Re: [Xen-devel] [PATCH 0/4] ns16550: enable support for Pericom controllers

[Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>]

On Tue, Feb 23, 2016 at 04:22:35AM -0700, Jan Beulich wrote:
> Patches 1 and 2 are meant to go in. Patch 3 is a prerequisite to patch
> 4 and may go in as well, but patch 4 is RFC because with the Pericom
> board I have MSI doesn't appear to function. Since it also does not
> work in baremetal Linux when doing the trivial adjustments needed in
> its driver, I suspect the feature doesn't work in general, which is
> supported by the observation that the device continues to assert
> INTx despite the MSI enable bit being set (causing unclaimed IRQs
> until that IRQ gets shut off). While I got some responses back from
> Pericom support, no actual statement of whether MSI is actually
> known to work on their boards was ever made by them. I _think_
> patch 4 is correct (and hence could go in), but I have no way of
> proving this by testing.
> 
> 1: ns16550: store pointer to config parameters for PCI
> 2: ns16550: enable Pericom controller support
> 3: console: adjust IRQ initialization
> 4: ns16550: enable use of PCI MSI
> 
> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
> 
> (Konrad, I'd appreciate if you could double check that I didn't
> accidentally break the Oxford controller support.)

Tested-by: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>

I tested it on my one-port serial card. Let me move the cables + card
tomorrow around to use the two port card and see how it behaves.

CC-ing Boris
Boris, I think the two port serial card is tst005 or such - so it may get wonky.

> 

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel