Netfilter fixes for net
Subsystems (3)
Pablo Neira Ayuso
maintainercover Netfilter fixes for net
Hi,
The following patchset contains Netfilter fixes for net:
1) Use DEBUG_NET_WARN_ON_ONCE() instead of WARN_ON() from the tproxy datapath, a recent bug found a way to reach WARN_ON from datapath due to insufficient validation of xt_TPROTO checkentry. From Fernando F. Mancera.
2) Similar to previous patch to replace WARN_ON_ONCE by DEBUG_NET_WARN_ON_ONCE() for connlimit. Not known issue, but since this patch has been around for a while, let's merge it. Also from Fernando.
3) Move nf_tables harware offload commit path after chain blob and audit to reduce chances of leaving the hardware in inconsistent state.
4) Add missing vzeroupper to nf_tables pipapo AVX2 to address performace degradation to later user of SSE code, from Eric Biggers.
5) Remove pr_debug() in x_tables extensions, a recent bogus found a way to print a unsanitized string in xt_IDLETIMER, many of these pr_debug() calls are there for historical reasons.
6) Use pr_info_ratelimited() in x_tables .checkentry.
7) Fix an imbalance in module refcount due to incorrect override expression logic with sets. Remove unnecessary clone in control plane, use the existing expressions provided by set or dynset expression. Release override expressions only.
9) Tigthen nf_tables device name removal, it is possible to remove prefix strings with exact device name. From Fernando F. Mancera.
9) Set on the set dead bit earlier, otherwise it is possible to call .commit on deleted sets. This also addresses the re-introduction of a bug.
10) Remove leftover definition of the local set_update_list. A recent patch made it per-netns but did not remove this which is now unused.
Please, pull these changes from:
git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf.git nf-26-08-21
Thanks.
----------------------------------------------------------------
The following changes since commit 746fc0787f616da418ffc04a110296fe95d53491:
net: usb: cdc_ncm: add Apple MacBook Pro USB product ID 0x1902 (2026-08-20 14:47:29 -0700)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf.git tags/nf-26-08-21
for you to fetch changes up to 878ca6f9739ab460a185fd1f0e64c8ef026057e4:
netfilter: nf_tables: remove leftover set_update_list (2026-08-21 17:12:59 +0200)
----------------------------------------------------------------
netfilter pull request 26-08-21
----------------------------------------------------------------
Eric Biggers (1):
netfilter: nft_set_pipapo_avx2: add missing vzeroupper
Fernando Fernandez Mancera (3):
netfilter: tproxy: use DEBUG_NET_WARN_ON_ONCE for protocol fallbacks
netfilter: conncount: use DEBUG_NET_WARN_ON_ONCE on reaching count limit
netfilter: nf_tables: fix device name and prefix match in hook lookup
Pablo Neira Ayuso (6):
netfilter: nf_tables: move hardware offload step after building the chain blob
netfilter: x_tables: remove pr_debug
netfilter: xt_cgroup: use pr_info_ratelimited()
netfilter: nf_tables: skip double clone set expressions on element insert
netfilter: nf_tables: set on dead bit when performing early element removal
netfilter: nf_tables: remove leftover set_update_list
include/net/netfilter/nf_tables.h | 2 - net/ipv4/netfilter/ipt_ah.c | 10 +---- net/ipv4/netfilter/nf_tproxy_ipv4.c | 2 +- net/ipv6/netfilter/ip6t_ah.c | 27 +----------- net/ipv6/netfilter/ip6t_frag.c | 41 +----------------- net/ipv6/netfilter/ip6t_hbh.c | 40 ++++-------------- net/ipv6/netfilter/ip6t_mh.c | 3 -- net/ipv6/netfilter/ip6t_rt.c | 6 +-- net/ipv6/netfilter/nf_tproxy_ipv6.c | 2 +- net/netfilter/nf_conncount.c | 3 +- net/netfilter/nf_tables_api.c | 84 +++++++++++++++++-------------------- net/netfilter/nft_dynset.c | 25 ++++++----- net/netfilter/nft_set_pipapo_avx2.c | 17 ++++---- net/netfilter/xt_IDLETIMER.c | 68 ++++++------------------------ net/netfilter/xt_LOG.c | 4 +- net/netfilter/xt_MASQUERADE.c | 4 +- net/netfilter/xt_NETMAP.c | 4 +- net/netfilter/xt_REDIRECT.c | 4 +- net/netfilter/xt_cgroup.c | 10 ++--- net/netfilter/xt_esp.c | 10 +---- net/netfilter/xt_ipcomp.c | 8 +--- net/netfilter/xt_iprange.c | 32 ++------------ net/netfilter/xt_ipvs.c | 1 - net/netfilter/xt_multiport.c | 4 -- net/netfilter/xt_sctp.c | 19 +-------- net/netfilter/xt_tcpudp.c | 8 +--- 26 files changed, 117 insertions(+), 321 deletions(-)
Pablo Neira Ayuso
maintainerPATCH 1/10 netfilter: tproxy: use DEBUG_NET_WARN_ON_ONCE for protocol fallbacks
From: Fernando Fernandez Mancera <[email protected]>
Replace WARN_ON calls with DEBUG_NET_WARN_ON_ONCE in the default switch blocks of nf_tproxy_get_sock_v4 and v6. Unsupported transport protocols are already safely handled by returning a NULL socket pointer. This prevents unnecessary system panics when panic_on_warn=1 is enabled in production systems.
--- net/ipv4/netfilter/nf_tproxy_ipv4.c | 2 +- net/ipv6/netfilter/nf_tproxy_ipv6.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
net/ipv4/netfilter/nf_tproxy_ipv4.c +1 -1
@@ -137,7 +137,7 @@ nf_tproxy_get_sock_v4(struct net *net, struct sk_buff *skb, }break ;default : -WARN_ON (1 ); +DEBUG_NET_WARN_ON_ONCE (1 ); sk =NULL ; }
net/ipv6/netfilter/nf_tproxy_ipv6.c +1 -1
@@ -136,7 +136,7 @@ nf_tproxy_get_sock_v6(struct net *net, struct sk_buff *skb, int thoff, }break ;default : -WARN_ON (1 ); +DEBUG_NET_WARN_ON_ONCE (1 ); sk =NULL ; }
Signature · 1 line
2.47.3
Pablo Neira Ayuso
maintainerPATCH 2/10 netfilter: conncount: use DEBUG_NET_WARN_ON_ONCE on reaching count limit
In reply to Pablo Neira AyusoFrom: Fernando Fernandez Mancera <[email protected]>
Replace WARN_ON_ONCE with DEBUG_NET_WARN_ON_ONCE in __nf_conncount_add. The function handles count limit breaches safely by returning -EOVERFLOW, so a production backtrace is not needed. This prevents unnecessary system panics when panic_on_warn=1 is enabled in production systems.
--- net/netfilter/nf_conncount.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
net/netfilter/nf_conncount.c +2 -1
@@ -251,7 +251,8 @@ static int __nf_conncount_add(struct net *net, list->last_gc_count = list->count; add_new_node: -if (WARN_ON_ONCE (list->count >INT_MAX )) { +if (unlikely (list->count >INT_MAX )) { +DEBUG_NET_WARN_ON_ONCE (1 ); err = -EOVERFLOW ; goto out_put; }
Signature · 1 line
2.47.3
Pablo Neira Ayuso
maintainerPATCH 3/10 netfilter: nf_tables: move hardware offload step after building the chain blob
In reply to Pablo Neira AyusoAllocate the chain blob before the ruleset offload to reduce chances of entering an inconsistent state where the offloaded ruleset in the nic and the software ruleset differ.
--- net/netfilter/nf_tables_api.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-)
net/netfilter/nf_tables_api.c +10 -4
@@ -10982,10 +10982,6 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)return -EAGAIN ; } - err =nft_flow_rule_offload_commit (net); -if (err <0 ) -return err; -/* 1. Allocate space for next generation rules_gen_X[] */ list_for_each_entry_safe (trans, next, &nft_net->commit_list, list) {struct nft_table *table = trans->table;
@@ -11010,6 +11006,16 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb) } } +/* must be last, so audit and chain blob set up does not leave hardware +* in consistent state. +*/ + err =nft_flow_rule_offload_commit (net); +if (err <0 ) { +nf_tables_commit_chain_prepare_cancel (net); +nf_tables_commit_audit_free (&adl); +return err; + } +/* step 2. Make rules_gen_X visible to packet path */ nft_set_commit_update (&ctx, nft_net);
Signature · 1 line
2.47.3
Pablo Neira Ayuso
maintainerPATCH 4/10 netfilter: nft_set_pipapo_avx2: add missing vzeroupper
In reply to Pablo Neira AyusoFrom: Eric Biggers <[email protected]>
Since pipapo_get_avx2() uses YMM registers, execute vzeroupper before returning from it. This is needed to avoid degrading the performance of any later SSE code that may happen to be executed.
--- net/netfilter/nft_set_pipapo_avx2.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-)
net/netfilter/nft_set_pipapo_avx2.c +8 -9
@@ -1134,6 +1134,7 @@ struct nft_pipapo_elem *pipapo_get_avx2(const struct nft_pipapo_match *m,struct nft_pipapo_scratch *scratch;const struct nft_pipapo_field *f;unsigned long *res, *fill, *map; +struct nft_pipapo_elem *e;bool map_index;int ret =0 ;int i;
@@ -1207,14 +1208,11 @@ struct nft_pipapo_elem *pipapo_get_avx2(const struct nft_pipapo_match *m, next_match:if (ret <0 ) { scratch->map_index = map_index; -kernel_fpu_end (); -__local_unlock_nested_bh (&scratch->bh_lock); -return NULL ; + e =NULL ; + goto out; }if (last) { -struct nft_pipapo_elem *e; - e = f->mt[ret].e;if (unlikely (__nft_set_elem_expired (&e->ext, tstamp) || !nft_set_elem_active (&e->ext, genmask))) {
@@ -1224,9 +1222,7 @@ struct nft_pipapo_elem *pipapo_get_avx2(const struct nft_pipapo_match *m, } scratch->map_index = map_index; -kernel_fpu_end (); -__local_unlock_nested_bh (&scratch->bh_lock); -return e; + goto out; } map_index = !map_index;
@@ -1234,9 +1230,12 @@ struct nft_pipapo_elem *pipapo_get_avx2(const struct nft_pipapo_match *m, data +=NFT_PIPAPO_GROUPS_PADDED_SIZE (f); } + e =NULL ; +out: + asmvolatile ("vzeroupper" );kernel_fpu_end ();__local_unlock_nested_bh (&scratch->bh_lock); -return NULL ; +return e; }/**
Signature · 1 line
2.47.3
Pablo Neira Ayuso
maintainerPATCH 5/10 netfilter: x_tables: remove pr_debug
In reply to Pablo Neira AyusoRemove pr_debug() for these xtables extensions, these have no use these days. Still, turn pr_debug() into pr_info_ratelimited() in the .check path since this helps provide a hint via dmesg in legacy iptables.
--- net/ipv4/netfilter/ipt_ah.c | 10 +---- net/ipv6/netfilter/ip6t_ah.c | 27 +------------- net/ipv6/netfilter/ip6t_frag.c | 41 +------------------- net/ipv6/netfilter/ip6t_hbh.c | 40 +++++--------------- net/ipv6/netfilter/ip6t_mh.c | 3 -- net/ipv6/netfilter/ip6t_rt.c | 6 +-- net/netfilter/xt_IDLETIMER.c | 68 ++++++---------------------------- net/netfilter/xt_LOG.c | 4 +- net/netfilter/xt_MASQUERADE.c | 4 +- net/netfilter/xt_NETMAP.c | 4 +- net/netfilter/xt_REDIRECT.c | 4 +- net/netfilter/xt_esp.c | 10 +---- net/netfilter/xt_ipcomp.c | 8 +--- net/netfilter/xt_iprange.c | 32 ++-------------- net/netfilter/xt_ipvs.c | 1 - net/netfilter/xt_multiport.c | 4 -- net/netfilter/xt_sctp.c | 19 +--------- net/netfilter/xt_tcpudp.c | 8 +--- 18 files changed, 47 insertions(+), 246 deletions(-)
net/ipv4/netfilter/ipt_ah.c +2 -8
@@ -19,12 +19,7 @@ MODULE_DESCRIPTION("Xtables: IPv4 IPsec-AH SPI match");static inline bool spi_match (u_int32_t min,u_int32_t max,u_int32_t spi,bool invert) { -bool r; -pr_debug ("spi_match:%c 0x%x <= 0x%x <= 0x%x\n" , - invert ?'!' :' ' , min, spi, max); - r = (spi >= min && spi <= max) ^ invert; -pr_debug (" result %s\n" , r ?"PASS" :"FAILED" ); -return r; +return (spi >= min && spi <= max) ^ invert; }static bool ah_mt (const struct sk_buff *skb,struct xt_action_param *par)
@@ -42,7 +37,6 @@ static bool ah_mt(const struct sk_buff *skb, struct xt_action_param *par)/* We've been asked to examine this packet, and we * can't. Hence, no choice but to drop. */ -pr_debug ("Dropping evil AH tinygram.\n" ); par->hotdrop = true;return false; }
@@ -58,7 +52,7 @@ static int ah_mt_check(const struct xt_mtchk_param *par)/* Must specify no unknown invflags */ if (ahinfo->invflags & ~IPT_AH_INV_MASK ) { -pr_debug ("unknown flags %X\n" , ahinfo->invflags); +pr_info_ratelimited ("unknown flags %X\n" , ahinfo->invflags);return -EINVAL ; }return 0 ;
net/ipv6/netfilter/ip6t_ah.c +2 -25
@@ -24,13 +24,7 @@ MODULE_AUTHOR("Andras Kis-Szabo <[email protected]>");static inline bool spi_match (u_int32_t min,u_int32_t max,u_int32_t spi,bool invert) { -bool r; - -pr_debug ("spi_match:%c 0x%x <= 0x%x <= 0x%x\n" , - invert ?'!' :' ' , min, spi, max); - r = (spi >= min && spi <= max) ^ invert; -pr_debug (" result %s\n" , r ?"PASS" :"FAILED" ); -return r; +return (spi >= min && spi <= max) ^ invert; }static bool ah_mt6 (const struct sk_buff *skb,struct xt_action_param *par)
@@ -62,23 +56,6 @@ static bool ah_mt6(const struct sk_buff *skb, struct xt_action_param *par)return false; } -pr_debug ("IPv6 AH LEN %u %u " , hdrlen, ah->hdrlen); -pr_debug ("RES %04X " , ah->reserved); -pr_debug ("SPI %u %08X\n" ,ntohl (ah->spi),ntohl (ah->spi)); - -pr_debug ("IPv6 AH spi %02X " , -spi_match (ahinfo->spis[0 ], ahinfo->spis[1 ], -ntohl (ah->spi), - !!(ahinfo->invflags &IP6T_AH_INV_SPI ))); -pr_debug ("len %02X %04X %02X " , - ahinfo->hdrlen, hdrlen, - (!ahinfo->hdrlen || - (ahinfo->hdrlen == hdrlen) ^ - !!(ahinfo->invflags &IP6T_AH_INV_LEN ))); -pr_debug ("res %02X %04X %02X\n" , - ahinfo->hdrres, ah->reserved, - !(ahinfo->hdrres && ah->reserved)); -return spi_match (ahinfo->spis[0 ], ahinfo->spis[1 ],ntohl (ah->spi), !!(ahinfo->invflags & IP6T_AH_INV_SPI)) &&
@@ -93,7 +70,7 @@ static int ah_mt6_check(const struct xt_mtchk_param *par)const struct ip6t_ah *ahinfo = par->matchinfo;if (ahinfo->invflags & ~IP6T_AH_INV_MASK ) { -pr_debug ("unknown flags %X\n" , ahinfo->invflags); +pr_info_ratelimited ("unknown flags %X\n" , ahinfo->invflags);return -EINVAL ; }return 0 ;
net/ipv6/netfilter/ip6t_frag.c +2 -39
59 lines not on this page · raw message
net/ipv6/netfilter/ip6t_hbh.c +9 -31
@@ -79,14 +79,6 @@ hbh_mt6(const struct sk_buff *skb, struct xt_action_param *par)return false; } -pr_debug ("IPv6 OPTS LEN %u %u " , hdrlen, oh->hdrlen); - -pr_debug ("len %02X %04X %02X " , - optinfo->hdrlen, hdrlen, - (!(optinfo->flags &IP6T_OPTS_LEN ) || - ((optinfo->hdrlen == hdrlen) ^ - !!(optinfo->invflags &IP6T_OPTS_INV_LEN )))); - ret = (!(optinfo->flags &IP6T_OPTS_LEN ) || ((optinfo->hdrlen == hdrlen) ^ !!(optinfo->invflags &IP6T_OPTS_INV_LEN )));
@@ -96,8 +88,6 @@ hbh_mt6(const struct sk_buff *skb, struct xt_action_param *par)if (!(optinfo->flags &IP6T_OPTS_OPTS )) {return ret; }else { -pr_debug ("Strict " ); -pr_debug ("#%d " , optinfo->optsnr);for (temp =0 ; temp < optinfo->optsnr; temp++) {/* type field exists ? */ if (hdrlen <1 )
@@ -108,13 +98,9 @@ hbh_mt6(const struct sk_buff *skb, struct xt_action_param *par)break ;/* Type check */ -if (*tp != (optinfo->opts[temp] &0xFF00 ) >>8 ) { -pr_debug ("Tbad %02X %02X\n" , *tp, - (optinfo->opts[temp] &0xFF00 ) >>8 ); +if (*tp != (optinfo->opts[temp] &0xFF00 ) >>8 )return false; - }else { -pr_debug ("Tok " ); - } +/* Length check */ if (*tp) {u16 spec_len;
@@ -129,26 +115,18 @@ hbh_mt6(const struct sk_buff *skb, struct xt_action_param *par)break ; spec_len = optinfo->opts[temp] &0x00FF ; -if (spec_len !=0x00FF && spec_len != *lp) { -pr_debug ("Lbad %02X %04X\n" , *lp, - spec_len); +if (spec_len !=0x00FF && spec_len != *lp)return false; - } -pr_debug ("Lok " ); + optlen = *lp +2 ; }else { -pr_debug ("Pad1\n" ); optlen =1 ; } -/* Step to the next */ -pr_debug ("len%04X\n" , optlen); -if ((ptr > skb->len - optlen || hdrlen < optlen) && - temp < optinfo->optsnr -1 ) { -pr_debug ("new pointer is too large!\n" ); + temp < optinfo->optsnr -1 )break ; - } + ptr += optlen; hdrlen -= optlen; }
