Kernel Traffic
Latest | Archives | People | Topics
Latest | Archives | People | Topics
Latest | Archives | People | Topics
Home | News | RSS Feeds | Mailing Lists | Authors Info | Mirrors | Sleeping Cousins

Hurd Traffic #55 For 26 Jul 2000

By Zack Brown

Do you pine for the nice days of minix-1.1, when men were men and wrote their own device drivers?
Are you without a nice project and just dying to cut your teeth on an OS you can try to modify for your needs?
Are you finding it frustrating when everything works on minix? No more all-nighters to get a nifty program working?
Then this post might be just for you :-)
-- Linus Torvalds, 1991

Table Of Contents

Mailing List Stats For This Week

We looked at 141 posts in 506K.

There were 58 different contributors. 27 posted more than once. 6 posted last week too.

The top posters of the week were:


1. Trouble Booting Red Hat From Older GRUB Installations
14 Jul 2000 (2 posts) Archive Link: "ex2 filesystem and grub"
Topics: Bootloaders
People: Tomasz WegrzanowskiPankaj Kaushal

Pankaj Kaushal installed the Hurd several months before, and although GRUB had no problem seeing his Debian partition (and he was sure everything was properly set up), the Red Hat partition would not boot from the menu. Going to the console and typing tab at the '/debian-hurd/kernel=/index.html' prompt, he only got garbage entries. Tomasz Wegrzanowski replied that this was a known bug with older GRUB versions, and that Pankaj should upgrade to the latest version. End Of Thread.


2. Situation Of Hurd Developers And Mailing List Subscribers
14 Jul 2000 - 23 Jul 2000 (45 posts) Archive Link: "Mailing List traffic and membership"
Topics: Bootloaders, GGI
People: Marcus BrinkmannErik EngeOKUJI YoshinoriThierry LarondeTobias HungerFarid HajjiPhilip CharlesTomasz WegrzanowskiBrent FulghamEduardo Nahum OchsRoland McGrath

Marcus Brinkmann gave a pointer to a graph comparing maling list traffic to the number of subscribers. The graph showed steadily increasing membership, but only uneven traffic. He gave his interpretation:

The graph shows that there is a steadily increasing interest in the Hurd, and a decreasing amount of time for people to discuss it :)

Seriously, the decrease in mailing list traffic is a result in lack of time on the developers side. However, it is clear from the graph that the amount of traffic is undergoing heavy changes from month to month (looks a bit lie sawteeth curve).

Roland McGrath mentioned that he'd recently gotten a job, which had cut into his free time considerably, and Erik Enge replied:

Is this the case for many? And if it is, is it because people want to be able to put food on their tables (i.e. money-issue) or something else?

My organization is thinking about taking a (or several) Hurd hacker(s) and paying that person to continue full-time on the Hurd-development. Would this be useful?

OKUJI Yoshinori suggested hiring Roland immediately. Elsewhere, OKUJI also replied to the initial graph, saying:

what I'm more interested in is the reason why the number of developers doesn't increase, even though so many people subscribe to the list. This is not ironical but a purely academic interest for me.

Seriously speaking, seeing the graph, I felt I should come back to the Hurd development as soon as possible...

Marcus replied, "the number of developers does increase, but only very slow (one every few months or so). There have been contributions by new people then and when." Thierry Laronde also replied to OKUJI, "I think that, after the Zeroth Debian Conference and the speeches about the HURD, we are a couple of people interested in becoming HURD developers. The only problem is to find a way to start... and it will take some time." Tobias Hunger also replied to OKUJI with his own reservations. He said:

I am a long time lurker on this list. I even have the hurd installed and toy around with it from time to time:-) I can only give you my reasons for not being more active with the hurd development:

I am interessted in the idea of a micro-kernel and would like to use one -- because I like the idea so much -- but I am not interessted in low-level kernel-related programming and I don't know too much about it, so I will never be a real hurd-developer anyway. But that wouldn't stop me from porting stuff to the hurd, doesen't it?

That's where the next thing comes in: The hurd does not yet support my PCMCIA-network card. Marcus told me at the LinuxTag some weeks ago that it is getting there, but until then it is no fun for me to run the hurd. Booting the hurd, discovering something I forgot to download, so rebooting again to download under linux, rebooting, ... and rebooting takes ages because of those $?%$! scsi-controllers :-(

I could still cross-compile, but well, I would like to know the system a bit better before I start doing so. Just a personal thing.

Something I didn't like at all about the hurd-project at the time I started experimenting with it was the documentation. Thanks to the work of some people on this list that situation got a _lot_ better. Keep up the good work, especially in this area. As long as people can not get the hurd installed properly (I had only an outdated description of the installation process to guide me) they won't use it.

