I've created a mirror of the files at http://www.cran.org.uk/~brucec/freebsd/xen/ .
Could I copy the instructions to other formats such as the Wiki and possibly the Handbook? The current instructions have a broken link so I'd like to replace them.
There's a 'doxygen' target in /usr/src/Makefile which suggests that FreeBSD does use it:
# doxygen - Build API documentation of the kernel, needs doxygen.
The trick is to overwrite the very start and end of the partition, since GPT stores a backup table at the end too. That's why FreeBSD will still think the disk is partitioned using GPT if you just overwrite the first few sectors.
The FreeBSD SCTP code comes from sctpCVS; it has lots of applications to verify that it's working as expected before being committed to the FreeBSD repository.
I think it's possible to build a kernel which doesn't have any IP support, but I don't think you can currently build an INET6-only kernel. At least when I tried to build one on 9-CURRENT a few days ago I got a compile error due to an unused label in one C file, and once that was fixed I got...
Have you tried disabling/enabling offloading on the NIC? See the rxcsum, txcsum, tso and lro options in ifconfig(8). I think these can have an impact on performance.
The named.conf gets copied from /etc/named to /var/named/etc/namedb so bind can access it, so if you make /var/named writable then an attacker, despite not being able to escape from the sandbox, could still rewrite named.conf.
The only option is to see if you can enable APM instead or upgrade the BIOS so you can re-enable ACPI. If the BIOS doesn't configure the power button to route the button event to FreeBSD there's nothing it can do to prevent the power being lost.
Unfortunately that option was broken for a long time and has only recently been fixed in -CURRENT and -STABLE - until it was fixed there was always a single DNS lookup that couldn't be avoided. For details, see http://lists.freebsd.org/pipermail/svn-src-head/2008-October/000270.html
In this case the power button is being reconfigured to switch the power supply off without sending a signal to the OS. However, see shutdown(8) and the -n option for how to tell FreeBSD to do a real unclean shutdown :)
This isn't technically a bug, but it has been worked around in STABLE and will be in 7.1. See PR misc/128861 for details - http://www.freebsd.org/cgi/query-pr.cgi?pr=128862&cat=
There are people working on ACPI sleep modes for amd64 on -CURRENT. See http://lists.freebsd.org/pipermail/freebsd-acpi/2008-October/005159.html for details.
There seem to be several problems, and while one or two have been fixed recently, there are still others lurking to bring the system down. One fix was made earlier this year, and with the new USB stack (aka 'usb2') another class of problems should be fixed. However I think there are still...
I'm not sure what the Shift-lock key is (caps-lock?) but there's a known problem with certain key combinations: it's documented at the bottom of http://wiki.freebsd.org/JeremyChadwick/Commonly_reported_issues - the best information I could find about it is in this thread...
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.