RE: x86/random: Speculation to the rescue

[David Laight]

From: Pavel Machek
> Sent: 09 October 2019 09:03
> > NAND flash requires ECC so is likely to be async.
> > But I2C is clocked from the cpu end - so is fixed.
>
> RTC i2c may be clocked from the CPU end, but the time source needs to
> work when machine is off, so that has a separate crystal for
> timekeeping.

That only helps if the rtc chip lets you read its internal counters.
You get one read of a few bits of 'randomness'.

> > Also an embedded system could be booting off a large serial EEPROM.
> > These have fixed timings and are clocked from the cpu end.
> 
> Have you seen such system running Linux?

You can run Linux on the Nios cpu on an Altera/Intel FPGA.
The kernel is likely to be loaded from the same serial eeprom as the FPGA image.

I've not personally run such a setup, but there are examples for the dev boards
so I assume some people do.
I'm not sure I'd want to run Linux on a 100MHz cpu with a slow memory interface.
Better finding an fpga with an arm core in the corner!

(We do use the Nios cpu - but for standalone code that fits in small internal
memory blocks.)

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)