kernel-ark/arch/um/os-Linux/process.c
Jeff Dike 02dea0875b [PATCH] UML: Hotplug memory, take 2
Changes since first version
	added check for MADV_REMOVE support on the host
	fixed error return botch
	shrunk sprintf array by one character

This adds hotplug memory support to UML.  The mconsole syntax is
 	config mem=[+-]n[KMG]
In other words, add or subtract some number of kilobytes, megabytes, or
gigabytes.

Unplugged pages are allocated and then madvise(MADV_TRUNCATE), which is a
currently experimental madvise extension.  These pages are tracked so they
can be plugged back in later if the admin decides to give them back.  The
first page to be unplugged is used to keep track of about 4M of other
pages.  A list_head is the first thing on this page.  The rest is filled
with addresses of other unplugged pages.  This first page is not madvised,
obviously.

When this page is filled, the next page is used in a similar way and linked
onto a list with the first page.  Etc.  This whole process reverses when
pages are plugged back in.  When a tracking page no longer tracks any
unplugged pages, then it is next in line for plugging, which is done by
freeing pages back to the kernel.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-31 12:18:50 -08:00

279 lines
6.1 KiB
C

/*
* Copyright (C) 2002 Jeff Dike (jdike@addtoit.com)
* Licensed under the GPL
*/
#include <unistd.h>
#include <stdio.h>
#include <errno.h>
#include <signal.h>
#include <setjmp.h>
#include <linux/unistd.h>
#include <sys/mman.h>
#include <sys/wait.h>
#include <sys/mman.h>
#include "ptrace_user.h"
#include "os.h"
#include "user.h"
#include "user_util.h"
#include "process.h"
#include "irq_user.h"
#include "kern_util.h"
#include "longjmp.h"
#include "skas_ptrace.h"
#include "kern_constants.h"
#define ARBITRARY_ADDR -1
#define FAILURE_PID -1
#define STAT_PATH_LEN sizeof("/proc/#######/stat\0")
#define COMM_SCANF "%*[^)])"
unsigned long os_process_pc(int pid)
{
char proc_stat[STAT_PATH_LEN], buf[256];
unsigned long pc;
int fd, err;
sprintf(proc_stat, "/proc/%d/stat", pid);
fd = os_open_file(proc_stat, of_read(OPENFLAGS()), 0);
if(fd < 0){
printk("os_process_pc - couldn't open '%s', err = %d\n",
proc_stat, -fd);
return(ARBITRARY_ADDR);
}
err = os_read_file(fd, buf, sizeof(buf));
if(err < 0){
printk("os_process_pc - couldn't read '%s', err = %d\n",
proc_stat, -err);
os_close_file(fd);
return(ARBITRARY_ADDR);
}
os_close_file(fd);
pc = ARBITRARY_ADDR;
if(sscanf(buf, "%*d " COMM_SCANF " %*c %*d %*d %*d %*d %*d %*d %*d "
"%*d %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d "
"%*d %*d %*d %*d %*d %lu", &pc) != 1){
printk("os_process_pc - couldn't find pc in '%s'\n", buf);
}
return(pc);
}
int os_process_parent(int pid)
{
char stat[STAT_PATH_LEN];
char data[256];
int parent, n, fd;
if(pid == -1) return(-1);
snprintf(stat, sizeof(stat), "/proc/%d/stat", pid);
fd = os_open_file(stat, of_read(OPENFLAGS()), 0);
if(fd < 0){
printk("Couldn't open '%s', err = %d\n", stat, -fd);
return(FAILURE_PID);
}
n = os_read_file(fd, data, sizeof(data));
os_close_file(fd);
if(n < 0){
printk("Couldn't read '%s', err = %d\n", stat, -n);
return(FAILURE_PID);
}
parent = FAILURE_PID;
n = sscanf(data, "%*d " COMM_SCANF " %*c %d", &parent);
if(n != 1)
printk("Failed to scan '%s'\n", data);
return(parent);
}
void os_stop_process(int pid)
{
kill(pid, SIGSTOP);
}
void os_kill_process(int pid, int reap_child)
{
kill(pid, SIGKILL);
if(reap_child)
CATCH_EINTR(waitpid(pid, NULL, 0));
}
/* This is here uniquely to have access to the userspace errno, i.e. the one
* used by ptrace in case of error.
