man/man9/pmap.9

.\"	$NetBSD: pmap.9,v 1.44 2012/02/16 12:10:10 yamt Exp $
.\"
.\" Copyright (c) 2000, 2001, 2002 The NetBSD Foundation, Inc.
.\" All rights reserved.
.\"
.\" This code is derived from software contributed to The NetBSD Foundation
.\" by Jason R. Thorpe.
.\"
.\" 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 NETBSD FOUNDATION, INC. 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 FOUNDATION 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.
.\"
.Dd February 16, 2012
.Dt PMAP 9
.Os
.Sh NAME
.Nm pmap
.Nd machine-dependent portion of the virtual memory system
.Sh SYNOPSIS
.In sys/param.h
.In uvm/uvm_extern.h
.Ft void
.Fn "pmap_init" "void"
.Ft void
.Fn "pmap_virtual_space" "vaddr_t *vstartp" "vaddr_t *vendp"
.Ft vaddr_t
.Fn "pmap_steal_memory" "vsize_t size" "vaddr_t *vstartp" "vaddr_t *vendp"
.Ft pmap_t
.Fn "pmap_kernel" "void"
.Ft pmap_t
.Fn "pmap_create" "void"
.Ft void
.Fn "pmap_destroy" "pmap_t pmap"
.Ft void
.Fn "pmap_reference" "pmap_t pmap"
.Ft void
.Fn "pmap_fork" "pmap_t src_map" "pmap_t dst_map"
.Ft long
.Fn "pmap_resident_count" "pmap_t pmap"
.Ft long
.Fn "pmap_wired_count" "pmap_t pmap"
.Ft vaddr_t
.Fn "pmap_growkernel" "vaddr_t maxkvaddr"
.Ft int
.Fn "pmap_enter" "pmap_t pmap" "vaddr_t va" "paddr_t pa" "vm_prot_t prot" \
    "u_int flags"
.Ft void
.Fn "pmap_remove" "pmap_t pmap" "vaddr_t sva" "vaddr_t eva"
.Ft void
.Fn "pmap_remove_all" "pmap_t pmap"
.Ft void
.Fn "pmap_protect" "pmap_t pmap" "vaddr_t sva" "vaddr_t eva" "vm_prot_t prot"
.Ft void
.Fn "pmap_unwire" "pmap_t pmap" "vaddr_t va"
.Ft bool
.Fn "pmap_extract" "pmap_t pmap" "vaddr_t va" "paddr_t *pap"
.Ft void
.Fn "pmap_kenter_pa" "vaddr_t va" "paddr_t pa" "vm_prot_t prot" "u_int flags"
.Ft void
.Fn "pmap_kremove" "vaddr_t va" "vsize_t size"
.Ft void
.Fn "pmap_copy" "pmap_t dst_map" "pmap_t src_map" "vaddr_t dst_addr" \
   "vsize_t len" "vaddr_t src_addr"
.Ft void
.Fn "pmap_update" "pmap_t pmap"
.Ft void
.Fn "pmap_activate" "struct lwp *l"
.Ft void
.Fn "pmap_deactivate" "struct lwp *l"
.Ft void
.Fn "pmap_zero_page" "paddr_t pa"
.Ft void
.Fn "pmap_copy_page" "paddr_t src" "paddr_t dst"
.Ft void
.Fn "pmap_page_protect" "struct vm_page *pg" "vm_prot_t prot"
.Ft bool
.Fn "pmap_clear_modify" "struct vm_page *pg"
.Ft bool
.Fn "pmap_clear_reference" "struct vm_page *pg"
.Ft bool
.Fn "pmap_is_modified" "struct vm_page *pg"
.Ft bool
.Fn "pmap_is_referenced" "struct vm_page *pg"
.Ft paddr_t
.Fn "pmap_phys_address" "paddr_t cookie"
.Ft vaddr_t
.Fn "PMAP_MAP_POOLPAGE" "paddr_t pa"
.Ft paddr_t
.Fn "PMAP_UNMAP_POOLPAGE" "vaddr_t va"
.Ft void
.Fn "PMAP_PREFER" "vaddr_t hint" "vaddr_t *vap" "vsize_t sz" "int td"
.Sh DESCRIPTION
The
.Nm
module is the machine-dependent portion of the
.Nx
virtual memory system
.Xr uvm 9 .
The purpose of the
.Nm
module is to manage physical address maps, to program the
memory management hardware on the system, and perform any
cache operations necessary to ensure correct operation of
the virtual memory system.
The
.Nm
module is also responsible for maintaining certain information
required by
.Xr uvm 9 .
