From: John Tyner <hidden> Date: 2002-02-27 23:32:04
From andrew may [off-list ref]:
Added Am79C874 support
Improved find_phy function:
use Min/Max Phy addr defines (not sure if MAX_NUM_PHYS is still needed)
continue to look for phy if we find a phy we don't support instead of panic
detect/report missing pullup problem
removed unused phy_id_done
enet_open
checks for undetected phy and fails if no know phy is present
enet_open/close
checks for completion of soft reset (writing 0 to soft reset has no effect)
Added Am79C874 support
Improved find_phy function:
use Min/Max Phy addr defines (not sure if MAX_NUM_PHYS is still needed)
continue to look for phy if we find a phy we don't support instead of panic
detect/report missing pullup problem
removed unused phy_id_done
enet_open
checks for undetected phy and fails if no know phy is present
enet_open/close
checks for completion of soft reset (writing 0 to soft reset has no effect)
------------------------------------------------------------------------
Thanks,
I want to try in on a few boards I have before it goes in.
--armin
** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
From: andrew may <hidden> Date: 2002-03-04 21:17:03
On Wed, Feb 27, 2002 at 04:35:05PM +0000, Armin wrote:
Thanks,
I want to try in on a few boards I have before it goes in.
--armin
Have you gotten a chance to try it yet? The testing I have
done has not shown any problems and I have been able to test
just about all of the cases. Boards with/without the correct
pull-ups. Remove the correct phy-id from the table. Booting
without a link.
Here is another patch that for the driver. It removes some
cut and paste code from the rx interupts and breaks the work
into rx_clean and rx_fill functions. rx_fill is also called from
open.
If rx_fill fails to alloc a skb, it leave the rx_slot there, so
the next time through rx_clean skips over all the slots without
skb's and hopefully rx_fill with then be able to do the alloc.
This prevents taking a rxde interrupt for a single alloc failure.
I also added a skb_res module parm, I should have defaulted it
down to 0 for most people, but I forgot and left it at the value
I use.
On Wed, Feb 27, 2002 at 04:35:05PM +0000, Armin wrote:
quoted
Thanks,
I want to try in on a few boards I have before it goes in.
--armin
Have you gotten a chance to try it yet? The testing I have
done has not shown any problems and I have been able to test
just about all of the cases. Boards with/without the correct
pull-ups. Remove the correct phy-id from the table. Booting
without a link.
Here is another patch that for the driver. It removes some
cut and paste code from the rx interupts and breaks the work
into rx_clean and rx_fill functions. rx_fill is also called from
open.
Yes I have.
I added history info and sent it to Tom to check in.
I will look into this patch and will try it after Tom gets done.
armin
** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
From: andrew may <hidden> Date: 2002-03-05 03:53:28
On Mon, Mar 04, 2002 at 02:31:15PM +0000, Armin wrote:
andrew may wrote:
quoted
On Wed, Feb 27, 2002 at 04:35:05PM +0000, Armin wrote:
quoted
Thanks,
I want to try in on a few boards I have before it goes in.
--armin
Have you gotten a chance to try it yet? The testing I have
done has not shown any problems and I have been able to test
just about all of the cases. Boards with/without the correct
pull-ups. Remove the correct phy-id from the table. Booting
without a link.
Here is another patch that for the driver. It removes some
cut and paste code from the rx interupts and breaks the work
into rx_clean and rx_fill functions. rx_fill is also called from
open.
Yes I have.
I added history info and sent it to Tom to check in.
Thanks
I will look into this patch and will try it after Tom gets done.
Well I just noticed that building as a module is pretty far from
working at the moment. Are there any plans to fix this? I am not
sure how to cleanly get this working since there are all sorts of
exports/externs that need to be fixed up. The tulip driver should
be a good example, but it is broken off into it's own dir and I
don't know if we want to do this for the ocp_enet as well.
I can resend with a .h change and config.in change to pull the
skb_res value out as a build option, but I want to see if anyone
has any other comments before hand.
** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Well I just noticed that building as a module is pretty far from
working at the moment. Are there any plans to fix this?
Havn't thought about it but sound like it needs to addressed.
I am not
sure how to cleanly get this working since there are all sorts of
exports/externs that need to be fixed up.
I can add it to my TODO list unless you need it asap?
I
don't know if we want to do this for the ocp_enet as well.
fine with me.
I can resend with a .h change and config.in change to pull the
skb_res value out as a build option, but I want to see if anyone
has any other comments before hand.
From: andrew may <hidden> Date: 2002-03-06 22:15:00
On Tue, Mar 05, 2002 at 10:51:38AM +0000, Armin wrote:
andrew may wrote:
quoted
Well I just noticed that building as a module is pretty far from
working at the moment. Are there any plans to fix this?
Havn't thought about it but sound like it needs to addressed.
quoted
I am not
sure how to cleanly get this working since there are all sorts of
exports/externs that need to be fixed up.
I can add it to my TODO list unless you need it asap?
No there is no rush getting this working as a module. I would like
a nice way to set the amount to reserve on the skb dynamically
though.
quoted
I
don't know if we want to do this for the ocp_enet as well.
fine with me.
So would it be easier to just ask someone with bk access to the
tree to add the dir and move the files rather than do a patch.
** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
On Tue, Mar 05, 2002 at 10:51:38AM +0000, Armin wrote:
quoted
andrew may wrote:
quoted
Well I just noticed that building as a module is pretty far from
working at the moment. Are there any plans to fix this?
Havn't thought about it but sound like it needs to addressed.
quoted
I am not
sure how to cleanly get this working since there are all sorts of
exports/externs that need to be fixed up.
I can add it to my TODO list unless you need it asap?
No there is no rush getting this working as a module. I would like
a nice way to set the amount to reserve on the skb dynamically
though.
quoted
quoted
I
don't know if we want to do this for the ocp_enet as well.
fine with me.
So would it be easier to just ask someone with bk access to the
tree to add the dir and move the files rather than do a patch.
There should be a patch since we need to have our own Makefile and its
my responsiblilty to do the work not them. I will have it done in the
next few days.
--armin
** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/