boards/atmega256rfr2-xpro: update clock configuration/stdio baudrate + a test timeout

[aabadie]

Contribution description

This PR is improving the atmega256rfr2-xpro clock configuration by using the available external 16MHz oscillator. This allows to use a 115200 baudrate for stdio, which is fixing some timeout with the automatic tests. There is a commit that adds some doc about how to correctly setup the fuse bytes.

This PR also provides a timeout increase for tests/periph_gpio.

There are still failing automatic tests but they are not directly related to this board. My guess is that AVR support is not working in some cases, or some tests don't support this platform at all. I will report later in detail the problems I found.

Testing procedure

• A green Murdock at least
• Run the automatic tests on the boards (will report my results here later)

Issues/PRs references

Follow-up of #12639

[maribu]

This should also be updated with the new frequency:

/**
 * @name    xtimer configuration values
 *
 * Xtimer runs at 8MHz / 64 = 125kHz
 * @{
 */
#define XTIMER_DEV                  (0)
#define XTIMER_CHAN                 (0)
#define XTIMER_WIDTH                (16)
#define XTIMER_HZ                   (125000UL)
#define XTIMER_BACKOFF              (40)
/** @} */

The ATmega Arduinos which are clocked at 16 MHz are using 250 kHz (so also a clock divider of 64) for the timer. If you want to allow people testing with the internal oscillator instead of the external (as they maybe want to build a low part count board), you could go for:

#define XTIMER_HZ                   (CLOCK_CORECLOCK / 64)

That would yield the 250 kHz on 16 MHz and the current setting of 125 kHz for the internal oscillator running at 8MHz.

[aabadie]

This should also be updated with the new frequency:

Also updated by this PR: https://github.com/RIOT-OS/RIOT/pull/12649/files#diff-e605502ae9b812b7d1b514f864fcd682R41

Are you reviewing the right commit ?

[maribu]

Weird. Somehow only part of this PR were shown in the Browser. After I pressed F5 I now can see all changes.

[aabadie]

Somehow only part of this PR were shown in the Browser

It could be because I updated the branch several times (push forced) before clicking the open PR button.

[aabadie]

Here is the list of failing tests:
Failures during test:

tests/bitarithm_timings: looks like a context switching issue

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/bitarithm_timings test
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.


main(): This is RIOT! (Version: buildtest)
Start.
+ bitarithm_msb: 102096 iterations per second
Timeout in expect script at "child.expect('\+ bitarithm_lsb: \d+ iterations per second')" (tests/bitarithm_timings/tests/01-run.py:16)

tests/event_wait_timeout: looks like a context switching issue

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/event_wait_timeout test
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.
finished: 4/4 events and 1/4 timeouts recorded
[FAILED]


main(): This is RIOT! (Version: buildtest)
[START] event_wait_timeout test application.

finished: 4/4 events and 1/4 timeouts recorded
[FAILED]
Timeout in expect script at "child.expect_exact(u"[SUCCESS]")" (tests/event_wait_timeout/tests/01-run.py:15)

tests/gnrc_*: missing network configuration

- [tests/gnrc_ipv6_ext](tests/gnrc_ipv6_ext/test.failed)
- [tests/gnrc_ipv6_ext_frag](tests/gnrc_ipv6_ext_frag/test.failed)
- [tests/gnrc_ndp](tests/gnrc_ndp/test.failed)
- [tests/gnrc_rpl_srh](tests/gnrc_rpl_srh/test.failed)
- [tests/gnrc_sock_dns](tests/gnrc_sock_dns/test.failed)
- [tests/gnrc_tcp](tests/gnrc_tcp/test.failed)

tests/gnrc_ipv6_fwd_w_sub: maybe again a missing network configuration ?

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/gnrc_ipv6_fwd_w_sub test
run_test
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.


main(): This is RIOT! (Version: buildtest)
> Timeout in expect script at "child.expect(r"Forwarded Ethernet frame:")" (tests/gnrc_ipv6_fwd_w_sub/tests/01-run.py:16)

tests/gnrc_ndp: no idea of the problems

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/gnrc_ndp test
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.