.Pp
In order to cope with hardware architectures that make the
invalidation of virtual address mappings expensive (e.g.,
TLB invalidations, TLB shootdown operations for multiple
processors), the
.Nm
module is allowed to delay mapping invalidation or protection
operations until such time as they are actually necessary.
The functions that are allowed to delay such actions are
.Fn pmap_enter ,
.Fn pmap_remove ,
.Fn pmap_protect ,
.Fn pmap_kenter_pa ,
and
.Fn pmap_kremove .
Callers of these functions must use the
.Fn pmap_update
function to notify the
.Nm
module that the mappings need to be made correct.
Since the
.Nm
module is provided with information as to which processors are
using a given physical map, the
.Nm
module may use whatever optimizations it has available to reduce
the expense of virtual-to-physical mapping synchronization.
.Ss HEADER FILES AND DATA STRUCTURES
Machine-dependent code must provide the header file
.In machine/pmap.h .
This file contains the definition of the
.Dv pmap
structure:
.Bd -literal -offset indent
struct pmap {
        /* Contents defined by pmap implementation. */
};
typedef struct pmap *pmap_t;
.Ed
.Pp
This header file may also define other data structures that the
.Nm
implementation uses.
.Pp
Note that all prototypes for
.Nm
interface functions are provided by the header file
.In uvm/uvm_pmap.h .
It is possible to override this behavior by defining the
C pre-processor macro
.Dv PMAP_EXCLUDE_DECLS .
This may be used to add a layer of indirection to
.Nm
API calls, for handling different MMU types in a single
.Nm
module, for example.
If the
.Dv PMAP_EXCLUDE_DECLS
macro is defined,
.In machine/pmap.h
.Em must
provide function prototypes in a block like so:
.Bd -literal -offset indent
#ifdef _KERNEL /* not exposed to user namespace */
__BEGIN_DECLS  /* make safe for C++ */
/* Prototypes go here. */
__END_DECLS
#endif /* _KERNEL */
.Ed
.Pp
The header file
.In uvm/uvm_pmap.h
defines a structure for tracking
.Nm
statistics (see below).
This structure is defined as:
.Bd -literal -offset indent
struct pmap_statistics {
        long        resident_count; /* number of mapped pages */
        long        wired_count;    /* number of wired pages */
};
.Ed
.Ss WIRED MAPPINGS
The
.Nm
module is based on the premise that all information contained
in the physical maps it manages is redundant.
That is, physical map information may be
.Dq forgotten
by the
.Nm
module in the event that it is necessary to do so; it can be rebuilt
by
.Xr uvm 9
by taking a page fault.
There is one exception to this rule: so-called
.Dq wired
mappings may not be forgotten.
Wired mappings are those for which either no high-level information
exists with which to rebuild the mapping, or mappings which are needed
by critical sections of code where taking a page fault is unacceptable.
Information about which mappings are wired is provided to the
.Nm
module when a mapping is established.
.Ss MODIFIED/REFERENCED INFORMATION
The
.Nm
module is required to keep track of whether or not a page managed
by the virtual memory system has been referenced or modified.
This information is used by
.Xr uvm 9
to determine what happens to the page when scanned by the
pagedaemon.
.Pp
Many CPUs provide hardware support for tracking
modified/referenced information.
However, many CPUs, particularly modern RISC CPUs, do not.
On CPUs which lack hardware support for modified/referenced tracking, the
.Nm
module must emulate it in software.
There are several strategies for doing this, and the best strategy
depends on the CPU.
.Pp
The
.Dq referenced
attribute is used by the pagedaemon to determine if a page is
.Dq active .
Active pages are not candidates for re-use in the page replacement algorithm.
Accurate referenced information is not required for correct operation; if
supplying referenced information for a page is not feasible, then the
.Nm
implementation should always consider the
.Dq referenced
attribute to be
.Dv false .
.Pp
The
.Dq modified
attribute is used by the pagedaemon to determine if a page needs
to be cleaned (written to backing store; swap space, a regular file, etc.).
Accurate modified information
.Em must
be provided by the
.Nm
module for correct operation of the virtual memory system.
.Pp
Note that modified/referenced information is only tracked for
pages managed by the virtual memory system (i.e., pages for
which a vm_page structure exists).
In addition, only
.Dq managed
mappings of those pages have modified/referenced tracking.
Mappings entered with the
.Fn pmap_enter
function are
.Dq managed
mappings.
It is possible for
.Dq unmanaged
mappings of a page to be created, using the
.Fn pmap_kenter_pa
function.
