Hi, I've got a reproducible oops when using cryptoloop on vanilla 2.6.0, 2.6.1 and 2.6.2 (2.4.* works fine). Way to reproduce: dd if=/dev/urandom of=crypto bs=1024 count=some_size losetup -e some_cipher /dev/loop0 crypto #Any of those commands causes oops and hard lockup: cp /dev/loop0 /dev/null mkreiserfs /dev/loop0 mkfs.ext2 /dev/loop0 Loop without cryptoapi works fine: dd if=/dev/urandom of=crypto bs=1024 count=some_size losetup /dev/loop0 crypto cp /dev/loop0 /dev/null #ok, no oops Sometimes it writes 10KB sometimes 10MB before it oopses. It may be just coincidence but without HT enabled it lived longer. As it locks hard after oops I don't have oops messages in log. I have written down some: 1. init: panic: Segmentation violation at 0x40126A83! Sleeping for 30 sec... 2. Unable to handle kernel paging request at virtual address 87550c8a EIP is at cleanup_bitmap_list +0x48/0x68 ... 3. Unable to handle kernel paging request at virtual address EIP IS AT get_cnode +0x61/0x94 ... 4. Unable to handle kernel paging request at virtual address faad7260 EIP is at journal_mark_dirty +0x1b5 Calltrace: flush_old_commits +0x113/0x16f reiserfs_write_super +0x98/0x9a sync_supers +0xd0/0xea do_sync sys_sync syscall_call 5. syncing...(after mkfs so it was almost done) Unable to handle kernel paging request at virtual address 9756a996 EIP is at cleanup_bitmap_list +0xb/0xd4 Calltrace: autoremove_wake_function cleanup_freed_for_journal_list ... 6. attempted to kill init bad: schedulling while atomic EIP is at add_wait_queue And much more (mostly in FS code) Tested ciphers: aes, blowfish (both as module and compiled in) Filesystems with crypto file: ReiserFS, FAT Tested sizes of crypto file: 50MB, 10GB Tested disk drives and controllers: ICH5 with WD120JB (pata), external CMD649 (Kouwell 671A) with 80GB Seagate barracuda IV Tested chipsets: I865 (Asus P4P800), I875 (Abit IC7 G) Other hardware: P4 2.6 (HT, SMP kernel), 1GB RAM (High memory support 4GB). Thoroughly tested with memtest and mprime95. See lspci and config in attachment. Kernel is not tainted. Write me please if additional information is need. miho PS. Excuse my poor english.