d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
31ce00f662
freebsd-skq/sys/vm/vm_radix.c

802 lines
20 KiB
C
Raw Normal View History

Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
* Copyright (c) 2013 EMC Corp.
* Copyright (c) 2011 Jeffrey Roberson <jeff@freebsd.org>
* Copyright (c) 2008 Mayur Shardul <mayur.shardul@gmail.com>
* 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 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 AUTHOR 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.
*
*/
/*
* Path-compressed radix trie implementation.
* The following code is not generalized into a general purpose library
* because there are way too many parameters embedded that should really
* be decided by the library consumers. At the same time, consumers
* of this code must achieve highest possible performance.
*
* The implementation takes into account the following rationale:
* - Size of the nodes should be as small as possible but still big enough
* to avoid a large maximum depth for the trie. This is a balance
* between the necessity to not wire too much physical memory for the nodes
* and the necessity to avoid too much cache pollution during the trie
* operations.
* - There is not a huge bias toward the number of lookup operations over
* the number of insert and remove operations. This basically implies
* that optimizations supposedly helping one operation but hurting the
* other might be carefully evaluated.
* - On average not many nodes are expected to be fully populated, hence
* level compression may just complicate things.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_ddb.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/vmmeter.h>
#include <vm/uma.h>
#include <vm/vm.h>
#include <vm/vm_param.h>
#include <vm/vm_page.h>
#include <vm/vm_radix.h>
#ifdef DDB
#include <ddb/ddb.h>
#endif
/*
* These widths should allow the pointers to a node's children to fit within
* a single cache line. The extra levels from a narrow width should not be
* a problem thanks to path compression.
*/
#ifdef __LP64__
#define VM_RADIX_WIDTH 4
#else
#define VM_RADIX_WIDTH 3
#endif
#define VM_RADIX_COUNT (1 << VM_RADIX_WIDTH)
#define VM_RADIX_MASK (VM_RADIX_COUNT - 1)
#define VM_RADIX_LIMIT \
Cleanup redundant parenthesis from existing howmany()/roundup() macro uses.
2016-04-22 16:57:42 +00:00
(howmany(sizeof(vm_pindex_t) * NBBY, VM_RADIX_WIDTH) - 1)
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/* Flag bits stored in node pointers. */
#define VM_RADIX_ISLEAF 0x1
#define VM_RADIX_FLAGS 0x1
#define VM_RADIX_PAD VM_RADIX_FLAGS
/* Returns one unit associated with specified level. */
#define VM_RADIX_UNITLEVEL(lev) \
To reduce the amount of arithmetic performed in the various radix tree functions, reverse the numbering scheme for the levels. The highest numbered level in the tree now appears near the root instead of the leaves. Sponsored by: EMC / Isilon Storage Division
2013-05-11 18:01:41 +00:00
((vm_pindex_t)1 << ((lev) * VM_RADIX_WIDTH))
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
struct vm_radix_node {
vm_pindex_t rn_owner; /* Owner of record. */
uint16_t rn_count; /* Valid children. */
uint16_t rn_clev; /* Current level. */
Micro-optimize the order of struct vm_radix_node's fields. Specifically, arrange for all of the fields to start at a short offset from the beginning of the structure. Eliminate unnecessary masking of VM_RADIX_FLAGS from the root pointer in vm_radix_getroot(). Sponsored by: EMC / Isilon Storage Division
2013-04-07 01:30:51 +00:00
void *rn_child[VM_RADIX_COUNT]; /* Child nodes. */
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
};
static uma_zone_t vm_radix_node_zone;
/*
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
* Allocate a radix node.
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
*/
static __inline struct vm_radix_node *
vm_radix_node_get(vm_pindex_t owner, uint16_t count, uint16_t clevel)
{
struct vm_radix_node *rnode;
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
rnode = uma_zalloc(vm_radix_node_zone, M_NOWAIT | M_ZERO);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
if (rnode == NULL)
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
return (NULL);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
rnode->rn_owner = owner;
rnode->rn_count = count;
rnode->rn_clev = clevel;
return (rnode);
}
/*
* Free radix node.
*/
static __inline void
vm_radix_node_put(struct vm_radix_node *rnode)
{
uma_zfree(vm_radix_node_zone, rnode);
}
/*
* Return the position in the array for a given level.
*/
static __inline int
vm_radix_slot(vm_pindex_t index, uint16_t level)
{
To reduce the amount of arithmetic performed in the various radix tree functions, reverse the numbering scheme for the levels. The highest numbered level in the tree now appears near the root instead of the leaves. Sponsored by: EMC / Isilon Storage Division
2013-05-11 18:01:41 +00:00
return ((index >> (level * VM_RADIX_WIDTH)) & VM_RADIX_MASK);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
/* Trims the key after the specified level. */
static __inline vm_pindex_t
vm_radix_trimkey(vm_pindex_t index, uint16_t level)
{
vm_pindex_t ret;
ret = index;
To reduce the amount of arithmetic performed in the various radix tree functions, reverse the numbering scheme for the levels. The highest numbered level in the tree now appears near the root instead of the leaves. Sponsored by: EMC / Isilon Storage Division
2013-05-11 18:01:41 +00:00
if (level > 0) {
ret >>= level * VM_RADIX_WIDTH;
ret <<= level * VM_RADIX_WIDTH;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
return (ret);
}
/*
* Get the root node for a radix tree.
