home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.vms
- Path: sparky!uunet!eco.twg.com!shiva.eco.twg.com!larry
- From: larry@shiva.eco.twg.com (Lawrence B. Henry III)
- Subject: Re: QIO failing with ACCVIO (WIN-TCP)
- Message-ID: <1992Nov5.231903.10500@eco.twg.com>
- Lines: 65
- Sender: news@eco.twg.com (News)
- Nntp-Posting-Host: eco.twg.com
- Reply-To: larry@eco.twg.com
- Organization: The Wollongong Group (East Coast Operations)
- References: <9211042022.AA20329@relay2.UU.NET>
- Date: Thu, 5 Nov 92 23:19:03 GMT
- Lines: 65
-
-
- In article <9211042022.AA20329@relay2.UU.NET>, ddj@gradient.UUCP writes:
- |>Path: eco.twg.com!psinntp!sgigate!sgiblab!zaphod.mps.ohio-state.edu!uwm.edu!linac!att!ucbvax!gradient.UUCP!ddj
- |>From: ddj@gradient.UUCP
- |>Newsgroups: comp.os.vms
- |>Subject: QIO failing with ACCVIO (WIN-TCP)
- |>Message-ID: <9211042022.AA20329@relay2.UU.NET>
- |>Date: 4 Nov 92 20:22:21 GMT
- |>Sender: daemon@ucbvax.BERKELEY.EDU
- |>Distribution: world
- |>Organization: The Internet
- |>Lines: 43
- |>
- |>I have a network application that runs under VMS, and can use CMU,
- |>Fusion, UCX, and DECnet transports. After a "quick and easy" port
- |>to TWG's WIN-TCP package, I started to get errors from the
- |>SYS$QIO(... IO$_READVBLK...) call.
- |>
- |>The call comes back immediately with a status of ACCVIO. This does
- |>not happen right away -- the program succeeds in opening and exchanging
- |>packets over two or three network ports to various servers, but when
- |>it reaches the third (or fourth) port it fails.
- |>
- |>The worst part is that the problem does not occur when the program
- |>is run from VMS DEBUG. There is only one place in the program where
- |>this call is made, and the arguments all look valid (and in fact the
- |>call works properly for a couple dozen packets before the fourth
- |>socket is opened).
- |>
- |>The program uses UDP packets (except for the decnet transport),
- |>and is run as a detached process, with a bytlim of about 32k and
- |>a file limit of about 30. It is normally started at boot time,
- |>with system privs. However it still fails when I run it from
- |>a terminal.
- |>
- |>1) What could be different between run/nodebug (or not compiled
- |>for debugging), and running under the debugger?
- |>
- |>2) Is there any way to print the failed address, etc, when the
- |>access violation is passed in the system call return value (as
- |>opposed to getting the fault directly)?
- |>
- |>3) Has anyone else run into anything like this problem before?
- |>What can be done to avoid it? Is there a limit in the TWG network
- |>ACP that would prevent a process from creating more than two or
- |>three sockets and starting receives on them all?
- |>
- |>4) Is there a tool that can watch the quotas of a program while
- |>it is running?
- |>
- |>Thanks in advance!
- |>
- |> Dave Johnson
- |> ddj@gradient.com (or uunet!gradient!ddj)
- |>
- |>
- |>
- Dave,
- Why don't you post this question to our news group
- vmsnet.networks.tcp-ip.wintcp or send me mail directly.. we will be more than
- happy to address any questions you might have concerning our QIO interface
- or socket library..
-
- -Larry.
- The Wollongong Group.
-