On Sat, Jun 2, 2018 at 10:38 PM Christopher Schultz <
ch...@christopherschultz.net> wrote:

> Rémy,
>
> On 6/2/18 3:40 PM, r...@apache.org wrote:
> > Author: remm
> > Date: Sat Jun  2 19:40:31 2018
> > New Revision: 1832744
> >
> > URL: http://svn.apache.org/viewvc?rev=1832744&view=rev
> > Log:
> > Add more missing timeouts, submitted by Alex Marchevskiy.
> >
> > Modified:
> >     tomcat/trunk/java/org/apache/tomcat/util/net/SecureNio2Channel.java
> >
> > Modified:
> tomcat/trunk/java/org/apache/tomcat/util/net/SecureNio2Channel.java
> > URL:
> http://svn.apache.org/viewvc/tomcat/trunk/java/org/apache/tomcat/util/net/SecureNio2Channel.java?rev=1832744&r1=1832743&r2=1832744&view=diff
> >
> ==============================================================================
> > --- tomcat/trunk/java/org/apache/tomcat/util/net/SecureNio2Channel.java
> (original)
> > +++ tomcat/trunk/java/org/apache/tomcat/util/net/SecureNio2Channel.java
> Sat Jun  2 19:40:31 2018
> > @@ -371,7 +371,8 @@ public class SecureNio2Channel extends N
> >          // an optimisation for the typical case so we don't create an
> >          // SNIExtractor only to discover there is no data to process
> >          if (netInBuffer.position() == 0) {
> > -            sc.read(netInBuffer, socket,
> handshakeReadCompletionHandler);
> > +            sc.read(netInBuffer,
> Nio2Endpoint.toNio2Timeout(endpoint.getConnectionTimeout()),
> > +                    TimeUnit.MILLISECONDS, socket,
> handshakeReadCompletionHandler);
>
> Is "Nio2Endpoint.toNio2Timeout(endpoint.getConnectionTimeout())", etc.
> something worth computing once and storing? I know its a simple
> comparison, but it could improve CPU usage/performance a small amount.
>

Hum, no because only one of the reads will be called during each processSNI
invocation (it then returns 1 after it). And it's extremely inexpensive too.

Rémy

Reply via email to