[PATCH 4/6] libata: Add ata_host_flags

[PATCH 4/6] libata: Add ata_host_flags

[Brian King]

Since SATA devices attached to SAS HBAs will not have
a host_set, add a helper function to retrieve the host_set
flags which can check for the existence of a host_set.

Signed-off-by: Brian King <brking@us.ibm.com>
---

 libata-dev-bjking1/drivers/scsi/libata-core.c |    7 +++----
 libata-dev-bjking1/include/linux/libata.h     |    7 +++++++
 2 files changed, 10 insertions(+), 4 deletions(-)

diff -puN drivers/scsi/libata-core.c~libata_sas_no_simplex drivers/scsi/libata-core.c
--- libata-dev/drivers/scsi/libata-core.c~libata_sas_no_simplex	2006-06-06 08:55:43.000000000 -0500
+++ libata-dev-bjking1/drivers/scsi/libata-core.c	2006-06-06 08:55:43.000000000 -0500
@@ -2161,7 +2161,7 @@ int ata_set_mode(struct ata_port *ap, st
 	/* Record simplex status. If we selected DMA then the other
 	 * host channels are not permitted to do so.
 	 */
-	if (used_dma && (ap->host_set->flags & ATA_HOST_SIMPLEX))
+	if (used_dma && (ata_host_flags(ap) & ATA_HOST_SIMPLEX))
 		ap->host_set->simplex_claimed = 1;
 
 	/* step5: chip specific finalisation */
@@ -3012,7 +3012,6 @@ static int ata_dma_blacklisted(const str
 static void ata_dev_xfermask(struct ata_device *dev)
 {
 	struct ata_port *ap = dev->ap;
-	struct ata_host_set *hs = ap->host_set;
 	unsigned long xfer_mask;
 	int i;
 
@@ -3050,8 +3049,8 @@ static void ata_dev_xfermask(struct ata_
 		ata_dev_printk(dev, KERN_WARNING,
 			       "device is on DMA blacklist, disabling DMA\n");
 
-	if (hs->flags & ATA_HOST_SIMPLEX) {
-		if (hs->simplex_claimed)
+	if (ata_host_flags(ap) & ATA_HOST_SIMPLEX) {
+		if (ap->host_set->simplex_claimed)
 			xfer_mask &= ~(ATA_MASK_MWDMA | ATA_MASK_UDMA);
 	}
 
diff -puN include/linux/libata.h~libata_sas_no_simplex include/linux/libata.h
--- libata-dev/include/linux/libata.h~libata_sas_no_simplex	2006-06-06 08:55:43.000000000 -0500
+++ libata-dev-bjking1/include/linux/libata.h	2006-06-06 08:55:43.000000000 -0500
@@ -1103,4 +1103,11 @@ static inline struct ata_port *ata_shost
 	return (struct ata_port *) &host->hostdata[0];
 }
 
+static inline unsigned int ata_host_flags(struct ata_port *ap)
+{
+	if (ap->host_set)
+		return ap->host_set->flags;
+	return 0;
+}
+
 #endif /* __LINUX_LIBATA_H__ */
_

Re: [PATCH 4/6] libata: Add ata_host_flags

[Jeff Garzik]

Brian King wrote:
> Since SATA devices attached to SAS HBAs will not have
> a host_set, add a helper function to retrieve the host_set
> flags which can check for the existence of a host_set.
> 
> Signed-off-by: Brian King <brking@us.ibm.com>
> ---
> 
>  libata-dev-bjking1/drivers/scsi/libata-core.c |    7 +++----
>  libata-dev-bjking1/include/linux/libata.h     |    7 +++++++
>  2 files changed, 10 insertions(+), 4 deletions(-)
> 
> diff -puN drivers/scsi/libata-core.c~libata_sas_no_simplex drivers/scsi/libata-core.c
> --- libata-dev/drivers/scsi/libata-core.c~libata_sas_no_simplex	2006-06-06 08:55:43.000000000 -0500
> +++ libata-dev-bjking1/drivers/scsi/libata-core.c	2006-06-06 08:55:43.000000000 -0500
> @@ -2161,7 +2161,7 @@ int ata_set_mode(struct ata_port *ap, st
>  	/* Record simplex status. If we selected DMA then the other
>  	 * host channels are not permitted to do so.
>  	 */
> -	if (used_dma && (ap->host_set->flags & ATA_HOST_SIMPLEX))
> +	if (used_dma && (ata_host_flags(ap) & ATA_HOST_SIMPLEX))

I'm not sure I like this one.  Will have to think a bit more on this -- 
there might be a need for ipr to carry some cross-port libata state.

