lkml.org 
[lkml]   [1997]   [Apr]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: procfs problems
On Thu, 17 Apr 1997, DUPRE Christophe wrote:
> On Tue, 15 Apr 1997, Dan Hollis wrote:
> > Looking at /proc/pci, it got me to thinking about kernel bloat :-) If we
> > were to take all the text out of pci.c and put it in an external program
> > (call it "pciconfig" or "lspci") and have /proc/pci just output hex
> > numbers, we could save several kilobytes off the size of the kernel. It
> > would also probably be easier for programs to parse /proc/pci if they
> > needed configuration information.
> Let me object on this : most things in /proc must be human-readable
> without looking up a manpage.

I think this is bogus.

/proc/stat, /proc/uptime, /proc/net/icmp, /proc/net/snmp,
/proc/sys/vm/bdflush, etc are all obviously not for human consumption.

> I for one often use this information while setting up a system, using a
> boot/root disk. I don't what to add any software to my root disk which
> is already packed.

Whether the information is in an external program or the kernel, it is
going to take roughly the same amount of space. I for one don't want these
tables (which will continue to grow!) statically in the kernel in
non-swappable space.

There is currently about 5 kilobytes of static text in pci.c, and IMHO it
has absolutely no business being there.

> I often use /proc/pci to know the io address of a network card to
> insmod the driver to do an nfs install... If the text was to suddenly
> disappear from /proc/pci, I don't know how I could use it. Remember,
> it's when all hell breaks loose that /proc is the most usefull (and to
> monitor the system in normal conditions, agreed). And there are not
> many Ks of text in that file... Well, most of it is dynamically
> produced anyway.

In this case I think you are wrong.

-Dan


\
 
 \ /
  Last update: 2005-03-22 13:39    [from the cache]
©2003-2011 Jasper Spaans