Hello, Alex.
On Tue, Dec 22, 2015 at 07:06:41PM +0000, Alex Ng (LIS) wrote:
> > Can you please let me know the steps to reproduce the bug?
>
> I tried this on a Hyper-V VM hosted in Windows Server 2012R2 and ran
> the attached script. The script clones the linux-next tree in a
> random directory under /tmp in a tight loop.
>
> This panic is not always reproducible, and I have only hit it once
> after running the script about 10 times. A different kernel panic
> happens each time I run this script; and the panics always happen
> during the first iteration of the loop.
Heh, I don't get it. The script doesn't do anything cgroup specific.
Can you please apply the attached patch, reproduce the issue and
report the kernel log?
Thanks.
--
tejun
---
kernel/cgroup.c | 21 ++++++++++++++++++++-
1 file changed, 20 insertions(+), 1 deletion(-)
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -779,6 +779,22 @@ static inline void get_css_set(struct cs
atomic_inc(&cset->refcount);
}
+static void dump_cset(struct css_set *cset)
+{
+ struct cgrp_cset_link *link;
+
+ printk("XXX dumping cset %p\n", cset);
+ list_for_each_entry(link, &cset->cgrp_links, cgrp_link) {
+ struct cgroup *cgrp = link->cgrp;
+ struct cgroup_root *root = cgrp->root;
+
+ printk("root %d:0x%04x:%s ",
+ root->hierarchy_id, root->subsys_mask, root->name);
+ pr_cont_cgroup_path(cgrp);
+ pr_cont("\n");
+ }
+}
+
/**
* compare_css_sets - helper function for find_existing_css_set().
* @cset: candidate css_set being tested
@@ -831,7 +847,10 @@ static bool compare_css_sets(struct css_
cgrp1 = link1->cgrp;
cgrp2 = link2->cgrp;
/* Hierarchies should be linked in the same order. */
- BUG_ON(cgrp1->root != cgrp2->root);
+ if (WARN_ON(cgrp1->root != cgrp2->root)) {
+ dump_cset(cset);
+ dump_cset(old_cset);
+ }
/*
* If this hierarchy is the hierarchy of the cgroup