	Jeff

Re: [PATCH 4/6] libata: Add ata_host_flags

[Brian King]

Jeff Garzik wrote:
> Brian King wrote:
>> Since SATA devices attached to SAS HBAs will not have
>> a host_set, add a helper function to retrieve the host_set
>> flags which can check for the existence of a host_set.
>>
>> Signed-off-by: Brian King <brking@us.ibm.com>
>> ---
>>
>>  libata-dev-bjking1/drivers/scsi/libata-core.c |    7 +++----
>>  libata-dev-bjking1/include/linux/libata.h     |    7 +++++++
>>  2 files changed, 10 insertions(+), 4 deletions(-)
>>
>> diff -puN drivers/scsi/libata-core.c~libata_sas_no_simplex drivers/scsi/libata-core.c
>> --- libata-dev/drivers/scsi/libata-core.c~libata_sas_no_simplex	2006-06-06 08:55:43.000000000 -0500
>> +++ libata-dev-bjking1/drivers/scsi/libata-core.c	2006-06-06 08:55:43.000000000 -0500
>> @@ -2161,7 +2161,7 @@ int ata_set_mode(struct ata_port *ap, st
>>  	/* Record simplex status. If we selected DMA then the other
>>  	 * host channels are not permitted to do so.
>>  	 */
>> -	if (used_dma && (ap->host_set->flags & ATA_HOST_SIMPLEX))
>> +	if (used_dma && (ata_host_flags(ap) & ATA_HOST_SIMPLEX))
> 
> I'm not sure I like this one.  Will have to think a bit more on this -- 
> there might be a need for ipr to carry some cross-port libata state.

I can't think of any need at this point. Would it be reasonable to go ahead
with this change and add in the support for a cross-port libata state
if and when such a need comes along? 


-- 
Brian King
eServer Storage I/O
IBM Linux Technology Center

Re: [PATCH 4/6] libata: Add ata_host_flags

[Brian King]

Brian King wrote:
> Jeff Garzik wrote:
>> Brian King wrote:
>>> Since SATA devices attached to SAS HBAs will not have
>>> a host_set, add a helper function to retrieve the host_set
>>> flags which can check for the existence of a host_set.
>>>
>>> Signed-off-by: Brian King <brking@us.ibm.com>
>>> ---
>>>
>>>  libata-dev-bjking1/drivers/scsi/libata-core.c |    7 +++----
>>>  libata-dev-bjking1/include/linux/libata.h     |    7 +++++++
>>>  2 files changed, 10 insertions(+), 4 deletions(-)
>>>
>>> diff -puN drivers/scsi/libata-core.c~libata_sas_no_simplex drivers/scsi/libata-core.c
>>> --- libata-dev/drivers/scsi/libata-core.c~libata_sas_no_simplex	2006-06-06 08:55:43.000000000 -0500
>>> +++ libata-dev-bjking1/drivers/scsi/libata-core.c	2006-06-06 08:55:43.000000000 -0500
>>> @@ -2161,7 +2161,7 @@ int ata_set_mode(struct ata_port *ap, st
>>>  	/* Record simplex status. If we selected DMA then the other
>>>  	 * host channels are not permitted to do so.
>>>  	 */
>>> -	if (used_dma && (ap->host_set->flags & ATA_HOST_SIMPLEX))
>>> +	if (used_dma && (ata_host_flags(ap) & ATA_HOST_SIMPLEX))
>> I'm not sure I like this one.  Will have to think a bit more on this -- 
>> there might be a need for ipr to carry some cross-port libata state.
> 
> I can't think of any need at this point. Would it be reasonable to go ahead
> with this change and add in the support for a cross-port libata state
> if and when such a need comes along? 

Jeff, any further thoughts about this? I had been trying to avoid having
any host_set type of data for the SAS usage of libata in part due to
not seeing a need and also due to a comment Bartlomiej made a while back:

http://marc.theaimsgroup.com/?l=linux-ide&m=112842136431977&w=2


Brian

-- 
Brian King
eServer Storage I/O
IBM Linux Technology Center

Re: [PATCH 4/6] libata: Add ata_host_flags

[Alan Cox]

Ar Llu, 2006-06-26 am 13:56 -0500, ysgrifennodd Brian King:
> Jeff, any further thoughts about this? I had been trying to avoid having
> any host_set type of data for the SAS usage of libata in part due to
> not seeing a need and also due to a comment Bartlomiej made a while back:
> 
> http://marc.theaimsgroup.com/?l=linux-ide&m=112842136431977&w=2

Well if anything we seem to be going the other way, in order to get
things like Simplex working with legacy mode the legacy setup needs the
two ports on the same host_set with differing IRQ/ports.

Alan