Description of problem: While attempting to access my Sony Clie under Severn, X would eventually stop responding correctly, and the machine would lock up. After running similiar steps from the console, I found that the machine was panicing. Version-Release number of selected component (if applicable): 2.4.21-20.1.2024.2.1.nptl (Atlhon + i686) How reproducible: Every time, on two separate machines. Steps to Reproduce: 1.Load "Visor" kernel module. 2.Press "hotsync" on Clie cradle 3.Around when the Clie stops trying to sync, the machine panics. Actual results: Kernel panic Expected results: Successful sync. Additional info: Note that this is my first attempt to access the Clie under linux. I'm not sure if this panic is specific to Severn, or if it happens on earlier releases as well. I should be able to test that in a few days. Here is the panic: Aug 14 23:50:49 necronomicon kernel: usb-uhci.c: interrupt, status 3, frame# 551 Aug 14 23:50:49 necronomicon kernel: usb.c: USB disconnect on device 00:07.2-2 address 2 Aug 14 23:50:49 necronomicon kernel: visor.c: Bytes In = 382 Bytes Out = 362 Aug 14 23:50:49 necronomicon kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000998 Aug 14 23:50:49 necronomicon kernel: printing eip: Aug 14 23:50:49 necronomicon kernel: e4d634da Aug 14 23:50:49 necronomicon kernel: *pde = 1d23f067 Aug 14 23:50:49 necronomicon kernel: *pte = 00000000 Aug 14 23:50:49 necronomicon kernel: Oops: 0002 Aug 14 23:50:49 necronomicon kernel: visor usbserial radeon agpgart parport_pc lp parport button autofs tg3 ipt_REJECT iptable_filter ip_tables sg sr_mod floppy ide-scsi scsi_mod ide-cd cdrom key Aug 14 23:50:49 necronomicon kernel: CPU: 0 Aug 14 23:50:49 necronomicon kernel: EIP: 0060:[<e4d634da>] Not tainted Aug 14 23:50:49 necronomicon kernel: EFLAGS: 00010246 Aug 14 23:50:49 necronomicon kernel: Aug 14 23:50:49 necronomicon kernel: EIP is at usb_serial_disconnect [usbserial] 0x6a (2.4.21-20.1.2024.2.1.nptl) Aug 14 23:50:49 necronomicon kernel: eax: 00000000 ebx: ddae6400 ecx: c178c000 edx: dda5df7c Aug 14 23:50:49 necronomicon kernel: esi: ddae641c edi: 00000000 ebp: ddae6400 esp: c178df18 Aug 14 23:50:49 necronomicon kernel: ds: 0068 es: 0068 ss: 0068 Aug 14 23:50:49 necronomicon kernel: Process khubd (pid: 78, stackpage=c178d000) Aug 14 23:50:49 necronomicon kernel: Stack: ddae641c 00000000 00000010 e4d65480 e4d65460 00000000 dac29494 e0836274 Aug 14 23:50:49 necronomicon kernel: dd6ef800 ddae6400 dd6ef804 00000002 00000000 dd6ef800 00000100 0000000a Aug 14 23:50:49 necronomicon kernel: dfeade00 00000001 e08390cf dfeadf10 00000002 00000010 dfeade00 e0838b2c Aug 14 23:50:49 necronomicon kernel: Call Trace: [<e4d65480>] usb_serial_driver [usbserial] 0x20 (0xc178df24) Aug 14 23:50:49 necronomicon kernel: [<e4d65460>] usb_serial_driver [usbserial] 0x0 (0xc178df28) Aug 14 23:50:49 necronomicon kernel: [<e0836274>] usb_disconnect_R2a7790a1 [usbcore] 0x94 (0xc178df34) Aug 14 23:50:49 necronomicon kernel: [<e08390cf>] usb_hub_port_connect_change [usbcore] 0x26f (0xc178df60) Aug 14 23:50:49 necronomicon kernel: [<e0838b2c>] usb_hub_port_status [usbcore] 0x6c (0xc178df74) Aug 14 23:50:49 necronomicon kernel: [<e08393b8>] usb_hub_events [usbcore] 0x2d8 (0xc178df94) Aug 14 23:50:49 necronomicon kernel: [<e0839446>] usb_hub_thread [usbcore] 0x36 (0xc178dfc0) Aug 14 23:50:49 necronomicon kernel: [<e0839410>] usb_hub_thread [usbcore] 0x0 (0xc178dfc4) Aug 14 23:50:49 necronomicon kernel: [<e0839410>] usb_hub_thread [usbcore] 0x0 (0xc178dfe0) Aug 14 23:50:49 necronomicon kernel: [<c010727d>] kernel_thread_helper [kernel] 0x5 (0xc178dff0) Aug 14 23:50:49 necronomicon kernel: Aug 14 23:50:49 necronomicon kernel: Aug 14 23:50:49 necronomicon kernel: Code: c7 80 98 09 00 00 00 00 00 00 8d 4e 58 ff 43 74 0f 8e 74 05
Looks familiar. I found that driver particularly easy to break when I got my Clie too, and that was with a mainstream kernel. Chances are this is going to be fixed upstream before someone at Red Hat digs into it, so you may be better off trying to work with the upstream maintainer, especially as iirc, we don't apply any patches to that driver..
I just tested this with Redhat 9 with the 2.4.20-18.9 kernel, and could not reproduce the panic (although I still don't have gpilot setup properly.
I'm pretty sure my "helper process" fixes that. Closing.