*/
static __inline struct vm_radix_node *
vm_radix_getroot(struct vm_radix *rtree)
{
Micro-optimize the order of struct vm_radix_node's fields. Specifically, arrange for all of the fields to start at a short offset from the beginning of the structure. Eliminate unnecessary masking of VM_RADIX_FLAGS from the root pointer in vm_radix_getroot(). Sponsored by: EMC / Isilon Storage Division
2013-04-07 01:30:51 +00:00
return ((struct vm_radix_node *)rtree->rt_root);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
/*
* Set the root node for a radix tree.
*/
static __inline void
vm_radix_setroot(struct vm_radix *rtree, struct vm_radix_node *rnode)
{
rtree->rt_root = (uintptr_t)rnode;
}
Introduce vm_radix_isleaf() and use it in a couple places. As compared to using vm_radix_node_page() == NULL, the compiler is able to generate one less conditional branch when vm_radix_isleaf() is used. More use cases involving the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove() will follow. Reviewed by: attilio Sponsored by: EMC / Isilon Storage Division
2013-03-26 17:30:40 +00:00
/*
* Returns TRUE if the specified radix node is a leaf and FALSE otherwise.
*/
static __inline boolean_t
vm_radix_isleaf(struct vm_radix_node *rnode)
{
return (((uintptr_t)rnode & VM_RADIX_ISLEAF) != 0);
}
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
* Returns the associated page extracted from rnode.
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
*/
static __inline vm_page_t
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
vm_radix_topage(struct vm_radix_node *rnode)
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
{
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
return ((vm_page_t)((uintptr_t)rnode & ~VM_RADIX_FLAGS));
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
/*
* Adds the page as a child of the provided node.
*/
static __inline void
vm_radix_addpage(struct vm_radix_node *rnode, vm_pindex_t index, uint16_t clev,
vm_page_t page)
{
int slot;
slot = vm_radix_slot(index, clev);
rnode->rn_child[slot] = (void *)((uintptr_t)page | VM_RADIX_ISLEAF);
}
/*
* Returns the slot where two keys differ.
* It cannot accept 2 equal keys.
*/
static __inline uint16_t
vm_radix_keydiff(vm_pindex_t index1, vm_pindex_t index2)
{
uint16_t clev;
KASSERT(index1 != index2, ("%s: passing the same key value %jx",
__func__, (uintmax_t)index1));
index1 ^= index2;
To reduce the amount of arithmetic performed in the various radix tree functions, reverse the numbering scheme for the levels. The highest numbered level in the tree now appears near the root instead of the leaves. Sponsored by: EMC / Isilon Storage Division
2013-05-11 18:01:41 +00:00
for (clev = VM_RADIX_LIMIT;; clev--)
Remove a redundant call to panic() from vm_radix_keydiff(). The assertion before the loop accomplishes the same thing. Sponsored by: EMC / Isilon Storage Division
2013-05-07 18:45:34 +00:00
if (vm_radix_slot(index1, clev) != 0)
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
return (clev);
}
/*
* Returns TRUE if it can be determined that key does not belong to the
* specified rnode. Otherwise, returns FALSE.
*/
static __inline boolean_t
vm_radix_keybarr(struct vm_radix_node *rnode, vm_pindex_t idx)
{
To reduce the amount of arithmetic performed in the various radix tree functions, reverse the numbering scheme for the levels. The highest numbered level in the tree now appears near the root instead of the leaves. Sponsored by: EMC / Isilon Storage Division
2013-05-11 18:01:41 +00:00
if (rnode->rn_clev < VM_RADIX_LIMIT) {
idx = vm_radix_trimkey(idx, rnode->rn_clev + 1);
Simplify vm_radix_keybarr(). Sponsored by: EMC / Isilon Storage Division
2013-04-06 18:04:35 +00:00
return (idx != rnode->rn_owner);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
return (FALSE);
}
/*
* Internal helper for vm_radix_reclaim_allnodes().
* This function is recursive.
*/
static void
vm_radix_reclaim_allnodes_int(struct vm_radix_node *rnode)
{
int slot;
Micro-optimize the control flow in a few places. Eliminate a panic call that could never be reached in vm_radix_insert(). (If the pointer being checked by the panic call were ever NULL, the immmediately preceding loop would have already crashed on a NULL pointer dereference.) Reviewed by: attilio (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-03-24 16:43:07 +00:00
KASSERT(rnode->rn_count <= VM_RADIX_COUNT,
("vm_radix_reclaim_allnodes_int: bad count in rnode %p", rnode));
for (slot = 0; rnode->rn_count != 0; slot++) {
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
if (rnode->rn_child[slot] == NULL)
continue;
Introduce vm_radix_isleaf() and use it in a couple places. As compared to using vm_radix_node_page() == NULL, the compiler is able to generate one less conditional branch when vm_radix_isleaf() is used. More use cases involving the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove() will follow. Reviewed by: attilio Sponsored by: EMC / Isilon Storage Division
2013-03-26 17:30:40 +00:00
if (!vm_radix_isleaf(rnode->rn_child[slot]))
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
vm_radix_reclaim_allnodes_int(rnode->rn_child[slot]);
rnode->rn_child[slot] = NULL;
rnode->rn_count--;
}
vm_radix_node_put(rnode);
}
#ifdef INVARIANTS
/*
* Radix node zone destructor.
