[BACK]Return to uvm.9 CVS log [TXT][DIR] Up to [cvs.NetBSD.org] / src / share / man / man9

File: [cvs.NetBSD.org] / src / share / man / man9 / uvm.9 (download)

Revision 1.94.2.1, Wed May 13 19:19:16 2009 UTC (14 years, 10 months ago) by jym
Branch: jym-xensuspend
Changes since 1.94: +77 -142 lines

Sync with HEAD.

Third (and last) commit. See http://mail-index.netbsd.org/source-changes/2009/05/13/msg221222.html

.\"	$NetBSD: uvm.9,v 1.94.2.1 2009/05/13 19:19:16 jym Exp $
.\"
.\" Copyright (c) 1998 Matthew R. Green
.\" All rights reserved.
.\"
.\" 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.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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.
.\"
.Dd March 12, 2009
.Dt UVM 9
.Os
.Sh NAME
.Nm uvm
.Nd virtual memory system external interface
.Sh SYNOPSIS
.In sys/param.h
.In uvm/uvm.h
.Sh DESCRIPTION
The UVM virtual memory system manages access to the computer's memory
resources.
User processes and the kernel access these resources through
UVM's external interface.
UVM's external interface includes functions that:
.Pp
.Bl -hyphen -compact
.It
initialize UVM sub-systems
.It
manage virtual address spaces
.It
resolve page faults
.It
memory map files and devices
.It
perform uio-based I/O to virtual memory
.It
allocate and free kernel virtual memory
.It
allocate and free physical memory
.El
.Pp
In addition to exporting these services, UVM has two kernel-level processes:
pagedaemon and swapper.
The pagedaemon process sleeps until physical memory becomes scarce.
When that happens, pagedaemon is awoken.
It scans physical memory, paging out and freeing memory that has not
been recently used.
The swapper process swaps in runnable processes that are currently swapped
out, if there is room.
.Pp
There are also several miscellaneous functions.
.Sh INITIALIZATION
.Bl -ohang
.It Ft void
.Fn uvm_init "void" ;
.It Ft void
.Fn uvm_init_limits "struct lwp *l" ;
.It Ft void
.Fn uvm_setpagesize "void" ;
.It Ft void
.Fn uvm_swap_init "void" ;
.El
.Pp
.Fn uvm_init
sets up the UVM system at system boot time, after the
console has been setup.
It initializes global state, the page, map, kernel virtual memory state,
machine-dependent physical map, kernel memory allocator,
pager and anonymous memory sub-systems, and then enables
paging of kernel objects.
.Pp
.Fn uvm_init_limits
initializes process limits for the named process.
This is for use by the system startup for process zero, before any
other processes are created.
.Pp
.Fn uvm_setpagesize
initializes the uvmexp members pagesize (if not already done by
machine-dependent code), pageshift and pagemask.
It should be called by machine-dependent code early in the
.Fn pmap_init
call (see
.Xr pmap 9 ) .
.Pp
.Fn uvm_swap_init
initializes the swap sub-system.
.Sh VIRTUAL ADDRESS SPACE MANAGEMENT
.Bl -ohang
.It Ft int
.Fn uvm_map "struct vm_map *map" "vaddr_t *startp" "vsize_t size" "struct uvm_object *uobj" "voff_t uoffset" "vsize_t align" "uvm_flag_t flags" ;
.It Ft void
.Fn uvm_unmap "struct vm_map *map" "vaddr_t start" "vaddr_t end" ;
.It Ft int
.Fn uvm_map_pageable "struct vm_map *map" "vaddr_t start" "vaddr_t end" "bool new_pageable" "int lockflags" ;
.It Ft bool
.Fn uvm_map_checkprot "struct vm_map *map" "vaddr_t start" "vaddr_t end" "vm_prot_t protection" ;
.It Ft int
.Fn uvm_map_protect "struct vm_map *map" "vaddr_t start" "vaddr_t end" "vm_prot_t new_prot" "bool set_max" ;
.It Ft int
.Fn uvm_deallocate "struct vm_map *map" "vaddr_t start" "vsize_t size" ;
.It Ft struct vmspace *
.Fn uvmspace_alloc "vaddr_t min" "vaddr_t max" "int pageable" ;
.It Ft void
.Fn uvmspace_exec "struct lwp *l" "vaddr_t start" "vaddr_t end" ;
.It Ft struct vmspace *
.Fn uvmspace_fork "struct vmspace *vm" ;
.It Ft void
.Fn uvmspace_free "struct vmspace *vm1" ;
.It Ft void
.Fn uvmspace_share "struct proc *p1" "struct proc *p2" ;
.It Ft void
.Fn uvmspace_unshare "struct lwp *l" ;
.It Ft bool
.Fn uvm_uarea_alloc "vaddr_t *uaddrp" ;
.It  Ft void
.Fn uvm_uarea_free "vaddr_t uaddr" ;
.El
.Pp
.Fn uvm_map
establishes a valid mapping in map
.Fa map ,
which must be unlocked.
