Saturday, 1 June 2002	Echunga –> Olivaylle	Images for 1 June 2002
		Top of page
		next day
		last day

Up early this morning and off to Victoria with the horse float to visit the Yeardleys in the Wimmera. Had some minor problems with the float brakes, but apart from that and a little drizzle, all went well.

In the afternoon, had a look at the horses; between the Yeardleys and Jorge, they now have about 50 horses spread over a pretty large area. After that, got some horses and went riding. Yvonne still didn't have a horse to ride, so she tried taking Chris out of lending her Scarlett until she's ready for her next foal, in about a year's time. Scarlett's still pretty green, and Yvonne went off on her own while Chris, David and I went off on three of Jorge's Paso Finos, the first time I have ever ridden any kind of Paso. Managed to get my horse (they called her “Wanda”, but Jorge later told me her real name was Guantanamera) into quite a reasonable Paso, but it still wasn't good enough to tempt me to change from trotting horses.

Jorge came to dinner in the evening, and we had a long talk about DSLD. We seem to agree that Mary Donald is probably a little over the top with her concerns, though she has some valid points.

Sunday, 2 June 2002	Olivaylle –> Echunga
		Top of page
		previous day
		next day
		last day

Up relatively late this morning, and down to take some video of some of the more interesting horses. Chris rode Jorge's stallion Cali (Don Cristobal), who is 13 but still pretty green, and he gave her some interesting action before he finally settled down.

Left for home at 12:40, and all went well except for the fuel consumption. From Bordertown to OliVaylle and back to Coonalpyn, a distance of 260 km, we used 54.6 litres of petrol, an average of 21 l/100. Admittedly we made pretty good time, arriving back home at 15:40, 350 km in 3½ hours.

The main issue was the strong headwind. Normally we didn't use anything like that much petrol.

Monday, 3 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

The trouble with going away at weekends is that there's so much to do when you get home again. Spent most of the morning catching up with mail, then back to the JFS1 stuff. I'm learning more about patching Linux source trees, but it's still more work than I like. By the evening had a working kernel.

Tuesday, 4 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

More slogging through the Linux kernel today trying to find out why I'm not getting woken on completion. Little to show for it.

Back riding Darah this afternoon, for about 20 minutes. Firstly I didn't have more time, and secondly Darah was pretty hot after not having been ridden for a while. Yvonne and Diane Saunders also did some riding on Scarlett, who promises to be quite good when she's learnt a little more.

Wednesday, 5 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Another teleconference in the morning, mercifully a bit shorter as the result of us having putting in written reports yesterday. Then on with the JFS1 stuff, and I'm learning more than I ever need to know about synchronization in Linux device drivers.

Riding Darah again today. Half an hour is easy enough to manage, which is more than I can say for her. Still, with a bit of regular riding things should improve.

Spent a lot of time going through the submissions for the AUUG 2002 conference, where we mercifully have many more submissions than slots. Another 90 minute teleconference and we had at least hacked out most of the papers.

Thursday, 6 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Greg Rodgers had ignored my recommendation not to apply the diffs I sent to him a couple of days ago, and had built a PPC64 kernel with them applied. It failed with a supposedly corrupt root inode, and I spent some time investigating that. After some examination, but without building a kernel here, came to the conclusion that the original IBM headers were to blame: the lengths of int, long and short obviously don't correspond.

Rode Darah yet again. I don't know when I last rode on three consecutive days, or four times in a week.

Friday, 7 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Frustrating day. I slept badly and got up late, and spent much of my time chasing this strange JFS1 bug, but got nowhere. Every time I look at the problem, it seems to have changed. I wonder if I have hardware problems somewhere; this is the machine that I developed the FreeBSD SMPng interrupt threads on, and I've been having trouble running FreeBSD on it as well lately.

Saturday, 8 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Finally the rains seem to have come, none too early: usually they start in late April. Decided it was finally time to buy the plants we had intended to plant this winter, so off, stopping on the way in Mount Barker, where I bought my first pair of shoes in years, and then to Belair National Park, where we bought a number of native plants. We're planning to plant the circle in the middle of the driveway with Grevilleas and Banksias.

Back home, where I had planned to work on my book, but somehow felt spectactularly tired, and did little for the rest of the day. I must find interfacing with the Real World to be particularly tiring.

Sunday, 9 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Finally had some time to work on my book again today, and found that I'm getting to a stage where everything is dependent on everything else. It's a good thing that FreeBSD 4.6-RELEASE is as good as out. Took down zaphod, currently running Linux as pita for JFS1 development, and booted it as daemon, the 4-STABLE incarnation. It paniced with two different Ethernet cards when running ifconfig. Maybe there is something wrong with the machine. This was a standard 4.5-RELEASE kernel, so it's nothing I've put in there.

