[PATCH] tms380tr: move to DMA API

This patch makes tms380tr use the new DMA API.  Now that on Alpha, this API
also supports bus master DMA for ISA (platform) devices, i changed the
driver to use this new API.

This also works around a bug in the firmware loader: The example provided
in Documentation/firmware_class no longer works, as the firmware loader now
calls get_kobj_path_length() and the kernel promptly oopses, as the
home-grown device doesn't have a parent.  Of course, this doesn't happen
with a "real" device which has its bus (or pseudo bus in the case of
platform) as parent.

Converted tms380tr to use new DMA API:
  - proteon.c, skisa.c: use platform pseudo bus to create a struct device
  - Space.c: delete init hooks
  - abyss.c, tmspci.c: pass struct device to tms380tr.c
  - tms380tr.c, tms380tr.h: new DMA API, use real device fo firmware loader

Signed-off-by: Jochen Friedrich <jochen@scram.de>
Cc: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

1 files changed, 2 insertions, 2 deletions

diff --git a/drivers/net/tokenring/tmspci.c b/drivers/net/tokenring/tmspci.c
index 2e18c0a..0014aef 100644
--- a/drivers/net/tokenring/tmspci.c
+++ b/drivers/net/tokenring/tmspci.c
@@ -100,7 +100,7 @@ static int __devinit tms_pci_attach(struct pci_dev *pdev, const struct pci_devic
 	unsigned int pci_irq_line;
 	unsigned long pci_ioaddr;
 	struct card_info *cardinfo = &card_info_table[ent->driver_data];
-		
+
 	if (versionprinted++ == 0)
 		printk("%s", version);
 
@@ -143,7 +143,7 @@ static int __devinit tms_pci_attach(struct pci_dev *pdev, const struct pci_devic
 		printk(":%2.2x", dev->dev_addr[i]);
 	printk("\n");
 		
-	ret = tmsdev_init(dev, PCI_MAX_ADDRESS, pdev);
+	ret = tmsdev_init(dev, PCI_MAX_ADDRESS, &pdev->dev);
 	if (ret) {
 		printk("%s: unable to get memory for dev->priv.\n", dev->name);
 		goto err_out_irq;