Older blog entries for mike750 (starting at number 26)

Freaky, Friday 13th and I've got to catch a plane today, could be fun.

Been fighting with thinkpads recently. It appears that on the a22p the internal ethernet adapter is hardwired to irq 11. Failure to load a driver or activate the adapter stops all other interrupts from getting to the cardbus subsystem.
Need to see if there is a way to move the cardbus stuff around to use a different interrupt.

On a more positive note, all the latest olympic changes made it into 2.4.6, the ibmtr changes are still in the -ac patches and hopefully will make it in for 2.4.7.

Well, I'm shocked, my flight was 3 hours late on Friday so didn't get home until 4am. Then had to get up at 6am to take the car to the garage only to be told it would cost $1200 to fix the air conditioning. As the car is only worth $3000, it ain't getting fixed.

Due to this spent most of Saturday asleep and then spent 2 hours cutting the grass Sunday morning and went out to the wench's bosses place in the afternoon.

So, I still haven't got around to installing Linux on the HP even though I've now got hold of a cdrom that should boot the sucker.

More fun today, the latest pam packages from debian screw up badly, no login, no nothing. Not impressed.

Got delayed for over a day by the airline on Friday. Didn't make it back home until 1am Sunday morning. Not a happy camper

The 4.3g barracuda drive for the HP712/80 turned up so I spent some time installing it (exacto knife to the hard drive supports). Fired up hpux, it found the extra drive no trouble at all.

So, onto the installation of hppa-linux 0.9. No matter what I did, the machine would not boot from the cd. Remembering deep back into the subconscious, I seem to remember that the cd-rom drive has to have a certain block size to boot the machine. So onto ebay we go, found an old 4x toshiba external for $12.50 that should do the trick. Hopefully this will be delivered this week so I can try to get this up and running next weekend.

Received comments back from Anton for the ppc changes to ibmtr, just a small change and then it looks good. I really need to get an older powerbook so that I can test all this stuff, unfortunately only the later models support cardbus as well, and they are still too expensive.

I've actually had to do some work this week, so not too much on the hacking front.

Beat up framebuffer and it still won't work on the laptop. Seems that the ATI Mach64 chipset in the laptop is not properly supported, I just get a blank screen. Everything else continues working, so I'm listening to the harddrive so I can work out when it has finally booted and is at a login prompt to be able to reboot and go back to a regular kernel. This is annoying because mplayer won't scale the image on regular x11 output.

Olympic patch sent off for inclusion :) Of course there is the next stage of work to start on for olympic now.

Fixed up net-tools properly, it will now allow setting of the token ring hardware address (LAA) properly for both the 2.2 and 2.4 kernels.Although I did have to grep through all the 2.3 patches to find out at which point we changed the token ring definittion.

Time to relax for the afternoon and play with some gtk/ui stuff.

Web site update done. All uploaded and operational. Couple of dead links found, but nothing drastic. The source web site is completely dynamic now, so updating is just a matter of updating the database and not hacking raw html anymore :)

Final testing of the latest olympic patch is underway. This patch cleans up the driver a little and add the copy packet if len < 1500 to stop the socket code dropping all our packets.If this doesn't die it's getting submitted this weekend.

Next task for olympic is to enable the sending of mac frames and enhance the error routines, i.e. get the adapter to close gracefully and retry to open if possible.

Tested out the new copy on small packets code in the olympic_rx() function. Looks good.
Someone somewhere is having problems with olympic, every 2-3 hours the interface stops and an ifconfig down/up is needed to recover it. This is definately not the drivers fault, although we are going to have to deal with it. I'm suspecting pci bus errors at the moment, there maybe a revision of the adapter that is really susceptible to this. Time for some extended testing.

I have got *so* many things in the queue for the drivers now, I really need a week off work just to get it all sorted out. Things should be easier once the British Invasion go back home.

Fixed up the compile problems with 2.0.3x. Still needs the mkdep from 2.4.4, although installing the debian version of the 2.0.38 kernel source and gcc 2.7.2 got everything working.

Started banging away with glade. I'm writing a user app to interface with the t/r adapters. Annoyingly the autogen.sh scripts that glade creates links in libdb.so (for no good reason) so had to install libdb3-dev as well as libgnome-dev, very interesting over a 14.4 connection.

The Stanford checker has found one issue with ibmtr in 2.4.4, I'll have to take a look and see what's going on. It's in tr_rx where some very strange magic goes on.

30 May 2001 (updated 30 May 2001 at 14:18 UTC) »

The British Invasion #1 landed on Thursday so didn't get a lot of work done over the weekend :(

PCMCIA-CS breaks on 2.4.x because I've removed the drivers/net/tokenring/ibmtr.h file. He, he, I'm not putting it back, pcmcia-cs will have to be fixed. So giving in I'm going to test pcmcia-cs against all three levels of kernel (2.0,2.2 & 2.4) and see what needs changing to get it to work with all of them. Of course, 2.0.39 decides not to compile up on my machine. First off mkdep was SIGBUS'ing. Grabbing mkdep from 2.4 fixed that, but now all the includes are wrong. Debian doesn't symlink /usr/include to the kernel source (quite right too) but the 2.0.39 build scripts think they do, so I'll have to fix all that up first. *sigh*

Loaded up a copy of Elite:FFE to play and found that modern machines are just too quick and I'm getting blown away in seconds before I have a chance to zero in a missile the baddies. Perhaps if I play it under vmware it will slow it down enough :)

Gave up with NS 6 on the iMac, the extra memory speeds it up but it still sucks. So it's now got Mozilla 0.9 instead which is working fine as long as you don't try to quit with no windows open.

24 May 2001 (updated 24 May 2001 at 21:53 UTC) »

Hmmm, life was going well until I got a report that setting LAA doesn't work in 2.4. Took a quick look and the report is quite right. Hacked net-tools apart and found the problem. There are two fixes for this, one simple and one "right". So I've done the simple one for now, the "right" one will take a little longer.

Memory has turned up for the iMac, I hope this improves the performance of NS 6, otherwise I'm going to put Mozilla 0.9 on the machine.

*Sigh*, looks like my flight is going to be late tonight, again :(

17 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!