On 09/08/2014 07:43 PM, Greg Kroah-Hartman wrote:

On Mon, Sep 08, 2014 at 07:33:51PM -0400, Peter Hurley wrote:

quoted

On 09/08/2014 07:24 PM, Greg Kroah-Hartman wrote:

quoted

On Tue, Sep 02, 2014 at 05:39:09PM -0400, Peter Hurley wrote:

quoted

Hi Greg,

This patch series reworks the internals of tty and serial flow control to
fix multiple races in both START/STOP flow control and RTS/CTS flow control.

The main changes in this series are:
Patch 1
   Backs out the UPF_HARD_FLOW kludge for 8250. This revert should be
   for mainline and -next

Patches 3-7
   Fixes to x_char handling (ie., sending START/STOP) both in the serial
   core and to several UART drivers

Patches 10-15
   Rename and move ASYNC_CTS_FLOW and ASYNC_CHECK_CD statuses into
   a private field for the serial core, and add helper functions to test
   those statuses for UART drivers.

   This is a necessary step toward making tty port->flags SMP-safe, without
   introducing a new lock.

I stopped applying at patch 13, due to the objections to that patch.
Please redo the rest of the series and resend.

I'm going to wait for the fallout from patch 21 to shake out first.

Why, the earlier patches in this series should be ok to go, right?  They
don't depend on 21 to be "correct".

Oh, you want me to redo just 14 through 20, and resend that?

Regards,
Peter Hurley