@@ -166,16 +144,16 @@ static int hbh_mt6_check(const struct xt_mtchk_param *par)const struct ip6t_opts *optsinfo = par->matchinfo;if (optsinfo->invflags & ~IP6T_OPTS_INV_MASK ) { -pr_debug ("unknown flags %X\n" , optsinfo->invflags); +pr_info_ratelimited ("unknown flags %X\n" , optsinfo->invflags);return -EINVAL ; }if (optsinfo->optsnr >IP6T_OPTS_OPTSNR ) { -pr_debug ("too many supported opts specified\n" ); +pr_info_ratelimited ("too many supported opts specified\n" );return -EINVAL ; }if (optsinfo->flags &IP6T_OPTS_NSTRICT ) { -pr_debug ("Not strict - not implemented" ); +pr_info_ratelimited ("Not strict - not implemented" );return -EINVAL ; }
net/ipv6/netfilter/ip6t_mh.c +0 -3
@@ -42,14 +42,11 @@ static bool mh_mt6(const struct sk_buff *skb, struct xt_action_param *par)if (mh ==NULL ) {/* We've been asked to examine this packet, and we can't. Hence, no choice but to drop. */ -pr_debug ("Dropping evil MH tinygram.\n" ); par->hotdrop = true;return false; }if (mh->ip6mh_proto !=IPPROTO_NONE ) { -pr_debug ("Dropping invalid MH Payload Proto: %u\n" , - mh->ip6mh_proto); par->hotdrop = true;return false; }
net/ipv6/netfilter/ip6t_rt.c +3 -3
@@ -155,18 +155,18 @@ static int rt_mt6_check(const struct xt_mtchk_param *par)const struct ip6t_rt *rtinfo = par->matchinfo;if (rtinfo->invflags & ~IP6T_RT_INV_MASK ) { -pr_debug ("unknown flags %X\n" , rtinfo->invflags); +pr_info_ratelimited ("unknown flags %X\n" , rtinfo->invflags);return -EINVAL ; }if (rtinfo->addrnr >IP6T_RT_HOPS ) { -pr_debug ("too many addresses specified\n" ); +pr_info_ratelimited ("too many addresses specified\n" );return -EINVAL ; }if ((rtinfo->flags & (IP6T_RT_RES |IP6T_RT_FST_MASK )) && (!(rtinfo->flags &IP6T_RT_TYP ) || (rtinfo->rt_type !=0 ) || (rtinfo->invflags &IP6T_RT_INV_TYP ))) { -pr_debug ("`--rt-type 0' required before `--rt-0-*'" ); +pr_info_ratelimited ("`--rt-type 0' required before `--rt-0-*'" );return -EINVAL ; }
net/netfilter/xt_IDLETIMER.c +12 -56
227 lines not on this page · raw message
net/netfilter/xt_LOG.c +2 -2
@@ -50,12 +50,12 @@ static int log_tg_check(const struct xt_tgchk_param *par)return -EINVAL ;if (loginfo->level >=8 ) { -pr_debug ("level %u >= 8\n" , loginfo->level); +pr_info_ratelimited ("level %u >= 8\n" , loginfo->level);return -EINVAL ; }if (loginfo->prefix[sizeof (loginfo->prefix)-1 ] !='\0' ) { -pr_debug ("prefix is not null-terminated\n" ); +pr_info_ratelimited ("prefix is not null-terminated\n" );return -EINVAL ; }
net/netfilter/xt_MASQUERADE.c +2 -2
@@ -21,11 +21,11 @@ static int masquerade_tg_check(const struct xt_tgchk_param *par)const struct nf_nat_ipv4_multi_range_compat *mr = par->targinfo;if (mr->range[0 ].flags &NF_NAT_RANGE_MAP_IPS ) { -pr_debug ("bad MAP_IPS.\n" ); +pr_info_ratelimited ("bad MAP_IPS.\n" );return -EINVAL ; }if (mr->rangesize !=1 ) { -pr_debug ("bad rangesize %u\n" , mr->rangesize); +pr_info_ratelimited ("bad rangesize %u\n" , mr->rangesize);return -EINVAL ; }return nf_ct_netns_get (par->net, par->family);
net/netfilter/xt_NETMAP.c +2 -2
@@ -106,11 +106,11 @@ static int netmap_tg4_check(const struct xt_tgchk_param *par)const struct nf_nat_ipv4_multi_range_compat *mr = par->targinfo;if (!(mr->range[0 ].flags &NF_NAT_RANGE_MAP_IPS )) { -pr_debug ("bad MAP_IPS.\n" ); +pr_info_ratelimited ("bad MAP_IPS.\n" );return -EINVAL ; }if (mr->rangesize !=1 ) { -pr_debug ("bad rangesize %u.\n" , mr->rangesize); +pr_info_ratelimited ("bad rangesize %u.\n" , mr->rangesize);return -EINVAL ; }return nf_ct_netns_get (par->net, par->family);
net/netfilter/xt_REDIRECT.c +2 -2
@@ -51,11 +51,11 @@ static int redirect_tg4_check(const struct xt_tgchk_param *par)const struct nf_nat_ipv4_multi_range_compat *mr = par->targinfo;if (mr->range[0 ].flags &NF_NAT_RANGE_MAP_IPS ) { -pr_debug ("bad MAP_IPS.\n" ); +pr_info_ratelimited ("bad MAP_IPS.\n" );return -EINVAL ; }if (mr->rangesize !=1 ) { -pr_debug ("bad rangesize %u.\n" , mr->rangesize); +pr_info_ratelimited ("bad rangesize %u.\n" , mr->rangesize);return -EINVAL ; }return nf_ct_netns_get (par->net, par->family);
net/netfilter/xt_esp.c +2 -8
@@ -25,12 +25,7 @@ MODULE_ALIAS("ip6t_esp");static inline bool spi_match (u_int32_t min,u_int32_t max,u_int32_t spi,bool invert) { -bool r; -pr_debug ("spi_match:%c 0x%x <= 0x%x <= 0x%x\n" , - invert ?'!' :' ' , min, spi, max); - r = (spi >= min && spi <= max) ^ invert; -pr_debug (" result %s\n" , r ?"PASS" :"FAILED" ); -return r; +return (spi >= min && spi <= max) ^ invert; }static bool esp_mt (const struct sk_buff *skb,struct xt_action_param *par)
@@ -48,7 +43,6 @@ static bool esp_mt(const struct sk_buff *skb, struct xt_action_param *par)/* We've been asked to examine this packet, and we * can't. Hence, no choice but to drop. */ -pr_debug ("Dropping evil ESP tinygram.\n" ); par->hotdrop = true;return false; }
@@ -62,7 +56,7 @@ static int esp_mt_check(const struct xt_mtchk_param *par)const struct xt_esp *espinfo = par->matchinfo;if (espinfo->invflags & ~XT_ESP_INV_MASK ) { -pr_debug ("unknown flags %X\n" , espinfo->invflags); +pr_info_ratelimited ("unknown flags %X\n" , espinfo->invflags);return -EINVAL ; }
net/netfilter/xt_ipcomp.c +1 -7
@@ -29,12 +29,7 @@ MODULE_ALIAS("ip6t_ipcomp");static inline bool spi_match (u_int32_t min,u_int32_t max,u_int32_t spi,bool invert) { -bool r; -pr_debug ("spi_match:%c 0x%x <= 0x%x <= 0x%x\n" , - invert ?'!' :' ' , min, spi, max); - r = (spi >= min && spi <= max) ^ invert; -pr_debug (" result %s\n" , r ?"PASS" :"FAILED" ); -return r; +return (spi >= min && spi <= max) ^ invert; }static bool comp_mt (const struct sk_buff *skb,struct xt_action_param *par)
@@ -52,7 +47,6 @@ static bool comp_mt(const struct sk_buff *skb, struct xt_action_param *par)/* We've been asked to examine this packet, and we * can't. Hence, no choice but to drop. */ -pr_debug ("Dropping evil IPComp tinygram.\n" ); par->hotdrop = true;return false; }
net/netfilter/xt_iprange.c +4 -28
@@ -24,27 +24,15 @@ iprange_mt4(const struct sk_buff *skb, struct xt_action_param *par) m =ntohl (iph->saddr) <ntohl (info->src_min.ip); m |=ntohl (iph->saddr) >ntohl (info->src_max.ip); m ^= !!(info->flags &IPRANGE_SRC_INV ); -if (m) { -pr_debug ("src IP %pI4 NOT in range %s%pI4-%pI4\n" , - &iph->saddr, - (info->flags &IPRANGE_SRC_INV ) ?"(INV) " :"" , - &info->src_min.ip, - &info->src_max.ip); +if (m)return false; - } }if (info->flags &IPRANGE_DST ) { m =ntohl (iph->daddr) <ntohl (info->dst_min.ip); m |=ntohl (iph->daddr) >ntohl (info->dst_max.ip); m ^= !!(info->flags &IPRANGE_DST_INV ); -if (m) { -pr_debug ("dst IP %pI4 NOT in range %s%pI4-%pI4\n" , - &iph->daddr, - (info->flags &IPRANGE_DST_INV ) ?"(INV) " :"" , - &info->dst_min.ip, - &info->dst_max.ip); +if (m)return false; - } }return true; }
@@ -73,27 +61,15 @@ iprange_mt6(const struct sk_buff *skb, struct xt_action_param *par) m =iprange_ipv6_lt (&iph->saddr, &info->src_min.in6); m |=iprange_ipv6_lt (&info->src_max.in6, &iph->saddr); m ^= !!(info->flags &IPRANGE_SRC_INV ); -if (m) { -pr_debug ("src IP %pI6 NOT in range %s%pI6-%pI6\n" , - &iph->saddr, - (info->flags &IPRANGE_SRC_INV ) ?"(INV) " :"" , - &info->src_min.in6, - &info->src_max.in6); +if (m)return false; - } }if (info->flags &IPRANGE_DST ) { m =iprange_ipv6_lt (&iph->daddr, &info->dst_min.in6); m |=iprange_ipv6_lt (&info->dst_max.in6, &iph->daddr); m ^= !!(info->flags &IPRANGE_DST_INV ); -if (m) { -pr_debug ("dst IP %pI6 NOT in range %s%pI6-%pI6\n" , - &iph->daddr, - (info->flags &IPRANGE_DST_INV ) ?"(INV) " :"" , - &info->dst_min.in6, - &info->dst_max.in6); +if (m)return false; - } }return true; }
net/netfilter/xt_ipvs.c +0 -1
@@ -148,7 +148,6 @@ ipvs_mt(const struct sk_buff *skb, struct xt_action_param *par) out_put_cp:__ip_vs_conn_put (cp); out: -pr_debug ("match=%d\n" , match);return match; }
net/netfilter/xt_multiport.c +0 -4
@@ -37,7 +37,6 @@ ports_match_v1(const struct xt_multiport_v1 *minfo,if (minfo->pflags[i]) {/* range port matching */ e = minfo->ports[++i]; -pr_debug ("src or dst matches with %d-%d?\n" , s, e);switch (minfo->flags) {case XT_MULTIPORT_SOURCE :
@@ -58,8 +57,6 @@ ports_match_v1(const struct xt_multiport_v1 *minfo, } }else {/* exact port matching */ -pr_debug ("src or dst matches with %d?\n" , s); -switch (minfo->flags) {case XT_MULTIPORT_SOURCE :if (src == s)
@@ -97,7 +94,6 @@ multiport_mt(const struct sk_buff *skb, struct xt_action_param *par)/* We've been asked to examine this packet, and we * can't. Hence, no choice but to drop. */ -pr_debug ("Dropping evil offset=0 tinygram.\n" ); par->hotdrop = true;return false; }
net/netfilter/xt_sctp.c +1 -18
@@ -48,30 +48,17 @@ match_packet(const struct sk_buff *skb,const struct xt_sctp_flag_info *flag_info = info->flag_info;int flag_count = info->flag_count; -#ifdef DEBUG -int i =0 ; -#endif -if (chunk_match_type ==SCTP_CHUNK_MATCH_ALL )SCTP_CHUNKMAP_COPY (chunkmapcopy, info->chunkmap); do { sch = skb_header_pointer(skb, offset, sizeof(_sch), &_sch); if (sch == NULL || sch->length == 0) { - pr_debug("Dropping invalid SCTP packet.\n"); *hotdrop = true; return false; } -#ifdef DEBUG - pr_debug("Chunk num: %d\toffset: %d\ttype: %d\tlength: %d" - "\tflags: %x\n", - ++i, offset, sch->type, htons(sch->length), - sch->flags); -#endif offset += SCTP_PAD4(ntohs(sch->length)); - pr_debug("skb->len: %d\toffset: %d\n", skb->len, offset); - if (SCTP_CHUNKMAP_IS_SET(info->chunkmap, sch->type)) { switch (chunk_match_type) { case SCTP_CHUNK_MATCH_ANY:
@@ -121,18 +108,14 @@ sctp_mt(const struct sk_buff *skb, struct xt_action_param *par)const struct sctphdr *sh;struct sctphdr _sh; -if (par->fragoff !=0 ) { -pr_debug ("Dropping non-first fragment.. FIXME\n" ); +if (par->fragoff !=0 )return false; - } sh =skb_header_pointer (skb, par->thoff,sizeof (_sh), &_sh);if (sh ==NULL ) { -pr_debug ("Dropping evil TCP offset=0 tinygram.\n" ); par->hotdrop = true;return false; } -pr_debug ("spt: %d\tdpt: %d\n" ,ntohs (sh->source),ntohs (sh->dest)); return SCCHECK(ntohs(sh->source) >= info->spts[0] && ntohs(sh->source) <= info->spts[1],
net/netfilter/xt_tcpudp.c +1 -7
@@ -44,8 +44,6 @@ tcp_find_option(u_int8_t option,u_int8_t _opt[60 -sizeof (struct tcphdr )];unsigned int i; -pr_debug ("finding option\n" ); -if (!optlen)return invert;
@@ -81,10 +79,8 @@ static bool tcp_mt(const struct sk_buff *skb, struct xt_action_param *par)causes this. Its a cracker trying to break in by doing a flag overwrite to pass the direction checks. */ -if (par->fragoff ==1 ) { -pr_debug ("Dropping evil TCP offset=1 frag.\n" ); +if (par->fragoff ==1 ) par->hotdrop = true; - }/* Must not be a fragment. */ return false; }
@@ -93,7 +89,6 @@ static bool tcp_mt(const struct sk_buff *skb, struct xt_action_param *par)if (th ==NULL ) {/* We've been asked to examine this packet, and we can't. Hence, no choice but to drop. */ -pr_debug ("Dropping evil TCP offset=0 tinygram.\n" ); par->hotdrop = true;return false; }
@@ -145,7 +140,6 @@ static bool udp_mt(const struct sk_buff *skb, struct xt_action_param *par)if (uh ==NULL ) {/* We've been asked to examine this packet, and we can't. Hence, no choice but to drop. */ -pr_debug ("Dropping evil UDP tinygram.\n" ); par->hotdrop = true;return false; }
Signature · 1 line
2.47.3
Jakub Kicinski
maintainerRe: [PATCH net 05/10] netfilter: x_tables: remove pr_debug
This is an AI-generated review of your patch. The human sending this email has considered the AI review valid, or at least plausible.
---
netfilter: x_tables: remove pr_debug
This drops the pr_debug() calls from a set of xtables matches and targets, and converts the ones in the .check paths into pr_info_ratelimited() so that legacy iptables users get a hint in dmesg when a rule is rejected.
A few observations below.
Quoted 4 lines
net/ipv6/netfilter/ip6t_hbh.c +0
-0
[ ... ]
Quoted 9 lines
net/ipv6/netfilter/ip6t_hbh.c +1
-1@@ -166,16 +144,16 @@ static int hbh_mt6_check(const struct xt_mtchk_param *par)const struct ip6t_opts *optsinfo = par->matchinfo;if (optsinfo->invflags & ~IP6T_OPTS_INV_MASK ) { -pr_debug ("unknown flags %X\n" , optsinfo->invflags); +pr_info_ratelimited ("unknown flags %X\n" , optsinfo->invflags);return -EINVAL ; }if (optsinfo->optsnr >IP6T_OPTS_OPTSNR ) {
net/ipv6/netfilter/ip6t_hbh.c static int hbh_mt6_check(const struct xt_mtchk_param *par) · line 158
Quoted 10 lines
net/ipv6/netfilter/ip6t_hbh.c +2
-2-pr_debug ("too many supported opts specified\n" ); +pr_info_ratelimited ("too many supported opts specified\n" );return -EINVAL ; }if (optsinfo->flags &IP6T_OPTS_NSTRICT ) { -pr_debug ("Not strict - not implemented" ); +pr_info_ratelimited ("Not strict - not implemented" );return -EINVAL ; }
[Severity: Low]
Should this string get a trailing newline like its two siblings in hbh_mt6_check()?
The missing '\n' was harmless while this was pr_debug() (compiled out or dynamic-debug gated), but as an unconditional pr_info_ratelimited() the record is left open for continuation:
kernel/printk/printk.c:printk_sprint() {
/* Mark and strip a trailing newline. */
if (text_len && text[text_len - 1] == '\n') {
text_len--;
*flags |= LOG_NEWLINE;
}
}kernel/printk/printk.c:vprintk_store() {
/* A message without a trailing newline can be continued. */
if (!(flags & LOG_NEWLINE))
prb_commit(&e);
else
prb_final_commit(&e);
}So an iptables-legacy rule with IP6T_OPTS_NSTRICT set produces a dmesg line that gets glued to whatever is printed next, which works against the "hint via dmesg" goal in the changelog.
The same missing '\n' appears in three more of the converted strings, noted below.
Quoted 15 lines
net/ipv6/netfilter/ip6t_rt.c +2
-2@@ -155,18 +155,18 @@ static int rt_mt6_check(const struct xt_mtchk_param *par)const struct ip6t_rt *rtinfo = par->matchinfo;if (rtinfo->invflags & ~IP6T_RT_INV_MASK ) { -pr_debug ("unknown flags %X\n" , rtinfo->invflags); +pr_info_ratelimited ("unknown flags %X\n" , rtinfo->invflags);return -EINVAL ; }if (rtinfo->addrnr >IP6T_RT_HOPS ) { -pr_debug ("too many addresses specified\n" ); +pr_info_ratelimited ("too many addresses specified\n" );
net/ipv6/netfilter/ip6t_rt.c static int rt_mt6_check(const struct xt_mtchk_param *par) · line 171
Quoted 10 lines
net/ipv6/netfilter/ip6t_rt.c +1
-1return -EINVAL ; }if ((rtinfo->flags & (IP6T_RT_RES |IP6T_RT_FST_MASK )) && (!(rtinfo->flags &IP6T_RT_TYP ) || (rtinfo->rt_type !=0 ) || (rtinfo->invflags &IP6T_RT_INV_TYP ))) { -pr_debug ("`--rt-type 0' required before `--rt-0-*'" ); +pr_info_ratelimited ("`--rt-type 0' required before `--rt-0-*'" );return -EINVAL ; }
[Severity: Low]
Same question here in rt_mt6_check(): the two messages above it end in '\n' but this one does not, so a rule with IP6T_RT_RES or IP6T_RT_FST_MASK and rt_type != 0 leaves the record un-terminated.
