d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
518d5f5152
freebsd-skq/sys/kern/kern_resource.c

1416 lines
32 KiB
C
Raw Normal View History

BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*-
* Copyright (c) 1982, 1986, 1991, 1993
* The Regents of the University of California. All rights reserved.
* (c) UNIX System Laboratories, Inc.
* All or some portions of this file are derived from material licensed
* to the University of California by American Telephone and Telegraph
* Co. or Unix System Laboratories, Inc. and are reproduced herein with
* the permission of UNIX System Laboratories, Inc.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)kern_resource.c 8.5 (Berkeley) 1/21/94
*/
Use __FBSDID().
2003-06-11 00:56:59 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
Make COMPAT_43 and COMPAT_SUNOS new-style options.
1997-12-16 17:40:42 +00:00
#include "opt_compat.h"
Make DFLDSIZ and MAXDSIZ fully-supported options. "Don't forget to do a ``make depend''" :-)
1996-12-22 23:17:09 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
The big 4.4BSD Lite to FreeBSD 2.0.0 (Development) patch. Reviewed by: Rodney W. Grimes Submitted by: John Dyson and David Greenman
1994-05-25 09:21:21 +00:00
#include <sys/systm.h>
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#include <sys/sysproto.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/file.h>
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
#include <sys/kernel.h>
Catch up to header include changes: - <sys/mutex.h> now requires <sys/systm.h> - <sys/mutex.h> and <sys/sx.h> now require <sys/lock.h>
2001-03-28 09:17:56 +00:00
#include <sys/lock.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/malloc.h>
Translate alfred to english. Submitted by: bde
2000-12-01 06:59:18 +00:00
#include <sys/mutex.h>
Sweep kernel replacing suser(9) calls with priv(9) calls, assigning specific privilege names to a broad range of privileges. These may require some future tweaking. Sponsored by: nCircle Network Security, Inc. Obtained from: TrustedBSD Project Discussed on: arch@ Reviewed (at least in part) by: mlaier, jmg, pjd, bde, ceri, Alex Lyashkov <umka at sevcity dot net>, Skip Ford <skip dot ford at verizon dot net>, Antoine Brodin <antoine dot brodin at laposte dot net>
2006-11-06 13:42:10 +00:00
#include <sys/priv.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/proc.h>
Use the reference count API to manage the reference counts for process limit structures rather than using pool mutexes to protect the reference counts. Tested on: i386, alpha, sparc64
2005-09-27 18:07:05 +00:00
#include <sys/refcount.h>
Add racct. It's an API to keep per-process, per-jail, per-loginclass and per-loginclass resource accounting information, to be used by the new resource limits code. It's connected to the build, but the code that actually calls the new functions will come later. Sponsored by: The FreeBSD Foundation Reviewed by: kib (earlier version)
2011-03-29 17:47:25 +00:00
#include <sys/racct.h>
Catch up to header include changes: - <sys/mutex.h> now requires <sys/systm.h> - <sys/mutex.h> and <sys/sx.h> now require <sys/lock.h>
2001-03-28 09:17:56 +00:00
#include <sys/resourcevar.h>
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
#include <sys/rwlock.h>
- Create a new scheduler api that is defined in sys/sched.h - Begin moving scheduler specific functionality into sched_4bsd.c - Replace direct manipulation of scheduler data with hooks provided by the new api. - Remove KSE specific state modifications and single runq assumptions from kern_switch.c Reviewed by: -arch
2002-10-12 05:32:24 +00:00
#include <sys/sched.h>
Convert the allproc and proctree locks from lockmgr locks to sx locks.
2001-03-28 11:52:56 +00:00
#include <sys/sx.h>
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
#include <sys/syscallsubr.h>
Use the fields in the sysentvec and in the vm map header in place of the constants VM_MIN_ADDRESS, VM_MAXUSER_ADDRESS, USRSTACK and PS_STRINGS. This is mainly so that they can be variable even for the native abi, based on different machine types. Get stack protections from the sysentvec too. This makes it trivial to map the stack non-executable for certain abis, on machines that support it.
2002-09-21 22:07:17 +00:00
#include <sys/sysent.h>
Add a bit of sanity checking and problem avoidance in case the timecounter hardware is bogus. This will produce a new warning "microuptime() went backwards" and try to not screw up the process resource accounting.
1999-11-29 11:29:04 +00:00
#include <sys/time.h>
Fix LOR of thread lock and umtx's priority propagation mutex due to the reworking of scheduler lock. MFC: after 3 days
2007-12-11 08:25:36 +00:00
#include <sys/umtx.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <vm/vm.h>
Untangled the vm.h include file spaghetti.
1995-12-07 12:48:31 +00:00
#include <vm/vm_param.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Cosmetics. related to getting prototypes into view.
1994-10-10 01:00:49 +00:00
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
static MALLOC_DEFINE(M_PLIMIT, "plimit", "plimit structures");
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
static MALLOC_DEFINE(M_UIDINFO, "uidinfo", "uidinfo structures");
#define UIHASH(uid) (&uihashtbl[(uid) & uihash])
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
static struct rwlock uihashtbl_lock;
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
static LIST_HEAD(uihashhead, uidinfo) *uihashtbl;
static u_long uihash; /* size of hash table - 1 */
Split calcru() back into a calcru1() function shared with calccru() and a calcru() wrapper that passes a local rusage_ext on the stack that is a snapshot to do the calculations on. Now we can pass p->p_crux to calcru1() in calccru() again which fixes the issues with runtime going backwards messages when dead processes are harvested by init. Reviewed by: phk Tested by: Stefan Ehmann shoesoft at gmx dot net
2006-02-21 21:47:46 +00:00
static void calcru1(struct proc *p, struct rusage_ext *ruxp,
struct timeval *up, struct timeval *sp);
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
static int donice(struct thread *td, struct proc *chgp, int n);
static struct uidinfo *uilookup(uid_t uid);
Fix the double counting of the last process thread td_incruntime on exit, that is done once in thread_exit() and the second time in proc_reap(), by clearing td_incruntime. Use the opportunity to revert to the pre-RUSAGE_THREAD exporting of ruxagg() instead of ruxagg_locked() and use it from thread_exit(). Diagnosed and tested by: neel MFC after: 3 days
2010-05-24 10:23:49 +00:00
static void ruxagg_locked(struct rusage_ext *rux, struct thread *td);
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Resource controls and accounting.
*/
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#ifndef _SYS_SYSPROTO_H_
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct getpriority_args {
int which;
int who;
};
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#endif
The big 4.4BSD Lite to FreeBSD 2.0.0 (Development) patch. Reviewed by: Rodney W. Grimes Submitted by: John Dyson and David Greenman
1994-05-25 09:21:21 +00:00
int
In order to maximize the re-usability of kernel code in user space this patch modifies makesyscalls.sh to prefix all of the non-compatibility calls (e.g. not linux_, freebsd32_) with sys_ and updates the kernel entry points and all places in the code that use them. It also fixes an additional name space collision between the kernel function psignal and the libc function of the same name by renaming the kernel psignal kern_psignal(). By introducing this change now we will ease future MFCs that change syscalls. Reviewed by: rwatson Approved by: re (bz)
2011-09-16 13:58:51 +00:00
sys_getpriority(td, uap)
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
struct thread *td;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
register struct getpriority_args *uap;
{
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
struct proc *p;
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
struct pgrp *pg;
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
int error, low;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
error = 0;
low = PRIO_MAX + 1;
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
switch (uap->which) {
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
case PRIO_PROCESS:
Back out previous. The locking here needs a rethink.
2003-03-13 00:54:53 +00:00
if (uap->who == 0)
Nice, is a property of a process as a whole.. I mistakenly moved it to the ksegroup when breaking up the process structure. Put it back in the proc structure.
2004-06-16 00:26:31 +00:00
low = td->td_proc->p_nice;
Back out previous. The locking here needs a rethink.
2003-03-13 00:54:53 +00:00
else {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
p = pfind(uap->who);
Change the pfind() and zpfind() functions to lock the process that they find before releasing the allproc lock and returning. Reviewed by: -smp, dfr, jake
2001-04-24 00:51:53 +00:00
if (p == NULL)
break;
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
if (p_cansee(td, p) == 0)
Nice, is a property of a process as a whole.. I mistakenly moved it to the ksegroup when breaking up the process structure. Put it back in the proc structure.
2004-06-16 00:26:31 +00:00
low = p->p_nice;
Change the pfind() and zpfind() functions to lock the process that they find before releasing the allproc lock and returning. Reviewed by: -smp, dfr, jake
2001-04-24 00:51:53 +00:00
PROC_UNLOCK(p);
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
case PRIO_PGRP:
Lock proctree_lock instead of pgrpsess_lock.
2002-04-16 17:11:34 +00:00
sx_slock(&proctree_lock);
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
if (uap->who == 0) {
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
pg = td->td_proc->p_pgrp;
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
PGRP_LOCK(pg);
} else {
pg = pgfind(uap->who);
if (pg == NULL) {
Lock proctree_lock instead of pgrpsess_lock.
2002-04-16 17:11:34 +00:00
sx_sunlock(&proctree_lock);
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
break;
}
}
Lock proctree_lock instead of pgrpsess_lock.
2002-04-16 17:11:34 +00:00
sx_sunlock(&proctree_lock);
This is a partial commit of the patch from PR 14914: Alot of the code in sys/kern directly accesses the *Q_HEAD and *Q_ENTRY structures for list operations. This patch makes all list operations in sys/kern use the queue(3) macros, rather than directly accessing the *Q_{HEAD,ENTRY} structures. This batch of changes compile to the same object files. Reviewed by: phk Submitted by: Jake Burkholder <jake@checker.org> PR: 14914
1999-11-16 10:56:05 +00:00
LIST_FOREACH(p, &pg->pg_members, p_pglist) {
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
PROC_LOCK(p);
Fix several places to ignore processes that are not yet fully constructed. MFC after: 1 week
2011-04-06 17:47:22 +00:00
if (p->p_state == PRS_NORMAL &&
p_cansee(td, p) == 0) {
Nice, is a property of a process as a whole.. I mistakenly moved it to the ksegroup when breaking up the process structure. Put it back in the proc structure.
2004-06-16 00:26:31 +00:00
if (p->p_nice < low)
low = p->p_nice;
Allocate KSEs and KSEGRPs separatly and remove them from the proc structure. next step is to allow > 1 to be allocated per process. This would give multi-processor threads. (when the rest of the infrastructure is in place) While doing this I noticed libkvm and sys/kern/kern_proc.c:fill_kinfo_proc are diverging more than they should.. corrective action needed soon.
2002-09-15 23:52:25 +00:00
}
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
PROC_UNLOCK(p);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
PGRP_UNLOCK(pg);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
case PRIO_USER:
if (uap->who == 0)
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
uap->who = td->td_ucred->cr_uid;
Convert the allproc and proctree locks from lockmgr locks to sx locks.
2001-03-28 11:52:56 +00:00
sx_slock(&allproc_lock);
Use FOREACH_PROC_IN_SYSTEM instead of using its unrolled form.
2007-01-17 14:58:53 +00:00
FOREACH_PROC_IN_SYSTEM(p) {
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
PROC_LOCK(p);
Fix some locking nits with the p_state field of struct proc: - Hold the proc lock while changing the state from PRS_NEW to PRS_NORMAL in fork to honor the locking requirements. While here, expand the scope of the PROC_LOCK() on the new process (p2) to avoid some LORs. Previously the code was locking the new child process (p2) after it had locked the parent process (p1). However, when locking two processes, the safe order is to lock the child first, then the parent. - Fix various places that were checking p_state against PRS_NEW without having the process locked to use PROC_LOCK(). Every place was already locking the process, just after the PRS_NEW check. - Remove or reduce the use of PROC_SLOCK() for places that were checking p_state against PRS_NEW. The PROC_LOCK() alone is sufficient for reading the current state. - Reorder fill_kinfo_proc() slightly so it only acquires PROC_SLOCK() once. MFC after: 1 week
2011-03-24 18:40:11 +00:00
if (p->p_state == PRS_NORMAL &&
p_cansee(td, p) == 0 &&
Allocate KSEs and KSEGRPs separatly and remove them from the proc structure. next step is to allow > 1 to be allocated per process. This would give multi-processor threads. (when the rest of the infrastructure is in place) While doing this I noticed libkvm and sys/kern/kern_proc.c:fill_kinfo_proc are diverging more than they should.. corrective action needed soon.
2002-09-15 23:52:25 +00:00
p->p_ucred->cr_uid == uap->who) {
Nice, is a property of a process as a whole.. I mistakenly moved it to the ksegroup when breaking up the process structure. Put it back in the proc structure.
2004-06-16 00:26:31 +00:00
if (p->p_nice < low)
low = p->p_nice;
Allocate KSEs and KSEGRPs separatly and remove them from the proc structure. next step is to allow > 1 to be allocated per process. This would give multi-processor threads. (when the rest of the infrastructure is in place) While doing this I noticed libkvm and sys/kern/kern_proc.c:fill_kinfo_proc are diverging more than they should.. corrective action needed soon.
2002-09-15 23:52:25 +00:00
}
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
PROC_UNLOCK(p);
}
Convert the allproc and proctree locks from lockmgr locks to sx locks.
2001-03-28 11:52:56 +00:00
sx_sunlock(&allproc_lock);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
default:
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
error = EINVAL;
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
if (low == PRIO_MAX + 1 && error == 0)
error = ESRCH;
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
td->td_retval[0] = low;
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#ifndef _SYS_SYSPROTO_H_
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct setpriority_args {
int which;
int who;
int prio;
};
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#endif
The big 4.4BSD Lite to FreeBSD 2.0.0 (Development) patch. Reviewed by: Rodney W. Grimes Submitted by: John Dyson and David Greenman
1994-05-25 09:21:21 +00:00
int
In order to maximize the re-usability of kernel code in user space this patch modifies makesyscalls.sh to prefix all of the non-compatibility calls (e.g. not linux_, freebsd32_) with sys_ and updates the kernel entry points and all places in the code that use them. It also fixes an additional name space collision between the kernel function psignal and the libc function of the same name by renaming the kernel psignal kern_psignal(). By introducing this change now we will ease future MFCs that change syscalls. Reviewed by: rwatson Approved by: re (bz)
2011-09-16 13:58:51 +00:00
sys_setpriority(td, uap)
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
struct thread *td;
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
struct setpriority_args *uap;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
struct proc *curp, *p;
struct pgrp *pg;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
int found = 0, error = 0;
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
curp = td->td_proc;
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
switch (uap->which) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
case PRIO_PROCESS:
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
if (uap->who == 0) {
PROC_LOCK(curp);
error = donice(td, curp, uap->prio);
PROC_UNLOCK(curp);
} else {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
p = pfind(uap->who);
Style fixes.
2008-03-16 18:26:59 +00:00
if (p == NULL)
Change the pfind() and zpfind() functions to lock the process that they find before releasing the allproc lock and returning. Reviewed by: -smp, dfr, jake
2001-04-24 00:51:53 +00:00
break;
Fix information leak. We can find PIDs of running processes from within a jail, etc. by simply calling setpriority(PRIO_PROCESS, <PID>, 0) and checking the return value: 0 means that the process exists and -1 that it doesn't exist. Reviewed by: rwatson MFC after: 1 week
2008-03-16 17:55:06 +00:00
error = p_cansee(td, p);
if (error == 0)
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
error = donice(td, p, uap->prio);
Change the pfind() and zpfind() functions to lock the process that they find before releasing the allproc lock and returning. Reviewed by: -smp, dfr, jake
2001-04-24 00:51:53 +00:00
PROC_UNLOCK(p);
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
found++;
break;
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
case PRIO_PGRP:
Lock proctree_lock instead of pgrpsess_lock.
2002-04-16 17:11:34 +00:00
sx_slock(&proctree_lock);
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
if (uap->who == 0) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
pg = curp->p_pgrp;
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
PGRP_LOCK(pg);
} else {
pg = pgfind(uap->who);
if (pg == NULL) {
Lock proctree_lock instead of pgrpsess_lock.
2002-04-16 17:11:34 +00:00
sx_sunlock(&proctree_lock);
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
break;
}
}
Lock proctree_lock instead of pgrpsess_lock.
2002-04-16 17:11:34 +00:00
sx_sunlock(&proctree_lock);
This is a partial commit of the patch from PR 14914: Alot of the code in sys/kern directly accesses the *Q_HEAD and *Q_ENTRY structures for list operations. This patch makes all list operations in sys/kern use the queue(3) macros, rather than directly accessing the *Q_{HEAD,ENTRY} structures. This batch of changes compile to the same object files. Reviewed by: phk Submitted by: Jake Burkholder <jake@checker.org> PR: 14914
1999-11-16 10:56:05 +00:00
LIST_FOREACH(p, &pg->pg_members, p_pglist) {
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
PROC_LOCK(p);
Fix several places to ignore processes that are not yet fully constructed. MFC after: 1 week
2011-04-06 17:47:22 +00:00
if (p->p_state == PRS_NORMAL &&
p_cansee(td, p) == 0) {
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
error = donice(td, p, uap->prio);
o Modify jail to limit creation of sockets to UNIX domain sockets, TCP/IP (v4) sockets, and routing sockets. Previously, interaction with IPv6 was not well-defined, and might be inappropriate for some environments. Similarly, sysctl MIB entries providing interface information also give out only addresses from those protocol domains. For the time being, this functionality is enabled by default, and toggleable using the sysctl variable jail.socket_unixiproute_only. In the future, protocol domains will be able to determine whether or not they are ``jail aware''. o Further limitations on process use of getpriority() and setpriority() by jailed processes. Addresses problem described in kern/17878. Reviewed by: phk, jmg
2000-06-04 04:28:31 +00:00
found++;
}
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
PROC_UNLOCK(p);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Lock struct pgrp, session and sigio. New locks are: - pgrpsess_lock which locks the whole pgrps and sessions, - pg_mtx which protects the pgrp members, and - s_mtx which protects the session members. Please refer to sys/proc.h for the coverage of these locks. Changes on the pgrp/session interface: - pgfind() needs the pgrpsess_lock held. - The caller of enterpgrp() is responsible to allocate a new pgrp and session. - Call enterthispgrp() in order to enter an existing pgrp. - pgsignal() requires a pgrp lock held. Reviewed by: jhb, alfred Tested on: cvsup.jp.FreeBSD.org (which is a quad-CPU machine running -current)
2002-02-23 11:12:57 +00:00
PGRP_UNLOCK(pg);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
case PRIO_USER:
if (uap->who == 0)
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
uap->who = td->td_ucred->cr_uid;
Convert the allproc and proctree locks from lockmgr locks to sx locks.
