From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752980AbdJSLdg (ORCPT ); Thu, 19 Oct 2017 07:33:36 -0400 Received: from mail.efficios.com ([167.114.142.141]:48367 "EHLO mail.efficios.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752738AbdJSLde (ORCPT ); Thu, 19 Oct 2017 07:33:34 -0400 Date: Thu, 19 Oct 2017 11:35:42 +0000 (UTC) From: Mathieu Desnoyers To: Ben Maurer , "Carlos O'Donell" , Florian Weimer Cc: "Paul E. McKenney" , Boqun Feng , Peter Zijlstra , Paul Turner , Andrew Hunter , Andy Lutomirski , Dave Watson , Josh Triplett , Will Deacon , linux-kernel , Thomas Gleixner , Andi Kleen , Chris Lameter , Ingo Molnar , "H. Peter Anvin" , rostedt , Linus Torvalds , Andrew Morton , Russell King , Catalin Marinas , Michael Kerrisk , Alexander Viro , linux-api Message-ID: <14399397.44601.1508412942855.JavaMail.zimbra@efficios.com> In-Reply-To: <515879378.43966.1508350299712.JavaMail.zimbra@efficios.com> References: <20171012230326.19984-1-mathieu.desnoyers@efficios.com> <20171012230326.19984-2-mathieu.desnoyers@efficios.com> <515879378.43966.1508350299712.JavaMail.zimbra@efficios.com> Subject: Re: [RFC PATCH v9 for 4.15 01/14] Restartable sequences system call MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-Originating-IP: [167.114.142.141] X-Mailer: Zimbra 8.7.11_GA_1854 (ZimbraWebClient - FF52 (Linux)/8.7.11_GA_1854) Thread-Topic: Restartable sequences system call Thread-Index: AQHTQ6515cQ6nxXTN0eVL251OZT19KLkKPJz5kNtRjeqVFQ6rg== Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Speaking of optimization, I think the rseq.c helper library (and eventually glibc) should define the __rseq_abi TLS variable with __attribute__((tls_model("initial-exec"))). It provides faster, and signal-safe, accesses to the TLS variable from libraries. The idea you were suggesting where the application could override the glibc symbol with its own just ends up enforcing an initial-exec model, but it looks like we can do this directly from the library. Thoughts ? Thanks, Mathieu -- Mathieu Desnoyers EfficiOS Inc. http://www.efficios.com From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mathieu Desnoyers Subject: Re: [RFC PATCH v9 for 4.15 01/14] Restartable sequences system call Date: Thu, 19 Oct 2017 11:35:42 +0000 (UTC) Message-ID: <14399397.44601.1508412942855.JavaMail.zimbra@efficios.com> References: <20171012230326.19984-1-mathieu.desnoyers@efficios.com> <20171012230326.19984-2-mathieu.desnoyers@efficios.com> <515879378.43966.1508350299712.JavaMail.zimbra@efficios.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <515879378.43966.1508350299712.JavaMail.zimbra-vg+e7yoeK/dWk0Htik3J/w@public.gmane.org> Sender: linux-api-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Ben Maurer , Carlos O'Donell , Florian Weimer Cc: "Paul E. McKenney" , Boqun Feng , Peter Zijlstra , Paul Turner , Andrew Hunter , Andy Lutomirski , Dave Watson , Josh Triplett , Will Deacon , linux-kernel , Thomas Gleixner , Andi Kleen , Chris Lameter , Ingo Molnar , "H. Peter Anvin" , rostedt , Linus Torvalds , Andrew Morton , Russell King , Catalin Marinas , Michael List-Id: linux-api@vger.kernel.org Speaking of optimization, I think the rseq.c helper library (and eventually glibc) should define the __rseq_abi TLS variable with __attribute__((tls_model("initial-exec"))). It provides faster, and signal-safe, accesses to the TLS variable from libraries. The idea you were suggesting where the application could override the glibc symbol with its own just ends up enforcing an initial-exec model, but it looks like we can do this directly from the library. Thoughts ? Thanks, Mathieu -- Mathieu Desnoyers EfficiOS Inc. http://www.efficios.com