THM 2005-05-19 Summary

Notes taken by kuroneko

Status reports:



  • It "takes the keys of the users on the other nodes and uses them for the tickets."
  • All done except for factotum
  • waiting on to come back online
  • Authsrv + backend are working.
  • Code available in
  • Still Private Key based
  • rsc9 suggests that this should be replaced to use public keys

Modularised Installer.

dtLinux v2.6 with v9fs.


(via boyd) OzInferno is done but release on hold until docs are writen, "doesn't need much doc 'cause the vita stuff will largely do".

Will IRC first week of june when in paris.


(See ericvh for details)


Summer projects:

Interested in:


(See kuroneko for details)




GDIAZ stalled due to workload.

Will talk to 20h about cross-domain auth.


Bad hand slowing work.

Compilers fix for to handle large macro expansions:

  • "99.9% done. the 0.1% is subtle"
  • cmd/8c and cmd/cc are on contrib/boyd


Running tip9ug servers, "everyone behaves good on mordor."

Talking about multi authdom in tip9ug.

vt3 adds: "working on japanese fonts, finding security issues with venti, just being modest"


Interested in USB work, modified usbd to automatically load usbmouse when a mouse gets plugged in.


Cifs, adding ntlmv2 auth, message signing, dfs support.

Seperating libasn1 from libsec, leading up to ldapfs, snmpfs and krb in factotum.

Chatfs as a /net service providing msim, yahoo, jabber, etc.

All incomplete - published when ready.


Working on usb storage improvements soon.

  • usb itself needs work as its too slow to be really usable for storage
  • many devices don't boey the spec either
  • poor performance attributed to "only being able to send one packet per usb frame"
  • forsyth may be looking at usb as well
  • usb booting should work if supported by bios(not sure about 9load)


Port to Xen3.0

  • goes really slowly

Working with vic zandy'x xcpu:

  • "xcpu is really really nice"
  • server for remote execution
  • not released yet
  • for lightweight cluster nodes.
  • similar to linux bproc stuff


  • enet driver has issue

Try to work with k8 once jmk releases initial kernel code.


Wkj has almost working sshv2.

Working on venti rewrite:

  • almost in a releaseable state
  • jmk + rsc9 will be deploying a new venti to test it
  • considerably faster than old venti
  • p9p release first, then plan9 native
  • available in the CVS as p9p


Pointed out buffer overflow vulnerability report: rsc9 said he would look into it if he had some time, but asked for volunteers to do it.


(See uriel for details)

Hacking on wiki and docs.