From: Madhuparna Bhowmik <[email protected]>
This patch fixes the following warning:
=============================
WARNING: suspicious RCU usage
5.7.0-rc4-next-20200507-syzkaller #0 Not tainted
-----------------------------
net/ipv6/ip6mr.c:124 RCU-list traversed in non-reader section!!
ipmr_new_table() returns an existing table, but there is no table at
init. Therefore the condition: either holding rtnl or the list is empty
is used.
Fixes: d13fee049f ("Default enable RCU list lockdep debugging with .."): WARNING: suspicious RCU usage
Reported-by: kernel test robot <[email protected]>
Suggested-by: Jakub Kicinski <[email protected]>
Signed-off-by: Madhuparna Bhowmik <[email protected]>
---
net/ipv6/ip6mr.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c
index 65a54d74acc1..fbe282bb8036 100644
--- a/net/ipv6/ip6mr.c
+++ b/net/ipv6/ip6mr.c
@@ -98,7 +98,7 @@ static void ipmr_expire_process(struct timer_list *t);
#ifdef CONFIG_IPV6_MROUTE_MULTIPLE_TABLES
#define ip6mr_for_each_table(mrt, net) \
list_for_each_entry_rcu(mrt, &net->ipv6.mr6_tables, list, \
- lockdep_rtnl_is_held())
+ lockdep_rtnl_is_held() || list_empty(&net->ipv6.mr6_tables))
static struct mr_table *ip6mr_mr_table_iter(struct net *net,
struct mr_table *mrt)
--
2.17.1
On Thu, 14 May 2020 12:32:04 +0530 [email protected] wrote:
> From: Madhuparna Bhowmik <[email protected]>
>
> This patch fixes the following warning:
>
> =============================
> WARNING: suspicious RCU usage
> 5.7.0-rc4-next-20200507-syzkaller #0 Not tainted
> -----------------------------
> net/ipv6/ip6mr.c:124 RCU-list traversed in non-reader section!!
>
> ipmr_new_table() returns an existing table, but there is no table at
> init. Therefore the condition: either holding rtnl or the list is empty
> is used.
>
> Fixes: d13fee049f ("Default enable RCU list lockdep debugging with .."): WARNING: suspicious RCU usage
Fixes tag: Fixes: d13fee049f ("Default enable RCU list lockdep debugging with .."): WARNING: suspicious RCU usage
Has these problem(s):
- Target SHA1 does not exist
I think the message at the end is confusing automation, could you use
the standard Fixes tag format, please?
> Reported-by: kernel test robot <[email protected]>
> Suggested-by: Jakub Kicinski <[email protected]>
> Signed-off-by: Madhuparna Bhowmik <[email protected]>
> ---
> net/ipv6/ip6mr.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c
> index 65a54d74acc1..fbe282bb8036 100644
> --- a/net/ipv6/ip6mr.c
> +++ b/net/ipv6/ip6mr.c
> @@ -98,7 +98,7 @@ static void ipmr_expire_process(struct timer_list *t);
> #ifdef CONFIG_IPV6_MROUTE_MULTIPLE_TABLES
> #define ip6mr_for_each_table(mrt, net) \
> list_for_each_entry_rcu(mrt, &net->ipv6.mr6_tables, list, \
> - lockdep_rtnl_is_held())
> + lockdep_rtnl_is_held() || list_empty(&net->ipv6.mr6_tables))
double space, line over 80 chars
> static struct mr_table *ip6mr_mr_table_iter(struct net *net,
> struct mr_table *mrt)
Other than these nits looks good, thanks!
On Thu, May 14, 2020 at 09:40:08AM -0700, Jakub Kicinski wrote:
> On Thu, 14 May 2020 12:32:04 +0530 [email protected] wrote:
> > From: Madhuparna Bhowmik <[email protected]>
> >
> > This patch fixes the following warning:
> >
> > =============================
> > WARNING: suspicious RCU usage
> > 5.7.0-rc4-next-20200507-syzkaller #0 Not tainted
> > -----------------------------
> > net/ipv6/ip6mr.c:124 RCU-list traversed in non-reader section!!
> >
> > ipmr_new_table() returns an existing table, but there is no table at
> > init. Therefore the condition: either holding rtnl or the list is empty
> > is used.
> >
> > Fixes: d13fee049f ("Default enable RCU list lockdep debugging with .."): WARNING: suspicious RCU usage
>
> Fixes tag: Fixes: d13fee049f ("Default enable RCU list lockdep debugging with .."): WARNING: suspicious RCU usage
> Has these problem(s):
> - Target SHA1 does not exist
>
I got this report: https://lkml.org/lkml/2020/5/12/1358
That's why I used this SHA1 with the fixes tag.
> I think the message at the end is confusing automation, could you use
> the standard Fixes tag format, please?
>
Sure, I will do that.
> > Reported-by: kernel test robot <[email protected]>
> > Suggested-by: Jakub Kicinski <[email protected]>
> > Signed-off-by: Madhuparna Bhowmik <[email protected]>
> > ---
> > net/ipv6/ip6mr.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c
> > index 65a54d74acc1..fbe282bb8036 100644
> > --- a/net/ipv6/ip6mr.c
> > +++ b/net/ipv6/ip6mr.c
> > @@ -98,7 +98,7 @@ static void ipmr_expire_process(struct timer_list *t);
> > #ifdef CONFIG_IPV6_MROUTE_MULTIPLE_TABLES
> > #define ip6mr_for_each_table(mrt, net) \
> > list_for_each_entry_rcu(mrt, &net->ipv6.mr6_tables, list, \
> > - lockdep_rtnl_is_held())
> > + lockdep_rtnl_is_held() || list_empty(&net->ipv6.mr6_tables))
>
> double space, line over 80 chars
Okay, I will make these changes and send the patch.
Regards,
Madhuparna
>
> > static struct mr_table *ip6mr_mr_table_iter(struct net *net,
> > struct mr_table *mrt)
>
> Other than these nits looks good, thanks!