Re: /dev/mem

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 





On Thu, Oct 28, 2010 at 2:56 PM, StephanT <stman937-linewbie@xxxxxxxxx> wrote:
John,

> I do not plan to argue this further in advance.

:-)   It wasn't in my intention to argue anything. Just knowledge exchange.

Why yes, a very fine line indeed.
 

> In a microkernel drivers could have their own global namespace and not have
>direct access to memory.

As long as the processor doesn't provide any specific mechanism to protect the
memory access from itself - I mean here "kernel-mode" - namespace or not the
code executing in kernel mode will do whatever it wants with the memory. Even
if the driver has not direct access to memory some other code has to have it.
The driver will request memory modification  via a message. If the request is
malicious the executing part has no way to know it. Or if it does it means the
memory management is in kernel mode - and this is no longer a microkernel ...

In my opinion the solution is not to isolate the memory access from the
executing code but to provide some protection mechanism in H/W. The MMU
does half this work today. It remains to invent the other half.

I thought you were aware of such mechanism when I asked you to elaborate...

Sorry if my knowledge was less then expected :)  My thinking was that having the insmod command 
is the same as having /dev/mem in a Linux kernel.  I am a constant student of the kernel and after 10 years
I realize every year that the more I learn the less I know!  The biggest mistake you can make in Linux 
kernel programming is assuming you already know something.
 

Anyhow, thanks for sharing.


And thank you back...

--
John

[Index of Archives]     [Newbies FAQ]     [Linux Kernel Mentors]     [Linux Kernel Development]     [IETF Annouce]     [Git]     [Networking]     [Security]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux RAID]     [Linux SCSI]     [Linux ACPI]
  Powered by Linux