Starting with GLib 2.14.1, there is new inline assembly code that implements
atomic spinlocks on ARM. This code causes the DM6446 to crash almost
A patch file is currently in place to work around this issue by disabling the
ARM assembly. This shouldn't degrade performance, as previous version of
didn't have this code. We need to investigate further to see if this code
be made to run properly so we can take advantage of the prformance
The patch file with the workaround is named
is also attached to this ticket.