Re: [9fans] intel ethernet

Re: [9fans] intel ethernet

[erik quanstrom]

that's a wireless card 3945ABG intel pro/wireless.
there's no driver for that that i know of.

- erik

Re: [9fans] intel ethernet

[lejatorn]

Yes, there is that but there is also an ethernet 82566MM card, that's
what I'm talking about.

On Wed, Jan 02, 2008 at 08:38:52AM -0500, erik quanstrom wrote:
> that's a wireless card 3945ABG intel pro/wireless.
> there's no driver for that that i know of.
> 
> - erik

-- 
GPG key on subkeys.pgp.net:

KeyID:	| Fingerprint:
683DE5F3 | 4324 5818 39AA 9545 95C6 09AF B0A4 DFEA 683D E5F3
--

[9fans] intel ethernet

[lejatorn]

Hi all,

My problem is I can't have the network up on my laptop; I get a
message stating there is no /net/ether0 when I try to ip/ipconfig
for example. According to the wiki, that means the ethernet card in my
laptop does not seem to be supported under plan9.

Given it's an intel gigabit pro/1000 (which is supposed to be supported)
I found it surprising, so I tried to force the use of the igbe/igbepcie
drivers with an entry in plan9.ini but to no avail.

Am I missing something here? Or is it just that my specific chipset
(it's a 82566MM) is not yet supported by these drivers? 
If yes, how hard would it be to have it so? Where should I start to do
that?

Thanks,
Mathieu.

-- 
GPG key on subkeys.pgp.net:

KeyID:	| Fingerprint:
683DE5F3 | 4324 5818 39AA 9545 95C6 09AF B0A4 DFEA 683D E5F3
--

Re: [9fans] intel ethernet

[ron minnich]

please send the output of the pci command. It could be something as
simple as a devid change, which intel loves to do all the time.
ron

Re: [9fans] intel ethernet

[Mathieu Lonjaret]

[-- Attachment #1: Type: text/plain, Size: 400 bytes --]

Ah thx.
Here is the output attached then.

Mathieu.

On Jan 2, 2008 6:33 AM, ron minnich <rminnich@gmail.com> wrote:
> please send the output of the pci command. It could be something as
> simple as a devid change, which intel loves to do all the time.
> ron
>



-- 
GPG key on subkeys.pgp.net:

KeyID:   | Fingerprint:
683DE5F3 | 4324 5818 39AA 9545 95C6 09AF B0A4 DFEA 683D E5F3

[-- Attachment #2: pci_out.txt --]
[-- Type: text/plain, Size: 1497 bytes --]

0.0.0:	brg  06.00.00 8086/2a00   0
0.2.0:	vid  03.00.00 8086/2a02  10 0:f8100004 1048576 1:00000000 16 2:e000000c 268435456 3:00000000 16 4:00001801 16
0.2.1:	vid  03.80.00 8086/2a03   0 0:f8200004 1048576 1:00000000 16
0.25.0:	net  02.00.00 8086/1049  11 0:fe000000 131072 1:fe025000 4096 2:00001841 32
0.26.0:	usb 0c.03.00 8086/2834  11 4:00001861 32
0.26.1:	usb 0c.03.00 8086/2835  11 4:00001881 32
0.26.7:	usb 0c.03.20 8086/283a  11 0:fe226400 1024
0.27.0:	---  04.03.00 8086/284b  11 0:fe020004 16384 1:00000000 16
0.28.0:	brg  06.04.00 8086/283f  11
0.28.1:	brg  06.04.00 8086/2841  11
0.28.2:	brg  06.04.00 8086/2843  11
0.28.3:	brg  06.04.00 8086/2845  11
0.28.4:	brg  06.04.00 8086/2847  11
0.29.0:	usb 0c.03.00 8086/2830  10 4:000018a1 32
0.29.1:	usb 0c.03.00 8086/2831  11 4:000018c1 32
0.29.2:	usb 0c.03.00 8086/2832  11 4:000018e1 32
0.29.7:	usb 0c.03.20 8086/2836  11 0:fe226800 1024
0.30.0:	brg  06.04.01 8086/2448 255
0.31.0:	brg  06.01.00 8086/2811   0
0.31.2:	disk 01.01.80 8086/2828 255 0:00000001 16 1:00000001 16 2:00000001 16 3:00000001 16 4:00001c31 16 5:00001c21 16
0.31.3:	smb 0c.05.00 8086/283e  11 0:fe226c00 256 4:00001c41 32
21.0.0:	brg  06.07.00 1180/0476  10
21.0.1:	---  0c.00.10 1180/0832  11 0:f8301000 2048
21.0.2:	---  08.05.00 1180/0822  11 0:f8301800 256
21.0.3:	---  08.80.00 1180/0843  11 0:f8301c00 256
21.0.4:	---  08.80.00 1180/0592  11 0:f8302000 256
21.0.5:	---  08.80.00 1180/0852  11 0:f8302400 256
3.0.0:	net  02.80.00 8086/4227  11 0:df3ff000 4096

