From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.33) id 1BdRzq-0003D9-Pi for qemu-devel@nongnu.org; Thu, 24 Jun 2004 07:02:22 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.33) id 1BdRzp-0003Cq-0f for qemu-devel@nongnu.org; Thu, 24 Jun 2004 07:02:22 -0400 Received: from [199.232.76.173] (helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.33) id 1BdRzo-0003Cn-Uy for qemu-devel@nongnu.org; Thu, 24 Jun 2004 07:02:20 -0400 Received: from [194.226.236.116] (helo=ums.usu.ru) by monty-python.gnu.org with esmtp (TLSv1:DES-CBC3-SHA:168) (Exim 4.34) id 1BdRy3-0004nn-UF for qemu-devel@nongnu.org; Thu, 24 Jun 2004 07:00:32 -0400 Message-ID: <40DAB30E.10605@ums.usu.ru> Date: Thu, 24 Jun 2004 16:55:10 +0600 From: "Alexander E. Patrakov" MIME-Version: 1.0 Subject: Re: [Qemu-devel] Fedora Core 2 patch References: <200406241114.45544.jm@poure.com> <200406241231.43037.hetz@softier.com> In-Reply-To: <200406241231.43037.hetz@softier.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Reply-To: qemu-devel@nongnu.org List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Hetz Ben Hamo wrote: > I'm removing the patch for fedora. > > The solution is simple. instead of just running qemu and sticking the > parameters, you should do: > > LD_ASSUME_KERNEL=2.2.5 qemu > > Feel free to create an alias in your .bashrc or whatever you favorite shell > is ;) > > I tested it on Fedora Core 2 and in Rawhide - works perfectly OK. Unfortunately, this solution is not acceptable for me. The problem is that it relies upon RedHat-specific glibc setup. RedHat provides separate glibc libraries for TLS and (obsolete) non-TLS systems. The LD_ASSUME_KERNEL variable selects between them and forces qemu to use the old library. On my system (LFS CVS HEAD as of April 2004) there is just NO non-TLS version of glibc, so this "solution" doesn't work. The same applies to most of modern source-based distros (e.g. Gentoo). Probably we have to insert into the ./configure script a test for the brokenness that forces us to use __libc_sigaction on older systems, and use __libc_sigaction ONLY if the regular sigaction is broken. Unfortunately, I can't do this myself because I don't know what exactly is broken and therefore I can't write the testcase. Yes, this means that there will be no single portable binary distribution. But it is not worse than the current situation with non-working qemu binary. -- Alexander E. Patrakov