Skip to contentSkip to conversation

skge: limit DMA to 32 bit

RFC In review Stanislaw Gruszka 2 participants 3 messages Sep 21, 2010 – Sep 22, 2010

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

RFC 1/1 skge: limit DMA to 32 bit

We have old fedora bug report about not working skge device on some Gigabyte motherboard:
https://bugzilla.redhat.com/show_bug.cgi?id=447489

In dmesg there are messages:

skge 0000:02:0b.0: PCI error cmd=0x7 status=0x22b0 skge 0000:02:0b.0: unable to clear error (so ignoring them)

Status value 0x22b0 include PCI_STATUS_REC_MASTER_ABORT bit set, what mean that DMA transfer on PCI bus was aborted.

I provided to user modified driver with module parameter, that allow to experiment with different dma_mask's. He told that maximum working dma_mask for him is 32 bits set.

Is 32bit DMA limit for every skge device acceptable patch?

Note, patch change possible driver DMA address space as below:
- 32bit - no change, still low mem only (896MB)
- 32bit-PAE - down from 36bit to low mem (896MB)
- 64bit - down from 64bit to 32bit

drivers/net/skge.c +6 -14
index a8a6358..1a674e3 100644
@@ -3797,8 +3797,7 @@ static const struct net_device_ops skge_netdev_ops = {
 
 
 /* Initialize network device */
-static struct net_device *skge_devinit(struct skge_hw *hw, int port,
-				       int highmem)
+static struct net_device *skge_devinit(struct skge_hw *hw, int port)
 {
 	struct skge_port *skge;
 	struct net_device *dev = alloc_etherdev(sizeof(*skge));
@@ -3814,9 +3813,6 @@ static struct net_device *skge_devinit(struct skge_hw *hw, int port,
 	dev->watchdog_timeo = TX_WATCHDOG;
 	dev->irq = hw->pdev->irq;
 
-	if (highmem)
-		dev->features |= NETIF_F_HIGHDMA;
-
 	skge = netdev_priv(dev);
 	netif_napi_add(dev, &skge->napi, skge_poll, NAPI_WEIGHT);
 	skge->netdev = dev;
@@ -3874,7 +3870,7 @@ static int __devinit skge_probe(struct pci_dev *pdev,
 {
 	struct net_device *dev, *dev1;
 	struct skge_hw *hw;
-	int err, using_dac = 0;
+	int err;
 
 	err = pci_enable_device(pdev);
 	if (err) {
@@ -3890,13 +3886,9 @@ static int __devinit skge_probe(struct pci_dev *pdev,
 
 	pci_set_master(pdev);
 
-	if (!pci_set_dma_mask(pdev, DMA_BIT_MASK(64))) {
-		using_dac = 1;
-		err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
-	} else if (!(err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32)))) {
-		using_dac = 0;
+	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
+	if (!err)
 		err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32));
-	}
 
 	if (err) {
 		dev_err(&pdev->dev, "no usable DMA configuration\n");
@@ -3944,7 +3936,7 @@ static int __devinit skge_probe(struct pci_dev *pdev,
 		(unsigned long long)pci_resource_start(pdev, 0), pdev->irq,
 		skge_board_name(hw), hw->chip_rev);
 
-	dev = skge_devinit(hw, 0, using_dac);
+	dev = skge_devinit(hw, 0);
 	if (!dev)
 		goto err_out_led_off;
 
@@ -3967,7 +3959,7 @@ static int __devinit skge_probe(struct pci_dev *pdev,
 	skge_show_addr(dev);
 
 	if (hw->ports > 1) {
-		dev1 = skge_devinit(hw, 1, using_dac);
+		dev1 = skge_devinit(hw, 1);
 		if (dev1 && register_netdev(dev1) == 0)
 			skge_show_addr(dev1);
 		else {
#1 lore raw
Quoted 9 lines · On Tue, 21 Sep 2010 09:15:19 +0200 Stanislaw Gruszka <[email protected]>…

We have old fedora bug report about not working skge device on some Gigabyte motherboard:
https://bugzilla.redhat.com/show_bug.cgi?id=447489

In dmesg there are messages:

skge 0000:02:0b.0: PCI error cmd=0x7 status=0x22b0 skge 0000:02:0b.0: unable to clear error (so ignoring them)

Quoted 8 lines

Status value 0x22b0 include PCI_STATUS_REC_MASTER_ABORT bit set, what mean that DMA transfer on PCI bus was aborted.

I provided to user modified driver with module parameter, that allow to experiment with different dma_mask's. He told that maximum working dma_mask for him is 32 bits set.

Is 32bit DMA limit for every skge device acceptable patch?

No. The device does support 64 bit DMA, the problem is a crap implementation of the device the motherboard. Therefore a PCI quirk or module parameter is a better solution.

#2 lore raw
Quoted 5 lines · Stephen Hemminger
Quoted 1 line

Is 32bit DMA limit for every skge device acceptable patch?

No. The device does support 64 bit DMA, the problem is a crap implementation of the device the motherboard. Therefore a PCI quirk or module parameter is a better solution.

Ok, I will use dmi_check_system() to recognize broken board and limit DMA address space when detected.

Thanks
Stanislaw