Re: [9fans] intel ethernet

[erik quanstrom]

i missed your 82566mm in my previous
post.  ron's speculation is basically right.
intel change device ids between the initial
documentation and the actual product.
in /sys/src/9/pc/ether82563.c makee this change
and recompile a kernel:

			break;
		case 0x1049:		/* mm */
+ 		case 0x1047:
		case 0x104a:		/* dm */
		case 0x104d:		/* v */
			type = i82566;
			break;
		case 0x10a4:
		case 0x105e:

i'll submit a patch.

- erik

Re: [9fans] intel ethernet

[lejatorn]

Thx, but as I told you in pv, I can't test if that works since any
kernel I build with that change (which I don't think is the cause)
freezes in the same way I described in the other thread (about iso not
booting on T61). Even in pata mode.
I tried building from the sources which had been installed by the iso I
used (from 07/12/07) to install originally. I also tried with the
sources from an iso from 22/12/07; same problem.
I'm currently getting the latest iso, but I don't think it should be
pretty different from the one of 22th of Dec, shouldn't it?

I will attach the pci -v output in a mail in the other thread.

Any idea?

Mathieu.

On Wed, Jan 02, 2008 at 08:45:03AM -0500, erik quanstrom wrote:
> i missed your 82566mm in my previous
> post.  ron's speculation is basically right.
> intel change device ids between the initial
> documentation and the actual product.
> in /sys/src/9/pc/ether82563.c makee this change
> and recompile a kernel:
> 
> 			break;
> 		case 0x1049:		/* mm */
> + 		case 0x1047:
> 		case 0x104a:		/* dm */
> 		case 0x104d:		/* v */
> 			type = i82566;
> 			break;
> 		case 0x10a4:
> 		case 0x105e:
> 
> i'll submit a patch.
> 
> - erik

-- 
GPG key on subkeys.pgp.net:

KeyID:	| Fingerprint:
683DE5F3 | 4324 5818 39AA 9545 95C6 09AF B0A4 DFEA 683D E5F3
--

Re: [9fans] intel ethernet

[erik quanstrom]

okay.  contact me off list with details.  especially what is
printed before it freezes.  this may be due to a bug in 9load.
there's a patch submitted.

- erik

Re: [9fans] intel ethernet

[Fazlul Shahriar]

> in /sys/src/9/pc/ether82563.c makee this change
> and recompile a kernel:
>
>                         break;
>                 case 0x1049:            /* mm */
> +               case 0x1047:
>                 case 0x104a:            /* dm */
>                 case 0x104d:            /* v */
>                         type = i82566;
>                         break;
>                 case 0x10a4:
>                 case 0x105e:

I don't see 1047 in the pci output.

I have a similar laptop (thinkpad R61i) with the same NIC. Installed
Plan 9 from the latest ISO and the driver doesn't seem to work.
/dev/kmesg says:

Plan 9
E820: 00000000 0009d800 memory
E820: 0009d800 000a0000 reserved
E820: 000d2000 000d4000 reserved
E820: 000e0000 00100000 reserved
E820: 00100000 3d6b0000 memory
E820: 3d6b0000 3d6cc000 acpi reclaim
E820: 3d6cc000 3d700000 acpi nvs
E820: 3d700000 3d800000 reserved
E820: 3d800000 3e000000 reserved
E820: f0000000 f4000000 reserved
E820: fec00000 fec10000 reserved
E820: fed00000 fed00400 reserved
E820: fed14000 fed18000 reserved
E820: fed18000 fed19000 reserved
E820: fed19000 fed1a000 reserved
E820: fed1c000 fed20000 reserved
126 holes free
00017000 0009d000 548864
003b7000 10000000 264540160
265089024 bytes free
cpu0: 1496MHz GenuineIntel Xeon5000-series (cpuid: AX 0x06FD DX 0xBFEBFBFF)
ELCR: 0C00
pcirouting: Cannot find south bridge PCI.255.31.7
#Y0: Ricoh 476 PCI/Cardbus bridge, F8300000 intl 10
i82566: unusable CLS
i82566: unusable CLS
i82566: unusable CLS
i82566: unusable CLS
i82566: unusable CLS
i82566: unusable CLS
#U/usb0: uhci: port 0x1860 irq 11
#U/usb1: uhci: port 0x1880 irq 11
#U/usb2: uhci: port 0x18A0 irq 10
#U/usb3: uhci: port 0x18C0 irq 11
#U/usb4: uhci: port 0x18E0 irq 11
983M memory: 256M kernel data, 726M user, 1351M swap


