|
From: Nicholas N. <n.n...@gm...> - 2009-07-20 03:56:47
|
On Sat, Jul 18, 2009 at 4:59 PM, Bart Van Assche<bar...@gm...> wrote: > On Fri, Jul 17, 2009 at 2:50 AM, Nicholas > Nethercote<n.n...@gm...> wrote: >> On Fri, Jul 17, 2009 at 2:40 AM, Bart Van >> +valgrind: >> m_scheduler/scheduler.c:1199 (vgPlain_scheduler): the 'impossible' >> happened. >>>> +valgrind: VG_(scheduler), phase 3: run_innerloop detected host state invariant failure >>>> + at 0x........: report_and_quit (m_libcassert.c:?) >>>> + by 0x........: vgPlain_assert_fail (m_libcassert.c:?) >>>> + by 0x........: vgPlain_scheduler (scheduler.c:?) >>>> + by 0x........: run_a_thread_NORETURN (syswrap-linux.c:91) >>> >>> Is the above reproducible ? Have these results been obtained on a >>> non-virtual or on a virtual machine ? >> >> I get it occasionally. This is a non-virtual machine. See >> https://bugs.kde.org/show_bug.cgi?id=143888. It is supposedly a >> kernel bug that has been fixed but it still comes up. > > Which kernel version did this issue occur on (dpkg-query -s > linux-image-generic | grep Version) ? I assume that this system was > running an unmodified Ubuntu kernel ? I'm going to have a look at the > Debian and Ubuntu patches. Version: 2.6.28.13.17 It's a vanilla Ubuntu 9.04 x86-64 system. The only kernel-related change I have is that I have "vdso32=0" as an additional kernel boot parameter. (And I'm about to remove that parameter as I don't need it any more.) Nick |