Resolve TLS handshake failing issue for REALTEK_RTL8195AM

[Archcady]

The issue this PR trying to resolve, relating to REALTEK_RTL8195AM wifi threads priority, is making blocking TCP socket failing.

[Archcady]

This shall resolve TSL handshake issue mentioned in #4700 #4683

[0xc0170]

@Archcady Can you please rebase, the earlier PR caused the conflicts here

[Archcady]

Hi @0xc0170 , PR rebased, thanks.

[0xc0170]

/morph test

[JanneKiiskila]

Unforunately the mbed-os-example-client still doesn't connect over TCP.

ROM Version: 0.3

Build ToolChain Version: gcc v==============
Check boot type form eFuse
SPI Initial
Image1 length: 0x2d64, Image Addr: 0x10000bc8
Image1 Validate OK, Going jump to Image1
K
K
K
K
K

Starting mbed Client example
[EasyConnect] IPv4 mode
[EasyConnect] Using WiFi (RTW)
[EasyConnect] Connecting to WiFi jannek.iki.fi
[EasyConnect] Connected to Network successfully
[EasyConnect] MAC address f0:03:8c:e7:a9:cd
[EasyConnect] IP address 192.168.1.149

SOCKET_MODE : TCP
Connecting to coap://api.connector.mbed.com:5684
simulate button_click, device not registered
simulate button_click, device not registered

This branch + cherry-picked the enable_sdram fix cherry-picked on top of this.

* d2113fe Yuguo Zou - enable sdram usage of REALTEK_RTL8195AM
* 2ec8b04 Yuguo Zou - Resolve TLS handshake failing issue
*   692d905 Martin Kojtal - Merge pull request #4725 from theotherjimmy/extra_targets_exporters
|\  
| * e6bdc32 Jimmy Brisson - Allow custom_targets.json exporting
| * 81446f6 Jimmy Brisson - Generate exporter suppored lists when needed
* |   c295187 Martin Kojtal - Merge pull request #4634 from nvlsianpu/bugfix/issue_4357_I2C_SPI_simultaneously
|\ \  
| * | 3236336 Andrzej Puzdrowski - nrf5 spi_api - coding style enhancement
| * | c1870af Andrzej Puzdrowski - Bugfix: #4357 simultaneous using of I2C and SPI.
* | |   b685fe8 Martin Kojtal - Merge pull request #4592 from KariHaapalehto/connect_timeout
|\ \ \  
| * | | 3c44a20 Kari - [ONME-3089] - Adjust lowpan ND interface connect timeout
* | | |   1849370 Martin Kojtal - Merge pull request #4652 from geky/fat-fix-unaligned-ioctl

[JanneKiiskila]

On 2nd trial - WiFi connection failed.

Starting mbed Client example
[EasyConnect] IPv4 mode
[EasyConnect] Using WiFi (RTW)
[EasyConnect] Connecting to WiFi jannek.iki.fi
failed: -1
[EasyConnect] MAC address f0:03:8c:e7:a9:cd
[EasyConnect] Connection to Network Failed -3004!

[JanneKiiskila]

With traces enabled I got it to log, so some sort of timing issue perhaps?
However, the device seemed to crash when one tried some operations on it via connector.mbed.com API Console.

Also, even with traces enabled the success rate over TCP is still not 100%, 2nd trial failed already.

[theotherjimmy]

/morph test

[Archcady]

@JanneKiiskila, yes this could be timing issue, putting code to SDRAM really slow down the performance. I'll also trace the code and maybe move some time-intensive code back from SDRAM to BDRAM in this PR.

[mbed-bot]

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

/morph test

Output

mbed Build Number: 833

All builds and test passed!

[0xc0170]

@JanneKiiskila, yes this could be timing issue, putting code to SDRAM really slow down the performance. I'll also trace the code and maybe move some time-intensive code back from SDRAM to BDRAM in this PR.

@Archcady what would you propose? this is actually fixing an issue, and the problems @JanneKiiskila is experiencing are related to different issue, thus should not block this PR?

[Archcady]

@0xc0170 I think this PR should not be blocked. And for the problems from @JanneKiiskila, I'll submit another to fix.