RE: ethtool -d no longer works for e1000

RE: ethtool -d no longer works for e1000

[cramerj]

Perhaps the network interface enumeration changed with these kernels?
Is the e1000 part still eth0?

-Jeb

> -----Original Message-----
> From: netdev-bounce@oss.sgi.com [mailto:netdev-bounce@oss.sgi.com] On
> Behalf Of Ben Greear
> Sent: Wednesday, March 09, 2005 1:36 PM
> To: 'netdev@oss.sgi.com'
> Subject: ethtool -d no longer works for e1000
> 
> I tried this with FC2's 2.6.10-1.770_FC2smp kernel and with my own
> slightly hacked 2.6.11 kernel.  Both give this response:
> 
> ethtool -d eth0
> Cannot dump registers:  Success
> 
> I am quite certain this used to work, and used to show me
> the PCI bus speed and other nice details...
> 
> Thanks,
> Ben
> 
> --
> Ben Greear <greearb@candelatech.com>
> Candela Technologies Inc  http://www.candelatech.com
> 

Re: ethtool -d no longer works for e1000

[Ben Greear]

cramerj wrote:
> Perhaps the network interface enumeration changed with these kernels?
> Is the e1000 part still eth0?

Nope, nothing but 4 e1000 interfaces in this machine.


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

RE: ethtool -d no longer works for e1000

[cramerj]

Ah!  Yes, if you're running ethtool 1.8, then this was an issue (or
rather, an annoyance) in e1000.c where we would default to mac_type =
e1000_undefined for any newer hardware.  Continuous updates to this file
for every piece of hardware we released got old pretty quick.  We've
since changed the default mac_type to e1000_82543, which has a basic set
of registers that haven't changed with each hardware release.  And
although the e1000_82546 mac_type exists, it isn't yet being assigned to
the 82546GB hardware (again, see e1000.c), thus the error.

Long story short, an updated ethtool should do the trick.

Thanks,
-Jeb

> -----Original Message-----
> From: netdev-bounce@oss.sgi.com [mailto:netdev-bounce@oss.sgi.com] On
> Behalf Of Ben Greear
> Sent: Wednesday, March 09, 2005 3:42 PM
> To: Jon Mason
> Cc: Ganesh Venkatesan; netdev@oss.sgi.com
> Subject: Re: ethtool -d no longer works for e1000
> 
> Ben Greear wrote:
> > Jon Mason wrote:
> >
> >> I don't see this problem at all on my 2.6.11-rc4-mm1 kernel
(Athlon64
> >> proc).
> >
> >
> > I think I see the problem.  ethtool -d eth0 works for me,
> > but ethtool -d eth1 does not, even with both are e1000
> > NICs.  It appears it cannot handle reading the second NIC for
> > some reason?
> 
> Errr, my bad.  The problem is more that the dual-port pro/1000 NICs
> don't seem to work, but the built-in e1000s do.
> 
> The chipset that does not work correctly is: 82546GB
> 
> The NICs with chipset: 82541EI seem to work just fine.
> 
> Ben
> 
> 
> --
> Ben Greear <greearb@candelatech.com>
> Candela Technologies Inc  http://www.candelatech.com
> 

Re: ethtool -d no longer works for e1000

[Jon Mason]

On Wednesday 09 March 2005 06:35 pm, cramerj wrote:
> Ah!  Yes, if you're running ethtool 1.8, then this was an issue (or
> rather, an annoyance) in e1000.c where we would default to mac_type =
> e1000_undefined for any newer hardware.  Continuous updates to this file
> for every piece of hardware we released got old pretty quick.  We've
> since changed the default mac_type to e1000_82543, which has a basic set
> of registers that haven't changed with each hardware release.  And
> although the e1000_82546 mac_type exists, it isn't yet being assigned to
> the 82546GB hardware (again, see e1000.c), thus the error.
>
> Long story short, an updated ethtool should do the trick.
>
> Thanks,
> -Jeb

Ah, I am running ethtool 3 on a 82545 chip, that is probably why mine works.

Thanks,
Jon

ethtool -d no longer works for e1000

[Ben Greear]

I tried this with FC2's 2.6.10-1.770_FC2smp kernel and with my own
slightly hacked 2.6.11 kernel.  Both give this response:

ethtool -d eth0
Cannot dump registers:  Success

I am quite certain this used to work, and used to show me
the PCI bus speed and other nice details...

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

Re: ethtool -d no longer works for e1000

[Ganesh Venkatesan]

Ben:

Are you using the e1000 that is included in the 2.6.11 kernel? ethtool
-d eth? works fine for me. What else is different in your env?

ganesh.


