[gentoo-user] (dual) booting windows kills the ethernet device for linux

[gentoo-user] (dual) booting windows kills the ethernet device for linux

[Allan Gottlieb]

I feel strange writing this since I can hardly believe it.
However, it seems to be quite repeatable.

I have a new dell latitude E6500 that I set up for dual booting:
windows 7 and gentoo linux.

The machine seems to have two "hardware" states determined by whether
windows has been run since power on.

If you power the machine on it goes into what I call State A.
Now if I either select linux from grub or just use the default we get

linux boots and eth0 works
reboot
linux boots and eth0 works
...
reboot
linux boots and eth0 works

but now reboot into windows and we get State B
windows boots and eth0 works
reboot to linux
linux boots but eth0 fails
reboot
linux boots but eth0 fails
...
reboot
linux boots but eth0 fails.

If I then power the machine off instead of simply rebooting
we get back to State A

power on
linux boots and eth0 works
reboot
linux boots and eth0 works

etc.

This is quite repeatable.  I would greatly appreciate an explanation.

thanks,
allan

[gentoo-user] Re: (dual) booting windows kills the ethernet device for linux

[Nikos Chantziaras]

On 06/17/2010 03:33 AM, Allan Gottlieb wrote:
> I feel strange writing this since I can hardly believe it.
> However, it seems to be quite repeatable.
>
> I have a new dell latitude E6500 that I set up for dual booting:
> windows 7 and gentoo linux.
>
> The machine seems to have two "hardware" states determined by whether
> windows has been run since power on.
>
> If you power the machine on it goes into what I call State A.
> Now if I either select linux from grub or just use the default we get
>
> linux boots and eth0 works
> reboot
> linux boots and eth0 works
> ...
> reboot
> linux boots and eth0 works
>
> but now reboot into windows and we get State B
> windows boots and eth0 works
> reboot to linux
> linux boots but eth0 fails
> reboot
> linux boots but eth0 fails
> ...
> reboot
> linux boots but eth0 fails.
>
> If I then power the machine off instead of simply rebooting
> we get back to State A
>
> power on
> linux boots and eth0 works
> reboot
> linux boots and eth0 works
>
> etc.
>
> This is quite repeatable.  I would greatly appreciate an explanation.

The explanation is that the Windows driver leaves the hardware in a 
state that the Linux driver doesn't expect.  A warm reboot doesn't reset 
it correctly.  Only a cold one does.  This is usually a bug in the Linux 
driver.  Might be worth reporting upstream.

Re: [gentoo-user] (dual) booting windows kills the ethernet device for linux

[Dale]

Allan Gottlieb wrote:
> I feel strange writing this since I can hardly believe it.
> However, it seems to be quite repeatable.
>
> I have a new dell latitude E6500 that I set up for dual booting:
> windows 7 and gentoo linux.
>
> The machine seems to have two "hardware" states determined by whether
> windows has been run since power on.
>
> If you power the machine on it goes into what I call State A.
> Now if I either select linux from grub or just use the default we get
>
> linux boots and eth0 works
> reboot
> linux boots and eth0 works
> ...
> reboot
> linux boots and eth0 works
>
> but now reboot into windows and we get State B
> windows boots and eth0 works
> reboot to linux
> linux boots but eth0 fails
> reboot
> linux boots but eth0 fails
> ...
> reboot
> linux boots but eth0 fails.
>
> If I then power the machine off instead of simply rebooting
> we get back to State A
>
> power on
> linux boots and eth0 works
> reboot
> linux boots and eth0 works
>
> etc.
>
> This is quite repeatable.  I would greatly appreciate an explanation.
>
> thanks,
> allan
>
>    

I read about this ages ago.  I *think* it was windoze doing some sort of 
a shutdown on the card, disabling it or putting it into sleep mode if 
you want to call it that.  It appears that Linux isn't "waking" the card 
up but a power cycle does.  It also appears that windoze "wakes" the 
card up when it boots.

I'm pretty sure they found a fix but I can't recall what they did.  It 
seems they changed some sort of setting in windoze but not real sure.  
May want to google the mailing list archives and see if you can find 
it.  I know it is older than the archives I have here.  I keep them for 
a year then it dumps them.  So it's been a good while.  May have to dig 
back a while.

