Re: Moving a tempo mark to the right

[David Kastrup]

Anthony Youngman <address@hidden> writes:

> On 01/09/17 22:14, David Kastrup wrote:
>> Change the memory for known good memory, and the kernel
>> compiled fine.  No idea what Gcc does that memory test programs fail to
>> account for.
>
> Have you come across the memory smashing exploit? I can't remember
> much about it, but if you can hammer memory in your own VM, you can
> actually corrupt memory in the next door VM. Even worse, you can
> control the corruption with the intent of hacking into the VM!
>
> I'm pretty certain there are proofs of concept out there. So I guess
> gcc might be doing exactly that by accident to cheap memory (my RAM is
> the "value" brand - paid about £13 per 4GB stick). When I think back
> the first memory I ever bought was £50 for a 32*M*B stick :-)

The first memory I bought was about DM800 for 64kByte and I had to
solder it to the PCB myself, all 32 DIP packages.  And it needed +12V
and -5V rails in addition to the +5V rail to run.

The first external data media I worked with was cardboard, and the
biggest longterm data storage peril were mice.  And I don't mean the
input devices but the rodents.

-- 
David Kastrup