On Wed, 09 Mar 2005 13:35:51 -0800, Ben Greear <greearb@candelatech.com> wrote:
> I tried this with FC2's 2.6.10-1.770_FC2smp kernel and with my own
> slightly hacked 2.6.11 kernel.  Both give this response:
> 
> ethtool -d eth0
> Cannot dump registers:  Success
> 
> I am quite certain this used to work, and used to show me
> the PCI bus speed and other nice details...
> 
> Thanks,
> Ben
> 
> --
> Ben Greear <greearb@candelatech.com>
> Candela Technologies Inc  http://www.candelatech.com
> 
>

Re: ethtool -d no longer works for e1000

[Ben Greear]

Ganesh Venkatesan wrote:
> Ben:
> 
> Are you using the e1000 that is included in the 2.6.11 kernel? ethtool
> -d eth? works fine for me. What else is different in your env?

I have patched the e1000 in my 2.6.11 kernel, but I see this -d problem
on other un-patched systems as well, so it can't be (just) my patches
that are the problem.

I also tried this on my x86-64 opteron system, running FC3.  I
have not hacked up this kernel or user-space tools at all :)

[root@grok lanforge]# ethtool -i eth3
driver: e1000
version: 5.5.4-k2-NAPI
firmware-version: N/A
bus-info: 0000:02:08.0
[root@grok lanforge]# ethtool -d eth3
Cannot dump registers: Success
[root@grok lanforge]# uname -a
Linux grok 2.6.10-1.766_FC3smp #1 SMP Wed Feb 9 23:17:48 EST 2005 x86_64 x86_64 x86_64 GNU/Linux


I'm using ethtool 1.8 on this system.


What version of ethtool are you using?


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

Re: ethtool -d no longer works for e1000

[Jon Mason]

I don't see this problem at all on my 2.6.11-rc4-mm1 kernel (Athlon64 proc).

# ethtool -i eth0
driver: e1000
version: 5.7.6-k2
firmware-version: N/A
bus-info: 0000:00:07.0

# ethtool -d eth0
MAC Registers
-------------
0x00000: CTRL (Device control register)  0x00F00249
      Duplex:                            full
      Endian mode (buffers):             little
      Link reset:                        reset
      Set link up:                       1
      Invert Loss-Of-Signal:             no
      Receive flow control:              disabled
      Transmit flow control:             disabled
      VLAN mode:                         disabled
      Auto speed detect:                 disabled
      Speed select:                      1000Mb/s
      Force speed:                       no
      Force duplex:                      no
0x00008: STATUS (Device status register) 0x0000C343
      Duplex:                            full
      Link up:                           link config
      TBI mode:                          disabled
      Link speed:                        100Mb/s
      Bus type:                          PCI
      Bus speed:                         33MHz
      Bus width:                         32-bit
0x00100: RCTL (Receive control register) 0x00008002
      Receiver:                          enabled
      Store bad packets:                 disabled
      Unicast promiscuous:               disabled
      Multicast promiscuous:             disabled
      Long packet:                       disabled
      Descriptor minimum threshold size: 1/2
      Broadcast accept mode:             accept
      VLAN filter:                       disabled
      Cononical form indicator:          disabled
      Discard pause frames:              filtered
      Pass MAC control frames:           don't pass
      Receive buffer size:               2048
0x02808: RDLEN (Receive desc length)     0x00001000
0x02810: RDH   (Receive desc head)       0x00000003
0x02818: RDT   (Receive desc tail)       0x00000000
0x02820: RDTR  (Receive delay timer)     0x00000000
0x00400: TCTL (Transmit ctrl register)   0x000400FA
      Transmitter:                       enabled
      Pad short packets:                 enabled
      Software XOFF Transmission:        disabled
      Re-transmit on late collision:     disabled
0x03808: TDLEN (Transmit desc length)    0x00001000
0x03810: TDH   (Transmit desc head)      0x000000D8
0x03818: TDT   (Transmit desc tail)      0x000000D8
0x03820: TIDV  (Transmit delay timer)    0x00000040
PHY type:                               M88

# uname -a
Linux victory 2.6.11-rc4-mm1 #4 Fri Feb 25 18:19:43 CST 2005 x86_64 AMD 
Athlon(tm) 64 Processor 3200+ AuthenticAMD GNU/Linux

