Add driver API reentrancy support to UART shim drivers

[zephyrbot]

Reported by Kuo-Lang Tseng:

Global data and device IO registers being set by the driver API can be messed up in the middle of the API call when a concurrent call (from other fiber/task) to the same API is made.

(Imported from Jira ZEP-454)

[zephyrbot]

by Kuo-Lang Tseng:

This is on hold until GH-2047 is root-caused.

[zephyrbot]

by ethan gao:

It's being blocked by issue GH-2047 now and QA will try to verify it as soon as it's available !

[zephyrbot]

by Mark Linkmeyer:

ethan gao , is the verification of this story still blocked by GH-2047? I see that Kuo-Lang Tseng has lowered the priority of GH-2047 to medium with this comment, "Updating the priority to medium as the patch reverting the reentrancy support was merged to unblock the issue for now. Baohong will debug the issue when he is done with other high bugs." If GH-2047 is truly blocking the verification of this story (which is needed for the 1.5 release), then it seems to me the priority of GH-2047 needs to be raised back up so it gets fixed. Please advise.
Sharron LIU Anagha Jamthe

[zephyrbot]

by Kuo-Lang Tseng:

Mark Linkmeyer GH-2047 requires more investigation and debug, before we can pull the reentrancy function back to UART shim driver. I wuild suggest we move both GH-2047 and GH-2003 to 1.6 so we have more time to investigate the issue.

[zephyrbot]

by Gajinder Vij:

The attempt to make the UART driver re-entrant conflicts with the existing Bluetooth driver. As UART re-entrancy is not a critical feature for 1.5, the feature has been reverted and a solution will be re-proposed for the next Zephyr release.

[zephyrbot]

by Mark Linkmeyer:

Gajinder Vij , I think you meant to uncheck v1.5.0 in the Fix Version/s field, so I'm doing it for you now.

[zephyrbot]

Related to GH-2047