Upgrading systems seems to be getting worse all the time. Somehow I can't make friends with this mergemaster utility, and in any case, it should run automatically from make install. Specifically, there are new groups and users in 4.6 which weren't in 4.5, and if you don't add them before installing, the install will fail. There has to be a better way.

Monday, 10 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Today was one of the many birthdays of the Queen of England, the one we celebrate in Australia with the exception of WA. If she were one year older for every birthday celebrations, she'd be over 500 years old by now. Anyway, it was a public holiday, so I had the opportunity to work on my book.

Made some pretty good progress for a change, and nearly finished the disk chapter. That's more work than it seems: the tools have changed. sysinstall has become more usable, though it seems to have caught a bug in cursor positioning, and fdisk and disklabel are bitrotting badly. fdisk is close to being non-functional. Took a look at fixing it, and decided it would probably make more sense to import a newer version from NetBSD or OpenBSD.

Tuesday, 11 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Yana started work pruning vines at Langhorne Creek today, which means that for the next few months she'll have to leave home at 5:50 am. Decided that we might as well go along with the timing, especially since I've been sleeping badly anyway lately, and because it's easier to keep in contact with the USA at this time in the morning, so got up at 5:30 am.

More work on the JFS1 problems. I'm completely baffled by the problem, and to make matters worse there appears to be a bug in gdb which stops me from putting breakpoints on certain locations (well, I can do it, but it crashes the system). As a result, spent most of the day reading code.

Another AUUG 2002 teleconference in the afternoon. We have more or less finished the programme details.

Wednesday, 12 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Up early again this morning, and had my daily housekeeping over and done by the time we had our teleconference this morning. I note that nobody understands how to attack my JFS1 bug, which doesn't help much, but it does confirm that I'm probably not being completely stupid. Greg Rodgers reports that it behaves exactly the same way on the RS/6000, so any concerns about flaky hardware are probably incorrect. Decided to take a day off the issue. Maybe I'll have better ideas tomorrow.

AUUG SA meeting this evening, with a talk by Wayne Piekarski, who did a very interesting talk about the wearable computers and Augmented Reality (a cross between virtual reality and real reality) that they're developing at UniSA. It's early days yet, but they're doing some impressive work.

Thursday, 13 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Up early again this morning. Maybe all these years I've been complaining about not being able to sleep, I really should have realized that I didn't need to sleep any longer. Let's see how long I can keep up like this.

I still didn't have any idea what to do with the JFS1 stuff, so left it for another day, and on looking at the debug environment instead. Jerry Dunham recently sent me an old Dell laptop, and it seemed a better candidate than sydney to use for debugging, but I have been having problems getting it upgraded to FreeBSD 5-CURRENT. It's getting more difficult all the time to upgrade a 4-STABLE machine to 5-CURRENT. Decided to bootstrap off zaphod, and first tried to build a recent snapshot of 5-CURRENT on that machine, but ran into surprising problems there too.

In the evening discovered that I had lost my alias addresses for the reverse lookup name servers. They're hard-wired by DENIC to be allegro.lemis.com and freebie.lemis.com, both of which no longer exist under those names, but I haven't been able to get them to change. Added an alias for freebie to echunga, and noted extreme degradation in response times. On investigation, I discovered I had inadvertently set the net mask to 255.255.255.255, effectively locking echunga out of the network. It worked anyway, because echunga is my default gateway, so packets from echunga to the network went out the PPP link, across the Pacific, up to PanAm Sat 8 and back via sat-gw to the network, a total distance of about 85,000 km and 500 ms. It's fun being on the Internet.

Friday, 14 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Finally got onto Daniel Phillips on IRC and discussed my JFS1 problems with him. He was at least as baffled as I, and made some suggestions. I offered to show him my code, but of course I first wanted to test his suggestions, and in the process discovered that I seemed to be going into an infinite loop in the directory conversion routine. Quite possibly all the strange things I have been experiencing before have been related to bugs in the remote serial gdb debugging code. sigh

Playing around with Diane Saunders' ancient PC again today. Yesterday I had replaced the power supply, the only thing that was really wrong with it, and it worked as fine as “Windows” 3.1 has ever worked, but today I couldn't get it to boot—the ominous Missing Operating System message. It booted most of the time on zaphod, but I couldn't get it to do so reliably. Considering it was only 100 MB, I decided to replace it with a not-quite-so-ancient 300 MB drive, but I couldn't get Microsoft to recognize it. Maybe it had something to do with the compressed file system on the old drive. How I hate Microsoft.

Saturday, 15 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

It's been raining all week, and looks like continuing for the next week or so, but Yvonne was burning to go out riding in Kuitpo Forest, so set off relatively early with Darah and Scarlett. It wasn't the best ride we have ever had: Darah wanted to go much faster than I wanted her to, Scarlett's still pretty green, and of course it started to rain. Still, it's been a while, and it wasn't too bad. Hopefully we'll be able to float them next time.

