Too Busy For Words - the PaulWay Blog

I think I've discovered the cause of the problem I'm having on my laptop, where DVDs play in stuttering, jerky fashion under Fedora Core 6. This is a classic symptom of the drive not being in DMA mode, confirmed by the large quantities of CPU time used during playback. The message I was getting back from hdparm -d 1 /dev/hdc was "Operation not permitted", and for a while I suspected that the SELinux context on the device wasn't allowing root to change its parameters (despite this being very unusual for Fedora Core's SELinux setup). But nothing was showing up in the kernel messages, so I had to go further.

Using strace on the above command revealed the ioctl call that was throwing the error, and as luck would have it this turned out to have been seen by one or two other people that had noted it somewhere that Google could catalogue. The salient comment was from none other than Mark Lord, who not only happens to be "The IDE Guy" but also the guy who build the excellent Hijack Kernel for the Empeg Car and Rio Car players. He asked if the correct kernel module had been loaded for the IDE interface.

I spent some time digging around in lspci and dmesg to find out what driver was being loaded, and if there was one for the PATA device that runs the NEC DVD drive (as opposed to the SATA hard disk, which is using the correct PIIX driver). But I didn't have much luck. I'm not even really sure that this is the correct line of investigation; but it's all I have to go on at the moment. According to Intel's documentation on the 945PM chipset, the 82801GBM provides one PATA channel and two SATA channels on the same chip. Given that information, and the fact that the DVD drive is coming up as /dev/hdc rather than /dev/sdc, I don't think I need to void my warranty open the case up and looking at the cables to determine that it's a PATA drive.

I still need to find out whether the correct driver is being used for the PATA interface. But how?