The new mapping has size
.Fa size ,
which must be a multiple of
.Dv PAGE_SIZE .
The
.Fa uobj
and
.Fa uoffset
arguments can have four meanings.
When
.Fa uobj
is
.Dv NULL
and
.Fa uoffset
is
.Dv UVM_UNKNOWN_OFFSET ,
.Fn uvm_map
does not use the machine-dependent
.Dv PMAP_PREFER
function.
If
.Fa uoffset
is any other value, it is used as the hint to
.Dv PMAP_PREFER .
When
.Fa uobj
is not
.Dv NULL
and
.Fa uoffset
is
.Dv UVM_UNKNOWN_OFFSET ,
.Fn uvm_map
finds the offset based upon the virtual address, passed as
.Fa startp .
If
.Fa uoffset
is any other value, we are doing a normal mapping at this offset.
The start address of the map will be returned in
.Fa startp .
.Pp
.Fa align
specifies alignment of mapping unless
.Dv UVM_FLAG_FIXED
is specified in
.Fa flags .
.Fa align
must be a power of 2.
.Pp
.Fa flags
passed to
.Fn uvm_map
are typically created using the
.Fn UVM_MAPFLAG "vm_prot_t prot" "vm_prot_t maxprot" "vm_inherit_t inh" "int advice" "int flags"
macro, which uses the following values.
The
.Fa prot
and
.Fa maxprot
can take are:
.Bd -literal
#define UVM_PROT_MASK   0x07    /* protection mask */
#define UVM_PROT_NONE   0x00    /* protection none */
#define UVM_PROT_ALL    0x07    /* everything */
#define UVM_PROT_READ   0x01    /* read */
#define UVM_PROT_WRITE  0x02    /* write */
#define UVM_PROT_EXEC   0x04    /* exec */
#define UVM_PROT_R      0x01    /* read */
#define UVM_PROT_W      0x02    /* write */
#define UVM_PROT_RW     0x03    /* read-write */
#define UVM_PROT_X      0x04    /* exec */
#define UVM_PROT_RX     0x05    /* read-exec */
#define UVM_PROT_WX     0x06    /* write-exec */
#define UVM_PROT_RWX    0x07    /* read-write-exec */
.Ed
.Pp
The values that
.Fa inh
can take are:
.Bd -literal
#define UVM_INH_MASK    0x30    /* inherit mask */
#define UVM_INH_SHARE   0x00    /* "share" */
#define UVM_INH_COPY    0x10    /* "copy" */
#define UVM_INH_NONE    0x20    /* "none" */
#define UVM_INH_DONATE  0x30    /* "donate" \*[Lt]\*[Lt] not used */
.Ed
.Pp
The values that
.Fa advice
can take are:
.Bd -literal
#define UVM_ADV_NORMAL     0x0  /* 'normal' */
#define UVM_ADV_RANDOM     0x1  /* 'random' */
#define UVM_ADV_SEQUENTIAL 0x2  /* 'sequential' */
#define UVM_ADV_MASK       0x7  /* mask */
.Ed
.Pp
The values that
.Fa flags
can take are:
.Bd -literal
#define UVM_FLAG_FIXED   0x010000 /* find space */
#define UVM_FLAG_OVERLAY 0x020000 /* establish overlay */
#define UVM_FLAG_NOMERGE 0x040000 /* don't merge map entries */
#define UVM_FLAG_COPYONW 0x080000 /* set copy_on_write flag */
#define UVM_FLAG_AMAPPAD 0x100000 /* for bss: pad amap to reduce malloc() */
#define UVM_FLAG_TRYLOCK 0x200000 /* fail if we can not lock map */
.Ed
.Pp
The
.Dv UVM_MAPFLAG
macro arguments can be combined with an or operator.
There are several special purpose macros for checking protection
combinations, e.g., the
.Dv UVM_PROT_WX
macro.
There are also some additional macros to extract bits from the flags.
The
.Dv UVM_PROTECTION ,
.Dv UVM_INHERIT ,
.Dv UVM_MAXPROTECTION
and
.Dv UVM_ADVICE
macros return the protection, inheritance, maximum protection and advice,
respectively.
.Fn uvm_map
returns a standard UVM return value.
