On Sat, 10 Jun 2000, Dave Hayes wrote:
> [ Please CC me on responses. Thank you. ]
>
> On 3.3-STABLE the following kernel message appeared recently:
>
> pmap_collect: collecting pv entries -- suggest increasing PMAP_SHPGPERPROC
>
> Is there a place I can get more specific information as to what this
> means, so I can perhaps embark on the road to a correct solution
> and/or interpretation of what happened? Alternatively, can someone
> explain?
>
Sounds like you are running some programs that make heavy use of shared
memory. Do as it says. Basically, pv_entries map logical address to physical
addresses. With lots of shared memory segments, lots of pv_entries exist to
map all of the per-process address mappings to the physical memory backing it.
There are only a certain number of pv_entries that exist in the kernel, and
you are running out (actually, you get this warning when over 90% of them are
used). If you up PMAP_SHPGPERPROC, you increase the number of pv_entries
created at boot time. However, I am not informed enough to say how high you
can safely increase PMAP_SHPGPERPROC.
> On a related note, I took a look at the code and saw the following
> disturbing thing:
>
[code snipped]
>
> The machine in question has 57 days of uptime. This code appears to
> imply that I only get to see this 5 times during this period. :)
How many times do you need to be told? :)
--
Kelly Yancey - [EMAIL PROTECTED] - Belmont, CA
System Administrator, eGroups.com http://www.egroups.com/
Maintainer, BSD Driver Database http://www.posi.net/freebsd/drivers/
Coordinator, Team FreeBSD http://www.posi.net/freebsd/Team-FreeBSD/
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
