Re: 3.0.1: pagevec_lookup+0x1d/0x30, SLAB issues? [again w/DEBUG_VM& SLAB_DEBUG enabled]

[Justin Piszcz]

On Mon, 12 Sep 2011, Lin Ming wrote:

> > [27336.007233]  [<ffffffff81089fc3>] kswapd+0xb3/0x250
> > [27336.007237]  [<ffffffff8104fea0>] ? abort_exclusive_wait+0xb0/0xb0
> > [27336.007241]  [<ffffffff81089f10>] ? balance_pgdat+0x6a0/0x6a0
> > [27336.007244]  [<ffffffff8104f457>] kthread+0x87/0x90
> > [27336.007248]  [<ffffffff81648814>] kernel_thread_helper+0x4/0x10
> > [27336.007252]  [<ffffffff8104f3d0>] ? kthread_flush_work_fn+0x10/0x10
> > [27336.007256]  [<ffffffff81648810>] ? gs_change+0xb/0xb
>
> No more info got.
>
> Could you try to make a test case which can trigger this bug reliably?
> You may also try some benchmarks, for example, fio, ffsb, etc.
>
> Lin Ming

Hi,

I do not unfortunately, it seems to occur at random times during heavy I/O+
mem and processing time, are there any other debug options that would help us
here?

Justin.