2001-03-28 11:52:56 +00:00
sx_slock(&allproc_lock);
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
FOREACH_PROC_IN_SYSTEM(p) {
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
PROC_LOCK(p);
Fix several places to ignore processes that are not yet fully constructed. MFC after: 1 week
2011-04-06 17:47:22 +00:00
if (p->p_state == PRS_NORMAL &&
p->p_ucred->cr_uid == uap->who &&
Style fixes.
2008-03-16 18:26:59 +00:00
p_cansee(td, p) == 0) {
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
error = donice(td, p, uap->prio);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
found++;
}
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
PROC_UNLOCK(p);
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
}
Convert the allproc and proctree locks from lockmgr locks to sx locks.
2001-03-28 11:52:56 +00:00
sx_sunlock(&allproc_lock);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
default:
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
error = EINVAL;
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
if (found == 0 && error == 0)
error = ESRCH;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return (error);
}
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
/*
Nice, is a property of a process as a whole.. I mistakenly moved it to the ksegroup when breaking up the process structure. Put it back in the proc structure.
2004-06-16 00:26:31 +00:00
* Set "nice" for a (whole) process.
Allocate KSEs and KSEGRPs separatly and remove them from the proc structure. next step is to allow > 1 to be allocated per process. This would give multi-processor threads. (when the rest of the infrastructure is in place) While doing this I noticed libkvm and sys/kern/kern_proc.c:fill_kinfo_proc are diverging more than they should.. corrective action needed soon.
2002-09-15 23:52:25 +00:00
*/
Staticize.
1998-02-09 06:11:36 +00:00
static int
Allocate KSEs and KSEGRPs separatly and remove them from the proc structure. next step is to allow > 1 to be allocated per process. This would give multi-processor threads. (when the rest of the infrastructure is in place) While doing this I noticed libkvm and sys/kern/kern_proc.c:fill_kinfo_proc are diverging more than they should.. corrective action needed soon.
2002-09-15 23:52:25 +00:00
donice(struct thread *td, struct proc *p, int n)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
Nice, is a property of a process as a whole.. I mistakenly moved it to the ksegroup when breaking up the process structure. Put it back in the proc structure.
2004-06-16 00:26:31 +00:00
int error;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Allocate KSEs and KSEGRPs separatly and remove them from the proc structure. next step is to allow > 1 to be allocated per process. This would give multi-processor threads. (when the rest of the infrastructure is in place) While doing this I noticed libkvm and sys/kern/kern_proc.c:fill_kinfo_proc are diverging more than they should.. corrective action needed soon.
2002-09-15 23:52:25 +00:00
PROC_LOCK_ASSERT(p, MA_OWNED);
if ((error = p_cansched(td, p)))
o Centralize inter-process access control, introducing: int p_can(p1, p2, operation, privused) which allows specification of subject process, object process, inter-process operation, and an optional call-by-reference privused flag, allowing the caller to determine if privilege was required for the call to succeed. This allows jail, kern.ps_showallprocs and regular credential-based interaction checks to occur in one block of code. Possible operations are P_CAN_SEE, P_CAN_SCHED, P_CAN_KILL, and P_CAN_DEBUG. p_can currently breaks out as a wrapper to a series of static function checks in kern_prot, which should not be invoked directly. o Commented out capabilities entries are included for some checks. o Update most inter-process authorization to make use of p_can() instead of manual checks, PRISON_CHECK(), P_TRESPASS(), and kern.ps_showallprocs. o Modify suser{,_xxx} to use const arguments, as it no longer modifies process flags due to the disabling of ASU. o Modify some checks/errors in procfs so that ENOENT is returned instead of ESRCH, further improving concealment of processes that should not be visible to other processes. Also introduce new access checks to improve hiding of processes for procfs_lookup(), procfs_getattr(), procfs_readdir(). Correct a bug reported by bp concerning not handling the CREATE case in procfs_lookup(). Remove volatile flag in procfs that caused apparently spurious qualifier warnigns (approved by bde). o Add comment noting that ktrace() has not been updated, as its access control checks are different from ptrace(), whereas they should probably be the same. Further discussion should happen on this topic. Reviewed by: bde, green, phk, freebsd-security, others Approved by: bde Obtained from: TrustedBSD Project
2000-08-30 04:49:09 +00:00
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (n > PRIO_MAX)
n = PRIO_MAX;
if (n < PRIO_MIN)
n = PRIO_MIN;
Whitespace cleanups.
2008-03-16 21:32:20 +00:00
if (n < p->p_nice && priv_check(td, PRIV_SCHED_SETPRIORITY) != 0)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return (EACCES);
Nice, is a property of a process as a whole.. I mistakenly moved it to the ksegroup when breaking up the process structure. Put it back in the proc structure.
2004-06-16 00:26:31 +00:00
sched_nice(p, n);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return (0);
}
Replace system call thr_getscheduler, thr_setscheduler, thr_setschedparam with rtprio_thread, while rtprio system call is for process only, the new system call rtprio_thread is responsible for LWP.
2006-09-21 04:18:46 +00:00
/*
* Set realtime priority for LWP.
*/
#ifndef _SYS_SYSPROTO_H_
struct rtprio_thread_args {
int function;
lwpid_t lwpid;
struct rtprio *rtp;
};
#endif
int
In order to maximize the re-usability of kernel code in user space this patch modifies makesyscalls.sh to prefix all of the non-compatibility calls (e.g. not linux_, freebsd32_) with sys_ and updates the kernel entry points and all places in the code that use them. It also fixes an additional name space collision between the kernel function psignal and the libc function of the same name by renaming the kernel psignal kern_psignal(). By introducing this change now we will ease future MFCs that change syscalls. Reviewed by: rwatson Approved by: re (bz)
2011-09-16 13:58:51 +00:00
sys_rtprio_thread(struct thread *td, struct rtprio_thread_args *uap)
Replace system call thr_getscheduler, thr_setscheduler, thr_setschedparam with rtprio_thread, while rtprio system call is for process only, the new system call rtprio_thread is responsible for LWP.
2006-09-21 04:18:46 +00:00
{
struct proc *p;
struct rtprio rtp;
struct thread *td1;
int cierror, error;
/* Perform copyin before acquiring locks if needed. */
if (uap->function == RTP_SET)
cierror = copyin(uap->rtp, &rtp, sizeof(struct rtprio));
else
cierror = 0;
Create a global thread hash table to speed up thread lookup, use rwlock to protect the table. In old code, thread lookup is done with process lock held, to find a thread, kernel has to iterate through process and thread list, this is quite inefficient. With this change, test shows in extreme case performance is dramatically improved. Earlier patch was reviewed by: jhb, julian
2010-10-09 02:50:23 +00:00
if (uap->lwpid == 0 || uap->lwpid == td->td_tid) {
p = td->td_proc;
td1 = td;
PROC_LOCK(p);
} else {
/* Only look up thread in current process */
td1 = tdfind(uap->lwpid, curproc->p_pid);
if (td1 == NULL)
return (ESRCH);
p = td1->td_proc;
}
Replace system call thr_getscheduler, thr_setscheduler, thr_setschedparam with rtprio_thread, while rtprio system call is for process only, the new system call rtprio_thread is responsible for LWP.
2006-09-21 04:18:46 +00:00
switch (uap->function) {
case RTP_LOOKUP:
if ((error = p_cansee(td, p)))
break;
Create a global thread hash table to speed up thread lookup, use rwlock to protect the table. In old code, thread lookup is done with process lock held, to find a thread, kernel has to iterate through process and thread list, this is quite inefficient. With this change, test shows in extreme case performance is dramatically improved. Earlier patch was reviewed by: jhb, julian
2010-10-09 02:50:23 +00:00
pri_to_rtp(td1, &rtp);
Replace system call thr_getscheduler, thr_setscheduler, thr_setschedparam with rtprio_thread, while rtprio system call is for process only, the new system call rtprio_thread is responsible for LWP.
2006-09-21 04:18:46 +00:00
PROC_UNLOCK(p);
return (copyout(&rtp, uap->rtp, sizeof(struct rtprio)));
case RTP_SET:
if ((error = p_cansched(td, p)) || (error = cierror))
break;
/* Disallow setting rtprio in most cases if not superuser. */
/*
* Realtime priority has to be restricted for reasons which should be
* obvious. However, for idle priority, there is a potential for
* system deadlock if an idleprio process gains a lock on a resource
* that other processes need (and the idleprio process can't run
* due to a CPU-bound normal process). Fix me! XXX
*/
#if 0
Whitespace cleanups.
2008-03-16 21:32:20 +00:00
if (RTP_PRIO_IS_REALTIME(rtp.type)) {
Replace system call thr_getscheduler, thr_setscheduler, thr_setschedparam with rtprio_thread, while rtprio system call is for process only, the new system call rtprio_thread is responsible for LWP.
2006-09-21 04:18:46 +00:00
#else
Remove the restriction that rtprio(2) cannot be used to set the realtime or idle priority of another process owned by the same user. This means that privilege in rtprio(2) (and rtprio_thread(2)) is required indirectly via p_cansched(9) or directly to set realtime/idle privilege, rather than directly affecting target process authorization.
2007-06-14 23:31:52 +00:00
if (rtp.type != RTP_PRIO_NORMAL) {
Replace system call thr_getscheduler, thr_setscheduler, thr_setschedparam with rtprio_thread, while rtprio system call is for process only, the new system call rtprio_thread is responsible for LWP.
2006-09-21 04:18:46 +00:00
#endif
Remove the restriction that rtprio(2) cannot be used to set the realtime or idle priority of another process owned by the same user. This means that privilege in rtprio(2) (and rtprio_thread(2)) is required indirectly via p_cansched(9) or directly to set realtime/idle privilege, rather than directly affecting target process authorization.
2007-06-14 23:31:52 +00:00
error = priv_check(td, PRIV_SCHED_RTPRIO);
if (error)
Replace system call thr_getscheduler, thr_setscheduler, thr_setschedparam with rtprio_thread, while rtprio system call is for process only, the new system call rtprio_thread is responsible for LWP.
2006-09-21 04:18:46 +00:00
break;
}
Create a global thread hash table to speed up thread lookup, use rwlock to protect the table. In old code, thread lookup is done with process lock held, to find a thread, kernel has to iterate through process and thread list, this is quite inefficient. With this change, test shows in extreme case performance is dramatically improved. Earlier patch was reviewed by: jhb, julian
2010-10-09 02:50:23 +00:00
error = rtp_to_pri(&rtp, td1);
Replace system call thr_getscheduler, thr_setscheduler, thr_setschedparam with rtprio_thread, while rtprio system call is for process only, the new system call rtprio_thread is responsible for LWP.
2006-09-21 04:18:46 +00:00
break;
default:
error = EINVAL;
break;
}
PROC_UNLOCK(p);
return (error);
}
A few more style fixes from Bruce including a few I missed last time. Submitted by: bde
2004-02-06 19:25:34 +00:00
/*
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
* Set realtime priority.
A few more style fixes from Bruce including a few I missed last time. Submitted by: bde
2004-02-06 19:25:34 +00:00
*/
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#ifndef _SYS_SYSPROTO_H_
Realtime priority scheduling support. Submitted by: Henrik Vestergaard Draboel
1994-09-01 05:12:53 +00:00
struct rtprio_args {
"idle priority" support. Based on code from Henrik Vestergaard Draboel, but substantially rewritten by me.
1994-10-02 04:48:21 +00:00
int function;
pid_t pid;
Fixed types of rtprio(), osetrlimit() and setrlimit(). The args struct tag and/or member names conflicted with the machine generated ones in <sys/sysproto.h>.
1995-11-11 01:48:17 +00:00
struct rtprio *rtp;
Realtime priority scheduling support. Submitted by: Henrik Vestergaard Draboel
1994-09-01 05:12:53 +00:00
};
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#endif
Realtime priority scheduling support. Submitted by: Henrik Vestergaard Draboel
1994-09-01 05:12:53 +00:00
int
In order to maximize the re-usability of kernel code in user space this patch modifies makesyscalls.sh to prefix all of the non-compatibility calls (e.g. not linux_, freebsd32_) with sys_ and updates the kernel entry points and all places in the code that use them. It also fixes an additional name space collision between the kernel function psignal and the libc function of the same name by renaming the kernel psignal kern_psignal(). By introducing this change now we will ease future MFCs that change syscalls. Reviewed by: rwatson Approved by: re (bz)
2011-09-16 13:58:51 +00:00
sys_rtprio(td, uap)
Fix rtprio() to do sensible things when called from threaded processes. It's not quite correct from a posix Point Of view, but it is a lot better than what was there before. This will be revisited later when we decide what form our priority extensions will take. Posix doesn't specify how a system scope thread can change its priority so you need to add non-standard extensions to be able to do it.. For now make this slightly non standard to allow it to be done. Submitted by: Dan Eischen originally, changed by myself.
2004-05-08 08:56:05 +00:00
struct thread *td; /* curthread */
Realtime priority scheduling support. Submitted by: Henrik Vestergaard Draboel
1994-09-01 05:12:53 +00:00
register struct rtprio_args *uap;
{
Fix rtprio() to do sensible things when called from threaded processes. It's not quite correct from a posix Point Of view, but it is a lot better than what was there before. This will be revisited later when we decide what form our priority extensions will take. Posix doesn't specify how a system scope thread can change its priority so you need to add non-standard extensions to be able to do it.. For now make this slightly non standard to allow it to be done. Submitted by: Dan Eischen originally, changed by myself.
2004-05-08 08:56:05 +00:00
struct proc *p;
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
struct thread *tdp;
"idle priority" support. Based on code from Henrik Vestergaard Draboel, but substantially rewritten by me.
1994-10-02 04:48:21 +00:00
struct rtprio rtp;
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
int cierror, error;
"idle priority" support. Based on code from Henrik Vestergaard Draboel, but substantially rewritten by me.
1994-10-02 04:48:21 +00:00
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
/* Perform copyin before acquiring locks if needed. */
if (uap->function == RTP_SET)
cierror = copyin(uap->rtp, &rtp, sizeof(struct rtprio));
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
else
cierror = 0;
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
Change the pfind() and zpfind() functions to lock the process that they find before releasing the allproc lock and returning. Reviewed by: -smp, dfr, jake
2001-04-24 00:51:53 +00:00
if (uap->pid == 0) {
Style fixes.
2008-03-16 18:26:59 +00:00
p = td->td_proc;
Change the pfind() and zpfind() functions to lock the process that they find before releasing the allproc lock and returning. Reviewed by: -smp, dfr, jake
2001-04-24 00:51:53 +00:00
PROC_LOCK(p);
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
} else {
"idle priority" support. Based on code from Henrik Vestergaard Draboel, but substantially rewritten by me.
1994-10-02 04:48:21 +00:00
p = pfind(uap->pid);
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
if (p == NULL)
return (ESRCH);
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
}
Realtime priority scheduling support. Submitted by: Henrik Vestergaard Draboel
1994-09-01 05:12:53 +00:00
"idle priority" support. Based on code from Henrik Vestergaard Draboel, but substantially rewritten by me.
1994-10-02 04:48:21 +00:00
switch (uap->function) {
case RTP_LOOKUP:
Change p_can{debug,see,sched,signal}()'s first argument to be a thread pointer instead of a proc pointer and require the process pointed to by the second argument to be locked. We now use the thread ucred reference for the credential checks in p_can*() as a result. p_canfoo() should now no longer need Giant.
2002-05-19 00:14:50 +00:00
if ((error = p_cansee(td, p)))
Make rtprio work again. - add a missing break which caused RTP_SET to always return EINVAL - break instead of returning if p_can fails so proc_lock is always dropped correctly - only copyin data that is actually needed - use break instead of goto - make rtp_to_pri return EINVAL instead of -1 if the values are out or range so we don't have to translate
2001-04-29 22:09:26 +00:00
break;
Fix rtprio() to do sensible things when called from threaded processes. It's not quite correct from a posix Point Of view, but it is a lot better than what was there before. This will be revisited later when we decide what form our priority extensions will take. Posix doesn't specify how a system scope thread can change its priority so you need to add non-standard extensions to be able to do it.. For now make this slightly non standard to allow it to be done. Submitted by: Dan Eischen originally, changed by myself.
2004-05-08 08:56:05 +00:00
/*
* Return OUR priority if no pid specified,
* or if one is, report the highest priority
Whitespace cleanups.
2008-03-16 21:32:20 +00:00
* in the process. There isn't much more you can do as
Fix rtprio() to do sensible things when called from threaded processes. It's not quite correct from a posix Point Of view, but it is a lot better than what was there before. This will be revisited later when we decide what form our priority extensions will take. Posix doesn't specify how a system scope thread can change its priority so you need to add non-standard extensions to be able to do it.. For now make this slightly non standard to allow it to be done. Submitted by: Dan Eischen originally, changed by myself.
2004-05-08 08:56:05 +00:00
* there is only room to return a single priority.
* Note: specifying our own pid is not the same
* as leaving it zero.
*/
if (uap->pid == 0) {
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
pri_to_rtp(td, &rtp);
Fix rtprio() to do sensible things when called from threaded processes. It's not quite correct from a posix Point Of view, but it is a lot better than what was there before. This will be revisited later when we decide what form our priority extensions will take. Posix doesn't specify how a system scope thread can change its priority so you need to add non-standard extensions to be able to do it.. For now make this slightly non standard to allow it to be done. Submitted by: Dan Eischen originally, changed by myself.
2004-05-08 08:56:05 +00:00
} else {
struct rtprio rtp2;
rtp.type = RTP_PRIO_IDLE;
rtp.prio = RTP_PRIO_MAX;
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
FOREACH_THREAD_IN_PROC(p, tdp) {
pri_to_rtp(tdp, &rtp2);
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
if (rtp2.type < rtp.type ||
(rtp2.type == rtp.type &&
rtp2.prio < rtp.prio)) {
Fix rtprio() to do sensible things when called from threaded processes. It's not quite correct from a posix Point Of view, but it is a lot better than what was there before. This will be revisited later when we decide what form our priority extensions will take. Posix doesn't specify how a system scope thread can change its priority so you need to add non-standard extensions to be able to do it.. For now make this slightly non standard to allow it to be done. Submitted by: Dan Eischen originally, changed by myself.
2004-05-08 08:56:05 +00:00
rtp.type = rtp2.type;
rtp.prio = rtp2.prio;
}
}
}
- Change donice() to take a thread as the first argument instead of a process so it can use td_ucred. - Require the target process of donice() to be locked when donice() is called. - Use td_ucred. - Lock the target process of p_cansee() and while reading the credentials of a process. - Change the logic of rtprio() slightly so it does it's copyin() if needed prior to locking the target process. - rtprio() no longer needs Giant. In theory with full KSE it would still need Giant to protect p_ucred of curproc for the p_canfoo() functions but p_canfoo() will be changing to using td_ucred of curthread before full KSE hits the tree.
2002-04-13 23:28:23 +00:00
PROC_UNLOCK(p);
return (copyout(&rtp, uap->rtp, sizeof(struct rtprio)));
"idle priority" support. Based on code from Henrik Vestergaard Draboel, but substantially rewritten by me.
1994-10-02 04:48:21 +00:00
case RTP_SET:
Change p_can{debug,see,sched,signal}()'s first argument to be a thread pointer instead of a proc pointer and require the process pointed to by the second argument to be locked. We now use the thread ucred reference for the credential checks in p_can*() as a result. p_canfoo() should now no longer need Giant.
2002-05-19 00:14:50 +00:00
if ((error = p_cansched(td, p)) || (error = cierror))
Make rtprio work again. - add a missing break which caused RTP_SET to always return EINVAL - break instead of returning if p_can fails so proc_lock is always dropped correctly - only copyin data that is actually needed - use break instead of goto - make rtp_to_pri return EINVAL instead of -1 if the values are out or range so we don't have to translate
2001-04-29 22:09:26 +00:00
break;
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
/* Disallow setting rtprio in most cases if not superuser. */
Restrict idleprio to superuser: Realtime priority has to be restricted for reasons which should be obvious. However, for idle priority, there is a potential for system deadlock if an idleprio process gains a lock on a resource that other processes need (and the idleprio process can't run due to a CPU-bound normal process). Fix me! XXX PR: 5639
1998-02-04 18:43:10 +00:00
/*
* Realtime priority has to be restricted for reasons which should be
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
* obvious. However, for idle priority, there is a potential for
Restrict idleprio to superuser: Realtime priority has to be restricted for reasons which should be obvious. However, for idle priority, there is a potential for system deadlock if an idleprio process gains a lock on a resource that other processes need (and the idleprio process can't run due to a CPU-bound normal process). Fix me! XXX PR: 5639
1998-02-04 18:43:10 +00:00
* system deadlock if an idleprio process gains a lock on a resource
* that other processes need (and the idleprio process can't run
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
* due to a CPU-bound normal process). Fix me! XXX
Restrict idleprio to superuser: Realtime priority has to be restricted for reasons which should be obvious. However, for idle priority, there is a potential for system deadlock if an idleprio process gains a lock on a resource that other processes need (and the idleprio process can't run due to a CPU-bound normal process). Fix me! XXX PR: 5639
1998-02-04 18:43:10 +00:00
*/
#if 0
Remove the restriction that rtprio(2) cannot be used to set the realtime or idle priority of another process owned by the same user. This means that privilege in rtprio(2) (and rtprio_thread(2)) is required indirectly via p_cansched(9) or directly to set realtime/idle privilege, rather than directly affecting target process authorization.
2007-06-14 23:31:52 +00:00
if (RTP_PRIO_IS_REALTIME(rtp.type)) {
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
#else
Remove the restriction that rtprio(2) cannot be used to set the realtime or idle priority of another process owned by the same user. This means that privilege in rtprio(2) (and rtprio_thread(2)) is required indirectly via p_cansched(9) or directly to set realtime/idle privilege, rather than directly affecting target process authorization.
2007-06-14 23:31:52 +00:00
if (rtp.type != RTP_PRIO_NORMAL) {
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
#endif
Remove the restriction that rtprio(2) cannot be used to set the realtime or idle priority of another process owned by the same user. This means that privilege in rtprio(2) (and rtprio_thread(2)) is required indirectly via p_cansched(9) or directly to set realtime/idle privilege, rather than directly affecting target process authorization.
2007-06-14 23:31:52 +00:00
error = priv_check(td, PRIV_SCHED_RTPRIO);
if (error)
Make rtprio work again. - add a missing break which caused RTP_SET to always return EINVAL - break instead of returning if p_can fails so proc_lock is always dropped correctly - only copyin data that is actually needed - use break instead of goto - make rtp_to_pri return EINVAL instead of -1 if the values are out or range so we don't have to translate
2001-04-29 22:09:26 +00:00
break;
"idle priority" support. Based on code from Henrik Vestergaard Draboel, but substantially rewritten by me.
1994-10-02 04:48:21 +00:00
}
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
/*
* If we are setting our own priority, set just our
* thread but if we are doing another process,
* do all the threads on that process. If we
* specify our own pid we do the latter.
*/
Fix rtprio() to do sensible things when called from threaded processes. It's not quite correct from a posix Point Of view, but it is a lot better than what was there before. This will be revisited later when we decide what form our priority extensions will take. Posix doesn't specify how a system scope thread can change its priority so you need to add non-standard extensions to be able to do it.. For now make this slightly non standard to allow it to be done. Submitted by: Dan Eischen originally, changed by myself.
2004-05-08 08:56:05 +00:00
if (uap->pid == 0) {
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
error = rtp_to_pri(&rtp, td);
Fix rtprio() to do sensible things when called from threaded processes. It's not quite correct from a posix Point Of view, but it is a lot better than what was there before. This will be revisited later when we decide what form our priority extensions will take. Posix doesn't specify how a system scope thread can change its priority so you need to add non-standard extensions to be able to do it.. For now make this slightly non standard to allow it to be done. Submitted by: Dan Eischen originally, changed by myself.
2004-05-08 08:56:05 +00:00
} else {
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
FOREACH_THREAD_IN_PROC(p, td) {
if ((error = rtp_to_pri(&rtp, td)) != 0)
break;
Fix rtprio() to do sensible things when called from threaded processes. It's not quite correct from a posix Point Of view, but it is a lot better than what was there before. This will be revisited later when we decide what form our priority extensions will take. Posix doesn't specify how a system scope thread can change its priority so you need to add non-standard extensions to be able to do it.. For now make this slightly non standard to allow it to be done. Submitted by: Dan Eischen originally, changed by myself.
2004-05-08 08:56:05 +00:00
}
}
Make rtprio work again. - add a missing break which caused RTP_SET to always return EINVAL - break instead of returning if p_can fails so proc_lock is always dropped correctly - only copyin data that is actually needed - use break instead of goto - make rtp_to_pri return EINVAL instead of -1 if the values are out or range so we don't have to translate
2001-04-29 22:09:26 +00:00
break;
"idle priority" support. Based on code from Henrik Vestergaard Draboel, but substantially rewritten by me.
1994-10-02 04:48:21 +00:00
default:
Change the pfind() and zpfind() functions to lock the process that they find before releasing the allproc lock and returning. Reviewed by: -smp, dfr, jake
2001-04-24 00:51:53 +00:00
error = EINVAL;
Make rtprio work again. - add a missing break which caused RTP_SET to always return EINVAL - break instead of returning if p_can fails so proc_lock is always dropped correctly - only copyin data that is actually needed - use break instead of goto - make rtp_to_pri return EINVAL instead of -1 if the values are out or range so we don't have to translate
2001-04-29 22:09:26 +00:00
break;
"idle priority" support. Based on code from Henrik Vestergaard Draboel, but substantially rewritten by me.
1994-10-02 04:48:21 +00:00
}
Change the pfind() and zpfind() functions to lock the process that they find before releasing the allproc lock and returning. Reviewed by: -smp, dfr, jake
2001-04-24 00:51:53 +00:00
PROC_UNLOCK(p);
return (error);
"idle priority" support. Based on code from Henrik Vestergaard Draboel, but substantially rewritten by me.
1994-10-02 04:48:21 +00:00
}
Realtime priority scheduling support. Submitted by: Henrik Vestergaard Draboel
1994-09-01 05:12:53 +00:00
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
int
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
rtp_to_pri(struct rtprio *rtp, struct thread *td)
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
{
Use scheduler API sched_user_prio() to adjust thread's userland priority, use td_base_user_prio to get real userland priority since POSIX priority mutex may adjust td_user_pri which is an effective priority.
2006-11-20 05:50:59 +00:00
u_char newpri;
Fix LOR of thread lock and umtx's priority propagation mutex due to the reworking of scheduler lock. MFC: after 3 days
2007-12-11 08:25:36 +00:00
u_char oldpri;
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
switch (RTP_PRIO_BASE(rtp->type)) {
case RTP_PRIO_REALTIME:
Add back a bounds check on valid idle priorities that was lost in an earlier commit. While here, move the thread lock down in rtp_to_pri(). It is not needed for all of the priority value checks and the computation of newpri. Reported by: swell.k @ gmail MFC after: 3 days
2010-12-17 16:29:06 +00:00
if (rtp->prio > RTP_PRIO_MAX)
sched_getparam was just plain broke for time-share processes. It did not return an error but instead just let garbage be passed back. This I fix so it actually properly translates the priority the process is at to a posix's high means more priority. I also fix it so that if the ULE scheduler has bumped it up to a realtime process you get back a sane value i.e. the highest priority (63 for time-share). sched_setscheduler() had the setting of the timeshare class priority disabled. With some notes about rejecting the posix high numbers is greater priority and use nice instead. This fix also adjusts that to work, with the cavet that a t-s process may well get bumped up or down i.e. the setscheduler() will NOT change the nice value only the current priority. I think this is reasonable considering if the user wants to play with nice then he can. At least all the posix'ish interfaces now respond sanely. MFC after: 3 weeks
2010-03-03 21:46:51 +00:00
return (EINVAL);
Use scheduler API sched_user_prio() to adjust thread's userland priority, use td_base_user_prio to get real userland priority since POSIX priority mutex may adjust td_user_pri which is an effective priority.
2006-11-20 05:50:59 +00:00
newpri = PRI_MIN_REALTIME + rtp->prio;
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
break;
case RTP_PRIO_NORMAL:
Add back a bounds check on valid idle priorities that was lost in an earlier commit. While here, move the thread lock down in rtp_to_pri(). It is not needed for all of the priority value checks and the computation of newpri. Reported by: swell.k @ gmail MFC after: 3 days
2010-12-17 16:29:06 +00:00
if (rtp->prio > (PRI_MAX_TIMESHARE - PRI_MIN_TIMESHARE))
sched_getparam was just plain broke for time-share processes. It did not return an error but instead just let garbage be passed back. This I fix so it actually properly translates the priority the process is at to a posix's high means more priority. I also fix it so that if the ULE scheduler has bumped it up to a realtime process you get back a sane value i.e. the highest priority (63 for time-share). sched_setscheduler() had the setting of the timeshare class priority disabled. With some notes about rejecting the posix high numbers is greater priority and use nice instead. This fix also adjusts that to work, with the cavet that a t-s process may well get bumped up or down i.e. the setscheduler() will NOT change the nice value only the current priority. I think this is reasonable considering if the user wants to play with nice then he can. At least all the posix'ish interfaces now respond sanely. MFC after: 3 weeks
2010-03-03 21:46:51 +00:00
return (EINVAL);
Use scheduler API sched_user_prio() to adjust thread's userland priority, use td_base_user_prio to get real userland priority since POSIX priority mutex may adjust td_user_pri which is an effective priority.
2006-11-20 05:50:59 +00:00
newpri = PRI_MIN_TIMESHARE + rtp->prio;
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
break;
case RTP_PRIO_IDLE:
Add back a bounds check on valid idle priorities that was lost in an earlier commit. While here, move the thread lock down in rtp_to_pri(). It is not needed for all of the priority value checks and the computation of newpri. Reported by: swell.k @ gmail MFC after: 3 days
2010-12-17 16:29:06 +00:00
if (rtp->prio > RTP_PRIO_MAX)
return (EINVAL);
Use scheduler API sched_user_prio() to adjust thread's userland priority, use td_base_user_prio to get real userland priority since POSIX priority mutex may adjust td_user_pri which is an effective priority.
2006-11-20 05:50:59 +00:00
newpri = PRI_MIN_IDLE + rtp->prio;
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
break;
default:
Make rtprio work again. - add a missing break which caused RTP_SET to always return EINVAL - break instead of returning if p_can fails so proc_lock is always dropped correctly - only copyin data that is actually needed - use break instead of goto - make rtp_to_pri return EINVAL instead of -1 if the values are out or range so we don't have to translate
2001-04-29 22:09:26 +00:00
return (EINVAL);
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
}
Add back a bounds check on valid idle priorities that was lost in an earlier commit. While here, move the thread lock down in rtp_to_pri(). It is not needed for all of the priority value checks and the computation of newpri. Reported by: swell.k @ gmail MFC after: 3 days
2010-12-17 16:29:06 +00:00
thread_lock(td);
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
sched_class(td, rtp->type); /* XXX fix */
Fix LOR of thread lock and umtx's priority propagation mutex due to the reworking of scheduler lock. MFC: after 3 days
2007-12-11 08:25:36 +00:00
oldpri = td->td_user_pri;
Use scheduler API sched_user_prio() to adjust thread's userland priority, use td_base_user_prio to get real userland priority since POSIX priority mutex may adjust td_user_pri which is an effective priority.
2006-11-20 05:50:59 +00:00
sched_user_prio(td, newpri);
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
if (curthread == td)
sched_prio(curthread, td->td_user_pri); /* XXX dubious */
Fix LOR of thread lock and umtx's priority propagation mutex due to the reworking of scheduler lock. MFC: after 3 days
2007-12-11 08:25:36 +00:00
if (TD_ON_UPILOCK(td) && oldpri != newpri) {
- Follow r216313, the sched_unlend_user_prio is no longer needed, always use sched_lend_user_prio to set lent priority. - Improve pthread priority-inherit mutex, when a contender's priority is lowered, repropagete priorities, this may cause mutex owner's priority to be lowerd, in old code, mutex owner's priority is rise-only.
2010-12-29 09:26:46 +00:00
critical_enter();
Fix LOR of thread lock and umtx's priority propagation mutex due to the reworking of scheduler lock. MFC: after 3 days
2007-12-11 08:25:36 +00:00
thread_unlock(td);
umtx_pi_adjust(td, oldpri);
- Follow r216313, the sched_unlend_user_prio is no longer needed, always use sched_lend_user_prio to set lent priority. - Improve pthread priority-inherit mutex, when a contender's priority is lowered, repropagete priorities, this may cause mutex owner's priority to be lowerd, in old code, mutex owner's priority is rise-only.
2010-12-29 09:26:46 +00:00
critical_exit();
Fix LOR of thread lock and umtx's priority propagation mutex due to the reworking of scheduler lock. MFC: after 3 days
2007-12-11 08:25:36 +00:00
} else
thread_unlock(td);
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
return (0);
}
void
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
pri_to_rtp(struct thread *td, struct rtprio *rtp)
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
{
Commit 14/14 of sched_lock decomposition. - Use thread_lock() rather than sched_lock for per-thread scheduling sychronization. - Use the per-process spinlock rather than the sched_lock for per-process scheduling synchronization. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-05 00:00:57 +00:00
thread_lock(td);
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
switch (PRI_BASE(td->td_pri_class)) {
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
case PRI_REALTIME:
Use scheduler API sched_user_prio() to adjust thread's userland priority, use td_base_user_prio to get real userland priority since POSIX priority mutex may adjust td_user_pri which is an effective priority.
2006-11-20 05:50:59 +00:00
rtp->prio = td->td_base_user_pri - PRI_MIN_REALTIME;
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
break;
case PRI_TIMESHARE:
Use scheduler API sched_user_prio() to adjust thread's userland priority, use td_base_user_prio to get real userland priority since POSIX priority mutex may adjust td_user_pri which is an effective priority.
2006-11-20 05:50:59 +00:00
rtp->prio = td->td_base_user_pri - PRI_MIN_TIMESHARE;
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
break;
case PRI_IDLE:
Use scheduler API sched_user_prio() to adjust thread's userland priority, use td_base_user_prio to get real userland priority since POSIX priority mutex may adjust td_user_pri which is an effective priority.
2006-11-20 05:50:59 +00:00
rtp->prio = td->td_base_user_pri - PRI_MIN_IDLE;
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
break;
default:
break;
}
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
rtp->type = td->td_pri_class;
Commit 14/14 of sched_lock decomposition. - Use thread_lock() rather than sched_lock for per-thread scheduling sychronization. - Use the per-process spinlock rather than the sched_lock for per-process scheduling synchronization. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-05 00:00:57 +00:00
thread_unlock(td);
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
}
Deorbit COMPAT_SUNOS. We inherited this from the sparc32 port of BSD4.4-Lite1. We have neither a sparc32 port nor a SunOS4.x compatibility desire these days.
2004-06-11 11:16:26 +00:00
#if defined(COMPAT_43)
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#ifndef _SYS_SYSPROTO_H_
Fixed types of rtprio(), osetrlimit() and setrlimit(). The args struct tag and/or member names conflicted with the machine generated ones in <sys/sysproto.h>.
1995-11-11 01:48:17 +00:00
struct osetrlimit_args {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
u_int which;
Fixed types of rtprio(), osetrlimit() and setrlimit(). The args struct tag and/or member names conflicted with the machine generated ones in <sys/sysproto.h>.
1995-11-11 01:48:17 +00:00
struct orlimit *rlp;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
};
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#endif
The big 4.4BSD Lite to FreeBSD 2.0.0 (Development) patch. Reviewed by: Rodney W. Grimes Submitted by: John Dyson and David Greenman
1994-05-25 09:21:21 +00:00
int
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
osetrlimit(td, uap)
struct thread *td;
Fixed types of rtprio(), osetrlimit() and setrlimit(). The args struct tag and/or member names conflicted with the machine generated ones in <sys/sysproto.h>.
1995-11-11 01:48:17 +00:00
register struct osetrlimit_args *uap;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
struct orlimit olim;
struct rlimit lim;
int error;
more caddr_t removal.
2002-06-29 02:00:02 +00:00
if ((error = copyin(uap->rlp, &olim, sizeof(struct orlimit))))
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return (error);
lim.rlim_cur = olim.rlim_cur;
lim.rlim_max = olim.rlim_max;
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
error = kern_setrlimit(td, uap->which, &lim);
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#ifndef _SYS_SYSPROTO_H_
Fixed types of rtprio(), osetrlimit() and setrlimit(). The args struct tag and/or member names conflicted with the machine generated ones in <sys/sysproto.h>.
1995-11-11 01:48:17 +00:00
struct ogetrlimit_args {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
u_int which;
struct orlimit *rlp;
};
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#endif
The big 4.4BSD Lite to FreeBSD 2.0.0 (Development) patch. Reviewed by: Rodney W. Grimes Submitted by: John Dyson and David Greenman
1994-05-25 09:21:21 +00:00
int
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
ogetrlimit(td, uap)
struct thread *td;
Fixed types of rtprio(), osetrlimit() and setrlimit(). The args struct tag and/or member names conflicted with the machine generated ones in <sys/sysproto.h>.
1995-11-11 01:48:17 +00:00
register struct ogetrlimit_args *uap;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
struct orlimit olim;
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
struct rlimit rl;
struct proc *p;
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
int error;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (uap->which >= RLIM_NLIMITS)
return (EINVAL);
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
p = td->td_proc;
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
PROC_LOCK(p);
lim_rlimit(p, uap->which, &rl);
PROC_UNLOCK(p);
- Correct the translation of old rlimit values to properly handle the old RLIM_INFINITY case for ogetrlimit(). - Use %jd and intmax_t to output negative time in usec in calcru(). - Rework getrusage() to make a copy of the rusage struct into a local variable while holding Giant and then do the copyout from the local variable to avoid having to have the original process rusage struct locked while doing the copyout (which would not be safe). This also includes a few style fixes from Bruce to getrusage(). Submitted by: bde (1, parts of 3) Suggested by: bde (2)
2004-02-06 19:30:12 +00:00
/*
* XXX would be more correct to convert only RLIM_INFINITY to the
* old RLIM_INFINITY and fail with EOVERFLOW for other larger
* values. Most 64->32 and 32->16 conversions, including not
* unimportant ones of uids are even more broken than what we
* do here (they blindly truncate). We don't do this correctly
* here since we have little experience with EOVERFLOW yet.
* Elsewhere, getuid() can't fail...
*/
olim.rlim_cur = rl.rlim_cur > 0x7fffffff ? 0x7fffffff : rl.rlim_cur;
olim.rlim_max = rl.rlim_max > 0x7fffffff ? 0x7fffffff : rl.rlim_max;
more caddr_t removal.
2002-06-29 02:00:02 +00:00
error = copyout(&olim, uap->rlp, sizeof(olim));
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Deorbit COMPAT_SUNOS. We inherited this from the sparc32 port of BSD4.4-Lite1. We have neither a sparc32 port nor a SunOS4.x compatibility desire these days.
2004-06-11 11:16:26 +00:00
#endif /* COMPAT_43 */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#ifndef _SYS_SYSPROTO_H_
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct __setrlimit_args {
u_int which;
Fixed types of rtprio(), osetrlimit() and setrlimit(). The args struct tag and/or member names conflicted with the machine generated ones in <sys/sysproto.h>.
1995-11-11 01:48:17 +00:00
struct rlimit *rlp;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
};
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#endif
The big 4.4BSD Lite to FreeBSD 2.0.0 (Development) patch. Reviewed by: Rodney W. Grimes Submitted by: John Dyson and David Greenman
1994-05-25 09:21:21 +00:00
int
In order to maximize the re-usability of kernel code in user space this patch modifies makesyscalls.sh to prefix all of the non-compatibility calls (e.g. not linux_, freebsd32_) with sys_ and updates the kernel entry points and all places in the code that use them. It also fixes an additional name space collision between the kernel function psignal and the libc function of the same name by renaming the kernel psignal kern_psignal(). By introducing this change now we will ease future MFCs that change syscalls. Reviewed by: rwatson Approved by: re (bz)
2011-09-16 13:58:51 +00:00
sys_setrlimit(td, uap)
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
struct thread *td;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
register struct __setrlimit_args *uap;
{
struct rlimit alim;
int error;
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
if ((error = copyin(uap->rlp, &alim, sizeof(struct rlimit))))
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return (error);
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
error = kern_setrlimit(td, uap->which, &alim);
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
static void
lim_cb(void *arg)
{
struct rlimit rlim;
struct thread *td;
struct proc *p;
p = arg;
PROC_LOCK_ASSERT(p, MA_OWNED);
/*
* Check if the process exceeds its cpu resource allocation. If
* it reaches the max, arrange to kill the process in ast().
*/
if (p->p_cpulimit == RLIM_INFINITY)
return;
Commit 14/14 of sched_lock decomposition. - Use thread_lock() rather than sched_lock for per-thread scheduling sychronization. - Use the per-process spinlock rather than the sched_lock for per-process scheduling synchronization. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-05 00:00:57 +00:00
PROC_SLOCK(p);
FOREACH_THREAD_IN_PROC(p, td) {
Implement RUSAGE_THREAD. Add td_rux to keep extended runtime and ticks information for thread to allow calcru1() (re)use. Rename ruxagg()->ruxagg_locked(), ruxagg_tlock()->ruxagg() [1]. The ruxagg_locked() function no longer clears thread ticks nor td_incruntime. Requested by: attilio [1] Discussed with: attilio, bde Reviewed by: bde Based on submission by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week X-MFC-Note: td_rux shall be moved to the end of struct thread
2010-05-04 05:55:37 +00:00
ruxagg(p, td);
Commit 14/14 of sched_lock decomposition. - Use thread_lock() rather than sched_lock for per-thread scheduling sychronization. - Use the per-process spinlock rather than the sched_lock for per-process scheduling synchronization. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-05 00:00:57 +00:00
}
PROC_SUNLOCK(p);
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
if (p->p_rux.rux_runtime > p->p_cpulimit * cpu_tickrate()) {
lim_rlimit(p, RLIMIT_CPU, &rlim);
if (p->p_rux.rux_runtime >= rlim.rlim_max * cpu_tickrate()) {
killproc(p, "exceeded maximum CPU limit");
} else {
if (p->p_cpulimit < rlim.rlim_max)
p->p_cpulimit += 5;
In order to maximize the re-usability of kernel code in user space this patch modifies makesyscalls.sh to prefix all of the non-compatibility calls (e.g. not linux_, freebsd32_) with sys_ and updates the kernel entry points and all places in the code that use them. It also fixes an additional name space collision between the kernel function psignal and the libc function of the same name by renaming the kernel psignal kern_psignal(). By introducing this change now we will ease future MFCs that change syscalls. Reviewed by: rwatson Approved by: re (bz)
2011-09-16 13:58:51 +00:00
kern_psignal(p, SIGXCPU);
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
}
}
Don't rearm callout if the process is exiting, it may leak a callout because callout_drain() only waits for running callout, but not disable it if it is rearmed.
2008-10-24 01:09:24 +00:00
if ((p->p_flag & P_WEXIT) == 0)
callout_reset(&p->p_limco, hz, lim_cb, p);
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
}
Unstaticized routines which are needed by the svr4 KLD and the streams garbage needed to support SysVR4 networking.
1999-01-30 06:25:00 +00:00
int
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
kern_setrlimit(td, which, limp)
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
struct thread *td;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
u_int which;
struct rlimit *limp;
{
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
struct plimit *newlim, *oldlim;
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
struct proc *p;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
register struct rlimit *alimp;
Fix a couple of issues with the stack limit for 32-bit processes on 64-bit kernels exposed by the recent fixes to resource limits for 32-bit processes on 64-bit kernels: - Let ABIs expose their maximum stack size via a new pointer in sysentvec and use that in preference to maxssiz during exec() rather than always using maxssiz for all processses. - Apply the ABI's limit fixup to the previous stack size when adjusting RLIMIT_STACK to determine if the existing mapping for the stack needs to be grown or shrunk (as well as how much it should be grown or shrunk). Approved by: re (kensmith)
2007-07-12 18:01:31 +00:00
struct rlimit oldssiz;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
int error;
if (which >= RLIM_NLIMITS)
return (EINVAL);
Don't allow negative limits at all. Convert them to RLIM_INFINITY instead of returning EINVAL since something may depend on them being broken. Allowing negative limits caused bugs almost everywhere. The recent fixes for MAXSSIZ checked the limits too late to stop anyone defeating limits set by root...
1994-12-06 22:53:37 +00:00
/*
* Preserve historical bugs by treating negative limits as unsigned.
*/
if (limp->rlim_cur < 0)
limp->rlim_cur = RLIM_INFINITY;
if (limp->rlim_max < 0)
limp->rlim_max = RLIM_INFINITY;
Fix a couple of issues with the stack limit for 32-bit processes on 64-bit kernels exposed by the recent fixes to resource limits for 32-bit processes on 64-bit kernels: - Let ABIs expose their maximum stack size via a new pointer in sysentvec and use that in preference to maxssiz during exec() rather than always using maxssiz for all processses. - Apply the ABI's limit fixup to the previous stack size when adjusting RLIMIT_STACK to determine if the existing mapping for the stack needs to be grown or shrunk (as well as how much it should be grown or shrunk). Approved by: re (kensmith)
2007-07-12 18:01:31 +00:00
oldssiz.rlim_cur = 0;
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
p = td->td_proc;
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
newlim = lim_alloc();
PROC_LOCK(p);
oldlim = p->p_limit;
alimp = &oldlim->pl_rlimit[which];
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
if (limp->rlim_cur > alimp->rlim_max ||
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
limp->rlim_max > alimp->rlim_max)
Eliminate now-unused SUSER_ALLOWJAIL arguments to priv_check_cred(); in some cases, move to priv_check() if it was an operation on a thread and no other flags were present. Eliminate caller-side jail exception checking (also now-unused); jail privilege exception code now goes solely in kern_jail.c. We can't yet eliminate suser() due to some cases in the KAME code where a privilege check is performed and then used in many different deferred paths. Do, however, move those prototypes to priv.h. Reviewed by: csjp Obtained from: TrustedBSD Project
2007-06-12 00:12:01 +00:00
if ((error = priv_check(td, PRIV_PROC_SETRLIMIT))) {
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
PROC_UNLOCK(p);
lim_free(newlim);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return (error);
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (limp->rlim_cur > limp->rlim_max)
limp->rlim_cur = limp->rlim_max;
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
lim_copy(newlim, oldlim);
alimp = &newlim->pl_rlimit[which];
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
switch (which) {
Some cleanups related to timecounters and weird ifdefs in <sys/time.h>. Clean up (or if antipodic: down) some of the msgbuf stuff. Use an inline function rather than a macro for timecounter delta. Maintain process "on-cpu" time as 64 bits of microseconds to avoid needless second rollover overhead. Avoid calling microuptime the second time in mi_switch() if we do not pass through _idle in cpu_switch() This should reduce our context-switch overhead a bit, in particular on pre-P5 and SMP systems. WARNING: Programs which muck about with struct proc in userland will have to be fixed. Reviewed, but found imperfect by: bde
1998-05-28 09:30:28 +00:00
case RLIMIT_CPU:
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
if (limp->rlim_cur != RLIM_INFINITY &&
p->p_cpulimit == RLIM_INFINITY)
callout_reset(&p->p_limco, hz, lim_cb, p);
- Move p_cpulimit to struct proc from struct plimit and protect it with sched_lock. This means that we no longer access p_limit in mi_switch() and the p_limit pointer can be protected by the proc lock. - Remove PRS_ZOMBIE check from CPU limit test in mi_switch(). PRS_ZOMBIE processes don't call mi_switch(), and even if they did there is no longer the danger of p_limit being NULL (which is what the original zombie check was added for). - When we bump the current processes soft CPU limit in ast(), just bump the private p_cpulimit instead of the shared rlimit. This fixes an XXX for some value of fix. There is still a (probably benign) bug in that this code doesn't check that the new soft limit exceeds the hard limit. Inspired by: bde (2)
2002-10-09 17:17:24 +00:00
p->p_cpulimit = limp->rlim_cur;
Some cleanups related to timecounters and weird ifdefs in <sys/time.h>. Clean up (or if antipodic: down) some of the msgbuf stuff. Use an inline function rather than a macro for timecounter delta. Maintain process "on-cpu" time as 64 bits of microseconds to avoid needless second rollover overhead. Avoid calling microuptime the second time in mi_switch() if we do not pass through _idle in cpu_switch() This should reduce our context-switch overhead a bit, in particular on pre-P5 and SMP systems. WARNING: Programs which muck about with struct proc in userland will have to be fixed. Reviewed, but found imperfect by: bde
1998-05-28 09:30:28 +00:00
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
case RLIMIT_DATA:
Make MAXTSIZ, DFLDSIZ, MAXDSIZ, DFLSSIZ, MAXSSIZ, SGROWSIZ loader tunable. Reviewed by: peter MFC after: 2 weeks
2001-10-10 23:06:54 +00:00
if (limp->rlim_cur > maxdsiz)
limp->rlim_cur = maxdsiz;
if (limp->rlim_max > maxdsiz)
limp->rlim_max = maxdsiz;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
case RLIMIT_STACK:
Make MAXTSIZ, DFLDSIZ, MAXDSIZ, DFLSSIZ, MAXSSIZ, SGROWSIZ loader tunable. Reviewed by: peter MFC after: 2 weeks
2001-10-10 23:06:54 +00:00
if (limp->rlim_cur > maxssiz)
limp->rlim_cur = maxssiz;
if (limp->rlim_max > maxssiz)
limp->rlim_max = maxssiz;
Fix a couple of issues with the stack limit for 32-bit processes on 64-bit kernels exposed by the recent fixes to resource limits for 32-bit processes on 64-bit kernels: - Let ABIs expose their maximum stack size via a new pointer in sysentvec and use that in preference to maxssiz during exec() rather than always using maxssiz for all processses. - Apply the ABI's limit fixup to the previous stack size when adjusting RLIMIT_STACK to determine if the existing mapping for the stack needs to be grown or shrunk (as well as how much it should be grown or shrunk). Approved by: re (kensmith)
2007-07-12 18:01:31 +00:00
oldssiz = *alimp;
We've already set p = td->td_proc, so use it.
2010-10-18 15:46:58 +00:00
if (p->p_sysent->sv_fixlimit != NULL)
p->p_sysent->sv_fixlimit(&oldssiz,
Fix a couple of issues with the stack limit for 32-bit processes on 64-bit kernels exposed by the recent fixes to resource limits for 32-bit processes on 64-bit kernels: - Let ABIs expose their maximum stack size via a new pointer in sysentvec and use that in preference to maxssiz during exec() rather than always using maxssiz for all processses. - Apply the ABI's limit fixup to the previous stack size when adjusting RLIMIT_STACK to determine if the existing mapping for the stack needs to be grown or shrunk (as well as how much it should be grown or shrunk). Approved by: re (kensmith)
2007-07-12 18:01:31 +00:00
RLIMIT_STACK);
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
break;
case RLIMIT_NOFILE:
if (limp->rlim_cur > maxfilesperproc)
limp->rlim_cur = maxfilesperproc;
if (limp->rlim_max > maxfilesperproc)
limp->rlim_max = maxfilesperproc;
break;
case RLIMIT_NPROC:
if (limp->rlim_cur > maxprocperuid)
limp->rlim_cur = maxprocperuid;
if (limp->rlim_max > maxprocperuid)
limp->rlim_max = maxprocperuid;
if (limp->rlim_cur < 1)
limp->rlim_cur = 1;
if (limp->rlim_max < 1)
limp->rlim_max = 1;
break;
}
We've already set p = td->td_proc, so use it.
2010-10-18 15:46:58 +00:00
if (p->p_sysent->sv_fixlimit != NULL)
p->p_sysent->sv_fixlimit(limp, which);
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
*alimp = *limp;
p->p_limit = newlim;
PROC_UNLOCK(p);
lim_free(oldlim);
if (which == RLIMIT_STACK) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Stack is allocated to the max at exec time with only
* "rlim_cur" bytes accessible. If stack limit is going
* up make more accessible, if going down make inaccessible.
*/
Fix a couple of issues with the stack limit for 32-bit processes on 64-bit kernels exposed by the recent fixes to resource limits for 32-bit processes on 64-bit kernels: - Let ABIs expose their maximum stack size via a new pointer in sysentvec and use that in preference to maxssiz during exec() rather than always using maxssiz for all processses. - Apply the ABI's limit fixup to the previous stack size when adjusting RLIMIT_STACK to determine if the existing mapping for the stack needs to be grown or shrunk (as well as how much it should be grown or shrunk). Approved by: re (kensmith)
2007-07-12 18:01:31 +00:00
if (limp->rlim_cur != oldssiz.rlim_cur) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
vm_offset_t addr;
vm_size_t size;
vm_prot_t prot;
Fix a couple of issues with the stack limit for 32-bit processes on 64-bit kernels exposed by the recent fixes to resource limits for 32-bit processes on 64-bit kernels: - Let ABIs expose their maximum stack size via a new pointer in sysentvec and use that in preference to maxssiz during exec() rather than always using maxssiz for all processses. - Apply the ABI's limit fixup to the previous stack size when adjusting RLIMIT_STACK to determine if the existing mapping for the stack needs to be grown or shrunk (as well as how much it should be grown or shrunk). Approved by: re (kensmith)
2007-07-12 18:01:31 +00:00
if (limp->rlim_cur > oldssiz.rlim_cur) {
Use the fields in the sysentvec and in the vm map header in place of the constants VM_MIN_ADDRESS, VM_MAXUSER_ADDRESS, USRSTACK and PS_STRINGS. This is mainly so that they can be variable even for the native abi, based on different machine types. Get stack protections from the sysentvec too. This makes it trivial to map the stack non-executable for certain abis, on machines that support it.
2002-09-21 22:07:17 +00:00
prot = p->p_sysent->sv_stackprot;
Fix a couple of issues with the stack limit for 32-bit processes on 64-bit kernels exposed by the recent fixes to resource limits for 32-bit processes on 64-bit kernels: - Let ABIs expose their maximum stack size via a new pointer in sysentvec and use that in preference to maxssiz during exec() rather than always using maxssiz for all processses. - Apply the ABI's limit fixup to the previous stack size when adjusting RLIMIT_STACK to determine if the existing mapping for the stack needs to be grown or shrunk (as well as how much it should be grown or shrunk). Approved by: re (kensmith)
2007-07-12 18:01:31 +00:00
size = limp->rlim_cur - oldssiz.rlim_cur;
Use the fields in the sysentvec and in the vm map header in place of the constants VM_MIN_ADDRESS, VM_MAXUSER_ADDRESS, USRSTACK and PS_STRINGS. This is mainly so that they can be variable even for the native abi, based on different machine types. Get stack protections from the sysentvec too. This makes it trivial to map the stack non-executable for certain abis, on machines that support it.
2002-09-21 22:07:17 +00:00
addr = p->p_sysent->sv_usrstack -
limp->rlim_cur;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
} else {
prot = VM_PROT_NONE;
Fix a couple of issues with the stack limit for 32-bit processes on 64-bit kernels exposed by the recent fixes to resource limits for 32-bit processes on 64-bit kernels: - Let ABIs expose their maximum stack size via a new pointer in sysentvec and use that in preference to maxssiz during exec() rather than always using maxssiz for all processses. - Apply the ABI's limit fixup to the previous stack size when adjusting RLIMIT_STACK to determine if the existing mapping for the stack needs to be grown or shrunk (as well as how much it should be grown or shrunk). Approved by: re (kensmith)
2007-07-12 18:01:31 +00:00
size = oldssiz.rlim_cur - limp->rlim_cur;
addr = p->p_sysent->sv_usrstack -
oldssiz.rlim_cur;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
addr = trunc_page(addr);
size = round_page(size);
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
(void)vm_map_protect(&p->p_vmspace->vm_map,
addr, addr + size, prot, FALSE);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
}
Calling setrlimit from 32bit apps could potentially increase certain limits beyond what should be capiable in a 32bit process, so we must fixup the limits. Reviewed by: jhb
2005-11-02 21:18:07 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return (0);
}
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#ifndef _SYS_SYSPROTO_H_
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct __getrlimit_args {
u_int which;
struct rlimit *rlp;
};
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#endif
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/* ARGSUSED */
The big 4.4BSD Lite to FreeBSD 2.0.0 (Development) patch. Reviewed by: Rodney W. Grimes Submitted by: John Dyson and David Greenman
1994-05-25 09:21:21 +00:00
int
In order to maximize the re-usability of kernel code in user space this patch modifies makesyscalls.sh to prefix all of the non-compatibility calls (e.g. not linux_, freebsd32_) with sys_ and updates the kernel entry points and all places in the code that use them. It also fixes an additional name space collision between the kernel function psignal and the libc function of the same name by renaming the kernel psignal kern_psignal(). By introducing this change now we will ease future MFCs that change syscalls. Reviewed by: rwatson Approved by: re (bz)
2011-09-16 13:58:51 +00:00
sys_getrlimit(td, uap)
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
struct thread *td;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
register struct __getrlimit_args *uap;
{
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
struct rlimit rlim;
A few more style fixes from Bruce including a few I missed last time. Submitted by: bde
2004-02-06 19:25:34 +00:00
struct proc *p;
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
int error;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (uap->which >= RLIM_NLIMITS)
return (EINVAL);
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
p = td->td_proc;
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
PROC_LOCK(p);
lim_rlimit(p, uap->which, &rlim);
PROC_UNLOCK(p);
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
error = copyout(&rlim, uap->rlp, sizeof(struct rlimit));
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Split calcru() back into a calcru1() function shared with calccru() and a calcru() wrapper that passes a local rusage_ext on the stack that is a snapshot to do the calculations on. Now we can pass p->p_crux to calcru1() in calccru() again which fixes the issues with runtime going backwards messages when dead processes are harvested by init. Reviewed by: phk Tested by: Stefan Ehmann shoesoft at gmx dot net
2006-02-21 21:47:46 +00:00
/*
Various style and comment fixes. Submitted by: bde
2006-02-22 16:58:48 +00:00
* Transform the running time and tick information for children of proc p
* into user and system time usage.
Split calcru() back into a calcru1() function shared with calccru() and a calcru() wrapper that passes a local rusage_ext on the stack that is a snapshot to do the calculations on. Now we can pass p->p_crux to calcru1() in calccru() again which fixes the issues with runtime going backwards messages when dead processes are harvested by init. Reviewed by: phk Tested by: Stefan Ehmann shoesoft at gmx dot net
2006-02-21 21:47:46 +00:00
*/
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
void
calccru(p, up, sp)
struct proc *p;
struct timeval *up;
struct timeval *sp;
{
PROC_LOCK_ASSERT(p, MA_OWNED);
Split calcru() back into a calcru1() function shared with calccru() and a calcru() wrapper that passes a local rusage_ext on the stack that is a snapshot to do the calculations on. Now we can pass p->p_crux to calcru1() in calccru() again which fixes the issues with runtime going backwards messages when dead processes are harvested by init. Reviewed by: phk Tested by: Stefan Ehmann shoesoft at gmx dot net
2006-02-21 21:47:46 +00:00
calcru1(p, &p->p_crux, up, sp);
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
}
CPU time accounting speedup (step 2) Keep accounting time (in per-cpu) cputicks and the statistics counts in the thread and summarize into struct proc when at context switch. Don't reach across CPUs in calcru(). Add code to calibrate the top speed of cpu_tickrate() for variable cpu_tick hardware (like TSC on power managed machines). Don't enforce monotonicity (at least for now) in calcru. While the calibrated cpu_tickrate ramps up it may not be true. Use 27MHz counter on i386/Geode. Use TSC on amd64 & i386 if present. Use tick counter on sparc64
2006-02-11 09:33:07 +00:00
/*
Various style and comment fixes. Submitted by: bde
2006-02-22 16:58:48 +00:00
* Transform the running time and tick information in proc p into user
* and system time usage. If appropriate, include the current time slice
* on this CPU.
CPU time accounting speedup (step 2) Keep accounting time (in per-cpu) cputicks and the statistics counts in the thread and summarize into struct proc when at context switch. Don't reach across CPUs in calcru(). Add code to calibrate the top speed of cpu_tickrate() for variable cpu_tick hardware (like TSC on power managed machines). Don't enforce monotonicity (at least for now) in calcru. While the calibrated cpu_tickrate ramps up it may not be true. Use 27MHz counter on i386/Geode. Use TSC on amd64 & i386 if present. Use tick counter on sparc64
2006-02-11 09:33:07 +00:00
*/
void
calcru(struct proc *p, struct timeval *up, struct timeval *sp)
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
{
Various style and comment fixes. Submitted by: bde
2006-02-22 16:58:48 +00:00
struct thread *td;
- Export each thread's individual resource usage in in struct kinfo_proc's ki_rusage member when KERN_PROC_INC_THREAD is passed to one of the process sysctls. - Correctly account for the current thread's cputime in the thread when doing the runtime fixup in calcru(). - Use TIDs as the key to lookup the previous thread to compute IO stat deltas in IO mode in top when thread display is enabled. Reviewed by: kib Approved by: re (kib)
2011-07-18 17:33:08 +00:00
uint64_t runtime, u;
- Move ke_sticks, ke_iticks, ke_uticks, ke_uu, ke_su, and ke_iu back into the proc. These counters are only examined through calcru. Submitted by: davidxu Tested on: x86, alpha, UP/SMP
2003-02-17 02:19:58 +00:00
CPU time accounting speedup (step 2) Keep accounting time (in per-cpu) cputicks and the statistics counts in the thread and summarize into struct proc when at context switch. Don't reach across CPUs in calcru(). Add code to calibrate the top speed of cpu_tickrate() for variable cpu_tick hardware (like TSC on power managed machines). Don't enforce monotonicity (at least for now) in calcru. While the calibrated cpu_tickrate ramps up it may not be true. Use 27MHz counter on i386/Geode. Use TSC on amd64 & i386 if present. Use tick counter on sparc64
2006-02-11 09:33:07 +00:00
PROC_LOCK_ASSERT(p, MA_OWNED);
rufetch and calcru sometimes should be called atomically together. This patch fixes places where they should be called atomically changing their locking requirements (both assume per-proc spinlock held) and introducing rufetchcalc which wrappers both calls to be performed in atomic way. Reviewed by: jeff Approved by: jeff (mentor)
2007-06-09 21:48:44 +00:00
PROC_SLOCK_ASSERT(p, MA_OWNED);
Split calcru() back into a calcru1() function shared with calccru() and a calcru() wrapper that passes a local rusage_ext on the stack that is a snapshot to do the calculations on. Now we can pass p->p_crux to calcru1() in calccru() again which fixes the issues with runtime going backwards messages when dead processes are harvested by init. Reviewed by: phk Tested by: Stefan Ehmann shoesoft at gmx dot net
2006-02-21 21:47:46 +00:00
/*
* If we are getting stats for the current process, then add in the
* stats that this thread has accumulated in its current time slice.
* We reset the thread and CPU state as if we had performed a context
* switch right here.
*/
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
td = curthread;
if (td->td_proc == p) {
CPU time accounting speedup (step 2) Keep accounting time (in per-cpu) cputicks and the statistics counts in the thread and summarize into struct proc when at context switch. Don't reach across CPUs in calcru(). Add code to calibrate the top speed of cpu_tickrate() for variable cpu_tick hardware (like TSC on power managed machines). Don't enforce monotonicity (at least for now) in calcru. While the calibrated cpu_tickrate ramps up it may not be true. Use 27MHz counter on i386/Geode. Use TSC on amd64 & i386 if present. Use tick counter on sparc64
2006-02-11 09:33:07 +00:00
u = cpu_ticks();
- Export each thread's individual resource usage in in struct kinfo_proc's ki_rusage member when KERN_PROC_INC_THREAD is passed to one of the process sysctls. - Correctly account for the current thread's cputime in the thread when doing the runtime fixup in calcru(). - Use TIDs as the key to lookup the previous thread to compute IO stat deltas in IO mode in top when thread display is enabled. Reviewed by: kib Approved by: re (kib)
2011-07-18 17:33:08 +00:00
runtime = u - PCPU_GET(switchtime);
td->td_runtime += runtime;
td->td_incruntime += runtime;
CPU time accounting speedup (step 2) Keep accounting time (in per-cpu) cputicks and the statistics counts in the thread and summarize into struct proc when at context switch. Don't reach across CPUs in calcru(). Add code to calibrate the top speed of cpu_tickrate() for variable cpu_tick hardware (like TSC on power managed machines). Don't enforce monotonicity (at least for now) in calcru. While the calibrated cpu_tickrate ramps up it may not be true. Use 27MHz counter on i386/Geode. Use TSC on amd64 & i386 if present. Use tick counter on sparc64
2006-02-11 09:33:07 +00:00
PCPU_SET(switchtime, u);
}
- Use ruxagg() in calcru() to make sure we have current tick information from all threads. Discussed with: bde, attilio Approved by: re
2007-07-17 01:08:09 +00:00
/* Make sure the per-thread stats are current. */
FOREACH_THREAD_IN_PROC(p, td) {
Don't zero td_runtime when billing thread CPU usage to the process; maintain a separate td_incruntime to hold unbilled CPU usage for the thread that has the previous properties of td_runtime. When thread information is requested using the thread monitoring sysctls, export thread td_runtime instead of process rusage runtime in kinfo_proc. This restores the display of individual ithread and other kernel thread CPU usage since inception in ps -H and top -SH, as well for libthr user threads, valuable debugging information lost with the move to try kthreads since they are no longer independent processes. There is universal agreement that we should rewrite the process and thread export sysctls, but this commit gets things going a bit better in the mean time. Likewise, there are resevations about the continued validity of statclock given the speed of modern processors. Reviewed by: attilio, emaste, jhb, julian
2008-01-10 22:11:20 +00:00
if (td->td_incruntime == 0)
- Use ruxagg() in calcru() to make sure we have current tick information from all threads. Discussed with: bde, attilio Approved by: re
2007-07-17 01:08:09 +00:00
continue;
Implement RUSAGE_THREAD. Add td_rux to keep extended runtime and ticks information for thread to allow calcru1() (re)use. Rename ruxagg()->ruxagg_locked(), ruxagg_tlock()->ruxagg() [1]. The ruxagg_locked() function no longer clears thread ticks nor td_incruntime. Requested by: attilio [1] Discussed with: attilio, bde Reviewed by: bde Based on submission by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week X-MFC-Note: td_rux shall be moved to the end of struct thread
2010-05-04 05:55:37 +00:00
ruxagg(p, td);
- Use ruxagg() in calcru() to make sure we have current tick information from all threads. Discussed with: bde, attilio Approved by: re
2007-07-17 01:08:09 +00:00
}
rufetch and calcru sometimes should be called atomically together. This patch fixes places where they should be called atomically changing their locking requirements (both assume per-proc spinlock held) and introducing rufetchcalc which wrappers both calls to be performed in atomic way. Reviewed by: jeff Approved by: jeff (mentor)
2007-06-09 21:48:44 +00:00
calcru1(p, &p->p_rux, up, sp);
Split calcru() back into a calcru1() function shared with calccru() and a calcru() wrapper that passes a local rusage_ext on the stack that is a snapshot to do the calculations on. Now we can pass p->p_crux to calcru1() in calccru() again which fixes the issues with runtime going backwards messages when dead processes are harvested by init. Reviewed by: phk Tested by: Stefan Ehmann shoesoft at gmx dot net
2006-02-21 21:47:46 +00:00
}
- Export each thread's individual resource usage in in struct kinfo_proc's ki_rusage member when KERN_PROC_INC_THREAD is passed to one of the process sysctls. - Correctly account for the current thread's cputime in the thread when doing the runtime fixup in calcru(). - Use TIDs as the key to lookup the previous thread to compute IO stat deltas in IO mode in top when thread display is enabled. Reviewed by: kib Approved by: re (kib)
2011-07-18 17:33:08 +00:00
/* Collect resource usage for a single thread. */
void
rufetchtd(struct thread *td, struct rusage *ru)
{
struct proc *p;
uint64_t runtime, u;
p = td->td_proc;
PROC_SLOCK_ASSERT(p, MA_OWNED);
THREAD_LOCK_ASSERT(td, MA_OWNED);
/*
* If we are getting stats for the current thread, then add in the
* stats that this thread has accumulated in its current time slice.
* We reset the thread and CPU state as if we had performed a context
* switch right here.
*/
if (td == curthread) {
u = cpu_ticks();
runtime = u - PCPU_GET(switchtime);
td->td_runtime += runtime;
td->td_incruntime += runtime;
PCPU_SET(switchtime, u);
}
ruxagg(p, td);
*ru = td->td_ru;
calcru1(p, &td->td_rux, &ru->ru_utime, &ru->ru_stime);
}
Split calcru() back into a calcru1() function shared with calccru() and a calcru() wrapper that passes a local rusage_ext on the stack that is a snapshot to do the calculations on. Now we can pass p->p_crux to calcru1() in calccru() again which fixes the issues with runtime going backwards messages when dead processes are harvested by init. Reviewed by: phk Tested by: Stefan Ehmann shoesoft at gmx dot net
2006-02-21 21:47:46 +00:00
static void
calcru1(struct proc *p, struct rusage_ext *ruxp, struct timeval *up,
struct timeval *sp)
{
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
/* {user, system, interrupt, total} {ticks, usec}: */
Use ISO C99 integer types in sys/kern where possible. There are only about 100 occurences of the BSD-specific u_int*_t datatypes in sys/kern. The ISO C99 integer types are used here more often.
2010-06-21 09:55:56 +00:00
uint64_t ut, uu, st, su, it, tt, tu;
CPU time accounting speedup (step 2) Keep accounting time (in per-cpu) cputicks and the statistics counts in the thread and summarize into struct proc when at context switch. Don't reach across CPUs in calcru(). Add code to calibrate the top speed of cpu_tickrate() for variable cpu_tick hardware (like TSC on power managed machines). Don't enforce monotonicity (at least for now) in calcru. While the calibrated cpu_tickrate ramps up it may not be true. Use 27MHz counter on i386/Geode. Use TSC on amd64 & i386 if present. Use tick counter on sparc64
2006-02-11 09:33:07 +00:00
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
ut = ruxp->rux_uticks;
st = ruxp->rux_sticks;
it = ruxp->rux_iticks;
- Move ke_sticks, ke_iticks, ke_uticks, ke_uu, ke_su, and ke_iu back into the proc. These counters are only examined through calcru. Submitted by: davidxu Tested on: x86, alpha, UP/SMP
2003-02-17 02:19:58 +00:00
tt = ut + st + it;
if (tt == 0) {
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
/* Avoid divide by zero */
- Move ke_sticks, ke_iticks, ke_uticks, ke_uu, ke_su, and ke_iu back into the proc. These counters are only examined through calcru. Submitted by: davidxu Tested on: x86, alpha, UP/SMP
2003-02-17 02:19:58 +00:00
st = 1;
tt = 1;
}
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
tu = cputick2usec(ruxp->rux_runtime);
Turned off the "calcru: negative time" warning for certain SMP cases where it is known to detect a problem but the problem is not very easy to fix. The warning became very common recently after a call to calcru() was added to fill_kinfo_thread(). Another (much older) cause of "negative times" (actually non-monotonic times) was fixed in rev.1.237 of kern_exit.c. Print separate messages for non-monotonic and negative times.
2004-06-21 17:46:27 +00:00
if ((int64_t)tu < 0) {
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
/* XXX: this should be an assert /phk */
Turned off the "calcru: negative time" warning for certain SMP cases where it is known to detect a problem but the problem is not very easy to fix. The warning became very common recently after a call to calcru() was added to fill_kinfo_thread(). Another (much older) cause of "negative times" (actually non-monotonic times) was fixed in rev.1.237 of kern_exit.c. Print separate messages for non-monotonic and negative times.
2004-06-21 17:46:27 +00:00
printf("calcru: negative runtime of %jd usec for pid %d (%s)\n",
- Correct the translation of old rlimit values to properly handle the old RLIM_INFINITY case for ogetrlimit(). - Use %jd and intmax_t to output negative time in usec in calcru(). - Rework getrusage() to make a copy of the rusage struct into a local variable while holding Giant and then do the copyout from the local variable to avoid having to have the original process rusage struct locked while doing the copyout (which would not be safe). This also includes a few style fixes from Bruce to getrusage(). Submitted by: bde (1, parts of 3) Suggested by: bde (2)
2004-02-06 19:30:12 +00:00
(intmax_t)tu, p->p_pid, p->p_comm);
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
tu = ruxp->rux_tu;
- Move ke_sticks, ke_iticks, ke_uticks, ke_uu, ke_su, and ke_iu back into the proc. These counters are only examined through calcru. Submitted by: davidxu Tested on: x86, alpha, UP/SMP
2003-02-17 02:19:58 +00:00
}
Enforce monotonicity of apparent process user, system and interrupt times. PR: 975, 10402
1999-03-13 19:46:13 +00:00
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
if (tu >= ruxp->rux_tu) {
/*
* The normal case, time increased.
* Enforce monotonicity of bucketed numbers.
*/
uu = (tu * ut) / tt;
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
if (uu < ruxp->rux_uu)
uu = ruxp->rux_uu;
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
su = (tu * st) / tt;
if (su < ruxp->rux_su)
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
su = ruxp->rux_su;
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
} else if (tu + 3 > ruxp->rux_tu || 101 * tu > 100 * ruxp->rux_tu) {
Whitespace cleanups.
2008-03-16 21:32:20 +00:00
/*
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
* When we calibrate the cputicker, it is not uncommon to
* see the presumably fixed frequency increase slightly over
* time as a result of thermal stabilization and NTP
* discipline (of the reference clock). We therefore ignore
* a bit of backwards slop because we expect to catch up
Whitespace cleanups.
2008-03-16 21:32:20 +00:00
* shortly. We use a 3 microsecond limit to catch low
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
* counts and a 1% limit for high counts.
*/
uu = ruxp->rux_uu;
su = ruxp->rux_su;
tu = ruxp->rux_tu;
} else { /* tu < ruxp->rux_tu */
/*
Fix a small typo in a comment in calcru1(). The word "happene" should read "happened". Submitted by: Jille Timmermans <jille quis cx>
2008-09-05 15:55:06 +00:00
* What happened here was likely that a laptop, which ran at
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
* a reduced clock frequency at boot, kicked into high gear.
* The wisdom of spamming this message in that case is
* dubious, but it might also be indicative of something
* serious, so lets keep it and hope laptops can be made
* more truthful about their CPU speed via ACPI.
*/
printf("calcru: runtime went backwards from %ju usec "
"to %ju usec for pid %d (%s)\n",
(uintmax_t)ruxp->rux_tu, (uintmax_t)tu,
p->p_pid, p->p_comm);
uu = (tu * ut) / tt;
su = (tu * st) / tt;
- Move ke_sticks, ke_iticks, ke_uticks, ke_uu, ke_su, and ke_iu back into the proc. These counters are only examined through calcru. Submitted by: davidxu Tested on: x86, alpha, UP/SMP
2003-02-17 02:19:58 +00:00
}
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
ruxp->rux_uu = uu;
ruxp->rux_su = su;
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
ruxp->rux_tu = tu;
- Move ke_sticks, ke_iticks, ke_uticks, ke_uu, ke_su, and ke_iu back into the proc. These counters are only examined through calcru. Submitted by: davidxu Tested on: x86, alpha, UP/SMP
2003-02-17 02:19:58 +00:00
up->tv_sec = uu / 1000000;
up->tv_usec = uu % 1000000;
sp->tv_sec = su / 1000000;
sp->tv_usec = su % 1000000;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#ifndef _SYS_SYSPROTO_H_
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct getrusage_args {
int who;
struct rusage *rusage;
};
Included <sys/sysproto.h> to get central declarations for syscall args structs and prototypes for syscalls. Ifdefed duplicated decentralized declarations of args structs. It's convenient to have this visible but they are hard to maintain. Some are already different from the central declarations. 4.4lite2 puts them in comments in the function headers but I wanted to avoid the large changes for that.
1995-11-12 06:43:28 +00:00
#endif
The big 4.4BSD Lite to FreeBSD 2.0.0 (Development) patch. Reviewed by: Rodney W. Grimes Submitted by: John Dyson and David Greenman
1994-05-25 09:21:21 +00:00
int
In order to maximize the re-usability of kernel code in user space this patch modifies makesyscalls.sh to prefix all of the non-compatibility calls (e.g. not linux_, freebsd32_) with sys_ and updates the kernel entry points and all places in the code that use them. It also fixes an additional name space collision between the kernel function psignal and the libc function of the same name by renaming the kernel psignal kern_psignal(). By introducing this change now we will ease future MFCs that change syscalls. Reviewed by: rwatson Approved by: re (bz)
2011-09-16 13:58:51 +00:00
sys_getrusage(td, uap)
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
register struct thread *td;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
register struct getrusage_args *uap;
{
- Correct the translation of old rlimit values to properly handle the old RLIM_INFINITY case for ogetrlimit(). - Use %jd and intmax_t to output negative time in usec in calcru(). - Rework getrusage() to make a copy of the rusage struct into a local variable while holding Giant and then do the copyout from the local variable to avoid having to have the original process rusage struct locked while doing the copyout (which would not be safe). This also includes a few style fixes from Bruce to getrusage(). Submitted by: bde (1, parts of 3) Suggested by: bde (2)
2004-02-06 19:30:12 +00:00
struct rusage ru;
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
int error;
error = kern_getrusage(td, uap->who, &ru);
if (error == 0)
error = copyout(&ru, uap->rusage, sizeof(struct rusage));
return (error);
}
int
Extract thread_lock()/ruxagg()/thread_unlock() fragment into utility function ruxagg_tlock(). Convert the definition of kern_getrusage() to ANSI C. Submitted by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week
2010-05-01 14:46:17 +00:00
kern_getrusage(struct thread *td, int who, struct rusage *rup)
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
{
- Correct the translation of old rlimit values to properly handle the old RLIM_INFINITY case for ogetrlimit(). - Use %jd and intmax_t to output negative time in usec in calcru(). - Rework getrusage() to make a copy of the rusage struct into a local variable while holding Giant and then do the copyout from the local variable to avoid having to have the original process rusage struct locked while doing the copyout (which would not be safe). This also includes a few style fixes from Bruce to getrusage(). Submitted by: bde (1, parts of 3) Suggested by: bde (2)
2004-02-06 19:30:12 +00:00
struct proc *p;
- Relax requirements for p_numthreads, p_threads, p_swtick, and p_nice from requiring the per-process spinlock to only requiring the process lock. - Reflect these changes in the proc.h documentation and consumers throughout the kernel. This is a substantial reduction in locking cost for these fields and was made possible by recent changes to threading support.
2008-03-19 06:19:01 +00:00
int error;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
- Relax requirements for p_numthreads, p_threads, p_swtick, and p_nice from requiring the per-process spinlock to only requiring the process lock. - Reflect these changes in the proc.h documentation and consumers throughout the kernel. This is a substantial reduction in locking cost for these fields and was made possible by recent changes to threading support.
2008-03-19 06:19:01 +00:00
error = 0;
- Correct the translation of old rlimit values to properly handle the old RLIM_INFINITY case for ogetrlimit(). - Use %jd and intmax_t to output negative time in usec in calcru(). - Rework getrusage() to make a copy of the rusage struct into a local variable while holding Giant and then do the copyout from the local variable to avoid having to have the original process rusage struct locked while doing the copyout (which would not be safe). This also includes a few style fixes from Bruce to getrusage(). Submitted by: bde (1, parts of 3) Suggested by: bde (2)
2004-02-06 19:30:12 +00:00
p = td->td_proc;
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
PROC_LOCK(p);
switch (who) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
case RUSAGE_SELF:
rufetch and calcru sometimes should be called atomically together. This patch fixes places where they should be called atomically changing their locking requirements (both assume per-proc spinlock held) and introducing rufetchcalc which wrappers both calls to be performed in atomic way. Reviewed by: jeff Approved by: jeff (mentor)
2007-06-09 21:48:44 +00:00
rufetchcalc(p, rup, &rup->ru_utime,
&rup->ru_stime);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
case RUSAGE_CHILDREN:
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
*rup = p->p_stats->p_cru;
calccru(p, &rup->ru_utime, &rup->ru_stime);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
Implement RUSAGE_THREAD. Add td_rux to keep extended runtime and ticks information for thread to allow calcru1() (re)use. Rename ruxagg()->ruxagg_locked(), ruxagg_tlock()->ruxagg() [1]. The ruxagg_locked() function no longer clears thread ticks nor td_incruntime. Requested by: attilio [1] Discussed with: attilio, bde Reviewed by: bde Based on submission by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week X-MFC-Note: td_rux shall be moved to the end of struct thread
2010-05-04 05:55:37 +00:00
case RUSAGE_THREAD:
PROC_SLOCK(p);
thread_lock(td);
- Export each thread's individual resource usage in in struct kinfo_proc's ki_rusage member when KERN_PROC_INC_THREAD is passed to one of the process sysctls. - Correctly account for the current thread's cputime in the thread when doing the runtime fixup in calcru(). - Use TIDs as the key to lookup the previous thread to compute IO stat deltas in IO mode in top when thread display is enabled. Reviewed by: kib Approved by: re (kib)
2011-07-18 17:33:08 +00:00
rufetchtd(td, rup);
Implement RUSAGE_THREAD. Add td_rux to keep extended runtime and ticks information for thread to allow calcru1() (re)use. Rename ruxagg()->ruxagg_locked(), ruxagg_tlock()->ruxagg() [1]. The ruxagg_locked() function no longer clears thread ticks nor td_incruntime. Requested by: attilio [1] Discussed with: attilio, bde Reviewed by: bde Based on submission by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week X-MFC-Note: td_rux shall be moved to the end of struct thread
2010-05-04 05:55:37 +00:00
thread_unlock(td);
- Export each thread's individual resource usage in in struct kinfo_proc's ki_rusage member when KERN_PROC_INC_THREAD is passed to one of the process sysctls. - Correctly account for the current thread's cputime in the thread when doing the runtime fixup in calcru(). - Use TIDs as the key to lookup the previous thread to compute IO stat deltas in IO mode in top when thread display is enabled. Reviewed by: kib Approved by: re (kib)
2011-07-18 17:33:08 +00:00
PROC_SUNLOCK(p);
Implement RUSAGE_THREAD. Add td_rux to keep extended runtime and ticks information for thread to allow calcru1() (re)use. Rename ruxagg()->ruxagg_locked(), ruxagg_tlock()->ruxagg() [1]. The ruxagg_locked() function no longer clears thread ticks nor td_incruntime. Requested by: attilio [1] Discussed with: attilio, bde Reviewed by: bde Based on submission by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week X-MFC-Note: td_rux shall be moved to the end of struct thread
2010-05-04 05:55:37 +00:00
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
default:
- Relax requirements for p_numthreads, p_threads, p_swtick, and p_nice from requiring the per-process spinlock to only requiring the process lock. - Reflect these changes in the proc.h documentation and consumers throughout the kernel. This is a substantial reduction in locking cost for these fields and was made possible by recent changes to threading support.
2008-03-19 06:19:01 +00:00
error = EINVAL;
Giant Pushdown. Saved the worst P4 tree breakage for last. reboot() getpriority() setpriority() rtprio() osetrlimit() ogetrlimit() setrlimit() getrlimit() getrusage() getpid() getppid() getpgrp() getpgid() getsid() getgid() getegid() getgroups() setsid() setpgid() setuid() seteuid() setgid() setegid() setgroups() setreuid() setregid() setresuid() setresgid() getresuid() getresgid () __setugid() getlogin() setlogin() modnext() modfnext() modstat() modfind() kldload() kldunload() kldfind() kldnext() kldstat() kldfirstmod() kldsym() getdtablesize() dup2() dup() fcntl() close() ofstat() fstat() nfsstat() fpathconf() flock()
2001-09-01 19:04:37 +00:00
}
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
PROC_UNLOCK(p);
- Relax requirements for p_numthreads, p_threads, p_swtick, and p_nice from requiring the per-process spinlock to only requiring the process lock. - Reflect these changes in the proc.h documentation and consumers throughout the kernel. This is a substantial reduction in locking cost for these fields and was made possible by recent changes to threading support.
2008-03-19 06:19:01 +00:00
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
The big 4.4BSD Lite to FreeBSD 2.0.0 (Development) patch. Reviewed by: Rodney W. Grimes Submitted by: John Dyson and David Greenman
1994-05-25 09:21:21 +00:00
void
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
rucollect(struct rusage *ru, struct rusage *ru2)
{
long *ip, *ip2;
int i;
if (ru->ru_maxrss < ru2->ru_maxrss)
ru->ru_maxrss = ru2->ru_maxrss;
ip = &ru->ru_first;
ip2 = &ru2->ru_first;
for (i = &ru->ru_last - &ru->ru_first; i >= 0; i--)
*ip++ += *ip2++;
}
void
ruadd(struct rusage *ru, struct rusage_ext *rux, struct rusage *ru2,
struct rusage_ext *rux2)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
Modify the way we account for CPU time spent (step 1) Keep track of time spent by the cpu in various contexts in units of "cputicks" and scale to real-world microsec^H^H^H^H^H^H^H^Hclock_t only when somebody wants to inspect the numbers. For now "cputicks" are still derived from the current timecounter and therefore things should by definition remain sensible also on SMP machines. (The main reason for this first milestone commit is to verify that hypothesis.) On slower machines, the avoided multiplications to normalize timestams at every context switch, comes out as a 5-7% better score on the unixbench/context1 microbenchmark. On more modern hardware no change in performance is seen.
2006-02-07 21:22:02 +00:00
rux->rux_runtime += rux2->rux_runtime;
Rework how we store process times in the kernel such that we always store the raw values including for child process statistics and only compute the system and user timevals on demand. - Fix the various kern_wait() syscall wrappers to only pass in a rusage pointer if they are going to use the result. - Add a kern_getrusage() function for the ABI syscalls to use so that they don't have to play stackgap games to call getrusage(). - Fix the svr4_sys_times() syscall to just call calcru() to calculate the times it needs rather than calling getrusage() twice with associated stackgap, etc. - Add a new rusage_ext structure to store raw time stats such as tick counts for user, system, and interrupt time as well as a bintime of the total runtime. A new p_rux field in struct proc replaces the same inline fields from struct proc (i.e. p_[isu]ticks, p_[isu]u, and p_runtime). A new p_crux field in struct proc contains the "raw" child time usage statistics. ruadd() has been changed to handle adding the associated rusage_ext structures as well as the values in rusage. Effectively, the values in rusage_ext replace the ru_utime and ru_stime values in struct rusage. These two fields in struct rusage are no longer used in the kernel. - calcru() has been split into a static worker function calcru1() that calculates appropriate timevals for user and system time as well as updating the rux_[isu]u fields of a passed in rusage_ext structure. calcru() uses a copy of the process' p_rux structure to compute the timevals after updating the runtime appropriately if any of the threads in that process are currently executing. It also now only locks sched_lock internally while doing the rux_runtime fixup. calcru() now only requires the caller to hold the proc lock and calcru1() only requires the proc lock internally. calcru() also no longer allows callers to ask for an interrupt timeval since none of them actually did. - calcru() now correctly handles threads executing on other CPUs. - A new calccru() function computes the child system and user timevals by calling calcru1() on p_crux. Note that this means that any code that wants child times must now call this function rather than reading from p_cru directly. This function also requires the proc lock. - This finishes the locking for rusage and friends so some of the Giant locks in exit1() and kern_wait() are now gone. - The locking in ttyinfo() has been tweaked so that a shared lock of the proctree lock is used to protect the process group rather than the process group lock. By holding this lock until the end of the function we now ensure that the process/thread that we pick to dump info about will no longer vanish while we are trying to output its info to the console. Submitted by: bde (mostly) MFC after: 1 month
2004-10-05 18:51:11 +00:00
rux->rux_uticks += rux2->rux_uticks;
rux->rux_sticks += rux2->rux_sticks;
rux->rux_iticks += rux2->rux_iticks;
rux->rux_uu += rux2->rux_uu;
rux->rux_su += rux2->rux_su;
Go over calcru and friends once more. Reintroduce the monotonicity for the normal case and make the two special cases behave in what is belived to be the most sensible fasion.
2006-03-11 10:48:19 +00:00
rux->rux_tu += rux2->rux_tu;
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
rucollect(ru, ru2);
}
/*
* Aggregate tick counts into the proc's rusage_ext.
*/
Fix the double counting of the last process thread td_incruntime on exit, that is done once in thread_exit() and the second time in proc_reap(), by clearing td_incruntime. Use the opportunity to revert to the pre-RUSAGE_THREAD exporting of ruxagg() instead of ruxagg_locked() and use it from thread_exit(). Diagnosed and tested by: neel MFC after: 3 days
2010-05-24 10:23:49 +00:00
static void
Implement RUSAGE_THREAD. Add td_rux to keep extended runtime and ticks information for thread to allow calcru1() (re)use. Rename ruxagg()->ruxagg_locked(), ruxagg_tlock()->ruxagg() [1]. The ruxagg_locked() function no longer clears thread ticks nor td_incruntime. Requested by: attilio [1] Discussed with: attilio, bde Reviewed by: bde Based on submission by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week X-MFC-Note: td_rux shall be moved to the end of struct thread
2010-05-04 05:55:37 +00:00
ruxagg_locked(struct rusage_ext *rux, struct thread *td)
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
{
Commit 14/14 of sched_lock decomposition. - Use thread_lock() rather than sched_lock for per-thread scheduling sychronization. - Use the per-process spinlock rather than the sched_lock for per-process scheduling synchronization. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-05 00:00:57 +00:00
THREAD_LOCK_ASSERT(td, MA_OWNED);
PROC_SLOCK_ASSERT(td->td_proc, MA_OWNED);
Don't zero td_runtime when billing thread CPU usage to the process; maintain a separate td_incruntime to hold unbilled CPU usage for the thread that has the previous properties of td_runtime. When thread information is requested using the thread monitoring sysctls, export thread td_runtime instead of process rusage runtime in kinfo_proc. This restores the display of individual ithread and other kernel thread CPU usage since inception in ps -H and top -SH, as well for libthr user threads, valuable debugging information lost with the move to try kthreads since they are no longer independent processes. There is universal agreement that we should rewrite the process and thread export sysctls, but this commit gets things going a bit better in the mean time. Likewise, there are resevations about the continued validity of statclock given the speed of modern processors. Reviewed by: attilio, emaste, jhb, julian
2008-01-10 22:11:20 +00:00
rux->rux_runtime += td->td_incruntime;
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
rux->rux_uticks += td->td_uticks;
rux->rux_sticks += td->td_sticks;
rux->rux_iticks += td->td_iticks;
}
Fix the double counting of the last process thread td_incruntime on exit, that is done once in thread_exit() and the second time in proc_reap(), by clearing td_incruntime. Use the opportunity to revert to the pre-RUSAGE_THREAD exporting of ruxagg() instead of ruxagg_locked() and use it from thread_exit(). Diagnosed and tested by: neel MFC after: 3 days
2010-05-24 10:23:49 +00:00
void
Implement RUSAGE_THREAD. Add td_rux to keep extended runtime and ticks information for thread to allow calcru1() (re)use. Rename ruxagg()->ruxagg_locked(), ruxagg_tlock()->ruxagg() [1]. The ruxagg_locked() function no longer clears thread ticks nor td_incruntime. Requested by: attilio [1] Discussed with: attilio, bde Reviewed by: bde Based on submission by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week X-MFC-Note: td_rux shall be moved to the end of struct thread
2010-05-04 05:55:37 +00:00
ruxagg(struct proc *p, struct thread *td)
Extract thread_lock()/ruxagg()/thread_unlock() fragment into utility function ruxagg_tlock(). Convert the definition of kern_getrusage() to ANSI C. Submitted by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week
2010-05-01 14:46:17 +00:00
{
thread_lock(td);
Implement RUSAGE_THREAD. Add td_rux to keep extended runtime and ticks information for thread to allow calcru1() (re)use. Rename ruxagg()->ruxagg_locked(), ruxagg_tlock()->ruxagg() [1]. The ruxagg_locked() function no longer clears thread ticks nor td_incruntime. Requested by: attilio [1] Discussed with: attilio, bde Reviewed by: bde Based on submission by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week X-MFC-Note: td_rux shall be moved to the end of struct thread
2010-05-04 05:55:37 +00:00
ruxagg_locked(&p->p_rux, td);
ruxagg_locked(&td->td_rux, td);
td->td_incruntime = 0;
td->td_uticks = 0;
td->td_iticks = 0;
td->td_sticks = 0;
Extract thread_lock()/ruxagg()/thread_unlock() fragment into utility function ruxagg_tlock(). Convert the definition of kern_getrusage() to ANSI C. Submitted by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week
2010-05-01 14:46:17 +00:00
thread_unlock(td);
}
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
/*
* Update the rusage_ext structure and fetch a valid aggregate rusage
* for proc p if storage for one is supplied.
*/
void
rufetch(struct proc *p, struct rusage *ru)
{
struct thread *td;
rufetch and calcru sometimes should be called atomically together. This patch fixes places where they should be called atomically changing their locking requirements (both assume per-proc spinlock held) and introducing rufetchcalc which wrappers both calls to be performed in atomic way. Reviewed by: jeff Approved by: jeff (mentor)
2007-06-09 21:48:44 +00:00
PROC_SLOCK_ASSERT(p, MA_OWNED);
The current rusage code show peculiar problems: - Unsafeness on ruadd() in thread_exit() - Unatomicity of thread_exiit() in the exit1() operations This patch addresses these problems allocating p_fd as part of the process and modifying the way it is accessed. A small chunk of this patch, resolves a race about p_state in kern_wait(), since we have to be sure about the zombif-ing process. Submitted by: jeff Approved by: jeff (mentor)
2007-06-09 18:56:11 +00:00
*ru = p->p_ru;
if (p->p_numthreads > 0) {
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
FOREACH_THREAD_IN_PROC(p, td) {
Implement RUSAGE_THREAD. Add td_rux to keep extended runtime and ticks information for thread to allow calcru1() (re)use. Rename ruxagg()->ruxagg_locked(), ruxagg_tlock()->ruxagg() [1]. The ruxagg_locked() function no longer clears thread ticks nor td_incruntime. Requested by: attilio [1] Discussed with: attilio, bde Reviewed by: bde Based on submission by: Alexander Krizhanovsky <ak natsys-lab com> MFC after: 1 week X-MFC-Note: td_rux shall be moved to the end of struct thread
2010-05-04 05:55:37 +00:00
ruxagg(p, td);
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
rucollect(ru, &td->td_ru);
}
The current rusage code show peculiar problems: - Unsafeness on ruadd() in thread_exit() - Unatomicity of thread_exiit() in the exit1() operations This patch addresses these problems allocating p_fd as part of the process and modifying the way it is accessed. A small chunk of this patch, resolves a race about p_state in kern_wait(), since we have to be sure about the zombif-ing process. Submitted by: jeff Approved by: jeff (mentor)
2007-06-09 18:56:11 +00:00
}
rufetch and calcru sometimes should be called atomically together. This patch fixes places where they should be called atomically changing their locking requirements (both assume per-proc spinlock held) and introducing rufetchcalc which wrappers both calls to be performed in atomic way. Reviewed by: jeff Approved by: jeff (mentor)
2007-06-09 21:48:44 +00:00
}
/*
* Atomically perform a rufetch and a calcru together.
* Consumers, can safely assume the calcru is executed only once
* rufetch is completed.
*/
void
rufetchcalc(struct proc *p, struct rusage *ru, struct timeval *up,
struct timeval *sp)
{
PROC_SLOCK(p);
rufetch(p, ru);
calcru(p, up, sp);
Commit 14/14 of sched_lock decomposition. - Use thread_lock() rather than sched_lock for per-thread scheduling sychronization. - Use the per-process spinlock rather than the sched_lock for per-process scheduling synchronization. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-05 00:00:57 +00:00
PROC_SUNLOCK(p);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
/*
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
* Allocate a new resource limits structure and initialize its
* reference count and mutex pointer.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
struct plimit *
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
lim_alloc()
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
{
struct plimit *limp;
A modest collection of various and sundry style, spelling, and whitespace fixes. Submitted by: bde (mostly)
2004-09-24 00:38:15 +00:00
limp = malloc(sizeof(struct plimit), M_PLIMIT, M_WAITOK);
Use the reference count API to manage the reference counts for process limit structures rather than using pool mutexes to protect the reference counts. Tested on: i386, alpha, sparc64
2005-09-27 18:07:05 +00:00
refcount_init(&limp->pl_refcnt, 1);
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
return (limp);
}
struct plimit *
lim_hold(limp)
struct plimit *limp;
{
Use the reference count API to manage the reference counts for process limit structures rather than using pool mutexes to protect the reference counts. Tested on: i386, alpha, sparc64
2005-09-27 18:07:05 +00:00
refcount_acquire(&limp->pl_refcnt);
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
return (limp);
}
- Move rusage from being per-process in struct pstats to per-thread in td_ru. This removes the requirement for per-process synchronization in statclock() and mi_switch(). This was previously supported by sched_lock which is going away. All modifications to rusage are now done in the context of the owning thread. reads proceed without locks. - Aggregate exiting threads rusage in thread_exit() such that the exiting thread's rusage is not lost. - Provide a new routine, rufetch() to fetch an aggregate of all rusage structures from all threads in a process. This routine must be used in any place requiring a rusage from a process prior to it's exit. The exited process's rusage is still available via p_ru. - Aggregate tick statistics only on demand via rufetch() or when a thread exits. Tick statistics are kept in the thread and protected by sched_lock until it exits. Initial patch by: attilio Reviewed by: attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
void
lim_fork(struct proc *p1, struct proc *p2)
{
p2->p_limit = lim_hold(p1->p_limit);
callout_init_mtx(&p2->p_limco, &p2->p_mtx, 0);
if (p1->p_cpulimit != RLIM_INFINITY)
callout_reset(&p2->p_limco, hz, lim_cb, p2);
}
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
void
lim_free(limp)
struct plimit *limp;
{
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
KASSERT(limp->pl_refcnt > 0, ("plimit refcnt underflow"));
Use the reference count API to manage the reference counts for process limit structures rather than using pool mutexes to protect the reference counts. Tested on: i386, alpha, sparc64
2005-09-27 18:07:05 +00:00
if (refcount_release(&limp->pl_refcnt))
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
free((void *)limp, M_PLIMIT);
}
/*
* Make a copy of the plimit structure.
* We share these structures copy-on-write after fork.
*/
void
lim_copy(dst, src)
struct plimit *dst, *src;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
KASSERT(dst->pl_refcnt == 1, ("lim_copy to shared limit"));
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
bcopy(src->pl_rlimit, dst->pl_rlimit, sizeof(src->pl_rlimit));
}
/*
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
* Return the hard limit for a particular system resource. The
* which parameter specifies the index into the rlimit array.
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
*/
rlim_t
lim_max(struct proc *p, int which)
{
struct rlimit rl;
lim_rlimit(p, which, &rl);
return (rl.rlim_max);
}
/*
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
* Return the current (soft) limit for a particular system resource.
* The which parameter which specifies the index into the rlimit array
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
*/
rlim_t
lim_cur(struct proc *p, int which)
{
struct rlimit rl;
lim_rlimit(p, which, &rl);
Argh! Fix a bogon. lim_cur() was returning the hard (max) limit rather than the soft (cur) limit. Submitted by: bde
2004-02-11 18:04:13 +00:00
return (rl.rlim_cur);
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
}
/*
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
* Return a copy of the entire rlimit structure for the system limit
* specified by 'which' in the rlimit structure pointed to by 'rlp'.
Locking for the per-process resource limits structure. - struct plimit includes a mutex to protect a reference count. The plimit structure is treated similarly to struct ucred in that is is always copy on write, so having a reference to a structure is sufficient to read from it without needing a further lock. - The proc lock protects the p_limit pointer and must be held while reading limits from a process to keep the limit structure from changing out from under you while reading from it. - Various global limits that are ints are not protected by a lock since int writes are atomic on all the archs we support and thus a lock wouldn't buy us anything. - All accesses to individual resource limits from a process are abstracted behind a simple lim_rlimit(), lim_max(), and lim_cur() API that return either an rlimit, or the current or max individual limit of the specified resource from a process. - dosetrlimit() was renamed to kern_setrlimit() to match existing style of other similar syscall helper functions. - The alpha OSF/1 compat layer no longer calls getrlimit() and setrlimit() (it didn't used the stackgap when it should have) but uses lim_rlimit() and kern_setrlimit() instead. - The svr4 compat no longer uses the stackgap for resource limits calls, but uses lim_rlimit() and kern_setrlimit() instead. - The ibcs2 compat no longer uses the stackgap for resource limits. It also no longer uses the stackgap for accessing sysctl's for the ibcs2_sysconf() syscall but uses kernel_sysctl() instead. As a result, ibcs2_sysconf() no longer needs Giant. - The p_rlimit macro no longer exists. Submitted by: mtm (mostly, I only did a few cleanups and catchups) Tested on: i386 Compiled on: alpha, amd64
2004-02-04 21:52:57 +00:00
*/
void
lim_rlimit(struct proc *p, int which, struct rlimit *rlp)
{
PROC_LOCK_ASSERT(p, MA_OWNED);
KASSERT(which >= 0 && which < RLIM_NLIMITS,
("request for invalid resource limit"));
*rlp = p->p_limit->pl_rlimit[which];
Rework the support for ABIs to override resource limits (used by 32-bit processes under 64-bit kernels). Previously, each 32-bit process overwrote its resource limits at exec() time. The problem with this approach is that the new limits affect all child processes of the 32-bit process, including if the child process forks and execs a 64-bit process. To fix this, don't ovewrite the resource limits during exec(). Instead, sv_fixlimits() is now replaced with a different function sv_fixlimit() which asks the ABI to sanitize a single resource limit. We then use this when querying and setting resource limits. Thus, if a 32-bit process sets a limit, then that new limit will be inherited by future children. However, if the 32-bit process doesn't change a limit, then a future 64-bit child will see the "full" 64-bit limit rather than the 32-bit limit. MFC is tentative since it will break the ABI of old linux.ko modules (no other modules are affected). MFC after: 1 week
2007-05-14 22:40:04 +00:00
if (p->p_sysent->sv_fixlimit != NULL)
p->p_sysent->sv_fixlimit(rlp, which);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
void
uihashinit()
{
Make uidinfo subsystem mpsafe use a mutex lock when looking up/deleting entries on the hashlist use a mutex lock on each uidinfo when updating fields make uifree() a void function rather than 'int' since no one cares allocate uidinfo structs with the M_ZERO flag and don't explicitly initialize them Assisted by: eivind, jhb, jakeb
2000-11-26 12:08:17 +00:00
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
uihashtbl = hashinit(maxproc / 16, M_UIDINFO, &uihash);
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
rw_init(&uihashtbl_lock, "uidinfo hash");
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
}
Make uidinfo subsystem mpsafe use a mutex lock when looking up/deleting entries on the hashlist use a mutex lock on each uidinfo when updating fields make uifree() a void function rather than 'int' since no one cares allocate uidinfo structs with the M_ZERO flag and don't explicitly initialize them Assisted by: eivind, jhb, jakeb
2000-11-26 12:08:17 +00:00
/*
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
* Look up a uidinfo struct for the parameter uid.
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
* uihashtbl_lock must be locked.
Make uidinfo subsystem mpsafe use a mutex lock when looking up/deleting entries on the hashlist use a mutex lock on each uidinfo when updating fields make uifree() a void function rather than 'int' since no one cares allocate uidinfo structs with the M_ZERO flag and don't explicitly initialize them Assisted by: eivind, jhb, jakeb
2000-11-26 12:08:17 +00:00
*/
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
static struct uidinfo *
uilookup(uid)
uid_t uid;
{
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
struct uihashhead *uipp;
struct uidinfo *uip;
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
rw_assert(&uihashtbl_lock, RA_LOCKED);
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
uipp = UIHASH(uid);
LIST_FOREACH(uip, uipp, ui_hash)
if (uip->ui_uid == uid)
break;
return (uip);
}
Make uidinfo subsystem mpsafe use a mutex lock when looking up/deleting entries on the hashlist use a mutex lock on each uidinfo when updating fields make uifree() a void function rather than 'int' since no one cares allocate uidinfo structs with the M_ZERO flag and don't explicitly initialize them Assisted by: eivind, jhb, jakeb
2000-11-26 12:08:17 +00:00
/*
Translate alfred to english. Submitted by: bde
2000-12-01 06:59:18 +00:00
* Find or allocate a struct uidinfo for a particular uid.
* Increase refcount on uidinfo struct returned.
Make uidinfo subsystem mpsafe use a mutex lock when looking up/deleting entries on the hashlist use a mutex lock on each uidinfo when updating fields make uifree() a void function rather than 'int' since no one cares allocate uidinfo structs with the M_ZERO flag and don't explicitly initialize them Assisted by: eivind, jhb, jakeb
2000-11-26 12:08:17 +00:00
* uifree() should be called on a struct uidinfo when released.
*/
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
struct uidinfo *
uifind(uid)
uid_t uid;
{
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
struct uidinfo *old_uip, *uip;
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
rw_rlock(&uihashtbl_lock);
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
uip = uilookup(uid);
Don't call malloc with M_WAITOK while holding a mutex.
