Jepp, the party goes on.
As was already prophecised by knowledgeable people in jan/feb last year, here is not a single security flaw that can be workarounded or fixed, but an essential design weakness that opens up a huge can of worms and will keep us occupied for years to come.
While I agree to most in this well-argued-for text by
ralphbsz , there is one important thing to consider:
The general concept of interaction of caches with side-channels has been known for at least a decade; I discussed it with colleagues in the context of storage subsystems (disks and RAID controllers) a long time ago, but I was not involved in CPU design. We decided that the risk and bandwidth were so ludicrously low for storage that it wasn't worth thinking about. Maybe people involved in CPU design also thought about it and rejected it being a significant risk; or maybe they didn't think about it. The important thing to remember is: those people are not evil, nor stupid. They are trying really hard to give CPU customers that what CPU customers have been clamoring for: chips that are "good, cheap, fast".
"Those people" do not exist as a uniform entity. Companies like Intel employ many engineers, and people are esentially different from each other. So, while many of them might not have thought about the issue, for a few of them, those with a stronger tendency to think-outside-the-box, this huge can of worms would without doubt have somehow adumbrated (just like
ralphbsz got the idea to consider it concerning those storage devices).
But then, at that point, what should they have done? To whom should they have talked?
The engineering consequences at that point would have been the same as they are today: to really get rid of the general issue, a fundamental re-design of the cpu-architecture (as it has developed over the last 50 years) would be necessary - a task that is practically impossible for a single company.
So even if the whole intel shop would have known about the issue, there was no option to do much about it. The product roadmap is already fixed for the next years, the investments are done:
the business must go on, or we all loose our money! There is
no alternative.
And this is not the only such development. Our commodities are under an increasing danger of cyber-attacks, and with the IoT virtually everything will come under such danger. At the same time, producers' responsibilities are delegated away over dozens of daisy-chained sub-contractors distributed all over the globe, so that you no longer can figure out anybody who would be respnsible for any decisions made. And all this happens, because there is
no alternative (to the ever-increasing speed of the globalized turbocapitalistic madness).
