Hi Jens, Marcelo, Doing an "eject" with my somewhat damaged SuSE 5.3 CD gave me the appended reproducible oops. This bug is present in both 2.4 (visual inspection) and 2.5. Here is the code path before the oops ide-cd.c static int cdrom_queue_packet_command(ide_drive_t *drive, struct packet_command *pc) { if (ide_do_drive_cmd (drive, &req, ide_wait)) { <== [1] printk("%s: do_drive_cmd returned stat=%02x,err=%02x\n", drive->name, req.buffer[0], req.buffer[1]); /* FIXME: we should probably abort/retry or something */ [1] ide_do_drive_cmd returns -EIO so we end up doing the printk. Tadow! NULL dereference because of (char *)req.buffer. Then again, this printk seems quite redundant IMO. Patch diffed and tested on 2.5.2-pre10, should also apply to 2.4.17+ --- linux-2.5.2-pre10/drivers/ide/ide-cd.c.orig Wed Jan 9 09:37:24 2002 +++ linux-2.5.2-pre10/drivers/ide/ide-cd.c Wed Jan 9 09:48:15 2002 @@ -1399,8 +1399,8 @@ req.flags = REQ_PC; req.special = (char *) pc; if (ide_do_drive_cmd (drive, &req, ide_wait)) { - printk("%s: do_drive_cmd returned stat=%02x,err=%02x\n", - drive->name, req.buffer[0], req.buffer[1]); + printk(KERN_DEBUG "%s: do_drive_cmd returned -EIO\n", + drive->name); /* FIXME: we should probably abort/retry or something */ } if (pc->stat != 0) { Regards, Zwane Mwaikambo [root@mondecino root]# eject Jan 1 02:01:25 mondecino kernel: VFS: Disk change detected on device ide1(22,64) Jan 1 02:01:35 mondecino kernel: hdd: irq timeout: status=0xd0 { Busy } Unable to handle kernel NULL pointer dereference at virtual address 00000001 *pde = 0a5d5001 *pte = 00000000 Oops: 0000 CPU: 0 EIP: 0010:[] Not tainted EFLAGS: 00010282 eax: fffffffb ebx: ca5d7dc0 ecx: 013de100 edx: 00000000 esi: 0000000a edi: ca5d7cec ebp: c03de138 esp: ca5d7cec ds: 0018 es: 0018 ss: 0018 Process eject (pid: 704, stackpage=ca5d7000) Stack: c03de138 c03de138 00002000 00000000 00000000 00000000 00000000 00000220 ffffffff 00001640 00000001 00000000 00000000 00000000 00000000 00000000 00000000 00000000 ca5d7dc0 00000000 ca5d7ca8 00000000 00000000 00000000 Call Trace: [] [] [] [] [] [] [] [] [] [] [] [] [] [] [] [] [] [] [] [] Code: 0f be 42 01 50 0f be 02 50 8d 85 34 01 00 00 50 68 40 4a 2f >>EIP; c02268ae <===== Trace; c022714d Trace; c02271f3 Trace; c01506d5 Trace; c0228a7d Trace; c0171640 Trace; c0162ef0 Trace; c0172ce7 Trace; c0171640 Trace; c021bfb9 Trace; c0150e69 Trace; c02311c8 Trace; c02289d0 Trace; c021c18e Trace; c0150f89 Trace; c0151125 Trace; c01472a6 Trace; c01471ad Trace; c015478e Trace; c0147517 Trace; c010910b Code; c02268ae 00000000 <_EIP>: Code; c02268ae <===== 0: 0f be 42 01 movsbl 0x1(%edx),%eax <===== Code; c02268b2 4: 50 push %eax Code; c02268b3 5: 0f be 02 movsbl (%edx),%eax Code; c02268b6 8: 50 push %eax Code; c02268b7 9: 8d 85 34 01 00 00 lea 0x134(%ebp),%eax Code; c02268bd f: 50 push %eax Code; c02268be 10: 68 40 4a 2f 00 push $0x2f4a40 - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/