Patrick Thomas wrote:
Incidently, looking at the PV entry angle for a moment. Suppose you
create a 1GB sysvshm (pageable) segment. That's 262144 pages. Mapping thi
s
once means you consume 262144 PV entries. At 28 bytes each, that is
about 7.3MB of KVM. Now, fork this process
Peter Wemm wrote:
Even more importantly it would be nice if we could share compatible
pmap pages, then we would have no need for 4MB pages... 50 mappings
of the same shared memory segment would wind up using the same pmap
pages as if only one mapping had been made. Such a
Hmm. I'm fairly sure that Linux does not quite do it that way. I
believe the 2-level page tables are copy-on-write, but that only
gives you shareability across a fork() and then only for a little
while. I'm fairly certain that Linux cannot share page tables
for post-fork
Patrick Thomas wrote:
Uhh, Terry, neither form of SHM uses KVA. Either pageable or physically
backed. The memory is only mapped into processes and is NOT mapped into
KVA anywhere. (*)
You and Alfred are right.
It looks like this was corrected by John Dyson, shortly after he
Patrick Thomas wrote:
Ok, I'm confused now - so I understood you to originally say that SHM does
not eat into KVA regardless of whether I set the kern.ipc.shm_use_phys to
'1' or not.
Don't be confused. It takes 1/1024th of the KVA space I thought
it did, times the number of processes atached
Matthew Dillon wrote:
Hmm. I'm fairly sure that Linux does not quite do it that way. I
believe the 2-level page tables are copy-on-write, but that only
gives you shareability across a fork() and then only for a little
while. I'm fairly certain that Linux cannot share page
: http://old.lwn.net/2002/0124/kernel.php3
: What Rik van Riel is up to.
:
:FWIW: In the original Article:
:
: http://lwn.net/Articles/3327/
:
:they say they've gon to a 3 level page table scheme for the Hammer
:port.
:
:-- Terry
They are still using a 2 level page table
:[commenting live from ottawa]
:
:On Wed, Jun 26, 2002 at 10:11:54AM -0700, Matthew Dillon wrote:
: They are still using a 2 level page table scheme in the linux
: kernel. That is, they didn't try to integrate the 3-level hardware
: in the hammer with the 2 level kernel
:[commenting live from ottawa]
Pictures! We want pictures!
-Matt
Matthew Dillon
[EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe
On Wed, Jun 26, 2002 at 11:24:47AM -0700, Matthew Dillon wrote:
:[commenting live from ottawa]
Pictures! We want pictures!
It's pretty cool that the Linux camp has decided to do the
Summit stuff too (I'm assuming that this is a relatively new
phenomenon).
* Terry Lambert [EMAIL PROTECTED] [020624 19:58] wrote:
Alfred Perlstein wrote:
* Terry Lambert [EMAIL PROTECTED] [020624 19:17] wrote:
System V shared memory is allocated out of KVA space (annoying,
but true).
You keep saying this but the backing object allocated for sysvshm
Alfred Perlstein wrote:
You keep saying this but the backing object allocated for sysvshm
is taken from either an OBJT_PHYS or OBJT_SWAP object.
Uh, it's only ever an OBJT_SWAP; see line 532 of kern/sysv_shm.c.
Your sources seem to be really out of date...
Yes. I was referring to
* Terry Lambert [EMAIL PROTECTED] [020625 13:08] wrote:
At what point does it eat KVA that is other than for the backing
data structures?
It eats address space, not RAM. And even if the mappings are not
active (which they usually are, because of LRU and processes
accessing
:
:This is also the _default_ for how solaris manages sysv segments,
:although it would be nice if we could get the OBJT_PHYS stuff to
:use 4meg pages (unless someone already did that?)...
:
:Anyhow, I'm glad we corrected your misconception and we now have
:a more accurate understanding of how
Incidently, looking at the PV entry angle for a moment. Suppose you
create a 1GB sysvshm (pageable) segment. That's 262144 pages. Mapping this
once means you consume 262144 PV entries. At 28 bytes each, that is
about 7.3MB of KVM. Now, fork this process 300 times. The numbers become
As a splinter to the ongoing KVA/crash/memory discussion, I am wondering:
- given a machine that will run 250+ httpds and another ~800 misc.
processes, what system tunings would any of you suggest other than the
ones I have done:
In my kernel: maxusers=256 (was 512, change to 256
On Mon, 24 Jun 2002, Patrick Thomas wrote:
Two things about this interested me:
a) watching `top` output anytime of the day, i see several httpd processes
in sbwait - granted I can only see 40 lines of processes or so in `top`,
but usually at least two show sbwait. Worrisome ?
sbwait is
Patrick Thomas wrote:
As a splinter to the ongoing KVA/crash/memory discussion, I am wondering:
- given a machine that will run 250+ httpds and another ~800 misc.
processes, what system tunings would any of you suggest other than the
ones I have done:
Can't comment. Best guess: whatever
18 matches
Mail list logo