The use of
.Dq unmanaged
mappings should be limited to code which may execute in interrupt context
(for example, the kernel memory allocator), or to enter mappings for
physical addresses which are not managed by the virtual memory system.
.Dq Unmanaged
mappings may only be entered into the kernel's virtual address space.
This constraint is placed on the callers of the
.Fn pmap_kenter_pa
and
.Fn pmap_kremove
functions so that the
.Nm
implementation need not block interrupts when manipulating data
structures or holding locks.
.Pp
Also note that the modified/referenced information must be tracked
on a per-page basis; they are not attributes of a mapping, but attributes
of a page.
Therefore, even after all mappings for a given page have
been removed, the modified/referenced information for that page
.Em must
be preserved.
The only time the modified/referenced attributes may
be cleared is when the virtual memory system explicitly calls the
.Fn pmap_clear_modify
and
.Fn pmap_clear_reference
functions.
These functions must also change any internal state necessary to detect
the page being modified or referenced again after the modified or
referenced state is cleared.
(Prior to
.Nx 1.6 ,
.Nm
implementations could get away without this because UVM (and Mach VM
before that) always called
.Fn pmap_page_protect
before clearing the modified or referenced state, but UVM has been changed
to not do this anymore, so all
.Nm
implementations must now handle this.)
.Ss STATISTICS
The
.Nm
is required to keep statistics as to the number of
.Dq resident
pages and the number of
.Dq wired
pages.
.Pp
A
.Dq resident
page is one for which a mapping exists.
This statistic is used to compute the resident size of a process and
enforce resource limits.
Only pages (whether managed by the virtual memory system or not)
which are mapped into a physical map should be counted in the resident
count.
.Pp
A
.Dq wired
page is one for which a wired mapping exists.
This statistic is used to enforce resource limits.
.Pp
Note that it is recommended (though not required) that the
.Nm
implementation use the
.Dv pmap_statistics
structure in the tracking of
.Nm
statistics by placing it inside the
.Dv pmap
structure and adjusting the counts when mappings are established, changed,
or removed.
This avoids potentially expensive data structure traversals when the
statistics are queried.
.Ss REQUIRED FUNCTIONS
This section describes functions that a
.Nm
module must provide to the virtual memory system.
.Bl -tag -width indent -offset indent
.It void Fn "pmap_init" "void"
This function initializes the
.Nm
module.
It is called by
.Fn uvm_init
to initialize any data structures that the module needs to
manage physical maps.
.It pmap_t Fn "pmap_kernel" "void"
A machine independent macro which expands to
.Va kernel_pmap_ptr .
This variable must be exported by the platform's pmap module and it
must point to the kernel pmap.
.It void Fn "pmap_virtual_space" "vaddr_t *vstartp" "vaddr_t *vendp"
The
.Fn pmap_virtual_space
function is called to determine the initial kernel virtual address
space beginning and end.
These values are used to create the kernel's virtual memory map.
The function must set
.Fa *vstartp
to the first kernel virtual address that will be managed by
.Xr uvm 9 ,
and must set
.Fa *vendp
to the last kernel virtual address that will be managed by
.Xr uvm 9 .
.Pp
If the
.Fn pmap_growkernel
feature is used by a
.Nm
implementation, then
.Fa *vendp
should be set to the maximum kernel virtual address allowed by the
implementation.
If
.Fn pmap_growkernel
is not used, then
.Fa *vendp
.Em must
be set to the maximum kernel virtual address that can be mapped with
the resources currently allocated to map the kernel virtual address
space.
.It pmap_t Fn "pmap_create" "void"
Create a physical map and return it to the caller.
The reference count on the new map is 1.
.It void Fn "pmap_destroy" "pmap_t pmap"
Drop the reference count on the specified physical map.
If the reference count drops to 0, all resources associated with the
physical map are released and the physical map destroyed.
In the case of a drop-to-0, no mappings will exist in the map.
The
.Nm
implementation may assert this.
.It void Fn "pmap_reference" "pmap_t pmap"
Increment the reference count on the specified physical map.
.It long Fn "pmap_resident_count" "pmap_t pmap"
Query the
.Dq resident pages
statistic for
.Fa pmap .
.Pp
Note that this function may be provided as a C pre-processor macro.
.It long Fn "pmap_wired_count" "pmap_t pmap"
Query the
.Dq wired pages
statistic for
.Fa pmap .
.Pp
Note that this function may be provided as a C pre-processor macro.
.It int Fn "pmap_enter" "pmap_t pmap" "vaddr_t va" "paddr_t pa" \
    "vm_prot_t prot" "u_int flags"
Create a mapping in physical map
.Fa pmap
for the physical address
.Fa pa
at the virtual address
.Fa va
with protection specified by bits in
.Fa prot :
.Bl -tag -width "VM_PROT_EXECUTE  " -offset indent
.It VM_PROT_READ
The mapping must allow reading.
.It VM_PROT_WRITE
The mapping must allow writing.
.It VM_PROT_EXECUTE
The page mapped contains instructions that will be executed by the
processor.
.El
.Pp
The
.Fa flags
argument contains protection bits (the same bits as used in the
.Fa prot
argument) indicating the type of access that caused the mapping to
be created.
This information may be used to seed modified/referenced
information for the page being mapped, possibly avoiding redundant faults
on platforms that track modified/referenced information in software.
Other information provided by
.Fa flags :
.Bl -tag -width "PMAP_CANFAIL  " -offset indent
.It PMAP_WIRED
The mapping being created is a wired mapping.
.It PMAP_CANFAIL
The call to
.Fn pmap_enter
is allowed to fail.
If this flag is
.Em not
set, and the
.Fn pmap_enter
call is unable to create the mapping, perhaps due to insufficient
resources, the
.Nm
module must panic.
.It PMAP_NOCACHE
The mapping being created is
.Em not
cached.
Write accesses have a write-through policy.
No speculative memory accesses.
.It PMAP_WRITE_COMBINE
The mapping being created is
.Em not
cached.
Writes are combined and done in one burst.
Speculative read accesses may be allowed.
.It PMAP_WRITE_BACK
All accesses to the created mapping are cached.
On reads, cachelines become shared or exclusive if allocated on cache miss.
On writes, cachelines become modified on a cache miss.
.It PMAP_NOCACHE_OVR
Same as PMAP_NOCACHE but mapping is overrideable (e.g. on x86 by MTRRs).
.El
.Pp
The access type provided in the
.Fa flags
argument will never exceed the protection specified by
.Fa prot .
The
.Nm
implementation may assert this.
Note that on systems that do not provide hardware support for
tracking modified/referenced information, modified/referenced
information for the page
.Em must
be seeded with the access type provided in
.Fa flags
if the
.Dv PMAP_WIRED
flag is set.
This is to prevent a fault for the purpose of tracking
modified/referenced information from occurring while the system is in
a critical section where a fault would be unacceptable.
.Pp
Note that
.Fn pmap_enter
is sometimes called to enter a mapping at a virtual address
for which a mapping already exists.
In this situation, the implementation must take whatever action is
necessary to invalidate the previous mapping before entering the new one.
.Pp
Also note that
.Fn pmap_enter
is sometimes called to change the protection for a pre-existing
mapping, or to change the
.Dq wired
attribute for a pre-existing mapping.
.Pp
The
.Fn pmap_enter
function returns 0 on success or an error code indicating the mode
of failure.
.It void Fn "pmap_remove" "pmap_t pmap" "vaddr_t sva" "vaddr_t eva"
Remove mappings from the virtual address range
.Fa sva
to
.Fa eva
from the specified physical map.
.It void Fn "pmap_remove_all" "pmap_t pmap"
This function is a hint to the
.Nm pmap
implementation that all entries in
.Fa pmap
will be removed before any more entries are entered.
Following this call, there will be
.Fn pmap_remove
calls resulting in every mapping being removed, followed by either
.Fn pmap_destroy
or
.Fn pmap_update .
No other
.Nm pmap
interfaces which take
.Fa pmap
as an argument will be called during this process.
Other interfaces which might need to access
.Fa pmap
(such as
.Fn pmap_page_protect )
are permitted during this process.
.Pp
The
.Nm pmap
implementation is free to either remove all the
.Nm pmap Ns 's
mappings immediately in
.Fn pmap_remove_all ,
or to use the knowledge of the upcoming
.Fn pmap_remove
calls to optimize the removals (or to just ignore this call).
.Pp
.It void Fn "pmap_protect" "pmap_t pmap" "vaddr_t sva" "vaddr_t eva" \
    "vm_prot_t prot"
Set the protection of the mappings in the virtual address range
.Fa sva
to
.Fa eva
in the specified physical map.
.It void Fn "pmap_unwire" "pmap_t pmap" "vaddr_t va"
Clear the
.Dq wired
attribute on the mapping for virtual address
.Fa va .
.It bool Fn "pmap_extract" "pmap_t pmap" "vaddr_t va" "paddr_t *pap"
This function extracts a mapping from the specified physical map.
It serves two purposes: to determine if a mapping exists for the specified
virtual address, and to determine what physical address is mapped at the
specified virtual address.
The
.Fn pmap_extract
should return the physical address for any kernel-accessible address,
including KSEG-style direct-mapped kernel addresses.
.Pp
The
.Fn pmap_extract
function returns
.Dv false
if a mapping for
.Fa va
does not exist.
Otherwise, it returns
.Dv true
and places the physical address mapped at
.Fa va
into
.Fa *pap
if the
.Fa pap
argument is non-NULL.
.It void Fn "pmap_kenter_pa" "vaddr_t va" "paddr_t pa" "vm_prot_t prot" \
    "u_int flags"
Enter an
.Dq unmanaged
mapping for physical address
.Fa pa
at virtual address
.Fa va
with protection specified by bits in
.Fa prot :
.Bl -tag -width "VM_PROT_EXECUTE  " -offset indent
.It VM_PROT_READ
The mapping must allow reading.
.It VM_PROT_WRITE
The mapping must allow writing.
.It VM_PROT_EXECUTE
The page mapped contains instructions that will be executed by the
processor.
.El
.Pp
Information provided by
.Fa flags :
.Bl -tag -width "PMAP_NOCACHE  " -offset indent
.It PMAP_NOCACHE
The mapping being created is
.Em not
cached.
Write accesses have a write-through policy.
No speculative memory accesses.
.It PMAP_WRITE_COMBINE
The mapping being created is
.Em not
cached.
Writes are combined and done in one burst.
Speculative read accesses may be allowed.
.It PMAP_WRITE_BACK
All accesses to the created mapping are cached.
On reads, cachelines become shared or exclusive if allocated on cache miss.
On writes, cachelines become modified on a cache miss.
.It PMAP_NOCACHE_OVR
Same as PMAP_NOCACHE but mapping is overrideable (e.g. on x86 by MTRRs).
.El
.Pp
Mappings of this type are always
.Dq wired ,
and are unaffected by routines that alter the protection of pages
(such as
.Fn pmap_page_protect ) .
Such mappings are also not included in the gathering of modified/referenced
information about a page.
Mappings entered with
.Fn pmap_kenter_pa
by machine-independent code
.Em must not
have execute permission, as the
data structures required to track execute permission of a page may not
be available to
.Fn pmap_kenter_pa .
Machine-independent code is not allowed to enter a mapping with
.Fn pmap_kenter_pa
at a virtual address for which a valid mapping already exists.
Mappings created with
.Fn pmap_kenter_pa
may be removed only with a call to
.Fn pmap_kremove .
.Pp
Note that
.Fn pmap_kenter_pa
must be safe for use in interrupt context.
.Fn splvm
blocks interrupts that might cause
.Fn pmap_kenter_pa
to be called.
.It void Fn "pmap_kremove" "vaddr_t va" "vsize_t size"
Remove all mappings starting at virtual address
.Fa va
for
.Fa size
bytes from the kernel physical map.
All mappings that are removed must be the
.Dq unmanaged
type created with
.Fn pmap_kenter_pa .
The implementation may assert this.
.It void Fn "pmap_copy" "pmap_t dst_map" "pmap_t src_map" "vaddr_t dst_addr" \
    "vsize_t len" "vaddr_t src_addr"
This function copies the mappings starting at
.Fa src_addr
in
.Fa src_map
for
.Fa len
bytes into
.Fa dst_map
starting at
.Fa dst_addr .
.Pp
Note that while this function is required to be provided by a
.Nm
implementation, it is not actually required to do anything.
.Fn pmap_copy
is merely advisory (it is used in the
.Xr fork 2
path to
.Dq pre-fault
the child's address space).
.It void Fn "pmap_update" "pmap_t pmap"
This function is used to inform the
.Nm
module that all physical mappings, for the specified pmap, must now be
correct.
That is, all delayed virtual-to-physical mappings updates (such as TLB
invalidation or address space identifier updates) must be completed.
This routine must be used after calls to
.Fn pmap_enter ,
.Fn pmap_remove ,
.Fn pmap_protect ,
.Fn pmap_kenter_pa ,
and
.Fn pmap_kremove
in order to ensure correct operation of the virtual memory system.
.Pp
If a
.Nm
implementation does not delay virtual-to-physical mapping updates,
.Fn pmap_update
has no operation.
In this case, the call may be deleted using a C pre-processor macro in
.In machine/pmap.h .
.It void Fn "pmap_activate" "struct lwp *l"
Activate the physical map used by the process behind lwp
.Fa l .
This is called by the virtual memory system when the
virtual memory context for a process is changed, and is also
often used by machine-dependent context switch code to program
the memory management hardware with the process's page table
base, etc.
Note that
.Fn pmap_activate
may not always be called when
.Fa l
is the current lwp.
.Fn pmap_activate
must be able to handle this scenario.
.It void Fn "pmap_deactivate" "struct lwp *l"
Deactivate the physical map used by the process behind lwp
.Fa l .
It is generally used in conjunction with
.Fn pmap_activate .
Like
.Fn pmap_activate ,
.Fn pmap_deactivate
may not always be called when
.Fa l
is the current lwp.
.It void Fn "pmap_zero_page" "paddr_t pa"
Zero the PAGE_SIZE sized region starting at physical address
.Fa pa .
The
.Nm
implementation must take whatever steps are necessary to map the
page to a kernel-accessible address and zero the page.
It is suggested that implementations use an optimized zeroing algorithm,
as the performance of this function directly impacts page fault performance.
The implementation may assume that the region is
PAGE_SIZE aligned and exactly PAGE_SIZE bytes in length.
.Pp
Note that the cache configuration of the platform should also be
considered in the implementation of
.Fn pmap_zero_page .
For example, on systems with a physically-addressed cache, the cache
load caused by zeroing the page will not be wasted, as the zeroing is
usually done on-demand.
However, on systems with a virtually-addressed cached, the cache load
caused by zeroing the page
.Em will
be wasted, as the page will be mapped at a virtual address which is
different from that used to zero the page.
In the virtually-addressed cache case, care should also be taken to
avoid cache alias problems.
.It void Fn "pmap_copy_page" "paddr_t src" "paddr_t dst"
Copy the PAGE_SIZE sized region starting at physical address
.Fa src
to the same sized region starting at physical address
.Fa dst .
The
.Nm
implementation must take whatever steps are necessary to map the
source and destination pages to a kernel-accessible address and
perform the copy.
It is suggested that implementations use an optimized copy algorithm,
as the performance of this function directly impacts page fault performance.
The implementation may assume that both regions are PAGE_SIZE aligned
and exactly PAGE_SIZE bytes in length.
.Pp
The same cache considerations that apply to
.Fn pmap_zero_page
apply to
.Fn pmap_copy_page .
.It void Fn "pmap_page_protect" "struct vm_page *pg" "vm_prot_t prot"
Lower the permissions for all mappings of the page
.Fa pg
to
.Fa prot .
This function is used by the virtual memory system to implement
copy-on-write (called with VM_PROT_READ set in
.Fa prot )
and to revoke all mappings when cleaning a page (called with
no bits set in
.Fa prot ) .
Access permissions must never be added to a page as a result of
this call.
.It bool Fn "pmap_clear_modify" "struct vm_page *pg"
Clear the
.Dq modified
attribute on the page
.Fa pg .
.Pp
The
.Fn pmap_clear_modify
function returns
.Dv true
or
.Dv false
indicating whether or not the
.Dq modified
attribute was set on the page before it was cleared.
.Pp
Note that this function may be provided as a C pre-processor macro.
.It bool Fn "pmap_clear_reference" "struct vm_page *pg"
Clear the
.Dq referenced
attribute on the page
.Fa pg .
.Pp
The
.Fn pmap_clear_reference
function returns
.Dv true
or
.Dv false
indicating whether or not the
.Dq referenced
attribute was set on the page before it was cleared.
.Pp
Note that this function may be provided as a C pre-processor macro.
.It bool Fn "pmap_is_modified" "struct vm_page *pg"
Test whether or not the
.Dq modified
attribute is set on page
.Fa pg .
.Pp
Note that this function may be provided as a C pre-processor macro.
.It bool Fn "pmap_is_referenced" "struct vm_page *pg"
Test whether or not the
.Dq referenced
attribute is set on page
.Fa pg .
.Pp
Note that this function may be provided as a C pre-processor macro.
.It paddr_t Fn "pmap_phys_address" "paddr_t cookie"
Convert a cookie returned by a device
.Fn mmap
function into a physical address.
This function is provided to accommodate systems which have physical
address spaces larger than can be directly addressed by the platform's
.Fa paddr_t
type.
The existence of this function is highly dubious, and it is
expected that this function will be removed from the
.Nm pmap
API in a future release of
.Nx .
.Pp
Note that this function may be provided as a C pre-processor macro.
.El
.Ss OPTIONAL FUNCTIONS
This section describes several optional functions in the
.Nm
API.
.Bl -tag -width indent -offset indent
.It vaddr_t Fn "pmap_steal_memory" "vsize_t size" "vaddr_t *vstartp" \
    "vaddr_t *vendp"
This function is a bootstrap memory allocator, which may be provided
as an alternative to the bootstrap memory allocator used within
.Xr uvm 9
itself.
