mmp (05/16/82)
When bringing up a new 4.1bsd system, we tried to swap on /dev/hp3b and
swapon complained that the device didn't exist. There is a table in
swaphphp.c that contains /dev/hp0b and /dev/hp1b and apparently restricts
swapping to those two devices. We added hp2b and hp3b assuming that
there would be no swapping done unless the device is specified in fstab.
(Incidentally, zeppo is now happily swapping on hp3b...)
A closer look at swaphphp.c and swaphp.c revealed that the only real difference
between the two is the table containing the swap devices -- swaphp.c has
only one entry in the table (hp0b). The question is: Why are there two
different files? Configuring the system with the hphp option but only having
one swap area specified in fstab seems to be equivalent to configuring with
the hp option.
I couldn't find much on this subject in the either the user manual or
the installation/operation document.
It seems that you should be able to specify swap areas in the system
configuration file and have the table in swaphphp.c be built dynamically (like
everything else in the system!).
Comments?
Michelle Peetz
BTL Whippany
harpo!mmp
[ two less significant problems:
1) When recompiling the system lately, newvers.sh puts a newline
in the version string at column 80. Cc doesn't like this.
We're up to version .123. Has someone encountered this?
2) We seem to have both .../sys/sys/sys4.c and .../sys/dev/sys4.c.
They differ in one line. Is this our goof or is there a reason
for having two versions?
-ED ]