Re: 2.6.13-mm3

[Martin J. Bligh]

--Sonny Rao <sonny@xxxxxxxxxxx> wrote (on Tuesday, September 13, 2005 03:02:04 -0400):

> On Tue, Sep 13, 2005 at 05:14:34AM +0000, Danny ter Haar wrote:
>> Sonny Rao  <sonny@xxxxxxxxxxx> wrote:
>> > Are you using jumbo frames or anything like that?
>> 
>> Not as far as i know.
>> 
>> I gave the kernel some more buffer as stated on
>> http://home.cern.ch/~jes/gige/acenic.html
>> 
>> echo 256144 > /proc/sys/net/core/rmem_max
>> echo 262144 > /proc/sys/net/core/wmem_max
> 
> Not sure if this could lead to higher order allocations -- the only
> place I think it might happen is in sock_kmalloc() 
> 
> Dunno, Martin?

Not sure. Throw a dump_stack in __alloc_pages conditional on if order > 0.
If it spews too much output, ratelimit it by incrementing a static variable
each time, and only printing if mod 100 or 1000 or something, but I suspect
it'll be OK at full rate.

M.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/