On Sun, Nov 20, 2005 at 02:06:24PM -0600, Bob Brose wrote:
> Hi All,
> Maybe the esp problem is timing related. I got an interesting result when I
> moved the kernel from compiling on an OSX machine to directly on the performa.
> With the same tree, the same .config BUT a different compiler (osx=2.95,
> performa = 3.3.5) I get somewhat different results! (these are with
> both of Finn's latest b and c patches:
>
> patch_b-mac68k_cvs_via_cleanup_and_fix2.diff
> patch_c-interrupts-again-again3.diff
>
> The kernel compiled on the performa with the adb active spits out the
> esp error MUCH less. In fact I had to do a sg_verify on the whole drive
> to get a "esp0: Warning, live target 0 not responding to selection" message.
> And even at that it went several thousands sectors into the drive before
> throwing the message as opposed to a few dozen with the kernel compiled on
> OSX.

That seems odd. Could you send me both kernels (in a private email) so I
can take a look at them? It's possible one compiler or the other has a
bug in the code generation. It's been known to happen before.

> Also another odd thing, from adb. I was holding down the up arrow key to go
> through the history under bash and the computer though at some point that I
> never released the key. I couldn't do anything to get it to stop by repushing
> the down key or any other key for that matter.

Do any other ADB events register? If no other keys do anything at all, it's
possible that the driver lost communication with the bus, and isn't seeing
anything to end the key repeat. Most likely this would just be a bug in
the driver.

Brad Boyer
flar@...

_______________________________________________
linux-mac68k mailing list
linux-mac68k@...
http://lists.purplehat.net/listinfo/linux-mac68k