Skip to contentSkip to conversation

avoid using inet_ntoa() and inet_aton()

PATCH Applied Davide Caratti 2 participants 11 messages Jun 3, 2021 – Jun 10, 2021

lore
Share & tools
lore
#0 lore raw

cover avoid using inet_ntoa() and inet_aton()

some static checkers, like rpminspect, complain about use of "forbidden" functions inet_aton() and inet_ntoa(). Use inet_pton() and inet_ntop() instead.

Davide Caratti (2):
avoid using inet_ntoa()
avoid using inet_aton()

 ell/acd.c                    |  5 +++--
 ell/dhcp-lease.c             |  3 ++-
 ell/dhcp-server.c            | 25 ++++++++++++++-----------
 ell/dhcp.c                   | 17 +++++++++++++++--
 ell/rtnl.c                   |  7 +++++--
 examples/https-server-test.c |  4 +++-
 6 files changed, 42 insertions(+), 19 deletions(-)
Signature · 1 line
2.31.1
#1 lore raw

PATCH 1/2 avoid using inet_ntoa()

use inet_ntop() and a temporary buffer, like it's done elsewhere for IPv6.

---
 ell/acd.c                    |  5 +++--
 ell/dhcp-lease.c             |  3 ++-
 ell/dhcp-server.c            | 11 +++++++----
 ell/dhcp.c                   | 17 +++++++++++++++--
 ell/rtnl.c                   |  7 +++++--
 examples/https-server-test.c |  4 +++-
 6 files changed, 35 insertions(+), 12 deletions(-)
ell/acd.c +3 -2
index 6989f82..e267a96 100644
@@ -57,10 +57,11 @@
 
 #define IP_STR(uint_ip) \
 ({ \
+	char _buf[INET_ADDRSTRLEN]; \
 	struct in_addr _in; \
-	char *_out; \
+	const char *_out; \
 	_in.s_addr = uint_ip; \
-	_out = inet_ntoa(_in); \
+	_out = inet_ntop(AF_INET, &_in, _buf, INET_ADDRSTRLEN) ?: "(inv)"; \
 	_out; \
 })
 
ell/dhcp-lease.c +2 -1
index 44c815f..94b67b4 100644
@@ -178,12 +178,13 @@ error:
 static inline char *get_ip(uint32_t ip)
 {
 	struct in_addr addr;
+	char buf[INET_ADDRSTRLEN];
 
 	if (ip == 0)
 		return NULL;
 
 	addr.s_addr = ip;
-	return l_strdup(inet_ntoa(addr));
+	return l_strdup(inet_ntop(AF_INET, &addr, buf, INET_ADDRSTRLEN));
 }
 
 LIB_EXPORT char *l_dhcp_lease_get_address(const struct l_dhcp_lease *lease)
ell/dhcp-server.c +7 -4
index ebd4438..43f06b3 100644
@@ -91,10 +91,11 @@ struct l_dhcp_server {
 
 #define IP_STR(uint_ip) \
 ({ \
+	char _buf[INET_ADDRSTRLEN]; \
 	struct in_addr _in; \
-	char *_out; \
+	const char *_out; \
 	_in.s_addr = uint_ip; \
-	_out = inet_ntoa(_in); \
+	_out = inet_ntop(AF_INET, &_in, _buf, INET_ADDRSTRLEN) ?: "(inv)"; \
 	_out; \
 })
 
@@ -750,6 +751,7 @@ LIB_EXPORT void l_dhcp_server_destroy(struct l_dhcp_server *server)
 
 LIB_EXPORT bool l_dhcp_server_start(struct l_dhcp_server *server)
 {
+	char buf[INET_ADDRSTRLEN];
 	struct in_addr ia;
 
 	if (unlikely(!server))
@@ -838,9 +840,10 @@ LIB_EXPORT bool l_dhcp_server_start(struct l_dhcp_server *server)
 	ia.s_addr = server->address;
 
 	/* In case of unit testing we don't want this to be a fatal error */
-	if (!l_acd_start(server->acd, inet_ntoa(ia))) {
+	if (!inet_ntop(AF_INET, &ia, buf, INET_ADDRSTRLEN) ||
+	    !l_acd_start(server->acd, buf)) {
 		SERVER_DEBUG("Failed to start ACD on %s, continuing without",
-				IP_STR(server->address));
+			     IP_STR(server->address));
 
 		l_acd_destroy(server->acd);
 		server->acd = NULL;
ell/dhcp.c +15 -2
index a4afa95..d7809cb 100644
@@ -55,6 +55,17 @@
 
 #define BITS_PER_LONG (sizeof(unsigned long) * 8)
 
