From mboxrd@z Thu Jan 1 00:00:00 1970 MIME-Version: 1.0 In-Reply-To: <82348218-84c9-47c7-800f-392261157389@googlegroups.com> References: <89ec9d50-9736-4d93-985f-bf32e6fd9247@n2g2000prj.googlegroups.com> <4169f20d-b962-455b-8afe-8aa52db190ab@googlegroups.com> <82348218-84c9-47c7-800f-392261157389@googlegroups.com> Date: Tue, 4 Nov 2014 17:13:54 -0800 Message-ID: From: Skip Tavakkolian To: Chris McGee Content-Type: multipart/alternative; boundary=001a11417c4420b14e0507124b25 Cc: Fans of the OS Plan 9 from Bell Labs <9fans@9fans.net>, golang-nuts Subject: Re: [9fans] [go-nuts] Re: 9p protocol go implementation Topicbox-Message-UUID: 230d872a-ead9-11e9-9d60-3106f5b1d025 --001a11417c4420b14e0507124b25 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable where would PostMountSrv reside? it isn't a syscall. it is not difficult to do by hand; this version of go9p's timefs example posts itself to /srv (plus some code to fake a few unix'isms on Plan 9). there is no authentication; permissions on the /srv file determine if a user can mount it: https://github.com/9nut/plan9/tree/master/go9p_timefs a package that wraps the factotum-to-app rpc protocol (like libauth) would be useful. -Skip On Mon, Nov 3, 2014 at 7:04 PM, wrote: > Thanks, I think I'll have to do a bit more reading to understand this. > I'll check out ramfs and 9pcon. > > Is this simpler if written in C using "postmountsrv" with a mount point? > I'm assuming that it doesn't require a tcp port and explicit authenticati= on > handling using libauth on both ends. > > If so, then I wonder why postmountsrv is not exposed via the Go 9P > libraries? > > Chris > > On Monday, November 3, 2014 4:29:10 AM UTC-5, Skip wrote: >> >> short version: you need libauth in Go (or start the go9p client/server b= y >> C programs that do the auth). >> >> 9P facilitates authentication (but doesn't define or dictate the method)= . >> intro(5), auth(2) and factotum(4) will be helpful. basically Tauth is us= ed >> to request a fid to negotiate authentication (a.k.a. afid). >> Tread's/Twrite's to afid are proxy-delivered to the factotums >> (authentication agents) of the sever and of the client by each side. onc= e >> server's factotum is convinced, the server is granted the system privile= ge >> to change its process id to the authenticated user. the client attaches >> (Tattach) to the server's namespace by providing the afid in addition to >> other parameters. tools like 'ramfs -D' and aux/9pcon are very handy fo= r >> watching 9P in action. >> >> i'm copying to 9fans; it might be a better place to continue. >> >> >> On Sun, Nov 2, 2014 at 6:27 PM, wrote: >> >>> I see that go9p supports authentication. Assuming the client and server >>> are both plan9 (even the same system), how does one hook up the OS's >>> authentication? >>> >>> Chris >>> >>> >>> On Monday, March 21, 2011 10:16:16 PM UTC-4, peterGo wrote: >>>> >>>> Mauricio, >>>> >>>> go9p - Package to write 9P clients and servers in Go >>>> http://code.google.com/p/go9p/ >>>> >>>> Peter >>>> >>>> On Mar 21, 9:57 pm, Maur=C3=ADcio CA wrot= e: >>>> > Hi, all, >>>> > >>>> > I see at page below that there exists go9p, "A 9P library in the Go >>>> > programming language, by Andrey Mirtchovski and Latchesar >>>> > Ionkov. Now part of the official Go distribution." >>>> > >>>> > http://9p.cat-v.org/implementations >>>> > >>>> > I can't find any implementation of 9p at this page, though, which, I >>>> > believe, is the official list of current standard go packages: >>>> > >>>> > http://golang.org/pkg >>>> > >>>> > Is there really a 9p implementation in the official go distribution? >>>> > >>>> > Thanks, >>>> > >>>> > Maur=C3=ADcio >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "golang-nuts" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to golang-nuts...@googlegroups.com. >>> For more options, visit https://groups.google.com/d/optout. >>> >> >> -- > You received this message because you are subscribed to the Google Groups > "golang-nuts" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to golang-nuts+unsubscribe@googlegroups.com. > For more options, visit https://groups.google.com/d/optout. > --001a11417c4420b14e0507124b25 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
where would PostMountSrv reside? it isn't a sysca= ll.

it is not difficult to do by hand; this versio= n of go9p's timefs example posts itself to /srv (plus some code to fake= a few unix'isms on Plan 9).=C2=A0there is no authentication; permissio= ns on the /srv file determine if a user can mount it:

<= div>
a package that wraps the= factotum-to-app rpc protocol (like libauth) would be useful.

-Skip

On Mon, Nov 3, 2014 at 7:04 PM, <= ;newton688@gmail.c= om> wrote:
Thanks, I thi= nk I'll have to do a bit more reading to understand this. I'll chec= k out ramfs and 9pcon.

Is this simpler if written in C u= sing "postmountsrv" with a mount point? I'm assuming that it = doesn't require a tcp port and explicit authentication handling using l= ibauth on both ends.

If so, then I wonder why post= mountsrv is not exposed via the Go 9P libraries?

C= hris

On Monday, November 3, 2014 4:29:10 AM UTC-5, Skip wrote:=
short version: you need liba= uth in Go (or start the go9p client/server by C programs that do the auth).=

9P facilitates authentication (but doesn't def= ine or dictate the method). intro(5), auth(2) and factotum(4) will be helpf= ul. basically Tauth is used to request a fid to negotiate authentication (a= .k.a. afid). Tread's/Twrite's to afid are proxy-delivered to the fa= ctotums (authentication agents) of the sever and of the client by each side= . once server's factotum is convinced, the server is granted the system= privilege to change its process id to the authenticated user. =C2=A0the cl= ient attaches (Tattach) to the server's namespace by providing the afid= in addition to other parameters. tools like 'ramfs =C2=A0-D' and a= ux/9pcon are very handy for watching 9P in action.

i'm copying to 9fans; it might be a better place to continue.


On= Sun, Nov 2, 2014 at 6:27 PM, <newt...@gmail.com> wrote:
I se= e that go9p supports authentication. Assuming the client and server are bot= h plan9 (even the same system), how does one hook up the OS's authentic= ation?

Chris


On Monday, March 21, 2011 = 10:16:16 PM UTC-4, peterGo wrote:
Mauricio,

go9p - Package to write 9P clients and servers in Go
http://cod= e.google.com/p/go9p/

Peter

On Mar 21, 9:57=C2=A0pm, Maur=C3=ADcio CA <mauricio.antu...@gmail= .com> wrote:
> Hi, all,
>
> I see at page below that there exists go9p, "A 9P library in = the Go
> programming language, by Andrey Mirtchovski and Latchesar
> Ionkov. Now part of the official Go distribution."
>
> =C2=A0 =C2=A0http://9p.cat-v.org/implementations
>
> I can't find any implementation of 9p at this page, though, wh= ich, I
> believe, is the official list of current standard go packages:
>
> =C2=A0 =C2=A0h= ttp://golang.org/pkg
>
> Is there really a 9p implementation in the official go distributio= n?
>
> Thanks,
>
> Maur=C3=ADcio

--
You received this message because you are subscribed to the Google Groups &= quot;golang-nuts" group.
 To unsubscribe from this group and stop receiving emails from it, send an e= mail to golang-nuts...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups &= quot;golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an e= mail to golang-nuts+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--001a11417c4420b14e0507124b25--