Fuse on OpenBSD

[Hugo Osvaldo Barrera]

About a month ago, I followed up on tech@ that some fuse support had
been merged into the kernel, but disable by default.
(By the way, congrats and thanks to the devs for that! :D)

I'm wondering if there's any timeframe for this getting enabled by default
- I'd love to have fuse support, but I don't think I'm ready to void my
warranty just yet ;)

Is there more testing needed, or exactly what's necessary for it to
move forward?

On a somewhat related note; might this mean we might be able to port
fuse drivers (like aufs) into BSD? :D

Thanks,

--
Hugo Osvaldo Barrera

[demime 1.01d removed an attachment of type application/pgp-signature]

[Theo de Raadt]

> About a month ago, I followed up on tech@ that some fuse support had
> been merged into the kernel, but disable by default.
> (By the way, congrats and thanks to the devs for that! :D)
>
> I'm wondering if there's any timeframe for this getting enabled by default
> - I'd love to have fuse support, but I don't think I'm ready to void my
> warranty just yet ;)
>
> Is there more testing needed, or exactly what's necessary for it to
> move forward?
>
> On a somewhat related note; might this mean we might be able to port
> fuse drivers (like aufs) into BSD? :D

Good grief.

You can enable it yourself, right now.

You can test it.  You can find bugs.  You can report them.  You can
even try to fix them.  You can communicate directly with developers
trying to bring it to fruition.

Instead, what is your mail -- is it a rah rah please enable it
tomorrow?  Is it a statement of "event if there is a major screw up
hiding, enable it tomorrow please please please rah rah rah?"

Hugo, grow up.  This is a participation community.  The process is not
opaque.  Opportunities for participation at all levels are highly
visible.  Participate in development, to your own form.

The email you sent above is not a form of participation.  It is at
the level of "fanboy".

[Eric Oyen]

Theo,
Don't you just love it when folks ask questions they already know the answers to?

Still, FUSE is a wonderful idea. It certainly would make OpenBSD more versatile (and even allow it to wend its way further into both the user and corporate market segments.

anyway, hope you are having a nice summer up there (its roasting here at or above 115).

keep cool and don't let the buggers get you down. :)

-Eric

On Jul 3, 2013, at 5:55 PM, Theo de Raadt wrote:

>> About a month ago, I followed up on tech@ that some fuse support had
>> been merged into the kernel, but disable by default.
>> (By the way, congrats and thanks to the devs for that! :D)
>>
>> I'm wondering if there's any timeframe for this getting enabled by default
>> - I'd love to have fuse support, but I don't think I'm ready to void my
>> warranty just yet ;)
>>
>> Is there more testing needed, or exactly what's necessary for it to
>> move forward?
>>
>> On a somewhat related note; might this mean we might be able to port
>> fuse drivers (like aufs) into BSD? :D
>
> Good grief.
>
> You can enable it yourself, right now.
>
> You can test it.  You can find bugs.  You can report them.  You can
> even try to fix them.  You can communicate directly with developers
> trying to bring it to fruition.
>
> Instead, what is your mail -- is it a rah rah please enable it
> tomorrow?  Is it a statement of "event if there is a major screw up
> hiding, enable it tomorrow please please please rah rah rah?"
>
> Hugo, grow up.  This is a participation community.  The process is not
> opaque.  Opportunities for participation at all levels are highly
> visible.  Participate in development, to your own form.
>
> The email you sent above is not a form of participation.  It is at
> the level of "fanboy".

[Theo de Raadt]

> Still, FUSE is a wonderful idea. It certainly would make OpenBSD
> more versatile (and even allow it to wend its way further into both
> the user and corporate market segments.

So we should enable it right now, today, when it is brand new code?

Skip the testing period?  Start from go, and immediately assume
all the potential downsides?

Eric, nice try, but you are an idiot.

[Eric Oyen]

Did I say that it had to be run today? Funny, I only remember remarking that its a wonderful idea.

As for my being an idiot, the jury is still out on that one. :) I know that this is your way to motivate others into doing for themselves. There are better ways to do this, but you are you and I am what I am. :)

Anyway, thanks for the motivation. :)

-eric

On Jul 3, 2013, at 6:18 PM, Theo de Raadt wrote:

>> Still, FUSE is a wonderful idea. It certainly would make OpenBSD
>> more versatile (and even allow it to wend its way further into both
>> the user and corporate market segments.
>
> So we should enable it right now, today, when it is brand new code?
>
> Skip the testing period?  Start from go, and immediately assume
> all the potential downsides?
>
> Eric, nice try, but you are an idiot.

[opendaddy]

Why do we need FUSE anyway?

O.D.

