Operating Systems
CMPSC 473
Virtual Memory Management (4)
November 18 2010 – Lecture 22
Instructor: Bhuvan Urgaonkar
Allocating Kernel Memory
• Treated differently from user memory
• Often allocated from a free-memory pool
– Kernel requests memory for structures of varying sizes
– Some kernel memory needs to be contiguous
Buddy System
• Allocates memory from fixed-size segment consisting of
physically-contiguous pages
• Memory allocated using power-of-2 allocator
– Satisfies requests in units sized as power of 2
– Request rounded up to next highest power of 2
– When smaller allocation needed than is available, current chunk split
into two buddies of next-lower power of 2
• Continue until appropriate sized chunk available
Buddy System Allocator
• Alternate strategy
Slab Allocator
• Slab is one or more physically contiguous pages
• Cache consists of one or more slabs
• Single cache for each unique kernel data structure
– Each cache filled with objects – instantiations of the data structure
• When cache created, filled with objects marked as free
• When structures stored, objects marked as used
• If slab is full of used objects, next object allocated from empty slab
– If no empty slabs, new slab allocated
• Benefits include no fragmentation, fast memory request satisfaction
Slab Allocation
How does malloc work?
• Your process has a heap spanning from virtual
address x to virtual address y
– All your “malloced” data lives here
• malloc() is part of a user-level library that keeps
some data structures, say a list, of all the free
chunks of space in the heap
• When you call malloc, it looks through the list
for a chunk that is big enough for your request,
returns a pointer to it, and records the fact that
it is not free anymore as well as how big it is
malloc (contd.)
• When you call free() with the same pointer, free() looks up
how big that chunk is and adds it back into the list of free
chunks()
• If you call malloc() and it can't find any large enough chunk in
the heap, it uses the brk() syscall to grow the heap, i.e. increase
address y and cause all the addresses between the old y and the
new y to be valid memory
– brk() must be a syscall; there is no way to do the same thing entirely
from userspace
NAME brk() system call
brk, sbrk - change the amount of space allocated for the calling process's data segment
SYNOPSIS
#include
int brk(void *endds);
void *sbrk(intptr_t incr);
DESCRIPTION
The brk() and sbrk() functions are used to change dynamically the amount of space
allocated for the calling process's data segment (see exec(2)). The change is made by
resetting the process's break value and allocating the appropriate amount of space.
The
break value is the address of the first location beyond the end of the data segment. The
amount of allocated space increases as the break value increases. Newly allocated space is
set to zero. If, how-ever, the same memory space is reallocated to the same process its
Memory-Mapped Files
• Memory-mapped file I/O allows file I/O to be treated as
routine memory access by mapping a disk block to a page in
memory
• A file is initially read using demand paging. A page-sized
portion of the file is read from the file system into a physical
page. Subsequent reads/writes to/from the file are treated as
ordinary memory accesses.
• Simplifies file access by treating file I/O through memory
rather than read() write() system calls
• Also allows several processes to map the same file allowing
the pages in memory to be shared
mmap vs. read/write:
advantages
• Reading from and writing to a memory-mapped file avoids
the extra copy that occurs when using the read() and write
system calls where data must be copied to and from a user-
space buffer
• Aside from any potential page faults, reading from and
writing to a memory-mapped file does not incur any system
call or context switch overhead
mmap advantages (contd.)
• When multiple processes map the same object into memory, data is
shared among all the processes
• Read-only and shared writable mappings are shared in their
entirety
• Private writable mappings have their not-yet COW pages shared
• Seeking around the mapping involves trivial pointer manipulations.
There is no need for the lseek() system call
mmap vs. read/write:
disadvantages
• Memory mappings are always at granularity of page =>
Wastage for small files
• Very large number of various-sized mappings can cause
fragmentation of address space making it hard to find large
free contiguous regions
• Overhead in creating and maintaining the mappings and
associated data structures inside kernel
Memory Mapped Files
Quiz
• Q1: Explain the difference between internal and external fragmentation. What is the
most internal fragmentation if address spaces > 4kB, 4kB <= page size <= 8kB
• Q2: Consider a paging system with the page table stored in memory
– If a memory ref. takes 200 nsec, how long does a paged memory ref. take?
– If we add TLBs, and TLB hit ratio is 90%, what is the effective mem. ref. time?
Assume TLB access takes 0 time
• Q3: How much memory is needed to store an entire page table for 32-bit address
space if
– Page size = 4kB, single-level page table
– Page size = 4kB, 3-level page table, p1=10, p2=10, d=12
Quiz
• Q4: Consider a demand-paging system with the following time-measured utilizations:
CPU util = 20%, Paging disk (swap device) = 97.7%, Other I/O devices = 5%. For
each of the following, say whether it will (or is likely to) improve CPU utilization.
Explain your answer.
– Install a faster CPU
– Install a bigger swap device
– Increase the degree of multi-programming
– Decrease the degree of multi-programming
– Install more main memory
– Install a faster hard disk (on which the swapping device resides)
– Increase the page size
Quiz
• Q5: Consider a system with the following #memory refs vs. LRU queue position curve
1000
0 1000
If the following two computers cost the same, which would you buy
1. CPU = 1GHz, RAM = 400pages
2. CPU = 2GHz, RAM = 500pages
Assume memory access time of 100 cycles and swap access time of 100,000 cycles. Ignore
memory needed to store page tables, ignore the TLB.