Commit Graph

231 Commits

Author SHA1 Message Date
Alan Cox
b6e48e0372 - Acquire the vm_object's lock when performing vm_object_page_clean().
- Add a parameter to vm_pageout_flush() that tells vm_pageout_flush()
   whether its caller has locked the vm_object.  (This is a temporary
   measure to bootstrap vm_object locking.)
2003-04-24 04:31:25 +00:00
John Baldwin
897ecacd64 Lock the proc to check p_flag and several other related tests in
vm_daemon().  We don't need to hold sched_lock as long now as a result.
2003-04-22 20:03:08 +00:00
Alan Cox
72ba747d16 - Lock the vm_object when performing vm_object_pip_wakeup().
- Merge two identical cases in a switch statement.
2003-04-20 20:37:14 +00:00
Alan Cox
d22bc7101c - Lock the vm_object when performing vm_object_pip_add(). 2003-04-20 03:41:21 +00:00
Wes Peters
f4cf2141f6 Add a facility allowing processes to inform the VM subsystem they are
critical and should not be killed when pageout is looking for more
memory pages in all the wrong places.

Reviewed by:	arch@
Sponsored by:	St. Bernard Software
2003-03-31 21:09:57 +00:00
David Schultz
72d97679ff - When the VM daemon is out of swap space and looking for a
process to kill, don't block on a map lock while holding the
  process lock.  Instead, skip processes whose map locks are held
  and find something else to kill.
- Add vm_map_trylock_read() to support the above.

Reviewed by:	alc, mike (mentor)
2003-03-12 23:13:16 +00:00
Alan Cox
6b4b77ad34 Add a comment describing how pagedaemon_wakeup() should be used and
synchronized.

Suggested by:	tegge
2003-02-09 20:40:36 +00:00
Alan Cox
a1c0a78518 - It's more accurate to say that vm_paging_needed() returns TRUE
than a positive number.
 - In pagedaemon_wakeup(), set vm_pages_needed to 1 rather than
   incrementing it to accomplish the same.
2003-02-02 07:16:40 +00:00
Alan Cox
8e1d8de578 - Convert vm_pageout()'s tsleep()s to msleep()s with the page queue lock. 2003-02-02 01:11:21 +00:00
Alan Cox
8b24576748 - Remove (some) unnecessary explicit initializations to zero.
- Style changes to vm_pageout(): declarations and white-space.
2003-02-01 21:55:30 +00:00
Alan Cox
b0ef8c5fe4 - Update vm_pageout_deficit using atomic operations. It's a simple
counter outside the scope of existing locks.
 - Eliminate a redundant clearing of vm_pageout_deficit.
2003-01-14 06:57:03 +00:00
Alan Cox
ff2023a5df Make vm_pageout_page_free() static. 2003-01-14 02:28:39 +00:00
Poul-Henning Kamp
c410df597b Avoid extern decls in .c files by putting them in the vm/swap_pager.h
include file where they belong.
Share the dmmax_mask variable.
2003-01-03 14:30:46 +00:00
Matthew Dillon
40bb4f4bcf vm_pager_put_pages() takes VM_PAGER_* flags, not OBJPC_* flags. It just
so happens that OBJPC_SYNC has the same value as VM_PAGER_PUT_SYNC so no
harm done.  But fix it :-)

No operational changes.

MFC after:	1 day
2002-12-28 21:15:39 +00:00
Alan Cox
b365ea9e30 Hold the page queues lock when performing vm_page_flag_set(). 2002-12-18 04:02:02 +00:00
Alan Cox
38857e7f73 Hold the page queues lock when calling pmap_protect(); it updates fields
of the vm_page structure.  Nearby, remove an unnecessary semicolon and
return statement.

Approved by:	re (blanket)
2002-12-01 05:40:18 +00:00
Alan Cox
78f7187d01 Increase the scope of the page queue lock in vm_pageout_scan().
Approved by:	re (blanket)
2002-12-01 00:02:39 +00:00
Alan Cox
ba0208b945 Assert that the page queues lock rather than Giant is held in
vm_pageout_page_free().

Approved by:	re
2002-11-23 08:08:54 +00:00
Jeff Roberson
855a310fcb - Add an event that is triggered when the system is low on memory. This is
intended to be used by significant memory consumers so that they may drain
   some of their caches.

