linux-ide.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* MSI question
@ 2005-07-10  0:07 kallol
  0 siblings, 0 replies; 3+ messages in thread
From: kallol @ 2005-07-10  0:07 UTC (permalink / raw)
  To: tom.l.nguyen; +Cc: Jeff Garzik, linux-ide@vger.kernel.org

Hello Tom,
         I have been trying to enable MSI interrupt on a MIPs based network
processor plaform.

The adapter is based on Prommise PCI-X SATA controller. The configuration
is described below

CPU --- HOST Bridge ---- device bus ---- PCI-X bridge --- PCI-X slot


The PCI-X bridge is programmed to accept write transactions (address, and data)
from the Promise adapter and convert them to generate an interrupt to the PIC
that is on the host bridge.

I have been successful to generate one MSI interrupt. The CPU gets the
interrupt.
With a PCI-X analyzer I see the Promise adapter is writing a predefined pattern
(0x5a5a) to a predefined address. The PCI-X bridge claims the transaction and
then I see an interrupt on the CPU.


The problem is that the card generates the MSI transaction only once. I guess
we have to enable the card to generate another MSI transaction again, right?

I do not have access to the card's manual, so not sure what has to be done to
get MSI working again.

On this system, the pin based interrupt mechanism is not working yet, so MSI has
to work.

Any reply will be greatly appreciated.

Kallol



^ permalink raw reply	[flat|nested] 3+ messages in thread

* RE: MSI question
@ 2005-07-18 17:37 Nguyen, Tom L
  2005-07-19 14:50 ` Kallol Biswas
  0 siblings, 1 reply; 3+ messages in thread
From: Nguyen, Tom L @ 2005-07-18 17:37 UTC (permalink / raw)
  To: kallol; +Cc: Jeff Garzik, linux-ide

On Saturday, July 09, 2005 5:08 PM, Kallol wrote:
> The problem is that the card generates the MSI transaction only once.
I
> guess we have to enable the card to generate another MSI transaction 
> again, right?

Have you verified whether the Promise adapter driver's interrupt service
routine associated with MSI vector is called once the card generates the
MSI? Does this driver handle MSI generated from its device
appropriately? If yes, you can enable the card to generate another MSI
transaction.

Thanks,
Long

^ permalink raw reply	[flat|nested] 3+ messages in thread

* RE: MSI question
  2005-07-18 17:37 MSI question Nguyen, Tom L
@ 2005-07-19 14:50 ` Kallol Biswas
  0 siblings, 0 replies; 3+ messages in thread
From: Kallol Biswas @ 2005-07-19 14:50 UTC (permalink / raw)
  To: Nguyen, Tom L; +Cc: Jeff Garzik, linux-ide

Hello Long,
          Here are the  steps I had followed:

1) Enabled MSI on the card, also programed the address and data
register.
2) Programmed the bridge  to accept the transaction to the address with
the data and convert them to a MSI interrupt.
3) Changed the pic_ack routine in kernel to acknowledge the MSI
interrupt indication.
4) Changed the isr to return IRQ handled always.

The  PCI-X bus analyzer showed that the card did generate the write
transaction for the interrupt.

The bridge converted the write and set a MSI interrupt indication
on its interrupt status register.

The driver isr got called, then the pic_ack routine cleared the MSI
interrupt indication on the bridge.

At this point, the card put more status to the memory, but no more
MSI transaction.

I did not get time to check if MSI got disabled on the adapter. Not
sure how to re-enable the MSI. I do not have access to the adapter's
manual. I will try to set the MSI enable bit again on config space.

Thanks for help,
Kallol

On Mon, 2005-07-18 at 10:37, Nguyen, Tom L wrote:
> On Saturday, July 09, 2005 5:08 PM, Kallol wrote:
> > The problem is that the card generates the MSI transaction only once.
> I
> > guess we have to enable the card to generate another MSI transaction 
> > again, right?
> 
> Have you verified whether the Promise adapter driver's interrupt service
> routine associated with MSI vector is called once the card generates the
> MSI? Does this driver handle MSI generated from its device
> appropriately? If yes, you can enable the card to generate another MSI
> transaction.
> 
> Thanks,
> Long
> -
> To unsubscribe from this list: send the line "unsubscribe linux-ide" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2005-07-19 14:53 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-07-18 17:37 MSI question Nguyen, Tom L
2005-07-19 14:50 ` Kallol Biswas
  -- strict thread matches above, loose matches on Subject: below --
2005-07-10  0:07 kallol

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).