[vortex] eth0: Too much work in interrupt, ...

[Hamid Reza Shahriari]

On Wed, 13 Sep 2000, Bogdan Costescu wrote:

> On Tue, 12 Sep 2000, Hamid Reza Shahriari wrote:
> 
> > Also I got some messages about insufficient memory from server (not
> > kernel).
> 
> The memory is used in common by the userland applications and the
> kernel; if you run short of memory and the kernel needs to allocate some
> (which is the case in the network driver), it will fail. This might
> explain the sudden halts. Can you try to increase the RAM size (just
> temporarily, by borrowing some memory modules from other computer) and
> test again ?
> 
Unfortunately, I used my all two motherboard mem slots by each 200MB. I
have not 512MB RAM chip. So I must buy a new 512MB chip.
It's 'free' output when testing: 
# free
             total       used       free     shared    buffers     cached
Mem:        516436     513428       3008      13100     215288      44808
-/+ buffers/cache:     253332     263104
Swap:        72252       2176      70076

> > I increased max_interrupt_work to 32 and 64, and TX_RING_SIZE to 128
> > Now it's better, but after two hours the server halts without any kernel
> > message. It seems that all interrupts were disabled.
> 
> What do you mean by "it's better" : do the messages still appear ? Do you
> get better throughput ?
>

I means that error messages were reduced, and in high loads, it halts
later.
The main bottleneck for me is kernel halt that is without any message when
halts! Is this due to disabling interrupts in NIC driver?

Sincerely,
--Hamid