From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S266361AbUBLG4x (ORCPT ); Thu, 12 Feb 2004 01:56:53 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S266365AbUBLG4x (ORCPT ); Thu, 12 Feb 2004 01:56:53 -0500 Received: from defout.telus.net ([199.185.220.240]:37114 "EHLO priv-edtnes46.telusplanet.net") by vger.kernel.org with ESMTP id S266361AbUBLG4u (ORCPT ); Thu, 12 Feb 2004 01:56:50 -0500 Subject: [2.6.3-rc2 bk] ieee1394 oops on bootup From: Bob Gill To: Linux Kernel Mailing List Content-Type: text/plain Message-Id: <1076569092.3550.53.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 (1.4.5-7) Date: Wed, 11 Feb 2004 23:58:12 -0700 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi. I tried your suggestion (but without success). I built 2.6.3-rc2-bk1 (and on boot I get the badness message). I backed up /lib/modules/2.6.3-rc2-bk1/modules.dep and deleted all of the 1394 references from it. I then booted 2.6.3-rc2-bk1 (no error messages but no 1394 modules either). Since insmod/modprobe/depmod will tell you no 1394 modules exist (at this point), I copied modules.dep.bak to modules.dep (and then ran modprobe ohci1394). Unfortunately, I got: ohci1394: $Rev: 1097 $ Ben Collins ohci1394: fw-host0: OHCI-1394 1.0 (PCI): IRQ=[9] MMIO=[eb005000-eb0057ff] Max Packet=[2048] Badness in kobject_get at lib/kobject.c:431 Call Trace: [] kobject_get+0x4c/0x4e [] get_device+0x18/0x21 [] bus_for_each_dev+0x63/0xba [] nodemgr_node_probe+0x4d/0x124 [ieee1394] [] nodemgr_probe_ne_cb+0x0/0x8d [ieee1394] [] nodemgr_host_thread+0x17d/0x1a1 [ieee1394] [] nodemgr_host_thread+0x0/0x1a1 [ieee1394] [] kernel_thread_helper+0x5/0xb Unable to handle kernel paging request at virtual address b828ec83 printing eip: b828ec83 *pde = 00000000 modprobe: FATAL: Module sbp2 already in kernel. Oops: 0000 [#1] ieee1394.agent[2413]: ... can't load module sbp2 kernel: CPU: 0 ieee1394.agent[2413]: missing kernel or user mode driver sbp2 kernel: EIP: 0060:[] Not tainted EFLAGS: 00010282 EIP is at 0xb828ec83 eax: b828ec83 ebx: f8c9d3c4 ecx: f75d7f9c edx: 00000000 esi: f8c938ff edi: 00000000 ebp: f8c92669 esp: f75d7f40 ds: 007b es: 007b ss: 0068 Process knodemgrd_0 (pid: 2411, threadinfo=f75d6000 task=f725e080) Stack: c0192fbc f8c9d3c4 f8c9d3a0 f8c9d3a8 f8c9d300 f70bf244 c01d82cc f8c9d3c4 f75d7f9c f8c9d34c 00000000 f70bf23c f75d7f9c f7155398 f75d7f9c f8c9401e f8c9d300 f70bf23c f75d7f9c f8c93ee7 00000004 f8c9cd40 f75cc000 f7155398 Call Trace: [] kobject_cleanup+0x98/0x9a [] bus_for_each_dev+0x80/0xba [] nodemgr_node_probe+0x4d/0x124 [ieee1394] [] nodemgr_probe_ne_cb+0x0/0x8d [ieee1394] [] nodemgr_host_thread+0x17d/0x1a1 [ieee1394] [] nodemgr_host_thread+0x0/0x1a1 [ieee1394] [] kernel_thread_helper+0x5/0xb Code: Bad EIP value. <6>sbp2: $Rev: 1096 $ Ben Collins The only really weird thing is that I prefer to keep things as modular as possible (and if for whatever reason initrd-yadda.img isn't created after the kernel compile, I run mkinitrd myself). Here, the messages are telling me that sbp2 is already in the kernel (and so sbp2 can't be loaded), but my build script shows:CONFIG_IEEE1394_SBP2=m CONFIG_IEEE1394_SBP2_PHYS_DMA=y and when I look in /lib/modules/`uname -r`/kernel/drivers/ie* I get sbp2.ko along with everything else. I double checked that the line loading sbp2 was not in the version of modules.dep that I booted with (and that it was in the one I finally loaded the 1394 modules with). The system seems to be confused as to whether sbp2 is compiled into the kernel or available as a module. The logic race theory was a good test, but I still get an Oops when loading modules long after boot time (sorry). Bob