mchain library. It can be either loaded as KLD or compiled into kernel
options         LIBMCHAIN               #mbuf management library

# Kernel side iconv library
options         LIBICONV

options         SMBFS

Plug a Microsoft USB mouse in and try to mount an SMB share after the USB
mouse has detected. This will return a "no buffer space available"
message. All other networking functions are unimpaired.

Tested this with all but umass compiled directly into the kernel, and no
modules. I will be able to do further tests on monday, e.g. does an
existing share stop responding when the mouse is plugged in? Does SMBFS work if the USB mouse is plugged in at system startup?

Disconnecting the mouse does not solve the issue; shares remain
non-mountable.

Using 4.4-STABLE cvsuped to Friday, and smbfs userland version 1.4.1, from
ports, an EtherExpress Pro 10/100, and both Windows 98 and Windows 2000
AS servers.

Let me know if there is anything more I should look for and report.

Brad

Fix: 

Don't use USB if you need to access SMBFS shares.
How-To-Repeat: Compile both USB and SMBFS fully into the kernel, plug in a USB device, and try to mount a share using mount_smbfs. I will test out modularization and see if this changes matters, but the flaw exists nonetheless in this configuration.
Comment 1 Sheldon Hearn freebsd_committer freebsd_triage 2001-12-27 12:41:59 UTC
State Changed
From-To: open->feedback

Have you asked around on the freebsd-stable mailing list to see whether 
anyone else is seeing this behaviour?  It really sounds like a hardware 
conflict, but that's just a guess.
Comment 2 Kris Kennaway freebsd_committer freebsd_triage 2002-01-06 03:59:40 UTC
State Changed
From-To: feedback->closed

Feedback timeout