Bug 33016

Summary: mount_smbfs
Product: Base System Reporter: Aaron Kaplan <aaron>
Component: binAssignee: Sheldon Hearn <sheldonh>
Status: Closed FIXED    
Severity: Affects Only Me    
Priority: Normal    
Version: 4.4-STABLE   
Hardware: Any   
OS: Any   

Description Aaron Kaplan 2001-12-19 21:30:01 UTC
	mount_smbfs will not work with smbfs module unless smbfs module
        loads libmchain module.

Fix: 

see ideas on how to fix it above.
How-To-Repeat: 	do the following: compile a -STABLE kernel without LIBMCHAIN
	and without NETSMB options. Include the LIBICONV option 
	(this does not seem to compile as module, ok).

	Now mount_smbfs some filesystem.
	It will fail with error message: 
		smbfs: vfsload(smbfs): Exec format error
	then kldload libmchain
	re-try mount_smbfs command... bingo everything works again.

	As a side note: If I compile options NETSMB, option LIBMCHAIN
	and option LIBICONV statically into the kernel then 
	mount_smbfs will fail with: smbfs: vfsload(smbfs): File exists.

	Seems like the vfsload will try to load the module but if its already
	there statically - it does not detect this fact and gives up.
	I recon some detection code would be in place here.
Comment 1 ru freebsd_committer freebsd_triage 2001-12-20 10:03:07 UTC
Responsible Changed
From-To: freebsd-bugs->sheldonh

Sheldon's last commits may have fixed part of this PR.
Comment 2 Sheldon Hearn freebsd_committer freebsd_triage 2001-12-20 10:22:55 UTC
State Changed
From-To: open->feedback

1) libiconv is now available as a kernel loadable module in the latest 
-STABLE.  Please update your system and try again. 

2) mount_smbfs does what all the other mount_* utilities do with 
respect to vfsload(), so if there's a problem with vfsload(), 
it'll need to be addressed globally.  It's probably beyond the 
scope of what I can manage right now, so I'll probably close this 
PR once you've confirmed that the first issue is resolved.
Comment 3 Sheldon Hearn freebsd_committer freebsd_triage 2002-01-07 13:29:45 UTC
State Changed
From-To: feedback->closed

The smbfs-specific problem reported by the originator is resolved. 
Another problem, namely suspected weirdness in vfsload(), should 
be addressed in a new PR when more information is available.
Comment 4 Sheldon Hearn freebsd_committer freebsd_triage 2002-01-09 11:46:18 UTC
State Changed
From-To: closed->feedback

Please try out the patch in kern/33625 and let me know whether it 
solves your problem. 

Please copy your feedback to <freebsd-gnats-submit@FreeBSD.org>, 
using the subject line of this message.
Comment 5 Sheldon Hearn 2002-01-09 11:59:59 UTC
On Wed, 09 Jan 2002 03:48:02 PST, sheldonh@FreeBSD.org wrote:

> State-Changed-From-To: closed->feedback
> State-Changed-By: sheldonh
> State-Changed-When: Wed Jan 9 03:46:18 PST 2002
> State-Changed-Why: 
> Please try out the patch in kern/33625 and let me know whether it
> solves your problem.

Note that "your problem" here refers to the second part of the problem,
which I previously miscredited to a general vfsload() problem.

> Please copy your feedback to <freebsd-gnats-submit@FreeBSD.org>,
> using the subject line of this message.

This still applies. :-)

Ciao,
Sheldon.
Comment 6 Sheldon Hearn freebsd_committer freebsd_triage 2002-01-21 22:37:11 UTC
State Changed
From-To: feedback->closed

In this case, feedback was more of a formality than anything else. 
We know the problem's solved.