buildworld failure in /usr/src/sys/boot/i386/boot2 this morning

[Bruce Evans]

On Mon, 16 Feb 2004, Fritz Heinrichmeyer wrote:

> as  --defsym FLAGS=0x80  --defsym SIOPRT=0x3f8  --defsym SIOFMT=0x3  --defsym
> SIOSPD=9600 -o boot2.o boot2.s
> cc -Os  -fno-guess-branch-probability -fomit-frame-pointer
> -mno-align-long-strings  -mrtd  -DUFS1_AND_UFS2
> -I/usr/src/sys/boot/i386/boot2/../../common
> -I/usr/src/sys/boot/i386/boot2/../btx/lib -I.  -Wall -Waggregate-return
> -Wbad-function-cast -Wcast-align  -Wmissing-declarations -Wmissing-prototypes
> -Wnested-externs  -Wpointer-arith -Wshadow -Wstrict-prototypes
> -Wwrite-strings -ffreestanding -mpreferred-stack-boundary=2
> -c /usr/src/sys/boot/i386/boot2/sio.S
> /var/tmp//ccrYtMTN.s: Assembler messages:
> /var/tmp//ccrYtMTN.s:84: Error: undefined symbol `SIOSPD' in operation setting
> `SIO_DIV'
> /var/tmp//ccrYtMTN.s:35: Error: undefined symbol `SIO_FMT' in operation
> /var/tmp//ccrYtMTN.s:39: Error: value of 115200 too large for field of 2 bytes
> at 13
>
> cc also should be informed about the defined symbols ....

cc shouldn't be run to build sio.o.

This problem seems to be caused by a recent repo-copy of sio.S from sio.s.
sio.s works, but sio.S is preferred although it has not been connected to
the build.

boot1.S was repo-copied not so recently from boot1.s, but this doesn't
caused problems because there is an explicit rule for building boot1.o
from boot1.s, so boot1.S is not preferred.

Bruce