AfterNET has noticed while using the (soon to be obsolete) X2 5.0.3 that when a supergline was set to catch sub7 connections (and it went off a lot over a period of time) that some kind of corruption caused segfaults in other seemingly random places.
A supergline was added on *@*$sub7server and a rather large population of connections with sub7server as there 'realname' flagged 1 hour glines based on the sgline. Over a period of between 5 minutes and 9 hours, X2 would crash, often in IDENTIFY or AUTH commands, while writing to files (sometimes deep inside the fopen() function).
I've scanned over related code and I can find no obvious explination. We suspected for a while that some file descriptors were not being closed properly, but I can find no use of fopen() relating to glines in any way other then the gline data file functions, which appear to close the file descriptor correctly.
Because of the near completion of x2 5.1, i'm not going to persue this any further. Its advised anyone using X2 5.0 not use the sgline feature.
An identical sgline added to pre 5.1 test bot did not result in a crash after more then 24 hours of similar exposure, so the problem is assumed repaired in the new version anyway.. this is just a heads up in case anyones using the old version.
-Rubin
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
AfterNET has noticed while using the (soon to be obsolete) X2 5.0.3 that when a supergline was set to catch sub7 connections (and it went off a lot over a period of time) that some kind of corruption caused segfaults in other seemingly random places.
A supergline was added on *@*$sub7server and a rather large population of connections with sub7server as there 'realname' flagged 1 hour glines based on the sgline. Over a period of between 5 minutes and 9 hours, X2 would crash, often in IDENTIFY or AUTH commands, while writing to files (sometimes deep inside the fopen() function).
I've scanned over related code and I can find no obvious explination. We suspected for a while that some file descriptors were not being closed properly, but I can find no use of fopen() relating to glines in any way other then the gline data file functions, which appear to close the file descriptor correctly.
Because of the near completion of x2 5.1, i'm not going to persue this any further. Its advised anyone using X2 5.0 not use the sgline feature.
An identical sgline added to pre 5.1 test bot did not result in a crash after more then 24 hours of similar exposure, so the problem is assumed repaired in the new version anyway.. this is just a heads up in case anyones using the old version.
-Rubin