pci output:

0.0.0:	brg  06.00.00 8086/2a00   0
0.2.0:	vid  03.00.00 8086/2a02  10 0:f8100004 1048576 1:00000000 16
2:e000000c 268435456 3:00000000 16 4:00001801 16
0.2.1:	vid  03.80.00 8086/2a03   0 0:f8200004 1048576 1:00000000 16
0.25.0:	net  02.00.00 8086/1049  11 0:fe000000 131072 1:fe025000 4096
2:00001841 32
0.26.0:	usb 0c.03.00 8086/2834  11 4:00001861 32
0.26.1:	usb 0c.03.00 8086/2835  11 4:00001881 32
0.26.7:	usb 0c.03.20 8086/283a  11 0:fe226400 1024
0.27.0:	---  04.03.00 8086/284b  11 0:fe020004 16384 1:00000000 16
0.28.0:	brg  06.04.00 8086/283f  11
0.28.1:	brg  06.04.00 8086/2841  11
0.28.2:	brg  06.04.00 8086/2843  11
0.28.3:	brg  06.04.00 8086/2845  11
0.28.4:	brg  06.04.00 8086/2847  11
0.29.0:	usb 0c.03.00 8086/2830  10 4:000018a1 32
0.29.1:	usb 0c.03.00 8086/2831  11 4:000018c1 32
0.29.2:	usb 0c.03.00 8086/2832  11 4:000018e1 32
0.29.7:	usb 0c.03.20 8086/2836  11 0:fe226800 1024
0.30.0:	brg  06.04.01 8086/2448 255
0.31.0:	brg  06.01.00 8086/2811   0
0.31.2:	disk 01.01.80 8086/2828 255 0:00000001 16 1:00000001 16
2:00000001 16 3:00000001 16 4:00001c31 16 5:00001c21 16
0.31.3:	smb 0c.05.00 8086/283e  11 0:fe226c00 256 4:00001c41 32
21.0.0:	brg  06.07.00 1180/0476  10
21.0.1:	---  0c.00.10 1180/0832  11 0:f8301000 2048
3.0.0:	net  02.80.00 8086/4227  11 0:df3ff000 4096

Re: [9fans] intel ethernet

[ron minnich]

On Jan 2, 2008 2:23 PM, Fazlul Shahriar <fshahriar@gmail.com> wrote:
> > in /sys/src/9/pc/ether82563.c makee this change
> > and recompile a kernel:
> >
> >                         break;
> >                 case 0x1049:            /* mm */

Right, you have this:

        1049  82566MM Gigabit Network Connection

And it's clearly covered in the driver. Hmm.

what do you get when you cat /dev/drivers

ron

Re: [9fans] intel ethernet

[Fazlul Shahriar]

> Right, you have this:
>
>         1049  82566MM Gigabit Network Connection
>
> And it's clearly covered in the driver. Hmm.
>
> what do you get when you cat /dev/drivers

Ok, I meant lejatorn's pci output didn't have 1047 either. Maybe I wasn't clear.

Anyway, here is /dev/drivers:

	#/ root
	#c cons
	#P arch
	#$ pnp
	#e env
	#| pipe
	#p proc
	#M mnt
	#s srv
	#d dup
	#r rtc
	#D ssl
	#a tls
	#¤ cap
	#K kprof
	#k devfs
	#l ether
	#I ip
	#i draw
	#m mouse
	#v vga
	#κ kbmap
	#S sd
	#f floppy
	#L lpt
	#A audio
	#Y cardbus
	#y i82365
	#t uart
	#U usb

Re: [9fans] intel ethernet

[Mathieu Lonjaret]

 Hmm, sorry for being such a noob in decoding those logs, but where do
you see whether or not I have 1047?
Anyway, it's not even my main issue since I can't boot on any other
kernel than the one which is already installed.
So even if adding the 1047 to the driver is the solution to the
network part, I won't even be able to try it...

