On Wed, Jul 25, 2001 at 12:40:05PM -0400, Andrew Cagney wrote:
> [...]
> while the new file had that test moved to near the end of the function. 
>   Looking at my dmesg:
> 
> cd0(wdc1:0:0): using PIO mode 4, DMA mode 2 (using DMA data transfers)
> wdc2 at obio0 offset 0x21000 irq 21: DMA transfer
> piomode -1, dmamode -1
> 
> one begins to suspect that the code might have been doing something 
> useful.  Examining 1.13, you find that it is going to eventually execute 
> the statements:
> 
>          min_cycle = pio_timing[piomode].cycle;
>          min_active = pio_timing[piomode].active;
> 
> where piomode==-1.  Er, outch!!

No, in your dmesg, piomode==-1 for wdc2, which has no drives attached.
This shouldn't matter much for wdc0 or wdc1
Still reading an array with negative offset isn't good. I'll correct this.

--
Manuel Bouyer <bouyer@antioche.eu.org>
--