.Pp
.Fn uvm_unmap
removes a valid mapping,
from
.Fa start
to
.Fa end ,
in map
.Fa map ,
which must be unlocked.
.Pp
.Fn uvm_map_pageable
changes the pageability of the pages in the range from
.Fa start
to
.Fa end
in map
.Fa map
to
.Fa new_pageable .
.Fn uvm_map_pageable
returns a standard UVM return value.
.Pp
.Fn uvm_map_checkprot
checks the protection of the range from
.Fa start
to
.Fa end
in map
.Fa map
against
.Fa protection .
This returns either
.Dv true
or
.Dv false .
.Pp
.Fn uvm_map_protect
changes the protection
.Fa start
to
.Fa end
in map
.Fa map
to
.Fa new_prot ,
also setting the maximum protection to the region to
.Fa new_prot
if
.Fa set_max
is true.
This function returns a standard UVM return value.
.Pp
.Fn uvm_deallocate
deallocates kernel memory in map
.Fa map
from address
.Fa start
to
.Fa start + size .
.Pp
.Fn uvmspace_alloc
allocates and returns a new address space, with ranges from
.Fa min
to
.Fa max ,
setting the pageability of the address space to
.Fa pageable .
.Pp
.Fn uvmspace_exec
either reuses the address space of lwp
.Fa l
if there are no other references to it, or creates
a new one with
.Fn uvmspace_alloc .
The range of valid addresses in the address space is reset to
.Fa start
through
.Fa end .
.Pp
.Fn uvmspace_fork
creates and returns a new address space based upon the
.Fa vm1
address space, typically used when allocating an address space for a
child process.
.Pp
.Fn uvmspace_free
lowers the reference count on the address space
.Fa vm ,
freeing the data structures if there are no other references.
.Pp
.Fn uvmspace_share
causes process
.Pa p2
to share the address space of
.Fa p1 .
.Pp
.Fn uvmspace_unshare
ensures that lwp
.Fa l
has its own, unshared address space, by creating a new one if
necessary by calling
.Fn uvmspace_fork .
.Pp
.Fn uvm_uarea_alloc
allocates virtual space for a u-area (i.e., a kernel stack) and stores
its virtual address in
.Fa *uaddrp .
The return value is
.Dv true
if the u-area is already backed by wired physical memory, otherwise
.Dv false .
.Pp
.Fn uvm_uarea_free
frees a u-area allocated with
.Fn uvm_uarea_alloc ,
freeing both the virtual space and any physical pages which may have been
allocated to back that virtual space later.
.Sh PAGE FAULT HANDLING
.Bl -ohang
.It Ft int
.Fn uvm_fault "struct vm_map *orig_map" "vaddr_t vaddr" "vm_prot_t access_type" ;
.El
.Pp
.Fn uvm_fault
is the main entry point for faults.
It takes
.Fa orig_map
as the map the fault originated in, a
.Fa vaddr
offset into the map the fault occurred, and
.Fa access_type
describing the type of access requested.
.Fn uvm_fault
returns a standard UVM return value.
.Sh MEMORY MAPPING FILES AND DEVICES
.Bl -ohang
.It Ft void
.Fn uvm_vnp_setsize "struct vnode *vp" "voff_t newsize" ;
.It Ft void *
.Fn ubc_alloc "struct uvm_object *uobj" "voff_t offset" "vsize_t *lenp" \
"int advice" "int flags" ;
.It Ft void
.Fn ubc_release "void *va" "int flags" ;
.It Ft int
.Fn ubc_uiomove "struct uvm_object *uobj" "struct uio *uio" "vsize_t todo" \
"int advice" "int flags" ;
.El
.Pp
.Fn uvm_vnp_setsize
sets the size of vnode
.Fa vp
to
.Fa newsize .
Caller must hold a reference to the vnode.
If the vnode shrinks, pages no longer used are discarded.
.Pp
.Fn ubc_alloc
creates a kernel mapping of
.Fa uobj
starting at offset
.Fa offset .
The desired length of the mapping is pointed to by
.Fa lenp ,
but the actual mapping may be smaller than this.
.Fa lenp
is updated to contain the actual length mapped.
.Fa advice
is the access pattern hint, which must be one of
.Pp
.Bl -tag -offset indent -width "UVM_ADV_SEQUENTIAL" -compact
.It UVM_ADV_NORMAL
No hint
.It UVM_ADV_RANDOM
Random access hint
.It UVM_ADV_SEQUENTIAL
Sequential access hint (from lower offset to higher offset)
.El
.Pp
The possible
.Fa flags
are
.Pp
.Bl -tag -offset indent -width "UVM_ADV_SEQUENTIAL" -compact
.It UBC_READ
Mapping will be accessed for read.