main(): This is RIOT! (Version: buildtest)
............................
tests.test_nbr_sol_send__pktbuf_full4 (tests/gnrc_ndp/main.c 552) gnrc_pktbuf_is_empty()
.
tests.test_nbr_adv_send__foreign_tgt_unspecified_dst_no_supply_tl2a_no_ext_opts (tests/gnrc_ndp/main.c 571) src != pkt->src
.
tests.test_nbr_adv_send__foreign_tgt_unspecified_dst_no_supply_tl2a_ext_opts (tests/gnrc_ndp/main.c 568) exp -1 was 0
.
tests.test_nbr_adv_send__foreign_tgt_unspecified_dst_supply_tl2a_no_ext_opts (tests/gnrc_ndp/main.c 568) exp 5 was 0
.
tests.test_nbr_adv_send__foreign_tgt_unspecified_dst_supply_tl2a_ext_opts (tests/gnrc_ndp/main.c 568) exp -1 was 0
.
tests.test_nbr_adv_send__foreign_tgt_specified_dst_no_supply_tl2a_no_ext_opts (tests/gnrc_ndp/main.c 574) exp 2 was 8
.
tests.test_nbr_adv_send__foreign_tgt_specified_dst_no_supply_tl2a_ext_opts (tests/gnrc_ndp/main.c 568) exp -1 was 0
.
tests.test_nbr_adv_send__foreign_tgt_specified_dst_supply_tl2a_no_ext_opts (tests/gnrc_ndp/main.c 568) exp -1 was 104
.
tests.test_nbr_adv_send__foreign_tgt_specified_dst_supply_tl2a_ext_opts (tests/gnrc_ndp/main.c 568) exp -1 was 0
.
tests.test_nbr_adv_send__src_tgt_unspecified_dst_no_supply_tl2a_no_ext_opts (tests/gnrc_ndp/main.c 568) exp 5 was 0
.
tests.test_nbr_adv_send__src_tgt_unspecified_dst_no_supply_tl2a_ext_opts (tests/gnrc_ndp/main.c 568) exp -1 was 0
.
tests.test_nbr_adv_send__src_tgt_unspecified_dst_supply_tl2a_no_ext_opts (tests/gnrc_ndp/main.c 568) exp 5 was 0
.
tests.test_nbr_adv_send__src_tgt_unspecified_dst_supply_tl2a_ext_opts (tests/gnrc_ndp/main.c 568) exp -1 was 0
.
tests.test_nbr_adv_send__src_tgt_specified_dst_no_supply_tl2a_no_ext_opts (tests/gnrc_ndp/main.c 574) exp 2 was 8
.
tests.test_nbr_adv_send__src_tgt_specified_dst_no_supply_tl2a_ext_opts (tests/gnrc_ndp/main.c 568) exp -1 was 0
.
tests.test_nbr_adv_send__src_tgt_specified_dst_supply_tl2a_no_ext_opts (tests/gnrc_ndp/main.c 568) exp -1 was 104
.
tests.test_nbr_adv_send__src_tgt_specified_dst_supply_tl2a_ext_opts (tests/gnrc_ndp/main.c 568) exp -1 was 0
....
tests.test_nbr_adv_send__pktbuf_full4 (tests/gnrc_ndp/main.c 763) gnrc_pktbuf_is_empty()
.
tests.test_rtr_sol_send__dst_NULL (tests/gnrc_ndp/main.c 784) exp 2 was 0
.
tests.test_rtr_sol_send__dst_local (tests/gnrc_ndp/main.c 780) src != pkt->src
.....
tests.test_rtr_sol_send__pktbuf_full4 (tests/gnrc_ndp/main.c 886) gnrc_pktbuf_is_empty()
.
tests.test_rtr_adv_send__src_NULL_dst_NULL_no_fin_no_ext_opts (tests/gnrc_ndp/main.c 903) exp -1 was 0
.
tests.test_rtr_adv_send__src_NULL_dst_NULL_no_fin_ext_opts (tests/gnrc_ndp/main.c 911) dst != pkt->dst
.
tests.test_rtr_adv_send__src_NULL_dst_NULL_fin_no_ext_opts (tests/gnrc_ndp/main.c 903) exp 5 was 8647
.
tests.test_rtr_adv_send__src_NULL_dst_NULL_fin_ext_opts (tests/gnrc_ndp/main.c 903) exp -1 was 0
.
tests.test_rtr_adv_send__src_NULL_dst_no_fin_no_ext_opts (tests/gnrc_ndp/main.c 903) exp 5 was 8647
.
tests.test_rtr_adv_send__src_NULL_dst_no_fin_ext_opts (tests/gnrc_ndp/main.c 903) exp -1 was 0
.
tests.test_rtr_adv_send__src_NULL_dst_fin_no_ext_opts (tests/gnrc_ndp/main.c 903) exp 5 was 8647
.
tests.test_rtr_adv_send__src_NULL_dst_fin_ext_opts (tests/gnrc_ndp/main.c 903) exp -1 was 0
.
tests.test_rtr_adv_send__src_dst_NULL_no_fin_no_ext_opts (tests/gnrc_ndp/main.c 903) exp 5 was 8647
.
tests.test_rtr_adv_send__src_dst_NULL_no_fin_ext_opts (tests/gnrc_ndp/main.c 903) exp -1 was 0
.
tests.test_rtr_adv_send__src_dst_NULL_fin_no_ext_opts (tests/gnrc_ndp/main.c 903) exp 5 was 8647
.
tests.test_rtr_adv_send__src_dst_NULL_fin_ext_opts (tests/gnrc_ndp/main.c 903) exp -1 was 0
.
tests.test_rtr_adv_send__src_dst_no_fin_no_ext_opts (tests/gnrc_ndp/main.c 903) exp 5 was 8647
.
tests.test_rtr_adv_send__src_dst_no_fin_ext_opts (tests/gnrc_ndp/main.c 903) exp -1 was 0
.
tests.test_rtr_adv_send__src_dst_fin_no_ext_opts (tests/gnrc_ndp/main.c 903) exp 5 was 8647
.
tests.test_rtr_adv_send__src_dst_fin_ext_opts (tests/gnrc_ndp/main.c 903) exp -1 was 0
....
tests.test_rtr_adv_send__pktbuf_full4 (tests/gnrc_ndp/main.c 1085) gnrc_pktbuf_is_empty()

run 75 failures 38
Timeout in expect script at "child.expect(r"OK \(\d+ tests\)")" (tests/gnrc_ndp/tests/01-run.py:16)

tests/isr_yield_higher: seems like a context switch issue again

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/isr_yield_higher test
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.
first thread done


main(): This is RIOT! (Version: buildtest)
first thread started
timer triggered
TEST FAILED
first thread done
Timeout in expect script at "child.expect('TEST SUCCESSFUL')" (tests/isr_yield_higher/tests/test.py:17)

tests/libfixmath: does it work on 8 bit architectures?

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/libfixmath test
r
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.


main(): This is RIOT! (Version: buildtest)
Help: Press s to start test, r to print it is ready
r
READY
s
START
Unary.
abs(-10.0000) = 10.0000
sq(-10.0000) = 100.0000
atan(-10.0000) = -1.4989

[...]

min(-4.2500, 1.2500) = -4.2500
max(-4.2500, 1.2500) = 1.2500
add(-4.2500, 1.5000) = -2.7500
sub(-4.2500, 1.5000) = -5.7500
mul(-4.2500, 1.5000) = -5.8750
Timeout in expect script at "child.expect('{}\(-?\d+.\d+\) = -?\d+.\d+'.format(op_name))" (tests/libfixmath/tests/01-run.py:14)

tests/periph_eeprom: shell is not ready when launching the test but it works when run manually (or when blocking until the application is ready)

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/periph_eeprom test
test
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.


main(): This is RIOT! (Version: buildtest)
EEPROM read write test

Please refer to the README.md for more details

EEPROM size:		8192
> Timeout in expect script at "child.expect('SUCCESS')" (tests/periph_eeprom/tests/01-run.py:15)

tests/pipe: could be several things, maybe again context switching ?

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/pipe test
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.


main(): This is RIOT! (Version: buildtest)
Start.
Middle read: <Middle read: <End read: <End read: <Middle read: <Middle read: <End read: <End read: <Middle read: <Middle read: <End read: <End read: <Middle read: <Middle read: <Main done.
End read: <End read: <Middle read: <Middle done.
End read: <End done.
Timeout in expect script at "child.expect_exact('Middle read: <ABCD> [0:4]')" (tests/pipe/tests/01-run.py:9)

tests/pkg_jsmn: Ouput is clearly invalid. does the package support AVR ?

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/pkg_jsmn test
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.


main(): This is RIOT! (Version: buildtest)
- User: - Admin: - UID: - Groups:
Timeout in expect script at "child.expect_exact('- User: johndoe')" (tests/pkg_jsmn/tests/01-run.py:8)

tests/pkg_lora-serialization: seems like a conversion issue. Maybe this package is not compatible with 8bit architecture?

