Hrvoje Popovski wrote in to alert us that Martin Pieuchot (mpi@) has written a new blog post entitled Faster vlan(4) forwarding?, which leads in with
Two years ago we observed that vlan(4) performances
suffered from the locks added to the queueing API.
At that time, the use of SRP was also
pointed out as a possible responsible for the regression. Since
dlg@recentlyreworkedif_enqueue() to allow pseudo-drivers to bypass the use of
queues, and their associated locks, let's dive into vlan(4) performances
Mike Larkin (mlarkin@) just committed support for 2TB of physical memory on the amd64 platform:
Module name: src
Changes by: firstname.lastname@example.org 2019/01/20 23:18:37
sys/arch/amd64/amd64: machdep.c pmap.c
Support 2TB phys mem
This change expands the direct map to 4 slots (512GB each), to support
machines with up to 2TB physical memory. Should further expansion be
required, this change provides the means to do that with a single #define
with help from and ok guenther
Those of you with machines with lots of memory are encouraged to try this out and report your dmesg!
from the virtually not a vax dept.
by Pratik Vyas (pd@),
vmm(4) support for i386 host systems has been deleted (one can still run i386 guests under vmm on an amd64 host).
The commit messages explain the reasoning behind this move: