Solved! (was: Dreadful xfer rates and TCPv4 bad checksum ....)

David G Hamblen (dave@afr-olt.com)
Sat, 7 Nov 1998 16:02:14 -0500 (EST)


On Sat, 7 Nov 1998, David G Hamblen wrote:

> On Sat, 7 Nov 1998, Alan Cox wrote:
>
> > > Yes, but the 16450 works on Win95, linux 2.0.33, and linux 2.1.xxx
> > > thru the mid 80's.
> >
> > Do you know which actual 2.1.8x broke it. That could be important since
> > it will identify which set of changes caused the problem
> >
>
>
> No, that was just a guess. I was running the 1.1.xxx series, and then

(snip)

Problem solved. It's the hdparm's on my ide drive. 2.1.125 defaults to
-i16 (multi-sector reads?), while 2.0.33 defaults to 0 (off). Turning on
the interrupt masking stuff with -u1 fixes it. I can make 2.0.33 go bad
by doing an `hdparm -i16 /dev/hda`, and can fix both kernels by executing
either an `hdparm -u1 /dev/hda` or hdparm -i0 /dev/hda`.

Thanks to all for the responses.

dave

-
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/