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/pkg_lora-serialization test
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.


main(): This is RIOT! (Version: buildtest)
Lora Serialization test

Test 1
Temperature and humidity
---------------------------------
- Writing temperature: 80.12
- Writing humidity: 99.99
- Encoded:  1f 4c 0f 27
- Expected: 1f 4c 0f 27
---------------------------------
SUCCESS

Test 2
Coordinates and unix time
---------------------------------
- Writing coordinates: -33.905052, 151.26641
- Writing unix time: 1467632413
- Encoded:  64 a6 ff ff 60 24 00 00 1d 4b 00 00
- Expected: 64 a6 fa fd 6a 24 04 09 1d 4b 7a 57
---------------------------------
FAILED
Timeout in expect script at "child.expect('SUCCESS')" (tests/pkg_lora-serialization/tests/01-run.py:12)

tests/pkg_micro-ecc: reboot loop

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/pkg_micro-ecc test
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.main(): This is RIOT! (Version: buildtest)
micro-ecc compiled!
Testing 16 random private key pairs and signature without using HWRNG
.

main(): This is RIOT! (Version: buildtest)
micro-ecc compiled!
Timeout in expect script at "child.expect_exact('................ done with 0 error(s)')" (tests/pkg_micro-ecc/tests/01-run.py:16)

tests/ps_schedstatistics: context switching issue again ?

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/ps_schedstatistics test
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.


main(): This is RIOT! (Version: buildtest)
Creating thread #0, next=1
Creating thread #1, next=2
Creating thread #2, next=3
Creating thread #3, next=4
Creating thread #4, next=0

Timeout in expect script at "child.expect('>')" (tests/ps_schedstatistics/tests/01-run.py:43)

tests/stdin: application not ready when sending the character to UART

make RIOT_CI_BUILD=1 CC_NOCOLOR=1 --no-print-directory -C ./tests/stdin test
O
/work/riot/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200" --noprefix --no-repeat-command-on-empty-line
Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.


main(): This is RIOT! (Version: buildtest)
Timeout in expect script at "child.expect_exact('You entered \'O\'')" (tests/stdin/tests/01-run.py:15)

[maribu]

tests/bitarithm_timings: looks like a context switching issue

This also fails on other ATmegas.

Adding a simple print before the tests starts

diff --git a/tests/bitarithm_timings/main.c b/tests/bitarithm_timings/main.c
index c718104d23..a5db16f7b3 100644
--- a/tests/bitarithm_timings/main.c
+++ b/tests/bitarithm_timings/main.c
@@ -54,6 +54,7 @@ static void run_test(const char *name, unsigned (*test)(unsigned))
     xtimer.callback = callback;
     xtimer.arg = (void *) &done;
 
+    puts("Starting");
     xtimer_set(&xtimer, TIMEOUT);
 
     do {

yields the following:

/home/maribu/Repos/software/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyUSB0" -b "9600" 
Twisted not available, please install it if you want to use pyterm's JSON capabilities
2019-11-06 10:31:12,250 # Connect to serial port /dev/ttyUSB0
Welcome to pyterm!
Type '/exit' to exit.
2019-11-06 10:31:13,262 # This is RIOT! (Version: 2020.01-devel-518-gd02ed)
2019-11-06 10:31:13,263 # Start.
2019-11-06 10:31:13,264 # Starting
2019-11-06 10:31:17,311 # + bitarithm_msb: 49264 iterations per second
2019-11-06 10:31:17,320 # Starting
2019-11-06 10:31:46,360 # Exiting Pyterm

So bitarithm_lsb() does actually start, but doesn't conclude. I'll investigate further.

However: Those failures seem to be unrelated to this PR. So I'd say this PR can go in and the issues should be tackled in different PRs. (Would you be so kind to open an issue with the list of failing tests? Feel free to assign me some of them to investigate. I already started with bitarithm_timings.)

[maribu]

Hmm. With ENABLE_DEBUG set to 1 in xtimer_core.c it works

[aabadie]

I opened #12651 about the failing tests on this board (and more generally on AVR).

[aabadie]

Those failures seem to be unrelated to this PR. So I'd say this PR can go in and the issues should be tackled in different PRs

I agree, so I think we are good here @maribu.

[maribu]

ACK

[maribu]

One moment, please before merging.

[maribu]

OK, the default baud rate of 115200 is defined in sys/include/stdio_uart.h. I just wanted to make sure where it is defined to make sure whether an #include would be needed when not defining it.