More work on the book, describing the startup procedures. It's deadening.

Sunday, 16 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

More book work today, and made some good progress. If I could keep this up regularly, it would be finished in no time at all. Now if I could only get JFS1 out of my hair.

Moving horses around in the afternoon. It took a while to find a combination which we thought would not ending up kicking each other to death. We ended up with Darah, Scarlett and Pebbles in one paddock, which should make riding and transportation easier too.

Monday, 17 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Back to work on JFS1 today and followed up the idea that the real issue might be a loop in the directory stuffing routines. Indeed, that was the case. It seems that Linux directories handle entry lengths differently, though it's difficult to tell, since you can't read a raw directory under Linux. Anyway, after that, it worked. Took out my hacks in block_read_full_page, and it hung again. sigh.

Deferred the fix for that and spent the rest of the day writing the stat routine, which was not as simple as it seems: JFS1 doesn't keep stat information in the super block, so I have to count the allocation bitmaps, and that takes a fair amount of code. Finally got that finished, and the resultant kernel hung about 10 seconds after a mount operation, apparently without any further file system call. I suspect there are some missing dependencies which resulted in various file system structures being inconsistent.

Tuesday, 18 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

More work on JFS1 today, and made good progress. By the end of the day, everything seems to be working except for the delayed I/O completion. I wish I knew how to look for this one, but it's not a show-stopper for testing, since the wait_on_buffer hack works around it. Now to catch up with other things I've been doing.

Our Davey HS 60-08 pump is playing up. Three months old and from time to time it stops working until it's power cycled. In that time we still haven't heard from the Davey representative who was supposed to come and find out why their pumps kept blowing up. Called Mount Barker Irrigation and discovered that Rick Babbidge is no longer there, and the new bloke, Brad, knew nothing about it. He did try to do something about the problem, but doesn't seem to understand the issues. Suggested that I call Robert Young; a lot of good that's going to do. It's about time to take more drastic measures.

Wednesday, 19 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

After the tiring work of the last few weeks, spent today catching up on things that got left over. The weather was dry and sunny despite the weather forecast, so got Michael Schirmer to come in and spray as many paddocks as possible for weeds. Spent a bit of time showing him things there, then back to work.

Yvonne wanted to go riding in Kuitpo Forest in the afternoon, but at the moment Scarlett panics when being floated, and I didn't feel like riding all the way down Duffield Road. So much for the advantages of being close to the forest.

Tried a couple of times to contact Davey, but the bloke didn't reply, and though I left a couple of messages, he didn't call back. I get the feeling that he doesn't want to. I need to take more drastic measures.

Thursday, 20 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Adelaide UNIX User Group lunch today, at a Belgian beer cafe of all places. It's interesting to note that of the ten people there, 7 use FreeBSD, and as far as I know nobody uses commercial UNIX any more. UNIX? In fact, it sounds as if, as a product name, UNIX is dead. We had a discussion about the fact, and decided that the last version of UNIX left is UnixWare, and not many people use that any more. How times are changing.

Friday, 21 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Midwinter's day today, and it rained to match. It was rather surprising to see Yana come home with sunburn. Obviously the weather in Langhorne Creek was better.

Didn't get much work done today; spent most of the time handling administrative details.

Saturday, 22 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Another slow day. Did some work on the book, but not as much as I had hoped.

In the evening planned to cook a choucroute garnie. That proved to be more difficult than I thought: the recipes I had all disagreed wildly on the ingredients and the cooking time, and even the lists of ingredients looked dubious. Looked on the web and found very little of use, and a lot of nonsense, like this recipe (http://www.germandeli.com/gd5.html>) which states “This traditional dish is a special favorite enjoyed throughout Germany all winter long.” I don't think I've ever seen it in Germany, though it's probably eaten here and there.

“This recipe” is long forgotten. Clearly that's not wrong.

After a lot of messing around, came up with a recipe which proved to be slightly unbalanced, as the photo shows:

Image title: choucroute          Dimensions:          1951 x 1396, 1022 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Sunday, 23 June 2002, thumbnails          All images taken on Sunday, 23 June 2002, small Diary entry for Sunday, 23 June 2002 Complete exposure details

 

It's time to come up with my own recipe for further reference.

Sunday, 23 June 2002	Echunga	Images for 23 June 2002
		Top of page
		previous day
		next day
		last day

Things have been slowing down lately, but today was really slow. Got up late and planned to work on my book, but somehow I didn't get very far. Then decided to do some work on my paper for AUUG 2002. In the process discovered that I had lots of duplicate large PostScript files, so set to deleting large unnecessary files. That developed into a session extending mklinks, which can now delete duplicates. Spent a few hours on that, and ended up with something like 5 GB less data on wantadilla. I'll look at echunga later; it has more disk, and I'm always surprised what I find there.