Well, I am waiting for the networking support and when I read that it works, then I will try to get Berlin to run on the hurd. That's the moment you can all point at me and laugh about me and my struggle ;-)

Farid Hajji also replied to OKUJI:

first of all, we need some kind of early release (or pre-release, if you prefer to call it so) that could be burned on a CD-R! Getting all those .debs over a modem link is not only awful, but in some countries really expensive. And I'm not yet talking about the difficulties in getting accurate and actual informations about the needed packages.

Marcus' tar-balls are here a real good starting point for first-time users, but it would be even better to provide splitted files for users with flaky connections (Marcus?). Actually, with Marcus' help, it is much more easy to get the pieces of a working Hurd together and I also mean the .debs. Thank you! You're doing a extremely valuable job at this.

Another important missing point is PPP. We really need PPP in the first place, so that it is possible to cvsup (or cvs update) current sources directly from a running Hurd system. One (probably quick) way to get PPP would be to use FreeBSD's _user-level_ ppp(8), which works through a tunnel device of the FreeBSD Kernel (an excellent, highly configurable and very user friendly tool, IMHO). A fix may be to provide a tunnel translator in the sense of FreeBSD, on which ppp(8) could dock. This approach may be even compatible with Mach's and the Hurd's philosophy. Any takers?

Once we get PPP up and running, it may be a good idea to start thinking (and experimenting) with some kind of installation over the 'Net (be it PPP or Ethernet). Here again, FreeBSD is a good example of how it can be done. Software that just needs one or two boot floppies to be downloaded manually and which then pulls itself off the 'Net is not only easy to setup, but also very practical in case of catasrophic failures etc.

In reply to this, a couple folks expressed interest in getting 'ppp' to work, and other folks suggested prioritizing other packages as well. At one point, Marcus remarked, "We are beyond the stage where certain software packages are more urgent than others (examples in the past have been perl, and the build chain), as the base system is almost complete. Preferences are set by the individuals doing the necessary "dirty" work."

Philip Charles also replied to Farid:

As a CD vendor I have been lurking in this group for about five months now. I have also added Hurd to my Debian archive and keep it up to date. My hope is that at some time I could assist the Hurd community by producing a suitable CD.

I haven't got round to a Hurd install up to now as I don't have much in the way of porting or development skills.

The best I could do at the moment would be to collect the various bits and pieces; grub, documentation etc and burn them onto a CD(s). I don't know how useful this would be. If people thought this would be useful I would be happy to take this on as a project - after the Debian 2.2 release.

Tomasz Wegrzanowski proposed:

You could burn CD containing :

  • Debian GNU/Hurd .debs
  • Hurd packages from
  • sources for Hurd-specific packages (burning packages which sources exist for Debian GNU/Linux would be redundant)
  • sources for Linux 2.4 (tones of drivers), KGI, ALSA, libGGI, GRUB etc.
  • all other sources of interest for OS-developers
  • documeniation for Hurd, Mach, OS and hardware programming

This would contain everything Hurd hacker would need and very little overhead. It probably could fit onto 1 CD.

Philip replied, "I will start looking at this in a few days. I have kept a number of messages to this group which will be useful pointers. However I may have to come back to people for help in identifying exactly what is needed and to hold my hand while I do test installs."

Brent Fulgham also replied to OKUJI with his own take on the low number of developers:

I think the following reasons apply:

  1. The ZERO HOLE problem. (Watch your work mysteriously vanish.)
  2. Lack of good developer documentation. This is changing, slowly. But it is still hard to figure out where things should be implemented, how the pieces of the Hurd fit together, what pieces of the standard libraries available under Linux (and other Unix's) are not implemented (and if they are not implemented on purpose, or just due to a lack of time).

    Roland and Thomas do a great job trying to answer questions from all of us. Given their limited time it's amazing how much response they manage to provide. However, but written documents are more long-lasting... ;-)

  3. Lack of leadership. We proceed in a very ad-hoc manner, because everyone is busy and we only work on the things that interest us. Consequently, people with different hardware are not always supported, or certain features are not worked on because they don't really matter to us (PPP for example). This is similar to how the Guile project was a few months ago, until a new lead maintainer took over. I'm not suggesting we do that here (of course), but I think it might help if we set some goals to work towards.

Now, none of the above are really problems. We have fun, and we do what we can. But it might explain why others aren't jumping on the bandwagon? However, it's heartening to see the many new list subscribers. :-)

Tomasz added an item 4, "Lack of hello-worlds. BIg hello-worlds archive could convert many hackers to Hurd. Having to spend two weeks reading sources before making anything interesting discourages" and an item 5, "Lack of PPP. For most Europeans, PPP is something every OS must have to be usable."

