home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!dtix!darwin.sura.net!haven.umd.edu!mimsy!afterlife!adm!news
- From: HERBER@fnalf.fnal.gov (Randolph J. Herber, CD/DCD/USS, x2966)
- Newsgroups: comp.sys.sgi
- Subject: Re: modem and background jobs
- Message-ID: <32243@adm.brl.mil>
- Date: 27 Aug 92 18:47:32 GMT
- Sender: news@adm.brl.mil
- Lines: 39
-
- +From: "Richard E. Gillilan" <reg@sdchemw2.ucsd.edu>
- +Newsgroups: comp.sys.sgi
- +Subject: modem and background jobs
- +Date: 10 Aug 92 20:13:27 GMT
-
- +Normally, when users log off of a modem line, the SGI hangs up the
- +phone. If, however, anyone has submitted a background job from that
- +line, our SGI will not terminate the call. Does anyone know what the
- +problem is here?
-
- +Thanks
-
- +Richard E. Gillilan
-
- Dave Olson of SGI, please correct me if I make any errors.
-
- This is a FAQ and a RTFM item. But, I must admit that it is hard-to-find
- RTFM item!
-
- 1. a regular file or special device file is not closed in the kernel until
- all open fd's for the file are closed either by explicit close()s or
- by implicit closes as a result of process termination.
-
- 2. an exec() replicates open fd's (except for those marked close-on-exec)
- so that both parent and child have open fd's. After the fork, the open
- status of the fd's are independent.
-
- 3. as long as any process has a open fd for the process, including control tty,
- the tty driver will not drop DTR to the modem. It is dropping of DTR that
- normally signals the disconnection of the telephone connection.
-
- 4. I believe that:
- nohup %cmd% </dev/null >& report.file &
- will result in the proper conditions for a background job that will release
- the modem by dropping DTR.
-
- Randolph J. Herber, herber@fnalf.fnal.gov, +1 708 840 2966
- (Speaking for myself and not for US, US DOE, FNAL nor URA.)
- (Product, trade, or service marks herein belong to their respective owners.)
-