[LEDE-DEV] umdns - TTL restricted to 255
Hi, When using umdns I was wondering why my mDNS query did not get any answer. I found the following reason: My mDNS query has TTL field set to 1. But LEDE umdns package expects a TTL value of 255 (see interface.c function read_socket4). According https://www.systutorials.com/docs/linux/man/5-avahi-daemon.conf/ (see check-response-ttl) this was necessary security behavior in older mDNS versions but will be incompatible with newer versions of mDNS implementations. Therefore my question here is: Is this by purpose restricted to 255 by LEDE umdns package or could this be changed? Regards Philipp ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] umdns - TTL restricted to 255
My question is about query (not response). LEDE is ignoring query when TTL != 255. Philipp On 09/25/2017 05:16 PM, John Crispin wrote: Hi, rfc6762 has the following ... 11. Source Address Check All Multicast DNS responses (including responses sent via unicast) SHOULD be sent with IP TTL set to 255. This is recommended to provide backwards-compatibility with older Multicast DNS queriers (implementing a draft version of this document, posted in February 2004) that check the IP TTL on reception to determine whether the packet originated on the local link. These older queriers discard all packets with TTLs other than 255. John On 25/09/17 16:44, Philipp Meier wrote: Hi, When using umdns I was wondering why my mDNS query did not get any answer. I found the following reason: My mDNS query has TTL field set to 1. But LEDE umdns package expects a TTL value of 255 (see interface.c function read_socket4). According https://www.systutorials.com/docs/linux/man/5-avahi-daemon.conf/ (see check-response-ttl) this was necessary security behavior in older mDNS versions but will be incompatible with newer versions of mDNS implementations. Therefore my question here is: Is this by purpose restricted to 255 by LEDE umdns package or could this be changed? Regards Philipp ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] [PATCH] Remove ttl==255 restriction for queries
This commit removes ttl==255 restriction for queries.
Newer mDNS implementations use ttl != 255 when sending queries.
Signed-off-by: Philipp Meier
---
interface.c | 6 --
1 file changed, 6 deletions(-)
diff --git a/interface.c b/interface.c
index 3904c89..7f814d2 100644
--- a/interface.c
+++ b/interface.c
@@ -233,9 +233,6 @@ read_socket4(struct uloop_fd *u, unsigned int events)
}
}
- if (ttl != 255)
- return;
-
if (debug > 1) {
char buf[256];
@@ -310,9 +307,6 @@ read_socket6(struct uloop_fd *u, unsigned int events)
}
}
- if (ttl != 255)
- return;
-
if (debug > 1) {
char buf[256];
--
2.7.4
___
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] [PATCH] umdns: Remove incorrect comma in http service json config
Remove trailing comma in http service json configuration file
Signed-off-by: Philipp Meier
---
json/http.json | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/json/http.json b/json/http.json
index cf60532..2a7b50a 100644
--- a/json/http.json
+++ b/json/http.json
@@ -1,3 +1,3 @@
{
-"http_80": { "service": "_http._tcp.local", "port": 80, "txt":
[ "foo=bar"] },
+"http_80": { "service": "_http._tcp.local", "port": 80, "txt":
[ "foo=bar"] }
}
--
2.7.4
___
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] [PATCH] umdns: Add debug output for service_timeout
Debug output for case where no response is sent out due to service_timeout.
This would have saved me some ours in finding out why an mDNS query did
not get any response.
Signed-off-by: Philipp Meier
---
service.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/service.c b/service.c
index 0a9e25d..97b6f91 100644
--- a/service.c
+++ b/service.c
@@ -121,8 +121,10 @@ service_timeout(struct service *s)
{
time_t t = monotonic_time();
-if (t - s->t <= TOUT_LOOKUP)
+if (t - s->t <= TOUT_LOOKUP) {
+DBG(2, "t=%lu, s->t=%lu, t - s->t = %lu\n", t, s->t, t
- s->t);
return 0;
+}
return t;
}
--
2.7.4
___
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] umdns - TTL restricted to 255
OK - I provided a patch for this change ([PATCH] Remove ttl==255 restriction for queries). Additionally I sent two other patches: 1. [PATCH] umdns: Remove incorrect comma in hhtp service json config 2. [PATCH] umdns: Add debug output for service_timeout Philipp On 09/25/2017 06:45 PM, John Crispin wrote: in that case its a bug and the code should be changed to only have that restriction on queries John On 25/09/17 17:21, Philipp Meier wrote: My question is about query (not response). LEDE is ignoring query when TTL != 255. Philipp On 09/25/2017 05:16 PM, John Crispin wrote: Hi, rfc6762 has the following ... 11. Source Address Check All Multicast DNS responses (including responses sent via unicast) SHOULD be sent with IP TTL set to 255. This is recommended to provide backwards-compatibility with older Multicast DNS queriers (implementing a draft version of this document, posted in February 2004) that check the IP TTL on reception to determine whether the packet originated on the local link. These older queriers discard all packets with TTLs other than 255. John On 25/09/17 16:44, Philipp Meier wrote: Hi, When using umdns I was wondering why my mDNS query did not get any answer. I found the following reason: My mDNS query has TTL field set to 1. But LEDE umdns package expects a TTL value of 255 (see interface.c function read_socket4). According https://www.systutorials.com/docs/linux/man/5-avahi-daemon.conf/ (see check-response-ttl) this was necessary security behavior in older mDNS versions but will be incompatible with newer versions of mDNS implementations. Therefore my question here is: Is this by purpose restricted to 255 by LEDE umdns package or could this be changed? Regards Philipp ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] umdns: query response only every 60s