Eduardo Nahum Ochs agreed with these, and added an item 6, "Lack of multiple VTs. If the Hurd had multiple VTs then I would use it often when I only need to edit text/programs in supported languages/etc, and I would explore the system in the intervals... Yes, I know that I could use X or lots of ^Zs (or maybe "screen" if it's working now) but that's a matter of taste..." and an item 7, "Lack of better docs on cross-compilation. I connect to the world via a 33.6Kbps modem and I don't mind downloading Debian source packages (I think it's one of the best uses for my connection time, in fact:-) but I hate downloading .debs, especially when their versions change very fast."

There was some more discussion, focusing mainly on how to use 'screen' to bypass the need for virtual terminals.


3. Improvements For 'pfinet' And 'ext2fs'
16 Jul 2000 (1 post) Archive Link: "old pfinet and hacked ext2fs to avoid zero hole bug"
Topics: FS: ext2
People: Marcus Brinkmann

Marcus Brinkmann posted a link to an older version of pfinet that didn't have the bugs the newer version had. He also gave a link to a version of ext2fs that did not have the bugs of the current version. There was no reply.


4. The Hurd Under Alpha
17 Jul 2000 - 19 Jul 2000 (3 posts) Archive Link: "Debian GNU/HURD for Alpha?"
People: Ron FarrerAd Buijsen

Ron Farrer wanted to see the Hurd on Alpha, and after some study and thought addressed the subject:

There seem to be different opinions on how one should start porting things over. Having read through a LOT of mail and docs I've outlined this:

  1. Port a microkernel to Alpha; Some say to use a previous Mach that was ported to Alpha, some say to go ahead and try to port GNU Mach, others say to find another microkernel (i.e. not a Mach-based one).
  2. Create/port some type of boot loader.
  3. Port the C library and friends.

After these three things have been done it should be fairly easy to get the rest to work. There is interest from others to port, but none seem to have gone further then an inquiry.

Ad Buijsen replied remarked, "The non-progress seems to me to be mainly caused by the fact that the Alpha is a little-used architecture. It's just too bloody expensive..............." He went on to reply to the idea of porting the microkernel:

The problem with using the existing Alpha Mach (provided the sources are available at all) is that it will probably be quite old: EV4 only. Porting it to something more modern will involve tweaking PALcode at least. Maybe, however, Mach's PALcode is closely related to DU's, so that could be used (like Linux does) or relatively easily adapted. Likewise, the porting of GNU Mach may or may not require the development of its own Privileged Architecture Library. This is a difficult task and you may need tools from Digital, I mean Compaq.

(For those who are wondering what I am on about: EV4 was the first generation of Alphas; the present one is EV6. As the hardware, even within the same generation, tends to vary vastly, there is a layer of software which hides these differences from the OS; this layer is the PAL. It has maximal privilege and thus maximal opportunity for wrecking the system.)

There seems to be some interest in getting the Hurd ported to the L4 microkernel, especially on Alpha. See:

Note that L4 doesn't resemble Mach much, so some amount of redesigning might be in order. Still, it is very micro...

Ron replied with a breakdown of Alpha generations by processor number:

For those that are interested the different generations break down as follows:

  • EV4 - 21064 processor
  • EV45 - 21066
  • EV5 - 21164
  • EV6 - 21264
  • EV67 - 21264A
  • EV68 - 21264B

Compaq/DEC are usually happy to provide loaner systems to people porting to Alpha. I don't think this will be much of a problem.

For an earlier discussion of porting the Hurd to Alpha, see Issue #40, Section #5  (19 Mar 2000: Hurd On The Alpha)


5. Ending Hurd's Dependency On Linux
18 Jul 2000 (7 posts) Archive Link: "Installing the Hurd without preinstalling Linux"
People: Farid HajjiRoland McGrathMarcus Brinkmann

Farid Hajji liked the idea of installing the Hurd on a raw system, without requiring a partition for Linux or any other OS. He exhorted, "Even at this early stage of development, we should strive to make the Hurd as free as possible from other OS (well, at least for bootstrapping!). The installation procedure should be at least self contained and independent of any other OS." Marcus Brinkmann replied that he'd been working on a set of Hurd boot disks, but that other things had eventually taken priority. Roland McGrath also replied to Farid, suggesting, "The way to do this is boot something from a floppy (or over the ethernet if you can do that) that you can use to set up the disk. There are multitudinous things ot choose from. The boot floppies from your linux distribution of choice will almost certainly suffice to get you a root shell on which you can run fdisk and mke2fs and so forth. There is at least one "linux on a single floppy" distribution whose sole purpose is to fit useful things you would need to diddle your disk or fetch things over the network on a bootable floppy. is one."


