-
Emmanuel Grumbach authored
Move the sysfs_create_group to iwl_ucode_callback after we have safely got the firmware. The motivation to do this comes from a warning from lockdep which detected that we request priv->mutex while holding s_active during a sysfs request (show_statistics in the example copy pasted). The reverse order exists upon request_firmware: request_firmware which is a sysfs operation that requires s_active is run under priv->mutex. This ensures that we don't get sysfs request before we finish to request the firmware, avoiding this deadlock. ======================================================= [ INFO: possible circular locking dependency detected ] ------------------------------------------------------- cat/2595 is trying to acquire lock: (&priv->mutex){+.+.+.}, at: [<facfa598>] show_statistics+0x48/0x100 [iwlagn] but task is already holding lock: (s_active){++++.+}, at: [<c0580ebd>] sysfs_get_active_two+0x1d/0x50 which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #1 (s_active){++++.+}: [<c0489b74>] __lock_acquire+0xc44/0x1230 [<c048a1ed>] lock_acquire+0x8d/0x110 [<c0581499>] sysfs_addrm_finish+0xe9/0x180 [<c057f64a>] sysfs_hash_and_remove+0x4a/0x80 [<c05829d4>] sysfs_remove_group+0x44/0xd0 [<c0714b75>] dpm_sysfs_remove+0x15/0x20 [<c070dac8>] device_del+0x38/0x170 [<c070dc1e>] device_unregister+0x1e/0x60 [<c071838d>] _request_firmware+0x29d/0x550 [<c07186c7>] request_firmware+0x17/0x20 [<fad01bf1>] iwl_mac_start+0xb1/0x1230 [iwlagn] [<fa46ba06>] ieee80211_open+0x436/0x6f0 [mac80211] [<c0808cd2>] dev_open+0x92/0xf0 [<c0808b2b>] dev_change_flags+0x7b/0x190 [<c08148e8>] do_setlink+0x178/0x3b0 [<c0815169>] rtnl_setlink+0xf9/0x130 [<c081453b>] rtnetlink_rcv_msg+0x1bb/0x1f0 [<c0827ce6>] netlink_rcv_skb+0x86/0xa0 [<c081436c>] rtnetlink_rcv+0x1c/0x30 [<c08279c3>] netlink_unicast+0x263/0x290 [<c0828768>] netlink_sendmsg+0x1c8/0x2a0 [<c07f85fd>] sock_sendmsg+0xcd/0x100 [<c07f964d>] sys_sendmsg+0x15d/0x290 [<c07f9e6b>] sys_socketcall+0xeb/0x2a0 [<c040ad9f>] sysenter_do_call+0x12/0x38 -> #0 (&priv->mutex){+.+.+.}: [<c0489f84>] __lock_acquire+0x1054/0x1230 [<c048a1ed>] lock_acquire+0x8d/0x110 [<c08bb358>] __mutex_lock_common+0x58/0x470 [<c08bb84a>] mutex_lock_nested+0x3a/0x50 [<facfa598>] show_statistics+0x48/0x100 [iwlagn] [<c070d219>] dev_attr_show+0x29/0x50 [<c057fecd>] sysfs_read_file+0xdd/0x190 [<c052880f>] vfs_read+0x9f/0x190 [<c0528d22>] sys_read+0x42/0x70 [<c040ad9f>] sysenter_do_call+0x12/0x38 other info that might help us debug this: 3 locks held by cat/2595: #0: (&buffer->mutex){+.+.+.}, at: [<c057fe25>] sysfs_read_file+0x35/0x190 #1: (s_active){++++.+}, at: [<c0580ecd>] sysfs_get_active_two+0x2d/0x50 #2: (s_active){++++.+}, at: [<c0580ebd>] sysfs_get_active_two+0x1d/0x50 stack backtrace: Pid: 2595, comm: cat Not tainted 2.6.33-tp-rc4 #2 Call Trace: [<c08b99ab>] ? printk+0x1d/0x22 [<c0487752>] print_circular_bug+0xc2/0xd0 [<c0489f84>] __lock_acquire+0x1054/0x1230 [<c0478d81>] ? sched_clock_cpu+0x121/0x180 [<c048a1ed>] lock_acquire+0x8d/0x110 [<facfa598>] ? show_statistics+0x48/0x100 [iwlagn] [<c08bb358>] __mutex_lock_common+0x58/0x470 [<facfa598>] ? show_statistics+0x48/0x100 [iwlagn] [<c08bb84a>] mutex_lock_nested+0x3a/0x50 [<facfa598>] ? show_statistics+0x48/0x100 [iwlagn] [<facfa598>] show_statistics+0x48/0x100 [iwlagn] [<c0580cf9>] ? sysfs_get_active+0x69/0xb0 [<facfa550>] ? show_statistics+0x0/0x100 [iwlagn] [<c070d219>] dev_attr_show+0x29/0x50 [<c057fecd>] sysfs_read_file+0xdd/0x190 [<c05ff314>] ? security_file_permission+0x14/0x20 [<c0528242>] ? rw_verify_area+0x62/0xd0 [<c052880f>] vfs_read+0x9f/0x190 [<c047745b>] ? up_read+0x1b/0x30 [<c057fdf0>] ? sysfs_read_file+0x0/0x190 [<c04af3b4>] ? audit_syscall_entry+0x1f4/0x220 [<c0528d22>] sys_read+0x42/0x70 [<c040ad9f>] sysenter_do_call+0x12/0x38 Signed-off-by:
Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by:
Reinette Chatre <reinette.chatre@intel.com>
7d47618a