DragonFly BSD developer stung by Opteron bug

Matthew Dillon, the lead developer behind the DragonFly BSD fork of the open source FreeBSD Unix variant, had some issues with crashes on Opteron-based systems running his operating system for more than a year - and now Advanced Micro Devices says it's a bug in earlier generations of Opteron processors. The errata does not …

[Notas Badoff]

"While there are millions of these processors in the field, no other cases have been reported."

'We have millions of people dying, but no one has found the causative agent yet, so it must just be bad lifestyle choices.' Medicine, before viruses were 'discovered'.

With something so esoteric how many people *could* generate a well-defined, well-described 'case'? That only _then_ are you forced to admit is true? Until then a company just falls back on 'have you tried powering it off, then on?'

"To see this observation multiple events needed to happen concurrently and required a certain BSD-derivative environment (BSD is based on the Unix operating system) that uses a unique GCC compiler build."

You fervently hope...

[Tom 38]

Spot on. AMD are trying to make out like Matt is a nutty weirdo doing crazy things with their processor. He's not, their processor has an errata. End of story.

[olli]

All processors have erratas. I have yet (in the embedded field, mostly using PowerPC) to see anyone without an errata.

[Ilgaz]

Sony rootkit

Sony CD scandal may have happened millions of CDs later if those unlucky bastards didn't pick the band World's most advanced Windows kernel hacker happens to be fan of.

Who do they expect report from? I mean which user or how many developers there can spot the issue?

For lots of people (including developers), it is "ah damn thing crashed"

[Anonymous Coward]

just be bad lifestyle choices

... and having done the research, we now know that the top killers, heart disease and cancer are essentially life-style driven.

The irony...

[IT Builders LLC]

Re: just be bad lifestyle choices

Largely because we no longer face untreatable pandemics like bubonic plague, cholera, influenza, leprosy, malaria, measles, smallpox, syphilis, tuberculosis, typhoid fever, typhus, yellow fever...

[Anonymous Coward]

Or maybe...

... they meant to say that nobody needs to rush out to their supplier to have their CPUs replaced now.

They're not saying they're ignoring the report, obviously they are taking it seriously since they are publishing an errata and announcing.

They're not forced to admit it's true: they didn't know about it themselves, they are *confirming* it's true.

Your metaphor is very inaccurate: death concerns everybody. This particular small bug impacts only a very small number of users (because if all those millions of processors were crashing, we'd have heard of it).

No, you cannot blame all of Windows' BSOD on that single bug.

Sheesh, some people have a knack for turning a company publishing a fix into some kind of conspiracy.

[Anonymous Coward]

This sounds similar to 94

If amd press office/ contact read some cpu history, they wouldn't sound like Intel back in 1994

" Publicly, Intel acknowledged the floating point flaw but claimed that it was not serious and would not affect most users."

Understand why people are a bit touchy when you acknowledge the issue in similar language?

Also opteron is a server/hpc optimized chip. Owners of these systems (especially *bsd camp) rarely makes huge changes like changing compiler.

[Ramazan]

@Matthew Dillon

My kudos go to Matthew, who has spent a year in debugging hell. Back in 90's they used ICEs for the purpose and it was much easier then (http://www.rcollins.org/ddj/Jul97/Jul97.html)