This occurred while testing a storage target implementation and after having stopped the storage target. An ext4 filesystem had been created on top of the initiator device. After the target had been stopped, CPU load went up to one and stayed there. Killing the processes (via killall -9) that were using the filesystem was not possible. Also, the following complaints appeared periodically in the system log: kernel: sd 9:0:0:0: [sdc] Unhandled error code kernel: sd 9:0:0:0: [sdc] Result: hostbyte=DID_NO_CONNECT driverbyte=DRIVER_OK kernel: sd 9:0:0:0: [sdc] CDB: Write(10): 2a 00 00 08 a0 58 00 00 08 00 kernel: end_request: I/O error, dev sdc, sector 565336 kernel: ------------[ cut here ]------------ kernel: WARNING: at fs/ext4/extents.c:3765 ext4_convert_unwritten_extents+0xf7/0x120 [ext4]() kernel: Hardware name: P5Q DELUXE kernel: Modules linked in: ext4 jbd2 crc16 ib_srp scsi_transport_srp scsi_tgt fuse ip6t_LOG ipt_MASQUERADE xt_pkttype xt_TCPMSS xt_tcpudp ipt_LOG xt_limit iptable_nat nf_nat radeon ttm drm_kms_helper drm i2c_algo_bit snd_pcm_os eq snd_seq_device vmnet vmblock vsock vmci vmmon af_packet rdma_ucm ip6t_REJECT nf_conntrack_ipv6 nf_defrag_ipv6 ip6table_raw xt_NOTRACK ipt_REJECT iptable_raw iptable_filter ip6table_mangle nf_conntrack_netbios_ns nf_conntrack ip_tables xt_conntrack nf_conntrack ip6table_filter ip6_tables x_tables ib_ipoib ib_uverbs ib_umad mlx4_ib cpufreq_conservative cpufreq_userspace cpufreq_powersave microcode acpi_cpufreq mperf dm_mod coretemp snd_hda_codec_hdmi snd_hda_intel snd_hda_codec snd_hwdep snd_pcm snd_timer joydev hid_belkin snd i2c_i801 soundcore intel_agp sg snd_page_alloc mlx4_core i2c_core sr_mod intel_gtt pcspkr serio_raw cdrom asus_atk0110 button raid456 async_raid6_re id raid6_pq kernel: async_xor xor async_memcpy async_tx raid10 raid0 uhci_hcd ehci_hcd rtc_cmos sd_mod rtc_core usbcore rtc_lib edd raid1 ext3 mbcache jbd fan ide_pci_generic ide_core ata_generic ata_piix pata_marvell ahci libahci libata t mal_sys hwmon [last unloaded: ib_srp] kernel: Pid: 9217, comm: kworker/0:1 Not tainted 2.6.38.7-scst #1 kernel: Call Trace: kernel: [<ffffffff8104c86f>] ? warn_slowpath_common+0x7f/0xc0 kernel: [<ffffffff8104c8ca>] ? warn_slowpath_null+0x1a/0x20 kernel: [<ffffffffa06640b7>] ? ext4_convert_unwritten_extents+0xf7/0x120 [ext4] kernel: [<ffffffff813c04e9>] ? trace_hardirqs_on_thunk+0x3a/0x3c kernel: [<ffffffffa064676a>] ? ext4_free_io_end+0xba/0xd0 [ext4] kernel: [<ffffffffa06467d0>] ? ext4_end_io_nolock+0x50/0x120 [ext4] kernel: [<ffffffffa06468df>] ? ext4_end_io_work+0x3f/0xb0 [ext4] kernel: [<ffffffffa06468a0>] ? ext4_end_io_work+0x0/0xb0 [ext4] kernel: [<ffffffff8106766b>] ? process_one_work+0x11b/0x490 kernel: [<ffffffff8106914f>] ? worker_thread+0x16f/0x350 kernel: [<ffffffff81068fe0>] ? worker_thread+0x0/0x350 kernel: [<ffffffff8106d8d6>] ? kthread+0x96/0xa0 kernel: [<ffffffff81003cd4>] ? kernel_thread_helper+0x4/0x10 kernel: [<ffffffff8103d7b1>] ? finish_task_switch+0x51/0xb0 kernel: [<ffffffff813c07ac>] ? _raw_spin_unlock_irq+0x1c/0x40 kernel: [<ffffffff813c13f1>] ? restore_args+0x0/0x30 kernel: [<ffffffff8106d840>] ? kthread+0x0/0xa0 kernel: [<ffffffff81003cd0>] ? kernel_thread_helper+0x0/0x10 kernel: ---[ end trace f53df40954754bc1 ]--- kernel: ext4_convert_unwritten_extents: ext4_ext_map_blocks returned error inode#27, block=329, max_blocks=1 kernel: EXT4-fs (sdc): previous I/O error to superblock detected kernel: EXT4-fs error (device sdc): __ext4_get_inode_loc:4779: inode #27: block 97: comm kworker/0:1: unable to read itable block kernel: EXT4-fs error (device sdc) in ext4_reserve_inode_write:5619: IO failure kernel: EXT4-fs (sdc): previous I/O error to superblock detected kernel: ext4_end_io_nolock: failed to convert unwritten extents to written extents, error is -5 io is still on inode 27 aio dio list