Date: Tue, 06 May 2008 08:01:39 +0100 From: Dieter <freebsd@sopwith.solgatos.com> To: Giorgos Keramidas <keramida@ceid.upatras.gr>, Mel <fbsd.questions@rachie.is-a-geek.net>, freebsd-questions@freebsd.org Subject: Re: 6.2 -> 7.0 now mlock(2) fails Message-ID: <200805061501.PAA11211@sopwith.solgatos.com> In-Reply-To: Your message of "Sun, 13 Apr 2008 01:00:41 %2B0300." <87ve2m7mye.fsf@kobe.laptop>
next in thread | previous in thread | raw e-mail | index | archive | help
> It is. The pageout daemon initializes max_wired with:
>
> /* XXX does not really belong here */
> if (vm_page_max_wired == 0)
> vm_page_max_wired = cnt.v_free_count / 3;
>
I installed the following:
diff -r1.1 vm_pageout.c
1421c1421,1425
< vm_page_max_wired = cnt.v_free_count / 3;
---
> {
> printf("vm_page_max_wired was 0. cnt.v_free_count / 3 = %d\n", cnt.v_free_count / 3);
> printf("setting vm_page_max_wired to cnt.v_free_count / 2 = %d\n", cnt.v_free_count / 2);
> vm_page_max_wired = cnt.v_free_count / 2;
> }
Which gives:
vm_page_max_wired was 0. cnt.v_free_count / 3 = 165410
setting vm_page_max_wired to cnt.v_free_count / 2 = 248115
No mlock failures in ~2.5 weeks.
The largest value logged for vm.stats.vm.v_wire_count is 188614.
Thanks!
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200805061501.PAA11211>