| Message ID | 51A611CF02000078000D99A6@nat28.tlf.novell.com |
|---|---|
| State | Not Applicable, archived |
| Delegated to: | David Miller |
| Headers | show |
On Wed, May 29, 2013 at 01:33:51PM +0100, Jan Beulich wrote: > kfree_rcu() requires offsetof(..., rcu_head) < 4096, which can get > violated with a sufficiently high CONFIG_IP_VS_SH_TAB_BITS. > > Signed-off-by: Jan Beulich <jbeulich@suse.com> Thanks Jan. Pablo, could you take this one directly into your tree? Signed-off-by: Simon Horman <horms@verge.net.au> > > --- > net/netfilter/ipvs/ip_vs_sh.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > --- 3.10-rc3/net/netfilter/ipvs/ip_vs_sh.c > +++ 3.10-rc3-ip_vs_sh-build/net/netfilter/ipvs/ip_vs_sh.c > @@ -67,8 +67,8 @@ struct ip_vs_sh_bucket { > #define IP_VS_SH_TAB_MASK (IP_VS_SH_TAB_SIZE - 1) > > struct ip_vs_sh_state { > - struct ip_vs_sh_bucket buckets[IP_VS_SH_TAB_SIZE]; > struct rcu_head rcu_head; > + struct ip_vs_sh_bucket buckets[IP_VS_SH_TAB_SIZE]; > }; > > /* > > > > -- > To unsubscribe from this list: send the line "unsubscribe lvs-devel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, May 29, 2013 at 11:59:06PM +0900, Simon Horman wrote: > On Wed, May 29, 2013 at 01:33:51PM +0100, Jan Beulich wrote: > > kfree_rcu() requires offsetof(..., rcu_head) < 4096, which can get > > violated with a sufficiently high CONFIG_IP_VS_SH_TAB_BITS. > > > > Signed-off-by: Jan Beulich <jbeulich@suse.com> > > Thanks Jan. > > Pablo, could you take this one directly into your tree? > > Signed-off-by: Simon Horman <horms@verge.net.au> Applied, thanks Simon. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Hello, On Wed, 29 May 2013, Jan Beulich wrote: > kfree_rcu() requires offsetof(..., rcu_head) < 4096, which can get > violated with a sufficiently high CONFIG_IP_VS_SH_TAB_BITS. > > Signed-off-by: Jan Beulich <jbeulich@suse.com> > > --- > net/netfilter/ipvs/ip_vs_sh.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > --- 3.10-rc3/net/netfilter/ipvs/ip_vs_sh.c > +++ 3.10-rc3-ip_vs_sh-build/net/netfilter/ipvs/ip_vs_sh.c > @@ -67,8 +67,8 @@ struct ip_vs_sh_bucket { > #define IP_VS_SH_TAB_MASK (IP_VS_SH_TAB_SIZE - 1) > > struct ip_vs_sh_state { > - struct ip_vs_sh_bucket buckets[IP_VS_SH_TAB_SIZE]; > struct rcu_head rcu_head; > + struct ip_vs_sh_bucket buckets[IP_VS_SH_TAB_SIZE]; Your fix reminds me about the same problem I noticed for LBLC (may be due to 10 bits) during development but forgot to check the other schedulers. The same solution is needed also for DH. OTOH, LBLC and LBLCR are correct. Do you prefer to provide similar fix for ip_vs_dh.c or I have to take care? > }; > > /* Regards -- Julian Anastasov <ja@ssi.bg> -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
>>> Julian Anastasov <ja@ssi.bg> 05/29/13 9:28 PM >>> >On Wed, 29 May 2013, Jan Beulich wrote: >> --- 3.10-rc3/net/netfilter/ipvs/ip_vs_sh.c >> +++ 3.10-rc3-ip_vs_sh-build/net/netfilter/ipvs/ip_vs_sh.c >> @@ -67,8 +67,8 @@ struct ip_vs_sh_bucket { >> #define IP_VS_SH_TAB_MASK (IP_VS_SH_TAB_SIZE - 1) >> >> struct ip_vs_sh_state { >> - struct ip_vs_sh_bucket buckets[IP_VS_SH_TAB_SIZE]; >> struct rcu_head rcu_head; >> + struct ip_vs_sh_bucket buckets[IP_VS_SH_TAB_SIZE]; > > Your fix reminds me about the same problem >I noticed for LBLC (may be due to 10 bits) during development >but forgot to check the other schedulers. The same solution >is needed also for DH. OTOH, LBLC and LBLCR are correct. >Do you prefer to provide similar fix for ip_vs_dh.c or >I have to take care? CONFIG_IP_VS_DH_TAB_BITS is always 8 (due to not coming from any Kconfig setting), and hence not a problem at this point. If it is being planned to make configurable, then that change should at once flip the structure members. Jan -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Hello, On Thu, 30 May 2013, Jan Beulich wrote: > CONFIG_IP_VS_DH_TAB_BITS is always 8 (due to not coming from any > Kconfig setting), and hence not a problem at this point. If it is being planned > to make configurable, then that change should at once flip the structure > members. Aha, OK. Regards -- Julian Anastasov <ja@ssi.bg> -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
--- 3.10-rc3/net/netfilter/ipvs/ip_vs_sh.c +++ 3.10-rc3-ip_vs_sh-build/net/netfilter/ipvs/ip_vs_sh.c @@ -67,8 +67,8 @@ struct ip_vs_sh_bucket { #define IP_VS_SH_TAB_MASK (IP_VS_SH_TAB_SIZE - 1) struct ip_vs_sh_state { - struct ip_vs_sh_bucket buckets[IP_VS_SH_TAB_SIZE]; struct rcu_head rcu_head; + struct ip_vs_sh_bucket buckets[IP_VS_SH_TAB_SIZE]; }; /*
kfree_rcu() requires offsetof(..., rcu_head) < 4096, which can get violated with a sufficiently high CONFIG_IP_VS_SH_TAB_BITS. Signed-off-by: Jan Beulich <jbeulich@suse.com> --- net/netfilter/ipvs/ip_vs_sh.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html