Do not set DMA mask to 32 bit first if adapter only supports 31

From: Rolf Eike Beer <[email protected]>
Date: Wed, 1 Jul 2009 22:43:39 +0200

---
 drivers/scsi/aacraid/linit.c |   14 ++++++++------
 1 files changed, 8 insertions(+), 6 deletions(-)

James, it looks like people at Adaptec don't care if you send patches in. That also applies to at least one patch from Julia Lawall that is waiting since December. What next?

index 33898b6..cad6f9a 100644
@@ -1091,6 +1091,7 @@ static int __devinit aac_probe_one(struct pci_dev *pdev,
 	struct list_head *insert = &aac_devices;
 	int error = -ENODEV;
 	int unique_id = 0;
+	u64 dmamask;
 
 	list_for_each_entry(aac, &aac_devices, entry) {
 		if (aac->id > unique_id)

@@ -1104,17 +1105,18 @@ static int __devinit aac_probe_one(struct pci_dev 

*pdev,
goto out;
error = -ENODEV;

- if (pci_set_dma_mask(pdev, DMA_BIT_MASK(32)) ||
- pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32)))
- goto out_disable_pdev;
/*
* If the quirk31 bit is set, the adapter needs adapter
* to driver communication memory to be allocated below 2gig */
if (aac_drivers[index].quirks & AAC_QUIRK_31BIT)
- if (pci_set_dma_mask(pdev, DMA_BIT_MASK(31)) ||
- pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(31)))
- goto out_disable_pdev;
+ dmamask = DMA_BIT_MASK(31);
+ else
+ dmamask = DMA_BIT_MASK(32);
+
+ if (pci_set_dma_mask(pdev, dmamask) ||
+ pci_set_consistent_dma_mask(pdev, dmamask))
+ goto out_disable_pdev;

pci_set_master(pdev);

1.7.1

Hi Rolf,

The patch looks good for me so from my side we can include it. Sorry for the delay - I will try to give quick answers for future requests. Please pay attention that my personal email address has changed because we are now part of PMC-Sierra Inc. But [email protected] remains active.

Thanks,
Achim Leubner
PMC-Sierra Inc.

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Rolf Eike Beer
Sent: Monday, July 26, 2010 8:08 AM
To: [email protected]
Cc: [email protected]; Andrew Morton
Subject: [PATCH][Resend #2] Do not set DMA mask to 32 bit first if adapter only supports 31

From: Rolf Eike Beer <[email protected]>
Date: Wed, 1 Jul 2009 22:43:39 +0200

---
 drivers/scsi/aacraid/linit.c |   14 ++++++++------
 1 files changed, 8 insertions(+), 6 deletions(-)

James, it looks like people at Adaptec don't care if you send patches in. That
also applies to at least one patch from Julia Lawall that is waiting since
December. What next?

index 33898b6..cad6f9a 100644
@@ -1091,6 +1091,7 @@ static int __devinit aac_probe_one(struct pci_dev

*pdev,
struct list_head *insert = &aac_devices;
int error = -ENODEV;
int unique_id = 0;
+ u64 dmamask;

list_for_each_entry(aac, &aac_devices, entry) {
if (aac->id > unique_id)
@@ -1104,17 +1105,18 @@ static int __devinit aac_probe_one(struct pci_dev
*pdev,
goto out;
error = -ENODEV;

- if (pci_set_dma_mask(pdev, DMA_BIT_MASK(32)) ||
- pci_set_consistent_dma_mask(pdev,
DMA_BIT_MASK(32)))
- goto out_disable_pdev;
/*
* If the quirk31 bit is set, the adapter needs adapter
* to driver communication memory to be allocated below 2gig */
if (aac_drivers[index].quirks & AAC_QUIRK_31BIT)
- if (pci_set_dma_mask(pdev, DMA_BIT_MASK(31)) ||
- pci_set_consistent_dma_mask(pdev,
DMA_BIT_MASK(31)))
- goto out_disable_pdev;
+ dmamask = DMA_BIT_MASK(31);
+ else
+ dmamask = DMA_BIT_MASK(32);
+
+ if (pci_set_dma_mask(pdev, dmamask) ||
+ pci_set_consistent_dma_mask(pdev, dmamask))
+ goto out_disable_pdev;

pci_set_master(pdev);

1.7.1

--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Can I get an Acked-by of any of you so James can push this for the next merge window?

Also please have a look at the patch at

http://marc.info/?l=linux-scsi&m=127472836901937&w=2

and give that an Acked-by, too.

Eike

Actually, this is sufficient for me to translate into an acked-by

James

Whatever works ;)

Eike

Looks reasonable too. I can also give an Acked-by for Julias patch.

Thanks,
Achim

-----Original Message-----
From: Rolf Eike Beer [mailto:[email protected]]
Sent: Monday, July 26, 2010 5:46 PM
To: Achim Leubner
Cc: [email protected]; [email protected]; Andrew Morton; Julia Lawall
Subject: Re: [PATCH][Resend #2] Do not set DMA mask to 32 bit first if adapter only supports 31

for

Please

are

Can I get an Acked-by of any of you so James can push this for the next merge
window?

Also please have a look at the patch at

http://marc.info/?l=linux-scsi&m=127472836901937&w=2

and give that an Acked-by, too.

Eike