EmbeddedRelated.com
Forums
The 2024 Embedded Online Conference

RCM4300 ethernet cable detect problem

Started by mfawzy79 July 2, 2009
Hi,

I'm using the RCm4300 and i have a problem that when i download the code to the module then i connect the Ethernet cable to the module, the module isn't able to detect that the cable is plugged, i have to turn off the power then ON to detect the problem.

by the way, i found a problem in the fat library with the ucos,which is
the _fat_automount function is calling fat_UnmountDevice inline 3437 but it should call _fat_UnmountDevice.

when calling fat_automount under the ucos, it should pend on a semaphore and if it got it, the FAT API _fat_automount will be started the problem that the _fat_automount is calling fat_UnmountDevice which is pending also on the semaphore so will timeout and break the code.
i hope it help
Best Regards
MFawzy

for more info, this problem appears random not all the time
--- In r..., "mfawzy79" wrote:
>
> Hi,
>
> I'm using the RCm4300 and i have a problem that when i download the code to the module then i connect the Ethernet cable to the module, the module isn't able to detect that the cable is plugged, i have to turn off the power then ON to detect the Ethernet Cable.
>
> by the way, i found a problem in the fat library with the ucos,which is
> the _fat_automount function is calling fat_UnmountDevice inline 3437 but it should call _fat_UnmountDevice.
>
> when calling fat_automount under the ucos, it should pend on a semaphore and if it got it, the FAT API _fat_automount will be started the problem that the _fat_automount is calling fat_UnmountDevice which is pending also on the semaphore so will timeout and break the code.
> i hope it help
> Best Regards
> MFawzy
>

On Jul 1, 2009, at 10:49 PM, mfawzy79 wrote:
> I'm using the RCm4300 and i have a problem that when i download the
> code to the module then i connect the Ethernet cable to the module,
> the module isn't able to detect that the cable is plugged, i have
> to turn off the power then ON to detect the problem.

Try Dynamic C 10.54. They fixed a bug with the ASIX driver (RCM4200,
RCM4300) that affected some network setups (I saw it with a 10Base-T
hub).

- Defect 30186 fixed. Boards equipped with ASIX Ethernet now
reliably recover link after a cable unplug, replug sequence.

-Tom
actually, I'm already using DC 10.54 and i'm still having the problem

Best Regards
MFawzy

--- In r..., Tom Collins wrote:
>
> On Jul 1, 2009, at 10:49 PM, mfawzy79 wrote:
> > I'm using the RCm4300 and i have a problem that when i download the
> > code to the module then i connect the Ethernet cable to the module,
> > the module isn't able to detect that the cable is plugged, i have
> > to turn off the power then ON to detect the problem.
>
> Try Dynamic C 10.54. They fixed a bug with the ASIX driver (RCM4200,
> RCM4300) that affected some network setups (I saw it with a 10Base-T
> hub).
>
> - Defect 30186 fixed. Boards equipped with ASIX Ethernet now
> reliably recover link after a cable unplug, replug sequence.
>
> -Tom
>

mfawzy79 wrote:
> actually, I'm already using DC 10.54 and i'm still having the problem
>
>
There is a bug in the ASIX chip that will cause it to not negotiate
correctly with some hubs/switches. The fix was to use a very long reset
to the chip, something like 3 seconds. It would work on a power-up but
not on later resets.

The bug was fixed in the Softools port of the stack. AFAIK, ZW never
implemented the fix. I also notice on the newer modules the no longer
feed a pin to the ASIX chip to use as a reset. So, your only option may
be to get a different hub/switch.

--
------
| Scott G. Henion| s...@shdesigns.org |
| Consultant | Stone Mountain, GA |
| SHDesigns http://www.shdesigns.org |
------

I've, unfortunately, just come across this problem. I'm really stuck because
I'm connecting directly to a PC via a crossover cable.

Anyone got any opinions on the following.

Adding code to try to cycle the interface if it detects it's down for more
than a few seconds.

Allowing an external reset to the board if the interface is down for more
than a few minutes.

Reconfiguring the Gigabit Ethernet port on the PC to lock it at 100FDX???
Any ideas appreciated..

PK

-----Original Message-----
From: r... [mailto:r...] On
Behalf Of Scott Henion
Sent: Sunday, 5 July 2009 3:15 PM
To: r...
Subject: Re: [rabbit-semi] Re: RCM4300 ethernet cable detect problem

mfawzy79 wrote:
> actually, I'm already using DC 10.54 and i'm still having the problem
>
>
There is a bug in the ASIX chip that will cause it to not negotiate
correctly with some hubs/switches. The fix was to use a very long reset
to the chip, something like 3 seconds. It would work on a power-up but
not on later resets.

The bug was fixed in the Softools port of the stack. AFAIK, ZW never
implemented the fix. I also notice on the newer modules the no longer
feed a pin to the ASIX chip to use as a reset. So, your only option may
be to get a different hub/switch.

--
------
| Scott G. Henion| s...@shdesigns.org |
| Consultant | Stone Mountain, GA |
| SHDesigns http://www.shdesigns.org |
------
An update. This problem occurred on a 3375 module, I was able to fix it by
locking the Gigabit Ethernet port on the PC to 100FDX.

PK

-----Original Message-----
From: r... [mailto:r...] On
Behalf Of Paul Kelly
Sent: Thursday, 6 August 2009 5:49 PM
To: r...
Subject: RE: [rabbit-semi] Re: RCM4300 ethernet cable detect problem

I've, unfortunately, just come across this problem. I'm really stuck because
I'm connecting directly to a PC via a crossover cable.

Anyone got any opinions on the following.

Adding code to try to cycle the interface if it detects it's down for more
than a few seconds.

Allowing an external reset to the board if the interface is down for more
than a few minutes.

Reconfiguring the Gigabit Ethernet port on the PC to lock it at 100FDX???
Any ideas appreciated..

PK

-----Original Message-----
From: r... [mailto:r...] On
Behalf Of Scott Henion
Sent: Sunday, 5 July 2009 3:15 PM
To: r...
Subject: Re: [rabbit-semi] Re: RCM4300 ethernet cable detect problem

mfawzy79 wrote:
> actually, I'm already using DC 10.54 and i'm still having the problem
>
>
There is a bug in the ASIX chip that will cause it to not negotiate
correctly with some hubs/switches. The fix was to use a very long reset
to the chip, something like 3 seconds. It would work on a power-up but
not on later resets.

The bug was fixed in the Softools port of the stack. AFAIK, ZW never
implemented the fix. I also notice on the newer modules the no longer
feed a pin to the ASIX chip to use as a reset. So, your only option may
be to get a different hub/switch.

--
------
| Scott G. Henion| s...@shdesigns.org |
| Consultant | Stone Mountain, GA |
| SHDesigns http://www.shdesigns.org |
------

The 2024 Embedded Online Conference