[cap-talk] Petname definition: MinorCtkrFs, are these petnames?

Rob Meijer capibara at xs4all.nl
Thu Jan 29 05:38:40 EST 2009


With the resent discussion on the definition of petnames, I started to doubt
if what I use and call petnames in MinorCtkrFs are actualy petnames.

Let me tell what MinorCtkrFs would do in this context.

Each unattenuated MinorCapFs gives access to a directory where attenuation
(caretaker) nodes can be created for this node and its namespace.

 attr -g ctrl <UNATTENUATED NODE SPARSECAP>

returns

  <ATTENUATION DIRECTORY NODE SPARSECAP>

Within this directory a process can create a 'petnamed' caretaker node

  mknod  <ATTENUATION DIRECTORY NODE SPARSECAP>/<PETNAME>

It can set the initial attenuation:

  chmod 0400 <ATTENUATION DIRECTORY NODE SPARSECAP>/<PETNAME>

And get an attenuated node sparsecap from it:

    attr -g attn <ATTENUATION DIRECTORY NODE SPARSECAP>/<PETNAME>

That returns

  <ATTENUATED SPARSECAP>

It can later use the petname to revoke the initialy given privileges.

  chmod 0000 <ATTENUATION DIRECTORY NODE SPARSECAP>/<PETNAME>


Does my usage of the 'petname' term concur with how others feel petnames
are defined, or am I stretching the petname concept beyond its meaning?

Rob



More information about the cap-talk mailing list