.It UBC_WRITE
Mapping will be accessed for write.
.It UBC_FAULTBUSY
Fault in window's pages already during mapping operation.
Makes sense only for write.
.El
.Pp
Once the mapping is created, it must be accessed only by methods that can
handle faults, such as
.Fn uiomove
or
.Fn kcopy .
Page faults on the mapping will result in the object's pager
method being called to resolve the fault.
.Pp
.Fn ubc_release
frees the mapping at
.Fa va
for reuse.
The mapping may be cached to speed future accesses to the same region
of the object.
The flags can be any of
.Pp
.Bl -tag -offset indent -width "UVM_ADV_SEQUENTIAL" -compact
.It UBC_UNMAP
Do not cache mapping.
.El
.Pp
.Fn ubc_uiomove
allocates an UBC memory window, performs I/O on it and unmaps the window.
The
.Fa advice
parameter takes the same values as the respective parameter in
.Fn ubc_alloc
and the
.Fa flags
parameter takes the same arguments as
.Fn ubc_alloc
and
.Fn ubc_unmap .
Additionally, the flag
.Dv UBC_PARTIALOK
can be provided to indicate that it is acceptable to return if an error
occurs mid-transfer.
.Sh VIRTUAL MEMORY I/O
.Bl -ohang
.It Ft int
.Fn uvm_io "struct vm_map *map" "struct uio *uio" ;
.El
.Pp
.Fn uvm_io
performs the I/O described in
.Fa uio
on the memory described in
.Fa map .
.Sh ALLOCATION OF KERNEL MEMORY
.Bl -ohang
.It Ft vaddr_t
.Fn uvm_km_alloc "struct vm_map *map" "vsize_t size" "vsize_t align" "uvm_flag_t flags" ;
.It Ft void
.Fn uvm_km_free "struct vm_map *map" "vaddr_t addr" "vsize_t size" "uvm_flag_t flags" ;
.It Ft struct vm_map *
.Fn uvm_km_suballoc "struct vm_map *map" "vaddr_t *min" "vaddr_t *max" \
"vsize_t size" "int flags" "bool fixed" "struct vm_map *submap" ;
.El
.Pp
.Fn uvm_km_alloc
allocates
.Fa size
bytes of kernel memory in map
.Fa map .
The first address of the allocated memory range will be aligned according to the
.Fa align
argument
.Pq specify 0 if no alignment is necessary .
The alignment must be a multiple of page size.
The
.Fa flags
is a bitwise inclusive OR of the allocation type and operation flags.
.Pp
The allocation type should be one of:
.Bl -tag -width UVM_KMF_PAGEABLE
.It UVM_KMF_WIRED
Wired memory.
.It UVM_KMF_PAGEABLE
Demand-paged zero-filled memory.
.It UVM_KMF_VAONLY
Virtual address only.
No physical pages are mapped in the allocated region.
If necessary, it's the caller's responsibility to enter page mappings.
It's also the caller's responsibility to clean up the mappings before freeing
the address range.
.El
.Pp
The following operation flags are available:
.Bl -tag -width UVM_KMF_PAGEABLE
.It UVM_KMF_CANFAIL
Can fail even if
.Dv UVM_KMF_NOWAIT
is not specified and
.Dv UVM_KMF_WAITVA
is specified.
.It UVM_KMF_ZERO
Request zero-filled memory.
Only supported for
.Dv UVM_KMF_WIRED .
Shouldn't be used with other types.
.It UVM_KMF_TRYLOCK
Fail if we can't lock the map.
.It UVM_KMF_NOWAIT
Fail immediately if no memory is available.
.It UVM_KMF_WAITVA
Sleep to wait for the virtual address resources if needed.
.El
.Pp
(If neither
.Dv UVM_KMF_NOWAIT
nor
.Dv UVM_KMF_CANFAIL
are specified and
.Dv UVM_KMF_WAITVA
is specified,
.Fn uvm_km_alloc
will never fail, but rather sleep indefinitely until the allocation succeeds.)
.Pp
Pageability of the pages allocated with
.Dv UVM_KMF_PAGEABLE
can be changed by
.Fn uvm_map_pageable .
In that case, the entire range must be changed atomically.
Changing a part of the range is not supported.
.Pp
.Fn uvm_km_free
frees the memory range allocated by
.Fn uvm_km_alloc .
.Fa addr
must be an address returned by
.Fn uvm_km_alloc .
.Fa map
and
.Fa size
must be the same as the ones used for the corresponding
.Fn uvm_km_alloc .
