Technical > PC Hardware Discussion/Reviews > [rant] HDD is not valid

reply to koitsu

Re: [rant] HDD is not valid

That KB sounds like a bunch of intentionally vague nonsense.

Chances are this is the exact same BIOS design bug that HP/Compaq has/had in their BIOSes:

The BIOSes issue the necessary ATA CDBs to the disk to get all SMART attribute data. They they incorrectly assume that the SMART attributes returned contain attribute 194 (Temperature). Many SSDs (particularly Intel) do not include a thermistor on their PCBs, thus there is no SMART attribute 194. This causes the BIOS to spit out a bunch of hullabaloo warnings about "temperature failure" (in the case of the HP/Compaq systems) because the BIOS lacked proper handling code for disks without that attribute. I can point you to the thread on the Intel forums discussing this at length if you want to read it, where HP and Intel had to have private behind-closed-doors meetings to "figure out what to do about this" (while I continually brought up the fact that the BIOS developer made horrible/atrocious assumptions and that the check in question should be completely removed -- i.e. it wasn't Intel's fault). I have no idea what was done about it, but from what I can tell, HP did something because Intel never issued a firmware upgrade that, say, faked attribute 194 (and if they had, I would have flamed them over that!).

If this is what's happening to you with your Dell system, the result is a vague/ambiguous nastygram that tells you absolutely nothing useful. Way to go Dell (or whoever they outsource their BIOS development to).

Only Dell engineers -- more specifically, those who develop the BIOS -- would know if this is the case. The likelihood of getting through Tier 1 or even Tier 2 Support to talk to one of these people is practically nil.

The reason the KB probably says Western Digital and Samsung only, is because they're two brands which are known to not vendor-encode attribute 194, i.e. the 6-byte data attribute can be read and interpreted (effectively) as a 64-bit unsigned decimal number that represents the temperature of the drive in Celsius. Some other brands might encode this attribute data (thus when read in the same fashion, results in a number like 19485209282), thus would probably throw the same error.

Your choices:

1. Continually harass Dell Technical Support until you reach an engineer at a high enough level that understands the predicament and can talk to the appropriate folks that develop the BIOS to remove the code,

2. Don't buy this particular model of Dell system, or consider buying another brand/vendors products, or build your own,

3. Remove the SSD and comply with Dell's requirements,

4. Do nothing.
--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.

I turned off smart reporting, and that didn't make any difference.

I don't know if this was put in because of attribute 194 or thermal issues, but this particular case definitely has poor cooling and I'm fairly certain certain HDs would fry themselves. (the 2TB Red is at 30 - 31C, BTW) I need to look up the the HD that was in there, but based on its thickness and weight, I think it's a single platter drive.

Just for kicks, I'm thinking about putting in a 2TB RE4 and a 2TB Black to see if the BIOS likes that combination better than the current one...

In any case, since the need to hit F1 is now gone, I'm a happy camper.
--
Wacky Races 2012!

Have you tried a WD or Samsung drive?