netfilter updates for net-next
Subsystems (4)
Florian Westphal
maintainercover netfilter updates for net-next
Hello,
This batch contains a few updates for your *net-next* tree. Note that this includes two patches that make changes to lib/.
1. silence a harmless warning for CONFIG_NF_CONNTRACK_PROCFS=n builds, from Zhu Wang.
2, 3:
Allow NLA_POLICY_MASK to be used with BE16/BE32 types, and replace a few manual checks with nla_policy based one in nf_tables, from myself.
4: cleanup in ctnetlink to validate while parsing rather than
using two steps, from Lin Ma.
5: refactor boyer-moore textsearch by moving a small chunk to
a helper function, rom Jeremy Sowden.
The following changes since commit bc758ade614576d1c1b167af0246ada8c916c804:
net/mlx4: clean up a type issue (2023-07-26 22:08:44 -0700)
are available in the Git repository at:
https://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git tags/nf-next-23-07-27
for you to fetch changes up to 86e9c9aa2358a74bcc5e63f9fc69c2d01e64c002:
lib/ts_bm: add helper to reduce indentation and improve readability (2023-07-27 13:45:51 +0200)
----------------------------------------------------------------
netfilter net-next pull request 2023-07-27
----------------------------------------------------------------
Florian Westphal (2):
netlink: allow be16 and be32 types in all uint policy checks
netfilter: nf_tables: use NLA_POLICY_MASK to test for valid flag options
Jeremy Sowden (1):
lib/ts_bm: add helper to reduce indentation and improve readability
Lin Ma (1):
netfilter: conntrack: validate cta_ip via parsing
Zhu Wang (1):
nf_conntrack: fix -Wunused-const-variable=
include/net/netlink.h | 10 +++----- lib/nlattr.c | 6 +++++ lib/ts_bm.c | 43 +++++++++++++++++++++++---------- net/netfilter/nf_conntrack_netlink.c | 8 ++---- net/netfilter/nf_conntrack_proto_dccp.c | 2 ++ net/netfilter/nft_fib.c | 13 +++++----- net/netfilter/nft_lookup.c | 6 ++--- net/netfilter/nft_masq.c | 8 +++--- net/netfilter/nft_nat.c | 8 +++--- net/netfilter/nft_redir.c | 8 +++--- 10 files changed, 61 insertions(+), 51 deletions(-)
Florian Westphal
maintainerPATCH 1/5 nf_conntrack: fix -Wunused-const-variable=
From: Zhu Wang <[email protected]>
When building with W=1, the following warning occurs.
net/netfilter/nf_conntrack_proto_dccp.c:72:27: warning: ‘dccp_state_names’ defined but not used [-Wunused-const-variable=]
static const char * const dccp_state_names[] = {
We include dccp_state_names in the macro CONFIG_NF_CONNTRACK_PROCFS, since it is only used in the place which is included in the macro CONFIG_NF_CONNTRACK_PROCFS.
--- net/netfilter/nf_conntrack_proto_dccp.c | 2 ++ 1 file changed, 2 insertions(+)
net/netfilter/nf_conntrack_proto_dccp.c +2 -0
@@ -69,6 +69,7 @@#define DCCP_MSL (2 * 60 * HZ) +#ifdef CONFIG_NF_CONNTRACK_PROCFS static const char *const dccp_state_names[] = { [CT_DCCP_NONE ] ="NONE" , [CT_DCCP_REQUEST] = "REQUEST",
@@ -81,6 +82,7 @@ static const char * const dccp_state_names[] = { [CT_DCCP_IGNORE] = "IGNORE", [CT_DCCP_INVALID] = "INVALID", }; +#endif #define sNO CT_DCCP_NONE#define sRQ CT_DCCP_REQUEST
Signature · 1 line
2.41.0
Jakub Kicinski
maintainerRe: [PATCH net-next 1/5] nf_conntrack: fix -Wunused-const-variable=
Quoted 8 lines · On Thu, 27 Jul 2023 15:35:56 +0200 Florian Westphal wrote:
When building with W=1, the following warning occurs.
net/netfilter/nf_conntrack_proto_dccp.c:72:27: warning: ‘dccp_state_names’ defined but not used [-Wunused-const-variable=]
static const char * const dccp_state_names[] = {We include
dccp_state_namesin the macroCONFIG_NF_CONNTRACK_PROCFS, since it is only used in the place which is included in the macroCONFIG_NF_CONNTRACK_PROCFS.
FTR I can't say I see this with the versions of gcc / clang I have :S
Quoted 1 line
Fixes 2bc780499aa3 ("[NETFILTER]: nf_conntrack: add DCCP protocol support")
Nor that it's worth a Fixes tag?
Jakub Kicinski
maintainerRe: [PATCH net-next 1/5] nf_conntrack: fix -Wunused-const-variable=
Quoted 5 lines · On Thu, 27 Jul 2023 20:28:11 -0700 Jakub Kicinski wrote:
Quoted 3 lines
We include
dccp_state_namesin the macroCONFIG_NF_CONNTRACK_PROCFS, since it is only used in the place which is included in the macroCONFIG_NF_CONNTRACK_PROCFS.FTR I can't say I see this with the versions of gcc / clang I have :S
Ignore. Just my stupidity.
Re: [PATCH net-next 1/5] nf_conntrack: fix -Wunused-const-variable=
In reply to Florian WestphalThis series was applied to netdev/net-next.git (main) by Florian Westphal <[email protected]>:
Hello:
This series was applied to netdev/net-next.git (main) by Florian Westphal <[email protected]>:
Quoted 8 lines · On Thu, 27 Jul 2023 15:35:56 +0200 you wrote:
From: Zhu Wang <[email protected]>
When building with W=1, the following warning occurs.
net/netfilter/nf_conntrack_proto_dccp.c:72:27: warning: ‘dccp_state_names’ defined but not used [-Wunused-const-variable=]
static const char * const dccp_state_names[] = {[...]
Here is the summary with links:
- [net-next,1/5] nf_conntrack: fix -Wunused-const-variable=
https://git.kernel.org/netdev/net-next/c/a927d77778e3 - [net-next,2/5] netlink: allow be16 and be32 types in all uint policy checks https://git.kernel.org/netdev/net-next/c/5fac9b7c16c5 - [net-next,3/5] netfilter: nf_tables: use NLA_POLICY_MASK to test for valid flag options https://git.kernel.org/netdev/net-next/c/100a11b69842 - [net-next,4/5] netfilter: conntrack: validate cta_ip via parsing https://git.kernel.org/netdev/net-next/c/0c805e80e35d - [net-next,5/5] lib/ts_bm: add helper to reduce indentation and improve readability https://git.kernel.org/netdev/net-next/c/86e9c9aa2358
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html
Florian Westphal
maintainerPATCH 2/5 netlink: allow be16 and be32 types in all uint policy checks
In reply to Florian Westphal__NLA_IS_BEINT_TYPE(tp) isn't useful. NLA_BE16/32 are identical to NLA_U16/32, the only difference is that it tells the netlink validation functions that byteorder conversion might be needed before comparing the value to the policy min/max ones.
After this change all policy macros that can be used with UINT types, such as NLA_POLICY_MASK() can also be used with NLA_BE16/32.
This will be used to validate nf_tables flag attributes which are in bigendian byte order.
--- include/net/netlink.h | 10 +++------- lib/nlattr.c | 6 ++++++ 2 files changed, 9 insertions(+), 7 deletions(-)
include/net/netlink.h +3 -7
@@ -375,12 +375,11 @@ struct nla_policy {#define NLA_POLICY_BITFIELD32 (valid) \ { .type = NLA_BITFIELD32, .bitfield32_valid = valid } -#define __NLA_IS_UINT_TYPE (tp) \ - (tp == NLA_U8 || tp == NLA_U16 || tp == NLA_U32 || tp == NLA_U64) +#define __NLA_IS_UINT_TYPE (tp) \ + (tp == NLA_U8 || tp == NLA_U16 || tp == NLA_U32 || \ + tp == NLA_U64 || tp == NLA_BE16 || tp == NLA_BE32)#define __NLA_IS_SINT_TYPE (tp) \ (tp == NLA_S8 || tp == NLA_S16 || tp == NLA_S32 || tp == NLA_S64) -#define __NLA_IS_BEINT_TYPE (tp) \ - (tp == NLA_BE16 || tp == NLA_BE32)#define __NLA_ENSURE (condition) BUILD_BUG_ON_ZERO(!(condition))#define NLA_ENSURE_UINT_TYPE (tp) \
@@ -394,7 +393,6 @@ struct nla_policy {#define NLA_ENSURE_INT_OR_BINARY_TYPE (tp) \ (__NLA_ENSURE(__NLA_IS_UINT_TYPE(tp) || \ __NLA_IS_SINT_TYPE(tp) || \ - __NLA_IS_BEINT_TYPE(tp) || \ tp == NLA_MSECS || \ tp == NLA_BINARY) + tp)#define NLA_ENSURE_NO_VALIDATION_PTR (tp) \
@@ -402,8 +400,6 @@ struct nla_policy { tp != NLA_REJECT && \ tp != NLA_NESTED && \ tp != NLA_NESTED_ARRAY) + tp) -#define NLA_ENSURE_BEINT_TYPE (tp) \ - (__NLA_ENSURE(__NLA_IS_BEINT_TYPE(tp)) + tp)#define NLA_POLICY_RANGE (tp, _min, _max) { \ .type = NLA_ENSURE_INT_OR_BINARY_TYPE(tp), \
lib/nlattr.c +6 -0
@@ -355,6 +355,12 @@ static int nla_validate_mask(const struct nla_policy *pt,case NLA_U64 : value =nla_get_u64 (nla);break ; +case NLA_BE16 : + value =ntohs (nla_get_be16 (nla)); +break ; +case NLA_BE32 : + value =ntohl (nla_get_be32 (nla)); +break ;default :return -EINVAL ; }
Signature · 1 line
2.41.0
Florian Westphal
maintainerPATCH 3/5 netfilter: nf_tables: use NLA_POLICY_MASK to test for valid flag options
In reply to Florian Westphalnf_tables relies on manual test of netlink attributes coming from userspace even in cases where this could be handled via netlink policy.
Convert a bunch of 'flag' attributes to use NLA_POLICY_MASK checks.
--- net/netfilter/nft_fib.c | 13 +++++++------ net/netfilter/nft_lookup.c | 6 ++---- net/netfilter/nft_masq.c | 8 +++----- net/netfilter/nft_nat.c | 8 +++----- net/netfilter/nft_redir.c | 8 +++----- 5 files changed, 18 insertions(+), 25 deletions(-)
net/netfilter/nft_fib.c +7 -6
@@ -14,17 +14,18 @@#include <net/netfilter/nf_tables.h> #include <net/netfilter/nft_fib.h> +#define NFTA_FIB_F_ALL (NFTA_FIB_F_SADDR | NFTA_FIB_F_DADDR | \ + NFTA_FIB_F_MARK | NFTA_FIB_F_IIF | NFTA_FIB_F_OIF | \ + NFTA_FIB_F_PRESENT) +const struct nla_policy nft_fib_policy[NFTA_FIB_MAX +1 ] = { [NFTA_FIB_DREG ] = { .type =NLA_U32 }, [NFTA_FIB_RESULT ] = { .type =NLA_U32 }, - [NFTA_FIB_FLAGS ] = { .type =NLA_U32 }, + [NFTA_FIB_FLAGS ] = +NLA_POLICY_MASK (NLA_BE32 ,NFTA_FIB_F_ALL ), };EXPORT_SYMBOL (nft_fib_policy); -#define NFTA_FIB_F_ALL (NFTA_FIB_F_SADDR | NFTA_FIB_F_DADDR | \ - NFTA_FIB_F_MARK | NFTA_FIB_F_IIF | NFTA_FIB_F_OIF | \ - NFTA_FIB_F_PRESENT) -int nft_fib_validate (const struct nft_ctx *ctx,const struct nft_expr *expr,const struct nft_data **data) {
@@ -77,7 +78,7 @@ int nft_fib_init(const struct nft_ctx *ctx, const struct nft_expr *expr, priv->flags =ntohl (nla_get_be32 (tb[NFTA_FIB_FLAGS ])); -if (priv->flags ==0 || (priv->flags & ~NFTA_FIB_F_ALL )) +if (priv->flags ==0 )return -EINVAL ; if ((priv->flags & (NFTA_FIB_F_SADDR | NFTA_FIB_F_DADDR)) ==
net/netfilter/nft_lookup.c +2 -4
@@ -90,7 +90,8 @@ static const struct nla_policy nft_lookup_policy[NFTA_LOOKUP_MAX + 1] = { [NFTA_LOOKUP_SET_ID] = { .type = NLA_U32 }, [NFTA_LOOKUP_SREG] = { .type = NLA_U32 }, [NFTA_LOOKUP_DREG] = { .type = NLA_U32 }, - [NFTA_LOOKUP_FLAGS] = { .type = NLA_U32 }, + [NFTA_LOOKUP_FLAGS ] = + NLA_POLICY_MASK(NLA_BE32, NFT_LOOKUP_F_INV), }; static int nft_lookup_init(const struct nft_ctx *ctx,
@@ -120,9 +121,6 @@ static int nft_lookup_init(const struct nft_ctx *ctx,if (tb[NFTA_LOOKUP_FLAGS ]) { flags =ntohl (nla_get_be32 (tb[NFTA_LOOKUP_FLAGS ])); -if (flags & ~NFT_LOOKUP_F_INV ) -return -EINVAL ; -if (flags &NFT_LOOKUP_F_INV ) priv->invert = true; }
net/netfilter/nft_masq.c +3 -5
@@ -20,7 +20,8 @@ struct nft_masq { };static const struct nla_policy nft_masq_policy[NFTA_MASQ_MAX +1 ] = { - [NFTA_MASQ_FLAGS ] = { .type =NLA_U32 }, + [NFTA_MASQ_FLAGS ] = +NLA_POLICY_MASK (NLA_BE32 ,NF_NAT_RANGE_MASK ), [NFTA_MASQ_REG_PROTO_MIN ] = { .type =NLA_U32 }, [NFTA_MASQ_REG_PROTO_MAX ] = { .type =NLA_U32 }, };
@@ -47,11 +48,8 @@ static int nft_masq_init(const struct nft_ctx *ctx,struct nft_masq *priv =nft_expr_priv (expr);int err; -if (tb[NFTA_MASQ_FLAGS ]) { +if (tb[NFTA_MASQ_FLAGS ]) priv->flags =ntohl (nla_get_be32 (tb[NFTA_MASQ_FLAGS ])); -if (priv->flags & ~NF_NAT_RANGE_MASK ) -return -EINVAL ; - }if (tb[NFTA_MASQ_REG_PROTO_MIN ]) { err = nft_parse_register_load(tb[NFTA_MASQ_REG_PROTO_MIN],
net/netfilter/nft_nat.c +3 -5
@@ -132,7 +132,8 @@ static const struct nla_policy nft_nat_policy[NFTA_NAT_MAX + 1] = { [NFTA_NAT_REG_ADDR_MAX] = { .type = NLA_U32 }, [NFTA_NAT_REG_PROTO_MIN] = { .type = NLA_U32 }, [NFTA_NAT_REG_PROTO_MAX] = { .type = NLA_U32 }, - [NFTA_NAT_FLAGS] = { .type = NLA_U32 }, + [NFTA_NAT_FLAGS ] = + NLA_POLICY_MASK(NLA_BE32, NF_NAT_RANGE_MASK), }; static int nft_nat_validate(const struct nft_ctx *ctx,
@@ -246,11 +247,8 @@ static int nft_nat_init(const struct nft_ctx *ctx, const struct nft_expr *expr, priv->flags |=NF_NAT_RANGE_PROTO_SPECIFIED ; } -if (tb[NFTA_NAT_FLAGS ]) { +if (tb[NFTA_NAT_FLAGS ]) priv->flags |=ntohl (nla_get_be32 (tb[NFTA_NAT_FLAGS ])); -if (priv->flags & ~NF_NAT_RANGE_MASK ) -return -EOPNOTSUPP ; - }return nf_ct_netns_get (ctx->net, family); }
net/netfilter/nft_redir.c +3 -5
@@ -22,7 +22,8 @@ struct nft_redir {static const struct nla_policy nft_redir_policy[NFTA_REDIR_MAX +1 ] = { [NFTA_REDIR_REG_PROTO_MIN ] = { .type =NLA_U32 }, [NFTA_REDIR_REG_PROTO_MAX ] = { .type =NLA_U32 }, - [NFTA_REDIR_FLAGS ] = { .type =NLA_U32 }, + [NFTA_REDIR_FLAGS ] = +NLA_POLICY_MASK (NLA_BE32 ,NF_NAT_RANGE_MASK ), }; static int nft_redir_validate(const struct nft_ctx *ctx,
@@ -68,11 +69,8 @@ static int nft_redir_init(const struct nft_ctx *ctx, priv->flags |=NF_NAT_RANGE_PROTO_SPECIFIED ; } -if (tb[NFTA_REDIR_FLAGS ]) { +if (tb[NFTA_REDIR_FLAGS ]) priv->flags =ntohl (nla_get_be32 (tb[NFTA_REDIR_FLAGS ])); -if (priv->flags & ~NF_NAT_RANGE_MASK ) -return -EINVAL ; - }return nf_ct_netns_get (ctx->net, ctx->family); }
Signature · 1 line
2.41.0
Florian Westphal
maintainerPATCH 4/5 netfilter: conntrack: validate cta_ip via parsing
In reply to Florian WestphalFrom: Lin Ma <[email protected]>
In current ctnetlink_parse_tuple_ip() function, nested parsing and validation is splitting as two parts, which could be cleanup to a simplified form. As the nla_parse_nested_deprecated function supports validation in the fly. These two finially reach same place __nla_validate_parse with same validate flag.
