From owner-freebsd-stable@freebsd.org  Mon Aug  3 12:08:14 2015
Return-Path: <owner-freebsd-stable@freebsd.org>
Delivered-To: freebsd-stable@mailman.ysv.freebsd.org
Received: from mx1.freebsd.org (mx1.freebsd.org
 [IPv6:2001:1900:2254:206a::19:1])
 by mailman.ysv.freebsd.org (Postfix) with ESMTP id 2464E9B186A
 for <freebsd-stable@mailman.ysv.freebsd.org>;
 Mon,  3 Aug 2015 12:08:14 +0000 (UTC) (envelope-from ortadur@web.de)
Received: from mout.web.de (mout.web.de [212.227.17.12])
 (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits))
 (Client CN "mout.web.de", Issuer "TeleSec ServerPass DE-1" (verified OK))
 by mx1.freebsd.org (Postfix) with ESMTPS id 8EA441E95
 for <freebsd-stable@freebsd.org>; Mon,  3 Aug 2015 12:08:13 +0000 (UTC)
 (envelope-from ortadur@web.de)
Received: from [131.169.71.113] by 3capp-webde-bs15.server.lan (via HTTP);
 Mon, 3 Aug 2015 14:08:04 +0200
MIME-Version: 1.0
Message-ID: <trinity-4c7d55d5-dd1f-4f24-85d4-a05dff012c0e-1438603684866@3capp-webde-bs15>
From: "Andre Meiser" <ortadur@web.de>
To: freebsd-stable@freebsd.org
Subject: Re: Many core dumps in pthread_getspecific.
Content-Type: text/plain; charset=UTF-8
Date: Mon, 3 Aug 2015 14:08:04 +0200
Importance: normal
Sensitivity: Normal
In-Reply-To: <CAJ-VmomBKUzZ_wF37EG5VTUXU601-eFSVQzrWyaU2vkT8u+RWA@mail.gmail.com>
References: <20150603145838.GX2499@kib.kiev.ua>
 <trinity-15fcacbd-871c-4ea8-9257-5d11e7862ec0-1434103396559@3capp-webde-bs41>
 <20150614190504.GT2080@kib.kiev.ua>
 <trinity-e44527ae-e511-4ff3-bcdf-ee8426fc8a94-1434438565708@3capp-webde-bs53>
 <20150616073637.GO2080@kib.kiev.ua>
 <trinity-9d219acd-7aa9-4574-a9ad-458b52374069-1435936910016@3capp-webde-bs27>
 <20150703211111.GZ2080@kib.kiev.ua>
 <trinity-54f5f349-442e-4777-b35b-866867b763dc-1436977933500@3capp-webde-bs41>
 <20150719205722.GT2404@kib.kiev.ua>
 <trinity-a61eab47-c174-48a6-9991-1ffac1069609-1437558395543@3capp-webde-bs25>
 <20150722102055.GA2072@kib.kiev.ua>
 <trinity-9a9b5735-93e5-4593-95ee-73e09bf46daf-1437998577258@3capp-webde-bs52>, 
 <CAJ-VmomBKUzZ_wF37EG5VTUXU601-eFSVQzrWyaU2vkT8u+RWA@mail.gmail.com>
X-UI-Message-Type: mail
X-Priority: 3
X-Provags-ID: V03:K0:+rvCF1383jq+v44OlsJ0xJwyFK0lUdrNcS5Wr5CPRsi
 7H/iZ2Ha1WDXy62MoWw8dezooHLFDHGT3BeLF8aFaLF7HYcou6
 rSC0QG5Lu3ix7bWsn5n0UZCR6ldw8XNPy8uKwLfg2G0mTNhXbX
 3x4F2ILnKFdmDTG7aOUJBGpIqN15hckEBqVNILR2eY7U3XMgCK
 2+kLq+uyJPflIbu2FTt7ns7Nv9B0AyTC2MjjjojHzB6ByCEh9c
 5kYuyaYmkFUKUxUzQuMljhrd8WRpFpeJje0FToADDiPudZkWjM wcwIXE=
X-UI-Out-Filterresults: notjunk:1;V01:K0:3g+XU3BuLzg=:Zc3yWVjPPkc/4UQfi78zJl
 7PATQqD700yXowC45ayQO/2fuv1C1BmXO1PdNANzdquPEa/zNZ/x9sCmsB0OGDwHUKyYVlmvN
 gzCIreHUmXItmqyWQ0pvmvXPckzLXUMDKwCXWfpJuLD0jsTEqqQAshPQrL77RFgHUDkfWLiYb
 JbbP/XTvcc8CvMc7Oz/b3oAGK1txlbxU5jDzRaUJY069dowpYkGysOAnt1eTuUuEdCnrRS8IE
 b/8QkRqQtPnD1Er5lGM7LrePKu0PDpUMeod1JE4X7NQ8Mg2vqCRytD7iMmoEaEUu7U+Q11POJ
 WlSGLJuGEWP3uRP+OX7Ag2zzjs7yU0HDkLuDgelKoAXSDioatPgbwCMnwt+JpeiwiRlcw61Ht
 F/WaJzuWPmVne9BkBPq2XESQOQt7cHINWy8yTeyZEbfGyandvPNSl2PW2Q/kuhg61huUcQ5Oq
 HqWNVNIUXg==
X-BeenThere: freebsd-stable@freebsd.org
X-Mailman-Version: 2.1.20
Precedence: list
List-Id: Production branch of FreeBSD source code <freebsd-stable.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/options/freebsd-stable>,
 <mailto:freebsd-stable-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-stable/>
List-Post: <mailto:freebsd-stable@freebsd.org>
List-Help: <mailto:freebsd-stable-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-stable>,
 <mailto:freebsd-stable-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Mon, 03 Aug 2015 12:08:14 -0000

Hi,

On 27 July 2015 at 22:41, Adrian Chadd wrote:
> What was that option actually doing? How'd it break things?

old operating systems like Windows 98 can't handle a CPUID higher than 3.  But Pentium 4 with HTT (Hyper-Threading Technologie) has a higher value.  So the BIOS can limit it to 3.

The default value from the BIOS is set to limit the CPUID.  And this caused troubles with FreeBSD which used for all but the first CPU a far bigger data structure (ucontext_t) for threads.  So sometimes the stack didn't had enough space for this big structure and made a segmentation fault (see subject -> many core dumps).

This become fixed by Konstantin Belousov today.  So the next FreeBSD release won't be affected by this BIOS setting on P4 with HTT.

Sincerely yours Andre.