SpinRite

[CrazyTeeka]

Oh my goodness .. Now Peter Blaise has joined the party. I'm going to see if I can block Mr. Tap-dance, this guy gets on my nerves.

Mr. Tap-Dance is a fitting name. Why would he promote the use of FORCEBIOS when that's for a specific purpose.

[CrazyTeeka]

Reading the grc.com forums hurts my head. There is so much disinformation, and not just from Steve.

I actually feel the same way. Which is why I do my own testing.

[fzabkar]

I've just joined the GRC forum. Let's see how long before Steve gives me the boot.

[CrazyTeeka]

I've just joined the GRC forum. Let's see how long before Steve gives me the boot.

Your braver than me.

[fzabkar]

Yeah, his answer is totally unsatisfactory and I'm trying rubbing his nose in. Saying SMART is nonsense while at the same time: https://www.grc.com/sr/smart.htm .. You can not have it both ways.

I'll do it for you. I'll probably get the boot soon, anyway.

[lcoughey]

Here is a response from Serge:

We’ve noticed cases like this several years ago, i.e. drives which would read 1 of 8 logical sectors from a single physical sector. We thought it might be a bug with our tools, so we spent some time testing this, concluded that this is real drive behavior sometimes, and dropped the subject. It’s a bit hard to remember at this point, but I seem to recall that this 1 logical sector that comes back as successful isn’t actually a real read, i.e. it’s garbage data that’s coming back due to some firmware bug of the drive. You guys can check this out, i.e. read that suspicious good sector and check if hex values seem valid.

A key point of larger sector sizes is to combine the ECC together because less space can be spent on ECC that way, so it’s extremely unlikely that any drive is able to successfully correct only 1 logical sector from a physical sector, it should be all or nothing. IIRC drive firmware doesn’t even have a concept of logical sectors internally, there is just a conversion that takes place at the end.

[CrazyTeeka]

Here is a response from Serge:

We’ve noticed cases like this several years ago, i.e. drives which would read 1 of 8 logical sectors from a single physical sector. We thought it might be a bug with our tools, so we spent some time testing this, concluded that this is real drive behavior sometimes, and dropped the subject. It’s a bit hard to remember at this point, but I seem to recall that this 1 logical sector that comes back as successful isn’t actually a real read, i.e. it’s garbage data that’s coming back due to some firmware bug of the drive. You guys can check this out, i.e. read that suspicious good sector and check if hex values seem valid.

A key point of larger sector sizes is to combine the ECC together because less space can be spent on ECC that way, so it’s extremely unlikely that any drive is able to successfully correct only 1 logical sector from a physical sector, it should be all or nothing. IIRC drive firmware doesn’t even have a concept of logical sectors internally, there is just a conversion that takes place at the end.

It makes sense that it's a firmware bug. I'll do some more testing.

[fzabkar]

I've just joined the GRC forum. Let's see how long before Steve gives me the boot.

Your braver than me.

It didn't take long. He just called me an ignorant Internet troll.

[CrazyTeeka]

Hmm. New batch of drives arrived from eBay seller. One of them, a Quantum Fireball (IDE) has a burnt chip. Wonder how that happened?

[CrazyTeeka]

It didn't take long. He just called me an ignorant Internet troll.

You know way more that Steve does.