On Fri, 28 Nov 2008 20:47:00 -0800, "Peter Wemm" <peter_at_wemm.org> wrote: >> The xterm processes that get stuck seem to be spinning near line 1854 of >> sched_ule.c. Running `info threads' on a live kernel after xterm starts >> spinning on a CPU shows: After a bit of help from kib_at_ and ed_at_ this seems to be wrong. The xterm processes were spinning in userspace, calling ioctl(fd, FIONREAD, ...). On Sat, 29 Nov 2008 16:02:03 -0800, "Garrett Cooper" <yanefbsd_at_gmail.com> wrote: > Might be another compiler issue. There's an issue with > -fstrict-aliasing (at least) on all versions of g++ up to 4.2.3: > <http://lists.copyleft.no/pipermail/pyrex/2007-November/003071.html>. > Are there some compiler warnings when compiling kernel / screen? > Also, have you tried with just COPTFLAGS disabled? Today I managed to reproduce this with a kernel that doesn't use optimizations at all, so fortunately this is not a compiler bug :) I'm testing a patch by Ed Schouten to see if it fixes this, and I will post more details after a day or so of running with the patch.Received on Sun Nov 30 2008 - 13:09:02 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:38 UTC