When last heard from, I was saying that inefficiencies in ld.so could account
for only a small percentage of the pages I was seeing.

I've now accounted for the rest.  They seem to emerge from the way the Mach vm
system deals with anonymous mappings (meaning mallocs).  Apparently, at some
point it wants to start converting them into copy objects, with shadow objects
and duplicate pages.  I don't quite understand where or why yet, but I'm
assuming it has something to do with the swap pager.  (Brilliant deductions,
all.  Mind like a razor.)

Most of you probably knew this already.  I've been in and out of this mailing
list for a couple of years, and have a copy of CSRG book, but this point eluded
me until I saw it in action with ddb.

I gather it's one of the major points addressed by UVM, so I look forward to
seeing how UVM works.

Meanwhile, ddb did reveal something else.  Watching 'systat vmstat' (Cranor's
favorite), I see that each new process brings 6 wired pages with it.  Yet
looking at the actual vm_map with ddb, I see only 3 wired pages.  Can anyone
expain those different numbers?


--Jim Howard  <jiho@mail.c-zone.net>


----------------------------------
E-Mail: jiho@mail.c-zone.net
Date: 25-Mar-98
Time: 15:46:47

This message was sent by XFMail
----------------------------------