Skip to contentSkip to conversation

macb: get mac address from environment

U-Boot Andrzej Wolski 2 participants 5 messages Jul 11, 2009 – Jul 12, 2009 Network drivers

lore
Share & tools
lore
#0 lore raw

This fix macb driver broken by commit
56b555a644f3cbb1b3929cb52b61d3ce483885f5

---

drivers/net/macb.c +0 -0
index 6de0a04..0d2ef5e 100644
@@ -408,8 +408,7 @@ static int macb_init(struct eth_device *netdev, bd_t 

*bd)
{
struct macb_device *macb = to_macb(netdev); unsigned long paddr;
- u32 hwaddr_bottom;
- u16 hwaddr_top;
+ uchar enetaddr[6];
int i;

/*
@@ -439,10 +438,10 @@ static int macb_init(struct eth_device *netdev, bd_t *bd)
macb_writel(macb, TBQP, macb->tx_ring_dma);

/* set hardware address */
- hwaddr_bottom = cpu_to_le32(*((u32 *)netdev->enetaddr));
- macb_writel(macb, SA1B, hwaddr_bottom);
- hwaddr_top = cpu_to_le16(*((u16 *)(netdev->enetaddr + 4)));
- macb_writel(macb, SA1T, hwaddr_top);
+ eth_getenv_enetaddr("ethaddr", enetaddr);
+ macb_writel(macb, SA1B, (enetaddr[3] << 24) | (enetaddr[2] << 16)
+ | (enetaddr[1] << 8) | (enetaddr[0]));
+ macb_writel(macb, SA1T, (enetaddr[5] << 8) | (enetaddr[4]));

/* choose RMII or MII mode. This depends on the board */
#ifdef CONFIG_RMII

#1 lore raw
Quoted 2 lines · On Saturday 11 July 2009 19:05:40 Andrzej Wolski wrote:

This fix macb driver broken by commit
56b555a644f3cbb1b3929cb52b61d3ce483885f5

what exactly broke ? drivers that are ported to NET_MULTI should *not* be getting the mac from the environment. and from what i can see, the macb driver is properly supporting the NET_MULTI setup.

Quoted 28 lines
drivers/net/macb.c +1 -2
@@ -408,8 +408,7 @@ static int macb_init(struct eth_device *netdev, bd_t
 *bd)
  {
  	struct macb_device *macb = to_macb(netdev);
  	unsigned long paddr;
-	u32 hwaddr_bottom;
-	u16 hwaddr_top;
+	uchar enetaddr[6];
  	int i;
 

/*

drivers/net/macb.c +4 -4
@@ -439,10 +438,10 @@ static int macb_init(struct eth_device *netdev,
 bd_t *bd)
  	macb_writel(macb, TBQP, macb->tx_ring_dma);
 
  	/* set hardware address */
-	hwaddr_bottom = cpu_to_le32(*((u32 *)netdev->enetaddr));
-	macb_writel(macb, SA1B, hwaddr_bottom);
-	hwaddr_top = cpu_to_le16(*((u16 *)(netdev->enetaddr + 4)));
-	macb_writel(macb, SA1T, hwaddr_top);
+	eth_getenv_enetaddr("ethaddr", enetaddr);
+	macb_writel(macb, SA1B, (enetaddr[3] << 24) | (enetaddr[2] << 16)
+				 | (enetaddr[1] <<  8) | (enetaddr[0]));
+	macb_writel(macb, SA1T, (enetaddr[5] <<  8) | (enetaddr[4]));
 
  	/* choose RMII or MII mode. This depends on the board */
Quoted 1 line
  #ifdef CONFIG_RMII

looks to me like the existing code is correct and your new proposed code is not
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
Url : http://lists.denx.de/pipermail/u-boot/attachments/20090711/57d46a57/attachment.pgp

#2 lore raw
Quoted 10 lines · On Saturday 11 July 2009 19:05:40 Andrzej Wolski wrote:
Quoted 2 lines

This fix macb driver broken by commit
56b555a644f3cbb1b3929cb52b61d3ce483885f5

what exactly broke ? drivers that are ported to NET_MULTI should *not* be getting the mac from the environment. and from what i can see, the macb driver is properly supporting the NET_MULTI setup.

looks to me like the existing code is correct and your new proposed code is not

Hmm. Since pointed commit Ethernet just stopped to work, at least on my board similar to SAM9260EK. Maybe someone else can confirm it?

Moreover, since that patch, when ethaddr is not set in the environment, and I set it from command line, I get this while trying to ping:

U-Boot> ping 192.168.2.1
macb0: link up, 100Mbps full-duplex (lpa: 0x45e1)
*** ERROR: `ethaddr' not set
macb0: link up, 100Mbps full-duplex (lpa: 0x45e1)
ping failed; host 192.168.2.1 is not alive

Andrzej

#3 lore raw
Quoted 14 lines · On Saturday 11 July 2009 22:27:36 Andrzej Wolski wrote:
Quoted 10 lines · On Saturday 11 July 2009 19:05:40 Andrzej Wolski wrote:
Quoted 2 lines

This fix macb driver broken by commit
56b555a644f3cbb1b3929cb52b61d3ce483885f5

what exactly broke ? drivers that are ported to NET_MULTI should *not* be getting the mac from the environment. and from what i can see, the macb driver is properly supporting the NET_MULTI setup.

looks to me like the existing code is correct and your new proposed code is not

Hmm. Since pointed commit Ethernet just stopped to work, at least on my board similar to SAM9260EK. Maybe someone else can confirm it?

Quoted 8 lines

Moreover, since that patch, when ethaddr is not set in the environment, and I set it from command line, I get this while trying to ping:

U-Boot> ping 192.168.2.1
macb0: link up, 100Mbps full-duplex (lpa: 0x45e1)
*** ERROR: `ethaddr' not set
macb0: link up, 100Mbps full-duplex (lpa: 0x45e1)
ping failed; host 192.168.2.1 is not alive

do you have CONFIG_NET_MULTI defined ? if not, your board needs to switch over to it.
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
Url : http://lists.denx.de/pipermail/u-boot/attachments/20090711/2b70f245/attachment.pgp

#4 lore raw
Quoted 20 lines
Quoted 11 lines
Quoted 8 lines
Quoted 2 lines

This fix macb driver broken by commit
56b555a644f3cbb1b3929cb52b61d3ce483885f5

what exactly broke ? drivers that are ported to NET_MULTI should *not* be getting the mac from the environment. and from what i can see, the macb driver is properly supporting the NET_MULTI setup.

looks to me like the existing code is correct and your new proposed code is not

Hmm. Since pointed commit Ethernet just stopped to work, at least on my board similar to SAM9260EK. Maybe someone else can confirm it?

Quoted 8 lines

Moreover, since that patch, when ethaddr is not set in the environment, and I set it from command line, I get this while trying to ping:

U-Boot> ping 192.168.2.1
macb0: link up, 100Mbps full-duplex (lpa: 0x45e1)
*** ERROR: `ethaddr' not set
macb0: link up, 100Mbps full-duplex (lpa: 0x45e1)
ping failed; host 192.168.2.1 is not alive

Quoted 3 lines

do you have CONFIG_NET_MULTI defined ? if not, your board needs to switch over to it.
-mike

I use configuration for at91sam9260ek board already present in U-Boot and CONFIG_NET_MULTI is defined there.

Andrzej