Messages in this thread | | | From | Rusty Russell <> | Subject | Re: Please add to stable: module: don't unlink the module until we've removed all exposure. | Date | Tue, 04 Jun 2013 15:26:28 +0930 |
| |
Ben Greear <greearb@candelatech.com> writes: >> It at least works around the problem for me as well. But, a more rare >> migration/[0-3] (I think) related lockup still exists in 3.9.4 for me, >> so I will also try applying that other kobject patch and continue testing >> today... > > Well, that other kobject patch is already in 3.9.4, so I think it's still > a good idea to include the > "module: don't unlink the module until we've removed all exposure." > patch in stable. I have a decent test case to reproduce the crash, so if someone > wants me to test other patches instead, then I will do so.
OK, I cannot reproduce on 3.9.4. I #if 0'd out the WARNs in sysfs and kobject, and did this (which reliably broke on 3.8):
# M=`modinfo -F filename e1000` # for i in `seq 10000`; do insmod $M; rmmod e1000; done >/dev/null 2>&1 & for i in `seq 10000`; do insmod $M; rmmod e1000; done > /dev/null 2>&1 & for i in `seq 10000`; do insmod $M; rmmod e1000; done > /dev/null 2>&1 & for i in `seq 10000`; do insmod $M; rmmod e1000; done >/dev/null 2>&1 & #
This was under kvm, 4-way SMP, init=/bin/bash.
Do you have a backtrace of the 3.9.4 crash? You can add "CFLAGS_module.o = -O0" to get a clearer backtrace if you want...
Thanks, Rusty.
| |