Thread (14 messages) 14 messages, 5 authors, 2007-01-27

Re: [PATCH 4/4] atl1: Ancillary C files for Attansic L1 driver

From: Jeff Garzik <hidden>
Date: 2007-01-23 22:07:08
Also in: lkml

Randy Dunlap wrote:
Stephen Hemminger wrote:
quoted
On Tue, 23 Jan 2007 16:33:29 -0500
Jeff Garzik [off-list ref] wrote:
quoted
Stephen Hemminger wrote:
quoted
IMHO the MSI disabling should be removed from drivers and be done
in the PCI core.
That is the consensus opinion.

Currently drivers implement the MSI tests because the core PCI code 
hasn't been up to snuff.  I (and others) have been discouraging that, 
but when a user faces a choice between working and non-working 
network, the pragmatic solution wins.
Linus's remark (IIRC) was to not enable CONFIG_PCI_MSI then.
Most distros, especially cutting edge ones like Fedora, will enable 
CONFIG_PCI_MSI.

quoted
quoted
All efforts to get us to the point where we can remove the MSI tests 
from drivers are strongly supported...

    Jeff
So far, either MSI works for all devices or is broken, so it makes
sense to have a "msi=off" boot option (if there isn't already)
There is one, but it's spelled "pci=nomsi".
Yep.  Thanks for repeating this, and refreshing the collective memory.

	Jeff


Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help