msm_fb: HDMI: Fix lockdep warning

Locks must always be taken in the same order to prevent AB-BA
scenarios from occuring:

======================================================
[ INFO: possible circular locking dependency detected ]
3.4.0-g55e6e49-00052-g71783d6-dirty #3258 Tainted: G        W
-------------------------------------------------------
kworker/0:1/15 is trying to acquire lock:
 (hdmi_msm_state_mutex){+.+...}, at: [<c0340e98>]
 hdmi_msm_hpd_state_work+0x78/0x2d8

but task is already holding lock:
 (external_common_state_hpd_mutex){+.+...}, at: [<c0340e8c>]
 hdmi_msm_hpd_state_work+0x6c/0x2d8

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #1 (external_common_state_hpd_mutex){+.+...}:
       [<c00cd744>] __lock_acquire+0x950/0xa10
       [<c00cd990>] lock_acquire+0x18c/0x1e8
       [<c080c248>] mutex_lock_nested+0x68/0x3c4
       [<c0341510>] hdmi_msm_hpd_on.clone.0+0x230/0x2c4
       [<c07e5dfc>] hdmi_msm_probe+0x378/0x588
       [<c03dcc00>] platform_drv_probe+0x18/0x1c
       [<c03db920>] driver_probe_device+0x148/0x334
       [<c03d9f5c>] bus_for_each_drv+0x48/0x84
       [<c03db7ac>] device_attach+0x78/0xa4
       [<c03daa4c>] bus_probe_device+0x28/0x9c
       [<c03d97e4>] device_add+0x438/0x5e0
       [<c03dd13c>] platform_device_add+0x13c/0x1a0
       [<c0d255d8>] hdmi_msm_init+0x218/0x2d0
       [<c0008708>] do_one_initcall+0x94/0x168
       [<c0d002fc>] kernel_init+0xe4/0x1b8
       [<c000f4f8>] kernel_thread_exit+0x0/0x8

-> #0 (hdmi_msm_state_mutex){+.+...}:
       [<c00cc6b8>] validate_chain+0x944/0x1080
       [<c00cd744>] __lock_acquire+0x950/0xa10
       [<c00cd990>] lock_acquire+0x18c/0x1e8
       [<c080c248>] mutex_lock_nested+0x68/0x3c4
       [<c0340e98>] hdmi_msm_hpd_state_work+0x78/0x2d8
       [<c0097420>] process_one_work+0x37c/0x694
       [<c0099978>] worker_thread+0x22c/0x3bc
       [<c009ecdc>] kthread+0x8c/0x9c
       [<c000f4f8>] kernel_thread_exit+0x0/0x8

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(external_common_state_hpd_mutex);
                               lock(hdmi_msm_state_mutex);
                               lock(external_common_state_hpd_mutex);
  lock(hdmi_msm_state_mutex);

 *** DEADLOCK ***

3 locks held by kworker/0:1/15:
 #0:  (hdmi_hdcp){.+.+..}, at: [<c00972b8>] process_one_work+0x214/0x694
 #1:  ((&hdmi_msm_state->hpd_state_work)){+.+...}, at: [<c00972b8>]
 process_one_work+0x214/0x694
 #2:  (external_common_state_hpd_mutex){+.+...}, at: [<c0340e8c>]
 hdmi_msm_hpd_state_work+0x6c/0x2d8

stack backtrace:
[<c001521c>] (unwind_backtrace+0x0/0x12c) from [<c00cbb14>]
(print_circular_bug+0x288/0x2d4)
[<c00cbb14>] (print_circular_bug+0x288/0x2d4) from [<c00cc6b8>]
(validate_chain+0x944/0x1080)
[<c00cc6b8>] (validate_chain+0x944/0x1080) from [<c00cd744>]
(__lock_acquire+0x950/0xa10)
[<c00cd744>] (__lock_acquire+0x950/0xa10) from [<c00cd990>]
(lock_acquire+0x18c/0x1e8)
[<c00cd990>] (lock_acquire+0x18c/0x1e8) from [<c080c248>]
(mutex_lock_nested+0x68/0x3c4)
[<c080c248>] (mutex_lock_nested+0x68/0x3c4) from [<c0340e98>]
(hdmi_msm_hpd_state_work+0x78/0x2d8)
[<c0340e98>] (hdmi_msm_hpd_state_work+0x78/0x2d8) from [<c0097420>]
(process_one_work+0x37c/0x694)
[<c0097420>] (process_one_work+0x37c/0x694) from [<c0099978>]
(worker_thread+0x22c/0x3bc)
[<c0099978>] (worker_thread+0x22c/0x3bc) from [<c009ecdc>]
(kthread+0x8c/0x9c)
[<c009ecdc>] (kthread+0x8c/0x9c) from [<c000f4f8>]
(kernel_thread_exit+0x0/0x8)

Change-Id: I251049edbc97a5fe0e6c67801cf472fdc48722a7
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>