Linux 5.3 Moves Ahead With No Longer Advertising RdRand On Older AMD CPUs/APUs

[phoronix]

Phoronix: Linux 5.3 Moves Ahead With No Longer Advertising RdRand On Older AMD CPUs/APUs

Just prior to yesterday's Linux 5.3-rc6 kernel release, a change was pulled into the code-base that disables the advertising of RdRand support on older AMD CPUs/APUs...

http://www.phoronix.com/scan.php?page=news_item&px=Linux-5.3-Disabled-Older-AMD-Rd

[muncrief]

I've had a Gigabyte GA-990FXA-UD5 motherboard with an FX-6300 CPU for years, and had a few suspend/resume problems when Mesa didn't fully support my Sapphire Nitro R9-390 GPU, but I've never had any problems other than those that could be attributed to the GPU. So I guess I'm just lucky? But I'm waiting for a back ordered Ryzen CPU to finally replace the system so I guess it doesn't matter anymore.

In any case I compiled 5.3-rc6 on Manjaro yesterday and have been running it since then and haven't experienced any problems.

[AndyChow]

I've had a Gigabyte GA-990FXA-UD5 motherboard with an FX-6300 CPU for years, and had a few suspend/resume problems when Mesa didn't fully support my Sapphire Nitro R9-390 GPU, but I've never had any problems other than those that could be attributed to the GPU. So I guess I'm just lucky? But I'm waiting for a back ordered Ryzen CPU to finally replace the system so I guess it doesn't matter anymore.

In any case I compiled 5.3-rc6 on Manjaro yesterday and have been running it since then and haven't experienced any problems.

You won't experience any problems. It will just take longer to generate random numbers whenever code you run asks for random numbers. Which is usually rare.

[ElectricPrism]

Misread title as Linux 5.3 is depreciating RandR, I'm like hmm? okay? Then read post and comments and am like RdRand. Sometimes these naming schemas make me feel the fatigue.

I almost expect a future where there's KDE, Kay DE, KED and other confusing variations just to fuck wit cha.

[nanonyme]

You won't experience any problems. It will just take longer to generate random numbers whenever code you run asks for random numbers. Which is usually rare.

Well, even that is more common than it really happens. Your system also needs to be low on entropy. You don't typically hit this at all expect maybe during boot on a VM.

[starshipeleven]

Neat, so the ryzen rdrand bug was not only in Ryzen.

[xpue]

AMD FineWine™

[ms178]

AMD FineWine™

As was mentioned, it is not AMD's fault but buggy BIOS implementations. If you want to blame/shame someone, blame/shame them instead.

[Ardje]

You won't experience any problems. It will just take longer to generate random numbers whenever code you run asks for random numbers. Which is usually rare.

Actually the (pseudo) random number generator will keep on running full speed. Just it has less sources to do reseeding.
Only applications that explicitly use these instructions as the single source of randomness can be hit by this, but were probably flawed to begin with. There is only a single moment that the random generator on linux is slightly predictable, and that's the first 60s after boot, and that's highly depending on all the devices that add to the PRNG reseeding. A notebook might not even have this problem.

[torsionbar28]

Misread title as Linux 5.3 is depreciating RandR, I'm like hmm? okay? Then read post and comments and am like RdRand. Sometimes these naming schemas make me feel the fatigue.

I almost expect a future where there's KDE, Kay DE, KED and other confusing variations just to fuck wit cha.

Lunix kernel anyone?