Josh Saddler

Hello again, Planet. Another month, another week, another doc or three, another bug, another GMN. Etc.

* * *

For the last month I've been dealing with optical drive issues. First my IDE drive, then the new SATA drive.

The issues with the Samsung IDE optical drive seem to be resolved with kernel 2.6.25, and with the newest stable gstreamer packages. Audio applications can not only see the drive and the media inside, but can actually play the tracks.

Now, however, I'm having issues with the Asus SATA drive I bought when the IDE drive was acting up. It's giving my system fits, as you can see in bug 221145. Libata just hates this drive, no matter what kernel I use.

I did discover that the SATA cables I had been using were bad; they were the original cables packaged with my MSI motherboard. I ordered replacements and plugged 'em in. No more cryptic. I/O errors in /var/log/messages.

However, the drive is still no better off than it was. Applications can see what's in the drive, but can't read from it. The weird thing is that I can sometimes use the drive to burn discs. I was able to burn distro ISOs, and copy them from the Samsung to the SATA drive for on-the-fly burning. But reading is right out. Strange. There are no error messages; there are no unusual messages of any kind. For awhile, I wasn't sure if the errors were of the common variety (poor SATA cables; seems it's universal), or if the SATA ports on the motherboard itself were bad. Given that just swapping out cables removes the errors messages, I assume it was the former.

So basically, I've spent $53 on a drive (Asus DRW-2014L1T), SATA cables, and shipping, and I'm stuck with a piece of nonworking hardware. Maybe I should have gone for another IDE drive, but I only have one IDE port on my motherboard, and it's in use by the other drive. Besides, SATA is supposed to be the way forward. I'd like to eventually have just one kind of interface for everything. Better bandwidth than IDE, no master/slave hassle, etc. Alas, the kernel and my applications refuse to cooperate with the drive. And there's no updated firmware available from the manufacturer, either.

If anyone has any suggestions not already covered in the bug, lemme know. I'm about out of ideas. The only thing I've come up with is booting with some other distro CD, one with known good hardware detection, like Knoppix or *buntu, from the IDE drive, then try to play a disc in the SATA drive and see if it works. If it does, I'll have to hunt up the kernel config and version for the LiveCD.

* * *

Now on to the good news. Jeremy Olexa (darkside) has added wicd to the tree. And not just wicd -- a working version! So now my laptop is amazingly happy. As am I; I had been trying to make wicd work for a long time without success. Fortunately, upstream released 1.5, which creates a much simpler dependency chain, and introduces better networking scripts.

Wicd really makes networking much easier when jumping between networks. It removes all the guesswork from network configuration, as well as the long, arcane iwconfig and wpa_supplicant command sequences. Random public hotspots are no longer a challenge. Just point and click to connect. Wicd is faster and more reliable than NetworkManager, and it has fewer dependencies.

I filed a bug requesting configuration information to be added to the ebuild. Jeremy obliged, so do read the output after you've installed wicd. It really is simple to setup, though baselayout-2/openrc users will need to make a couple of changes, replacing /etc/conf.d/rc with /etc/rc.conf. Here's how I setup wicd for my laptop:

# rc-update del net.wlan0
# rc-update del net.eth0
# rc-update add wicd default battery
# nano -w /etc/conf.d/rc
RC_PLUG_SERVICES="!net.wlan0 !net.eth0"

I rebooted, just to test its autostart capabilities; previous versions could never start properly. 1.5 does; no issues so far. It displayed my network, asked for my key, and then connected. Simple, but oh-so-wonderful.

I'm now a proud wicd user.