Quoted 4 lines
net/netfilter/xt_IDLETIMER.c +0
-0
[ ... ]
net/netfilter/xt_IDLETIMER.c static int idletimer_tg_create(struct idletimer_tg_info *info) · line 173
Quoted 8 lines
net/netfilter/xt_IDLETIMER.c +1
-1@@ -171,7 +168,7 @@ static int idletimer_tg_create(struct idletimer_tg_info *info) ret =sysfs_create_file (idletimer_tg_kobj, &info->timer->attr.attr);if (ret <0 ) { -pr_debug ("couldn't add file to sysfs" ); +pr_info_ratelimited ("couldn't add file to sysfs" ); goto out_free_attr; }
[Severity: Low]
Would it be worth adding the '\n' here in idletimer_tg_create() too? This one is reached from idletimer_tg_checkentry() when sysfs_create_file() fails, and the other strings converted in this file
("timeout value is zero\n", "label is empty or not nul-terminated\n") are newline terminated.
net/netfilter/xt_IDLETIMER.c static int idletimer_tg_create_v1(struct idletimer_tg_info_v1 *info) · line 222
Quoted 8 lines
net/netfilter/xt_IDLETIMER.c +1
-1@@ -220,7 +217,7 @@ static int idletimer_tg_create_v1(struct idletimer_tg_info_v1 *info) ret =sysfs_create_file (idletimer_tg_kobj, &info->timer->attr.attr);if (ret <0 ) { -pr_debug ("couldn't add file to sysfs" ); +pr_info_ratelimited ("couldn't add file to sysfs" ); goto out_free_attr; }
[Severity: Low]
And the second copy of the same string, in idletimer_tg_create_v1(), reached from idletimer_tg_checkentry_v1().
[ ... ]
Quoted 19 lines
net/netfilter/xt_IDLETIMER.c +2
-2@@ -534,7 +490,7 @@ static int __init idletimer_tg_init(void) idletimer_tg_class =class_create ("xt_idletimer" ); err =PTR_ERR (idletimer_tg_class);if (IS_ERR (idletimer_tg_class)) { -pr_debug ("couldn't register device class\n" ); +pr_info_ratelimited ("couldn't register device class\n" ); goto out; }@@ -542,7 +498,7 @@ static int __init idletimer_tg_init(void) MKDEV(0, 0), NULL, "timers"); err =PTR_ERR (idletimer_tg_device);if (IS_ERR (idletimer_tg_device)) { -pr_debug ("couldn't register system device\n" ); +pr_info_ratelimited ("couldn't register system device\n" ); goto out_class; }
net/netfilter/xt_IDLETIMER.c static int __init idletimer_tg_init(void) · line 512
Quoted 7 lines
net/netfilter/xt_IDLETIMER.c +1
-1@@ -551,7 +507,7 @@ static int __init idletimer_tg_init(void) err =xt_register_targets (idletimer_tg,ARRAY_SIZE (idletimer_tg));if (err <0 ) { -pr_debug ("couldn't register xt target\n" ); +pr_info_ratelimited ("couldn't register xt target\n" ); goto out_dev; }
[Severity: Low]
The changelog says:
Still, turn pr_debug() into pr_info_ratelimited() in the .check path since this helps provide a hint via dmesg in legacy iptables.
These three conversions are in idletimer_tg_init(), which is __init and not an xtables checkentry callback, so they fall outside the ".check path" described above.
Were these intended? Ratelimiting can never trigger on a one-shot module init failure, and pr_err() would be the usual level for class_create(), device_create() and xt_register_targets() failing.
Quoted 8 lines
net/netfilter/xt_NETMAP.c +0
-0@@ -106,11 +106,11 @@ static int netmap_tg4_check(const struct xt_tgchk_param *par)const struct nf_nat_ipv4_multi_range_compat *mr = par->targinfo;if (!(mr->range[0 ].flags &NF_NAT_RANGE_MAP_IPS )) {
net/netfilter/xt_NETMAP.c static int netmap_tg4_check(const struct xt_tgchk_param *par) · line 116
Quoted 10 lines
net/netfilter/xt_NETMAP.c +2
-2-pr_debug ("bad MAP_IPS.\n" ); +pr_info_ratelimited ("bad MAP_IPS.\n" );return -EINVAL ; }if (mr->rangesize !=1 ) { -pr_debug ("bad rangesize %u.\n" , mr->rangesize); +pr_info_ratelimited ("bad rangesize %u.\n" , mr->rangesize);return -EINVAL ; }return nf_ct_netns_get (par->net, par->family);
[Severity: Low]
Do these new messages need a pr_fmt() in this file to be identifiable?
xt_NETMAP.c and xt_REDIRECT.c do not define pr_fmt(), so the fallback in include/linux/printk.h applies:
#define pr_fmt(fmt) fmt
while the peer site does define one:
net/netfilter/xt_MASQUERADE.c:
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
and prints the byte-identical strings "bad MAP_IPS." and "bad rangesize %u". The result is that MASQUERADE reports "xt_MASQUERADE: bad MAP_IPS." while NETMAP and REDIRECT report a bare "bad MAP_IPS.", so dmesg no longer says which target rejected the rule. With pr_debug() the dynamic debug machinery supplied module, file and function, so this did not matter before.
Quoted 8 lines
net/netfilter/xt_REDIRECT.c +0
-0@@ -51,11 +51,11 @@ static int redirect_tg4_check(const struct xt_tgchk_param *par)const struct nf_nat_ipv4_multi_range_compat *mr = par->targinfo;if (mr->range[0 ].flags &NF_NAT_RANGE_MAP_IPS ) {
net/netfilter/xt_REDIRECT.c static int redirect_tg4_check(const struct xt_tgchk_param *par) · line 61
Quoted 10 lines
net/netfilter/xt_REDIRECT.c +2
-2-pr_debug ("bad MAP_IPS.\n" ); +pr_info_ratelimited ("bad MAP_IPS.\n" );return -EINVAL ; }if (mr->rangesize !=1 ) { -pr_debug ("bad rangesize %u.\n" , mr->rangesize); +pr_info_ratelimited ("bad rangesize %u.\n" , mr->rangesize);return -EINVAL ; }return nf_ct_netns_get (par->net, par->family);
[Severity: Low]
Same for redirect_tg4_check() in xt_REDIRECT.c, which also has no pr_fmt() definition.
Pablo Neira Ayuso
maintainerRe: [PATCH net 05/10] netfilter: x_tables: remove pr_debug
Quoted 33 lines · Jakub Kicinski
This is an AI-generated review of your patch. The human sending this email has considered the AI review valid, or at least plausible.
---
netfilter:x_tables: removepr_debugThis drops the
pr_debug()calls from a set of xtables matches and targets, and converts the ones in the .check paths intopr_info_ratelimited()so that legacy iptables users get a hint in dmesg when a rule is rejected.A few observations below.
Quoted 4 lines
net/ipv6/netfilter/ip6t_hbh.c +0
-0[ ... ]
Quoted 14 lines
net/ipv6/netfilter/ip6t_hbh.c +2
-2@@ -166,16 +144,16 @@ static int hbh_mt6_check(const struct xt_mtchk_param *par)const struct ip6t_opts *optsinfo = par->matchinfo;if (optsinfo->invflags & ~IP6T_OPTS_INV_MASK ) { -pr_debug ("unknown flags %X\n" , optsinfo->invflags); +pr_info_ratelimited ("unknown flags %X\n" , optsinfo->invflags);return -EINVAL ; }if (optsinfo->optsnr >IP6T_OPTS_OPTSNR ) { -pr_debug ("too many supported opts specified\n" ); +pr_info_ratelimited ("too many supported opts specified\n" );return -EINVAL ; }
Quoted 9 lines
net/ipv6/netfilter/ip6t_hbh.c static int hbh_mt6_check(const struct xt_mtchk_param *par) · line 158
Quoted 5 lines
net/ipv6/netfilter/ip6t_hbh.c +1
-1if (optsinfo->flags &IP6T_OPTS_NSTRICT ) { -pr_debug ("Not strict - not implemented" ); +pr_info_ratelimited ("Not strict - not implemented" );return -EINVAL ; }[Severity: Low]
Should this string get a trailing newline like its two siblings inhbh_mt6_check()?
Yes, this needs \n in all these pr_debug() to pr_info_ratelimited().
Pablo Neira Ayuso
maintainerPATCH 6/10 netfilter: xt_cgroup: use pr_info_ratelimited()
In reply to Pablo Neira AyusoThis seems to be the last xtables extension not using the ratelimited version, update it.