.Fa flags
must be the allocation type used for the corresponding
.Fn uvm_km_alloc .
.Pp
.Fn uvm_km_free
is the only way to free memory ranges allocated by
.Fn uvm_km_alloc .
.Fn uvm_unmap
must not be used.
.Pp
.Fn uvm_km_suballoc
allocates submap from
.Fa map ,
creating a new map if
.Fa submap
is
.Dv NULL .
The addresses of the submap can be specified exactly by setting the
.Fa fixed
argument to true, which causes the
.Fa min
argument to specify the beginning of the address in the submap.
If
.Fa fixed
is false, any address of size
.Fa size
will be allocated from
.Fa map
and the start and end addresses returned in
.Fa min
and
.Fa max .
The
.Fa flags
are used to initialize the created submap.
The following flags could be set:
.Bl -tag -width VM_MAP_PAGEABLE
.It VM_MAP_PAGEABLE
Entries in the map may be paged out.
.It VM_MAP_INTRSAFE
Map should be interrupt-safe.
.It VM_MAP_TOPDOWN
A top-down mapping should be arranged.
.El
.Sh ALLOCATION OF PHYSICAL MEMORY
.Bl -ohang
.It Ft struct vm_page *
.Fn uvm_pagealloc "struct uvm_object *uobj" "voff_t off" "struct vm_anon *anon" "int flags" ;
.It Ft void
.Fn uvm_pagerealloc "struct vm_page *pg" "struct uvm_object *newobj" "voff_t newoff" ;
.It Ft void
.Fn uvm_pagefree "struct vm_page *pg" ;
.It Ft int
.Fn uvm_pglistalloc "psize_t size" "paddr_t low" "paddr_t high" "paddr_t alignment" "paddr_t boundary" "struct pglist *rlist" "int nsegs" "int waitok" ;
.It Ft void
.Fn uvm_pglistfree "struct pglist *list" ;
.It Ft void
.Fn uvm_page_physload "vaddr_t start" "vaddr_t end" "vaddr_t avail_start" "vaddr_t avail_end" "int free_list" ;
.El
.Pp
.Fn uvm_pagealloc
allocates a page of memory at virtual address
.Fa off
in either the object
.Fa uobj
or the anonymous memory
.Fa anon ,
which must be locked by the caller.
Only one of
.Fa uobj
and
.Fa anon
can be non
.Dv NULL .
Returns
.Dv NULL
when no page can be found.
The flags can be any of
.Bd -literal
#define UVM_PGA_USERESERVE      0x0001  /* ok to use reserve pages */
#define UVM_PGA_ZERO            0x0002  /* returned page must be zero'd */
.Ed
.Pp
.Dv UVM_PGA_USERESERVE
means to allocate a page even if that will result in the number of free pages
being lower than
.Dv uvmexp.reserve_pagedaemon
(if the current thread is the pagedaemon) or
.Dv uvmexp.reserve_kernel
(if the current thread is not the pagedaemon).
.Dv UVM_PGA_ZERO
causes the returned page to be filled with zeroes, either by allocating it
from a pool of pre-zeroed pages or by zeroing it in-line as necessary.
.Pp
.Fn uvm_pagerealloc
reallocates page
.Fa pg
to a new object
.Fa newobj ,
at a new offset
.Fa newoff .
.Pp
.Fn uvm_pagefree
frees the physical page
.Fa pg .
If the content of the page is known to be zero-filled,
caller should set
.Dv PG_ZERO
in pg-\*[Gt]flags so that the page allocator will use
the page to serve future
.Dv UVM_PGA_ZERO
requests efficiently.
.Pp
.Fn uvm_pglistalloc
allocates a list of pages for size
.Fa size
byte under various constraints.
.Fa low
and
.Fa high
describe the lowest and highest addresses acceptable for the list.
If
.Fa alignment
is non-zero, it describes the required alignment of the list, in
power-of-two notation.
If
.Fa boundary
is non-zero, no segment of the list may cross this power-of-two
boundary, relative to zero.
.Fa nsegs
is the maximum number of physically contiguous segments.
If
.Fa waitok
is non-zero, the function may sleep until enough memory is available.
(It also may give up in some situations, so a non-zero
.Fa waitok
does not imply that
.Fn uvm_pglistalloc
cannot return an error.)
The allocated memory is returned in the
.Fa rlist
list; the caller has to provide storage only, the list is initialized by
.Fn uvm_pglistalloc .
.Pp
.Fn uvm_pglistfree
frees the list of pages pointed to by
.Fa list .
If the content of the page is known to be zero-filled,
caller should set
.Dv PG_ZERO
in pg-\*[Gt]flags so that the page allocator will use
the page to serve future
.Dv UVM_PGA_ZERO
requests efficiently.
