Project information - SMBv2+

CerebrosuS CerebrosuS at
Thu Dec 31 08:29:20 UTC 2020

 From the starting point I cannot imagine what might be the optimum in
investing time and fitting all needs of FreeBSD folks. Cause for now my
knowledge in samba/network and FreeBSD os-level programming is limited.
This will come with time while working on it. And it might be a question
the community should answer like:

Should we optimize a third party smb client library or should the smbfs
FreeBSD module be extended (or rewritten ... whatever)?

I don't know how such a question gets its answer in the FreeBSD community.

 From my point of view: an external library might have better future
support, cause developer of many os are working on it. But I would
prefer a fixed solution inside FreeBSD cause it is more easy to handle
and once implemented less influenced by outside opinions from evolution
of other operating systems.

Am 31.12.20 um 00:13 schrieb Alan Somers:
> I notice that smbnetfs is still using libfuse-2.  libfuse-3 uses a newer
> protocol, with more features and better performance.  If you want to help
> FreeBSD's SMB situation, your time might be more productively spent by
> upgrading smbnetfs to libfuse-3.
> -Alan
> On Wed, Dec 30, 2020 at 4:02 PM CerebrosuS <CerebrosuS at> wrote:
>> My experience with the fuse smbnetnfs is about a week old on FreeBSD 12.2.
>>> Am 30.12.2020 um 23:56 schrieb Daniel Ebdrup Jensen <debdrup at
>>> :
>>> On Wed, Dec 30, 2020 at 10:16:44PM +0100, Miroslav Lachman wrote:
>>>>> On 30/12/2020 20:24, CerebrosuS wrote:
>>>>> Am 30.12.20 um 20:05 schrieb Miroslav Lachman:
>>>>>> On 30/12/2020 18:57, CerebrosuS wrote:
>>>>>>> Hello at all,
>>>>>>> the community and developer at FreeBSD seem to know, that SMBv1 for
>>>>>>> clients is nearly over and that the included mount_smbfs doesn't
>> support
>>>>>>> newer versions. So good, so far...
>>>>>>> So I can find multiple information about the situation, but no clear
>>>>>>> path on how FreeBSD community and developer will go on to solve this
>>>>>>> missing function. (Just got the information on:
>> )
>>>>>>> This is what I am asking:
>>>>>>> - Is there a project existing for solving this problem (with whatever
>>>>>>> target)?
>>>>>>> - What is the way to go in future? Extend mount_smbfs or support the
>>>>>>> fuse-smbnetfs part to be stable and fast like mount_smbfs (buggy and
>>>>>>> laggy here)?
>>>>>>> - Who is mainly working on it, if a project already exist?
>>>>>>> I'am just interested, cause of not finding such information clearly.
>> Is
>>>>>>> there maybe a general project management list / team to see what
>>>>>>> projects are going on in whatever state?
>>>>>>> I am a hobby developer mainly coming from chemical engineering side,
>>>>>>> having some time to help. I've already written some cross platform
>>>>>>> software but never related to network or on os-level. So I am
>> motivated
>>>>>>> to invest some time in getting stuff into FreeBSD, but for me, there
>> is
>>>>>>> a lack on information (see above).
>>>>>>> Thank you in advance for information and help.
>>>>>> I was involved in the thread linked by Gleb. AFAIK nothing changed
>> from
>>>>>> that time. I tried something from ports but it has more problems
>> (shares
>>>>>> cannot be mounted on boot like mount_smbfs does).
>>>>>> If somebody has time and skills to try to bring SMBv2 or v3 to FreeBSD
>>>>>> then Apple or Solaris sources is good start. The both were using the
>>>>>> same mount_smbfs (v1) as FreeBSD so one can check their sources and
>> see
>>>>>> how they evolve to v2 / v3.
>>>>> They are both using exactly the same source code as a starting point
>> and
>>>>> extend it (or rewrite it) to SMBv2?
>>>> They are based on the ported code. Apple Mac OS X and Solaris have
>> different kernel so they needed modified port of the same code as was in
>> FreeBSD back in the days (there is the same copyright header). Apple
>> sources or Solaris sources cannot be used directly on FreeBSD but some
>> skilled developer can look in to those sources to see their evolution. But
>> as was already noted v2 and v3 are very different from v1. It will be hard
>> to port but not impossible.
>>>> Current solutions in ports (fusefs) are almost useless in server
>> environment.
>>>> Miroslav Lachman
>>>> _______________________________________________
>>>> freebsd-hackers at mailing list
>>>> To unsubscribe, send any mail to "
>> freebsd-hackers-unsubscribe at"
>>> Hi folks,
>>> Assuming that the reasons for not using fuse in a server environment are
>> related primarily to performance and that the implementation that was in
>> base used to be quite out-of-date, has this at all been reevalulated since
>> a new version was merged? [1]
>>> Yours,
>>> Daniel Ebdrup Jensen
>>> [1]:
>> _______________________________________________
>> freebsd-hackers at mailing list
>> To unsubscribe, send any mail to "freebsd-hackers-unsubscribe at"
> _______________________________________________
> freebsd-hackers at mailing list
> To unsubscribe, send any mail to "freebsd-hackers-unsubscribe at"

More information about the freebsd-hackers mailing list