make check error on arm64 Fedora 39

ElementsProject / lightning

Core Lightning — Lightning Network implementation focusing on spec compliance and performance

Other

2.81k stars 887 forks source link

make check error on arm64 Fedora 39 #7030

Open grubles opened 7 months ago

grubles commented 7 months ago

Fedora 39 on arm64. Occurs with both v23.11.2 and master using./configure --enable-rust.

VALGRIND=1 valgrind -q --error-exitcode=7 --track-origins=yes --leak-check=full --show-reachable=yes --errors-for-leak-kinds=all common/test/run-tal_arr_randomize > /dev/null
==151138== Source and destination overlap in memcpy(0x4d69f08, 0x4d69f08, 8)
==151138==    at 0x48CB68C: __GI_memcpy (vg_replace_strmem.c:1147)
==151138==    by 0x41B50B: tal_arr_randomize_ (pseudorand.c:84)
==151138==    by 0x41BB07: main (run-tal_arr_randomize.c:166)
==151138== 
make: *** [Makefile:750: unittest/common/test/run-tal_arr_randomize] Error 7

vincenzopalazzo commented 7 months ago

--enable-rust it is needed? looks like that the tests is happening on C side

grubles commented 7 months ago

No it's not AFAIK. I just wanted to show there isn't anything special about my build.

vincenzopalazzo commented 7 months ago

Mh currently I am not able to reproduce on my raspberry-pi, I should try on my odroid when I find it

cdecker commented 6 months ago

This is likely just valgrind learning new tricks, including that overlapping memcpy source and destination may cause trouble. Furthermore, the memcpy source and destinations are actually being randomized, which would make this a very flaky error.

We might get away with a + 1 in on the iteration, or when setting j we could ensure that j != i in all cases, which as far as I can see is the only time those memcpy calls overlap.