From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by neteng.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id UAA48702 for ; Mon, 3 May 1999 20:37:45 -0700 (PDT) Return-Path: Received: (from majordomo-owner@localhost) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id UAA66685 for linux-list; Mon, 3 May 1999 20:34:04 -0700 (PDT) mail_from (owner-linux@relay.engr.sgi.com) Received: from sgi.com (sgi.engr.sgi.com [192.26.80.37]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id UAA53958 for ; Mon, 3 May 1999 20:34:02 -0700 (PDT) mail_from (clepple@foo.tho.org) Received: from foo.tho.org (pr250.pheasantrun.net [208.140.225.250]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id XAA00256 for ; Mon, 3 May 1999 23:33:57 -0400 (EDT) mail_from (clepple@foo.tho.org) Received: from foo.tho.org (clepple@sprocket.foo.tho.org [206.223.45.3]) by foo.tho.org (8.8.7/8.8.7) with ESMTP id XAA29966 for ; Mon, 3 May 1999 23:33:56 -0400 Message-ID: <372E6AA0.505A6071@foo.tho.org> Date: Tue, 04 May 1999 03:33:52 +0000 From: Charles Lepple X-Mailer: Mozilla 4.5 [en] (X11; U; Linux 2.0.36 i586) X-Accept-Language: en MIME-Version: 1.0 To: Linux/SGI Subject: building an elf64 R10k kernel Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux@cthulhu.engr.sgi.com Precedence: bulk Hey all, Does anyone out there know any details on building elf64 objects? I was all happy about seeing Andrew's Indigo2 patches, and decided that I _had_ to try and make it work on an Indigo2 Impact 10000... Suffice it to say that it isn't straightforward. I first tried to get the machine to boot an elf32 kernel. That failed miserably, since it doesn't recognize 32-bit ELF objects. Comparison of sashARCS and sash64 indicated that I needed to use elf64. My first inclination was to just try the same load address as the Indys and r4400 Indigo2s use. I can't remember the exact error message, but the PROM basically panicked while loading the first word at 0x88002000 (or wherever). So I tried using the 64 bit start address from sash64 (high hex nibble is 0xa). Needless to say, the elf32 linker options conflicted with that one, and the linker (from binutils-2.8.1 or so) seg faulted. When I finally recompiled with -mips3 and with a revised linker script, I noticed something curious. Instead of just truncating the 64-bit addresses, it rounded them off to 0xffffffff. Any thoughts on this? Am I approaching this wrong? And does anyone have a semi-offical memory map for the R10k Indigo2s? -- Charles Lepple System Administrator, Virginia Tech EE Workstation Labs clepple@ee.vt.edu || http://www.foo.tho.org/charles/