linux-kernel - Re: Subject: [PATCH 2/2] priority System V Semaphores

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <1324562457.24803.24.camel@twins>
Date:	Thu, 22 Dec 2011 15:00:57 +0100
From:	Peter Zijlstra <peterz@...radead.org>
To:	Raz <raziebe@...il.com>
Cc:	Manfred Spraul <manfred@...orfullife.com>,
	linux-rt-users@...r.kernel.org, linux-kernel@...r.kernel.org,
	Lior Brafman <LBrafman@...z.com>,
	Torsten Scherer <TScherer@...z.com>,
	Rasty Slutsker <RSlutsker@...z.com>
Subject: Re: Subject: [PATCH 2/2] priority System V Semaphores

On Thu, 2011-12-22 at 15:00 +0200, Raz wrote:

> please correct me if am wrong, " posix semaphores
> are implemented with pi mutex. ..?" 

Probably not, that would be pointless. They might be implemented using a
mutex, but I've already looked at glibc this month and my eyes can't
handle more.

> I need a counting semaphore.
> vxWorks priority/fifo semaphores are different from posix semaphores in
> that the behaviour is defined on the semaphore and not the thread.

That doesn't make them sane locking primitives.

> Q: what happens if I want one posix semahore to be FIFO and another
> posix semaphore to be PRIO while both are used by the same
> thread.should i to change policies each time ?

If you need a counting semaphore your program is very likely not a
correct real-time application. So who cares what order things are woken
up in.

Semaphores don't have lock owners, therefore priority inheritance and
related schemes don't work, therefore you suffer from priority inversion
and thus your program is invalid.

Seriously, forget semaphores exist, they're a hysterical accident.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/