2005-02-08 08:11:57

by Nico Schottelius

[permalink] [raw]
Subject: [PATCH] compile error: 2.6.10 / megaraid_mbox

Good morning!

I was trying to compile Megaraid on 2.6.10 and
noticed that pci_dma_sync_single and pci_dma_sync_sg
are deprecated. Greg seems to tried to patch it in 2.6.9
(http://lkml.org/lkml/2004/10/19/425), but it seems he didn't catch it
all.

A patch against vanialla 2.6.10 is attached.


Greetings,

Nico

--- megaraid_mbox.c.orig 2005-02-07 14:57:16.000000000 +0100
+++ megaraid_mbox.c 2005-02-07 15:03:00.000000000 +0100
@@ -1554,12 +1554,12 @@

if (scb->dma_direction == PCI_DMA_TODEVICE) {
if (!scb->scp->use_sg) { // sg list not used
- pci_dma_sync_single(adapter->pdev, ccb->buf_dma_h,
+ pci_dma_sync_single_for_cpu(adapter->pdev, ccb->buf_dma_h,
scb->scp->request_bufflen,
PCI_DMA_TODEVICE);
}
else {
- pci_dma_sync_sg(adapter->pdev, scb->scp->request_buffer,
+ pci_dma_sync_sg_for_cpu(adapter->pdev, scb->scp->request_buffer,
scb->scp->use_sg, PCI_DMA_TODEVICE);
}
}
@@ -2332,7 +2332,7 @@

case MRAID_DMA_WBUF:
if (scb->dma_direction == PCI_DMA_FROMDEVICE) {
- pci_dma_sync_single(adapter->pdev,
+ pci_dma_sync_single_for_cpu(adapter->pdev,
ccb->buf_dma_h,
scb->scp->request_bufflen,
PCI_DMA_FROMDEVICE);
@@ -2345,7 +2345,7 @@

case MRAID_DMA_WSG:
if (scb->dma_direction == PCI_DMA_FROMDEVICE) {
- pci_dma_sync_sg(adapter->pdev,
+ pci_dma_sync_sg_for_cpu(adapter->pdev,
scb->scp->request_buffer,
scb->scp->use_sg, PCI_DMA_FROMDEVICE);
}


2005-02-08 19:58:08

by Greg KH

[permalink] [raw]
Subject: Re: [PATCH] compile error: 2.6.10 / megaraid_mbox

On Tue, Feb 08, 2005 at 09:10:57AM +0100, Nico Schottelius wrote:
> Good morning!
>
> I was trying to compile Megaraid on 2.6.10 and
> noticed that pci_dma_sync_single and pci_dma_sync_sg
> are deprecated. Greg seems to tried to patch it in 2.6.9
> (http://lkml.org/lkml/2004/10/19/425), but it seems he didn't catch it
> all.
>
> A patch against vanialla 2.6.10 is attached.

Hm, can you redo the patch so that I can apply it with "-p1" to patch,
and add a Signed-off-by: line (also keep the same description as above.)

thanks,

greg k-h