--- net/netfilter/xt_cgroup.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)
net/netfilter/xt_cgroup.c +5 -5
@@ -33,7 +33,7 @@ static int cgroup_mt_check_v0(const struct xt_mtchk_param *par)return -EINVAL ;if (!IS_ENABLED (CONFIG_CGROUP_NET_CLASSID )) { -pr_info (NET_CLS_CLASSID_INVALID_MSG ); +pr_info_ratelimited (NET_CLS_CLASSID_INVALID_MSG );return -EINVAL ; }
@@ -49,7 +49,7 @@ static int cgroup_mt_check_v1(const struct xt_mtchk_param *par)return -EINVAL ;if (!info->has_path && !info->has_classid) { -pr_info ("xt_cgroup: no path or classid specified\n" ); +pr_info_ratelimited ("xt_cgroup: no path or classid specified\n" );return -EINVAL ; }
@@ -59,7 +59,7 @@ static int cgroup_mt_check_v1(const struct xt_mtchk_param *par) }if (info->has_classid && !IS_ENABLED (CONFIG_CGROUP_NET_CLASSID )) { -pr_info (NET_CLS_CLASSID_INVALID_MSG ); +pr_info_ratelimited (NET_CLS_CLASSID_INVALID_MSG );return -EINVAL ; }
@@ -89,7 +89,7 @@ static int cgroup_mt_check_v2(const struct xt_mtchk_param *par)return -EINVAL ;if (!info->has_path && !info->has_classid) { -pr_info ("xt_cgroup: no path or classid specified\n" ); +pr_info_ratelimited ("xt_cgroup: no path or classid specified\n" );return -EINVAL ; }
@@ -99,7 +99,7 @@ static int cgroup_mt_check_v2(const struct xt_mtchk_param *par) }if (info->has_classid && !IS_ENABLED (CONFIG_CGROUP_NET_CLASSID )) { -pr_info (NET_CLS_CLASSID_INVALID_MSG ); +pr_info_ratelimited (NET_CLS_CLASSID_INVALID_MSG );return -EINVAL ; }
Signature · 1 line
2.47.3
Pablo Neira Ayuso
maintainerPATCH 7/10 netfilter: nf_tables: skip double clone set expressions on element insert
In reply to Pablo Neira AyusoBoth the dynset and newsetelem path clone the existing set expressions when setting set element expressions if no override expressions are provided. This results in a double clone, once to clone the template set expressions then another clone on the new element. Add a flag to annotate if userspace provides a override expression (ie. expression of the same type of the set but different configuration), otherwise borrow the existing expression from the set. Add conditionals to release expression iif they represent an override. Use this new override_exprs flag to dump the dynset expression override to userspace.
This simplifies the existing logic and it also fixes a bug with the connlimit expression which results in a module refcount imbalance WARNING splat when resorting on the default set expressions.
--- include/net/netfilter/nf_tables.h | 2 -- net/netfilter/nf_tables_api.c | 56 ++++++++++--------------------- net/netfilter/nft_dynset.c | 25 ++++++++------ 3 files changed, 33 insertions(+), 50 deletions(-)
include/net/netfilter/nf_tables.h +0 -2
@@ -870,8 +870,6 @@ struct nft_elem_priv *nft_set_elem_init(const struct nft_set *set, const u32 *key, const u32 *key_end, const u32 *data, u64 timeout, u64 expiration, gfp_t gfp); -int nft_set_elem_expr_clone (const struct nft_ctx *ctx,struct nft_set *set, -struct nft_expr *expr_array[]);void nft_set_elem_expr_destroy (const struct nft_ctx *ctx,struct nft_set_elem_expr *elem_expr); void nft_set_elem_destroy(const struct nft_set *set,
net/netfilter/nf_tables_api.c +18 -38
114 lines not on this page · raw message
net/netfilter/nft_dynset.c +15 -10
@@ -19,7 +19,8 @@ struct nft_dynset {u8 sreg_key;u8 sreg_data;bool invert; -bool expr; + bool expr:1, + override_exprs:1;u8 num_exprs;u64 timeout;struct nft_expr *expr_array[NFT_SET_EXPR_MAX ];
@@ -257,6 +258,7 @@ static int nft_dynset_init(const struct nft_ctx *ctx, priv->num_exprs++; priv->expr_array[0 ] = dynset_expr; + priv->override_exprs = true; if (set->num_exprs > 1 || (set->num_exprs == 1 &&
@@ -289,6 +291,7 @@ static int nft_dynset_init(const struct nft_ctx *ctx, } priv->expr_array[i] = dynset_expr; priv->num_exprs++; + priv->override_exprs = true;if (set->num_exprs) {if (i >= set->num_exprs) {
@@ -307,9 +310,8 @@ static int nft_dynset_init(const struct nft_ctx *ctx, goto err_expr_free; } }else if (set->num_exprs >0 ) { - err =nft_set_elem_expr_clone (ctx, set, priv->expr_array); -if (err <0 ) -return err; +for (i =0 ; i < set->num_exprs; i++) + priv->expr_array[i] = set->exprs[i]; priv->num_exprs = set->num_exprs; }
@@ -339,8 +341,10 @@ static int nft_dynset_init(const struct nft_ctx *ctx,return 0 ; err_expr_free: -for (i =0 ; i < priv->num_exprs; i++) -nft_expr_destroy (ctx, priv->expr_array[i]); +if (priv->override_exprs) { +for (i =0 ; i < priv->num_exprs; i++) +nft_expr_destroy (ctx, priv->expr_array[i]); + }return err; }
@@ -367,9 +371,10 @@ static void nft_dynset_destroy(const struct nft_ctx *ctx,struct nft_dynset *priv =nft_expr_priv (expr);int i; -for (i =0 ; i < priv->num_exprs; i++) -nft_expr_destroy (ctx, priv->expr_array[i]); - +if (priv->override_exprs) { +for (i =0 ; i < priv->num_exprs; i++) +nft_expr_destroy (ctx, priv->expr_array[i]); + }nf_tables_destroy_set (ctx, priv->set); }
@@ -393,7 +398,7 @@ static int nft_dynset_dump(struct sk_buff *skb, nf_jiffies64_to_msecs(priv->timeout), NFTA_DYNSET_PAD)) goto nla_put_failure; -if (priv->set->num_exprs ==0 ) { +if (priv->set->num_exprs ==0 || priv->override_exprs) {if (priv->num_exprs ==1 ) {if (nft_expr_dump (skb,NFTA_DYNSET_EXPR , priv->expr_array[0 ], reset))
Signature · 1 line
2.47.3
Jakub Kicinski
maintainerRe: [PATCH net 07/10] netfilter: nf_tables: skip double clone set expressions on element insert
This is an AI-generated review of your patch. The human sending this email has considered the AI review valid, or at least plausible.
---
netfilter: nf_tables: skip double clone set expressions on element insert
This adds an override_exprs flag so that the dynset and newsetelem paths borrow the set's expressions instead of cloning them when userspace does not supply an override, and releases the expression array only when it really is an override. The same flag is now used to decide whether the dynset expression override is dumped back to userspace.
Quoted 14 lines
net/netfilter/nft_dynset.c +2
-1@@ -19,7 +19,8 @@ struct nft_dynset {u8 sreg_key;u8 sreg_data;bool invert; -bool expr; + bool expr:1, + override_exprs:1;u8 num_exprs;u64 timeout;struct nft_expr *expr_array[NFT_SET_EXPR_MAX ];
[ ... ]
net/netfilter/nft_dynset.c static int nft_dynset_dump(struct sk_buff *skb, · line 404
Quoted 9 lines
net/netfilter/nft_dynset.c +1
-1@@ -393,7 +398,7 @@ static int nft_dynset_dump(struct sk_buff *skb, nf_jiffies64_to_msecs(priv->timeout), NFTA_DYNSET_PAD)) goto nla_put_failure; -if (priv->set->num_exprs ==0 ) { +if (priv->set->num_exprs ==0 || priv->override_exprs) {if (priv->num_exprs ==1 ) {if (nft_expr_dump (skb,NFTA_DYNSET_EXPR , priv->expr_array[0 ], reset))
[Severity: Medium]
Can the message emitted by nft_dynset_dump() after this change be rejected by nft_dynset_init() when it is replayed?
With the widened condition, a dynset attached to a set that owns expressions (set->num_exprs > 0) now dumps its override. For priv->num_exprs > 1 that override is emitted as an NFTA_DYNSET_EXPRESSIONS nest:
net/netfilter/nft_dynset.c:nft_dynset_dump() {
...
u32 flags = priv->invert ? NFT_DYNSET_F_INV : 0;
...
} else if (priv->num_exprs > 1) {
struct nlattr *nest;
nest = nla_nest_start_noflag(skb, NFTA_DYNSET_EXPRESSIONS);
...
if (nla_put_be32(skb, NFTA_DYNSET_FLAGS, htonl(flags)))
...
}The dumped NFTA_DYNSET_FLAGS is computed only from priv->invert, so NFT_DYNSET_F_EXPR is never set, even though priv->expr records that userspace passed it at creation time. On the parse side the NFTA_DYNSET_EXPRESSIONS branch requires that flag:
net/netfilter/nft_dynset.c:nft_dynset_init() {
...
} else if (tb[NFTA_DYNSET_EXPRESSIONS]) {
...
if (!priv->expr)
return -EINVAL;
...
}So a rule with two override expressions on an NFT_SET_EVAL set that also has two set expressions (for example counter plus quota) appears to be dumped in a form that the kernel's own parser refuses with -EINVAL when the netlink message is sent back (raw netlink echo, or libnftnl parse-then-build restore paths). Before this patch that configuration emitted no expression attributes at all, so the replay was accepted and silently fell back to the set's expressions.
The missing NFT_DYNSET_F_EXPR in the dump path predates this patch (it is already reachable for priv->set->num_exprs == 0 with more than one expression), but the new condition extends the asymmetry to override-on-eval-set rules that used to round-trip. Should nft_dynset_dump() also report NFT_DYNSET_F_EXPR when priv->expr is set?
Pablo Neira Ayuso
maintainerRe: [PATCH net 07/10] netfilter: nf_tables: skip double clone set expressions on element insert
Quoted 82 lines · Jakub Kicinski
This is an AI-generated review of your patch. The human sending this email has considered the AI review valid, or at least plausible.