.Pp
.Fn uvm_page_physload
loads physical memory segments into VM space on the specified
.Fa free_list .
It must be called at system boot time to set up physical memory
management pages.
The arguments describe the
.Fa start
and
.Fa end
of the physical addresses of the segment, and the available start and end
addresses of pages not already in use.
If a system has memory banks of
different speeds the slower memory should be given a higher
.Fa free_list
value.
.\" XXX expand on "system boot time"!
.Sh PROCESSES
.Bl -ohang
.It Ft void
.Fn uvm_pageout "void" ;
.It Ft void
.Fn uvm_scheduler "void" ;
.It Ft void
.Fn uvm_swapin "struct lwp *l" ;
.El
.Pp
.Fn uvm_pageout
is the main loop for the page daemon.
.Pp
.Fn uvm_scheduler
is the process zero main loop, which is to be called after the
system has finished starting other processes.
It handles the swapping in of runnable, swapped out processes in priority
order.
.Pp
.Fn uvm_swapin
swaps in the named lwp.
.Sh PAGE LOAN
.Bl -ohang
.It Ft int
.Fn uvm_loan "struct vm_map *map" "vaddr_t start" "vsize_t len" "void *v" "int flags" ;
.It Ft void
.Fn uvm_unloan "void *v" "int npages" "int flags" ;
.El
.Pp
.Fn uvm_loan
loans pages in a map out to anons or to the kernel.
.Fa map
should be unlocked,
.Fa start
and
.Fa len
should be multiples of
.Dv PAGE_SIZE .
Argument
.Fa flags
should be one of
.Bd -literal
#define UVM_LOAN_TOANON       0x01    /* loan to anons */
#define UVM_LOAN_TOPAGE       0x02    /* loan to kernel */
.Ed
.Pp
.Fa v
should be pointer to array of pointers to
.Li struct anon
or
.Li struct vm_page ,
as appropriate.
The caller has to allocate memory for the array and
ensure it's big enough to hold
.Fa len / PAGE_SIZE
pointers.
Returns 0 for success, or appropriate error number otherwise.
Note that wired pages can't be loaned out and
.Fn uvm_loan
will fail in that case.
.Pp
.Fn uvm_unloan
kills loans on pages or anons.
The
.Fa v
must point to the array of pointers initialized by previous call to
.Fn uvm_loan .
.Fa npages
should match number of pages allocated for loan, this also matches
number of items in the array.
Argument
.Fa flags
should be one of
.Bd -literal
#define UVM_LOAN_TOANON       0x01    /* loan to anons */
#define UVM_LOAN_TOPAGE       0x02    /* loan to kernel */
.Ed
.Pp
and should match what was used for previous call to
.Fn uvm_loan .
.Sh MISCELLANEOUS FUNCTIONS
.Bl -ohang
.It Ft struct uvm_object *
.Fn uao_create "vsize_t size" "int flags" ;
.It Ft void
.Fn uao_detach "struct uvm_object *uobj" ;
.It Ft void
.Fn uao_reference "struct uvm_object *uobj" ;
.It Ft bool
.Fn uvm_chgkprot "void *addr" "size_t len" "int rw" ;
.It Ft void
.Fn uvm_kernacc "void *addr" "size_t len" "int rw" ;
.It Ft int
.Fn uvm_vslock "struct vmspace *vs" "void *addr" "size_t len" "vm_prot_t prot" ;
.It Ft void
.Fn uvm_vsunlock "struct vmspace *vs" "void *addr" "size_t len" ;
.It Ft void
.Fn uvm_meter "void" ;
.It Ft void
.Fn uvm_fork "struct lwp *l1" "struct lwp *l2" "bool shared" ;
.It Ft int
.Fn uvm_grow "struct proc *p" "vaddr_t sp" ;
.It Ft void
.Fn uvn_findpages "struct uvm_object *uobj" "voff_t offset" "int *npagesp" "struct vm_page **pps" "int flags" ;
.It Ft void
.Fn uvm_swap_stats "int cmd" "struct swapent *sep" "int sec" "register_t *retval" ;
.El
.Pp
The
.Fn uao_create ,
.Fn uao_detach ,
and
.Fn uao_reference
functions operate on anonymous memory objects, such as those used to support
System V shared memory.
.Fn uao_create
returns an object of size
.Fa size
with flags:
.Bd -literal
#define UAO_FLAG_KERNOBJ        0x1     /* create kernel object */
#define UAO_FLAG_KERNSWAP       0x2     /* enable kernel swap */
.Ed
.Pp
which can only be used once each at system boot time.
