Michael's musings

Sun, 28 Jan 2007

How to get Asterisk working under a XenU

Asterisk works in basic SIP mode, but you can't use any of the conference facilities, and or any of the announcements without a source of 1khz interrupts.

Normally, this comes from the zaptel driver in the kernel, if you have a stock FXO card. You can also use ztdummy on 2.4 kernels using the USB interface for timing, and on 2.6 kernels you can get 1khz timing from the RTC driver.

But, there is no RTC on Xen.

The solution is to use the ztdummy driver, but to REBUILD your XenU kernel with HZ=1000, so that it can use the stock timer interrupts.

In theory, the in-kernel HPET interface should be used instead, and I hope to adapt the code. I also expect to reduce the number of .h files, and most of the dependancies.

(Note, you need the CCITT CRC routines available in your kernel as well)

I have placed my tgz file of drivers/char/zaptel at:

http://www.sandelman.ca/software/zaptel-2.6-xenU.tgz

note that it overwrites drivers/char/Makefile and drivers/char/Kconfig. I will produce patch-o-matic and git trees soon.

I put a kernel that I'm using at:

http://www.sandelman.ca/software/vmlinuz-2.6.16.29-3.0.3b

If you use this, and it works, I'll clean it up a bit more.

The config items you need to get HZ=1000 are:

gimli-[xen/xen/xen-3.0.3_0-src/linux-2.6.16.29-xen] mcr 1062 %grep HZ .config
# CONFIG_HZ_100 is not set
# CONFIG_HZ_250 is not set
CONFIG_HZ_1000=y
CONFIG_HZ=1000
CONFIG_MACHZ_WDT=m
CONFIG_NO_IDLE_HZ=y

This is under the "Processor Type and Features" item in menuconfig, if you are looking that way. Near the bottom: "Timer frequency (100 HZ)" is the default.

1000HZ settings make the kernel a bit more intensive, and it might be that you have to also adjust the Xen scheduler at bit. I haven't done this yet, but seem to have good results.

With these changes, I can now deploy Asterisk into XenUs at a colo.

Mon, 15 Jan 2007

some notes on IBM OpenPower LPARs and VIO channels

I thought I'd share something I just learned about VIO disks by a process of experimentation, based upon some comments I found in a wiki, but couldn't confirm anyplace in documentation.

You can run:

mkvdev -vdev lv_cayenne_home -vadapter vhost9 -dev dev_cayenne_h
mkvdev -vdev lv_athabasca_sl -vadapter vhost9 -dev dev_athab_sl

I.e. set two virtual disks to be on the same host/server adapter. Linux, sees, properly:

root@cayenne:~ # dmesg | grep 'Attached scsi'
Attached scsi disk sda at scsi0, channel 0, id 1, lun 0
Attached scsi disk sdb at scsi1, channel 0, id 1, lun 0
Attached scsi disk sdc at scsi2, channel 0, id 1, lun 0
Attached scsi disk sdd at scsi3, channel 0, id 1, lun 0
Attached scsi disk sde at scsi3, channel 0, id 2, lun 0

This is a nice thing, because you can then add disks to running LPARs, and get the SCSI bus rescanned by doing:

root@cayenne:~ # echo >/sys/bus/scsi/devices/3:0:1:0/rescan

that means you can also do things like perform rescue operations on a sick LPAR's root disk by mounting it on another LPAR, and you can do this without rebooting.

In particular, if you have provisioned your VIO server with enough spare SCSI server adapters, you can easily add LPARs without restarting the VIO server.

The instructions we read suggesting creating a new VIO channel per disk, which seemed dumb. We much prefer to have a seperate LV on the VIO server per partition, because ideally we can resize them as well.

Some thoughts on the BSD license

http://www.groklaw.net/article.php?story=20070114093427179

writes some interesting things... which I excerpt:

4.2 Clause 3 simply talks about Redistributions of source code. A distribution of modified source is a distribution of source code, so clause 3 applies a copy of these terms must be included with the modified source code. The wording does not support identification of a specific subset of the source code (ie the portion which is unmodified) for the license to apply to. Assuming our analysis above applies, then by including the license terms with the modified source code will make those terms (in particular, the rights to distribute and use in clause 2) apply to the whole of the modified source code as redistributed.

...

6. Can BSD Code be Licensed Under a Different License/Closed Source License?

6.1 We asserted above the collective wisdom that BSD code can be licensed under other licenses. There does not appear to be any basis for this wisdom in the terms of the license. Further, the terms of the BSD license do not appear to allow more restrictive licensing terms. This is not to say that the BSD terms forbid them. Rather, the BSD's usage and distribution rights are so broad, that additional (parallel) restrictions would appear to be ineffective.15 There is not even any basis on which modifications to the code can be used to leverage a different license16 because, as we have concluded above, (if part of a redistribution) those parts which are the original with or without modification must be licensed under the BSD terms.

...

(d) when a company incorporates BSD code into some software, then (to the extent the binaries are modifications of that BSD code) the company is, by assumption, required to license that software (including their own modifications and the source code for their modifications) under the terms of the BSD. We saw above that this does not mean that the source code must be released at all. However, what of the situation where the source code subsequently becomes available? If that was the case, would there be no copyright infringement for copying and distributing that source code? This is discussed further below;

...

9.3 We must therefore pose the question can an employee (or, more likely, an ex- or soon-to-be-ex employee) who has access to the modified source code disclose that source code to others if they acquire a copy of the released BSD binary (i.e. as released by the organization)? Even without this requirement, if the license of a binary version can affect the confidential nature of the source code for that version, then, depending on the timing, this may affect the ability to apply for patents which relate to the source code version.

Thu, 11 Jan 2007

This video is illegal in some places

Today I posted a video on Youtube of my son Liam.

http://www.youtube.com/watch?v=HNV4AdZx4Qw

(If you want the AVI file, I can make that available. I'm told that there is a way to get a raw stream out of youtube)

Liam is 19 months old, and he has four major passions:

1 Chasing the cats and squealing loudly 2 sitting on my desk, talking on both my phones at the same time (a cisco voip phone, and a polycom PSTN two line phone). I think it would be funny to teach him to pick up random telephones, scream "Sell! Sell! Sell!" into them, and run away.

3 playing with his Thomas the Tank Engine toys (and trucks, and anything that can be pushed around the place)

4 watching Thomas and Bear in the Big Blue House videos

The video I took is illegal in some places.

But, not for the reasons you might think. While, yes, we do have some videos of him swimming in the tub that I won't post, it does seem reasonable that I can take a video of him... watching a video.

You'll notice in this video that Liam is watching TV. On TV, you can even hear Tutter the Mouse talking. There is a splash of sun over the screen.

But, in many places (not Canada, where we have a "Fair Use" regime) this would be considered illegal copying. If Hollywood has it's way, to plug this "analog hole" (me pointing my video camera at my TV set), they would force the makers of all video cameras to look for a watermark in the signal.

If my camera saw the watermark, it would BLANK the input.

So, if I pointed my camera at Hollywood content, even if my son was also in the picture, my camera would blank. Even 50 years (or was it 75? or was it 100?) after the video was produced, and it is supposed to go into public domain, I still would be unable to take a video (or maybe even a picture) of my son watching his favorite show.

(nice firehat too, eh?)