Ah yes. -i That's the ticket. Refresh the brain cache. Glad you figured it out. As I said if you run a cmp between that binary and one of the ones available in the wild I think you will find them to be the same.
Clem Sent from my iPhone > On Dec 10, 2015, at 1:10 AM, Will Senn <[email protected]> wrote: > > > >> On 12/9/15 10:01 PM, Clem cole wrote: >> It should not be a 411 code. It should be a standard 407. 11/34 and 11/60s >> are so called 40 class systems. I used v6tar with them just fine. >> >> Sent from my iPhone >> >>> On Dec 9, 2015, at 10:40 PM, Will Senn <[email protected]> wrote: >>> >>> use its v6tar on v6. It appears to require split I+D, > Clem, > > Thanks for persisting. I figured it out - on Keith Bostic's v7, using the > default makefile for v6tar: > make v6tar > cc -O -c tar.c > cc -c -O /usr/src/libc/v6/access.c > cc -c -O /usr/src/libc/v6/chown.c > cc -c -O /usr/src/libc/v6/execl.c > cc -c -O /usr/src/libc/v6/ftime.c > cc -c -O /usr/src/libc/v6/gtty.c > cc -c -O /usr/src/libc/v6/lseek.c > cc -c -O /usr/src/libc/v6/stat.c > cc -c -O /usr/src/libc/v6/syscall.s > cc -c -O /usr/src/libc/v6/time.s > cc -i -s -O *.o -o v6tar > # dd if=v6tar bs=128 count=1|od > 0000000 000411 037100 004224 027466 000000 000000 000000 000001 > ... > man cc says -i and -s are passed through to linker, so man ld says -i creates > separate I+D and -s strips symbol table. I compiled the binary without these > (I thought I would keep the symbol table as an excercise) using: > cc -O *.o -o v6tar > > The resulting binary is 407, yay!: > dd if=v6tar bs=128 count=1|od > 0000000 000407 037040 004224 027466 020744 000000 000000 000001 > ... > > I copied it over to v6, lo and behold, it works on a PDP11/40 without > separate I+D! > > Thanks, > > Will > _______________________________________________ Simh mailing list [email protected] http://mailman.trailing-edge.com/mailman/listinfo/simh