*/
static void
vm_radix_node_zone_dtor(void *mem, int size __unused, void *arg __unused)
{
struct vm_radix_node *rnode;
int slot;
rnode = mem;
KASSERT(rnode->rn_count == 0,
("vm_radix_node_put: rnode %p has %d children", rnode,
rnode->rn_count));
for (slot = 0; slot < VM_RADIX_COUNT; slot++)
KASSERT(rnode->rn_child[slot] == NULL,
("vm_radix_node_put: rnode %p has a child", rnode));
}
#endif
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
#ifndef UMA_MD_SMALL_ALLOC
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
* Reserve the KVA necessary to satisfy the node allocation.
* This is mandatory in architectures not supporting direct
* mapping as they will need otherwise to carve into the kernel maps for
* every node allocation, resulting into deadlocks for consumers already
* working with kernel maps.
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
*/
static void
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
vm_radix_reserve_kva(void *arg __unused)
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
{
When calculating the number of reserved nodes, discount the pages that will be used to store the nodes. Sponsored by: EMC / Isilon Storage Division
2013-04-18 05:34:33 +00:00
/*
* Calculate the number of reserved nodes, discounting the pages that
* are needed to store them.
*/
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
if (!uma_zone_reserve_kva(vm_radix_node_zone,
Rename global cnt to vm_cnt to avoid shadowing. To reduce the diff struct pcu.cnt field was not renamed, so PCPU_OP(cnt.field) is still used. pc_cnt and pcpu are also used in kvm(3) and vmstat(8). The goal was to not affect externally used KPI. Bump __FreeBSD_version_ in case some out-of-tree module/code relies on the the global cnt variable. Exp-run revealed no ports using it directly. No objection from: arch@ Sponsored by: EMC / Isilon Storage Division
2014-03-22 10:26:09 +00:00
((vm_paddr_t)vm_cnt.v_page_count * PAGE_SIZE) / (PAGE_SIZE +
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
sizeof(struct vm_radix_node))))
panic("%s: unable to reserve KVA", __func__);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
Pull in r267961 and r267973 again. Fix for issues reported will follow.
2014-06-28 03:56:17 +00:00
SYSINIT(vm_radix_reserve_kva, SI_SUB_KMEM, SI_ORDER_THIRD,
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
vm_radix_reserve_kva, NULL);
#endif
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
* Initialize the UMA slab zone.
*/
void
vm_radix_init(void)
{
vm_radix_node_zone = uma_zcreate("RADIX NODE",
sizeof(struct vm_radix_node), NULL,
#ifdef INVARIANTS
vm_radix_node_zone_dtor,
#else
NULL,
#endif
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
NULL, NULL, VM_RADIX_PAD, UMA_ZONE_VM);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
/*
* Inserts the key-value pair into the trie.
* Panics if the key already exists.
*/
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
int
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
vm_radix_insert(struct vm_radix *rtree, vm_page_t page)
{
vm_pindex_t index, newind;
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
void **parentp;
struct vm_radix_node *rnode, *tmp;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
vm_page_t m;
int slot;
uint16_t clev;
index = page->pindex;
/*
* The owner of record for root is not really important because it
* will never be used.
*/
rnode = vm_radix_getroot(rtree);
if (rnode == NULL) {
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
rtree->rt_root = (uintptr_t)page | VM_RADIX_ISLEAF;
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
return (0);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
parentp = (void **)&rtree->rt_root;
for (;;) {
if (vm_radix_isleaf(rnode)) {
m = vm_radix_topage(rnode);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
if (m->pindex == index)
panic("%s: key %jx is already present",
__func__, (uintmax_t)index);
clev = vm_radix_keydiff(m->pindex, index);
tmp = vm_radix_node_get(vm_radix_trimkey(index,
To reduce the amount of arithmetic performed in the various radix tree functions, reverse the numbering scheme for the levels. The highest numbered level in the tree now appears near the root instead of the leaves. Sponsored by: EMC / Isilon Storage Division
2013-05-11 18:01:41 +00:00
clev + 1), 2, clev);
During vm_page_cache()'s call to vm_radix_insert(), if vm_page_alloc() was called to allocate a new page of radix trie nodes, there could be a call to vm_radix_remove() on the same trie (of PG_CACHED pages) as the in-progress vm_radix_insert(). With the removal of PG_CACHED pages, we can simplify vm_radix_insert() and vm_radix_remove() by removing the flags on the root of the trie that were used to detect this case and the code for restarting vm_radix_insert() when it happened. Reviewed by: kib, markj Tested by: pho Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D8664
2016-12-01 17:26:37 +00:00
if (tmp == NULL)
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
return (ENOMEM);
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
*parentp = tmp;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
vm_radix_addpage(tmp, index, clev, page);
vm_radix_addpage(tmp, m->pindex, clev, m);
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
return (0);
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
} else if (vm_radix_keybarr(rnode, index))
break;
slot = vm_radix_slot(index, rnode->rn_clev);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
if (rnode->rn_child[slot] == NULL) {
rnode->rn_count++;
vm_radix_addpage(rnode, index, rnode->rn_clev, page);
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
return (0);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
parentp = &rnode->rn_child[slot];
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
rnode = rnode->rn_child[slot];
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
}
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
* A new node is needed because the right insertion level is reached.
