Date: Thu, 11 Feb 1999 02:16:12 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: dillon@apollo.backplane.com (Matthew Dillon) Cc: tlambert@primenet.com, kuku@gilberto.physik.RWTH-Aachen.DE, hackers@FreeBSD.ORG Subject: Re: portability of shm, mmap, pipes and socket IPC Message-ID: <199902110216.TAA08700@usr06.primenet.com> In-Reply-To: <199902110114.RAA87339@apollo.backplane.com> from "Matthew Dillon" at Feb 10, 99 05:14:58 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> :FreeBSD must maintain and correctly apply at process exit the value of
> :sem_adj. See the referenced Solaris man pages for semantic details.
>
> In other words, you have no idea whether there's a bug or not. This
> is more second hand ( or worse ) information.
No.
It is a bug.
SVR4, per SVID III, tracks sem_adj. Those were just the man pages I
had handy that you could read.
> Look, I don't have a problem with second hand information - but
> I do have a problem with it being put out as solid fact when it
> isn't. It's one thing to say that there might be a bug, quite
> another to start crying wolf over unproven and unresearched
> reports. It doesn't do anyone any good to condemn a system
> for having a 'bug' when you can't even describe what the bug is!
You must be referring th the INN stuff here, since the sem_adj failure
when a process core dumps *is* a bug, right? See kern/sysv_sem.c,
~line 689:
if (sopptr->sem_op < 0) {
if (semptr->semval + sopptr->sem_op < 0) {
#ifdef SEM_DEBUG
printf("semop: can't do it now\n");
#endif
break;
} else {
And from the man page:
sem_op is a negative integer; {ALTER}
. If semval is less than the absolute value of sem_op and
(sem_flg&IPC_NOWAIT) is false, semop() increments the
semncnt associated with the specified semaphore and
suspends execution of the calling process until one of
the following conditions occur:
. semval becomes greater than or equal to the absolute
value of sem_op. When this occurs, the value of
semncnt associated with the specified semaphore is
decremented, the absolute value of sem_op is sub
tracted from semval and, if (sem_flg&SEM_UNDO) is
true, the absolute value of sem_op is added to the
calling process's semadj value for the specified
semaphore.
. The semid for which the calling process is awaiting
action is removed from the system (see semctl(2)).
When this occurs, errno is set equal to EIDRM, and a
value of -1 is returned.
. The calling process receives a signal that is to be
caught. When this occurs, the value of semncnt
associated with the specified semaphore is decre
mented, and the calling process resumes execution in
the manner prescribed in signal(3C).
i.e.: "semop: can't do it now" doesn't cut it.
Terry Lambert
terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199902110216.TAA08700>