Skip to contentSkip to conversation

dco: move availability check to the end of check_option_conflict() function

PATCH Superseded Antonio Quartulli 3 participants 4 messages Aug 2, 2022 – Aug 18, 2022

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 dco: move availability check to the end of check_option_conflict() function

To better arrange the order DCO option conflict messages are printed, we decided to first perform all needed checks on provided options and, only at the end, if no conflict was detected, to check if DCO is really available on the system.

This way a user gets prompted with all warnings about their configuration first and, when everything is fixed, they will see if DCO is available or not.

While at it, compress the first check in just one if to make the code simpler.

---
 src/openvpn/dco.c | 22 +++++++++-------------
 1 file changed, 9 insertions(+), 13 deletions(-)
src/openvpn/dco.c +9 -13
index a6912d4e..fdf474ca 100644
@@ -268,18 +268,11 @@ dco_check_option_conflict_ce(const struct connection_entry *ce, int msglevel)
 bool
 dco_check_option_conflict(int msglevel, const struct options *o)
 {
-    if (o->tuntap_options.disable_dco)
-    {
-        /* already disabled by --disable-dco, no need to print warnings */
-        return false;
-    }
-
-    if (!dco_available(msglevel))
-    {
-        return false;
-    }
-
-    if (!o->dev)
+    /* check if DCO was already disabled by the user or if no dev name was
+     * specified at all. In the latter case, later logic will most likely stop
+     * OpenVPN, so no need to print any message here.
+     */
+    if (o->tuntap_options.disable_dco || !o->dev)
     {
         return false;
     }
@@ -361,7 +354,10 @@ dco_check_option_conflict(int msglevel, const struct options *o)
     }
     gc_free(&gc);
 
-    return true;
+    /* now that all options have been confirmed to be supported, check if DCO is
+     * truly available on the system
+     */
+    return dco_available(o);
 }
 
 bool
Signature · 1 line
2.35.1
#1 lore raw

PATCH v2 1/1 dco: move availability check to the end of check_option_conflict() function

To better arrange the order DCO option conflict messages are printed, we decided to first perform all needed checks on provided options and, only at the end, if no conflict was detected, to check if DCO is really available on the system.

This way a user gets prompted with all warnings about their configuration first and, when everything is fixed, they will see if DCO is available or not.

While at it, compress the first check in just one if to make the code simpler.

---

Changes from v1:
* pass proper argument to dco_available()

---
 src/openvpn/dco.c | 22 +++++++++-------------
 1 file changed, 9 insertions(+), 13 deletions(-)
src/openvpn/dco.c +9 -13
index a6912d4e..0877f0af 100644
@@ -268,18 +268,11 @@ dco_check_option_conflict_ce(const struct connection_entry *ce, int msglevel)
 bool
 dco_check_option_conflict(int msglevel, const struct options *o)
 {
-    if (o->tuntap_options.disable_dco)
-    {
-        /* already disabled by --disable-dco, no need to print warnings */
-        return false;
-    }
-
-    if (!dco_available(msglevel))
-    {
-        return false;
-    }
-
-    if (!o->dev)
+    /* check if DCO was already disabled by the user or if no dev name was
+     * specified at all. In the latter case, later logic will most likely stop
+     * OpenVPN, so no need to print any message here.
+     */
+    if (o->tuntap_options.disable_dco || !o->dev)
     {
         return false;
     }
@@ -361,7 +354,10 @@ dco_check_option_conflict(int msglevel, const struct options *o)
     }
     gc_free(&gc);
 
-    return true;
+    /* now that all options have been confirmed to be supported, check if DCO is
+     * truly available on the system
+     */
+    return dco_available(msglevel);
 }
 
 bool
Signature · 1 line
2.35.1
Acked-by on v2 1/1
#2 lore raw

FWIW I put this through the buildbot as a test for the new extended t_client tests on the docker workers and it caused no issues. Also ran the t_client tests on my DCO-enabled Ubuntu 22 laptop. I did not do any more specific tests.

Changes look sensible to me, so

Quoted 60 lines · Antonio Quartulli

To better arrange the order DCO option conflict messages are printed, we decided to first perform all needed checks on provided options and, only at the end, if no conflict was detected, to check if DCO is really available on the system.

This way a user gets prompted with all warnings about their configuration first and, when everything is fixed, they will see if DCO is available or not.

While at it, compress the first check in just one if to make the code simpler.

---

Changes from v1:
* pass proper argument to dco_available()

---
 src/openvpn/dco.c | 22 +++++++++-------------
 1 file changed, 9 insertions(+), 13 deletions(-)
src/openvpn/dco.c +9 -13
index a6912d4e..0877f0af 100644
@@ -268,18 +268,11 @@ dco_check_option_conflict_ce(const struct connection_entry *ce, int msglevel)
 bool
 dco_check_option_conflict(int msglevel, const struct options *o)
 {
-    if (o->tuntap_options.disable_dco)
-    {
-        /* already disabled by --disable-dco, no need to print warnings */
-        return false;
-    }
-
-    if (!dco_available(msglevel))
-    {
-        return false;
-    }
-
-    if (!o->dev)
+    /* check if DCO was already disabled by the user or if no dev name was
+     * specified at all. In the latter case, later logic will most likely stop
+     * OpenVPN, so no need to print any message here.
+     */
+    if (o->tuntap_options.disable_dco || !o->dev)
     {
         return false;
     }
@@ -361,7 +354,10 @@ dco_check_option_conflict(int msglevel, const struct options *o)
     }
     gc_free(&gc);
 
-    return true;
+    /* now that all options have been confirmed to be supported, check if DCO is
+     * truly available on the system
+     */
+    return dco_available(msglevel);
 }
 
Signature · 1 line
  Frank Lichtenheld
#3 lore raw
In reply to Antonio Quartulli

Thanks. This was left hanging in the cold for some reason... merged now. I've subjected it to the usual test for DCO related stuff (client with no-dco kernel, client with dco, server with dco) and verified that the same instances have DCO enabled that had before - glad for Arne's GLOBAL_STATS patch now :-)

This *should* be safe even on windows, as windows patch 1/7 changes the first condition to "if (!dco_enabled(o) ...)" so it won't abort on DCO-incompatible option in a config that does not have "windows-driver dco" in the first place.

I have rewrapped the comment to better fit into 80 char line constraints (since the second line was quite short)

Your patch has been applied to the master branch.

commit 78b8d0e162e1fa34780ee3d3ea84691539a0b1f3

Author: Antonio Quartulli
Date:   Tue Aug 2 15:03:12 2022 +0200
     dco: move availability check to the end of check_option_conflict() function

     Signed-off-by: Antonio Quartulli <[email protected]...>
     Acked-by: Frank Lichtenheld <[email protected]...>
     Message-Id: <[email protected]...>
     URL: https://www.mail-archive.com/[email protected]/msg24783.html
     Signed-off-by: Gert Doering <[email protected]...>

--
kind regards,

Gert Doering