Closed bmagistro closed 2 months ago
@bmagistro thanks for creating this public UHD issue. I've tagged it in our private / internal issues as you can see. I've also marked is as "escalated" since I think it's critical to get it fixed. We'll see what others think.
Any updates?
I still have this issue with Yocto Gatesgarth and GCC 10.2 on a Zynq-7000
I saw here some Debian people saying it was solved as a toolchain problem (https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=943396) but I still have the issue.
Hey, we are just in the progress of updating an E310 that was still running UHD 3.9.2 to a more recent version I found the same with UHD 4.0. Is there a solution yet? Is 4.1 any better? If not, what's the last known working version?
Thanks Andre
In the next release, we'll be moving to kirkstone which contains a newer gcc.
Opening an issue here for tracking/linking purposes. The initial discussion occurred on the mailing list (http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2020-December/063312.html).
Short version, when fftw is compiled against gcc 9.x and NEON is enabled, fftw may segfault. Another user opened an issue with the fftw project (https://github.com/FFTW/fftw3/issues/213). A possible solution would be to add gcc-8.x as a devtools recipe that could be used to override the default compiler shipped in Zeus (and likely Dunfell). There is an open request with the yocto project to have them revert a recent commit and continue including gcc 8.x due to this issue.
Believe the issue better resides in meta-ettus but don't see a way to add issues there.
cc: @michaelld @balister