Diane Saunders came to dinner in the evening to help us with the left-over choucroute. Early to bed to make up for getting up late.

Monday, 24 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

I was just getting ready to wonder what to do today when I got a call from Liz Carroll with various fires burning out of control. Spent most of the day trying to handle that. Spent the rest of the time trying to build a -CURRENT kernel on zaphod. Our recent changes to the C compiler seem to have strange and far-reaching consequences.

Tuesday, 25 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Chris Maltby announced the results of the AUUG elections today, dashing my hopes of getting anything else much done. It's interesting to note that once again, five of the nine members are from 08 territory (SA, WA and NT). In this case, though, four members are from SA, so it looks like we're going to have meetings here more often.

Also today a vulnerability in ssh was announced, and Stephen Rothwell (now newly elected to the AUUG board of directors) wanted to upgrade www.auug.org.au, which wasn't as simple as it seems, especially since the only access to the box is via ssh. It's rather like cutting off the branch you're sitting on.

Final dentist's session today. Thank God that's over.

Wednesday, 26 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

AUUG conference preparations are still taking a sizeable chunk out of my day. Today things were further complicated by the announcement of the results of the FreeBSD Core Team election. Rather to my surprise, I was reelected. Now we have more work there as well. It's a good thing I don't have anything really pressing else to do, though I still have to finish a book, a paper and an article for Daemon News. Where does the time go?

Got a mail message from Davey today, asking a minor question. They still haven't addressed any of my real questions. This is impressive, but not positively.

Thursday, 27 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Still more AUUG and FreeBSD stuff today. Somehow getting the conference brochure out has taken more time than I expected, and getting the new core team rolling also took up some time. In addition we had news of an exploit in ssh, which kept people both upgrading and discussing.

As if that wasn't enough, I've spent all week so far trying to upgrade zaphod to the latest FreeBSD-CURRENT. After some trouble discovered that it was leaving old binaries behind in /usr/libexec, causing compilations to fail with nonsensical error messages. Removed that and discovered after the reboot that I had removed /usr/libexec/ld.so, and that for some reason it no longer gets installed. sigh.

Yet another power failure this evening, taking out sat-gw. Finally got fed up with it and installed a UPS.

Friday, 28 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Still more admin work. This is getting frustrating; a whole week with nothing useful to show for it. Out of sheer despair, started preparing a duck terrine, but even that was frustrating. The recipes petered out into nothingness (what do you do with the wings? I have three recipes, none of which said anything about the issue). In addition, the quantities looked just plain wrong. Anyway, some of the meat require marinating overnight, so we'll have to wait to see how it turns out.

Saturday, 29 June 2002	Echunga
		Top of page
		previous day
		next day
		last day

Still more work with AUUG today. I'm beginning to wonder if I'm becoming single-threaded. At least I'm getting things sorted out, and today, instead of much-needed work on my book, I spent a lot of time tidying up the web site, which is still sorely in need of it.

Despite everything managed to get some riding done in the afternoon. For once, Darah wasn't too hot.

Started making a duck terrine today, and discovered that all the recipes I could find had incorrect quantities. It proved that the one I had chosen somehow wanted to make 2.5 kg of filling out of 2 kg of ingredients. I wonder how people write these recipe books. I suspect that they cook first, take the photos, and then guess what they put in. Grr.

Sunday, 30 June 2002	Echunga	Images for 30 June 2002
		Top of page
		previous day
		next day
		last day

End of the month and the financial year today. I had meant to write another article on audio ports for Daemon News, but I just couldn't get my act together, especially as keeping zaphod running proved to be a big problem. Wrote about xv instead, and for once was relatively happy with the results.

Off riding in Kuitpo Forest in the afternoon, the first time in some time. Just before I left, the UPS I put in on Friday died on me for no obvious reason, taking out the satellite receiver and thus most of my Internet connectivity. This UPS is by Deltec, and from time to time it starts beeping as if the power had failed. The beep needs to be reset manually, and if it's left long enough, it powers down the UPS. I really wonder what kind of substance abuse UPS designers practice.

The ride was fun, the longest we've had in the forest for some months, though still not as long as I'd like. Once Yvonne has Scarlett in better condition, things should improve.

Do you have a comment about something I have written? This is a diary, not a “blog”, and there is deliberately no provision for directly adding comments. It's also not a vehicle for third-party content. But I welcome feedback and try to reply to all messages I receive. See the diary overview for more details. If you do send me a message relating to something I have written, please indicate whether you'd prefer me not to mention your name. Otherwise I'll assume that it's OK to do so.