It is particularly useful on systems which provide for example a direct-mapped
memory segment.
This function works by stealing pages from the (to be) managed memory
pool, which has already been provided to
.Xr uvm 9
in the vm_physmem[] array.
The pages are then mapped, or otherwise made accessible to the kernel,
in a machine-dependent way.
The memory must be zeroed by
.Fn pmap_steal_memory .
Note that memory allocated with
.Fn pmap_steal_memory
will never be freed, and mappings made by
.Fn pmap_steal_memory
must never be
.Dq forgotten .
.Pp
Note that
.Fn pmap_steal_memory
should not be used as a general-purpose early-startup memory
allocation routine.
It is intended to be used only by the
.Fn uvm_pageboot_alloc
routine and its supporting routines.
If you need to allocate memory before the virtual memory system is
initialized, use
.Fn uvm_pageboot_alloc .
See
.Xr uvm 9
for more information.
.Pp
The
.Fn pmap_steal_memory
function returns the kernel-accessible address of the allocated memory.
If no memory can be allocated, or if allocated memory cannot be mapped,
the function must panic.
.Pp
If the
.Fn pmap_steal_memory
function uses address space from the range provided to
.Xr uvm 9
by the
.Fn pmap_virtual_space
call, then
.Fn pmap_steal_memory
must adjust
.Fa *vstartp
and
.Fa *vendp
upon return.
.Pp
The
.Fn pmap_steal_memory
function is enabled by defining the C pre-processor macro
.Dv PMAP_STEAL_MEMORY
in
.In machine/pmap.h .
.It vaddr_t Fn "pmap_growkernel" "vaddr_t maxkvaddr"
Management of the kernel virtual address space is complicated by the
fact that it is not always safe to wait for resources with which to
map a kernel virtual address.
However, it is not always desirable to pre-allocate all resources
necessary to map the entire kernel virtual address space.
.Pp
The
.Fn pmap_growkernel
interface is designed to help alleviate this problem.
The virtual memory startup code may choose to allocate an initial set
of mapping resources (e.g., page tables) and set an internal variable
indicating how much kernel virtual address space can be mapped using
those initial resources.
Then, when the virtual memory system wishes to map something
at an address beyond that initial limit, it calls
.Fn pmap_growkernel
to pre-allocate more sources with which to create the mapping.
Note that once additional kernel virtual address space mapping resources
have been allocated, they should not be freed; it is likely they will
be needed again.
.Pp
The
.Fn pmap_growkernel
function returns the new maximum kernel virtual address that can be mapped
with the resources it has available.
If new resources cannot be allocated,
.Fn pmap_growkernel
must panic.
.Pp
The
.Fn pmap_growkernel
function is enabled by defining the C pre-processor macro
.Dv PMAP_GROWKERNEL
in
.In machine/pmap.h .
.It void Fn "pmap_fork" "pmap_t src_map" "pmap_t dst_map"
Some
.Nm
implementations may need to keep track of other information not
directly related to the virtual address space.
For example, on the i386 port, the Local Descriptor Table state of a
process is associated with the pmap (this is due to the fact that
applications manipulate the Local Descriptor Table directly expect it
to be logically associated with the virtual memory state of the process).
.Pp
The
.Fn pmap_fork
function is provided as a way to associate information from
.Fa src_map
with
.Fa dst_map
when a
.Dv vmspace
is forked.
.Fn pmap_fork
is called from
.Fn uvmspace_fork .
.Pp
The
.Fn pmap_fork
function is enabled by defining the C pre-processor macro
.Dv PMAP_FORK
in
.In machine/pmap.h .
.It vaddr_t Fn "PMAP_MAP_POOLPAGE" "paddr_t pa"
This function is used by the
.Xr pool 9
memory pool manager.
Pools allocate backing pages one at a time.
This is provided as a means to use hardware features such as a
direct-mapped memory segment to map the pages used by the
.Xr pool 9
allocator.
This can lead to better performance by e.g. reducing TLB contention.
.Pp
.Fn PMAP_MAP_POOLPAGE
returns the kernel-accessible address of the page being mapped.
It must always succeed.
.Pp
The use of
.Fn PMAP_MAP_POOLPAGE
is enabled by defining it as a C pre-processor macro in
.In machine/pmap.h .
If
.Fn PMAP_MAP_POOLPAGE
is defined,
.Fn PMAP_UNMAP_POOLPAGE
must also be defined.
.Pp
The following is an example of how to define
.Fn PMAP_MAP_POOLPAGE :
.Bd -literal -offset indent
#define PMAP_MAP_POOLPAGE(pa)   MIPS_PHYS_TO_KSEG0((pa))
.Ed
.Pp
This takes the physical address of a page and returns the KSEG0
address of that page on a MIPS processor.
.It paddr_t Fn "PMAP_UNMAP_POOLPAGE" "vaddr_t va"
This function is the inverse of
.Fn PMAP_MAP_POOLPAGE .
.Pp
.Fn PMAP_UNMAP_POOLPAGE
returns the physical address of the page corresponding to the
provided kernel-accessible address.
.Pp
The use of
.Fn PMAP_UNMAP_POOLPAGE
is enabled by defining it as a C pre-processor macro in
.In machine/pmap.h .
If
.Fn PMAP_UNMAP_POOLPAGE
is defined,
.Fn PMAP_MAP_POOLPAGE
must also be defined.
.Pp
The following is an example of how to define
.Fn PMAP_UNMAP_POOLPAGE :
.Bd -literal -offset indent
#define PMAP_UNMAP_POOLPAGE(pa) MIPS_KSEG0_TO_PHYS((va))
.Ed
.Pp
This takes the KSEG0 address of a previously-mapped pool page
and returns the physical address of that page on a MIPS processor.
.It void Fn "PMAP_PREFER" "vaddr_t hint" "vaddr_t *vap" "vsize_t sz" "int td"
This function is used by
.Xr uvm_map 9
to adjust a virtual address being allocated in order to avoid
cache alias problems.
If necessary, the virtual address pointed by
.Fa vap
will be advanced.
.Fa hint
is an object offset which will be mapped into the resulting virtual address, and
.Fa sz
is size of the region being mapped in bytes.
.Fa td
indicates if the machine dependent pmap uses the topdown VM.
.Pp
The use of
.Fn PMAP_PREFER
is enabled by defining it as a C pre-processor macro in
.In machine/pmap.h .
.Pp
.It void Fn "pmap_procwr" "struct proc *p" "vaddr_t va" "vsize_t size"
Synchronize CPU instruction caches of the specified range.
The address space is designated by
.Fa p .
This function is typically used to flush instruction caches
after code modification.
.Pp
The use of
.Fn pmap_procwr
is enabled by defining a C pre-processor macro
.Dv PMAP_NEED_PROCWR
in
.In machine/pmap.h .
.El
.Sh SEE ALSO
.Xr uvm 9
.Sh HISTORY
The
.Nm
module was originally part of the design of the virtual memory system
in the Mach Operating System.
The goal was to provide a clean separation between the machine-independent
and the machine-dependent portions of the virtual memory system, in
stark contrast to the original
.Bx 3
virtual memory system, which was specific to the VAX.
.Pp
Between
.Bx 4.3
and
.Bx 4.4 ,
the Mach virtual memory system, including the
.Nm
API, was ported to
.Bx
and included in the
.Bx 4.4
release.
.Pp
.Nx
inherited the
.Bx
version of the Mach virtual memory system.
.Nx 1.4
was the first
.Nx
release with the new
.Xr uvm 9
virtual memory system, which included several changes to the
.Nm
API.
Since the introduction of
.Xr uvm 9 ,
the
.Nm
API has evolved further.
.Sh AUTHORS
The original Mach VAX
.Nm
module was written by
.An Avadis Tevanian, Jr.
and
.An Michael Wayne Young .
.Pp
.An Mike Hibler
did the integration of the Mach virtual memory system into
.Bx 4.4
and implemented a
.Nm
module for the Motorola 68020+68851/68030/68040.
.Pp
The
.Nm
API as it exists in
.Nx
is derived from
.Bx 4.4 ,
and has been modified by
.An Chuck Cranor ,
.An Charles M. Hannum ,
.An Chuck Silvers ,
.An Wolfgang Solfrank ,
.An Bill Sommerfeld ,
and
.An Jason R. Thorpe .
.Pp
The author of this document is
.An Jason R. Thorpe
.Aq thorpej@NetBSD.org .
.Sh BUGS
The use and definition of
.Fn pmap_activate
and
.Fn pmap_deactivate
needs to be reexamined.
.Pp
The use of
.Fn pmap_copy
needs to be reexamined.
Empirical evidence suggests that performance of the system suffers when
.Fn pmap_copy
actually performs its defined function.
This is largely due to the fact that the copy of the virtual-to-physical
mappings is wasted if the process calls
.Xr execve 2
after
.Xr fork 2 .
For this reason, it is recommended that
.Nm
implementations leave the body of the
.Fn pmap_copy
function empty for now.