Do device node auto cleanup in drivers/soc/ti/

Do "struct device_node" auto cleanup in soc/ti/. This patch series takes care of all the cases where this is possible.

Thanks Nishanth for the review on the previous round.

v3:

	https://lore.kernel.org/linux-arm-kernel/[email protected]/

Changes since v3:
- s/seperate/separate in the commit msgs througout.
- 1 / 4 - "return ret" instead of "goto rpm_put" as pm_runtime was never enabled by the time we call pruss_of_setup_memories() in case we return on err.

v2:

	https://lore.kernel.org/linux-arm-kernel/[email protected]/

Changes since v2:
- Split v2 1/3 into v3 1/4 and v3 2/4. The memory setup code is seperated out of the pruss_probe() function and put into 1/4 and the actual conversion to auto cleanup is done in 2/4.
- Replace dev_err() with dev_err_probe() in the code paths touched.

v1:

	https://lore.kernel.org/linux-arm-kernel/[email protected]/

Changes since v1:
- Refactor code so that it the scope of the pointers touched is reduced, making the code look more clean.
- The above also the side-effect of fixing the errors that clang emitted (but my local version of gcc didn't) for PATCH 2/3 during v1.
Kousik Sanagavarapu (4):
soc: ti: pruss: factor out memories setup
soc: ti: pruss: do device_node auto cleanup
soc: ti: knav_qmss_queue: do device_node auto cleanup
soc: ti: pm33xx: do device_node auto cleanup

 drivers/soc/ti/knav_qmss_queue.c | 100 +++++++++---------
 drivers/soc/ti/pm33xx.c          |  52 ++++-----
 drivers/soc/ti/pruss.c           | 176 ++++++++++++++-----------------
 3 files changed, 155 insertions(+), 173 deletions(-)

2.46.0.290.g4f02c709e8.dirty