via-velocity: printk of dev->name before name has been allocated

via-velocity: printk of dev->name before name has been allocated

[Jan Ceuleers]

I have the following in dmesg:

[   21.212838] VIA Networking Velocity Family Gigabit Ethernet Adapter 
Driver Ver. 1.14
[   21.212847] Copyright (c) 2002, 2003 VIA Networking Technologies, Inc.
[   21.212852] Copyright (c) 2004 Red Hat Inc.
[   21.212885] ACPI: PCI Interrupt 0000:00:0e.0[A] -> GSI 18 (level, 
low) -> IRQ 17
[   21.212937] eth%d: set value of parameter Interrupt service works to 64
[   21.213684] eth0: VIA Networking Velocity Family Gigabit Ethernet Adapter
[   21.213691] eth0: Ethernet Address: 00:40:63:F4:AF:DE

The notification that the module parameter has been taken into account 
is output before the device name has been allocated. This is with the 
following kernel version:

Linux fe2 2.6.22-16-generic #1 SMP Mon Jan 26 00:07:52 GMT 2009 i686 
GNU/Linux

which is quite "old", but when I looked at current git net-2.6 I think 
that this problem is still there.

I think that the cause is that in velocity_found1(), 
velocity_get_pci_info() is called (which outputs the "set value of 
parameter" message) before register_netdev() is (which I think allocates 
the device name).

I'm afraid I'm incapable of fixing this cosmetic problem myself; perhaps 
one of you kind folk can.

Thanks, Jan

Re: via-velocity: printk of dev->name before name has been allocated

[David Miller]

From: Jan Ceuleers <jan.ceuleers@computer.org>
Date: Sat, 21 Feb 2009 19:56:11 +0100

> I have the following in dmesg:
> 
> [   21.212838] VIA Networking Velocity Family Gigabit Ethernet Adapter Driver Ver. 1.14
> [   21.212847] Copyright (c) 2002, 2003 VIA Networking Technologies, Inc.
> [   21.212852] Copyright (c) 2004 Red Hat Inc.
> [   21.212885] ACPI: PCI Interrupt 0000:00:0e.0[A] -> GSI 18 (level, low) -> IRQ 17
> [   21.212937] eth%d: set value of parameter Interrupt service works to 64
> [   21.213684] eth0: VIA Networking Velocity Family Gigabit Ethernet Adapter
> [   21.213691] eth0: Ethernet Address: 00:40:63:F4:AF:DE

This is already fixed in current kernels.

Reporting such things against 6 release old kernels is very near to
pointless, please check against current releases first next time.

Thank you.

Re: via-velocity: printk of dev->name before name has been allocated

[Jan Ceuleers]

David Miller wrote:
>> I have the following in dmesg:
>>
>> [   21.212838] VIA Networking Velocity Family Gigabit Ethernet Adapter Driver Ver. 1.14
>> [   21.212847] Copyright (c) 2002, 2003 VIA Networking Technologies, Inc.
>> [   21.212852] Copyright (c) 2004 Red Hat Inc.
>> [   21.212885] ACPI: PCI Interrupt 0000:00:0e.0[A] -> GSI 18 (level, low) -> IRQ 17
>> [   21.212937] eth%d: set value of parameter Interrupt service works to 64
>> [   21.213684] eth0: VIA Networking Velocity Family Gigabit Ethernet Adapter
>> [   21.213691] eth0: Ethernet Address: 00:40:63:F4:AF:DE
> 
> This is already fixed in current kernels.
> 
> Reporting such things against 6 release old kernels is very near to
> pointless, please check against current releases first next time.

Ooh, excuse me!

As I wrote I did look at current git net-2.6 and thought the issue was 
still there. That was based on looking at the code though; I can't run a 
current kernel on that machine.

You guys ar going to have to get better at communicating with your user 
base...

Re: via-velocity: printk of dev->name before name has been allocated

[David Miller]

From: Jan Ceuleers <jan.ceuleers@computer.org>
Date: Sun, 22 Feb 2009 09:43:05 +0100

> As I wrote I did look at current git net-2.6 and thought the issue
> was still there. That was based on looking at the code though; I
> can't run a current kernel on that machine.
> 
> You guys ar going to have to get better at communicating with your
> user base...

You're the one who didn't mention that you looked at the current
code, that changes everything :)

Re: via-velocity: printk of dev->name before name has been allocated

[Jan Ceuleers]

David Miller wrote:
> From: Jan Ceuleers <jan.ceuleers@computer.org>
> Date: Sun, 22 Feb 2009 09:43:05 +0100
> 
>> As I wrote I did look at current git net-2.6 and thought the issue
>> was still there. That was based on looking at the code though; I
>> can't run a current kernel on that machine.
>>
>> You guys ar going to have to get better at communicating with your
>> user base...
> 
> You're the one who didn't mention that you looked at the current
> code, that changes everything :)

I did actually, you're the one who didn't read what I wrote.

Quoting from my original email:

"This is with the following kernel version:

Linux fe2 2.6.22-16-generic #1 SMP Mon Jan 26 00:07:52 GMT 2009 i686 
GNU/Linux

which is quite "old", but when I looked at current git net-2.6 I think 
that this problem is still there."

Re: via-velocity: printk of dev->name before name has been allocated

[Jan Ceuleers]

David Miller wrote:
>> I have the following in dmesg:
>>
>> [   21.212838] VIA Networking Velocity Family Gigabit Ethernet Adapter Driver Ver. 1.14
>> [   21.212847] Copyright (c) 2002, 2003 VIA Networking Technologies, Inc.
>> [   21.212852] Copyright (c) 2004 Red Hat Inc.
>> [   21.212885] ACPI: PCI Interrupt 0000:00:0e.0[A] -> GSI 18 (level, low) -> IRQ 17
>> [   21.212937] eth%d: set value of parameter Interrupt service works to 64
>> [   21.213684] eth0: VIA Networking Velocity Family Gigabit Ethernet Adapter
>> [   21.213691] eth0: Ethernet Address: 00:40:63:F4:AF:DE
> 
> This is already fixed in current kernels.

OK, found the commit (07b5f6a6fd0ce47390f7fbec966cd5c70127e597). Sorry 
to have bothered you.

Re: via-velocity: printk of dev->name before name has been allocated

[David Miller]

From: Jan Ceuleers <jan.ceuleers@computer.org>
Date: Sun, 22 Feb 2009 11:44:02 +0100

> David Miller wrote:
> > From: Jan Ceuleers <jan.ceuleers@computer.org>
> > Date: Sun, 22 Feb 2009 09:43:05 +0100
> > 
> >> As I wrote I did look at current git net-2.6 and thought the issue
> >> was still there. That was based on looking at the code though; I
> >> can't run a current kernel on that machine.
> >>
> >> You guys ar going to have to get better at communicating with your
> >> user base...
> > You're the one who didn't mention that you looked at the current
> > code, that changes everything :)
> 
> I did actually, you're the one who didn't read what I wrote.

My bad, sorry about that.