Description of problem: Attempting to run hddtemp with a SATA drive produces results like this: /dev/sda: ATA SAMSUNG SP2504C: S.M.A.R.T. not available Version-Release number of selected component (if applicable): 2.6.20-1.2300.fc5 How reproducible: Always Steps to Reproduce: 1. Boot 2.6.20-1.2300.fc5 2. Run hddtemp on a SATA drive. Actual results: /dev/sda: ATA SAMSUNG SP2504C: S.M.A.R.T. not available Expected results: /dev/sda: ATA SAMSUNG SP2504C: 40°C Additional info: Attempting to force the SATA drive type in /etc/sysconfig/hddtemp has no effect. smartctl -d ata -a /dev/sda produces a S.M.A.R.T. report It is possible that the bug is in hddtemp, but the problem only began with the new kernel.
What version of hddtemp do you have?
hddtemp-0.3-0.8.beta15.fc5 is installed on this system.
This is due to a bug in hddtemp. It needs this patch: http://lkml.org/lkml/2006/12/19/131 I submitted it to the hddtemp bug reporting system.
This also affects PATA drives on Promise SATA/PATA controllers with a kernel fixed for bug 232600, such as kernel-2.6.20-1.2932.fc6. (They also use libata, so I assume it's the same problem.) Is a fixed hddtemp forthcoming?
Adding hddtemp maintainer to CC. To sum up, hddtemp needs an update/patch to work with kernel >=2.6.20's libata, for all of FC5, FC6 and devel. Please see the previous comments for details. The bug should also be reassigned to hddtemp, but as I'm not the submitter nor in fedorabugs, I can't do that.
Looking at the CVS, it turns out this is already fixed in devel (first reported as #221100), and has been fixed in CVS for FC6 and FC5 a few minutes ago. Thanks Ville. So I guess this can be closed now (or once the builds get pushed).
Yep, Chuck was kind enough to ping me about this. I had been running the patched version locally ever since bug 221100 was fixed and was thus not bitten by forgetting to push the builds for <= FC6 when they got a 2.6.20 kernel, sorry about that. Builds are done now, and will be in the next push.