PERFORCE change 181136 for review
Edward Tomasz Napierala
trasz at FreeBSD.org
Sun Jul 18 12:56:14 UTC 2010
http://p4web.freebsd.org/@@181136?ac=10
Change 181136 by trasz at trasz_victim on 2010/07/18 12:55:21
Update TODO.
Affected files ...
.. //depot/projects/soc2009/trasz_limits/TODO#16 edit
Differences ...
==== //depot/projects/soc2009/trasz_limits/TODO#16 (text+ko) ====
@@ -1,27 +1,22 @@
Limits done:
- - HRL_RESOURCE_OPENFILES
- - HRL_RESOURCE_DATASIZE
- - HRL_RESOURCE_FILESIZE
- - HRL_RESOURCE_MAXPROCESSES
- - HRL_RESOURCE_PTY
+ - data size (RUSAGE_DATA), in megabytes
+ - maximum file size (RUSAGE_FSIZE), in megabytes
+ - number of processes (RUSAGE_NPROC)
+ - virtual memory usage (address space limit) (RUSAGE_VMEM), in megabytes
Limits to do:
Milestone 1:
- - CPU time, in seconds
- - data size, in megabytes
- - stack size, in megabytes,
- - resident set size (physical memory usage), in megabytes
- - locked memory usage, in megabytes
- - virtual memory usage (address space limit), in megabytes
- - swap usage, in megabytes
- - number of processes
- - number of file descriptors
- - amount of memory consumed by socket buffers, in megabytes
- - maximum core size, in megabytes (core, as in ELF program state dump)
- - maximum file size, in megabytes
+ - CPU time (RUSAGE_CPU), in seconds
+ - stack size (RUSAGE_STACK), in megabytes,
+ - resident set size (physical memory usage) (RUSAGE_RSS), in megabytes
+ - locked memory usage (RUSAGE_MEMLOCK), in megabytes
+ - swap usage (RUSAGE_SWAP), in megabytes
+ - number of file descriptors (RUSAGE_NOFILE)
+ - amount of memory consumed by socket buffers (RUSAGE_SBSIZE), in megabytes
+ - maximum core size (RUSAGE_CORE), in megabytes (core, as in ELF program state dump)
Milestone 2:
@@ -36,11 +31,14 @@
- number of SysV shared memory segments
- SysV shared memory size, in megabytes
+Milestone 3:
+
+ - inheritance: different resources need to sum things up differently
+ - per-jail containers
+ - per-jail resource limits
+
HRL-specific issues:
- - Make the limits lists protected by the subjects lock (e.g. process lock)
- instead of hrl_lock.
-
- Bring back per-group limits.
- Some things need to be accounted for per-euid, and some per-egid. Geez.
@@ -51,5 +49,7 @@
- Add sorting to hrl(8).
+ - Should priv(9) checking be done by HRL, or by the callers?
+
- Use expand_number(3) in hrl(8).
More information about the p4-projects
mailing list