.Fn uao_reference
creates an additional reference to the named anonymous memory object.
.Fn uao_detach
removes a reference from the named anonymous memory object, destroying
it if removing the last reference.
.Pp
.Fn uvm_chgkprot
changes the protection of kernel memory from
.Fa addr
to
.Fa addr + len
to the value of
.Fa rw .
This is primarily useful for debuggers, for setting breakpoints.
This function is only available with options
.Dv KGDB .
.Pp
.Fn uvm_kernacc
checks the access at address
.Fa addr
to
.Fa addr + len
for
.Fa rw
access in the kernel address space.
.Pp
.Fn uvm_vslock
and
.Fn uvm_vsunlock
control the wiring and unwiring of pages for process
.Fa p
from
.Fa addr
to
.Fa addr + len .
These functions are normally used to wire memory for I/O.
.Pp
.Fn uvm_meter
calculates the load average and wakes up the swapper if necessary.
.Pp
.Fn uvm_fork
forks a virtual address space for process' (old)
.Fa p1
and (new)
.Fa p2 .
If the
.Fa shared
argument is non zero, p1 shares its address space with p2,
otherwise a new address space is created.
This function currently has no return value, and thus cannot fail.
In the future, this function will be changed to allow it to
fail in low memory conditions.
.Pp
.Fn uvm_grow
increases the stack segment of process
.Fa p
to include
.Fa sp .
.Pp
.Fn uvn_findpages
looks up or creates pages in
.Fa uobj
at offset
.Fa offset ,
marks them busy and returns them in the
.Fa pps
array.
Currently
.Fa uobj
must be a vnode object.
The number of pages requested is pointed to by
.Fa npagesp ,
and this value is updated with the actual number of pages returned.
The flags can be
.Bd -literal
#define UFP_ALL         0x00    /* return all pages requested */
#define UFP_NOWAIT      0x01    /* don't sleep */
#define UFP_NOALLOC     0x02    /* don't allocate new pages */
#define UFP_NOCACHE     0x04    /* don't return pages which already exist */
#define UFP_NORDONLY    0x08    /* don't return PG_READONLY pages */
.Ed
.Pp
.Dv UFP_ALL
is a pseudo-flag meaning all requested pages should be returned.
.Dv UFP_NOWAIT
means that we must not sleep.
.Dv UFP_NOALLOC
causes any pages which do not already exist to be skipped.
.Dv UFP_NOCACHE
causes any pages which do already exist to be skipped.
.Dv UFP_NORDONLY
causes any pages which are marked PG_READONLY to be skipped.
.Pp
.Fn uvm_swap_stats
implements the
.Dv SWAP_STATS
and
.Dv SWAP_OSTATS
operation of the
.Xr swapctl 2
system call.
.Fa cmd
is the requested command,
.Dv SWAP_STATS
or
.Dv SWAP_OSTATS .
The function will copy no more than
.Fa sec
entries in the array pointed by
.Fa sep .
On return,
.Fa retval
holds the actual number of entries copied in the array.
.Sh SYSCTL
UVM provides support for the
.Dv CTL_VM
domain of the
.Xr sysctl 3
hierarchy.
It handles the
.Dv VM_LOADAVG ,
.Dv VM_METER ,
.Dv VM_UVMEXP ,
and
.Dv VM_UVMEXP2
nodes, which return the current load averages, calculates current VM
totals, returns the uvmexp structure, and a kernel version independent
view of the uvmexp structure, respectively.
It also exports a number of tunables that control how much VM space is
allowed to be consumed by various tasks.
The load averages are typically accessed from userland using the
.Xr getloadavg 3
function.
