Add a pthread barrier to synchronize the Rx/Tx threads and remove the
Tx pipe as it becomes useless.

This fixes the issue when the serial test is run inside bit.

BCS 746-000147

Add the possibility to use a TCP connection for the echo client.
The code is well prepared for this and you simply need to activate
the correct parameter parsing and connect the socket to the server.

Additionally to activating the TCP connection we have to wait for the
whole package to arrive.

As a refactoring step sue a simple division instead of a conditional
shift to get a loop boundary.

BCS 746-000127

The ```gettimeofday()``` function is defined in ```sys/time.h```. Thus
include it.

BCS 746-000125

For concurrency and synchronization issues we need a new approach.
The select() function and additional pipes should be sufficient to
timeout and synchronize the TX source and Rx sink thread.

* Add RS485 echo test for our EMC verification.
* The test is more flexible than the legacy RS485 ping pong test and will
  always recover from transmission errors due to EMI.
* The sent data includes a hard-coded 4-byte header and a 4-byte data length
  field, which is parsed by the receiver. The receiver always scans the
  received data for the header bytes in order to detect a new message.

BCS 746-000125

BCS 746-000120

can: Remove the usleep

See merge request garz-fricke/tools/gf-emc-test-suite!3

The usleep did not fix the broken Microchip MCP25XXFD CAN driver.
To resolve the issue we need to update the driver, thus remove the
sleep.

can: Move mutex unlock after signal.

See merge request garz-fricke/tools/gf-emc-test-suite!2

The pthread_cond_signal should be called with the mutex locked.
Thus move the unlock after the signal.

Add an usleep(1000) before closing the CAN socket. Some drivers have
some kind of race condition sending data and closing the socket.

iptest assumed sizeof(unsigned long) == 4, which fails on aarch64. fixed
this.

ltp-guf-test: Remove version generation as it is broken

See merge request garz-fricke/tools/gf-emc-test-suite!1

BCS 746-000110

* Our hardware supports more baud rates than the test did. Add support for baud rates above 230400.

Change-Id: Ic1f0ad2244b84881ef0a8bb952d3ec27825cd569

* disable spread spectrum in preparation of EMV test run

Change-Id: I8906db5d5fc11cb3bbc9c74f0aabf32462ea6bff

* enable bball_sdl in nallino runtest file to improve performance

Change-Id: Iba9dc03d850819c7b31d715722f49b7c3dfd9bb7

* added calls to shutdown customer application
* moved mount-call to the end to gain some time (USB-power on to successful mount takes variable amounts of time)
* added call to disable screensaver

Change-Id: Iacc91ee495689e2e8f2d93a31cb52cb31f9d5040
Note: This setup-script should be called on each boot before starting the @bit@ testsuite but only after the customer application has started completely (dialog to chose language is visible on screen). If the script is started before the customer app is fully started, some calls may not work, the USB-stick may fail to mount continuously, the @bit@ runtimes may increase significantly or the system may reboot automatically after some time.

* The @sanvito-custom@ testscript for the BELD OS still had some issues:
** Load on the 3 Ethernet and the 4 slow RS485 interfaces was too high, making the test too sensitive to scheduling variances and causing the test to break the 1s-raster. Reduced amount of data for both.
** Removed the @i2c@ test for the proximity sensor. While physically still present, it won't be used or supported anymore (only ambient light will be used)
** increased load on fast RS485 port and GPU (@bball@) because these tests still allowed higher data-rates without performance or temperature impact
** increased load on Analog-I/O test to 10 reads of 10 median-filtered samples, each, as specified by customer as application work-load (also still possible without impact on performance or temperature)

Change-Id: I635cc5c3652d01d90ff7380acbcbf231b0fe6d5e

* the new @sanvito-custom-setup.sh@ script was accidently commited with Windows line-breaks. Fixed now.

