Re: Problems with to large of a kernel

H. Peter Anvin (hpa@transmeta.com)
2 Dec 1998 00:48:36 GMT


Followup to: <m0zkxFu-0007U3C@the-village.bc.nu>
By author: alan@lxorguk.ukuu.org.uk (Alan Cox)
In newsgroup: linux.dev.kernel
>
> > is unreachable in the 16-bit segment:offset world. Since this function
> > is enabled and tested by the BIOS as part of the memory check, it is
> > not too likely that it is broken. There was a problem with stuck
>
> There have been numerous A20 gate problems with laptops and bootup - often
> the external caches werent wired to the A20 switch for example
>

I presume you mean internal caches... if A20 is done properly, i.e. it
feeds into the #A20G line on the CPU then it is applied before the L1
(internal) cache.

Unfortunately there is no standard way to flush the external caches,
unless they are wired properly and the WBINVD instruction works (on
486 and higher.) Part of the problem is that the WBINVD instruction
needs to be conditionalized, since it isn't there on a 386.

I think there is enough evidence that A20 is broken on enough machines
that putting it in the disambiguation wait loop probably makes sense.

-hpa

-- 
    PGP: 2047/2A960705 BA 03 D3 2C 14 A8 A8 BD  1E DF FE 69 EE 35 BD 74
    See http://www.zytor.com/~hpa/ for web page and full PGP public key
        I am Bahá'í -- ask me about it or see http://www.bahai.org/
   "To love another person is to see the face of God." -- Les Misérables

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/