“Intel Seamless Update” is a forthcoming feature for Intel platforms seemingly first being exposed by their new Linux kernel patches working on the functionality… Intel is working on being able to carry out system firmware updates such as UEFI updates but doing so at run-time and being able to avoid the reboot in the process.
Pretty cool, but sadly, it’s only for enterprise machines and upcoming Xeon processors.
Big corporations seem to have reinforcing the two tier “them and us” system down pat. Only when he kings in their castle have it all figured out will their deign to release it to the masses and only then a crumb at a time and at a price.
Yes I know about controlled environments and not confusing the market. But there is a suspicion in my mind this is Intel furthering their strategy of wanting to appear to be a market leader while burying their liabilities of Spectre and hoping everyone forgets about this. “Oooh look shiny!”
No, this is a classic case of you get what you pay for. Do you want thunderbolt? Then you’ll have to pay extra for some model that has the hardware configured to support it. It’s the same here, but also it’s because there are going to be almost zero home consumer market users that will actually use this feature. If you want server features, buy the more expensive server hardware.
dark2,
Outside of enterprise this would be an extremely niche feature. I already power off my home computers daily so the reboot is really not too much of an imposition.
In terms of 24×7 servers the use case is clearer. Although I don’t know that I would use it anyways. Firmware updates are relatively rare and I’m somewhat hesitant to apply potentially significant changes without planned downtime.
Reboots serve another more subtle purpose too, they test that a computer will boot from scratch. If you apply patches at run time and never test a reboot, it increases the risk that an unbootable firmware will go undetected at the time of deployment. Then when you eventually do reboot the firmware problem may come out at a very inopportune time. This could potentially be devastating if many servers are rebooted simultaneously, perhaps the result of a power failure.
Of course enterprise engineers work hard to make sure their firmware is robust, but nothing is 100%, so I think I’d still want to test with reboots.
It’s easy to criticize Intel for the Spectre and Meltdown hardware bugs, but ultimately they were oversights. It’s not like Intel knew about them before they were discovered, or that they knowingly sold insecure products for nefarious reasons. It was genuine, but massive oversight that affected multiple CPU architectures, not just Intel’s line of x86 processors.
Pinning the blame solely on Intel is unfair, especially since they’ve now actually developed a product that can mitigate issues like this through seamless firmware updates.
I wouldn’t be surprised to see this filter down into the desktop market eventually, it is just a feature most important to always on enterprise systems where uptime needs to be maximized. And you have to roll out the feature somewhere, it makes sense.
Remember when the idea of having the firmware be writeable by the OS was considered a bad idea because it leads to destructive malware such as Chernobyl/CIH?
https://en.wikipedia.org/wiki/CIH_(computer_virus)
Apparently, OS-writeable firmware now it’s considered a good idea. I am sure that whatever system Intel has in place to prevent OS malware from writing to the firmware won’t be compromised, ever.
Can’t we have one piece of software that is released in a reasonably tested state and doesn’t need constant updates? There is no reason why a non-enthusiast user should have to update their firmware. The whole point of having separate firmware is that it’s supposed to be a small, relatively simple, never-updated piece of software that boots the big, complex, updateable OS.