From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751543AbbJLNx3 (ORCPT ); Mon, 12 Oct 2015 09:53:29 -0400 Received: from eu-smtp-delivery-143.mimecast.com ([207.82.80.143]:61503 "EHLO eu-smtp-delivery-143.mimecast.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751608AbbJLNx1 convert rfc822-to-8bit (ORCPT ); Mon, 12 Oct 2015 09:53:27 -0400 Date: Mon, 12 Oct 2015 14:53:13 +0100 From: Dave Martin To: Arnd Bergmann Cc: linux-arm-kernel@lists.infradead.org, catalin.marinas@arm.com, will.deacon@arm.com, linux-kernel@vger.kernel.org, Akhilesh Kumar , Rohit Thapliyal , Andreas Schwab , Manjeet Pawar , pankaj.m@samsung.com Subject: Re: [PATCHv2] ARM64:Fix MINSIGSTKSZ and SIGSTKSZ Message-ID: <20151012135300.GA26786@e103592.cambridge.arm.com> References: <1444109743-8561-1-git-send-email-manjeet.p@samsung.com> <20151006103128.GN6281@e103592.cambridge.arm.com> <5340366.WTVksOGb0G@wuerfel> <5223992.qv9q1k7gFp@wuerfel> MIME-Version: 1.0 In-Reply-To: <5223992.qv9q1k7gFp@wuerfel> User-Agent: Mutt/1.5.21 (2010-09-15) X-OriginalArrivalTime: 12 Oct 2015 13:53:19.0952 (UTC) FILETIME=[5A55B500:01D104F5] X-MC-Unique: _ur6nlNsS6KvdmYfcJExcQ-1 Content-Type: text/plain; charset=WINDOWS-1252 Content-Transfer-Encoding: 8BIT Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Oct 09, 2015 at 10:41:43AM +0200, Arnd Bergmann wrote: > On Tuesday 06 October 2015 12:51:24 Arnd Bergmann wrote: > > > > I think it makes sense to stick with the traditional definition > > of MINSIGSTKSZ == "the minimum amount that you will always need, > > add whatever you require yourself" and SIGSTKSZ == "Should be > > enough for a couple of function calls". If we want to be conservative > > in the kernel, using 8192 and 32768, to stay with the x4 ratio > > that everyone else uses would be my first pick, though I'm not > > particularly attached to those values. > > > > > > On second thought, it really seems to late to make up our minds > about the size now that glibc has already established 5KB as the > minimum size. If we set it to 8KB/32KB, not just the testcase but > real applications would start failing when they use the 5KB > constant from glibc. I agree for MINSIGSTKSZ. We could still raise SIGSTKSZ if we think that will be more future-proof (SIGSTKSZ would be less than the magic 4*MINSIGSTKSZ that most arches assume, unless SIGSTKSZ is made >=20KB). Those might be independent changes. The definition of MINSIGSTKSZ is definitely broken right now, whereas SIGSTKSZ could be debated, but isn't actually broken. Cheers ---Dave