Richard A Downing wrote:
> Bruce Dubbs wrote:

> Not wishing to sow dissent in the ranks, but why do you prefer the kill
> method?  

Because I'm old and set in my ways.  :)

My method works reliably, and seems to be the documented method
> by which the userland tells the kernel module nfsd.o how many [nfsd]
> threads to run.  Check out 'man rpc.nfsd'.

I just stated a personal preference.  I'm OK with your method for the book.

> Impudent possibly, but I think using kill is plain wrong (even if RH do
> it).  Kill is not, je pense, intended to send signals to kernel processes.

Here I disagree.  The kill command is misnamed.  It really should be
named 'signal'.  In general, a signal is valid for processes running in
either kernel space or user space.

OTOH, I just tool a look at the source (what a concept :) ) and rpc.nfsd
sets the number of processes in /proc makes a syscall to the kernel to
make it happen.  After reviewing this, I will have to admit that your
method seems to be the most technically correct.

  -- Bruce

-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Reply via email to