+#define IP_STR(uint_ip) \
+({ \
+	char _buf[INET_ADDRSTRLEN]; \
+	struct in_addr _in; \
+	const char *_out; \
+	_in.s_addr = uint_ip; \
+	_out = inet_ntop(AF_INET, &_in, _buf, INET_ADDRSTRLEN) ?: "(inv)"; \
+	_out; \
+})
+
+
 enum dhcp_state {
 	DHCP_STATE_INIT,
 	DHCP_STATE_SELECTING,
@@ -778,6 +789,7 @@ static void dhcp_client_rx_message(const void *data, size_t len, void *userdata)
 	struct l_dhcp_client *client = userdata;
 	const struct dhcp_message *message = data;
 	struct dhcp_message_iter iter;
+	char buf[INET_ADDRSTRLEN];
 	uint8_t msg_type = 0;
 	uint8_t t, l;
 	const void *v;
@@ -911,9 +923,10 @@ static void dhcp_client_rx_message(const void *data, size_t len, void *userdata)
 		ia.s_addr = client->lease->address;
 
 		/* For unit testing we don't want this to be a fatal error */
-		if (!l_acd_start(client->acd, inet_ntoa(ia))) {
+		if (!inet_ntop(AF_INET, &ia, buf, INET_ADDRSTRLEN) ||
+		    !l_acd_start(client->acd, buf)) {
 			CLIENT_DEBUG("Failed to start ACD on %s, continuing",
-						inet_ntoa(ia));
+				     IP_STR(ia.s_addr));
 			l_acd_destroy(client->acd);
 			client->acd = NULL;
 		}
ell/rtnl.c +5 -2
index 957e749..82ad941 100644
@@ -752,6 +752,7 @@ LIB_EXPORT uint32_t l_rtnl_set_powered(struct l_netlink *rtnl, int ifindex,
 LIB_EXPORT void l_rtnl_ifaddr4_extract(const struct ifaddrmsg *ifa, int bytes,
 				char **label, char **ip, char **broadcast)
 {
+	char buf[INET_ADDRSTRLEN];
 	struct in_addr in_addr;
 	struct rtattr *attr;
 
@@ -763,7 +764,8 @@ LIB_EXPORT void l_rtnl_ifaddr4_extract(const struct ifaddrmsg *ifa, int bytes,
 				break;
 
 			in_addr = *((struct in_addr *) RTA_DATA(attr));
-			*ip = l_strdup(inet_ntoa(in_addr));
+			*ip = l_strdup(inet_ntop(AF_INET, &in_addr, buf,
+						 INET_ADDRSTRLEN));
 
 			break;
 		case IFA_BROADCAST:
@@ -771,7 +773,8 @@ LIB_EXPORT void l_rtnl_ifaddr4_extract(const struct ifaddrmsg *ifa, int bytes,
 				break;
 
 			in_addr = *((struct in_addr *) RTA_DATA(attr));
-			*broadcast = l_strdup(inet_ntoa(in_addr));
+			*broadcast = l_strdup(inet_ntop(AF_INET, &in_addr, buf,
+							INET_ADDRSTRLEN));
 
 			break;
 		case IFA_LABEL:
examples/https-server-test.c +3 -1
index 6362722..ddf326c 100644
@@ -199,12 +199,14 @@ int main(int argc, char *argv[])
 			https_tls_ready, https_tls_disconnected, NULL);
 
 	if (getenv("TLS_DEBUG")) {
+		char b[INET_ADDRSTRLEN];
 		char *str;
 
 		l_tls_set_debug(tls, https_tls_debug_cb, NULL, NULL);
 
 		str = l_strdup_printf("/tmp/ell-certchain-%s.pem",
-					inet_ntoa(client_addr.sin_addr));
+				      inet_ntop(AF_INET, &client_addr.sin_addr,
+						b, INET_ADDRSTRLEN) ?: "NA");
 		l_tls_set_cert_dump_path(tls, str);
 		l_free(str);
 	}
Signature · 1 line
2.31.1
#9 lore raw

PATCH 2/2 avoid using inet_aton()

In reply to Davide Caratti

use inet_pton(AF_INET,...), like it's done elsewhere for IPv6.

---
 ell/dhcp-server.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
ell/dhcp-server.c +7 -7
index 43f06b3..e39e456 100644
@@ -773,7 +773,7 @@ LIB_EXPORT bool l_dhcp_server_start(struct l_dhcp_server *server)
 
 	/* Assign a default netmask if not already */
 	if (!server->netmask) {
-		if (inet_aton("255.255.255.0", &ia) < 0)
+		if (inet_pton(AF_INET,"255.255.255.0", &ia) != 1)
 			return false;
 
 		server->netmask = ia.s_addr;
@@ -890,12 +890,12 @@ LIB_EXPORT bool l_dhcp_server_set_ip_range(struct l_dhcp_server *server,
 	if (unlikely(!server || !start_ip || !end_ip))
 		return false;
 
-	if (inet_aton(start_ip, &_host_addr) == 0)
+	if (inet_pton(AF_INET, start_ip, &_host_addr) != 1)
 		return false;
 
 	start = ntohl(_host_addr.s_addr);
 
-	if (inet_aton((const char *) end_ip, &_host_addr) == 0)
+	if (inet_pton(AF_INET, (const char *) end_ip, &_host_addr) != 1)
 		return false;
 
 	server->start_ip = start;
@@ -952,7 +952,7 @@ LIB_EXPORT bool l_dhcp_server_set_ip_address(struct l_dhcp_server *server,
 	if (unlikely(!server))
 		return false;
 
-	if (inet_aton(ip, &ia) < 0)
+	if (inet_pton(AF_INET, ip, &ia) != 1)
 		return false;
 
 	server->address = ia.s_addr;
@@ -980,7 +980,7 @@ LIB_EXPORT bool l_dhcp_server_set_netmask(struct l_dhcp_server *server,
 	if (unlikely(!server || !mask))
 		return false;
 
-	if (inet_aton(mask, &ia) < 0)
+	if (inet_pton(AF_INET, mask, &ia) != 1)
 		return false;
 
 	server->netmask = ia.s_addr;
@@ -996,7 +996,7 @@ LIB_EXPORT bool l_dhcp_server_set_gateway(struct l_dhcp_server *server,
 	if (unlikely(!server || !ip))
 		return false;
 
-	if (inet_aton(ip, &ia) < 0)
+	if (inet_pton(AF_INET, ip, &ia) != 1)
 		return false;
 
 	server->gateway = ia.s_addr;
@@ -1017,7 +1017,7 @@ LIB_EXPORT bool l_dhcp_server_set_dns(struct l_dhcp_server *server, char **dns)
 	for (i = 0; dns[i]; i++) {
 		struct in_addr ia;
 
-		if (inet_aton(dns[i], &ia) < 0)
+		if (inet_pton(AF_INET, dns[i], &ia) != 1)
 			goto failed;
 
 		dns_list[i] = ia.s_addr;
Signature · 1 line
2.31.1
#10 lore raw

Re: [PATCH 2/2] avoid using inet_aton()

Hi Davide,

Quoted 3 lines · Davide Caratti

use inet_pton(AF_INET,...), like it's done elsewhere for IPv6.

We do not use Signed-off-by here, so I dropped this.

Quoted 4 lines
---
  ell/dhcp-server.c | 14 +++++++-------
  1 file changed, 7 insertions(+), 7 deletions(-)

Applied, thanks.

Regards,
-Denis

#2 lore raw

Re: [PATCH 1/2] avoid using inet_ntoa()

In reply to Davide Caratti

Hi Davide,

Quoted 10 lines · Davide Caratti

use inet_ntop() and a temporary buffer, like it's done elsewhere for IPv6.

---
  ell/acd.c                    |  5 +++--
  ell/dhcp-lease.c             |  3 ++-
  ell/dhcp-server.c            | 11 +++++++----
  ell/dhcp.c                   | 17 +++++++++++++++--
  ell/rtnl.c                   |  7 +++++--
  examples/https-server-test.c |  4 +++-

Strictly speaking the examples/ change should be in its own patch. See HACKING.

Quoted 66 lines
  6 files changed, 35 insertions(+), 12 deletions(-)
ell/acd.c +3 -2
index 6989f82..e267a96 100644
@@ -57,10 +57,11 @@
 
  #define IP_STR(uint_ip) \
  ({ \
+	char _buf[INET_ADDRSTRLEN]; \
  	struct in_addr _in; \
-	char *_out; \
+	const char *_out; \
  	_in.s_addr = uint_ip; \
-	_out = inet_ntoa(_in); \
+	_out = inet_ntop(AF_INET, &_in, _buf, INET_ADDRSTRLEN) ?: "(inv)"; \
  	_out; \
  })
 
ell/dhcp-lease.c +2 -1
index 44c815f..94b67b4 100644
@@ -178,12 +178,13 @@ error:
  static inline char *get_ip(uint32_t ip)
  {
  	struct in_addr addr;
+	char buf[INET_ADDRSTRLEN];
 
  	if (ip == 0)
  		return NULL;
 
  	addr.s_addr = ip;
-	return l_strdup(inet_ntoa(addr));
+	return l_strdup(inet_ntop(AF_INET, &addr, buf, INET_ADDRSTRLEN));
  }
 
  LIB_EXPORT char *l_dhcp_lease_get_address(const struct l_dhcp_lease *lease)
ell/dhcp-server.c +4 -2
index ebd4438..43f06b3 100644
@@ -91,10 +91,11 @@ struct l_dhcp_server {
 
  #define IP_STR(uint_ip) \
  ({ \
+	char _buf[INET_ADDRSTRLEN]; \
  	struct in_addr _in; \
-	char *_out; \
+	const char *_out; \
  	_in.s_addr = uint_ip; \
-	_out = inet_ntoa(_in); \
+	_out = inet_ntop(AF_INET, &_in, _buf, INET_ADDRSTRLEN) ?: "(inv)"; \
  	_out; \
  })
 
@@ -750,6 +751,7 @@ LIB_EXPORT void l_dhcp_server_destroy(struct l_dhcp_server *server)
 
  LIB_EXPORT bool l_dhcp_server_start(struct l_dhcp_server *server)
  {
+	char buf[INET_ADDRSTRLEN];
  	struct in_addr ia;
 
  	if (unlikely(!server))

ell/dhcp-server.c LIB_EXPORT bool l_dhcp_server_start(struct l_dhcp_server *server) · line 844

Quoted 6 lines
ell/dhcp-server.c +2 -1
@@ -838,9 +840,10 @@ LIB_EXPORT bool l_dhcp_server_start(struct l_dhcp_server *server)
  	ia.s_addr = server->address;
 
  	/* In case of unit testing we don't want this to be a fatal error */
-	if (!l_acd_start(server->acd, inet_ntoa(ia))) {
+	if (!inet_ntop(AF_INET, &ia, buf, INET_ADDRSTRLEN) ||
+	    !l_acd_start(server->acd, buf)) {

Note that we use only tabs for indentation... See doc/coding-style.txt for details

ell/dhcp-server.c LIB_EXPORT bool l_dhcp_server_start(struct l_dhcp_server *server) · line 846

Quoted 3 lines
ell/dhcp-server.c +1 -1
  		SERVER_DEBUG("Failed to start ACD on %s, continuing without",
-				IP_STR(server->address));
+			     IP_STR(server->address));

as above

Quoted 12 lines
ell/dhcp-server.c +0 -0
  		l_acd_destroy(server->acd);
  		server->acd = NULL;
ell/dhcp.c +1 -0
index a4afa95..d7809cb 100644
@@ -55,6 +55,17 @@
 
  #define BITS_PER_LONG (sizeof(unsigned long) * 8)
 
+#define IP_STR(uint_ip) \

ell/dhcp.c line 68

Quoted 10 lines
ell/dhcp.c +10 -0
+({ \
+	char _buf[INET_ADDRSTRLEN]; \
+	struct in_addr _in; \
+	const char *_out; \
+	_in.s_addr = uint_ip; \
+	_out = inet_ntop(AF_INET, &_in, _buf, INET_ADDRSTRLEN) ?: "(inv)"; \
+	_out; \
+})
+
+

no double empty lines please

Quoted 12 lines
ell/dhcp.c +1 -0
  enum dhcp_state {
  	DHCP_STATE_INIT,
  	DHCP_STATE_SELECTING,
@@ -778,6 +789,7 @@ static void dhcp_client_rx_message(const void *data, size_t len, void *userdata)
  	struct l_dhcp_client *client = userdata;
  	const struct dhcp_message *message = data;
  	struct dhcp_message_iter iter;
+	char buf[INET_ADDRSTRLEN];
  	uint8_t msg_type = 0;
  	uint8_t t, l;
  	const void *v;

ell/dhcp.c static void dhcp_client_rx_message(const void *data, size_t len, void *userdata) · line 929

Quoted 9 lines
ell/dhcp.c +3 -2
@@ -911,9 +923,10 @@ static void dhcp_client_rx_message(const void *data, size_t len, void *userdata)
  		ia.s_addr = client->lease->address;
 
  		/* For unit testing we don't want this to be a fatal error */
-		if (!l_acd_start(client->acd, inet_ntoa(ia))) {
+		if (!inet_ntop(AF_INET, &ia, buf, INET_ADDRSTRLEN) ||
+		    !l_acd_start(client->acd, buf)) {
  			CLIENT_DEBUG("Failed to start ACD on %s, continuing",
-						inet_ntoa(ia));
+				     IP_STR(ia.s_addr));

Not sure introducing IP_STR macro just for this is worth it, but ok...

same here for indentation

Quoted 41 lines
ell/dhcp.c +0 -0
  			l_acd_destroy(client->acd);
  			client->acd = NULL;
  		}
ell/rtnl.c +5 -2
index 957e749..82ad941 100644
@@ -752,6 +752,7 @@ LIB_EXPORT uint32_t l_rtnl_set_powered(struct l_netlink *rtnl, int ifindex,
  LIB_EXPORT void l_rtnl_ifaddr4_extract(const struct ifaddrmsg *ifa, int bytes,
  				char **label, char **ip, char **broadcast)
  {
+	char buf[INET_ADDRSTRLEN];
  	struct in_addr in_addr;
  	struct rtattr *attr;
 
@@ -763,7 +764,8 @@ LIB_EXPORT void l_rtnl_ifaddr4_extract(const struct ifaddrmsg *ifa, int bytes,
  				break;
 
  			in_addr = *((struct in_addr *) RTA_DATA(attr));
-			*ip = l_strdup(inet_ntoa(in_addr));
+			*ip = l_strdup(inet_ntop(AF_INET, &in_addr, buf,
+						 INET_ADDRSTRLEN)); >
  			break;
  		case IFA_BROADCAST:
@@ -771,7 +773,8 @@ LIB_EXPORT void l_rtnl_ifaddr4_extract(const struct ifaddrmsg *ifa, int bytes,
  				break;
 
  			in_addr = *((struct in_addr *) RTA_DATA(attr));
-			*broadcast = l_strdup(inet_ntoa(in_addr));
+			*broadcast = l_strdup(inet_ntop(AF_INET, &in_addr, buf,
+							INET_ADDRSTRLEN));
 
  			break;
  		case IFA_LABEL:
examples/https-server-test.c +0 -0
index 6362722..ddf326c 100644
@@ -199,12 +199,14 @@ int main(int argc, char *argv[])
  			https_tls_ready, https_tls_disconnected, NULL);
 

examples/https-server-test.c int main(int argc, char *argv[]) · line 209

Quoted 10 lines
examples/https-server-test.c +3 -1
  	if (getenv("TLS_DEBUG")) {
+		char b[INET_ADDRSTRLEN];
  		char *str;
 
  		l_tls_set_debug(tls, https_tls_debug_cb, NULL, NULL);
 
  		str = l_strdup_printf("/tmp/ell-certchain-%s.pem",
-					inet_ntoa(client_addr.sin_addr));
+				      inet_ntop(AF_INET, &client_addr.sin_addr,
+						b, INET_ADDRSTRLEN) ?: "NA");

Looks like indentation problem here too. Maybe you want to put the inet_ntop business outside of l_strdup_printf call to make things more readable?

examples/https-server-test.c int main(int argc, char *argv[]) · line 212

Quoted 4 lines
examples/https-server-test.c +0 -0
  		l_tls_set_cert_dump_path(tls, str);
  		l_free(str);
  	}

Regards,
-Denis

author
#3 lore raw

Re: [PATCH 1/2] avoid using inet_ntoa()

hello Denis, thanks for reviewing!

Il giorno ven 4 giu 2021 alle ore 18:01 Denis Kenzior <[email protected]> ha scritto:

Quoted 2 lines

Hi Davide,

[...]

Quoted 3 lines

Strictly speaking the examples/ change should be in its own patch. See HACKING.

I should have read it, as well as the coding style rules, before submitting _ sorry for not doing that.
I will address coding style and remove SOB line in the v2.

[...]

Quoted 10 lines
(quoted diff) +0 -0
@@ -911,9 +923,10 @@ static void dhcp_client_rx_message(const void *data,
 size_t len, void *userdata)
 >               ia.s_addr = client->lease->address;
Quoted 2 lines
              /* For unit testing we don't want this to be a fatal error

*/

ell/dhcp.c static void dhcp_client_rx_message(const void *data, size_t len, void *userdata)

Quoted 4 lines
ell/dhcp.c +2 -1
-             if (!l_acd_start(client->acd, inet_ntoa(ia))) {
+             if (!inet_ntop(AF_INET, &ia, buf, INET_ADDRSTRLEN) ||
+                 !l_acd_start(client->acd, buf)) {
                      CLIENT_DEBUG("Failed to start ACD on %s,
Quoted 6 lines

continuing",

ell/dhcp.c static void dhcp_client_rx_message(const void *data, size_t len, void *userdata) · line 929

Quoted 2 lines
ell/dhcp.c +1 -1
-                                             inet_ntoa(ia));
+                                  IP_STR(ia.s_addr));

Not sure introducing IP_STR macro just for this is worth it, but ok...

that might be useful for other future CLIENT_DEBUG() users.

[...]

Quoted 4 lines

Looks like indentation problem here too. Maybe you want to put the inet_ntop
business outside of l_strdup_printf call to make things more readable?

yes, it seems to look better - I will also fix this.
By the way, looking at the inet_ntop() documentation [1], I think that I can also avoid things like

   inet_ntop(AF_INET, &ia, buf, INET_ADDRSTRLEN) ? : "(inv)";

because the first argument is hardcoded to AF_INET and also we are sure that 'buf' is INET_ADDRSTRLEN long.
Under these conditions, the return value of inet_ntop() should always be 'buf'. Correct?

--
davide

[1] https://man7.org/linux/man-pages/man3/inet_ntop.3.html
RispondiInoltra

#4 lore raw

Re: [PATCH 1/2] avoid using inet_ntoa()

Hi Davide,

Quoted 2 lines

that might be useful for other future CLIENT_DEBUG() users.

Fair enough, Wish there was a printf extension for ipv4 and ipv6 addresses like printk uses... Oh well.

Quoted 11 lines

yes, it seems to look better - I will also fix this.
By the way, looking at the inet_ntop() documentation [1], I think that I can also avoid things like

   inet_ntop(AF_INET, &ia, buf, INET_ADDRSTRLEN) ? : "(inv)";

because the first argument is hardcoded to AF_INET and also we are sure that 'buf' is INET_ADDRSTRLEN long.
Under these conditions, the return value of inet_ntop() should always be 'buf'. Correct?

I would assume so. You can tell for sure by looking at glibc implementation for example.

Oh, one other thing. Have you checked the scope rules for GCC statement expressions? inet_ntoa uses a static buffer (which glibc further enhances by marking it as thread-specific storage). So the returned pointer is guaranteed to be valid after inet_ntoa returns. The statement expression doesn't seem to do that, so that seems suspicious?

Regards,
-Denis

author
#5 lore raw

Re: [PATCH 1/2] avoid using inet_ntoa()

Il giorno lun 7 giu 2021 alle ore 20:41 Denis Kenzior <[email protected]> ha scritto:

Quoted 9 lines

I would assume so. You can tell for sure by looking at glibc implementation for example.

Oh, one other thing. Have you checked the scope rules for GCC statement expressions? inet_ntoa uses a static buffer (which glibc further enhances by marking it as thread-specific storage). So the returned pointer is guaranteed to be valid after inet_ntoa returns. The statement expression doesn't seem to do that, so that seems suspicious?

... you are right. GCC documentation [1] says:

"In a statement expression, any temporaries created within a statement are destroyed at that statement’s end."

so, we can't do IP_STR() that way. And using inet_ntop() + static buffer, like it's done by iproute2 [2] would actually silence the rpminspect warning I'm seeing, but not fix the actual
problem: AFAIK use of a static buffer is one of the 2 reasons why inet_ntoa() has been "obsoleted".

At this point, probably the cleanest thing to do is to replace IP_STR() with a proper call to inet_ntop(), preceded by declaration of buf[INET_ADDRSTRLEN] (like it's already done in some hunks of this patch). Luckily, it's just 10 places (versus 6 users of inet_ntoa(), still looks feasible).

Il giorno lun 7 giu 2021 alle ore 20:41 Denis Kenzior <[email protected]> ha scritto:

Quoted 6 lines
Quoted 3 lines

that might be useful for other future CLIENT_DEBUG() users.

Fair enough, Wish there was a printf extension for ipv4 and ipv6 addresses like printk uses... Oh well.

yes, that would really make this game less boring :)

[1] https://gcc.gnu.org/onlinedocs/gcc/Statement-Exprs.html
[2] https://github.com/shemminger/iproute2/blob/main/lib/utils.c#L980-L1020

#6 lore raw

Re: [PATCH 1/2] avoid using inet_ntoa()

Hi Davide,

Quoted 13 lines · d. caratti

Il giorno lun 7 giu 2021 alle ore 20:41 Denis Kenzior <[email protected]> ha scritto:

Quoted 9 lines

I would assume so. You can tell for sure by looking at glibc implementation for example.

Oh, one other thing. Have you checked the scope rules for GCC statement expressions? inet_ntoa uses a static buffer (which glibc further enhances by marking it as thread-specific storage). So the returned pointer is guaranteed to be valid after inet_ntoa returns. The statement expression doesn't seem to do that, so that seems suspicious?

Quoted 5 lines

... you are right. GCC documentation [1] says:

"In a statement expression, any temporaries created within a statement are destroyed at that statement’s end."

That is what I was afraid of.

Quoted 5 lines

so, we can't do IP_STR() that way. And using inet_ntop() + static buffer, like it's done by iproute2 [2] would actually silence the rpminspect warning I'm seeing, but not fix the actual
problem: AFAIK use of a static buffer is one of the 2 reasons why inet_ntoa() has been "obsoleted".

That would be unfortunate. I'm really not sure whether this needs to be 'fixed' anyway. ell is meant for single-threaded, event driven apps. So our use of inet_ntoa is just fine. In fact we do similar stuff all over the place.

Quoted 6 lines

At this point, probably the cleanest thing to do is to replace IP_STR() with a proper call to inet_ntop(), preceded by declaration of buf[INET_ADDRSTRLEN] (like it's already done in some hunks of this patch). Luckily, it's just 10 places (versus 6 users of inet_ntoa(), still looks feasible).

Perhaps we can do something similar to MAC_STR instead?

Quoted 7 lines

Il giorno lun 7 giu 2021 alle ore 20:41 Denis Kenzior <[email protected]> ha scritto:

Quoted 4 lines
Quoted 3 lines

that might be useful for other future CLIENT_DEBUG() users.

Quoted 8 lines
Quoted 2 lines

Fair enough, Wish there was a printf extension for ipv4 and ipv6 addresses like printk uses... Oh well.

yes, that would really make this game less boring :)

[1] https://gcc.gnu.org/onlinedocs/gcc/Statement-Exprs.html
[2] https://github.com/shemminger/iproute2/blob/main/lib/utils.c#L980-L1020

Yep, that's the other thing we can do. Just re-implement inet_ntoa or make our own version that covers v4/v6 with a static buffer. Not sure that is any better than just using inet_ntoa though ;)

Regards,
-Denis

author
#7 lore raw

Re: [PATCH 1/2] avoid using inet_ntoa()

Il giorno mar 8 giu 2021 alle ore 18:22 Denis Kenzior <[email protected]> ha scritto:

Quoted 10 lines

Hi Davide,

Quoted 6 lines · d. caratti

Il giorno lun 7 giu 2021 alle ore 20:41 Denis Kenzior <[email protected]> ha scritto:

Quoted 3 lines

I would assume so. You can tell for sure by looking at glibc implementation for example.

this:

https://sourceware.org/git/?p=glibc.git;a=blob;f=resolv/inet_ntop.c;h=c4d38c0f951013e51a4fc6eaa8a9b82e146abe5a;hb=HEAD

yes, the ternary operator is not needed as long as the address family is hardcoded to AF_INET and 'buf ' fits the size of 'tmp' in inet_ntop4().

Quoted 20 lines
Quoted 11 lines
Quoted 5 lines

Oh, one other thing. Have you checked the scope rules for GCC statement expressions? inet_ntoa uses a static buffer (which glibc further enhances by marking it as thread-specific storage). So the returned pointer is guaranteed to be valid after inet_ntoa returns. The statement expression doesn't seem to do that, so that seems suspicious?

... you are right. GCC documentation [1] says:

"In a statement expression, any temporaries created within a statement are destroyed at that statement’s end."

That is what I was afraid of.

Quoted 5 lines

so, we can't do IP_STR() that way. And using inet_ntop() + static buffer, like it's done by iproute2 [2] would actually silence the rpminspect warning I'm seeing, but not fix the actual
problem: AFAIK use of a static buffer is one of the 2 reasons why inet_ntoa() has been "obsoleted".

Quoted 3 lines

That would be unfortunate. I'm really not sure whether this needs to be 'fixed' anyway. ell is meant for single-threaded, event driven apps. So our use of inet_ntoa is just fine. In fact we do similar stuff all over the place.

there's nothing wrong with using inet_ntoa() as long as the caller doesn't need to be "IPv6 friendly" and concurrency is not an issue. However,
this would result in disabling the whole fedora "badfuncs" CI, that now is failing just because of inet_ntoa().

Quoted 7 lines
Quoted 5 lines

At this point, probably the cleanest thing to do is to replace IP_STR() with a proper call to inet_ntop(), preceded by declaration of buf[INET_ADDRSTRLEN] (like it's already done in some hunks of this patch). Luckily, it's just 10 places (versus 6 users of inet_ntoa(), still looks feasible).

Perhaps we can do something similar to MAC_STR instead?

that's similar to the NIPQUAD / NIPQUAD_FMT thing that was done in Linux before the %pI4 format specified was introduced in printk (
https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/commit/?id=3685f25de1b0447fff381c420de
) :

#include <stdio.h>
#include <netinet/in.h>
#include <arpa/inet.h>

#define IP "%u.%u.%u.%u"

#define IP_STR(uint_ip) ((unsigned char *) &uint_ip)[0], \
                        ((unsigned char *) &uint_ip)[1], \
                        ((unsigned char *) &uint_ip)[2], \
                        ((unsigned char *) &uint_ip)[3]

int main(int argc, const char *argv[])

{
        struct in_addr ia = { .s_addr = 0x010200c0 };
        char buf[INET_ADDRSTRLEN];

        printf("test 0 %s\n", inet_ntoa(ia));
        printf("test 1 %s\n", inet_ntop(AF_INET, &ia, buf, INET_ADDRSTRLEN));
        printf("test 2 " IP "\n", IP_STR(ia.s_addr));

        return 0;
}
Quoted 11 lines
Quoted 11 lines

Il giorno lun 7 giu 2021 alle ore 20:41 Denis Kenzior <[email protected]> ha scritto:

Quoted 6 lines
Quoted 3 lines

that might be useful for other future CLIENT_DEBUG() users.

Fair enough, Wish there was a printf extension for ipv4 and ipv6 addresses like printk uses... Oh well.

yes, that would really make this game less boring :)

Quoted 7 lines
Quoted 3 lines

[1] https://gcc.gnu.org/onlinedocs/gcc/Statement-Exprs.html
[2] https://github.com/shemminger/iproute2/blob/main/lib/utils.c#L980-L1020

Yep, that's the other thing we can do. Just re-implement inet_ntoa or make our own version that covers v4/v6 with a static buffer. Not sure that is any better than just using inet_ntoa though ;)

what about converting inet_ntoa() into inet_ntop(), and re-writing IP_STR() as above?
thanks!

Signature · 1 line
davide
#8 lore raw

Re: [PATCH 1/2] avoid using inet_ntoa()

Hi Davide,

Quoted 10 lines
Quoted 3 lines

That would be unfortunate. I'm really not sure whether this needs to be 'fixed' anyway. ell is meant for single-threaded, event driven apps. So our use of inet_ntoa is just fine. In fact we do similar stuff all over the place.

there's nothing wrong with using inet_ntoa() as long as the caller doesn't need to be "IPv6 friendly" and concurrency is not an issue. However,
this would result in disabling the whole fedora "badfuncs" CI, that now is failing just because of inet_ntoa().

Is this a worthy goal in and of itself though? What other 'badfuncs' are there?

Quoted 24 lines
Quoted 7 lines
Quoted 5 lines

At this point, probably the cleanest thing to do is to replace IP_STR() with a proper call to inet_ntop(), preceded by declaration of buf[INET_ADDRSTRLEN] (like it's already done in some hunks of this patch). Luckily, it's just 10 places (versus 6 users of inet_ntoa(), still looks feasible).

Perhaps we can do something similar to MAC_STR instead?

that's similar to the NIPQUAD / NIPQUAD_FMT thing that was done in Linux before the %pI4 format specified was introduced in printk (
https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/commit/?id=3685f25de1b0447fff381c420de
) :

#include <stdio.h>
#include <netinet/in.h>
#include <arpa/inet.h>

#define IP "%u.%u.%u.%u"

#define IP_STR(uint_ip) ((unsigned char *) &uint_ip)[0], \
                         ((unsigned char *) &uint_ip)[1], \
                         ((unsigned char *) &uint_ip)[2], \
                         ((unsigned char *) &uint_ip)[3]

int main(int argc, const char *argv[])

Quoted 11 lines
{
         struct in_addr ia = { .s_addr = 0x010200c0 };
         char buf[INET_ADDRSTRLEN];

         printf("test 0 %s\n", inet_ntoa(ia));
         printf("test 1 %s\n", inet_ntop(AF_INET, &ia, buf, INET_ADDRSTRLEN));
         printf("test 2 " IP "\n", IP_STR(ia.s_addr));

         return 0;
}

I think this or something like this would be quite reasonable to use in / for any debug messages where we print an ip.

Quoted 16 lines
Quoted 16 lines
Quoted 15 lines

Il giorno lun 7 giu 2021 alle ore 20:41 Denis Kenzior <[email protected]> ha scritto:

Quoted 6 lines
Quoted 3 lines

that might be useful for other future CLIENT_DEBUG() users.

Fair enough, Wish there was a printf extension for ipv4 and ipv6 addresses like printk uses... Oh well.

yes, that would really make this game less boring :)

[1] https://gcc.gnu.org/onlinedocs/gcc/Statement-Exprs.html
[2] https://github.com/shemminger/iproute2/blob/main/lib/utils.c#L980-L1020

Quoted 6 lines
Quoted 3 lines

Yep, that's the other thing we can do. Just re-implement inet_ntoa or make our own version that covers v4/v6 with a static buffer. Not sure that is any better than just using inet_ntoa though ;)

what about converting inet_ntoa() into inet_ntop(), and re-writing IP_STR() as above?

I guess that would be OK. We just have to remember to not use inet_ntoa in the future...

Regards,
-Denis