6. Suggested Reading
18 Jul 2000 - 21 Jul 2000 (3 posts) Archive Link: "Recommended Reading?"
People: Sune KirkebyJim FranklinTobias Hunger

Tobias Hunger felt he didn't know enough about the Hurd and microkernels, and asked for some pointers to good reading. Sune Kirkeby replied:

Depending on your interests various / all of the original Mach papers are recommendable reading, Kernel Principles and MIG - Mach Interface Generator both helped me along alot.

There is also an old appendix on Mach from the Operating System Principles book, you can download it from, I can recommend this for a 50K ft. birds-eye view of Mach.

If you want to see Real Hurd Code there some examples of translators in the CVS tree that make for nice simple introductions (hurd/trans/null.c and hurd/trans/hello*.c).

And Jim Franklin recommended, "try "Operating System: A design-objected approach" by Crowly. it deals with concepts of object-oriented design operating systems."


7. Packages Needed For Native Development Under The Hurd
20 Jul 2000 - 21 Jul 2000 (5 posts) Archive Link: "List of Development tools for the Hurd?"
People: Igor KhavkineFarid Hajji

Farid Hajji wanted a list of all .deb files needed to develop software under the Hurd. He found it frustrating to install a package, find out he needed something else as well, and have to reboot to Linux in order to download the new package. Igor Khavkine replied that after some experimentation, he had come up with just such a list:

Dimitry Andric also replied to Farid, and suggested using 'dselect' to choose packages and resolve all dependencies, and only then reboot and download the packages in one fell swoop.


8. Developers Join Forces For PPP
20 Jul 2000 - 21 Jul 2000 (4 posts) Archive Link: "ppp"
Topics: POSIX
People: Daniel E. BaumannChris Lingard

George A. Dowding wanted to get 'ppp' working for the Hurd. He asked who else was working on it and what work had been done. Daniel E. Baumann replied:

"I plan on working on ppp for the Hurd. Next quarter (Fall) I am taking a directed study to meet a course elective (for B.S. Computer Engineering at Milwaukee School of Engineering, MSOE). I plan on porting pppd as the final project." Since he wouldn't be starting for awhile, he invited anyone to do it sooner, or to collaborate with him. Chris Lingard also volunteered in reply to George, and added:

I have had a little look at UNIX pppd and BSD pppd; both need a lot of work.

Also there is the little matter of designing the translater. I read somewhere that the experts have some thoughts as to how it should be done, but cannot find details.

I have a years mail from the Hurd groups; and though many people have started work on this; nobody has published any results.

For a history and status of the ongoing 'ppp' saga, check out Issue #49, Section #4  (19 May 2000: PPP: Saga Continues) .


9. Sharing Swap With FreeBSD
21 Jul 2000 - 22 Jul 2000 (3 posts) Archive Link: "Sharing swap with FreeBSD"
People: Marcus BrinkmannR Joseph Wright

R Joseph Wright gave a pointer to a document on sharing swap between Linux and FreeBSD, and wanted to know if a similar thing could be done between FreeBSD and the Hurd. Marcus Brinkmann saw no problem with this, and replied:

GNU Mach should be able to read disklabels.

About the swap partitions: Neither BSD nor the Hurd needs a special format, so you can just add the BSD swap partition to /etc/fstab on your Hurd.

The Hurd can also recognize and honour a Linux swap signature, but it is not necessary. There are no relevant size limits either.

In short, in this respect BSD and GNU Mach/Hurd are very similar :)


10. 'top' For The Hurd
23 Jul 2000 (2 posts) Archive Link: "top : existed for Utah Mach4"
People: Thierry LarondeMarcus Brinkmann

Thierry Laronde remembered someone asking if 'top' was available for the Hurd, and mentioned, "I have searched for docs, etc... and found that a version of `top' existed for utah MACH4 (UK22). It's probably a good starting point (Uhm ?), and one can find a bunch of utils for MACH4 here :" Marcus Brinkmann replied shortly, "any version of top and other ps related utilities for the Hurd should make use of the ps-library included in the Hurd."







Sharon And Joy

Kernel Traffic is grateful to be developed on a computer donated by Professor Greg Benson and Professor Allan Cruse in the Department of Computer Science at the University of San Francisco. This is the same department that invented FlashMob Computing. Kernel Traffic is hosted by the generous folks at All pages on this site are copyright their original authors, and distributed under the terms of the GNU General Public License, version 2.0.