RE: [ECOS] ping is resulting in 50% packet loss.

["George Sosnowski" <george at stratalight dot com>]

Yes, as Gary mentioned this should most definitely solve your problem,
and use the updated cvs version if you can.

>From what I recall, it was ~4 or 5 months back, I could see the proper
ICMP responses being put in the tx buffers but the fec was never txing
some of them because it was skipping 50% of the buffer descriptor
entries. So the results you showed from the ping looked very familiar.

-----Original Message-----
From: Gary Thomas [mailto:gary@chez-thomas.org] 
Sent: Monday, July 22, 2002 9:50 AM
To: Khasim
Cc: George Sosnowski; eCos Discussion
Subject: Re: [ECOS] ping is resulting in 50% packet loss.

On Mon, 2002-07-22 at 10:43, Khasim wrote:
> Thanks George
> But according to me I find the problem in fec_Txevent
> I am little confused..
> The Fec transmitt is handled by interrupts here. If the tx interrupt
does
> not come then what will happen to the data.It will be delayed?
> 
> I will enable the change given by you .
> 
> 
> " Sorry GARY Thomas I am not using the new cvs updated code.may be
tomorrow
> or day after I will try this"
> But will this change help me?????
> 

Without it, the driver was quite broken.

Update to the latest CVS - that's why it's out there!

> Thanks again George..
> 
> With regards
> 
> Khasim
> :) :)
> ----- Original Message -----
> From: "George Sosnowski" <george@stratalight.com>
> To: <ecos-discuss@sources.redhat.com>
> Sent: Monday, July 22, 2002 8:50 PM
> Subject: RE: [ECOS] ping is resulting in 50% packet loss.
> 
> 
> I had a similar problem when using the 860T fec driver initially.
> What it turned out to be was the IP stack or upper level networking
code
> was calling fec_eth_start and fec_eth_init and both functions were
> setting the eConrtol_EN bit in the eControl byte. This was screwing up
> the state machine of the fec and causing it to act goofy. My work
around
> for this was in fec_eth_start to check if eControl_EN was not enabled
in
> the eControl byte before enabling it there.
> 
> if(!(qi->fec->eControl & eControl_EN))
> {
>   qi->fec->eControl |= eControl_EN;
>   qi->fec->RxUpdate = 0x0f0f0f0f;
> }
> 
> -----Original Message-----
> From: Khasim [mailto:khasim@iwavesystems.com]
> Sent: Sunday, July 21, 2002 9:19 PM
> To: ecos-discuss@sources.redhat.com
> Subject: [ECOS] ping is resulting in 50% packet loss.
> 
> Greetings
> 
> I saw same sort of question in the Discussion list but I didnt get the
> answer properly.
> 
> I have ported the TCP/IP stack it is running from FLASH no redboot.
When
> I
> ping from other windows machine to my boards ip address the ping is
> resulting in 60 or some times 50% packet loss.
> 
> * I have not enabled any prints or debug messages.
> * But when I see in ethreal each packet is replied.
> * I dont know which TCP/IP stack I am using old or new.How to know
> this???
> 
> Suggesstions please....
> 
> With regards
> Khasim..
> 
> Here is the output on cmd prompt
> 
> C:\>ping -n 100 192.168.1.21
> 
> Pinging 192.168.1.21 with 32 bytes of data:
> 
> Request timed out.
> Request timed out.
> Reply from 192.168.1.21: bytes=32 time=501ms TTL=255
> Request timed out.
> Request timed out.
> Reply from 192.168.1.21: bytes=32 time=250ms TTL=255
> Reply from 192.168.1.21: bytes=32 time=250ms TTL=255
> Request timed out.
> Request timed out.
> Request timed out.
> Reply from 192.168.1.21: bytes=32 time=661ms TTL=255
> Request timed out.
> Request timed out.
> Reply from 192.168.1.21: bytes=32 time=821ms TTL=255
> Reply from 192.168.1.21: bytes=32 time=250ms TTL=255
> Request timed out.
> Request timed out.
> Reply from 192.168.1.21: bytes=32 time=251ms TTL=255
> Reply from 192.168.1.21: bytes=32 time=250ms TTL=255
> Request timed out.
> Request timed out.
> Reply from 192.168.1.21: bytes=32 time=251ms TTL=255
> Reply from 192.168.1.21: bytes=32 time=250ms TTL=255
> Request timed out.
> Request timed out.
> Reply from 192.168.1.21: bytes=32 time=250ms TTL=255
> Reply from 192.168.1.21: bytes=32 time=251ms TTL=255
> Request timed out.
> Request timed out.
> Reply from 192.168.1.21: bytes=32 time=250ms TTL=255
> Reply from 192.168.1.21: bytes=32 time=250ms TTL=255
> Request timed out.
> Request timed out.
> Reply from 192.168.1.21: bytes=32 time=251ms TTL=255
> Reply from 192.168.1.21: bytes=32 time=250ms TTL=255
> Request timed out.
> Request timed out.
> 
> Ping statistics for 192.168.1.21:
>     Packets: Sent = 40, Received = 16, Lost = 24 (60% loss),
> Approximate round trip times in milli-seconds:
>     Minimum = 250ms, Maximum =  821ms, Average =  130ms
> Control-C
> 
> 
> --
> Before posting, please read the FAQ:
http://sources.redhat.com/fom/ecos
> and search the list archive: http://sources.redhat.com/ml/ecos-discuss
> 
> 
> --
> Before posting, please read the FAQ:
http://sources.redhat.com/fom/ecos
> and search the list archive: http://sources.redhat.com/ml/ecos-discuss
> 
> 
> 
> 
> -- 
> Before posting, please read the FAQ:
http://sources.redhat.com/fom/ecos
> and search the list archive: http://sources.redhat.com/ml/ecos-discuss



--
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss