* [PATCH v2 1/1] netfilter: load nf_log_syslog on enabling nf_conntrack_log_invalid
@ 2025-05-26 8:59 Lance Yang
2025-05-28 11:05 ` Florian Westphal
2025-07-24 12:24 ` Pablo Neira Ayuso
0 siblings, 2 replies; 7+ messages in thread
From: Lance Yang @ 2025-05-26 8:59 UTC (permalink / raw)
To: fw, pablo
Cc: coreteam, davem, edumazet, horms, kadlec, kuba, linux-kernel,
netfilter-devel, pabeni, zi.li, Lance Yang
From: Lance Yang <lance.yang@linux.dev>
When no logger is registered, nf_conntrack_log_invalid fails to log invalid
packets, leaving users unaware of actual invalid traffic. Improve this by
loading nf_log_syslog, similar to how 'iptables -I FORWARD 1 -m conntrack
--ctstate INVALID -j LOG' triggers it.
Suggested-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Zi Li <zi.li@linux.dev>
Signed-off-by: Lance Yang <lance.yang@linux.dev>
---
v1 -> v2:
- Add a new, simpler helper (per Florian)
- Load the module only when no logger is registered (per Florian)
- https://lore.kernel.org/all/20250514053751.2271-1-lance.yang@linux.dev/
include/net/netfilter/nf_log.h | 3 +++
net/netfilter/nf_conntrack_standalone.c | 26 +++++++++++++++++++++++-
net/netfilter/nf_log.c | 27 +++++++++++++++++++++++++
3 files changed, 55 insertions(+), 1 deletion(-)
diff --git a/include/net/netfilter/nf_log.h b/include/net/netfilter/nf_log.h
index e55eedc84ed7..00506792a06d 100644
--- a/include/net/netfilter/nf_log.h
+++ b/include/net/netfilter/nf_log.h
@@ -59,6 +59,9 @@ extern int sysctl_nf_log_all_netns;
int nf_log_register(u_int8_t pf, struct nf_logger *logger);
void nf_log_unregister(struct nf_logger *logger);
+/* Check if any logger is registered for a given protocol family. */
+bool nf_log_is_registered(u_int8_t pf);
+
int nf_log_set(struct net *net, u_int8_t pf, const struct nf_logger *logger);
void nf_log_unset(struct net *net, const struct nf_logger *logger);
diff --git a/net/netfilter/nf_conntrack_standalone.c b/net/netfilter/nf_conntrack_standalone.c
index 2f666751c7e7..cdc27424f84a 100644
--- a/net/netfilter/nf_conntrack_standalone.c
+++ b/net/netfilter/nf_conntrack_standalone.c
@@ -14,6 +14,7 @@
#include <linux/sysctl.h>
#endif
+#include <net/netfilter/nf_log.h>
#include <net/netfilter/nf_conntrack.h>
#include <net/netfilter/nf_conntrack_core.h>
#include <net/netfilter/nf_conntrack_l4proto.h>
@@ -543,6 +544,29 @@ nf_conntrack_hash_sysctl(const struct ctl_table *table, int write,
return ret;
}
+static int
+nf_conntrack_log_invalid_sysctl(const struct ctl_table *table, int write,
+ void *buffer, size_t *lenp, loff_t *ppos)
+{
+ int ret, i;
+
+ ret = proc_dou8vec_minmax(table, write, buffer, lenp, ppos);
+ if (ret < 0 || !write)
+ return ret;
+
+ if (*(u8 *)table->data == 0)
+ return ret;
+
+ /* Load nf_log_syslog only if no logger is currently registered */
+ for (i = 0; i < NFPROTO_NUMPROTO; i++) {
+ if (nf_log_is_registered(i))
+ return ret;
+ }
+ request_module("%s", "nf_log_syslog");
+
+ return ret;
+}
+
static struct ctl_table_header *nf_ct_netfilter_header;
enum nf_ct_sysctl_index {
@@ -649,7 +673,7 @@ static struct ctl_table nf_ct_sysctl_table[] = {
.data = &init_net.ct.sysctl_log_invalid,
.maxlen = sizeof(u8),
.mode = 0644,
- .proc_handler = proc_dou8vec_minmax,
+ .proc_handler = nf_conntrack_log_invalid_sysctl,
},
[NF_SYSCTL_CT_EXPECT_MAX] = {
.procname = "nf_conntrack_expect_max",
diff --git a/net/netfilter/nf_log.c b/net/netfilter/nf_log.c
index 6dd0de33eebd..c7dd5019a89d 100644
--- a/net/netfilter/nf_log.c
+++ b/net/netfilter/nf_log.c
@@ -125,6 +125,33 @@ void nf_log_unregister(struct nf_logger *logger)
}
EXPORT_SYMBOL(nf_log_unregister);
+/**
+ * nf_log_is_registered - Check if any logger is registered for a given
+ * protocol family.
+ *
+ * @pf: Protocol family
+ *
+ * Returns: true if at least one logger is active for @pf, false otherwise.
+ */
+bool nf_log_is_registered(u_int8_t pf)
+{
+ int i;
+
+ /* Out of bounds. */
+ if (pf >= NFPROTO_NUMPROTO) {
+ WARN_ON_ONCE(1);
+ return false;
+ }
+
+ for (i = 0; i < NF_LOG_TYPE_MAX; i++) {
+ if (rcu_access_pointer(loggers[pf][i]))
+ return true;
+ }
+
+ return false;
+}
+EXPORT_SYMBOL(nf_log_is_registered);
+
int nf_log_bind_pf(struct net *net, u_int8_t pf,
const struct nf_logger *logger)
{
--
2.49.0
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH v2 1/1] netfilter: load nf_log_syslog on enabling nf_conntrack_log_invalid
2025-05-26 8:59 [PATCH v2 1/1] netfilter: load nf_log_syslog on enabling nf_conntrack_log_invalid Lance Yang
@ 2025-05-28 11:05 ` Florian Westphal
2025-05-28 11:42 ` Lance Yang
2025-07-14 12:05 ` Lance Yang
2025-07-24 12:24 ` Pablo Neira Ayuso
1 sibling, 2 replies; 7+ messages in thread
From: Florian Westphal @ 2025-05-28 11:05 UTC (permalink / raw)
To: Lance Yang
Cc: pablo, coreteam, davem, edumazet, horms, kadlec, kuba,
linux-kernel, netfilter-devel, pabeni, zi.li, Lance Yang
Lance Yang <ioworker0@gmail.com> wrote:
> From: Lance Yang <lance.yang@linux.dev>
>
> When no logger is registered, nf_conntrack_log_invalid fails to log invalid
> packets, leaving users unaware of actual invalid traffic. Improve this by
> loading nf_log_syslog, similar to how 'iptables -I FORWARD 1 -m conntrack
> --ctstate INVALID -j LOG' triggers it.
Acked-by: Florian Westphal <fw@strlen.de>
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v2 1/1] netfilter: load nf_log_syslog on enabling nf_conntrack_log_invalid
2025-05-28 11:05 ` Florian Westphal
@ 2025-05-28 11:42 ` Lance Yang
2025-05-30 3:10 ` Lance Yang
2025-07-14 12:05 ` Lance Yang
1 sibling, 1 reply; 7+ messages in thread
From: Lance Yang @ 2025-05-28 11:42 UTC (permalink / raw)
To: Florian Westphal
Cc: pablo, coreteam, davem, Lance Yang, edumazet, horms, kadlec, kuba,
linux-kernel, netfilter-devel, pabeni, zi.li
Thanks for taking the time to review!
On 2025/5/28 19:05, Florian Westphal wrote:
> Lance Yang <ioworker0@gmail.com> wrote:
>> From: Lance Yang <lance.yang@linux.dev>
>>
>> When no logger is registered, nf_conntrack_log_invalid fails to log invalid
>> packets, leaving users unaware of actual invalid traffic. Improve this by
>> loading nf_log_syslog, similar to how 'iptables -I FORWARD 1 -m conntrack
>> --ctstate INVALID -j LOG' triggers it.
>
> Acked-by: Florian Westphal <fw@strlen.de>
Hmm... should this patch be backported to stable kernels? Without it,
nf_conntrack_log_invalid won't log invalid packets when no logger is
registered, causing unnecessary debugging effort ;)
Back then, I actually thought my machine wasn't seeing any invalid
packets... turns out they just weren't logged in dmesg :(
Thanks,
Lance
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v2 1/1] netfilter: load nf_log_syslog on enabling nf_conntrack_log_invalid
2025-05-28 11:42 ` Lance Yang
@ 2025-05-30 3:10 ` Lance Yang
0 siblings, 0 replies; 7+ messages in thread
From: Lance Yang @ 2025-05-30 3:10 UTC (permalink / raw)
To: Florian Westphal
Cc: pablo, coreteam, davem, Lance Yang, edumazet, horms, kadlec, kuba,
linux-kernel, netfilter-devel, pabeni, zi.li, stable
Cc: stable
On 2025/5/28 19:42, Lance Yang wrote:
>
> Thanks for taking the time to review!
>
> On 2025/5/28 19:05, Florian Westphal wrote:
>> Lance Yang <ioworker0@gmail.com> wrote:
>>> From: Lance Yang <lance.yang@linux.dev>
>>>
>>> When no logger is registered, nf_conntrack_log_invalid fails to log
>>> invalid
>>> packets, leaving users unaware of actual invalid traffic. Improve
>>> this by
>>> loading nf_log_syslog, similar to how 'iptables -I FORWARD 1 -m
>>> conntrack
>>> --ctstate INVALID -j LOG' triggers it.
>>
>> Acked-by: Florian Westphal <fw@strlen.de>
>
> Hmm... should this patch be backported to stable kernels? Without it,
> nf_conntrack_log_invalid won't log invalid packets when no logger is
> registered, causing unnecessary debugging effort ;)
>
> Back then, I actually thought my machine wasn't seeing any invalid
> packets... turns out they just weren't logged in dmesg :(
>
> Thanks,
> Lance
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v2 1/1] netfilter: load nf_log_syslog on enabling nf_conntrack_log_invalid
2025-05-28 11:05 ` Florian Westphal
2025-05-28 11:42 ` Lance Yang
@ 2025-07-14 12:05 ` Lance Yang
1 sibling, 0 replies; 7+ messages in thread
From: Lance Yang @ 2025-07-14 12:05 UTC (permalink / raw)
To: Florian Westphal
Cc: pablo, coreteam, davem, edumazet, horms, kadlec, kuba,
linux-kernel, netfilter-devel, pabeni, zi.li, Lance Yang
On 2025/5/28 19:05, Florian Westphal wrote:
> Lance Yang <ioworker0@gmail.com> wrote:
>> From: Lance Yang <lance.yang@linux.dev>
>>
>> When no logger is registered, nf_conntrack_log_invalid fails to log invalid
>> packets, leaving users unaware of actual invalid traffic. Improve this by
>> loading nf_log_syslog, similar to how 'iptables -I FORWARD 1 -m conntrack
>> --ctstate INVALID -j LOG' triggers it.
>
> Acked-by: Florian Westphal <fw@strlen.de>
A gentle reminder for this patch, as the merge window is approaching ;)
Please let me know if any changes are needed.
Thanks,
Lance
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v2 1/1] netfilter: load nf_log_syslog on enabling nf_conntrack_log_invalid
2025-05-26 8:59 [PATCH v2 1/1] netfilter: load nf_log_syslog on enabling nf_conntrack_log_invalid Lance Yang
2025-05-28 11:05 ` Florian Westphal
@ 2025-07-24 12:24 ` Pablo Neira Ayuso
2025-07-24 12:48 ` Lance Yang
1 sibling, 1 reply; 7+ messages in thread
From: Pablo Neira Ayuso @ 2025-07-24 12:24 UTC (permalink / raw)
To: Lance Yang
Cc: fw, coreteam, davem, edumazet, horms, kadlec, kuba, linux-kernel,
netfilter-devel, pabeni, zi.li, Lance Yang
Hi,
On Mon, May 26, 2025 at 04:59:02PM +0800, Lance Yang wrote:
> diff --git a/net/netfilter/nf_conntrack_standalone.c b/net/netfilter/nf_conntrack_standalone.c
> index 2f666751c7e7..cdc27424f84a 100644
> --- a/net/netfilter/nf_conntrack_standalone.c
> +++ b/net/netfilter/nf_conntrack_standalone.c
> @@ -14,6 +14,7 @@
> #include <linux/sysctl.h>
> #endif
>
> +#include <net/netfilter/nf_log.h>
> #include <net/netfilter/nf_conntrack.h>
> #include <net/netfilter/nf_conntrack_core.h>
> #include <net/netfilter/nf_conntrack_l4proto.h>
> @@ -543,6 +544,29 @@ nf_conntrack_hash_sysctl(const struct ctl_table *table, int write,
> return ret;
> }
>
> +static int
> +nf_conntrack_log_invalid_sysctl(const struct ctl_table *table, int write,
> + void *buffer, size_t *lenp, loff_t *ppos)
> +{
> + int ret, i;
> +
> + ret = proc_dou8vec_minmax(table, write, buffer, lenp, ppos);
> + if (ret < 0 || !write)
> + return ret;
> +
> + if (*(u8 *)table->data == 0)
> + return ret;
What is this table->data check for? I don't find any similar idiom
like this in the existing proc_dou8vec_minmax() callers.
> +
> + /* Load nf_log_syslog only if no logger is currently registered */
> + for (i = 0; i < NFPROTO_NUMPROTO; i++) {
> + if (nf_log_is_registered(i))
> + return ret;
> + }
> + request_module("%s", "nf_log_syslog");
> +
> + return ret;
> +}
> +
> static struct ctl_table_header *nf_ct_netfilter_header;
>
> enum nf_ct_sysctl_index {
> @@ -649,7 +673,7 @@ static struct ctl_table nf_ct_sysctl_table[] = {
> .data = &init_net.ct.sysctl_log_invalid,
> .maxlen = sizeof(u8),
> .mode = 0644,
> - .proc_handler = proc_dou8vec_minmax,
> + .proc_handler = nf_conntrack_log_invalid_sysctl,
> },
> [NF_SYSCTL_CT_EXPECT_MAX] = {
> .procname = "nf_conntrack_expect_max",
> diff --git a/net/netfilter/nf_log.c b/net/netfilter/nf_log.c
> index 6dd0de33eebd..c7dd5019a89d 100644
> --- a/net/netfilter/nf_log.c
> +++ b/net/netfilter/nf_log.c
> @@ -125,6 +125,33 @@ void nf_log_unregister(struct nf_logger *logger)
> }
> EXPORT_SYMBOL(nf_log_unregister);
>
> +/**
> + * nf_log_is_registered - Check if any logger is registered for a given
> + * protocol family.
> + *
> + * @pf: Protocol family
> + *
> + * Returns: true if at least one logger is active for @pf, false otherwise.
> + */
> +bool nf_log_is_registered(u_int8_t pf)
> +{
> + int i;
> +
> + /* Out of bounds. */
No need for this comment, please remove it.
> + if (pf >= NFPROTO_NUMPROTO) {
> + WARN_ON_ONCE(1);
> + return false;
> + }
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v2 1/1] netfilter: load nf_log_syslog on enabling nf_conntrack_log_invalid
2025-07-24 12:24 ` Pablo Neira Ayuso
@ 2025-07-24 12:48 ` Lance Yang
0 siblings, 0 replies; 7+ messages in thread
From: Lance Yang @ 2025-07-24 12:48 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: fw, coreteam, davem, edumazet, horms, kadlec, kuba, linux-kernel,
netfilter-devel, pabeni, zi.li, Lance Yang
On 2025/7/24 20:24, Pablo Neira Ayuso wrote:
> Hi,
>
> On Mon, May 26, 2025 at 04:59:02PM +0800, Lance Yang wrote:
>> diff --git a/net/netfilter/nf_conntrack_standalone.c b/net/netfilter/nf_conntrack_standalone.c
>> index 2f666751c7e7..cdc27424f84a 100644
>> --- a/net/netfilter/nf_conntrack_standalone.c
>> +++ b/net/netfilter/nf_conntrack_standalone.c
>> @@ -14,6 +14,7 @@
>> #include <linux/sysctl.h>
>> #endif
>>
>> +#include <net/netfilter/nf_log.h>
>> #include <net/netfilter/nf_conntrack.h>
>> #include <net/netfilter/nf_conntrack_core.h>
>> #include <net/netfilter/nf_conntrack_l4proto.h>
>> @@ -543,6 +544,29 @@ nf_conntrack_hash_sysctl(const struct ctl_table *table, int write,
>> return ret;
>> }
>>
>> +static int
>> +nf_conntrack_log_invalid_sysctl(const struct ctl_table *table, int write,
>> + void *buffer, size_t *lenp, loff_t *ppos)
>> +{
>> + int ret, i;
>> +
>> + ret = proc_dou8vec_minmax(table, write, buffer, lenp, ppos);
>> + if (ret < 0 || !write)
>> + return ret;
>> +
>> + if (*(u8 *)table->data == 0)
>> + return ret;
>
> What is this table->data check for? I don't find any similar idiom
> like this in the existing proc_dou8vec_minmax() callers.
My intention was to avoid the unnecessary nf_log_is_registered()
calls when a user disables the feature by writing '0' to the sysctl.
But it's not a big deal. I will remove this check in v3 ;)
>> +
>> + /* Load nf_log_syslog only if no logger is currently registered */
>> + for (i = 0; i < NFPROTO_NUMPROTO; i++) {
>> + if (nf_log_is_registered(i))
>> + return ret;
>> + }
>> + request_module("%s", "nf_log_syslog");
>> +
>> + return ret;
>> +}
>> +
>> static struct ctl_table_header *nf_ct_netfilter_header;
>>
>> enum nf_ct_sysctl_index {
>> @@ -649,7 +673,7 @@ static struct ctl_table nf_ct_sysctl_table[] = {
>> .data = &init_net.ct.sysctl_log_invalid,
>> .maxlen = sizeof(u8),
>> .mode = 0644,
>> - .proc_handler = proc_dou8vec_minmax,
>> + .proc_handler = nf_conntrack_log_invalid_sysctl,
>> },
>> [NF_SYSCTL_CT_EXPECT_MAX] = {
>> .procname = "nf_conntrack_expect_max",
>> diff --git a/net/netfilter/nf_log.c b/net/netfilter/nf_log.c
>> index 6dd0de33eebd..c7dd5019a89d 100644
>> --- a/net/netfilter/nf_log.c
>> +++ b/net/netfilter/nf_log.c
>> @@ -125,6 +125,33 @@ void nf_log_unregister(struct nf_logger *logger)
>> }
>> EXPORT_SYMBOL(nf_log_unregister);
>>
>> +/**
>> + * nf_log_is_registered - Check if any logger is registered for a given
>> + * protocol family.
>> + *
>> + * @pf: Protocol family
>> + *
>> + * Returns: true if at least one logger is active for @pf, false otherwise.
>> + */
>> +bool nf_log_is_registered(u_int8_t pf)
>> +{
>> + int i;
>> +
>> + /* Out of bounds. */
>
> No need for this comment, please remove it.
OK, will remove it.
Thanks,
Lance
>
>> + if (pf >= NFPROTO_NUMPROTO) {
>> + WARN_ON_ONCE(1);
>> + return false;
>> + }
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2025-07-24 12:48 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-05-26 8:59 [PATCH v2 1/1] netfilter: load nf_log_syslog on enabling nf_conntrack_log_invalid Lance Yang
2025-05-28 11:05 ` Florian Westphal
2025-05-28 11:42 ` Lance Yang
2025-05-30 3:10 ` Lance Yang
2025-07-14 12:05 ` Lance Yang
2025-07-24 12:24 ` Pablo Neira Ayuso
2025-07-24 12:48 ` Lance Yang
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).