On Wednesday 09 March 2005 05:09 pm, Ben Greear wrote:
> Ganesh Venkatesan wrote:
> > Ben:
> >
> > Are you using the e1000 that is included in the 2.6.11 kernel? ethtool
> > -d eth? works fine for me. What else is different in your env?
>
> I have patched the e1000 in my 2.6.11 kernel, but I see this -d problem
> on other un-patched systems as well, so it can't be (just) my patches
> that are the problem.
>
> I also tried this on my x86-64 opteron system, running FC3.  I
> have not hacked up this kernel or user-space tools at all :)
>
> [root@grok lanforge]# ethtool -i eth3
> driver: e1000
> version: 5.5.4-k2-NAPI
> firmware-version: N/A
> bus-info: 0000:02:08.0
> [root@grok lanforge]# ethtool -d eth3
> Cannot dump registers: Success
> [root@grok lanforge]# uname -a
> Linux grok 2.6.10-1.766_FC3smp #1 SMP Wed Feb 9 23:17:48 EST 2005 x86_64
> x86_64 x86_64 GNU/Linux
>
>
> I'm using ethtool 1.8 on this system.
>
>
> What version of ethtool are you using?

-- 
Jon Mason
jdmason@us.ibm.com

Re: ethtool -d no longer works for e1000

[Ben Greear]

Jon Mason wrote:
> I don't see this problem at all on my 2.6.11-rc4-mm1 kernel (Athlon64 proc).

I think I see the problem.  ethtool -d eth0 works for me,
but ethtool -d eth1 does not, even with both are e1000
NICs.  It appears it cannot handle reading the second NIC for
some reason?

Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

Re: ethtool -d no longer works for e1000

[Ben Greear]

Ben Greear wrote:
> Jon Mason wrote:
> 
>> I don't see this problem at all on my 2.6.11-rc4-mm1 kernel (Athlon64 
>> proc).
> 
> 
> I think I see the problem.  ethtool -d eth0 works for me,
> but ethtool -d eth1 does not, even with both are e1000
> NICs.  It appears it cannot handle reading the second NIC for
> some reason?

Errr, my bad.  The problem is more that the dual-port pro/1000 NICs
don't seem to work, but the built-in e1000s do.

The chipset that does not work correctly is: 82546GB

The NICs with chipset: 82541EI seem to work just fine.

Ben


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

Re: ethtool -d no longer works for e1000

[Lennert Buytenhek]

On Wed, Mar 09, 2005 at 03:42:03PM -0800, Ben Greear wrote:

> >>I don't see this problem at all on my 2.6.11-rc4-mm1 kernel (Athlon64 
> >>proc).
> >
> >I think I see the problem.  ethtool -d eth0 works for me,
> >but ethtool -d eth1 does not, even with both are e1000
> >NICs.  It appears it cannot handle reading the second NIC for
> >some reason?
> 
> Errr, my bad.  The problem is more that the dual-port pro/1000 NICs
> don't seem to work, but the built-in e1000s do.
> 
> The chipset that does not work correctly is: 82546GB
> 
> The NICs with chipset: 82541EI seem to work just fine.

"Me too."

Works fine: 82540EM (SE7505VB2 on-board e1000)
Doesn't work: 82546GB (dual port e1000 server NIC)

[root@phi ~]# uname -a
Linux phi.wantstofly.org 2.6.10-1.766_FC3smp #1 SMP Wed Feb 9 23:21:37 EST 2005 i686 i686 i386 GNU/Linux

[root@phi ~]# ethtool -d eth1
Cannot dump registers: Success
[root@phi ~]# ethtool -d eth2
Cannot dump registers: Success
[root@phi ~]# ethtool -d eth3 | head -3
MAC Registers
-------------
0x00000: CTRL (Device control register)  0x003C0249

Nothing appears in syslog.


--L

Re: ethtool -d no longer works for e1000

[Lennert Buytenhek]

On Wed, Mar 09, 2005 at 05:13:55PM -0600, Jon Mason wrote:

>       Cononical form indicator:          disabled

Shouldn't this be 'Canonical' ?


--L

Re: ethtool -d no longer works for e1000

[Andre Tomt]

Ben Greear wrote:
> I tried this with FC2's 2.6.10-1.770_FC2smp kernel and with my own
> slightly hacked 2.6.11 kernel.  Both give this response:
> 
> ethtool -d eth0
> Cannot dump registers:  Success
> 
> I am quite certain this used to work, and used to show me
> the PCI bus speed and other nice details...

Works fine here. Only thing is that CSA adapters seems to come up as 
32bit 33Mhz PCI.