* Setup the new intermediate node and add the 2 children: the
* new object and the older edge.
*/
Simplify vm_radix_insert(). Reviewed by: attilio Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-06 06:02:55 +00:00
newind = rnode->rn_owner;
clev = vm_radix_keydiff(newind, index);
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
tmp = vm_radix_node_get(vm_radix_trimkey(index, clev + 1), 2, clev);
During vm_page_cache()'s call to vm_radix_insert(), if vm_page_alloc() was called to allocate a new page of radix trie nodes, there could be a call to vm_radix_remove() on the same trie (of PG_CACHED pages) as the in-progress vm_radix_insert(). With the removal of PG_CACHED pages, we can simplify vm_radix_insert() and vm_radix_remove() by removing the flags on the root of the trie that were used to detect this case and the code for restarting vm_radix_insert() when it happened. Reviewed by: kib, markj Tested by: pho Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D8664
2016-12-01 17:26:37 +00:00
if (tmp == NULL)
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
return (ENOMEM);
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
*parentp = tmp;
Simplify vm_radix_insert(). Reviewed by: attilio Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-06 06:02:55 +00:00
vm_radix_addpage(tmp, index, clev, page);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
slot = vm_radix_slot(newind, clev);
Simplify vm_radix_insert(). Reviewed by: attilio Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-06 06:02:55 +00:00
tmp->rn_child[slot] = rnode;
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
return (0);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
Addendum to r254141: The call to vm_radix_insert() in vm_page_cache() can reclaim the last preexisting cached page in the object, resulting in a call to vdrop(). Detect this scenario so that the vnode's hold count is correctly maintained. Otherwise, we panic. Reported by: scottl Tested by: pho Discussed with: attilio, jeff, kib
2013-08-23 17:27:12 +00:00
/*
* Returns TRUE if the specified radix tree contains a single leaf and FALSE
* otherwise.
*/
boolean_t
vm_radix_is_singleton(struct vm_radix *rtree)
{
struct vm_radix_node *rnode;
rnode = vm_radix_getroot(rtree);
if (rnode == NULL)
return (FALSE);
return (vm_radix_isleaf(rnode));
}
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
* Returns the value stored at the index. If the index is not present,
* NULL is returned.
*/
vm_page_t
vm_radix_lookup(struct vm_radix *rtree, vm_pindex_t index)
{
struct vm_radix_node *rnode;
vm_page_t m;
int slot;
rnode = vm_radix_getroot(rtree);
while (rnode != NULL) {
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
if (vm_radix_isleaf(rnode)) {
m = vm_radix_topage(rnode);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
if (m->pindex == index)
return (m);
else
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we always create a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the first step in eliminating those unnecessary level zero interior nodes. Specifically, it updates all of the lookup functions so that they do not require a level zero interior node at the root. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-12 20:21:28 +00:00
break;
} else if (vm_radix_keybarr(rnode, index))
break;
slot = vm_radix_slot(index, rnode->rn_clev);
rnode = rnode->rn_child[slot];
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
return (NULL);
}
/*
* Look up the nearest entry at a position bigger than or equal to index.
*/
vm_page_t
vm_radix_lookup_ge(struct vm_radix *rtree, vm_pindex_t index)
{
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
struct vm_radix_node *stack[VM_RADIX_LIMIT];
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
vm_pindex_t inc;
vm_page_t m;
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
struct vm_radix_node *child, *rnode;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
#ifdef INVARIANTS
int loops = 0;
#endif
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
int slot, tos;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we always create a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the first step in eliminating those unnecessary level zero interior nodes. Specifically, it updates all of the lookup functions so that they do not require a level zero interior node at the root. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-12 20:21:28 +00:00
rnode = vm_radix_getroot(rtree);
if (rnode == NULL)
return (NULL);
else if (vm_radix_isleaf(rnode)) {
m = vm_radix_topage(rnode);
if (m->pindex >= index)
return (m);
else
return (NULL);
}
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
tos = 0;
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we always create a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the first step in eliminating those unnecessary level zero interior nodes. Specifically, it updates all of the lookup functions so that they do not require a level zero interior node at the root. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-12 20:21:28 +00:00
for (;;) {
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
Avoid some lookup restarts in vm_radix_lookup_{ge,le}(). Sponsored by: EMC / Isilon Storage Division
2013-04-27 16:44:59 +00:00
* If the keys differ before the current bisection node,
* then the search key might rollback to the earliest
* available bisection node or to the smallest key
* in the current node (if the owner is bigger than the
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
* search key).
*/
if (vm_radix_keybarr(rnode, index)) {
if (index > rnode->rn_owner) {
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
ascend:
KASSERT(++loops < 1000,
("vm_radix_lookup_ge: too many loops"));
/*
* Pop nodes from the stack until either the
* stack is empty or a node that could have a
* matching descendant is found.
*/
do {
if (tos == 0)
return (NULL);
rnode = stack[--tos];
} while (vm_radix_slot(index,
rnode->rn_clev) == (VM_RADIX_COUNT - 1));
/*
* The following computation cannot overflow
* because index's slot at the current level
* is less than VM_RADIX_COUNT - 1.
