bad system call on HP-UX 10.20

David DeSimone fox at
Fri Mar 24 11:17:01 CET 2000

Ralf Hildebrandt <R.Hildebrandt at> wrote:
> Well, when building as unprivileged user and testing as such, it works.

That's because, I think, GPG knows that it needs privileges in order to
perform the memory-lock operation, so it doesn't even attempt to do so
when run as a normal user.

> At least I don't get "bad system call" any more.  Nevertheless a
> program should never crash, but instead throw out an error message.

The problem is that for some reason, libc implements the system-call
callout, but the OS does not support the system-call.  The standard
method for dealing with a bad system-call is to send the process a
SIGSYS signal, which is what's happening.  The process basically gets
killed; the "bad system call" message comes from your shell, reporting
the method that caused the process to die.

SIGSYS is not one of the signals that GnuPG attempts to catch or handle.

David DeSimone   | "The doctrine of human equality reposes on this:
fox at   |  that there is no man really clever who has not
Hewlett-Packard  |  found that he is stupid." -- Gilbert K. Chesterson
Richardson IT    |    PGP: 5B 47 34 9F 3B 9A B0 0D  AB A6 15 F1 BB BE 8C 44

More information about the Gnupg-devel mailing list