*/
long os_ptrace_ldt(long pid, long addr, long data)
{
int ret;
ret = ptrace(PTRACE_LDT, pid, addr, data);
if (ret < 0)
return -errno;
return ret;
}
/* Kill off a ptraced child by all means available. kill it normally first,
* then PTRACE_KILL it, then PTRACE_CONT it in case it's in a run state from
* which it can't exit directly.
*/
void os_kill_ptraced_process(int pid, int reap_child)
{
kill(pid, SIGKILL);
ptrace(PTRACE_KILL, pid);
ptrace(PTRACE_CONT, pid);
if(reap_child)
CATCH_EINTR(waitpid(pid, NULL, 0));
}
void os_usr1_process(int pid)
{
kill(pid, SIGUSR1);
}
/* Don't use the glibc version, which caches the result in TLS. It misses some
* syscalls, and also breaks with clone(), which does not unshare the TLS.
*/
inline _syscall0(pid_t, getpid)
int os_getpid(void)
{
return(getpid());
}
int os_getpgrp(void)
{
return getpgrp();
}
int os_map_memory(void *virt, int fd, unsigned long long off, unsigned long len,
int r, int w, int x)
{
void *loc;
int prot;
prot = (r ? PROT_READ : 0) | (w ? PROT_WRITE : 0) |
(x ? PROT_EXEC : 0);
loc = mmap64((void *) virt, len, prot, MAP_SHARED | MAP_FIXED,
fd, off);
if(loc == MAP_FAILED)
return(-errno);
return(0);
}
int os_protect_memory(void *addr, unsigned long len, int r, int w, int x)
{
int prot = ((r ? PROT_READ : 0) | (w ? PROT_WRITE : 0) |
(x ? PROT_EXEC : 0));
if(mprotect(addr, len, prot) < 0)
return(-errno);
return(0);
}
int os_unmap_memory(void *addr, int len)
{
int err;
err = munmap(addr, len);
if(err < 0)
return(-errno);
return(0);
}
#ifndef MADV_REMOVE
#define MADV_REMOVE 0x5 /* remove these pages & resources */
#endif
int os_drop_memory(void *addr, int length)
{
int err;
err = madvise(addr, length, MADV_REMOVE);
if(err < 0)
err = -errno;
return err;
}
int can_drop_memory(void)
{
void *addr;
int fd;
printk("Checking host MADV_REMOVE support...");
fd = create_mem_file(UM_KERN_PAGE_SIZE);
if(fd < 0){
printk("Creating test memory file failed, err = %d\n", -fd);
return 0;
}
addr = mmap64(NULL, UM_KERN_PAGE_SIZE, PROT_READ | PROT_WRITE,
MAP_PRIVATE, fd, 0);
if(addr == MAP_FAILED){
printk("Mapping test memory file failed, err = %d\n", -errno);
return 0;
}
if(madvise(addr, UM_KERN_PAGE_SIZE, MADV_REMOVE) != 0){
printk("MADV_REMOVE failed, err = %d\n", -errno);
return 0;
}
printk("OK\n");
return 1;
}
void init_new_thread_stack(void *sig_stack, void (*usr1_handler)(int))
{
int flags = 0, pages;
if(sig_stack != NULL){
pages = (1 << UML_CONFIG_KERNEL_STACK_ORDER);
set_sigstack(sig_stack, pages * page_size());
flags = SA_ONSTACK;
}
if(usr1_handler) set_handler(SIGUSR1, usr1_handler, flags, -1);
}
void init_new_thread_signals(int altstack)
{
int flags = altstack ? SA_ONSTACK : 0;
set_handler(SIGSEGV, (__sighandler_t) sig_handler, flags,
SIGUSR1, SIGIO, SIGWINCH, SIGALRM, SIGVTALRM, -1);
set_handler(SIGTRAP, (__sighandler_t) sig_handler, flags,
SIGUSR1, SIGIO, SIGWINCH, SIGALRM, SIGVTALRM, -1);
set_handler(SIGFPE, (__sighandler_t) sig_handler, flags,
SIGUSR1, SIGIO, SIGWINCH, SIGALRM, SIGVTALRM, -1);
set_handler(SIGILL, (__sighandler_t) sig_handler, flags,
SIGUSR1, SIGIO, SIGWINCH, SIGALRM, SIGVTALRM, -1);
set_handler(SIGBUS, (__sighandler_t) sig_handler, flags,
SIGUSR1, SIGIO, SIGWINCH, SIGALRM, SIGVTALRM, -1);
set_handler(SIGUSR2, (__sighandler_t) sig_handler,
flags, SIGUSR1, SIGIO, SIGWINCH, SIGALRM, SIGVTALRM, -1);
signal(SIGHUP, SIG_IGN);
init_irq_signals(altstack);
}
int run_kernel_thread(int (*fn)(void *), void *arg, void **jmp_ptr)
{
sigjmp_buf buf;
int n, enable;
*jmp_ptr = &buf;
n = UML_SIGSETJMP(&buf, enable);
if(n != 0)
return(n);
(*fn)(arg);
return(0);
}