cvs commit: src/sys/dev/pci pci.c

[M. Warner Losh]

In message: <20040702142635.F63950 at odysseus.silby.com>
            Mike Silbersack <silby at silby.com> writes:
: 
: On Fri, 2 Jul 2004, Warner Losh wrote:
: 
: > imp         2004-07-02 13:42:36 UTC
: >
: >  FreeBSD src repository
: >
: >  Modified files:
: >    sys/dev/pci          pci.c
: >  Log:
: >  Disable native ata support for now, too much breaks
: 
: So that the rest of us know what to look for, could you give a little more 
: detail?  Does native ata mode provide any benefits that would be 
: noticeable over legacy mode?  Also, when it breaks, does that just mean 
: the devices don't probe, or does it mean that data corruption might occur?

Native ata mode gives little to no benefit over non-native mode.  It
allows one to map the device's resource better in terms of overall
system resource usage, but that's a minor addition.  I don't think
that there is any other wins to doing that.

Devices either don't probe or don't work.  I don't think that you'd
see 'working mostly, with some data corruption'.

The big win comes in when we have devices that don't run in legacy
mode at all...

: I ask because I have a CD-ROM that has never worked in DMA mode, and I was 
: wondering if native mode might fix that.  (After jhb fixed my pccard+ACPI 
: problem by adding better EISA support, I'm optimistic about everything.)

I doubt it.  Feel free to enable the ifdef and try :-)

Warner