1ppgsz(1) User Commands ppgsz(1)
2
6 ppgsz - set preferred page size for stack, heap, and/or other anonymous
7 segments
8
10 /usr/bin/ppgsz [-F] -o option[,option] cmd | -p pid...
11
14 The ppgsz utility sets the preferred page size for stack, heap, and/or
15 other anonymous segments for the target process(es), that is, the
16 launched cmd or the process(es) in the pid list. ppgsz stops the target
17 process(es) while changing the page size. See memcntl(2).
18
20 The following options are supported:
21 -F Force. Sets the preferred page size options(s)
23 for target process(es) even if controlled by
24 other process(es). Caution should be exercised
25 when using the -F flag. See proc(1).
26 -o option[,option] The options are:
29 heap=size This option specifies the preferred
31 page size for the heap of the tar‐
32 get process(es). heap is defined to
33 be the bss (uninitialized data) and
34 the brk area that immediately fol‐
35 lows the bss (see brk(2)). The pre‐
36 ferred heap page size is set for
37 the existing heap and for any addi‐
38 tional heap memory allocated in the
39 future. See NOTES.
40 stack=size This option specifies the preferred
43 page size for the stack of the tar‐
44 get process(es). The preferred
45 stack page size is set for the
46 existing stack and newly allocated
47 parts of the stack as it expands.
48 anon=size This option specifies the preferred
51 page size for all existing MAP_PRI‐
52 VATE anonymous segments of the tar‐
53 get process(es), other than heap
54 and stack, which are large enough
55 to fit at least one aligned page of
56 the specified size. For the seg‐
57 ments that are large enough, the
58 preferred page size is set starting
59 at the first size-aligned address
60 in the segment. The anon preferred
61 pagesize is not applied to MAP_PRI‐
62 VATE anonymous segments created in
63 the future. See MAP_ANON in
64 mmap(2).
65 Anonymous memory refers to MAP_PRI‐
67 VATE pages that are not directly
68 associated with a file in some
69 filesystem. The ppgsz command uses
70 memcntl(2) to set the preferred
71 page size for anonymous segments.
72 See MC_HAT_ADVISE in memcntl(2).
73 At least one of the above options must be speci‐
75 fied.
76 size must be a supported page size (see page‐
78 size(1)) or 0, in which case the system will
79 select an appropriate page size. See memcntl(2).
80 size defaults to bytes and can be specified in
82 octal (0), decimal, or hexadecimal (0x). The
83 numeric value can be qualified with K, M, G, or T
84 to specify Kilobytes, Megabytes, Gigabytes, or
85 Terabytes, respectively. 4194304, 0x400000,
86 4096K, 0x1000K, and 4M are different ways to
87 specify 4 Megabytes.
88 -p pid Sets the preferred page size option(s) for the
91 target process(es) in the process-id (pid) list
92 following the -p option. The pid list can also
93 consist of names in the /proc directory. Only the
94 process owner or the super-user is permitted to
95 set page size.
96 cmd is interpreted if -p is not specified. ppgsz
98 launches cmd and applies page size option(s) to
99 the new process.
100 The heap and stack preferred page sizes are
102 inherited. Child process(es) created (see
103 fork(2)) from the launched process or the target
104 process(es) in the pid list after ppgsz completes
105 will inherit the preferred heap and stack page
106 sizes. The preferred page sizes of all segments
107 are set back to the default system page size on
108 exec(2) (see getpagesize(3C)). The preferred page
109 size for all other anonymous segments is not
110 inherited by children of the launched or target
111 process(es).
112
115 Example 1 Setting the preferred heap and stack page size
116 The following example sets the preferred heap page size to 4M and the
119 preferred stack page size to 512K for all ora—owned processes running
120 commands that begin with ora:
121 example% ppgsz -o heap=4M,stack=512K -p `pgrep -u ora '^ora'`
124 Example 2 Setting the preferred anonymous page size
128 The following example sets the preferred page size of existing qualify‐
131 ing anonymous segments to 512k for process ID 953:
132 example% ppgsz -o anon=512k -p 953
135
139 If cmd is specified and successfully invoked (see exec(2)), the exit
140 status of ppgsz will be the exit status of cmd. Otherwise, ppgsz will
141 exit with one of the following values:
142 0 Successfully set preferred page size(s) for processes in the pid
144 list.
145 125 An error occurred in ppgsz. Errors include: invalid argument,
148 invalid page size(s) specified, and failure to set preferred
149 page size(s) for one or more processes in the pid list or cmd.
150 126 cmd was found but could not be invoked.
153 127 cmd could not be found.
156
159 /proc/* Process files.
160 /usr/lib/ld/map.bssalign A template link-editor mapfile for align‐
163 ing bss (see NOTES).
164
167 See attributes(5) for descriptions of the following attributes:
168 ┌─────────────────────────────┬─────────────────────────────┐
173 │ ATTRIBUTE TYPE │ ATTRIBUTE VALUE │
174 ├─────────────────────────────┼─────────────────────────────┤
175 │Availability │SUNWesu (32-bit) │
176 │ │SUNWesxu (64-bit) │
177 ├─────────────────────────────┼─────────────────────────────┤
178 │Interface Stability │Evolving │
179 └─────────────────────────────┴─────────────────────────────┘
180
182 ld(1), mpss.so.1(1), pagesize(1), pgrep(1), pmap(1), proc(1), brk(2),
183 exec(2), fork(2), memcntl(2), mmap(2), sbrk(2), getpagesize(3C),
184 proc(4), attributes(5)
185 Linker and Libraries Guide
188
190 Due to resource constraints, the setting of the preferred page size
191 does not necessarily guarantee that the target process(es) will get the
192 preferred page size. Use pmap(1) to view the actual heap and stack page
193 sizes of the target process(es) (see pmap -s option).
194 Large pages are required to be mapped at addresses that are multiples
197 of the size of the large page. Given that the heap is typically not
198 large page aligned, the starting portions of the heap (below the first
199 large page aligned address) are mapped with the system memory page
200 size. See getpagesize(3C).
201 To provide a heap that will be mapped with a large page size, an appli‐
204 cation can be built using a link-editor (ld(1)) mapfile containing the
205 bss segment declaration directive. Refer to the section Mapfile Option
206 in the Linker and Libraries Guide for more details of this directive
207 and the template mapfile provided in /usr/lib/ld/map.bssalign. Users
208 are cautioned that an alignment specification may be machine-specific
209 and may lose its benefit on different hardware platforms. A more flexi‐
210 ble means of requesting the most optimal underlying page size may
211 evolve in future releases.
212 mpss.so.1(1), a preloadable shared object, can also be used to set the
215 preferred stack and/or heap page sizes.
216SunOS 5.11 23 Jan 2003 ppgsz(1)