[Prism54-devel] Re: Interrupt Polling

Rodel Miguel rodel.miguel at eazix.com
Wed Aug 11 10:41:05 UTC 2004

Hi Margit,

How about performance? There are lots of workaround on the firmware
loading and hotplugging, i.e. initialization routines.



-----Original Message-----
From: Andriy Korud [mailto:a.korud at vector.com.pl] 
Sent: Wednesday, August 11, 2004 6:21 PM
To: Margit Schubert-While; Jim Thompson
Cc: prism54-devel at prism54.org; Rodel Miguel
Subject: RE: [Prism54-devel] Re: Interrupt Polling

Just my 2 cents: 
prism54 is working perfectly on PowerPC405/Embedded (however we're using
normal Linux, not uClinux)


-----Original Message-----
From:	prism54-devel-bounces at prism54.org on behalf of Margit
Sent:	Wed 11.08.2004 12:11
To:	Jim Thompson
Cc:	prism54-devel at prism54.org; rodel.miguel at eazix.com
Subject:	Re: [Prism54-devel] Re: Interrupt Polling

Jim scribeth:

>On Aug 10, 2004, at 11:23 PM, Margit Schubert-While wrote:
>>Jim scribeth:
>>>>Rodel, the current Prism54 driver is not particularly suited for 
>>>>embedded devices.
>>>Care to explain why?
>>Sure, the current hotplug mechanism for starters.
>I'm not sure why this is a show-stopper.
>I'm given to understand that the prism54 driver isn't that stable on 
>non-x86 architectures, and this certainly needs to be fixed in the
>term (if true).

What makes you think that it isn't stable on non-x86 arches ?

What do you mean "needs to be fixed" ?
The hotplug firmware loading mechanism in 2.4 and 2.6 is the preferred
way of doing things. It's just not suitable for embedded devices that
lack the hotplug infrastructure.


Prism54-devel mailing list
Prism54-devel at prism54.org

More information about the Prism54-devel mailing list