Re: ide-cs using 100% CPU

[Robert Hancock]

----- Original Message ----- 
From: "Hamie" <hamish@xxxxxxxxxxxxxxxxxx>


> Robert Hancock wrote:
>
> >It isn't that the CPU is doing so much work, it's mostly waiting. However
> >
> >
>
> That was my point... While waiting, shouldn't the CPU be off doing
> something else? Like giving X some attention...
>
> >with this type of PIO access, the CPU must do all the reads/writes from
the
> >buffer and while doing this the CPU is blocked and cannot do anything
else.
> >
> >
> >
>
> Or is the CF requirements such that it's spending it's time doing the
> actual reads & writes from the buffer, and it's the hardware inserting
> wait-states when it's being accessed?
>

That's basically what's happening, when the CPU does a write or a read of
some data from the buffer, it has to wait for that to go all the way across
the bus and to/from the card, meanwhile no other useful work can be done
while it is waiting. As Alan Cox mentioned, having a hyperthreaded CPU helps
tremendously in such cases, since the other "half" of the CPU can be doing
useful work while the first half is blocked.

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