*/
index = vm_radix_trimkey(index,
rnode->rn_clev);
index += VM_RADIX_UNITLEVEL(rnode->rn_clev);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
} else
Avoid some lookup restarts in vm_radix_lookup_{ge,le}(). Sponsored by: EMC / Isilon Storage Division
2013-04-27 16:44:59 +00:00
index = rnode->rn_owner;
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
KASSERT(!vm_radix_keybarr(rnode, index),
("vm_radix_lookup_ge: keybarr failed"));
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
slot = vm_radix_slot(index, rnode->rn_clev);
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
child = rnode->rn_child[slot];
if (vm_radix_isleaf(child)) {
m = vm_radix_topage(child);
if (m->pindex >= index)
return (m);
} else if (child != NULL)
goto descend;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
* Look for an available edge or page within the current
* bisection node.
*/
if (slot < (VM_RADIX_COUNT - 1)) {
inc = VM_RADIX_UNITLEVEL(rnode->rn_clev);
index = vm_radix_trimkey(index, rnode->rn_clev);
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
do {
index += inc;
slot++;
child = rnode->rn_child[slot];
if (vm_radix_isleaf(child)) {
m = vm_radix_topage(child);
if (m->pindex >= index)
return (m);
} else if (child != NULL)
goto descend;
} while (slot < (VM_RADIX_COUNT - 1));
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
KASSERT(child == NULL || vm_radix_isleaf(child),
("vm_radix_lookup_ge: child is radix node"));
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
* If a page or edge bigger than the search slot is not found
* in the current node, ascend to the next higher-level node.
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
*/
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
goto ascend;
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
descend:
To reduce the amount of arithmetic performed in the various radix tree functions, reverse the numbering scheme for the levels. The highest numbered level in the tree now appears near the root instead of the leaves. Sponsored by: EMC / Isilon Storage Division
2013-05-11 18:01:41 +00:00
KASSERT(rnode->rn_clev > 0,
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
("vm_radix_lookup_ge: pushing leaf's parent"));
KASSERT(tos < VM_RADIX_LIMIT,
("vm_radix_lookup_ge: stack overflow"));
stack[tos++] = rnode;
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
rnode = child;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
}
/*
* Look up the nearest entry at a position less than or equal to index.
*/
vm_page_t
vm_radix_lookup_le(struct vm_radix *rtree, vm_pindex_t index)
{
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
struct vm_radix_node *stack[VM_RADIX_LIMIT];
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
vm_pindex_t inc;
vm_page_t m;
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
struct vm_radix_node *child, *rnode;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
#ifdef INVARIANTS
int loops = 0;
#endif
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
int slot, tos;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we always create a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the first step in eliminating those unnecessary level zero interior nodes. Specifically, it updates all of the lookup functions so that they do not require a level zero interior node at the root. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-12 20:21:28 +00:00
rnode = vm_radix_getroot(rtree);
if (rnode == NULL)
return (NULL);
else if (vm_radix_isleaf(rnode)) {
m = vm_radix_topage(rnode);
if (m->pindex <= index)
return (m);
else
return (NULL);
}
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
tos = 0;
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we always create a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the first step in eliminating those unnecessary level zero interior nodes. Specifically, it updates all of the lookup functions so that they do not require a level zero interior node at the root. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-12 20:21:28 +00:00
for (;;) {
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
Avoid some lookup restarts in vm_radix_lookup_{ge,le}(). Sponsored by: EMC / Isilon Storage Division
2013-04-27 16:44:59 +00:00
* If the keys differ before the current bisection node,
* then the search key might rollback to the earliest
* available bisection node or to the largest key
* in the current node (if the owner is smaller than the
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
* search key).
*/
if (vm_radix_keybarr(rnode, index)) {
if (index > rnode->rn_owner) {
Avoid some lookup restarts in vm_radix_lookup_{ge,le}(). Sponsored by: EMC / Isilon Storage Division
2013-04-27 16:44:59 +00:00
index = rnode->rn_owner + VM_RADIX_COUNT *
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
VM_RADIX_UNITLEVEL(rnode->rn_clev);
Avoid some lookup restarts in vm_radix_lookup_{ge,le}(). Sponsored by: EMC / Isilon Storage Division
2013-04-27 16:44:59 +00:00
} else {
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
ascend:
KASSERT(++loops < 1000,
("vm_radix_lookup_le: too many loops"));
/*
* Pop nodes from the stack until either the
* stack is empty or a node that could have a
* matching descendant is found.
*/
do {
if (tos == 0)
return (NULL);
rnode = stack[--tos];
} while (vm_radix_slot(index,
rnode->rn_clev) == 0);
/*
* The following computation cannot overflow
* because index's slot at the current level
* is greater than 0.
*/
index = vm_radix_trimkey(index,
rnode->rn_clev);
Avoid some lookup restarts in vm_radix_lookup_{ge,le}(). Sponsored by: EMC / Isilon Storage Division
2013-04-27 16:44:59 +00:00
}
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
index--;
KASSERT(!vm_radix_keybarr(rnode, index),
("vm_radix_lookup_le: keybarr failed"));
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
slot = vm_radix_slot(index, rnode->rn_clev);
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
child = rnode->rn_child[slot];
if (vm_radix_isleaf(child)) {
m = vm_radix_topage(child);
if (m->pindex <= index)
return (m);
} else if (child != NULL)
goto descend;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
* Look for an available edge or page within the current
* bisection node.
