writeups for hotplugging S/390 chandev, PNP docking stations ?
Quoted 4 lines
I normally use the "Device Drivers and Installation Commands" document when I want to refresh my memory on a particular command for a particular type of device.
(http://www10.software.ibm.com/developerworks/opensource/linux390/docu/l390dd07.pdf)
Luckily Acrobat has a "find", and the hotplug discussion starts at the front of chapter 8. (It's not in the index or TOC ... :)
The hotplug script there doesn't follow the conventions of the linux-hotplug project; that'd involve "/etc/hotplug/chandev.agent" as the script doing the magic. Which means that the network hotplug events will get ignored ... why is that? Seems to me more natural to do network administration through network hotplug events, not chandev ones! (The /sbin/hotplug script normally delegates the work to /etc/hotplug/$1.agent ...)
Quoted 2 lines
This seems pretty clear and reasonably well-written. If Dave finds this more acceptable, then I guess we're done.
I believe there's a bunch left unsaid there about what the device names indicate. I'd hope that eth0/eth1/eth2 are all Ethernet, and that tr0 is token ring (it still lives? :), but I have no clue what "ctc0" would be or, in general, how and what those names are. Wouldn't disk drives show up there?
And tape backup units, etc?
What I'll do is stick a pointer to it, and update the linux-hotplug webpage about "chandev" to present some of that info. My understanding is that the management of mainframes has its own families of idiosyncracies, developed over decades, that it's not worth my time to master ... :) For now I'm happy just to keep the subsystem information current on the web
- Dave
_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel
Quoted 7 lines
The hotplug script there doesn't follow the conventions of the linux-hotplug project; that'd involve "
/etc/hotplug/chandev.agent" as the script doing the magic. Which means that the network hotplug events will get ignored ... why is that? Seems to me more natural to do network administration through network hotplug events, not chandev ones! (The/sbin/hotplugscript normally delegates the work to/etc/hotplug/$1.agent...)
I can't speak to this,
but I'm sure the S/390 subdivision of the Linux community would like to see it fixed.
Quoted 3 lines
I believe there's a bunch left unsaid there about what the device names indicate. I'd hope that eth0/eth1/eth2 are all Ethernet,
Yes.
Quoted 1 line
and that tr0 is token ring (it still lives? :), but I have
Token ring, yes. Still lives in some corners.
Quoted 3 lines
no clue what "ctc0" would be or, in general, how and what those names are. Wouldn't disk drives show up there? And tape backup units, etc?
CTC is "channel-to-channel" which is a point-to-point thing conceptually like two SCSI chains having a bridge between them.
Mainframe also has a network interface known as IUCV (iucv0, iucv1, etc), but that has no representation in real hardware so I don't know if it would really fit in the "chandev" suite.
Quoted 6 lines
What I'll do is stick a pointer to it, and update the linux-hotplug webpage about "chandev" to present some of that info. My understanding is that the management of mainframes has its own families of idiosyncracies, developed over decades, that it's not worth my time to master ... :) For now I'm happy just to keep the subsystem information current on the web
Thanks for keeping it current.
I for one very much value the commonality
and hope that S/390 will be able to use most of this infrastructure rather than re-invent it.
_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel