diff options
author | root <root> | 2010-10-31 20:20:20 +0000 |
---|---|---|
committer | root <root> | 2010-10-31 20:20:20 +0000 |
commit | f8e3d0d4f02d2d8e21d76f8b6c5183c1136f2e57 (patch) | |
tree | 844fedc6ac9a2a28a46969357c74a8fb16ca66c8 /ev.pod | |
parent | 84d2ac4060205af81c55bf81ab802cd57ce08c19 (diff) |
*** empty log message ***
Diffstat (limited to 'ev.pod')
-rw-r--r-- | ev.pod | 5 |
1 files changed, 3 insertions, 2 deletions
@@ -465,8 +465,9 @@ epoll scales either O(1) or O(active_fds). The epoll mechanism deserves honorable mention as the most misdesigned of the more advanced event mechanisms: mere annoyances include silently dropping file descriptors, requiring a system call per change per file -descriptor (and unnecessary guessing of parameters), problems with dup and -so on. The biggest issue is fork races, however - if a program forks then +descriptor (and unnecessary guessing of parameters), problems with dup, +returning before the timeout value requiring additional iterations and so +on. The biggest issue is fork races, however - if a program forks then I<both> parent and child process have to recreate the epoll set, which can take considerable time (one syscall per file descriptor) and is of course hard to detect. |