Skip to contentSkip to conversation

2.5.71 -- Lost second 3c509 card

DISCUSSION Pete Clements 3 participants 6 messages Jun 15, 2003

lore
Share & tools
lore
#0 lore raw

With 2.5.71, have lost second 3c509 card (i386 UP).

Quoted 1 line

From kernel boot log with 2.5.71 get

eth%%d: 3c5x9 at 0x300, BNC port, address 00 20 af 26 bf 2c, IRQ 10.

Quoted 1 line

From kernel boot log with 2.5.70 get

eth0: 3c5x9 at 0x300, BNC port, address 00 20 af 26 bf 2c, IRQ 10.
eth1: 3c5x9 at 0x310, 10baseT port, address 00 60 08 15 31 84, IRQ 9.

Checking on a single card system (SMP), the eth%%d: indication was introduced with 2.5.70-bk10.

Signature · 2 lines
Pete Clements 
[email protected]
#1 lore raw
Quoted 4 lines

With 2.5.71, have lost second 3c509 card (i386 UP).

Quoted 1 line

From kernel boot log with 2.5.71 get

Quoted 8 lines

eth%%d: 3c5x9 at 0x300, BNC port, address 00 20 af 26 bf 2c, IRQ 10.

Quoted 1 line

From kernel boot log with 2.5.70 get

eth0: 3c5x9 at 0x300, BNC port, address 00 20 af 26 bf 2c, IRQ 10.
eth1: 3c5x9 at 0x310, 10baseT port, address 00 60 08 15 31 84, IRQ 9.

Checking on a single card system (SMP), the eth%%d: indication was introduced with 2.5.70-bk10.

As a followup, reverted the 3c509 bk10 changes. Back in business with 2.5.71.

Signature · 2 lines
Pete Clements 
[email protected]
#2 lore raw
Quoted 2 lines · Pete Clements

As a followup, reverted the 3c509 bk10 changes. Back in business with 2.5.71.

Are you using command line port number specifications? If so, what do they look like?

Signature · 1 line
David S. Miller <[email protected]>
#3 lore raw

Quoting David S. Miller

Quoted 12 lines · Pete Clements
Quoted 2 lines

As a followup, reverted the 3c509 bk10 changes. Back in business with 2.5.71.

Are you using command line port number specifications? If so, what do they look like?

Signature · 4 lines
David S. Miller <[email protected]>


From boot log:

Kernel command line: auto BOOT_IMAGE=Linux ro root=341 ether=9,0x310,4,0x3c509,eth1

Signature · 2 lines
Pete Clements 
[email protected]
#4 lore raw

From: Pete Clements <[email protected]>
Date: Sun, 15 Jun 2003 12:50:25 -0400 (EDT)

Quoted 1 line

From boot log:

Kernel command line: auto BOOT_IMAGE=Linux ro root=341 ether=9,0x310,4,0x3c509,eth1

Yes, that explains why the second card isn't found.

The problem is:

1) the we can't assign a name to the device
until we've registered it with the networking

2) without a name the boot argument lookup doesn't work

3) we have to register the device with the networking only after it is initialized

Hmmm...

#5 lore raw
Quoted 14 lines · David S. Miller

From: Pete Clements <[email protected]>
Date: Sun, 15 Jun 2003 12:50:25 -0400 (EDT)

Quoted 1 line

From boot log:

Kernel command line: auto BOOT_IMAGE=Linux ro root=341 ether=9,0x310,4,0x3c509,eth1

Yes, that explains why the second card isn't found.

The problem is:

1) the we can't assign a name to the device
until we've registered it with the networking

Quoted 6 lines

2) without a name the boot argument lookup doesn't work

3) we have to register the device with the networking only after it is initialized

Hmmm...

Crap. AFAICS, the clean solution would be to pass these guys (blah_probe()) expected device name. And let them do allocation, etc., themselves.

Looking into it...