Skip to contentSkip to conversation

ide: don't set hwif->dma_ops in init_dma() method

PATCH In review Sergei Shtylyov 2 participants 2 messages Dec 31, 2008 – Jan 2, 2009 libata

lore
Share & tools
lore

Sergei Shtylyov

author
#0 lore raw

PATCH 1/4 ide: don't set hwif->dma_ops in init_dma() method

Replace setting of 'hwif->dma_ops' in the 'alim16x3' and 'palm_bk3710' drivers' init_dma() methods with initializing the corresponding member of their 'struct ide_port_info' instances and remove such setting from the 'hpt366' driver that just doesn't use 'sff_dma_ops'. Along with some code save, this prepares us for the next patch...

Signed-off-by Sergei Shtylyov <[email protected]>

---
Here are the unfortunate series of patches that hadn't gotten into 2.6.28 -- I decided not to leave it to the next year despite an extreme overload I had during the past several months... :-)
The patch is atop of the current pata-2.6 series...

 drivers/ide/alim15x3.c    |    3 +--
 drivers/ide/hpt366.c      |    2 --
 drivers/ide/palm_bk3710.c |    3 +--
 3 files changed, 2 insertions(+), 6 deletions(-)
drivers/ide/alim15x3.c +1 -2
Index: linux-2.6/drivers/ide/alim15x3.c
===================================================================
@@ -490,8 +490,6 @@ static int __devinit init_dma_ali15x3(id
 	if (ide_allocate_dma_engine(hwif))
 		return -1;
 
-	hwif->dma_ops = &sff_dma_ops;
-
 	return 0;
 }
 

@@ -519,6 +517,7 @@ static const struct ide_port_info ali15x
 	.init_hwif	= init_hwif_ali15x3,
 	.init_dma	= init_dma_ali15x3,
 	.port_ops	= &ali_port_ops,
+	.dma_ops	= &sff_dma_ops,
 	.pio_mask	= ATA_PIO5,
 	.swdma_mask	= ATA_SWDMA2,
 	.mwdma_mask	= ATA_MWDMA2,
drivers/ide/hpt366.c +0 -2
Index: linux-2.6/drivers/ide/hpt366.c
===================================================================
@@ -1349,8 +1349,6 @@ static int __devinit init_dma_hpt366(ide
 	if (ide_allocate_dma_engine(hwif))
 		return -1;
 
-	hwif->dma_ops = &sff_dma_ops;
-
 	return 0;
 }
drivers/ide/palm_bk3710.c +1 -2
Index: linux-2.6/drivers/ide/palm_bk3710.c
===================================================================
@@ -324,8 +324,6 @@ static int __devinit palm_bk3710_init_dm
 
 	hwif->dma_base = hwif->io_ports.data_addr - IDE_PALM_ATA_PRI_REG_OFFSET;
 
-	hwif->dma_ops = &sff_dma_ops;
-
 	return 0;
 }
 

@@ -338,6 +336,7 @@ static const struct ide_port_ops palm_bk
 static struct ide_port_info __devinitdata palm_bk3710_port_info = {
 	.init_dma		= palm_bk3710_init_dma,
 	.port_ops		= &palm_bk3710_ports_ops,
+	.dma_ops		= &sff_dma_ops,
 	.host_flags		= IDE_HFLAG_MMIO,
 	.pio_mask		= ATA_PIO4,
 	.mwdma_mask		= ATA_MWDMA2,
#1 lore raw
Quoted 6 lines · Sergei Shtylyov

Replace setting of 'hwif->dma_ops' in the 'alim16x3' and 'palm_bk3710' drivers' init_dma() methods with initializing the corresponding member of their 'struct ide_port_info' instances and remove such setting from the 'hpt366' driver that just doesn't use 'sff_dma_ops'. Along with some code save, this prepares us for the next patch...

Quoted 6 lines
Signed-off-by Sergei Shtylyov <[email protected]>

---
Here are the unfortunate series of patches that hadn't gotten into 2.6.28 -- I decided not to leave it to the next year despite an extreme overload I had during the past several months... :-)

Cool, I applied whole series (it indeed makes code a lot saner/cleaner :).