TSEC/PHY @ MPC834x stopped working

Ben,

after pulling u-boot today ethernet stopped working on my MPC8343 based board (mvBL-M7).
I'm using a Vitesse VSC8601 RGMII phy on TSEC0-1.

I know there's some rewriting/clean-up going on ...

What do I have to do to get it working again ?
Any new defines introduced lately ?

Obviously the MII interface stopped working, i.e. all reads give 0xFF. Switching back to older U-boot (3-4 weeks ago) works fine.

Any help is welcome.

regards,
Andre

MATRIX VISION GmbH, Talstra?e 16, DE-71570 Oppenweiler - Registergericht: Amtsgericht Stuttgart, HRB 271090 Gesch?ftsf?hrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner

Hi Andre,

Quoted 9 lines · Andre Schwarz

Ben,

after pulling u-boot today ethernet stopped working on my MPC8343 based board (mvBL-M7).
I'm using a Vitesse VSC8601 RGMII phy on TSEC0-1.

I know there's some rewriting/clean-up going on ...

Yeah, but none of that's in the main source tree.

Quoted 6 lines

What do I have to do to get it working again ?
Any new defines introduced lately ?

Obviously the MII interface stopped working, i.e. all reads give 0xFF. Switching back to older U-boot (3-4 weeks ago) works fine.

Looking here:
http://git.denx.de/?p=u-boot.git;a=history;f=drivers/net/tsec.c;h=6e0f2c6fd081baa19d80a985717d470a1f78bc2d;hb=HEAD

nothing new has been added except support for a new PHY, and that was about 5 weeks ago.

Quoted 3 lines

Any help is welcome.

Sorry, I don't have access to my MPC8349 board right now or I'd try to help more usefully. Maybe you can post some more information.

Quoted 3 lines

regards,
Andre

Quoted 4 lines

MATRIX VISION GmbH, Talstra?e 16, DE-71570 Oppenweiler - Registergericht: Amtsgericht Stuttgart, HRB 271090 Gesch?ftsf?hrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner

regards,
Ben

Ben,

thanks for your quick reply.
Looks like you're right - nothing changed regarding network yet.

Maybe it's a cpu specific problem.
Kim ? Can you help here ?

Quoted 5 lines · On Wed, 02 Jul 2008 16:36:11 +0200 Andre Schwarz…

Ben,

thanks for your quick reply.
Looks like you're right - nothing changed regarding network yet.

Quoted 3 lines

Maybe it's a cpu specific problem.
Kim ? Can you help here ?

I know of nothing (barring this being the ETHADDR not specified problem) that could have caused this. Can you try using git-bisect?

Kim

Obviously the MII interface stopped working, i.e. all reads give 0xFF.

Quoted 9 lines · On Wed, 02 Jul 2008 16:36:11 +0200 Andre Schwarz…

Ben,

thanks for your quick reply.
Looks like you're right - nothing changed regarding network yet.

Maybe it's a cpu specific problem.
Kim ? Can you help here ?

I know of nothing (barring this being the ETHADDR not specified problem) that could have caused this. Can you try using git-bisect?

Kim

Quoted 8 lines · Kim Phillips

Quoted 5 lines · On Wed, 02 Jul 2008 16:36:11 +0200 Andre Schwarz…

Ben,

thanks for your quick reply.
Looks like you're right - nothing changed regarding network yet.

Quoted 8 lines

Quoted 3 lines

Maybe it's a cpu specific problem.
Kim ? Can you help here ?

I know of nothing (barring this being the ETHADDR not specified problem) that could have caused this. Can you try using git-bisect?

Kim

Some people were caught by the somewhat recent requirement for an /aliases node and properties pointing to the "real" ethernet node (which is a Good Thing[tm] because it makes the code board agnostic, pushes the board configuration into the .dts where it belongs).

I don't associate Andre's assertion:

Quoted 1 line

Obviously the MII interface stopped working, i.e. all reads give 0xFF.

with the symptoms caused by a missing /aliases node. Maybe, maybe not, could be my lack of experience.

Best regards,
gvb

Jerry,

I don't understand this.

The code is far from using the device tree - the dtb is fetched by tftp right after bootp.

It's the initial bootp/dhcp that doesn't work....

regards,
Andre

this is a fully configured board, i.e. serial# and ethaddr/eth1addr set up correctly.

I'll give the bisect a try a soon as there'll be some time.

Quoted 6 lines · On Wed, 02 Jul 2008 18:01:01 +0200 Andre Schwarz…

this is a fully configured board, i.e. serial# and ethaddr/eth1addr set up correctly.

I'll give the bisect a try a soon as there'll be some time.

just to clarify, are you using the mainline u-boot.git tree, or u-boot-mpc83xx.git? you should be using the latter until I ask WD to pull. Actually, if you can confirm it corrects your problem, I have no reason to delay it further (I was waiting for WD to get back from vacation).

Kim