mbuf autotuning effect

[hiren panchasara]

On Sep 6, 2013 8:26 PM, "Warner Losh" <imp at bsdimp.com> wrote:
>
>
> On Sep 6, 2013, at 7:11 PM, Adrian Chadd wrote:
>
> > Yeah, why is VM_KMEM_SIZE only 12mbyte for MIPS? That's a little low
for a
> > platform that has a direct map that's slightly larger than 12mb :)
> >
> > Warner? Juli?
>
> All architectures have it at 12MB, except sparc64 where it is 16MB. This
can be changed with the options VM_KMEM_SIZE=xxxxx in the config file.

Right. Does that mean for any platform, if we do not have nmbclusters
pre-set in kmeminit() than we will always have pretty low value of
vm_kmem_size. And because of that, if maxmbufmem is not pre-set (via
loader.conf) inside tunable_mbinit() , we will have very low value for
maxmbufmem too.

I hope (partially believe) that my understanding is not entirely correct.
Because if its correct, we arw depending on loader.conf instead of actually
auto tuning.

Thanks,
Hiren
>
> So my guess as to why this is the case: cut and paste worked, so nobody
changed it after that.
>
> # Still need to reads hiren's email to comprehend it...
>
> Warner
>
>
> >
> >
> > -adrian
> >
> >
> >
> > On 6 September 2013 16:36, hiren panchasara <hiren.panchasara at gmail.com
>wrote:
> >
> >> We are seeing an interesting thing on a mips board with 32MB ram.
> >>
> >> We run out of mbuf very easily and looking at numbers it seems we are
only
> >> getting 6mb of maxmbufmem.
> >>
> >> # sysctl -a | grep hw | grep mem
> >> hw.physmem: 33554432
> >> hw.usermem: 21774336
> >> hw.realmem: 33554432
> >> #
> >> # sysctl -a | grep maxmbuf
> >> kern.ipc.maxmbufmem: 6291456
> >>
> >> I believe that number is very low for a board with 32mb of ram.
> >>
> >> Looking at the code:
> >>
> >> sys/kern/kern_mbuf.c : tunable_mbinit()
> >>
> >> 124         realmem = qmin((quad_t)physmem * PAGE_SIZE, vm_kmem_size);
> >> 125         maxmbufmem = realmem / 2;
> >> 126         TUNABLE_QUAD_FETCH("kern.ipc.maxmbufmem", &maxmbufmem);
> >> 127         if (maxmbufmem > realmem / 4 * 3)
> >> 128                 maxmbufmem = realmem / 4 * 3;
> >>
> >> So, realmem plays important role in determining maxmbufmem.
> >>
> >> physmem = 32mb
> >> PAGE_SIZE = 4096
> >>
> >> vm_kmem_size is calculated inside sys/kern/kern_malloc.c : kmeminit()
> >>
> >> 705         vm_kmem_size = VM_KMEM_SIZE + nmbclusters * PAGE_SIZE;
> >> 706         mem_size = cnt.v_page_count;
> >> 707
> >> 708 #if defined(VM_KMEM_SIZE_SCALE)
> >> 709         vm_kmem_size_scale = VM_KMEM_SIZE_SCALE;
> >> 710 #endif
> >> 711         TUNABLE_INT_FETCH("vm.kmem_size_scale",
&vm_kmem_size_scale);
> >> 712         if (vm_kmem_size_scale > 0 &&
> >> 713             (mem_size / vm_kmem_size_scale) > (vm_kmem_size /
> >> PAGE_SIZE))
> >> 714                 vm_kmem_size = (mem_size / vm_kmem_size_scale) *
> >> PAGE_SIZE;
> >>
> >> here,
> >> VM_KMEM_SIZE = 12*1024*1024
> >> nmbclusters = 0 (initially)
> >> PAGE_SIZE = 4096
> >> # sysctl -a | grep v_page_count
> >> vm.stats.vm.v_page_count: 7035
> >>
> >> and VM_KMEM_SIZE_SCALE = 3 for mips.
> >>
> >> So, vm_kmem_size = 12mb.
> >>
> >> Going back to tunable_mbinit(),
> >> we get realmem = 12mb.
> >> and masmbufmem = 6mb.
> >>
> >>
> >> Wanted to see if I am following the code correctly and how autotuning
> >> should work here.
> >>
> >> cheers,
> >> Hiren
> >> _______________________________________________
> >> freebsd-net at freebsd.org mailing list
> >> http://lists.freebsd.org/mailman/listinfo/freebsd-net
> >> To unsubscribe, send any mail to "freebsd-net-unsubscribe at freebsd.org"
> >>
> > _______________________________________________
> > freebsd-mips at freebsd.org mailing list
> > http://lists.freebsd.org/mailman/listinfo/freebsd-mips
> > To unsubscribe, send any mail to "freebsd-mips-unsubscribe at freebsd.org"
>