2001-03-09 18:40:34 +00:00
if (uip == NULL) {
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
rw_runlock(&uihashtbl_lock);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
uip = malloc(sizeof(*uip), M_UIDINFO, M_WAITOK | M_ZERO);
Add racct. It's an API to keep per-process, per-jail, per-loginclass and per-loginclass resource accounting information, to be used by the new resource limits code. It's connected to the build, but the code that actually calls the new functions will come later. Sponsored by: The FreeBSD Foundation Reviewed by: kib (earlier version)
2011-03-29 17:47:25 +00:00
racct_create(&uip->ui_racct);
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
rw_wlock(&uihashtbl_lock);
Don't call malloc with M_WAITOK while holding a mutex.
2001-03-09 18:40:34 +00:00
/*
* There's a chance someone created our uidinfo while we
* were in malloc and not holding the lock, so we have to
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
* make sure we don't insert a duplicate uidinfo.
Don't call malloc with M_WAITOK while holding a mutex.
2001-03-09 18:40:34 +00:00
*/
if ((old_uip = uilookup(uid)) != NULL) {
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
/* Someone else beat us to it. */
Add racct. It's an API to keep per-process, per-jail, per-loginclass and per-loginclass resource accounting information, to be used by the new resource limits code. It's connected to the build, but the code that actually calls the new functions will come later. Sponsored by: The FreeBSD Foundation Reviewed by: kib (earlier version)
2011-03-29 17:47:25 +00:00
racct_destroy(&uip->ui_racct);
Don't call malloc with M_WAITOK while holding a mutex.
2001-03-09 18:40:34 +00:00
free(uip, M_UIDINFO);
uip = old_uip;
} else {
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
refcount_init(&uip->ui_ref, 0);
Don't call malloc with M_WAITOK while holding a mutex.
2001-03-09 18:40:34 +00:00
uip->ui_uid = uid;
Implement global and per-uid accounting of the anonymous memory. Add rlimit RLIMIT_SWAP that limits the amount of swap that may be reserved for the uid. The accounting information (charge) is associated with either map entry, or vm object backing the entry, assuming the object is the first one in the shadow chain and entry does not require COW. Charge is moved from entry to object on allocation of the object, e.g. during the mmap, assuming the object is allocated, or on the first page fault on the entry. It moves back to the entry on forks due to COW setup. The per-entry granularity of accounting makes the charge process fair for processes that change uid during lifetime, and decrements charge for proper uid when region is unmapped. The interface of vm_pager_allocate(9) is extended by adding struct ucred *, that is used to charge appropriate uid when allocation if performed by kernel, e.g. md(4). Several syscalls, among them is fork(2), may now return ENOMEM when global or per-uid limits are enforced. In collaboration with: pho Reviewed by: alc Approved by: re (kensmith)
2009-06-23 20:45:22 +00:00
mtx_init(&uip->ui_vmsize_mtx, "ui_vmsize", NULL,
MTX_DEF);
Don't call malloc with M_WAITOK while holding a mutex.
2001-03-09 18:40:34 +00:00
LIST_INSERT_HEAD(UIHASH(uid), uip, ui_hash);
}
}
Make uidinfo subsystem mpsafe use a mutex lock when looking up/deleting entries on the hashlist use a mutex lock on each uidinfo when updating fields make uifree() a void function rather than 'int' since no one cares allocate uidinfo structs with the M_ZERO flag and don't explicitly initialize them Assisted by: eivind, jhb, jakeb
2000-11-26 12:08:17 +00:00
uihold(uip);
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
rw_unlock(&uihashtbl_lock);
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
return (uip);
}
use a oppurtunistic locking strategy with the uidinfo structures to avoid locking the global hash on each uifree() make struct uidinfo only visible to the kernel make uihold() a function rather than a macro to reduce bloat swap the order of a spl/mutex to maintain consistancy
2000-11-30 19:15:22 +00:00
/*
Translate alfred to english. Submitted by: bde
2000-12-01 06:59:18 +00:00
* Place another refcount on a uidinfo struct.
use a oppurtunistic locking strategy with the uidinfo structures to avoid locking the global hash on each uifree() make struct uidinfo only visible to the kernel make uihold() a function rather than a macro to reduce bloat swap the order of a spl/mutex to maintain consistancy
2000-11-30 19:15:22 +00:00
*/
void
uihold(uip)
Translate alfred to english. Submitted by: bde
2000-12-01 06:59:18 +00:00
struct uidinfo *uip;
use a oppurtunistic locking strategy with the uidinfo structures to avoid locking the global hash on each uifree() make struct uidinfo only visible to the kernel make uihold() a function rather than a macro to reduce bloat swap the order of a spl/mutex to maintain consistancy
2000-11-30 19:15:22 +00:00
{
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
refcount_acquire(&uip->ui_ref);
use a oppurtunistic locking strategy with the uidinfo structures to avoid locking the global hash on each uifree() make struct uidinfo only visible to the kernel make uihold() a function rather than a macro to reduce bloat swap the order of a spl/mutex to maintain consistancy
2000-11-30 19:15:22 +00:00
}
Revert r210225 - turns out I was wrong; the "/*-" is not license-only thing; it's also used to indicate that the comment should not be automatically rewrapped. Explained by: cperciva@
2010-07-18 20:57:53 +00:00
/*-
Translate alfred to english. Submitted by: bde
2000-12-01 06:59:18 +00:00
* Since uidinfo structs have a long lifetime, we use an
use a oppurtunistic locking strategy with the uidinfo structures to avoid locking the global hash on each uifree() make struct uidinfo only visible to the kernel make uihold() a function rather than a macro to reduce bloat swap the order of a spl/mutex to maintain consistancy
2000-11-30 19:15:22 +00:00
* opportunistic refcounting scheme to avoid locking the lookup hash
* for each release.
*
Translate alfred to english. Submitted by: bde
2000-12-01 06:59:18 +00:00
* If the refcount hits 0, we need to free the structure,
use a oppurtunistic locking strategy with the uidinfo structures to avoid locking the global hash on each uifree() make struct uidinfo only visible to the kernel make uihold() a function rather than a macro to reduce bloat swap the order of a spl/mutex to maintain consistancy
2000-11-30 19:15:22 +00:00
* which means we need to lock the hash.
Translate alfred to english. Submitted by: bde
2000-12-01 06:59:18 +00:00
* Optimal case:
* After locking the struct and lowering the refcount, if we find
* that we don't need to free, simply unlock and return.
* Suboptimal case:
* If refcount lowering results in need to free, bump the count
Universally adopt most conventional spelling of acquire.
2007-05-27 20:50:23 +00:00
* back up, lose the lock and acquire the locks in the proper
Translate alfred to english. Submitted by: bde
2000-12-01 06:59:18 +00:00
* order to try again.
Make uidinfo subsystem mpsafe use a mutex lock when looking up/deleting entries on the hashlist use a mutex lock on each uidinfo when updating fields make uifree() a void function rather than 'int' since no one cares allocate uidinfo structs with the M_ZERO flag and don't explicitly initialize them Assisted by: eivind, jhb, jakeb
2000-11-26 12:08:17 +00:00
*/
void
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
uifree(uip)
Translate alfred to english. Submitted by: bde
2000-12-01 06:59:18 +00:00
struct uidinfo *uip;
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
{
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
int old;
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
Translate alfred to english. Submitted by: bde
2000-12-01 06:59:18 +00:00
/* Prepare for optimal case. */
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
old = uip->ui_ref;
if (old > 1 && atomic_cmpset_int(&uip->ui_ref, old, old - 1))
use a oppurtunistic locking strategy with the uidinfo structures to avoid locking the global hash on each uifree() make struct uidinfo only visible to the kernel make uihold() a function rather than a macro to reduce bloat swap the order of a spl/mutex to maintain consistancy
2000-11-30 19:15:22 +00:00
return;
Translate alfred to english. Submitted by: bde
2000-12-01 06:59:18 +00:00
/* Prepare for suboptimal case. */
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
rw_wlock(&uihashtbl_lock);
if (refcount_release(&uip->ui_ref)) {
Add racct. It's an API to keep per-process, per-jail, per-loginclass and per-loginclass resource accounting information, to be used by the new resource limits code. It's connected to the build, but the code that actually calls the new functions will come later. Sponsored by: The FreeBSD Foundation Reviewed by: kib (earlier version)
2011-03-29 17:47:25 +00:00
racct_destroy(&uip->ui_racct);
use a oppurtunistic locking strategy with the uidinfo structures to avoid locking the global hash on each uifree() make struct uidinfo only visible to the kernel make uihold() a function rather than a macro to reduce bloat swap the order of a spl/mutex to maintain consistancy
2000-11-30 19:15:22 +00:00
LIST_REMOVE(uip, ui_hash);
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
rw_wunlock(&uihashtbl_lock);
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
if (uip->ui_sbsize != 0)
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
printf("freeing uidinfo: uid = %d, sbsize = %ld\n",
uip->ui_uid, uip->ui_sbsize);
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
if (uip->ui_proccnt != 0)
Add new line character to debugging printf's.
2000-09-18 17:03:03 +00:00
printf("freeing uidinfo: uid = %d, proccnt = %ld\n",
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
uip->ui_uid, uip->ui_proccnt);
Implement global and per-uid accounting of the anonymous memory. Add rlimit RLIMIT_SWAP that limits the amount of swap that may be reserved for the uid. The accounting information (charge) is associated with either map entry, or vm object backing the entry, assuming the object is the first one in the shadow chain and entry does not require COW. Charge is moved from entry to object on allocation of the object, e.g. during the mmap, assuming the object is allocated, or on the first page fault on the entry. It moves back to the entry on forks due to COW setup. The per-entry granularity of accounting makes the charge process fair for processes that change uid during lifetime, and decrements charge for proper uid when region is unmapped. The interface of vm_pager_allocate(9) is extended by adding struct ucred *, that is used to charge appropriate uid when allocation if performed by kernel, e.g. md(4). Several syscalls, among them is fork(2), may now return ENOMEM when global or per-uid limits are enforced. In collaboration with: pho Reviewed by: alc Approved by: re (kensmith)
2009-06-23 20:45:22 +00:00
if (uip->ui_vmsize != 0)
printf("freeing uidinfo: uid = %d, swapuse = %lld\n",
uip->ui_uid, (unsigned long long)uip->ui_vmsize);
mtx_destroy(&uip->ui_vmsize_mtx);
Retire the MALLOC and FREE macros. They are an abomination unto style(9). MFC after: 3 months
2008-10-23 15:53:51 +00:00
free(uip, M_UIDINFO);
Make uidinfo subsystem mpsafe use a mutex lock when looking up/deleting entries on the hashlist use a mutex lock on each uidinfo when updating fields make uifree() a void function rather than 'int' since no one cares allocate uidinfo structs with the M_ZERO flag and don't explicitly initialize them Assisted by: eivind, jhb, jakeb
2000-11-26 12:08:17 +00:00
return;
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
}
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
/*
* Someone added a reference between atomic_cmpset_int() and
* rw_wlock(&uihashtbl_lock).
*/
rw_wunlock(&uihashtbl_lock);
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
}
Add racct. It's an API to keep per-process, per-jail, per-loginclass and per-loginclass resource accounting information, to be used by the new resource limits code. It's connected to the build, but the code that actually calls the new functions will come later. Sponsored by: The FreeBSD Foundation Reviewed by: kib (earlier version)
2011-03-29 17:47:25 +00:00
void
ui_racct_foreach(void (*callback)(struct racct *racct,
void *arg2, void *arg3), void *arg2, void *arg3)
{
struct uidinfo *uip;
struct uihashhead *uih;
rw_rlock(&uihashtbl_lock);
for (uih = &uihashtbl[uihash]; uih >= uihashtbl; uih--) {
LIST_FOREACH(uip, uih, ui_hash) {
(callback)(uip->ui_racct, arg2, arg3);
}
}
rw_runlock(&uihashtbl_lock);
}
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
/*
* Change the count associated with number of processes
* a given user is using. When 'max' is 0, don't enforce a limit
*/
int
chgproccnt(uip, diff, max)
struct uidinfo *uip;
int diff;
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
rlim_t max;
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
{
Make uidinfo subsystem mpsafe use a mutex lock when looking up/deleting entries on the hashlist use a mutex lock on each uidinfo when updating fields make uifree() a void function rather than 'int' since no one cares allocate uidinfo structs with the M_ZERO flag and don't explicitly initialize them Assisted by: eivind, jhb, jakeb
2000-11-26 12:08:17 +00:00
- A lot of style and whitespace fixes. - Update a few comments regarding locking notes. Submitted by: bde (1, mostly)
2004-02-05 20:53:25 +00:00
/* Don't allow them to exceed max, but allow subtraction. */
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
if (diff > 0 && max != 0) {
if (atomic_fetchadd_long(&uip->ui_proccnt, (long)diff) + diff > max) {
atomic_subtract_long(&uip->ui_proccnt, (long)diff);
return (0);
}
} else {
atomic_add_long(&uip->ui_proccnt, (long)diff);
if (uip->ui_proccnt < 0)
printf("negative proccnt for uid = %d\n", uip->ui_uid);
Make uidinfo subsystem mpsafe use a mutex lock when looking up/deleting entries on the hashlist use a mutex lock on each uidinfo when updating fields make uifree() a void function rather than 'int' since no one cares allocate uidinfo structs with the M_ZERO flag and don't explicitly initialize them Assisted by: eivind, jhb, jakeb
2000-11-26 12:08:17 +00:00
}
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
return (1);
}
/*
* Change the total socket buffer size a user has used.
*/
int
chgsbsize(uip, hiwat, to, max)
struct uidinfo *uip;
Widen struct sockbuf's sb_timeo member to int from short. With non-default but reasonable values of hz this member overflowed, breaking NFS over UDP. Also, as long as I'm plowing up struct sockbuf ... Change certain members from u_long/long to u_int/int in order to reduce wasted space on 64-bit machines. This change was requested by Andrew Gallatin. Netstat and systat need to be rebuilt. I am incrementing __FreeBSD_version in case any ports need to change.
2002-07-24 03:02:43 +00:00
u_int *hiwat;
u_int to;
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
rlim_t max;
{
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
int diff;
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
- Use wait-free method to manage ui_sbsize and ui_proccnt fields in the uidinfo structure. This entirely removes contention observed on the ui_mtxp mutex (as it is now gone). - Convert the uihashtbl_mtx mutex to a rwlock, as most of the time we just need to read-lock it. Reviewed by: jhb, jeff, kris & others Tested by: kris
2008-03-16 21:29:02 +00:00
diff = to - *hiwat;
if (diff > 0) {
if (atomic_fetchadd_long(&uip->ui_sbsize, (long)diff) + diff > max) {
atomic_subtract_long(&uip->ui_sbsize, (long)diff);
return (0);
}
} else {
atomic_add_long(&uip->ui_sbsize, (long)diff);
if (uip->ui_sbsize < 0)
printf("negative sbsize for uid = %d\n", uip->ui_uid);
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
}
Whitespace cleanups.
2008-03-16 21:32:20 +00:00
*hiwat = to;
return (1);
Remove uidinfo hash table lookup and maintenance out of chgproccnt() and chgsbsize(), which are called rather frequently and may be called from an interrupt context in the case of chgsbsize(). Instead, do the hash table lookup and maintenance when credentials are changed, which is a lot less frequent. Add pointers to the uidinfo structures to the ucred and pcred structures for fast access. Pass a pointer to the credential to chgproccnt() and chgsbsize() instead of passing the uid. Add a reference count to the uidinfo structure and use it to decide when to free the structure rather than freeing the structure when the resource consumption drops to zero. Move the resource tracking code from kern_proc.c to kern_resource.c. Move some duplicate code sequences in kern_prot.c to separate helper functions. Change KASSERTs in this code to unconditional tests and calls to panic().
2000-09-05 22:11:13 +00:00
}
Integrate the new MPSAFE TTY layer to the FreeBSD operating system. The last half year I've been working on a replacement TTY layer for the FreeBSD kernel. The new TTY layer was designed to improve the following: - Improved driver model: The old TTY layer has a driver model that is not abstract enough to make it friendly to use. A good example is the output path, where the device drivers directly access the output buffers. This means that an in-kernel PPP implementation must always convert network buffers into TTY buffers. If a PPP implementation would be built on top of the new TTY layer (still needs a hooks layer, though), it would allow the PPP implementation to directly hand the data to the TTY driver. - Improved hotplugging: With the old TTY layer, it isn't entirely safe to destroy TTY's from the system. This implementation has a two-step destructing design, where the driver first abandons the TTY. After all threads have left the TTY, the TTY layer calls a routine in the driver, which can be used to free resources (unit numbers, etc). The pts(4) driver also implements this feature, which means posix_openpt() will now return PTY's that are created on the fly. - Improved performance: One of the major improvements is the per-TTY mutex, which is expected to improve scalability when compared to the old Giant locking. Another change is the unbuffered copying to userspace, which is both used on TTY device nodes and PTY masters. Upgrading should be quite straightforward. Unlike previous versions, existing kernel configuration files do not need to be changed, except when they reference device drivers that are listed in UPDATING. Obtained from: //depot/projects/mpsafetty/... Approved by: philip (ex-mentor) Discussed: on the lists, at BSDCan, at the DevSummit Sponsored by: Snow B.V., the Netherlands dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
/*
* Change the count associated with number of pseudo-terminals
* a given user is using. When 'max' is 0, don't enforce a limit
*/
int
chgptscnt(uip, diff, max)
struct uidinfo *uip;
int diff;
rlim_t max;
{
/* Don't allow them to exceed max, but allow subtraction. */
if (diff > 0 && max != 0) {
if (atomic_fetchadd_long(&uip->ui_ptscnt, (long)diff) + diff > max) {
atomic_subtract_long(&uip->ui_ptscnt, (long)diff);
return (0);
}
} else {
atomic_add_long(&uip->ui_ptscnt, (long)diff);
if (uip->ui_ptscnt < 0)
printf("negative ptscnt for uid = %d\n", uip->ui_uid);
}
return (1);
}
Reference in New Issue Copy Permalink