Biggest point of reply, you're not nuts.  Someone else had the same 
problem.  lol  Hard to believe I remember as much as I did tho.  o_O

Dale

:-)  :-)

Re: [gentoo-user] (dual) booting windows kills the ethernet device for linux

[Allan Gottlieb]

At Wed, 16 Jun 2010 20:04:23 -0500 Dale <rdalek1967@gmail.com> wrote:

> Allan Gottlieb wrote:
>> I feel strange writing this since I can hardly believe it.
>> However, it seems to be quite repeatable.
>>
>> I have a new dell latitude E6500 that I set up for dual booting:
>> windows 7 and gentoo linux.
>>
>> The machine seems to have two "hardware" states determined by whether
>> windows has been run since power on.
>>
>> If you power the machine on it goes into what I call State A.
>> Now if I either select linux from grub or just use the default we get
>>
>> linux boots and eth0 works
>> reboot
>> linux boots and eth0 works
>> ...
>> reboot
>> linux boots and eth0 works
>>
>> but now reboot into windows and we get State B
>> windows boots and eth0 works
>> reboot to linux
>> linux boots but eth0 fails
>> reboot
>> linux boots but eth0 fails
>> ...
>> reboot
>> linux boots but eth0 fails.
>>
>> If I then power the machine off instead of simply rebooting
>> we get back to State A
>>
>> power on
>> linux boots and eth0 works
>> reboot
>> linux boots and eth0 works
>>
>> etc.
>>
>> This is quite repeatable.  I would greatly appreciate an explanation.
>>
>> thanks,
>> allan
>>
>
> I read about this ages ago.  I *think* it was windoze doing some sort
> of a shutdown on the card, disabling it or putting it into sleep mode
> if you want to call it that.  It appears that Linux isn't "waking" the
> card up but a power cycle does.  It also appears that windoze "wakes"
> the card up when it boots.
>
> I'm pretty sure they found a fix but I can't recall what they did.  It
> seems they changed some sort of setting in windoze but not real sure.
> May want to google the mailing list archives and see if you can find
> it.  I know it is older than the archives I have here.  I keep them
> for a year then it dumps them.  So it's been a good while.  May have
> to dig back a while.

Thank you and thanks to nikos.  I have googled a little and the new ones
I see are slightly different and the old ones use older versions of the
kernel and hence driver.  I run gentoo-sources-2.6.34

> Biggest point of reply, you're not nuts.

Indeed, I was concerned about that.

I must say that the problem is easy for me to work around.  I hardly
ever use windows (maybe once every month or two once the systems are
stably installed), so having to power down is no big deal.  I was just
worried that I had something badly configured.

thank you both again,
allan

Re: [gentoo-user] (dual) booting windows kills the ethernet device for linux

[Alex Schuster]

Allan Gottlieb writes:

> The machine seems to have two "hardware" states determined by whether
> windows has been run since power on.

I _think_ I read about such a problem ages ago, and there was a 
workaround. Either in the BIOS, or in Windows, some Wake-On-LAN option. If 
activated, Windows would not shut down the device, so Wake-On-LAN will 
continue to work.

My memory on this is vague, but it might be worth a try to look into that.

	Wonko

Re: [gentoo-user] (dual) booting windows kills the ethernet device for linux

[Allan Gottlieb]

At Thu, 17 Jun 2010 11:27:38 +0200 Alex Schuster <wonko@wonkology.org> wrote:

> Allan Gottlieb writes:
>
>> The machine seems to have two "hardware" states determined by whether
>> windows has been run since power on.
>
> I _think_ I read about such a problem ages ago, and there was a 
> workaround. Either in the BIOS, or in Windows, some Wake-On-LAN option. If 
> activated, Windows would not shut down the device, so Wake-On-LAN will 
> continue to work.
>
> My memory on this is vague, but it might be worth a try to look into that.

Thanks.   I just looked and the google hits I looked were about trouble
recovering from hibernation.

thanks again,
allan