Sorry folks for making so much trouble, but there is still another locking issue with the kernel client. iput gets called with locks held, but iput can sleep and did in this occasion. It did not hang the system this time, but it can be quite nasty. Log attached as usual. Amon Ott -- Dr. Amon Ott m-privacy GmbH Tel: +49 30 24342334 Am Köllnischen Park 1 Fax: +49 30 24342336 10179 Berlin http://www.m-privacy.de Amtsgericht Charlottenburg, HRB 84946 Geschäftsführer: Dipl.-Kfm. Holger Maczkowsky, Roman Maczkowsky GnuPG-Key-ID: 0x2DD3A649
Dec 9 10:45:26 tgpro1 kernel: BUG: scheduling while atomic: firefox-bin/31447/0x10000002 Dec 9 10:45:26 tgpro1 kernel: 2 locks held by firefox-bin/31447: Dec 9 10:45:26 tgpro1 kernel: #0: (&sb->s_type->i_mutex_key#12){+.+.+.}, at: [<006c1e12>] ceph_dir_fsync+0x77/0x18a [ceph] Dec 9 10:45:26 tgpro1 kernel: #1: (&(&ci->i_unsafe_lock)->rlock){+.+...}, at: [<006c1ecc>] ceph_dir_fsync+0x131/0x18a [ceph] Dec 9 10:45:26 tgpro1 kernel: Modules linked in: parport_pc lp parport ceph libceph btrfs zlib_deflate lzo_decompress lzo_compress zlib_inflate libcrc32c fuse e1000e i2c_i801 i2c_core iTCO_wdt iTCO_vendor_support evdev igb rtc_cmos button Dec 9 10:45:26 tgpro1 kernel: Pid: 31447, comm: firefox-bin Tainted: G W 3.1.4-rsbac #1 Dec 9 10:45:26 tgpro1 kernel: Call Trace: Dec 9 10:45:26 tgpro1 kernel: [<00060591>] __schedule_bug+0x5d/0x64 Dec 9 10:45:26 tgpro1 kernel: [<0053ca0f>] __schedule+0x8e/0xb60 Dec 9 10:45:26 tgpro1 kernel: [<000025c0>] ? __switch_to+0x1a/0x1a Dec 9 10:45:26 tgpro1 kernel: [<0008e987>] ? mark_held_locks+0xac/0xc9 Dec 9 10:45:26 tgpro1 kernel: [<000a095a>] ? res_counter_uncharge+0x50/0x58 Dec 9 10:45:26 tgpro1 kernel: [<0008ebfb>] ? trace_hardirqs_on_caller+0x10b/0x13c Dec 9 10:45:26 tgpro1 kernel: [<0008ec37>] ? trace_hardirqs_on+0xb/0xd Dec 9 10:45:26 tgpro1 kernel: [<0053d555>] _cond_resched+0x24/0x37 Dec 9 10:45:26 tgpro1 kernel: [<000b9ef3>] truncate_inode_pages_range+0x10e/0x2d8 Dec 9 10:45:26 tgpro1 kernel: [<0007ebc5>] ? wake_up_bit+0x57/0x5b Dec 9 10:45:26 tgpro1 kernel: [<00100414>] ? evict+0x57/0x12c Dec 9 10:45:26 tgpro1 kernel: [<000ba0c9>] truncate_inode_pages+0xc/0x10 Dec 9 10:45:26 tgpro1 kernel: [<00100460>] evict+0xa3/0x12c Dec 9 10:45:26 tgpro1 kernel: [<0010063c>] iput+0x153/0x158 Dec 9 10:45:26 tgpro1 kernel: [<000fe5fe>] d_kill+0xc9/0xe2 Dec 9 10:45:26 tgpro1 kernel: [<000fed51>] dput+0x120/0x12b Dec 9 10:45:26 tgpro1 kernel: [<006d5883>] ? ceph_mdsc_create_request+0xf5/0xf5 [ceph] Dec 9 10:45:26 tgpro1 kernel: [<006d5912>] ceph_mdsc_release_request+0x8f/0xf9 [ceph] Dec 9 10:45:26 tgpro1 kernel: [<001b44a4>] kref_put+0x3f/0x48 Dec 9 10:45:26 tgpro1 kernel: [<006c1ed9>] ceph_dir_fsync+0x13e/0x18a [ceph] Dec 9 10:45:26 tgpro1 kernel: [<000237f0>] ? kunmap+0x7/0x4e Dec 9 10:45:26 tgpro1 kernel: [<006c1d9b>] ? ceph_finish_lookup+0x92/0x92 [ceph] Dec 9 10:45:26 tgpro1 kernel: [<0010c7cf>] vfs_fsync_range+0x23/0x35 Dec 9 10:45:26 tgpro1 kernel: [<0010c7f5>] vfs_fsync+0x14/0x16 Dec 9 10:45:26 tgpro1 kernel: [<0010ca00>] do_fsync+0x23/0x33 Dec 9 10:45:26 tgpro1 kernel: [<0010ca20>] sys_fdatasync+0x10/0x12 Dec 9 10:45:26 tgpro1 kernel: [<0053faba>] syscall_call+0x7/0xb Dec 9 10:45:26 tgpro1 kernel: [<00530000>] ? update_res+0xc0/0x179