Re: POSIX.1 semaphores vs message queues

[Masao Uebayashi <uebayasi%gmail.com@localhost>]

On Sat, Nov 14, 2015 at 7:35 AM, Paul Goyette <paul%vps1.whooppee.com@localhost> wrote:
> On Fri, 13 Nov 2015, David Holland wrote:
>
>> On Fri, Nov 13, 2015 at 04:41:18PM +0000, David Holland wrote:
>> > On Fri, Nov 13, 2015 at 08:05:25PM +0800, Paul Goyette wrote:
>> >  > One final attempt to summarize the objections that have been made:
>> >  > [snip]
>> >
>> > One other thing: posix semaphores used to be a module. That code was
>> > made the victim^W showpiece for demonstrating how the New World Order
>> > was going to be. When the New World Order turned out to be broken,
>> > that and other things were rolled back after a lengthy political
>> > squabble.
>> >
>> > Part of the resistance you're seeing is the legacy of that squabble,
>> > even though people probably don't really want to say so.
>>
>> One other point: the changes add mechanism, and that mechanism has a
>> cost (in maintenance if nothing else) -- if the changes are more or
>> less pointless as a number of people think, then we'd be paying that
>> cost for zero or negligible benefit. That does seem like a valid
>> reason to oppose such changes.
>
>
> The mechanisms are currently still included.  When the ksem module was
> removed, the only thing that was done was removal of the modular variant
> from the build.  All of the "module mechanisms" still remain.  The
> current ksem_modcmd() even continues to use syscall_establish() to
> "install" the syscalls, and those syscalls are still listed in the table
> of autoload-able syscalls (in sys/kern/syscalls_autoload.c) and are
> marked as such in syscalls.master.

I wonder if syscall_disestablish() can track syscall use
(l->l_sysent), when it return back to userland by tsleep(PCATCH), then
another syscall is called and l->l_sysent is overwritten?

>
>
>> for my part during the above-mentioned squabble I made it pretty clear
>> that I think aggressive use of kernel modules is at best tilting at
>> windmills; I stand by that position, but I'm not really interested in
>> fighting about it at the moment so I won't stick my oar in any
>> further.
>
>
> :)
>
>
>
> +------------------+--------------------------+-------------------------+
> | Paul Goyette     | PGP Key fingerprint:     | E-mail addresses:       |
> | (Retired)        | FA29 0E3B 35AF E8AE 6651 | paul at whooppee.com    |
> | Kernel Developer | 0786 F758 55DE 53BA 7731 | pgoyette at netbsd.org  |
> +------------------+--------------------------+-------------------------+