gonzalez@vax.bbn.com (Jim Gonzalez) (09/12/87)
I have started working with an HPSM that we had installed several weeks ago. The only test performed on it was successful login on each of the eight lines using a VT-220. In addition, I succeeded in completing a Kermit exchange over one of the lines, the week before last. Tonight I can't get at the lines. I've tried cat, the connect option of kermit and a short program that simply does an open on /dev/tty15. In the last case, I know that it goes into the open and does not return. Kermit isn't printing it's little connect message. On the one line that has getty enabled on it (/dev/tty8 == port 0) I can do cat and connect with Kermit, and send characters out (they appear on the 220). The lines (/dev/tty8 ... /dev/tty15) are all mode 622, owned by root. My test runs are performed as root (when attempted as non-root I get EPERM on the open, as expected). The HPSM seems to be working, since I can still login on the getty line, but I can't get at it. The wording for tty(4) seems to imply that my process is simply waiting for a response that never comes from the I/O systmem. Has anyone had this problem? My three regular lines are behaving normally, but of course they don't use the same device and probably don't use the same driver. Thanks in advance for any assistance. -Jim Gonzalez gonzalez@bbn.com BBN Labs, Inc. Cambridge, MA [Did you reload the HPSM software into the HPSM? What version of the HPSM software are you runnings? I think the latest is 3.0. -- sob]
gonzalez@vax.bbn.com (09/15/87)
Peg Bailey in Customer Service gave me the answer I needed. The HPSM is much more picky about control lines than most RS-232 implementations I've seen. Once DSR and DCD were asserted, I was able to open the port and I was able to do ioctl's once CTS was asserted. Apparently, Masscomp still doesn't have a formal description of the HPSM for customers to use, so this problem pops up on the hotline all the time. By the way, is there a list of special interest groups around? The MUS application asks what groups one might be interested in, but doesn't list which groups exist. Our group does ship-based applications, so I have an interest in groups covering graphics, data acquisition, etc. Also, I'd be interested in comparing notes on field operations (packing, power line conditioning, salt spray, ship motion, etc). -Jim Gonzalez gonzalez@bbn.com BBN Labs, Inc. Cambridge, MA
masscomp-request@soma.UUCP (09/17/87)
In article <3202@soma.bcm.tmc.edu>, gonzalez@vax.bbn.com (Jim Gonzalez) writes: > I have started working with an HPSM that we had installed several weeks ago. > ............ and Stan suggests: > [Did you reload the HPSM software into the HPSM? What version of the HPSM > software are you runnings? I think the latest is 3.0. -- sob] I just (last week ?) got HPSM software release 4.0, but have not installed it yet. I am hoping that it will fix some of the problems we have been having. The RELEASE NOTES claim to fix: -A flow control problem ... (which) would hang a tty port. -confusion resulting from NULLs in the data stream. and a collection of other errors in the previous releases We have had a number of problems with HPSM lines. The basic symptom is that when you kill one (however that happens) the only way to recover that line is to reboot. It may be possible to reload the HPSM, but you would probably have to do it in single user or at least assure that no one else is using it. The most common type of HPSM failure occurs when we transfer very long files through it. We run a large format digitizer with a Mac+, then move the resulting files from the Mac to a Masscomp with Kermit. I don't think that Kermit is at fault. It appears that the HPSM loses an interupt, and then can not recover. I probably should have spent this time installing the new release..... Dale Chayes Lamont-Doherty Geological Observatory Palisades, N.Y. 10964 (914) 359-2900 ....!philabs!lamont!dale