Re: KVM: memory ballooning bug?
From: Minchan Kim
Date: Wed Dec 23 2015 - 09:17:31 EST
On Wed, Dec 23, 2015 at 06:14:49AM -0500, Rafael Aquini wrote:
> On Wed, Dec 23, 2015 at 02:22:28PM +0900, Minchan Kim wrote:
> > During my compaction-related stuff, I encountered some problems with
> > ballooning.
> >
> > Firstly, with repeated inflating and deflating cycle, guest memory(ie,
> > cat /proc/meminfo | grep MemTotal) decreased and couldn't recover.
> >
> > When I review source code, balloon_lock should cover release_pages_balloon.
> > Otherwise, struct virtio_balloon fields could be overwritten by race
> > of fill_balloon(e,g, vb->*pfns could be critical).
> > Below patch fixed the problem.
> >
> > diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c
> > index 7efc32945810..7d3e5d0e9aa4 100644
> > --- a/drivers/virtio/virtio_balloon.c
> > +++ b/drivers/virtio/virtio_balloon.c
> > @@ -209,8 +209,8 @@ static unsigned leak_balloon(struct virtio_balloon *vb, size_t num)
> > */
> > if (vb->num_pfns != 0)
> > tell_host(vb, vb->deflate_vq);
> > - mutex_unlock(&vb->balloon_lock);
> > release_pages_balloon(vb);
> > + mutex_unlock(&vb->balloon_lock);
> > return num_freed_pages;
> > }
> >
> > Secondly, in balloon_page_dequeue, pages_lock should cover
> > list_for_each_entry_safe loop. Otherwise, the cursor page
> > could be isolated by compaction and then list_del by isolation
> > could poison the page->lru so the loop could access wrong address
> > like this.
> >
> > general protection fault: 0000 [#1] SMP
> > Dumping ftrace buffer:
> > (ftrace buffer empty)
> > Modules linked in:
> > CPU: 2 PID: 82 Comm: vballoon Not tainted 4.4.0-rc5-mm1+ #1906
> > Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
> > task: ffff8800a7ff0000 ti: ffff8800a7fec000 task.ti: ffff8800a7fec000
> > RIP: 0010:[<ffffffff8115e754>] [<ffffffff8115e754>] balloon_page_dequeue+0x54/0x130
> > RSP: 0018:ffff8800a7fefdc0 EFLAGS: 00010246
> > RAX: ffff88013fff9a70 RBX: ffffea000056fe00 RCX: 0000000000002b7d
> > RDX: ffff88013fff9a70 RSI: ffffea000056fe00 RDI: ffff88013fff9a68
> > RBP: ffff8800a7fefde8 R08: ffffea000056fda0 R09: 0000000000000000
> > R10: ffff8800a7fefd90 R11: 0000000000000001 R12: dead0000000000e0
> > R13: ffffea000056fe20 R14: ffff880138809070 R15: ffff880138809060
> > FS: 0000000000000000(0000) GS:ffff88013fc40000(0000) knlGS:0000000000000000
> > CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
> > CR2: 00007f229c10e000 CR3: 00000000b8b53000 CR4: 00000000000006a0
> > Stack:
> > 0000000000000100 ffff880138809088 ffff880138809000 ffff880138809060
> > 0000000000000046 ffff8800a7fefe28 ffffffff812c86d3 ffff880138809020
> > ffff880138809000 fffffffffff91900 0000000000000100 ffff880138809060
> > Call Trace:
> > [<ffffffff812c86d3>] leak_balloon+0x93/0x1a0
> > [<ffffffff812c8bc7>] balloon+0x217/0x2a0
> > [<ffffffff8143739e>] ? __schedule+0x31e/0x8b0
> > [<ffffffff81078160>] ? abort_exclusive_wait+0xb0/0xb0
> > [<ffffffff812c89b0>] ? update_balloon_stats+0xf0/0xf0
> > [<ffffffff8105b6e9>] kthread+0xc9/0xe0
> > [<ffffffff8105b620>] ? kthread_park+0x60/0x60
> > [<ffffffff8143b4af>] ret_from_fork+0x3f/0x70
> > [<ffffffff8105b620>] ? kthread_park+0x60/0x60
> > Code: 8d 60 e0 0f 84 af 00 00 00 48 8b 43 20 a8 01 75 3b 48 89 d8 f0 0f ba 28 00 72 10 48 8b 03 f6 c4 08 75 2f 48 89 df e8 8c 83 f9 ff <49> 8b 44 24 20 4d 8d 6c 24 20 48 83 e8 20 4d 39 f5 74 7a 4c 89
> > RIP [<ffffffff8115e754>] balloon_page_dequeue+0x54/0x130
> > RSP <ffff8800a7fefdc0>
> > ---[ end trace 43cf28060d708d5f ]---
> > Kernel panic - not syncing: Fatal exception
> > Dumping ftrace buffer:
> > (ftrace buffer empty)
> > Kernel Offset: disabled
> >
> > We could fix it by protecting the entire loop by pages_lock but
> > problem is irq latency during walking the list.
> > But I doubt how often such worst scenario happens because
> > in normal situation, the loop would exit easily via succeeding
> > trylock_page.
> >
> > Any comments?
>
> Nope, I think the simplest way to address both cases you stumbled
> across is by replacing the locking to extend those critical sections as
> you suggested.
I couldn't understand why you said "Nope" and which lock do you mean?
There are two locks to need to extend.
If you are on same page with me, I suggested this.
diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c
index 7efc329..7d3e5d0 100644
--- a/drivers/virtio/virtio_balloon.c
+++ b/drivers/virtio/virtio_balloon.c
@@ -209,8 +209,8 @@ static unsigned leak_balloon(struct virtio_balloon *vb, size_t num)
*/
if (vb->num_pfns != 0)
tell_host(vb, vb->deflate_vq);
- mutex_unlock(&vb->balloon_lock);
release_pages_balloon(vb);
+ mutex_unlock(&vb->balloon_lock);
return num_freed_pages;
}
diff --git a/mm/balloon_compaction.c b/mm/balloon_compaction.c
index d3116be..300117f 100644
--- a/mm/balloon_compaction.c
+++ b/mm/balloon_compaction.c
@@ -61,6 +61,7 @@ struct page *balloon_page_dequeue(struct balloon_dev_info *b_dev_info)
bool dequeued_page;
dequeued_page = false;
+ spin_lock_irqsave(&b_dev_info->pages_lock, flags);
list_for_each_entry_safe(page, tmp, &b_dev_info->pages, lru) {
/*
* Block others from accessing the 'page' while we get around
@@ -75,15 +76,14 @@ struct page *balloon_page_dequeue(struct balloon_dev_info *b_dev_info)
continue;
}
#endif
- spin_lock_irqsave(&b_dev_info->pages_lock, flags);
balloon_page_delete(page);
__count_vm_event(BALLOON_DEFLATE);
- spin_unlock_irqrestore(&b_dev_info->pages_lock, flags);
unlock_page(page);
dequeued_page = true;
break;
}
}
+ spin_unlock_irqrestore(&b_dev_info->pages_lock, flags);
if (!dequeued_page) {
/*
Do you agree this patch? If so, I will send patch after X-mas.
Otherwise, please elaborate it more.
Happy Happy Xmas!
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/