[PATCH 0/5] serial: imx: fixes

On Thu, May 29, 2014 at 06:58:55AM +0200, Dirk Behme wrote:

Am 28.05.2014 21:34, schrieb Greg KH:

quoted

On Fri, May 09, 2014 at 04:19:43PM +0100, dean_jenkins at mentor.com wrote:

quoted

From: Dean Jenkins <redacted>

The following set of patches were tested on an i.MX6 multi-core platform.

Patches 1 to 4 were tested using a Bluetooth BCSP UART connection at 115.2kbps
and 4Mbps. The fixes were needed because the i.MX6 UART driver caused spinlock
deadlocks with the HCI TTY layer.

Patch 5 improves the i.MX6 UART driver for use with kdb.

Patch 1: serial: imx: remove unneeded imx_transmit_buffer() from imx_start_tx()
Patch 2: serial: imx: remove uart_write_wakeup() from imx_transmit_buffer()
Patch 3: serial: imx: avoid spinlock recursion deadlock
Patch 4: serial: imx: move imx_transmit_buffer() into imx_txint()
Patch 5: serial: imx: clean up imx_poll_get_char()

The patches are based off 3.15-rc4.

----------------------------------------------------------------
Andy Lowe (1):
      serial: imx: avoid spinlock recursion deadlock

Dirk Behme (4):
      serial: imx: remove unneeded imx_transmit_buffer() from imx_start_tx()
      serial: imx: remove uart_write_wakeup() from imx_transmit_buffer()
      serial: imx: move imx_transmit_buffer() into imx_txint()
      serial: imx: clean up imx_poll_get_char()

 drivers/tty/serial/imx.c |   73 ++++++++++++++--------------------------------
 1 file changed, 22 insertions(+), 51 deletions(-)

Given the rather confusing threads based on these patches, I don't know
what to apply.

Dean, can you respin what patches you want applied

Well, we think patches #1-#4 are correct and should be applied, but
Huang has the opinion that this is fixed already in the BT stack. If
Huang doesn't change his mind, then let's drop #1-#4.

Yes, the BT stack has fixed it. :)

we can drop the #1-#4.	   

thanks
Huang Shijie