*/
if (slot > 0) {
inc = VM_RADIX_UNITLEVEL(rnode->rn_clev);
index |= inc - 1;
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
do {
index -= inc;
slot--;
child = rnode->rn_child[slot];
if (vm_radix_isleaf(child)) {
m = vm_radix_topage(child);
if (m->pindex <= index)
return (m);
} else if (child != NULL)
goto descend;
} while (slot > 0);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
KASSERT(child == NULL || vm_radix_isleaf(child),
("vm_radix_lookup_le: child is radix node"));
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
/*
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
* If a page or edge smaller than the search slot is not found
* in the current node, ascend to the next higher-level node.
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
*/
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
goto ascend;
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
descend:
To reduce the amount of arithmetic performed in the various radix tree functions, reverse the numbering scheme for the levels. The highest numbered level in the tree now appears near the root instead of the leaves. Sponsored by: EMC / Isilon Storage Division
2013-05-11 18:01:41 +00:00
KASSERT(rnode->rn_clev > 0,
Optimize vm_radix_lookup_ge() and vm_radix_lookup_le(). Specifically, change the way that these functions ascend the tree when the search for a matching leaf fails at an interior node. Rather than returning to the root of the tree and repeating the lookup with an updated key, maintain a stack of interior nodes that were visited during the descent and use that stack to resume the lookup at the closest ancestor that might have a matching descendant. Sponsored by: EMC / Isilon Storage Division Reviewed by: attilio Tested by: pho
2013-05-04 22:50:15 +00:00
("vm_radix_lookup_le: pushing leaf's parent"));
KASSERT(tos < VM_RADIX_LIMIT,
("vm_radix_lookup_le: stack overflow"));
stack[tos++] = rnode;
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
rnode = child;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
}
/*
Previously, vm_radix_remove() would panic if the radix trie didn't contain a vm_page_t at the specified index. However, with this change, vm_radix_remove() no longer panics. Instead, it returns NULL if there is no vm_page_t at the specified index. Otherwise, it returns the vm_page_t. The motivation for this change is that it simplifies the use of radix tries in the amd64, arm64, and i386 pmap implementations. Instead of performing a lookup before every remove, the pmap can simply perform the remove. Reviewed by: kib, markj Differential Revision: https://reviews.freebsd.org/D8708
2016-12-08 04:29:29 +00:00
* Remove the specified index from the trie, and return the value stored at
* that index. If the index is not present, return NULL.
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
*/
Previously, vm_radix_remove() would panic if the radix trie didn't contain a vm_page_t at the specified index. However, with this change, vm_radix_remove() no longer panics. Instead, it returns NULL if there is no vm_page_t at the specified index. Otherwise, it returns the vm_page_t. The motivation for this change is that it simplifies the use of radix tries in the amd64, arm64, and i386 pmap implementations. Instead of performing a lookup before every remove, the pmap can simply perform the remove. Reviewed by: kib, markj Differential Revision: https://reviews.freebsd.org/D8708
2016-12-08 04:29:29 +00:00
vm_page_t
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
vm_radix_remove(struct vm_radix *rtree, vm_pindex_t index)
{
struct vm_radix_node *rnode, *parent;
vm_page_t m;
int i, slot;
rnode = vm_radix_getroot(rtree);
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
if (vm_radix_isleaf(rnode)) {
m = vm_radix_topage(rnode);
if (m->pindex != index)
Previously, vm_radix_remove() would panic if the radix trie didn't contain a vm_page_t at the specified index. However, with this change, vm_radix_remove() no longer panics. Instead, it returns NULL if there is no vm_page_t at the specified index. Otherwise, it returns the vm_page_t. The motivation for this change is that it simplifies the use of radix tries in the amd64, arm64, and i386 pmap implementations. Instead of performing a lookup before every remove, the pmap can simply perform the remove. Reviewed by: kib, markj Differential Revision: https://reviews.freebsd.org/D8708
2016-12-08 04:29:29 +00:00
return (NULL);
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
vm_radix_setroot(rtree, NULL);
Previously, vm_radix_remove() would panic if the radix trie didn't contain a vm_page_t at the specified index. However, with this change, vm_radix_remove() no longer panics. Instead, it returns NULL if there is no vm_page_t at the specified index. Otherwise, it returns the vm_page_t. The motivation for this change is that it simplifies the use of radix tries in the amd64, arm64, and i386 pmap implementations. Instead of performing a lookup before every remove, the pmap can simply perform the remove. Reviewed by: kib, markj Differential Revision: https://reviews.freebsd.org/D8708
2016-12-08 04:29:29 +00:00
return (m);
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
}
parent = NULL;
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
for (;;) {
if (rnode == NULL)