On Jan 3, 2008 12:07 AM, Fazlul Shahriar <fshahriar@gmail.com> wrote:
> > Right, you have this:
> >
> >         1049  82566MM Gigabit Network Connection
> >
> > And it's clearly covered in the driver. Hmm.
> >
> > what do you get when you cat /dev/drivers
>
> Ok, I meant lejatorn's pci output didn't have 1047 either. Maybe I wasn't clear.
>
> Anyway, here is /dev/drivers:
>
>         #/ root
>         #c cons
>         #P arch
>         #$ pnp
>         #e env
>         #| pipe
>         #p proc
>         #M mnt
>         #s srv
>         #d dup
>         #r rtc
>         #D ssl
>         #a tls
>         #¤ cap
>         #K kprof
>         #k devfs
>         #l ether
>         #I ip
>         #i draw
>         #m mouse
>         #v vga
>         #κ kbmap
>         #S sd
>         #f floppy
>         #L lpt
>         #A audio
>         #Y cardbus
>         #y i82365
>         #t uart
>         #U usb
>



-- 
GPG key on subkeys.pgp.net:

KeyID:   | Fingerprint:
683DE5F3 | 4324 5818 39AA 9545 95C6 09AF B0A4 DFEA 683D E5F3

Re: [9fans] intel ethernet

[Mathieu Lonjaret]

ah nevermind the first question, now I indeed see the "net" line with 1049...

On Jan 3, 2008 12:20 AM, Mathieu Lonjaret <lejatorn@gmail.com> wrote:
>  Hmm, sorry for being such a noob in decoding those logs, but where do
> you see whether or not I have 1047?
> Anyway, it's not even my main issue since I can't boot on any other
> kernel than the one which is already installed.
> So even if adding the 1047 to the driver is the solution to the
> network part, I won't even be able to try it...
>
>
> On Jan 3, 2008 12:07 AM, Fazlul Shahriar <fshahriar@gmail.com> wrote:
> > > Right, you have this:
> > >
> > >         1049  82566MM Gigabit Network Connection
> > >
> > > And it's clearly covered in the driver. Hmm.
> > >
> > > what do you get when you cat /dev/drivers
> >
> > Ok, I meant lejatorn's pci output didn't have 1047 either. Maybe I wasn't clear.
> >
> > Anyway, here is /dev/drivers:
> >
> >         #/ root
> >         #c cons
> >         #P arch
> >         #$ pnp
> >         #e env
> >         #| pipe
> >         #p proc
> >         #M mnt
> >         #s srv
> >         #d dup
> >         #r rtc
> >         #D ssl
> >         #a tls
> >         #¤ cap
> >         #K kprof
> >         #k devfs
> >         #l ether
> >         #I ip
> >         #i draw
> >         #m mouse
> >         #v vga
> >         #κ kbmap
> >         #S sd
> >         #f floppy
> >         #L lpt
> >         #A audio
> >         #Y cardbus
> >         #y i82365
> >         #t uart
> >         #U usb
> >
>
>
>
>
> --
> GPG key on subkeys.pgp.net:
>
> KeyID:   | Fingerprint:
> 683DE5F3 | 4324 5818 39AA 9545 95C6 09AF B0A4 DFEA 683D E5F3
>



-- 
GPG key on subkeys.pgp.net:

KeyID:   | Fingerprint:
683DE5F3 | 4324 5818 39AA 9545 95C6 09AF B0A4 DFEA 683D E5F3

Re: [9fans] intel ethernet

[erik quanstrom]

the missing device was vid 8088 did 1047, i believe.

- erik

p.s. /dev/drivers is something unrelated.  all ethernet
are #l devices.  they are sort of a sub device thing.

Re: [9fans] intel ethernet

[erik quanstrom]

ah, i understand the problem now.

i subitted a patch for this some time ago. the unusable CLS
message is from the driver.  it's checking for cache line size.  an
anacronism from the pci driver.  pcie does not really have the
concept of cache line size, though the vestigal register remains.
the fix is to remobe any reference to cls in the driver and recompile.

- erik