I was playing with luks when once I got a oops and the whole crypt-things are broken upon reboot. Unfortunately I cant reproduce it. It stopped in the luksOpen operation (it was a container, bound by losetup and formated by luks with serpent-lrw-benbi). bug 183407 Unable to handle kernel NULL pointer dereference at 0000000000000000 RIP: [<ffffffff8022f288>] mempool_free+0x18/0xe0 PGD 275a1067 PUD aa8a067 PMD 0 Oops: 0000 [1] PREEMPT CPU 0 Modules linked in: serpent sha1 lrw blkcipher gf128mul cryptomgr dm_crypt aes crypto_algapi nls_iso8859_1 nls_cp437 vfat fat isofs nls_base edac_mc cpufreq_conservative saa7134_alsa tuner saa7134 video_buf ir_kbd_i2c ir_common it87 hwmon_vid i2c_isa autofs4 tun af_packet ipt_MASQUERADE xt_tcpudp iptable_nat nf_nat xt_condition iptable_filter xt_mac nf_conntrack_ipv4 xt_state nf_conntrack ipt_TARPIT ipt_LOG ipt_REJECT ip_tables x_tables snd_pcm_oss snd_mixer_oss snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device fuse uinput ide_disk ide_generic rtc cpufreq_ondemand powernow_k8 freq_table pwc usb_storage ide_cd cdrom parport_pc parport usbhid hid nvidia(P) 8250_pnp 8250 serial_core k8temp hwmon ehci_hcd compat_ioctl32 videodev v4l2_common v4l1_compat snd_intel8x0 snd_ac97_codec ac97_bus snd_pcm snd_timer snd soundcore snd_page_alloc prism54 firmware_class forcedeth generic amd74xx ide_core ohci_hcd usbcore i2c_nforce2 sg evdev unix Pid: 24238, comm: kcryptd/0 Tainted: P 2.6.20-gentoo-r8 #1 RIP: 0010:[<ffffffff8022f288>] [<ffffffff8022f288>] mempool_free+0x18/0xe0 RSP: 0018:ffff810004aefd70 EFLAGS: 00010282 RAX: 0000000000000000 RBX: ffff810004aefdc0 RCX: 0000000000000000 RDX: 0000000000000007 RSI: 0000000000000000 RDI: ffff81000e31afa8 RBP: 0000000000000000 R08: ffff810004aee000 R09: 00000000ffffffff R10: 0000000000000000 R11: ffffffff802455c0 R12: ffff81000e31afa8 R13: ffffffff88824a40 R14: ffff810046673570 R15: ffff810046673550 FS: 00002b7fe5ac8da0(0000) GS:ffffffff80508000(0000) knlGS:00000000f75926b0 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000000000 CR3: 00000000689f0000 CR4: 00000000000006e0 Process kcryptd/0 (pid: 24238, threadinfo ffff810004aee000, task ffff81007381a200) Stack: ffff810004aefdc0 0000000000000246 ffff810079a20000 ffffffff88824a9e ffffffff80260fb9 ffff81001050ccc0 ffff81000e31afa8 ffff81000cf76fe0 ffff810015499b00 ffff81007381a200 ffff81001050ccc0 ffff81001050ccc0 Call Trace: [<ffffffff88824a9e>] :dm_crypt:kcryptd_do_work+0x5e/0x3a0 [<ffffffff80260fb9>] preempt_schedule_irq+0x59/0x90 [<ffffffff88824a40>] :dm_crypt:kcryptd_do_work+0x0/0x3a0 [<ffffffff8024d91a>] run_workqueue+0xca/0x1b0 [<ffffffff8024a7c0>] worker_thread+0x0/0x190 [<ffffffff8024a90b>] worker_thread+0x14b/0x190 [<ffffffff8027f440>] default_wake_function+0x0/0x10 [<ffffffff8024a7c0>] worker_thread+0x0/0x190 [<ffffffff80233233>] kthread+0xd3/0x110 [<ffffffff80292210>] keventd_create_kthread+0x0/0xb0 [<ffffffff8025d718>] child_rip+0xa/0x12 [<ffffffff80292210>] keventd_create_kthread+0x0/0xb0 [<ffffffff80233160>] kthread+0x0/0x110 [<ffffffff8025d70e>] child_rip+0x0/0x12 Code: 8b 06 39 46 04 0f 8d 9d 00 00 00 9c 5b fa 65 48 8b 04 25 10 RIP [<ffffffff8022f288>] mempool_free+0x18/0xe0 RSP <ffff810004aefd70> CR2: 0000000000000000
Created attachment 123492 [details] emerge --info
We can't really do anything with your crash as your kernel is tainted by the nvidia module. Regardless, it looks like this bug: http://bugzilla.kernel.org/show_bug.cgi?id=7388
I somewhy expected such an answer... thats why I havent reported it on vanilla-kernel... but gentoo is supporting the nvidia driver, so it should take care... but never mind, I cant reproduce the problem, neigher with nor without nvidia loaded... (tried it 5000 times... and nothing happend...) Thats why I was posting the tained report... but the patches of upstream might be helpfull for the gentoo-patchset...