nla_parse_nested_deprecated
__nla_parse(.., NL_VALIDATE_LIBERAL, ..)
__nla_validate_parse
nla_validate_nested_deprecated
__nla_validate_nested(.., NL_VALIDATE_LIBERAL, ..)
__nla_validate
__nla_validate_parse
This commit removes the call to nla_validate_nested_deprecated and pass cta_ip_nla_policy when do parsing.
--- net/netfilter/nf_conntrack_netlink.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-)
net/netfilter/nf_conntrack_netlink.c +2 -6
@@ -1321,15 +1321,11 @@ static int ctnetlink_parse_tuple_ip(struct nlattr *attr,struct nlattr *tb[CTA_IP_MAX +1 ];int ret =0 ; - ret =nla_parse_nested_deprecated (tb,CTA_IP_MAX , attr,NULL ,NULL ); + ret =nla_parse_nested_deprecated (tb,CTA_IP_MAX , attr, + cta_ip_nla_policy,NULL );if (ret <0 )return ret; - ret =nla_validate_nested_deprecated (attr,CTA_IP_MAX , - cta_ip_nla_policy,NULL ); -if (ret) -return ret; -switch (tuple->src.l3num) {case NFPROTO_IPV4 : ret =ipv4_nlattr_to_tuple (tb, tuple, flags);
Signature · 1 line
2.41.0
Florian Westphal
maintainerPATCH 5/5 lib/ts_bm: add helper to reduce indentation and improve readability
In reply to Florian WestphalFrom: Jeremy Sowden <[email protected]>
The flow-control of bm_find is very deeply nested with a conditional comparing a ternary expression against the pattern inside a for-loop inside a while-loop inside a for-loop.
Move the inner for-loop into a helper function to reduce the amount of indentation and make the code easier to read.
Fix indentation and trailing white-space in preceding debug logging statement.
--- lib/ts_bm.c | 43 ++++++++++++++++++++++++++++++------------- 1 file changed, 30 insertions(+), 13 deletions(-)
lib/ts_bm.c +30 -13
@@ -55,6 +55,24 @@ struct ts_bmunsigned int good_shift[]; }; +static unsigned int matchpat (const u8 *pattern,unsigned int patlen, +const u8 *text,bool icase) +{ +unsigned int i; + +for (i =0 ; i < patlen; i++) { +u8 t = *(text-i); + +if (icase) + t =toupper (t); + +if (t != *(pattern-i)) +break ; + } + +return i; +} +static unsigned int bm_find (struct ts_config *conf,struct ts_state *state) {struct ts_bm *bm =ts_config_priv (conf);
@@ -72,19 +90,18 @@ static unsigned int bm_find(struct ts_config *conf, struct ts_state *state)break ;while (shift < text_len) { -DEBUGP ("Searching in position %d (%c)\n" , - shift, text[shift]); -for (i =0 ; i < bm->patlen; i++) -if ((icase ?toupper (text[shift-i]) - : text[shift-i]) - != bm->pattern[bm->patlen-1 -i]) - goto next; - -/* London calling... */ -DEBUGP ("found!\n" ); -return consumed + (shift-(bm->patlen-1 )); - -next: bs = bm->bad_shift[text[shift-i]]; +DEBUGP ("Searching in position %d (%c)\n" , + shift, text[shift]); + + i =matchpat (&bm->pattern[bm->patlen-1 ], bm->patlen, + &text[shift], icase); +if (i == bm->patlen) { +/* London calling... */ +DEBUGP ("found!\n" ); +return consumed + (shift-(bm->patlen-1 )); + } + + bs = bm->bad_shift[text[shift-i]];/* Now jumping to... */ shift =max_t (int, shift-i+bs, shift+bm->good_shift[i]);
Signature · 1 line
2.41.0