---
netfilter:nf_tables: skip double clone set expressions on element insertThis adds an
override_exprsflag so that the dynset and newsetelem paths borrow the set's expressions instead of cloning them when userspace does not supply an override, and releases the expression array only when it really is an override. The same flag is now used to decide whether the dynset expression override is dumped back to userspace.Quoted 14 lines
net/netfilter/nft_dynset.c +2
-1@@ -19,7 +19,8 @@ struct nft_dynset {u8 sreg_key;u8 sreg_data;bool invert; -bool expr; + bool expr:1, + override_exprs:1;u8 num_exprs;u64 timeout;struct nft_expr *expr_array[NFT_SET_EXPR_MAX ];[ ... ]
net/netfilter/nft_dynset.c static int nft_dynset_dump(struct sk_buff *skb, · line 404
Quoted 9 lines
net/netfilter/nft_dynset.c +1
-1@@ -393,7 +398,7 @@ static int nft_dynset_dump(struct sk_buff *skb, nf_jiffies64_to_msecs(priv->timeout), NFTA_DYNSET_PAD)) goto nla_put_failure; -if (priv->set->num_exprs ==0 ) { +if (priv->set->num_exprs ==0 || priv->override_exprs) {if (priv->num_exprs ==1 ) {if (nft_expr_dump (skb,NFTA_DYNSET_EXPR , priv->expr_array[0 ], reset))[Severity: Medium]
Can the message emitted bynft_dynset_dump()after this change be rejected bynft_dynset_init()when it is replayed?With the widened condition, a dynset attached to a set that owns expressions (
set->num_exprs> 0) now dumps its override. Forpriv->num_exprs> 1 that override is emitted as anNFTA_DYNSET_EXPRESSIONSnest:net/netfilter/nft_dynset.c:nft_dynset_dump() { ... u32 flags = priv->invert ? NFT_DYNSET_F_INV : 0; ... } else if (priv->num_exprs > 1) { struct nlattr *nest; nest = nla_nest_start_noflag(skb, NFTA_DYNSET_EXPRESSIONS); ... if (nla_put_be32(skb, NFTA_DYNSET_FLAGS, htonl(flags))) ... }The dumped
NFTA_DYNSET_FLAGSis computed only frompriv->invert, soNFT_DYNSET_F_EXPRis never set, even thoughpriv->exprrecords that userspace passed it at creation time. On the parse side theNFTA_DYNSET_EXPRESSIONSbranch requires that flag:net/netfilter/nft_dynset.c:nft_dynset_init() { ... } else if (tb[NFTA_DYNSET_EXPRESSIONS]) { ... if (!priv->expr) return -EINVAL; ... }So a rule with two override expressions on an
NFT_SET_EVALset that also has two set expressions (for example counter plus quota) appears to be dumped in a form that the kernel's own parser refuses with-EINVALwhen the netlink message is sent back (raw netlink echo, or libnftnl parse-then-build restore paths). Before this patch that configuration emitted no expression attributes at all, so the replay was accepted and silently fell back to the set's expressions.
Quoted 5 lines
The missing
NFT_DYNSET_F_EXPRin the dump path predates this patch (it is already reachable forpriv->set->num_exprs== 0 with more than one expression), but the new condition extends the asymmetry to override-on-eval-set rules that used to round-trip. Shouldnft_dynset_dump()also reportNFT_DYNSET_F_EXPRwhenpriv->expris set?
This sounds like a theotical issue:
1) Example with one single expression.
table ip x {
set y {
typeof ip saddr
flags dynamic
ct count over 2
}
chain z {
update @y { ip saddr ct count over 3 }
}
}2) Example with two expressions
table ip x {
set y {
typeof ip saddr
flags dynamic
counter ct count over 2
}
chain z {
update @y { ip saddr counter ct count over 3 }
}
}It works fine with userspace 'nft'.
It might be a correctness API issue to guarantee this symmetry, but we can follow up.
Pablo Neira Ayuso
maintainerPATCH 8/10 netfilter: nf_tables: fix device name and prefix match in hook lookup
In reply to Pablo Neira AyusoFrom: Fernando Fernandez Mancera <[email protected]>
Currently, a netdev chain or flowtable hooked to a device prefix can be unintentionally deleted or updated by a control-plane request targeting an exact device name or even a shorter one due to the usage of min() to calculate the length to match.
Fix this by making sure an exact device match never matches a prefix and that both the target and the candidate have the same length.
--- net/netfilter/nf_tables_api.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)
net/netfilter/nf_tables_api.c +6 -3
@@ -1978,7 +1978,7 @@ static int nft_dump_stats(struct sk_buff *skb, struct nft_stats __percpu *stats)return -ENOSPC ; } -static bool hook_is_prefix (struct nft_hook *hook) +static bool hook_is_prefix (const struct nft_hook *hook) {return strlen (hook->ifname) >= hook->ifnamelen; }
@@ -2445,8 +2445,11 @@ static struct nft_hook *nft_hook_list_find(struct list_head *hook_list,struct nft_hook *hook;list_for_each_entry (hook, hook_list, list) { -if (!strncmp (hook->ifname, this->ifname, -min (hook->ifnamelen, this->ifnamelen))) { +if (hook_is_prefix (hook) !=hook_is_prefix (this)) +continue ; +if (hook->ifnamelen != this->ifnamelen) +continue ; +if (!strncmp (hook->ifname, this->ifname, hook->ifnamelen)) {if (hook->flags &NFT_HOOK_REMOVE )continue ;
Signature · 1 line
2.47.3
Jakub Kicinski
maintainerRe: [PATCH net 08/10] netfilter: nf_tables: fix device name and prefix match in hook lookup
This is an AI-generated review of your patch. The human sending this email has considered the AI review valid, or at least plausible.
---
netfilter: nf_tables: fix device name and prefix match in hook lookup
This changes nft_hook_list_find() so a hook spec only matches another spec of the same kind (prefix vs exact name) and the same length, replacing the previous strncmp() over min(hook->ifnamelen, this->ifnamelen). hook_is_prefix() also gains a const argument.
Quoted 4 lines
net/netfilter/nf_tables_api.c +0
-0
[ ... ]
Quoted 4 lines
net/netfilter/nf_tables_api.c +0
-0@@ -2445,8 +2445,11 @@ static struct nft_hook *nft_hook_list_find(struct list_head *hook_list,struct nft_hook *hook;list_for_each_entry (hook, hook_list, list) {
net/netfilter/nf_tables_api.c static struct nft_hook *nft_hook_list_find(struct list_head *hook_list, · line 2455
Quoted 10 lines
net/netfilter/nf_tables_api.c +5
-2-if (!strncmp (hook->ifname, this->ifname, -min (hook->ifnamelen, this->ifnamelen))) { +if (hook_is_prefix (hook) !=hook_is_prefix (this)) +continue ; +if (hook->ifnamelen != this->ifnamelen) +continue ; +if (!strncmp (hook->ifname, this->ifname, hook->ifnamelen)) {if (hook->flags &NFT_HOOK_REMOVE )continue ;
[Severity: Critical]nft_hook_list_find() has two callers with different expectations, and this turns it into a pure identity comparator for both. The delete/get paths (nf_tables_delchain_hook(), nft_flowtable_del_hook()) want identity, but the add paths use it as an overlap rejector:
nf_tables_parse_netdev_hooks()
if (nft_hook_list_find(hook_list, hook)) {
NL_SET_BAD_ATTR(extack, tmp);
nft_netdev_hook_free(hook);
err = -EEXIST;
goto err_hook;
}and the same in nf_tables_updchain(), nft_register_flowtable_net_hooks() and nft_flowtable_update().
The Fixes-referenced commit 6d07a289504a states the min() compare was
deliberate: "When checking for duplicate hooks, compare the shortest prefix so a device may never match more than a single hook spec." Does this change intentionally drop that invariant?
Concretely, is this still rejected after the patch?
nft add chain netdev t c \
'{ type filter hook ingress devices = { "eth*", "eth0" } ; }'Before, prefix "eth" (ifnamelen 3) and exact "eth0" (ifnamelen 5) matched via strncmp(..., min(3, 5)) and gave -EEXIST. Now hook_is_prefix(hook) != hook_is_prefix(this) short-circuits, so both hooks end up in the same hook_list. Prefix "eth" versus prefix "eth0" is likewise separated by the ifnamelen compare.
nft_netdev_hook_alloc() then builds one nf_hook_ops per matching netdev per hook:
for_each_netdev(net, dev) {
if (strncmp(dev->name, hook->ifname, hook->ifnamelen))
continue;
...
ops->dev = dev;
list_add_tail(&ops->list, &hook->ops_list);
}so two distinct nft_hook objects each own an ops with ops->dev == eth0, and nft_netdev_register_hooks() registers both. nf_hook_entries_grow() only rejects duplicates for NF_HOOK_OP_BPF, so both land in eth0->nf_hooks_ingress and the base chain is evaluated twice per packet. For flowtables the cross-flowtable -EEXIST guard in nft_register_flowtable_net_hooks() is bypassed the same way, so the same device gets two FLOW_BLOCK_BIND calls.
Can this also leave a dangling struct net_device pointer? Both notifier loops service only the first hook that owns an ops for the device, because of the unconditional break after the switch:
net/netfilter/nft_chain_filter.c:nft_netdev_event() {
list_for_each_entry(hook, &basechain->hook_list, list) {
ops = nft_hook_find_ops(hook, dev);
...
switch (event) {
case NETDEV_UNREGISTER:
...
nf_unregister_net_hook(dev_net(dev), ops);
list_del_rcu(&ops->list);
kfree_rcu(ops, rcu);
break;
...
