On 15-Nov-2001 Craig Coles wrote: > I have got three different machines, out of serveral more, that are > reporting the message: > eth0: Too much work in interrupt, status e401 > > The message doesn't appear continually, but periodically through-out the > day, and sometimes a few times per minute. Most of the boxes I work with > use 'unstable' with 2.2.19 and 2.4.3 kernels, and one a different distro. > The problem is being seen on the 'other distro' box, and two of my > 'unstable' installs (I have many other unstable installs too that aren't > seeing the problem). > > Does anyone know what causes this, what is happening, or what I can do to > stop the messages? >
do: rgrep 'Too much work in interrup' * in your kernel source tree. What this is all about is each time the interrupt handler is called it has a loop which reads through the queue of waiting packets. If there are more items in the queue than the number of iterations this message happens. Basically, you are getting more data per interrupt than expected and getting backlogged. The docs and source explain some possible solutions.