Previously, vm_radix_remove() would panic if the radix trie didn't contain a vm_page_t at the specified index. However, with this change, vm_radix_remove() no longer panics. Instead, it returns NULL if there is no vm_page_t at the specified index. Otherwise, it returns the vm_page_t. The motivation for this change is that it simplifies the use of radix tries in the amd64, arm64, and i386 pmap implementations. Instead of performing a lookup before every remove, the pmap can simply perform the remove. Reviewed by: kib, markj Differential Revision: https://reviews.freebsd.org/D8708
2016-12-08 04:29:29 +00:00
return (NULL);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
slot = vm_radix_slot(index, rnode->rn_clev);
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
if (vm_radix_isleaf(rnode->rn_child[slot])) {
m = vm_radix_topage(rnode->rn_child[slot]);
if (m->pindex != index)
Previously, vm_radix_remove() would panic if the radix trie didn't contain a vm_page_t at the specified index. However, with this change, vm_radix_remove() no longer panics. Instead, it returns NULL if there is no vm_page_t at the specified index. Otherwise, it returns the vm_page_t. The motivation for this change is that it simplifies the use of radix tries in the amd64, arm64, and i386 pmap implementations. Instead of performing a lookup before every remove, the pmap can simply perform the remove. Reviewed by: kib, markj Differential Revision: https://reviews.freebsd.org/D8708
2016-12-08 04:29:29 +00:00
return (NULL);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
rnode->rn_child[slot] = NULL;
rnode->rn_count--;
if (rnode->rn_count > 1)
Previously, vm_radix_remove() would panic if the radix trie didn't contain a vm_page_t at the specified index. However, with this change, vm_radix_remove() no longer panics. Instead, it returns NULL if there is no vm_page_t at the specified index. Otherwise, it returns the vm_page_t. The motivation for this change is that it simplifies the use of radix tries in the amd64, arm64, and i386 pmap implementations. Instead of performing a lookup before every remove, the pmap can simply perform the remove. Reviewed by: kib, markj Differential Revision: https://reviews.freebsd.org/D8708
2016-12-08 04:29:29 +00:00
return (m);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
for (i = 0; i < VM_RADIX_COUNT; i++)
if (rnode->rn_child[i] != NULL)
break;
KASSERT(i != VM_RADIX_COUNT,
("%s: invalid node configuration", __func__));
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
if (parent == NULL)
vm_radix_setroot(rtree, rnode->rn_child[i]);
else {
slot = vm_radix_slot(index, parent->rn_clev);
KASSERT(parent->rn_child[slot] == rnode,
("%s: invalid child value", __func__));
parent->rn_child[slot] = rnode->rn_child[i];
}
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
rnode->rn_count--;
rnode->rn_child[i] = NULL;
vm_radix_node_put(rnode);
Previously, vm_radix_remove() would panic if the radix trie didn't contain a vm_page_t at the specified index. However, with this change, vm_radix_remove() no longer panics. Instead, it returns NULL if there is no vm_page_t at the specified index. Otherwise, it returns the vm_page_t. The motivation for this change is that it simplifies the use of radix tries in the amd64, arm64, and i386 pmap implementations. Instead of performing a lookup before every remove, the pmap can simply perform the remove. Reviewed by: kib, markj Differential Revision: https://reviews.freebsd.org/D8708
2016-12-08 04:29:29 +00:00
return (m);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
parent = rnode;
rnode = rnode->rn_child[slot];
}
}
/*
* Remove and free all the nodes from the radix tree.
* This function is recursive but there is a tight control on it as the
* maximum depth of the tree is fixed.
*/
void
vm_radix_reclaim_allnodes(struct vm_radix *rtree)
{
struct vm_radix_node *root;
root = vm_radix_getroot(rtree);
if (root == NULL)
return;
vm_radix_setroot(rtree, NULL);
Although we perform path compression to reduce the height of the trie and the number of interior nodes, we have previously created a level zero interior node at the root of every non-empty trie, even when that node is not strictly necessary, i.e., it has only one child. This change is the second (and final) step in eliminating those unnecessary level zero interior nodes. Specifically, it updates the deletion and insertion functions so that they do not require a level zero interior node at the root of the trie. For a "buildworld" workload, this change results in a 16.8% reduction in the number of interior nodes allocated and a similar reduction in the average execution time for lookup functions. For example, the average execution time for a call to vm_radix_lookup_ge() is reduced by 22.9%. Reviewed by: attilio, jeff (an earlier version) Sponsored by: EMC / Isilon Storage Division
2013-04-15 06:12:00 +00:00
if (!vm_radix_isleaf(root))
vm_radix_reclaim_allnodes_int(root);
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
}
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
/*
Eliminate a redundant parameter to vm_radix_replace(). Improve the wording of the comment describing vm_radix_replace(). Reviewed by: attilio MFC after: 6 weeks Sponsored by: EMC / Isilon Storage Division
2013-12-08 20:07:02 +00:00
* Replace an existing page in the trie with another one.
* Panics if there is not an old page in the trie at the new page's index.