On 4. juli 2013 at 2:10 AM, "eric oyen" <[hidden email]> wrote:

>
>Did I say that it had to be run today? Funny, I only remember
>remarking that its a wonderful idea.
>
>As for my being an idiot, the jury is still out on that one. :) I
>know that this is your way to motivate others into doing for
>themselves. There are better ways to do this, but you are you and
>I am what I am. :)
>
>Anyway, thanks for the motivation. :)
>
>-eric
>
>On Jul 3, 2013, at 6:18 PM, Theo de Raadt wrote:
>
>>> Still, FUSE is a wonderful idea. It certainly would make OpenBSD
>>> more versatile (and even allow it to wend its way further into
>both
>>> the user and corporate market segments.
>>
>> So we should enable it right now, today, when it is brand new
>code?
>>
>> Skip the testing period?  Start from go, and immediately assume
>> all the potential downsides?
>>
>> Eric, nice try, but you are an idiot.

[Brad Smith-14]

On 03/07/13 11:07 PM, [hidden email] wrote:
> Why do we need FUSE anyway?

To be able to utilize FUSE based filesystems.

--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

[Johan Beisser]

On Jul 3, 2013, at 20:23, Brad Smith <[hidden email]> wrote:

> On 03/07/13 11:07 PM, [hidden email] wrote:
>> Why do we need FUSE anyway?
>
> To be able to utilize FUSE based filesystems.
>

Fuse is a terrible hack.

But, a useful one that solves all kinds of problems.

Sent form my iFoe.

[Hugo Osvaldo Barrera]

On 2013-07-03 18:55, Theo de Raadt wrote:
> > About a month ago, I followed up on tech@ that some fuse support had
> > been merged into the kernel, but disable by default.
> > (By the way, congrats and thanks to the devs for that! :D)
> >
> > I'm wondering if there's any timeframe for this getting enabled by
default

> > - I'd love to have fuse support, but I don't think I'm ready to void my
> > warranty just yet ;)
> >
> > Is there more testing needed, or exactly what's necessary for it to
> > move forward?
> >
> > On a somewhat related note; might this mean we might be able to port
> > fuse drivers (like aufs) into BSD? :D
>
> Good grief.
>
> You can enable it yourself, right now.
>
> You can test it.  You can find bugs.  You can report them.  You can
> even try to fix them.  You can communicate directly with developers
> trying to bring it to fruition.

Well yeah, and that's basically the intention of the emails; an attempt
to communicate with the devs. I do feel it is slightly OT for tech@

>
> Instead, what is your mail -- is it a rah rah please enable it
> tomorrow?  Is it a statement of "event if there is a major screw up
> hiding, enable it tomorrow please please please rah rah rah?"
>

On the contrary, I'm not demanding it be enabled or tested right now; I'm
legitimately curious about's it's status, and wondering how close it is
to completition, how safe it is to use it for everyday use, etc. I've
no issue waiting either.

> Hugo, grow up.  This is a participation community.  The process is not
> opaque.  Opportunities for participation at all levels are highly
> visible.  Participate in development, to your own form.

Well, I did say "what's necessary for it to move forward?". I was being
quite sincere about my question. If the reply is "we think it's ok,
but just need more real-world testing", then I know I can use it. Maybe
the reply would have been "it breaks ocasionally and corrupts your stuff".

>
> The email you sent above is not a form of participation.  It is at
> the level of "fanboy".
>

Let me apologize if this sounds like a "please enable it" email. It
wasn't the intention and I was being quite sinciere about what I meant.

Anyway, I'll enable it on one of my laptops, and send any feedback I
can come across.

Cheers,

--
Hugo Osvaldo Barrera

[demime 1.01d removed an attachment of type application/pgp-signature]

[Henning Brauer]

* [hidden email] <[hidden email]> [2013-07-04 05:09]:
> Why do we need FUSE anyway?

it's a firewall between filesystem code written by people who
shouldn't write filesystem code and our kernel.

--
Henning Brauer, [hidden email], [hidden email]
BS Web Services, http://bsws.de, Full-Service ISP
Secure Hosting, Mail and DNS Services. Dedicated Servers, Root to Fully Managed
Henning Brauer Consulting, http://henningbrauer.com/

[Theo de Raadt]

> * [hidden email] <[hidden email]> [2013-07-04 05:09]:
> > Why do we need FUSE anyway?
>
> it's a firewall between filesystem code written by people who
> shouldn't write filesystem code and our kernel.

not really.

it is a simpler to understand interface, than the other userland
interface which already provides this functionality.

that other interface is the RPC one that amd(8) uses it.

FUSE is simpler.  as a result of being simpler, sure, it is a
bit of a firewall.

but henning, you just used the word firewall.  you're going to be
mocked forever.

[Henning Brauer]

* Theo de Raadt <[hidden email]> [2013-07-04 20:19]:
> but henning, you just used the word firewall.  you're going to be
> mocked forever.

firewall? me? I write packet filter code :)

--
Henning Brauer, [hidden email], [hidden email]
BS Web Services, http://bsws.de, Full-Service ISP
Secure Hosting, Mail and DNS Services. Dedicated Servers, Root to Fully Managed
Henning Brauer Consulting, http://henningbrauer.com/

[Martin Schröder]

2013/7/4 Henning Brauer <[hidden email]>:
> * [hidden email] <[hidden email]> [2013-07-04 05:09]:
>> Why do we need FUSE anyway?
>
> it's a firewall between filesystem code written by people who

It's also a firewall for licenses.

Best
   Martin