🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: ch.comp.os.linux
Article #73

Re: Network card problem - (Linux Kernel 2.2.16 - de4x5)

#73
From: Marc SCHAEFER
Date: Sat, 19 Aug 2000 00:00
56 lines
2712 bytes
In ch.comp.os.linux Marc Hauswirth <marc.hauswirth@urbanet.ch> wrote:
> Regullary (4-6 / year) we have problemes with them :
>   kernel: eth0: Insufficient memory; nuking packet.
>   last message repeated 430 times ...
>   (this probleme occurs to in in and off peak hours)

Did you try to change the VM settings so that more memory
is maintained in the `free' state ?

Documentation is in 

   /usr/src/linux/Documentation/sysctl/vm.txt
   /usr/src/linux/Documentation/proc.txt

Excerpt:

   freepages
      This file contains three values: min, low and high:
   
      min
        When the number of free pages in the system reaches this number,
        only the kernel can allocate more memory.
   
      low
        If the number of free pages gets below this point, the kernel
        starts swapping aggressively.
   
      high
        The kernel tries to keep up to this amount of memory free; if
        memory comes below this point, the kernel gently starts swapping
        in the hopes that it never has to do really aggressive swapping.

I suggest you try to change those values so they are a bit bigger. This
must be done at each reboot with something like

   echo 254 508 762 > /proc/sys/vm/freepages

But first try by hand with root to see if it improves the `free' output.
   
I also wanted to suggest upgrading to 2.2.17pre, since this contains
some TCP hang fixes, however I don't think it has any relevance to
your problem.

Does the problem resolve itself automatically (ie: once it happens,
there is again a pause of N months before the next, or once it
happens a reboot is required to fix it?  While the problem is
happening, do you still have some kind of service, or is the
service dead or just slower?)

> Did somebody else have the same problems ?
>   (with the same card or with others network cards) ?

It looks like the above is a resource-starvation problem, which is
not ethernet dependant. However the message you describe is printed
within drivers/net/de4x5, when memory allocation with dev_alloc_skb()
fails, using GFP_ATOMIC allocation class (ie within free memory).
Technical Details
Message-ID: <8nllks$d8n$1@vulcan.alphanet.ch>
Path: arc1-ca.pugleaf.net!unknown!not-for-mail
References: <399BF490.9D7B713C@urbanet.ch>

Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: fa123,11243a06494c675b
X-Google-Attributes: gidfa123,public
X-Google-Thread: f4b73,11243a06494c675b
X-Google-Attributes: gidf4b73,public
From: Marc SCHAEFER <schaefer@alphanet.ch>
Subject: Re: Network card problem - (Linux Kernel 2.2.16 - de4x5)
Date: 2000/08/19
Message-ID: <8nllks$d8n$1@vulcan.alphanet.ch>
X-Deja-AN: 660039105
References: <399BF490.9D7B713C@urbanet.ch>
Organization: ALPHANET NF -- Not for profit telecom research
Newsgroups: ch.comp.os.linux,ch.comp.networks