Inspired by:	phk
Approved by:	re
Tested on:	x86, alpha
2002-11-21 09:17:56 +00:00
Alan Cox
a12cc0e489 Remove vm_page_protect(). Instead, use pmap_page_protect() directly. 2002-11-18 04:05:22 +00:00
Alan Cox
4fec79bef8 Now that pmap_remove_all() is exported by our pmap implementations
use it directly.
2002-11-16 07:44:25 +00:00
Alan Cox
eea85e9bb6 Move pmap_collect() out of the machine-dependent code, rename it
to reflect its new location, and add page queue and flag locking.

Notes: (1) alpha, i386, and ia64 had identical implementations
of pmap_collect() in terms of machine-independent interfaces;
(2) sparc64 doesn't require it; (3) powerpc had it as a TODO.
2002-11-13 05:39:58 +00:00
Alan Cox
d154fb4fe6 When prot is VM_PROT_NONE, call pmap_page_protect() directly rather than
indirectly through vm_page_protect().  The one remaining page flag that
is updated by vm_page_protect() is already being updated by our various
pmap implementations.

Note: A later commit will similarly change the VM_PROT_READ case and
eliminate vm_page_protect().
2002-11-10 07:12:04 +00:00
Jeff Roberson
b43179fbe8 - 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
Jeff Roberson
3ef3e7c42b - Get rid of the unused LK_NOOBJ. 2002-09-25 01:24:58 +00:00
Jake Burkholder
05ba50f522 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
Julian Elischer
71fad9fdee Completely redo thread states.
Reviewed by:	davidxu@freebsd.org
2002-09-11 08:13:56 +00:00
Alan Cox
67ef391e00 o Lock page queue accesses by vm_page_activate(). 2002-08-10 23:53:59 +00:00
Alan Cox
299018d3b6 o Lock page queue accesses by vm_page_free().
o Increment cnt.v_dfree inside vm_pageout_page_free() rather than
   at each call.
2002-07-28 05:46:47 +00:00
Alan Cox
55df3298c6 o Require that the page queues lock is held on entry to vm_pageout_clean()
and vm_pageout_flush().
 o Acquire the page queues lock before calling vm_pageout_clean()
   or vm_pageout_flush().
2002-07-27 23:20:32 +00:00
Alan Cox
ce18aebde4 o Lock page queue accesses by vm_page_activate() and vm_page_deactivate()
in vm_pageout_object_deactivate_pages().
 o Apply some style fixes to vm_pageout_object_deactivate_pages().
2002-07-27 06:41:03 +00:00
Alan Cox
8ffc151979 o Extend the scope of the page queues lock in vm_pageout_scan()
to cover the traversal of the cache queue.
2002-07-23 02:42:25 +00:00
Alan Cox
40eab1e944 o Lock page queue accesses by vm_page_try_to_cache(). (The accesses
in kern/vfs_bio.c are already locked.)
 o Assert that the page queues lock is held in vm_page_try_to_cache().
2002-07-20 20:58:46 +00:00
Alan Cox
15a5d2108e o Lock page queue accesses by vm_page_cache() in vm_fault() and
vm_pageout_scan().  (The others are already locked.)
 o Assert that the page queues lock is held in vm_page_cache().
2002-07-20 19:34:21 +00:00
Alan Cox
48c0444c98 o Lock accesses to the active page queue in vm_pageout_scan() and
vm_pageout_page_stats().
2002-07-20 18:45:25 +00:00
Julian Elischer
e602ba25fd Part 1 of KSE-III
The ability to schedule multiple threads per process
(one one cpu) by making ALL system calls optionally asynchronous.
to come: ia64 and power-pc patches, patches for gdb, test program (in tools)

Reviewed by:	Almost everyone who counts
	(at various times, peter, jhb, matt, alfred, mini, bernd,
	and a cast of thousands)

	NOTE: this is still Beta code, and contains lots of debugging stuff.
	expect slight instability in signals..
2002-06-29 17:26:22 +00:00
Alan Cox
d974f03c69 o Introduce and use vm_map_trylock() to replace several direct uses
of lockmgr().
 o Add missing synchronization to vmspace_swap_count(): Obtain a read lock
   on the vm_map before traversing it.
2002-04-28 06:07:54 +00:00
Jeff Roberson
670d17b5c0 Remove references to vm_zone.h and switch over to the new uma API. 2002-03-20 04:02:59 +00:00
Alfred Perlstein
11caded34f Remove __P. 2002-03-19 22:20:14 +00:00
Jeff Roberson
8355f576a9 This is the first part of the new kernel memory allocator. This replaces
malloc(9) and vm_zone with a slab like allocator.

