> actually...while i know the strict "definition" of pipes is that they
> are unidirectional with a single read and a single write endpoint, in
> my (limited?) experience, i've always found pipes to be bidirectional.

> including ours.

When did you last test?  Looking at sys_pipe(), I note that it opens
one end FREAD-only and the other end FWRITE-only.

And I did a test just now, and find that the pipe is in fact
unidirectional.  Note in particular the marked kdump lines.

	% cat pipetest.c
	#include <unistd.h>
	#include <stdlib.h>
	
	int main(void);
	int main(void)
	{
	 int p[2];
	 char buf[32];
	
	 pipe(p);
	 write(p[0],"123",3);
	 write(p[1],"123",3);
	 read(p[0],&buf[0],32);
	 read(p[1],&buf[0],32);
	 exit(0);
	}
	% cc -o pipetest pipetest.c
	% ktrace pipetest
	% kdump | tail -20
	 25302 pipetest RET   mmap 269307904/0x100d5000
	 25302 pipetest CALL  munmap(0x100c0000,0xf000)
	 25302 pipetest RET   munmap 0
	 25302 pipetest CALL  close(0x3)
	 25302 pipetest RET   close 0
	 25302 pipetest CALL  pipe
	 25302 pipetest RET   pipe 3
	 25302 pipetest CALL  write(0x3,0x10aa0,0x3)
>>>	 25302 pipetest RET   write -1 errno 9 Bad file descriptor
	 25302 pipetest CALL  write(0x4,0x10aa0,0x3)
	 25302 pipetest GIO   fd 4 wrote 3 bytes
	       "123"
	 25302 pipetest RET   write 3
	 25302 pipetest CALL  read(0x3,0xeffff7f0,0x20)
	 25302 pipetest GIO   fd 3 read 3 bytes
	       "123"
	 25302 pipetest RET   read 3
	 25302 pipetest CALL  read(0x4,0xeffff7f0,0x20)
>>>	 25302 pipetest RET   read -1 errno 9 Bad file descriptor
	 25302 pipetest CALL  exit(0)

This test was done under 1.4T.

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B