Re: Using Asyn driver or modifying devLib (or related programs) of EPICSbase for any knid of cPCI cards

[Andrew Johnson <anj@aps.anl.gov>]

Jenny Chen wrote:
> We may have a cPCI crate with an Intel CPU and four DI/DO/AI/AO cards
> soon. The OS might be Fedora Linux2.6. Before we write any Asyn
> driver for EPICS, I think we need to know how to write/install the
> driver and implement test programs for the cards.

It would certainly be wise to understand the hardware before writing an 
Asyn driver for it.  However you should probably have at least some 
understanding of how an Asyn driver behaves before you finish the kernel 
driver for the hardware.

> We purchase the cards locally (might be ADLink).

Make sure that you check out the Comedi website that Emmanual mentioned, 
their supported hardware mentions a number of ADLink boards for which 
drivers are available: http://www.comedi.org/hardware.html

Has anyone else written Asyn drivers for boards using Comedi?

> We might also buy
> some other cards later (might from other company). The Linux kernel 
> 2.6 and 2.4 are quite different, right?

Correct, the internal Linux kernel APIs used by kernel drivers are 
significantly different between 2.4 and 2.6, and can change between 
different minor releases of 2.6 (2.6.23 is very different to 2.6.1).

> The way to install driver for cPCI cards is similar or not?

A cPCI driver will be very similar to any other in-kernel PCI driver.

> One more question, is it difficult to handle interrupt from cPCI
> cards?

As Emmanuel has suggested you probably don't want to try and bring 
interrupt information all the way up to user-space â the various 
mechanisms available for signaling interrupts to user-space are 
significantly slower than handling the interrupt completely inside your 
kernel driver, but debugging code that runs in the kernel can be much 
harder than code running in user-space.

Hopefully the Comedi drivers will save you from having to write much code.

- Andrew
--
When a distinguished but elderly scientist states that something is
possible, he is almost certainly right.  When he states that something
is impossible, he is very probably wrong.  -- Arthur C. Clarke