The uvmexp structure has all global state of the UVM system,
and has the following members:
.Bd -literal
/* vm_page constants */
int pagesize;   /* size of a page (PAGE_SIZE): must be power of 2 */
int pagemask;   /* page mask */
int pageshift;  /* page shift */

/* vm_page counters */
int npages;     /* number of pages we manage */
int free;       /* number of free pages */
int active;     /* number of active pages */
int inactive;   /* number of pages that we free'd but may want back */
int paging;     /* number of pages in the process of being paged out */
int wired;      /* number of wired pages */
int reserve_pagedaemon; /* number of pages reserved for pagedaemon */
int reserve_kernel; /* number of pages reserved for kernel */

/* pageout params */
int freemin;    /* min number of free pages */
int freetarg;   /* target number of free pages */
int inactarg;   /* target number of inactive pages */
int wiredmax;   /* max number of wired pages */

/* swap */
int nswapdev;   /* number of configured swap devices in system */
int swpages;    /* number of PAGE_SIZE'ed swap pages */
int swpginuse;  /* number of swap pages in use */
int nswget;     /* number of times fault calls uvm_swap_get() */
int nanon;      /* number total of anon's in system */
int nfreeanon;  /* number of free anon's */

/* stat counters */
int faults;             /* page fault count */
int traps;              /* trap count */
int intrs;              /* interrupt count */
int swtch;              /* context switch count */
int softs;              /* software interrupt count */
int syscalls;           /* system calls */
int pageins;            /* pagein operation count */
                        /* pageouts are in pdpageouts below */
int swapins;            /* swapins */
int swapouts;           /* swapouts */
int pgswapin;           /* pages swapped in */
int pgswapout;          /* pages swapped out */
int forks;              /* forks */
int forks_ppwait;       /* forks where parent waits */
int forks_sharevm;      /* forks where vmspace is shared */

/* fault subcounters */
int fltnoram;   /* number of times fault was out of ram */
int fltnoanon;  /* number of times fault was out of anons */
int fltpgwait;  /* number of times fault had to wait on a page */
int fltpgrele;  /* number of times fault found a released page */
int fltrelck;   /* number of times fault relock called */
int fltrelckok; /* number of times fault relock is a success */
int fltanget;   /* number of times fault gets anon page */
int fltanretry; /* number of times fault retrys an anon get */
int fltamcopy;  /* number of times fault clears "needs copy" */
int fltnamap;   /* number of times fault maps a neighbor anon page */
int fltnomap;   /* number of times fault maps a neighbor obj page */
int fltlget;    /* number of times fault does a locked pgo_get */
int fltget;     /* number of times fault does an unlocked get */
int flt_anon;   /* number of times fault anon (case 1a) */
int flt_acow;   /* number of times fault anon cow (case 1b) */
int flt_obj;    /* number of times fault is on object page (2a) */
int flt_prcopy; /* number of times fault promotes with copy (2b) */
int flt_przero; /* number of times fault promotes with zerofill (2b) */

/* daemon counters */
int pdwoke;     /* number of times daemon woke up */
int pdrevs;     /* number of times daemon rev'd clock hand */
int pdswout;    /* number of times daemon called for swapout */
int pdfreed;    /* number of pages daemon freed since boot */
int pdscans;    /* number of pages daemon scanned since boot */
int pdanscan;   /* number of anonymous pages scanned by daemon */
int pdobscan;   /* number of object pages scanned by daemon */
int pdreact;    /* number of pages daemon reactivated since boot */
int pdbusy;     /* number of times daemon found a busy page */
int pdpageouts; /* number of times daemon started a pageout */
int pdpending;  /* number of times daemon got a pending pageout */
int pddeact;    /* number of pages daemon deactivates */
.Ed
.Sh NOTES
.Fn uvm_chgkprot
is only available if the kernel has been compiled with options
.Dv KGDB .
.Pp
All structure and types whose names begin with
.Dq vm_
will be renamed to
.Dq uvm_ .
.Sh SEE ALSO
.Xr swapctl 2 ,
.Xr getloadavg 3 ,
.Xr kvm 3 ,
.Xr sysctl 3 ,
.Xr ddb 4 ,
.Xr options 4 ,
.Xr memoryallocators 9 ,
.Xr pmap 9
.Sh HISTORY
UVM is a new VM system developed at Washington University in St. Louis
(Missouri).
UVM's roots lie partly in the Mach-based
.Bx 4.4
VM system, the
.Fx
VM system, and the SunOS 4 VM system.
UVM's basic structure is based on the
.Bx 4.4
VM system.
UVM's new anonymous memory system is based on the
anonymous memory system found in the SunOS 4 VM (as described in papers
published by Sun Microsystems, Inc.).
UVM also includes a number of features new to
.Bx
including page loanout, map entry passing, simplified
copy-on-write, and clustered anonymous memory pageout.
UVM is also further documented in an August 1998 dissertation by
Charles D. Cranor.
.Pp
UVM appeared in
.Nx 1.4 .
.Sh AUTHORS
Charles D. Cranor
.Aq chuck@ccrc.wustl.edu
designed and implemented UVM.
.Pp
Matthew Green
.Aq mrg@eterna.com.au
wrote the swap-space management code and handled the logistical issues
involved with merging UVM into the
.Nx
source tree.
.Pp
Chuck Silvers
.Aq chuq@chuq.com
implemented the aobj pager, thus allowing UVM to support System V shared
memory and process swapping.
He also designed and implemented the UBC part of UVM, which uses UVM pages
to cache vnode data rather than the traditional buffer cache buffers.