}
break;
}
return 0;
}nft_flowtable_event() in nf_tables_api.c has the same trailing break. On ip link del eth0 only the first hook's ops is unregistered and freed; the second hook keeps a registered nf_hook_ops whose ops->dev points at the device that is then freed, since nf_hook_ops holds no device reference.
A later teardown (nft delete chain, nft delete table, dormant toggle, netns exit) walks every ops:
net/netfilter/nf_tables_api.c:nft_netdev_unregister_hooks() {
list_for_each_entry_safe(hook, next, hook_list, list) {
if (!(table->flags & NFT_TABLE_F_DORMANT)) {
list_for_each_entry(ops, &hook->ops_list, list)
nf_unregister_net_hook(net, ops);
}
...
}and __nf_unregister_net_hook() reaches nf_hook_entry_head(net, pf, reg->hooknum, reg->dev), which does dev_net(dev) and returns &dev->nf_hooks_ingress, i.e. a read and then a write through the stale pointer. For flowtables, nft_unregister_flowtable_ops() additionally calls ->setup(..., ops->dev, FLOW_BLOCK_UNBIND) on the freed device. The orphaned nf_hook_entries blob is also leaked.
On NETDEV_REGISTER the same break means only one of two covering hooks gains an ops, so the other stays permanently unhooked.
Would it work to keep the strict identity match only for the delete/update lookups and retain the prefix-overlap semantics for the add/-EEXIST paths, e.g. as two separate helpers? Alternatively, should the unconditional break in nft_netdev_event() and nft_flowtable_event() be dropped so every matching hook is serviced?
The commit message describes only the delete/update lookup semantics; the fact that configurations previously rejected with -EEXIST are now accepted is not mentioned. Could that be spelled out?
Re: [PATCH net 08/10] netfilter: nf_tables: fix device name and prefix match in hook lookup
Quoted 141 lines · Jakub Kicinski
This is an AI-generated review of your patch. The human sending this email has considered the AI review valid, or at least plausible.
---
netfilter:nf_tables: fix device name and prefix match in hook lookupThis changes
nft_hook_list_find()so a hook spec only matches another spec of the same kind (prefix vs exact name) and the same length, replacing the previousstrncmp()overmin(hook->ifnamelen, this->ifnamelen).hook_is_prefix()also gains a const argument.Quoted 4 lines
net/netfilter/nf_tables_api.c +0
-0[ ... ]
Quoted 4 lines
net/netfilter/nf_tables_api.c +0
-0@@ -2445,8 +2445,11 @@ static struct nft_hook *nft_hook_list_find(struct list_head *hook_list,struct nft_hook *hook;list_for_each_entry (hook, hook_list, list) {net/netfilter/nf_tables_api.c static struct nft_hook *nft_hook_list_find(struct list_head *hook_list, · line 2455
Quoted 10 lines
net/netfilter/nf_tables_api.c +5
-2-if (!strncmp (hook->ifname, this->ifname, -min (hook->ifnamelen, this->ifnamelen))) { +if (hook_is_prefix (hook) !=hook_is_prefix (this)) +continue ; +if (hook->ifnamelen != this->ifnamelen) +continue ; +if (!strncmp (hook->ifname, this->ifname, hook->ifnamelen)) {if (hook->flags &NFT_HOOK_REMOVE )continue ;[Severity: Critical]
nft_hook_list_find()has two callers with different expectations, and this turns it into a pure identity comparator for both. The delete/get paths (nf_tables_delchain_hook(),nft_flowtable_del_hook()) want identity, but the add paths use it as an overlap rejector:
nf_tables_parse_netdev_hooks()if (nft_hook_list_find(hook_list, hook)) { NL_SET_BAD_ATTR(extack, tmp); nft_netdev_hook_free(hook); err = -EEXIST; goto err_hook; }and the same in
nf_tables_updchain(),nft_register_flowtable_net_hooks()andnft_flowtable_update().The Fixes-referenced commit
6d07a289504astates themin()compare was
deliberate: "When checking for duplicate hooks, compare the shortest prefix so a device may never match more than a single hook spec." Does this change intentionally drop that invariant?Concretely, is this still rejected after the patch?
nft add chain netdev t c \ '{ type filter hook ingress devices = { "eth*", "eth0" } ; }'Before, prefix "eth" (ifnamelen 3) and exact "eth0" (ifnamelen 5) matched via
strncmp(..., min(3, 5))and gave-EEXIST. Nowhook_is_prefix(hook)!=hook_is_prefix(this)short-circuits, so both hooks end up in the samehook_list. Prefix "eth" versus prefix "eth0" is likewise separated by the ifnamelen compare.
nft_netdev_hook_alloc()then builds onenf_hook_opsper matching netdev per hook:for_each_netdev(net, dev) { if (strncmp(dev->name, hook->ifname, hook->ifnamelen)) continue; ... ops->dev = dev; list_add_tail(&ops->list, &hook->ops_list); }so two distinct
nft_hookobjects each own an ops withops->dev== eth0, andnft_netdev_register_hooks()registers both.nf_hook_entries_grow()only rejects duplicates forNF_HOOK_OP_BPF, so both land ineth0->nf_hooks_ingressand the base chain is evaluated twice per packet. For flowtables the cross-flowtable-EEXISTguard innft_register_flowtable_net_hooks()is bypassed the same way, so the same device gets twoFLOW_BLOCK_BINDcalls.Can this also leave a dangling struct
net_devicepointer? Both notifier loops service only the first hook that owns an ops for the device, because of the unconditional break after the switch:net/netfilter/nft_chain_filter.c:nft_netdev_event() { list_for_each_entry(hook, &basechain->hook_list, list) { ops = nft_hook_find_ops(hook, dev); ... switch (event) { case NETDEV_UNREGISTER: ... nf_unregister_net_hook(dev_net(dev), ops); list_del_rcu(&ops->list); kfree_rcu(ops, rcu); break; ... } break; } return 0; }
nft_flowtable_event()innf_tables_api.chas the same trailing break. On ip link del eth0 only the first hook's ops is unregistered and freed; the second hook keeps a registerednf_hook_opswhoseops->devpoints at the device that is then freed, sincenf_hook_opsholds no device reference.A later teardown (nft delete chain, nft delete table, dormant toggle, netns exit) walks every ops:
net/netfilter/nf_tables_api.c:nft_netdev_unregister_hooks() { list_for_each_entry_safe(hook, next, hook_list, list) { if (!(table->flags & NFT_TABLE_F_DORMANT)) { list_for_each_entry(ops, &hook->ops_list, list) nf_unregister_net_hook(net, ops); } ... }and
__nf_unregister_net_hook()reachesnf_hook_entry_head(net, pf, reg->hooknum, reg->dev), which doesdev_net(dev)and returns&dev->nf_hooks_ingress, i.e. a read and then a write through the stale pointer. For flowtables,nft_unregister_flowtable_ops()additionally calls ->setup(...,ops->dev,FLOW_BLOCK_UNBIND) on the freed device. The orphanednf_hook_entriesblob is also leaked.On
NETDEV_REGISTERthe same break means only one of two covering hooks gains an ops, so the other stays permanently unhooked.Would it work to keep the strict identity match only for the delete/update lookups and retain the prefix-overlap semantics for the add/-EEXIST paths, e.g. as two separate helpers? Alternatively, should the unconditional break in
nft_netdev_event()andnft_flowtable_event()be dropped so every matching hook is serviced?
Quoted 4 lines
The commit message describes only the delete/update lookup semantics; the fact that configurations previously rejected with
-EEXISTare now accepted is not mentioned. Could that be spelled out?
Hi,
About the add path it was intended, what it wasn't is the UAF. The AI review managed to catch it. Btw, sashiko.dev didn't catch it, it seems the NIPA instance is managing to catch more subtle bugs.
This patch needs a V2 enforcing an exact match on delete/update path while using the current approach on the add path. I am sending one tomorrow first thing in the morning.
Thanks,
Fernando.
Pablo Neira Ayuso
maintainerPATCH 9/10 netfilter: nf_tables: set on dead bit when performing early element removal
In reply to Pablo Neira Ayuso.commit call for sets is skipped if set->dead flag is set on, but this flag is set on later in the commit path.
This also reintroduces the bug fixed in commit 7315dc1e122c8 ("netfilter: nf_tables: skip set commit for deleted/destroyed sets").
--- net/netfilter/nf_tables_api.c | 4 ++++ 1 file changed, 4 insertions(+)
net/netfilter/nf_tables_api.c +4 -0
@@ -10877,6 +10877,10 @@ static void nft_set_commit_update(struct nft_ctx *ctx,nft_ctx_update (ctx, trans);switch (trans->msg_type) { +case NFT_MSG_DELSET : +case NFT_MSG_DESTROYSET : +nft_trans_set (trans)->dead =1 ; +break ;case NFT_MSG_DELSETELEM : te =nft_trans_container_elem (trans);if (!te->set->ops->commit)
Signature · 1 line
2.47.3
Pablo Neira Ayuso
maintainerPATCH 10/10 netfilter: nf_tables: remove leftover set_update_list
In reply to Pablo Neira AyusoThis list has been moved to per-netns, remove onstack list which is not used anymore.
--- net/netfilter/nf_tables_api.c | 1 - 1 file changed, 1 deletion(-)
net/netfilter/nf_tables_api.c +0 -1
@@ -11298,7 +11298,6 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action) {struct nftables_pernet *nft_net =nft_pernet (net);struct nft_trans *trans, *next; -LIST_HEAD (set_update_list);struct nft_trans_elem *te;struct nft_ctx ctx = { .net = net,
Signature · 1 line
2.47.3