Re: CPU Cost of KernelOControl
From: David Liao \(MS\) (davli_at_online.microsoft.com)
Date: 01/25/05
- Next message: David Liao \(MS\): "Re: Xscale PXA255 UART Driver"
- Previous message: Dan McCarty: "CPU Cost of KernelOControl"
- In reply to: Dan McCarty: "CPU Cost of KernelOControl"
- Next in thread: Dan McCarty: "Re: CPU Cost of KernelOControl"
- Reply: Dan McCarty: "Re: CPU Cost of KernelOControl"
- Messages sorted by: [ date ] [ thread ]
Date: Tue, 25 Jan 2005 11:51:27 -0800
It is PSL call. It obviously cause CPU time to make any PSL call.
Is this IO is shared between your application and kernel ? Why you need
KernelIOControl to read data from IO?
David Liao
"Dan McCarty" <dmccarty@gmail.com> wrote in message
news:1106681739.449895.285250@z14g2000cwz.googlegroups.com...
> Could anyone jump in and help settle an argument?
>
> What is the cost of calling KernelIOControl with custom IOCTL messages?
> Is it high? Is it negligible? I'm in the middle of an argument with
> no clear end in sight.
>
> I'd rather call KernelIOControl() many times to retrieve small
> structures of data being read from I/O. My illustrious opponents <g>
> would rather call a just a few times and retrieve huge structures, even
> though we may not need all that data.
>
> Any suggestions as to the better approach? Any ideas would be
> appreciated.
>
> TIA,
> Dan.
>
- Next message: David Liao \(MS\): "Re: Xscale PXA255 UART Driver"
- Previous message: Dan McCarty: "CPU Cost of KernelOControl"
- In reply to: Dan McCarty: "CPU Cost of KernelOControl"
- Next in thread: Dan McCarty: "Re: CPU Cost of KernelOControl"
- Reply: Dan McCarty: "Re: CPU Cost of KernelOControl"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
|