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
*/
vm_page_t
Eliminate a redundant parameter to vm_radix_replace(). Improve the wording of the comment describing vm_radix_replace(). Reviewed by: attilio MFC after: 6 weeks Sponsored by: EMC / Isilon Storage Division
2013-12-08 20:07:02 +00:00
vm_radix_replace(struct vm_radix *rtree, vm_page_t newpage)
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
{
struct vm_radix_node *rnode;
vm_page_t m;
Eliminate a redundant parameter to vm_radix_replace(). Improve the wording of the comment describing vm_radix_replace(). Reviewed by: attilio MFC after: 6 weeks Sponsored by: EMC / Isilon Storage Division
2013-12-08 20:07:02 +00:00
vm_pindex_t index;
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
int slot;
Eliminate a redundant parameter to vm_radix_replace(). Improve the wording of the comment describing vm_radix_replace(). Reviewed by: attilio MFC after: 6 weeks Sponsored by: EMC / Isilon Storage Division
2013-12-08 20:07:02 +00:00
index = newpage->pindex;
On all the architectures, avoid to preallocate the physical memory for nodes used in vm_radix. On architectures supporting direct mapping, also avoid to pre-allocate the KVA for such nodes. In order to do so make the operations derived from vm_radix_insert() to fail and handle all the deriving failure of those. vm_radix-wise introduce a new function called vm_radix_replace(), which can replace a leaf node, already present, with a new one, and take into account the possibility, during vm_radix_insert() allocation, that the operations on the radix trie can recurse. This means that if operations in vm_radix_insert() recursed vm_radix_insert() will start from scratch again. Sponsored by: EMC / Isilon storage division Reviewed by: alc (older version) Reviewed by: jeff Tested by: pho, scottl
2013-08-09 11:28:55 +00:00
rnode = vm_radix_getroot(rtree);
if (rnode == NULL)
panic("%s: replacing page on an empty trie", __func__);
if (vm_radix_isleaf(rnode)) {
m = vm_radix_topage(rnode);
if (m->pindex != index)
panic("%s: original replacing root key not found",
__func__);
rtree->rt_root = (uintptr_t)newpage | VM_RADIX_ISLEAF;
return (m);
}
for (;;) {
slot = vm_radix_slot(index, rnode->rn_clev);
if (vm_radix_isleaf(rnode->rn_child[slot])) {
m = vm_radix_topage(rnode->rn_child[slot]);
if (m->pindex == index) {
rnode->rn_child[slot] =
(void *)((uintptr_t)newpage |
VM_RADIX_ISLEAF);
return (m);
} else
break;
} else if (rnode->rn_child[slot] == NULL ||
vm_radix_keybarr(rnode->rn_child[slot], index))
break;
rnode = rnode->rn_child[slot];
}
panic("%s: original replacing page not found", __func__);
}
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
#ifdef DDB
/*
* Show details about the given radix node.
*/
DB_SHOW_COMMAND(radixnode, db_show_radixnode)
{
struct vm_radix_node *rnode;
int i;
if (!have_addr)
return;
rnode = (struct vm_radix_node *)addr;
db_printf("radixnode %p, owner %jx, children count %u, level %u:\n",
(void *)rnode, (uintmax_t)rnode->rn_owner, rnode->rn_count,
rnode->rn_clev);
for (i = 0; i < VM_RADIX_COUNT; i++)
if (rnode->rn_child[i] != NULL)
db_printf("slot: %d, val: %p, page: %p, clev: %d\n",
i, (void *)rnode->rn_child[i],
Replace the remaining uses of vm_radix_node_page() by vm_radix_isleaf() and vm_radix_topage(). This transformation eliminates some unnecessary conditional branches from the inner loops of vm_radix_insert(), vm_radix_lookup{,_ge,_le}(), and vm_radix_remove(). Simplify the control flow of vm_radix_lookup_{ge,le}(). Reviewed by: attilio (an earlier version) Tested by: pho Sponsored by: EMC / Isilon Storage Division
2013-04-03 06:37:25 +00:00
vm_radix_isleaf(rnode->rn_child[i]) ?
vm_radix_topage(rnode->rn_child[i]) : NULL,
Sync back vmcontention branch into HEAD: Replace the per-object resident and cached pages splay tree with a path-compressed multi-digit radix trie. Along with this, switch also the x86-specific handling of idle page tables to using the radix trie. This change is supposed to do the following: - Allowing the acquisition of read locking for lookup operations of the resident/cached pages collections as the per-vm_page_t splay iterators are now removed. - Increase the scalability of the operations on the page collections. The radix trie does rely on the consumers locking to ensure atomicity of its operations. In order to avoid deadlocks the bisection nodes are pre-allocated in the UMA zone. This can be done safely because the algorithm needs at maximum one new node per insert which means the maximum number of the desired nodes is the number of available physical frames themselves. However, not all the times a new bisection node is really needed. The radix trie implements path-compression because UFS indirect blocks can lead to several objects with a very sparse trie, increasing the number of levels to usually scan. It also helps in the nodes pre-fetching by introducing the single node per-insert property. This code is not generalized (yet) because of the possible loss of performance by having much of the sizes in play configurable. However, efforts to make this code more general and then reusable in further different consumers might be really done. The only KPI change is the removal of the function vm_page_splay() which is now reaped. The only KBI change, instead, is the removal of the left/right iterators from struct vm_page, which are now reaped. Further technical notes broken into mealpieces can be retrieved from the svn branch: http://svn.freebsd.org/base/user/attilio/vmcontention/ Sponsored by: EMC / Isilon storage division In collaboration with: alc, jeff Tested by: flo, pho, jhb, davide Tested by: ian (arm) Tested by: andreast (powerpc)
2013-03-18 00:25:02 +00:00
rnode->rn_clev);
}
#endif /* DDB */
Reference in New Issue Copy Permalink