Fighting USB

Mar 30, 2006 / 0 comments

Earlier this week my Amstrad E3 2.6 patch was accepted into the linux-omap-2.6 tree. This is pretty cool; I'm a big believer in pushing changes to upstream where possible and thus it's great when they take them.

I'd thought that once this base patch was in it would be a simple matter of cleaning up my patches for the LCD, USB, keymap and NAND support, ready for submission. I have all of these working fine on my 2.6.15 tree, so it should have been easy, right? Of course not. I've spent most of the week (well, the bits I wasn't working or doing late night router upgrades) trying to solve the ohci: Unlink after no-IRQ? Controller is probably using the wrong IRQ. problem. No, I know it's not the wrong IRQ. It works fine with 2.4 and 2.6.15. Gah.

What's worse is Google only showed up people having the same issue; a number apparently x86 with ACPI being suggested as the culprit, but there were a few that were also with the OMAP5910 (the E3's processor). The most recent from Ladis but he's already turned up on #emailer to ask me how I'd got it working in 2.6.15.

Anyway, I think I've finally solved it (the 5910 needs the local bus clock enabled for USB according to the datasheet and this wasn't being done), though I'm waiting to see if my solution works for anyone else before deciding it's a success. Assuming it does I can actually start to look at getting the other bits ready too; I think the LCD stuff has changed a bit and I dread to think what the MTD guys are going to say about my attempts at a driver.

Nominet EGM

Mar 20, 2006 / 0 comments

I was in Oxford on Thursday for the Nominet EGM. There were a variety of changes that they wanted to make and while in general the principle behind them seemed good the detail was a bit worrying, and I believe had been badly communicated. However I was still surprised to find that the vote was defeated; I'd expected a narrow victory. There are some more details on The Register, who were the only people to send someone I believe.

What's scary is the percentage of members who didn't vote. Nominet has just under 3000 members. I wouldn't have said more than 60 people turned up to the EGM, and some of them were Nominet staff or multiple people from the same member organisation. Only 10% of members actually bothered to vote. That's quite a pathetic number. Even if you don't take on the view point that Nominet are a member run organisation (and I do and I think they work very well in general and try to do the right thing and promote good practises), this was a vote that would have given the board the right to modify pricing as they wished. Surely that's something you're interested in even if all you do is treat Nominet as a normal supplier? There's more to the changes than that obviously, but I would have thought price changes were something that might at least make beancounters sit up and take notice.

Do you work for an organisation who's a Nominet member? Were you aware there was an EGM taking place? Did your organisation vote? Are you aware there's a PAB election starting from this Friday? Are you planning to vote? Have you ever been to an AGM? If not, why not? What about the members lunches?

It's somewhat disappointing to see such a low level of member involvement. Nominet do seem to try to talk to people and get their opinions, but you have to actually turn up for that to happen...

Neighbour table overflow

Mar 17, 2006 / 0 comments

This is mainly for Google, as I found a lot of things talking about the error with not a lot of solutions when I looked for it.

I had a machine that was outputing the Neighbour table overflow message with reasonable frequency, and while it didn't actually seem to be causing problems it was disconcerting. Digging around seemed to suggest it was an ARP issue, with some pages mentioning Blaster - presumably because the scanning that this sort of worm can do is likely to fill up the ARP table?

Anyway, this machine has over a /22 of IPv4 space hanging off its various interfaces, so it seemed logical that I might need a larger cache than the default. I found this SuSE post which solves the problem perfectly. So I've added the appropriate magic to /etc/sysctl.conf:

net/ipv4/neigh/default/gc_thresh1=1024
net/ipv4/neigh/default/gc_thresh2=2048
net/ipv4/neigh/default/gc_thresh3=4096

(Oh, and the message comes from net/ipv4/route.c in the kernel source tree.)

Open Source Accessibility

Mar 16, 2006 / 0 comments

While driving to Oxford yesterday (more on that tomorrow I feel) I was listening to Radio 1. The Scott Mills show had someone ring in asking for Laura's autograph. (I'm not sure exactly who she is other than someone who's having her diary from university read out on the show and appears to work for it somehow.) Scott then proceeded to try and get people to ring in with more pointless autographs that they had.

This got me thinking. I don't think I've asked anyone for an autograph since Eddie Izzard signed my jacket when I was 15. I don't think I'd be that bothered about getting anyone's these days. But there are probably a bunch of people I'd quite like to have a drink and a chat with.

And that led me down the path of the people I consider famous in the Free Software world that I've met and had what I consider proper conversations with. They probably don't remember them or me, but it's a level of access that I don't think is mirrored in other industries, such as the music biz. If I were a lowly guy with a band in my garage then I'm unlikely to get to meet the movers and shakers of the record industry and even if I managed to be in the same room as them they're unlikely to have a decent conversation with me.

Yet in the Free Software community even a lowly unknown like me can rub shoulders with people who really make things happen and actually have meaningful exchanges with them. How cool is that? All you have to tend to do is make the effort to turn up to events; sometimes just your local LUG if you're lucky, or maybe a free LinuxExpo, or maybe something like Debconf. I think this is a pretty low barrier to entry. And I think it's a real strength of the community. In fact, it's part of what helps us have the community.

(Maybe I'm just easily impressed by people though. And to be fair the people I'm viewing as famous probably aren't widely known to the non IT world. I still think it's a remarkable thing about the way in which we operate, something to be proud of and to not forget about.)

Mmmmm. Munin.

Mar 15, 2006 / 0 comments

I've been using MRTG for many years now to draw pretty graphs. It's a bit inflexible though and I'd written a kludgy script or two to graph things that wanted more than 2 lines, or that I wanted to dynamically configure. I'd been on the looking for something simple that would do this all for me and replace my MRTG + scripts setup. All of what I'd seen seemed a bit complicated to get up and running, and what I had was BALGE, so I hadn't really looked that hard.

At the start of the week, however, I got pointed at someone else's stats page, generated by Munin. The graphs were presented in a nice clear format rather than a million levels down a tree, and all the graphs for a host were on one page for at at-a-glance view. Of course there's a Debian package, so I thought I'd try it out. Very quickly I was up and running with my local machine being monitored. A little extra work and I had SNMP monitoring of my switch, allowing me to ditch MRTG. Slightly more work and I had my UPS being graphed too. Finally I installed snmpd on my wgt634u running OpenWRT and had that sorted too. Excellent. A lot less work than I expected and a much nicer setup than before. I'm sure there are other good options out there, but I found this one worked for me.

Oh, and while I was at it I discovered mbrowse, an SNMP MIB browser. I think I'd rather have something that showed me the values read from the device in tree form, but I still found it quite handy.

subscribe via RSS