Hi When using umdns I experience the follow not expected behavior: There is no mDNS response sent for a service where an announcement or a response has been sent in the last 60 seconds (see service.c function service_timeout). 1. I started the LEDE device 2. Then I started a PC application 3. In the PC app I sent out an mDNS query 4. I wondered why I did not get a response Then by looking at the source code I detected the service_timeout function in service.c and now see that when I wait long enough with above step 3, then I get the expected response from LEDE. In the mDSN RFC I cannot directly see any such requirement. But I might interpret it wrong. Why is LEDE umdns only answering queries for a certain service every 60s? Regards Philipp ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] [PATCH 2/3] Remove incorrect comma in http service json config
Signed-off-by: Philipp Meier
---
json/http.json | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/json/http.json b/json/http.json
index cf60532..2a7b50a 100644
--- a/json/http.json
+++ b/json/http.json
@@ -1,3 +1,3 @@
{
- "http_80": { "service": "_http._tcp.local", "port": 80, "txt": [
"foo=bar"] },
+ "http_80": { "service": "_http._tcp.local", "port": 80, "txt": [
"foo=bar"] }
}
--
2.7.4
___
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] [PATCH 1/3] Remove ttl==255 restriction for queries
Signed-off-by: Philipp Meier
---
interface.c | 6 --
1 file changed, 6 deletions(-)
diff --git a/interface.c b/interface.c
index 3904c89..7f814d2 100644
--- a/interface.c
+++ b/interface.c
@@ -233,9 +233,6 @@ read_socket4(struct uloop_fd *u, unsigned int events)
}
}
- if (ttl != 255)
- return;
-
if (debug > 1) {
char buf[256];
@@ -310,9 +307,6 @@ read_socket6(struct uloop_fd *u, unsigned int events)
}
}
- if (ttl != 255)
- return;
-
if (debug > 1) {
char buf[256];
--
2.7.4
___
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] [PATCH 3/3] Add debug output for service_timeout
Signed-off-by: Philipp Meier
---
service.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/service.c b/service.c
index 0a9e25d..97b6f91 100644
--- a/service.c
+++ b/service.c
@@ -121,8 +121,10 @@ service_timeout(struct service *s)
{
time_t t = monotonic_time();
- if (t - s->t <= TOUT_LOOKUP)
+ if (t - s->t <= TOUT_LOOKUP) {
+ DBG(2, "t=%lu, s->t=%lu, t - s->t = %lu\n", t, s->t, t - s->t);
return 0;
+ }
return t;
}
--
2.7.4
___
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH] umdns: Remove incorrect comma in http service json config
On 26/09/17 10:55, Philipp Meier wrote:
Remove trailing comma in http service json configuration file
Signed-off-by: Philipp Meier
---
json/http.json | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/json/http.json b/json/http.json
index cf60532..2a7b50a 100644
--- a/json/http.json
+++ b/json/http.json
@@ -1,3 +1,3 @@
{
-"http_80": { "service": "_http._tcp.local", "port": 80,
"txt": [ "foo=bar"] },
+"http_80": { "service": "_http._tcp.local", "port": 80,
"txt": [ "foo=bar"] }
}
--
2.7.4
Hi Philipp,
your mail client chewed up all tabs and replaced them with spaces on
patches that you sent. please resend them using git send-email.
John
Hi John,
I resent all 3 patches for umdns using git send-email as you requested
for the http.json minor fix.
I have seen that the subject now misses the "umdns: ...". If this is a
problem please tell me.
Philipp
___
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH] umdns: Remove incorrect comma in http service json config
On 26/09/17 10:55, Philipp Meier wrote:
Remove trailing comma in http service json configuration file
Signed-off-by: Philipp Meier
---
json/http.json | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/json/http.json b/json/http.json
index cf60532..2a7b50a 100644
--- a/json/http.json
+++ b/json/http.json
@@ -1,3 +1,3 @@
{
-"http_80": { "service": "_http._tcp.local", "port": 80,
"txt": [ "foo=bar"] },
+"http_80": { "service": "_http._tcp.local", "port": 80,
"txt": [ "foo=bar"] }
}
--
2.7.4
Hi Philipp,
your mail client chewed up all tabs and replaced them with spaces on
patches that you sent. please resend them using git send-email.
John
Hi John,
I resent all 3 patches for umdns using git send-email as you requested
for the http.json minor fix.
I have seen that the subject now misses the "umdns: ...". If this is a
problem please tell me.
Philipp
___
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
Hi John,
would it help you if I resend the patch mails once again with better
subject ("umdns: " in front of subject line) and not as series of patch
but as 3 single patch mails?
Philipp
___
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH 3/3] Add debug output for service_timeout
On 09/28/2017 09:35 PM, Philip Prindeville wrote:
Inline
Sent from my iPhone
On Sep 28, 2017, at 1:09 AM, Philipp Meier wrote:
Signed-off-by: Philipp Meier
---
service.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/service.c b/service.c
index 0a9e25d..97b6f91 100644
--- a/service.c
+++ b/service.c
@@ -121,8 +121,10 @@ service_timeout(struct service *s)
{
time_t t = monotonic_time();
-if (t - s->t <= TOUT_LOOKUP)
+if (t - s->t <= TOUT_LOOKUP) {
+DBG(2, "t=%lu, s->t=%lu, t - s->t = %lu\n", t, s->t, t - s->t);
Do you need to write "t - s->t" or would "elapsed", "remaining", or even
"delta" be more descriptive?
return 0;
+}
return t;
}
--
2.7.4
___
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
Hi Philip,
No - I do not "need" to write "t - s->t" - if John (maintainer) would
like to have a more descriptive debug message I would be happy to
provide a patch with these changes.
If not: it's anyway a debug only message which is only meaningful for
people looking at the source code.
Regards,
Philipp
___
Lede-dev mailing list
Lede-dev@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/lede-dev