Change-Id: I5e855c26cf8ac82313155e945802cf0361e8c696

* added "sanvito-custom-setup.sh". This must be @source@d on each login before starting the test
* added LED test steps
* fixed SD-Card test

Change-Id: I6c2adc905161a72df8cfa9c50d230cd119271f32

* (temporarily) added contents of @/etc/config/network@ for properly working ethernet tests
* added @MMC_write@-test for µSD-Card

Note:
* LED tests are still not working
* USB test is still unstable due to timing issues when enabling the port and mounting the stick

Change-Id: I60db484c5119c013b1db970a646435400fcb820e

* @aio@ tool can now use device-names of the form
** e.g. @iio:device0/in_voltage_raw@ to specify devices relative to @/sys/bus/iio/devices/@, as before
** or e.g. @/dev/ain1/in_voltage_raw@ to specify absolute device names

Change-Id: Id0e38ea614e03b3f3363209a5ce3af21c5a7fb0a

* added new parameters @-f@ and @-g@ to override the IIO scale values reported by the system

Change-Id: I02e4bce2a4ab7a22bf9be54470890857e811ab70

* enabled @bball@-test again, because @bball@ is rumored to be working in the next @BELD@ OS

Change-Id: Ib10f35726317b600c97563b3fc020567dfc62942

* changed default input-filter for @aio@ test from 20 dummy reads to 5-sample median filter

Change-Id: I6081186ea624ca21a36e5f9cf46edabbb6053e24

* renamed "sanvito-custom.txt" to "sanvito-custom"

Change-Id: I5cb2b87d70d8c13f57c1d8059ccda53c4a69a1ed

* added option @-m@ to apply a median filter to the input
* added option @-a@ to apply an average filter to the input

Change-Id: Ic8a6e7de08f4280de418f12f7126da003b395e26
Note: The new option are mutually exclusive with each other and the @-r@ (dummy read) option

* added new SANVITO script for the custom BELD-OS

Note: still not fully working on BELD v0.3 snapshot because some OS-changes are required
* @aio@ fails often due to Kernel bug in IIO SysFS-interface (DAC-output not working for output values >= 100)
* @mmc@ test not working, because not yet supported by BELD OS v0.3 (supposed to be already fixed in newer versions)
* @fingerpaint@ not working, because of Qt-version conflict in BELD OS v0.3 (supposed to be already fixed in newer versions)
* @touch@ not working, because no driver support in BELD OS v0.3 (will hopefully be fixed with next BELD release, as well)
* @bball@ still crashing on all BELD versions (even new ones). Reason still unknown.

Change-Id: I6989010ef49265ba2e80715d94a0c79bbdba818f

* added new "-d" parameter to specify a delay (in milliseconds) between setting the output and reading the input to allow for slower hardware (esp. galvanically isolated DIOs).

Change-Id: I982196b1d57584d6d069e973a8dc1faa675a1c36

* renamed the "-v"-parameter to specify the value to be set/expected to "-e" instead to fix a parameter collision with "-v" for verbose output.
* replace usage of "-v"-paramter in runtest files with new "-e"-paramter

Change-Id: I9c31d2d3840ce9e1a1ba0baa1f5fecf76188fa5b

* renamed "-i"-parameter to set the IIO input-device to "-s" to fix collision with LTP "-i <iterations>"-parameter

Change-Id: I7d1904c5ea28ef8c6f400f9c612d4d421955da9c

Change-Id: I14f42da8aec76165600ce95ba244622d6847c115

Change-Id: Ie7b1d91a512a09801d07500632eff06a566ab8ef

Change-Id: Id41988386ac73ccc112cbada23c4e770bd3eb410

Change-Id: I716a3423e0a198ad7c0f8b96c796c37a657eaa45

Change-Id: If8e2e46c473208194e9125fe34d08716006d7a76

Change-Id: I2829521e20d8bbee230b8fec13e646bcd8acc72a

Change-Id: Iee6a519d469b596210eb82dc1425f1b032464d29

Change-Id: I302bf87d02a3192faad1fd23b3bc29a669e07566