On Tue, 2011-02-01 at 00:13 +0100, Andrea Arcangeli wrote: > On Mon, Jan 31, 2011 at 02:40:16PM -0800, Dave Hansen wrote: > > Still not a very good data point, but I ran a heavy swap load for an > > hour or so without reproducing this. But, it happened again after I > > enabled transparent huge pages. I managed to get a sysrq-t dump out of > > it: > > > > http://sr71.net/~dave/ibm/2.6.38-rc2-hang-0.txt > > > > khugepaged is one of the three running tasks. Note, I set both its > > sleep timeouts to zero to stress it out a bit. > > sysrq+l? sysrq+t doesn't provide interesting info for running tasks. I'll try to get a sysrq-l dump. > > I'll keep trying to reproduce without THP. > > BTW, do you have prove locking enabled? With recent git I get a > deadlock inside _raw_spin_unlock_irq with prove locking enabled and it > goes away when I disable it. Nope: $ grep PROVE ../mhp-build/x86_64-elm3b82/.config # CONFIG_PROVE_LOCKING is not set Full .config is here http://sr71.net/~dave/ibm/config-v2.6.38-rc2 -- Dave -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxxx For more info on Linux MM, see: http://www.linux-mm.org/ . Fight unfair telecom policy in Canada: sign http://dissolvethecrtc.ca/ Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>