the block comment at the top of the file:
This module provides kernel support for testing network
throughput from the perspective of the kernel. It is
similar in spirit to the classic ttcp network benchmark
program, the main difference being that with kttcp, the
kernel is the source and sink of the data.
Testing like this is useful for a few reasons:
1. This allows us to know what kind of performance we can
expect from network applications that run in the kernel
space, such as the NFS server or the NFS client. These
applications don't have to move the data to/from userspace,
and so benchmark programs which run in userspace don't
give us an accurate model.
2. Since data received is just thrown away, the receiver
is very fast. This can provide better exercise for the
sender at the other end.
3. Since the NetBSD kernel currently uses a run-to-completion
scheduling model, kttcp provides a benchmark model where
preemption of the benchmark program is not an issue.
There is a companion "kttcp" user program which uses the kttcp
pseudo-device.
Largely written by Frank van der Linden, with some modifications
from me.
-treat the builtin font like any other font at runtime
-for that, copy it to malloc()'d memory during attach()
-in early console initialization, if we have to consider a broken card
(VGA_CONSOLE_ATI_BROKEN_FONTSEL), copy the builtin font to another
location in font ram; the attach() code will do the rest
put the "quirk" code into effect again
sidestep it using troff macro trick from FreeBSD version of this manpage.
xref psignal(3) and strsignal(3)
This fixes lib/12260 by Ada Lim and lib/17406 by Greg A. Woods.
bit and then read it back in a loop (with appropriate delays) waiting
for it to read back clear.
This fixes a problem where the bus would hang when bringing down
the interface or changing interface flags on a system with a
sufficiently fast CPU (e.g. 2GHz P4 Xeon).
mappings bus_dma(9) states: "In the event that the DMA handle contains
a valid mapping, the mapping will be unloaded via the same mechanism
used by bus_dmamap_unload()." And some drivers do mean to skip the
unload step.
isn't addressed with a negative offset when back at the top of the tree.
This caused pax -M on sparc64 to generate corrupt tar files.
Problem found by Tim Goodwin <tjg@star.le.ac.uk> in [bin/17412].
- if SRC_DIR or DEST_DIR == /, set to "" so that various pathnames
don't end up with a leading "//". when specifically printing just
the xxx_DIR variable, use ${xxx_DIR:-/} to DTRT
- cosmetic coding tweaks
(generally it's 20k). Adjust mblklen temporarily to the size of the block
required and allocate one. This avoids coredumps when mapping in identifiers
that have huge values. (In my example it was a char[] for a 640k pixmap).
rafal