summaryrefslogtreecommitdiffstats
path: root/runtime/nsdpoll_ptcp.c
Commit message (Collapse)AuthorAgeFilesLines
* bugfix (kind of): memory leak with tcp reception epoll handlerRainer Gerhards2011-04-131-2/+12
| | | | | This was an extremely unlikely leak and, if it happend, quite small. Still it is better to handle this border case.
* Merge branch 'v5-devel'Rainer Gerhards2011-03-141-5/+8
|\ | | | | | | | | Conflicts: tests/diag.sh
| * Fall back to epoll_create() if epoll_create1() is not availableMichael Biebl2011-03-141-5/+8
| | | | | | | | | | | | | | | | | | epoll_create1() was introduced in Linux kernel 2.6.27. If rsyslog was compiled on a newer kernel but run on a kernel older than 2.6.27, remote syslog fails. Apply a runtime check for epoll_create1() and fall back to epoll_create() in this case. Bug-Debian: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=617996
* | fixing regression: multi-threading had racesRainer Gerhards2011-01-311-0/+8
| |
* | resolved problem with intermixing requestsRainer Gerhards2011-01-271-2/+2
| |
* | experimental: added thread pool to tcpsrv epoll handlerRainer Gerhards2011-01-271-1/+2
| | | | | | | | | | this seems to work in lab, but is brand-new code. needs practice drill.
* | interim commit: refactored epoll processingRainer Gerhards2011-01-261-5/+4
| | | | | | | | this is a perquisite for multi-threading the input handler
* | improved tcpsrv performance by enabling multiple-entry epollRainer Gerhards2011-01-251-10/+18
|/ | | | | | so far, we always pulled a single event from the epoll interface. Now 128, what should result in performance improvement (less API calls) on busy systems. Most importantly affects imtcp.
* bugfix: compile failed with --enable-unlimited-selectvarmojfekoj2010-11-051-1/+0
| | | | Signed-off-by: Rainer Gerhards <rgerhards@adiscon.com>
* bugfix: failed to compile on systems without epoll supportRainer Gerhards2010-03-021-0/+4
|
* solved some platform issues with nsdpoll_ptcpRainer Gerhards2009-11-251-3/+3
|
* fixing some potential segfault conditionsRainer Gerhards2009-11-251-5/+10
|
* milestone commit: first working version with epoll/tcpRainer Gerhards2009-11-231-10/+50
| | | | ... but not well-tested, so there may be many hidden bugs.
* milestone commit: first driver layer call doneRainer Gerhards2009-11-181-4/+123
| | | | | ... does not really run. We can now call into the epoll driver, but not handle epoll(). The driver also needs more modifications.
* added skeleton for supporting epoll() API in netstream subsystemRainer Gerhards2009-11-181-0/+120