On Fri, 2007-07-06 at 16:21 +0200, Patrick McHardy wrote: > Ranko Zivojnovic wrote: > > BUG: spinlock lockup on CPU#0, swapper/0, c03eff80 > > [<c01ed1fe>] _raw_spin_lock+0x108/0x13c > > [<c02a8468>] __qdisc_run+0x97/0x1b0 > > [<c02a96f3>] qdisc_watchdog+0x19/0x58 > > [<c02fe5e7>] __lock_text_start+0x37/0x43 > > [<c02a9730>] qdisc_watchdog+0x56/0x58 > > [<c02a96da>] qdisc_watchdog+0x0/0x58 > > [<c0135d84>] run_hrtimer_softirq+0x58/0xb5 > > [...] > > > BUG: spinlock lockup on CPU#1, swapper/0, c03eff80 > > [<c01ed1fe>] _raw_spin_lock+0x108/0x13c > > [<c0298b9b>] est_timer+0x53/0x148 > > [<c01294b3>] run_timer_softirq+0x30/0x184 > > [<c01295a4>] run_timer_softirq+0x121/0x184 > > [<c0126252>] __do_softirq+0x66/0xf3 > > [<c0298b48>] est_timer+0x0/0x148 > > [...] > > > There is at least one ABBA deadlock, est_timer does: > > read_lock(&est_lock) > spin_lock(e->stats_lock) (which is dev->queue_lock) > > and qdisc_destroy calls htb_destroy under dev->queue_lock, which > calls htb_destroy_class, then gen_kill_estimator and this > write_locks est_lock. > > I can't see the problem above though, the qdisc_run path only takes > dev->queue_lock. Please enable lockdep and post the output if any.
That was with lockdep already enabled: $ zgrep LOCKDEP /proc/config.gz CONFIG_LOCKDEP_SUPPORT=y CONFIG_LOCKDEP=y # CONFIG_DEBUG_LOCKDEP is not set also with: CONFIG_PROVE_LOCKING=y CONFIG_DEBUG_LOCK_ALLOC=y that Jarek suggested, and including: CONFIG_DETECT_SOFTLOCKUP=y CONFIG_DEBUG_SPINLOCK=y I will give it another spin as it is, just to see if it will show a different path - and then I will try 2.6.21.5 with these two patches. R. - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