Reviewed by:	arch@
2002-03-19 09:11:49 +00:00
Brian Feldman
25adb370be Back out the modification of vm_map locks from lockmgr to sx locks. The
best path forward now is likely to change the lockmgr locks to simple
sleep mutexes, then see if any extra contention it generates is greater
than removed overhead of managing local locking state information,
cost of extra calls into lockmgr, etc.

Additionally, making the vm_map lock a mutex and respecting it properly
will put us much closer to not needing Giant magic in vm.
2002-03-18 15:08:09 +00:00
Brian Feldman
0e0af8ecda Rename SI_SUB_MUTEX to SI_SUB_MTX_POOL to make the name at all accurate.
While doing this, move it earlier in the sysinit boot process so that the
VM system can use it.

After that, the system is now able to use sx locks instead of lockmgr
locks in the VM system.  To accomplish this, some of the more
questionable uses of the locks (such as testing whether they are
owned or not, as well as allowing shared+exclusive recursion) are
removed, and simpler logic throughout is used so locks should also be
easier to understand.

This has been tested on my laptop for months, and has not shown any
problems on SMP systems, either, so appears quite safe.  One more
user of lockmgr down, many more to go :)
2002-03-13 23:48:08 +00:00
Eivind Eklund
a128794977 - Remove a number of extra newlines that do not belong here according to
style(9)
- Minor space adjustment in cases where we have "( ", " )", if(), return(),
  while(), for(), etc.
- Add /* SYMBOL */ after a few #endifs.

Reviewed by:	alc
2002-03-10 21:52:48 +00:00
Mike Silbersack
7f3a40933b Fix a horribly suboptimal algorithm in the vm_daemon.
In order to determine what to page out, the vm_daemon checks
reference bits on all pages belonging to all processes.  Unfortunately,
the algorithm used reacted badly with shared pages; each shared page
would be checked once per process sharing it; this caused an O(N^2)
growth of tlb invalidations.  The algorithm has been changed so that
each page will be checked only 16 times.

Prior to this change, a fork/sleepbomb of 1300 processes could cause
the vm_daemon to take over 60 seconds to complete, effectively
freezing the system for that time period.  With this change
in place, the vm_daemon completes in less than a second.  Any system
with hundreds of processes sharing pages should benefit from this change.

Note that the vm_daemon is only run when the system is under extreme
memory pressure.  It is likely that many people with loaded systems saw
no symptoms of this problem until they reached the point where swapping
began.

Special thanks go to dillon, peter, and Chuck Cranor, who helped me
get up to speed with vm internals.

PR:		33542, 20393
Reviewed by:	dillon
MFC after:	1 week
2002-02-27 18:03:02 +00:00
Mike Silbersack
ef6020d187 Changes to make the OOM killer much more effective:
- Allow the OOM killer to target processes currently locked in
  memory.  These very often are the ones doing the memory hogging.
- Drop the wakeup priority of processes currently sleeping while
  waiting for their page fault to complete.  In order for the OOM
  killer to work well, the killed process and other system processes
  waiting on memory must be allowed to wakeup first.

Reviewed by:	dillon
MFC after:	1 week
2002-02-19 18:34:02 +00:00
Matthew Dillon
027df6bdd7 GC P_BUFEXHAUST leftovers, we've had a new mechanism to avoid buffer
cache lockups for over a year now.

MFC after:		0 days
2002-01-31 18:39:44 +00:00
Matthew Dillon
23b590188f Fix a BUF_TIMELOCK race against BUF_LOCK and fix a deadlock in vget()
against VM_WAIT in the pageout code.  Both fixes involve adjusting
the lockmgr's timeout capability so locks obtained with timeouts do not
interfere with locks obtained without a timeout.

Hopefully MFC: before the 4.5 release
2001-12-20 22:42:27 +00:00
Matthew Dillon
57601bcb5d Syntax cleanup and documentation, no operational changes.
MFC after:	1 day
2001-10-21 06:12:06 +00:00
Tor Egge
30105b9ec4 Don't remove all mappings of a swapped out process if the vm map contained
wired entries.  vm_fault_unwire() depends on the mapping being intact.

Reviewed by:	dillon
2001-10-14 20:51:14 +00:00
Julian Elischer
b40ce4165d 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