e041c68341
The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
506 lines
11 KiB
C
506 lines
11 KiB
C
/*
|
|
* Copyright (C) 2000, 2002 Jeff Dike (jdike@karaya.com)
|
|
* Licensed under the GPL
|
|
*/
|
|
|
|
#include "linux/config.h"
|
|
#include "linux/kernel.h"
|
|
#include "linux/sched.h"
|
|
#include "linux/notifier.h"
|
|
#include "linux/mm.h"
|
|
#include "linux/types.h"
|
|
#include "linux/tty.h"
|
|
#include "linux/init.h"
|
|
#include "linux/bootmem.h"
|
|
#include "linux/spinlock.h"
|
|
#include "linux/utsname.h"
|
|
#include "linux/sysrq.h"
|
|
#include "linux/seq_file.h"
|
|
#include "linux/delay.h"
|
|
#include "linux/module.h"
|
|
#include "asm/page.h"
|
|
#include "asm/pgtable.h"
|
|
#include "asm/ptrace.h"
|
|
#include "asm/elf.h"
|
|
#include "asm/user.h"
|
|
#include "asm/setup.h"
|
|
#include "ubd_user.h"
|
|
#include "asm/current.h"
|
|
#include "user_util.h"
|
|
#include "kern_util.h"
|
|
#include "kern.h"
|
|
#include "mem_user.h"
|
|
#include "mem.h"
|
|
#include "umid.h"
|
|
#include "initrd.h"
|
|
#include "init.h"
|
|
#include "os.h"
|
|
#include "choose-mode.h"
|
|
#include "mode_kern.h"
|
|
#include "mode.h"
|
|
#ifdef UML_CONFIG_MODE_SKAS
|
|
#include "skas.h"
|
|
#endif
|
|
|
|
#define DEFAULT_COMMAND_LINE "root=98:0"
|
|
|
|
/* Changed in linux_main and setup_arch, which run before SMP is started */
|
|
static char command_line[COMMAND_LINE_SIZE] = { 0 };
|
|
|
|
static void add_arg(char *arg)
|
|
{
|
|
if (strlen(command_line) + strlen(arg) + 1 > COMMAND_LINE_SIZE) {
|
|
printf("add_arg: Too many command line arguments!\n");
|
|
exit(1);
|
|
}
|
|
if(strlen(command_line) > 0)
|
|
strcat(command_line, " ");
|
|
strcat(command_line, arg);
|
|
}
|
|
|
|
struct cpuinfo_um boot_cpu_data = {
|
|
.loops_per_jiffy = 0,
|
|
.ipi_pipe = { -1, -1 }
|
|
};
|
|
|
|
unsigned long thread_saved_pc(struct task_struct *task)
|
|
{
|
|
return(os_process_pc(CHOOSE_MODE_PROC(thread_pid_tt, thread_pid_skas,
|
|
task)));
|
|
}
|
|
|
|
static int show_cpuinfo(struct seq_file *m, void *v)
|
|
{
|
|
int index = 0;
|
|
|
|
#ifdef CONFIG_SMP
|
|
index = (struct cpuinfo_um *) v - cpu_data;
|
|
if (!cpu_online(index))
|
|
return 0;
|
|
#endif
|
|
|
|
seq_printf(m, "processor\t: %d\n", index);
|
|
seq_printf(m, "vendor_id\t: User Mode Linux\n");
|
|
seq_printf(m, "model name\t: UML\n");
|
|
seq_printf(m, "mode\t\t: %s\n", CHOOSE_MODE("tt", "skas"));
|
|
seq_printf(m, "host\t\t: %s\n", host_info);
|
|
seq_printf(m, "bogomips\t: %lu.%02lu\n\n",
|
|
loops_per_jiffy/(500000/HZ),
|
|
(loops_per_jiffy/(5000/HZ)) % 100);
|
|
|
|
return(0);
|
|
}
|
|
|
|
static void *c_start(struct seq_file *m, loff_t *pos)
|
|
{
|
|
return *pos < NR_CPUS ? cpu_data + *pos : NULL;
|
|
}
|
|
|
|
static void *c_next(struct seq_file *m, void *v, loff_t *pos)
|
|
{
|
|
++*pos;
|
|
return c_start(m, pos);
|
|
}
|
|
|
|
static void c_stop(struct seq_file *m, void *v)
|
|
{
|
|
}
|
|
|
|
struct seq_operations cpuinfo_op = {
|
|
.start = c_start,
|
|
.next = c_next,
|
|
.stop = c_stop,
|
|
.show = show_cpuinfo,
|
|
};
|
|
|
|
/* Set in linux_main */
|
|
unsigned long host_task_size;
|
|
unsigned long task_size;
|
|
|
|
unsigned long uml_start;
|
|
|
|
/* Set in early boot */
|
|
unsigned long uml_physmem;
|
|
unsigned long uml_reserved;
|
|
unsigned long start_vm;
|
|
unsigned long end_vm;
|
|
int ncpus = 1;
|
|
|
|
#ifdef CONFIG_CMDLINE_ON_HOST
|
|
/* Pointer set in linux_main, the array itself is private to each thread,
|
|
* and changed at address space creation time so this poses no concurrency
|
|
* problems.
|
|
*/
|
|
static char *argv1_begin = NULL;
|
|
static char *argv1_end = NULL;
|
|
#endif
|
|
|
|
/* Set in early boot */
|
|
static int have_root __initdata = 0;
|
|
long long physmem_size = 32 * 1024 * 1024;
|
|
|
|
void set_cmdline(char *cmd)
|
|
{
|
|
#ifdef CONFIG_CMDLINE_ON_HOST
|
|
char *umid, *ptr;
|
|
|
|
if(CHOOSE_MODE(honeypot, 0)) return;
|
|
|
|
umid = get_umid();
|
|
if(*umid != '\0'){
|
|
snprintf(argv1_begin,
|
|
(argv1_end - argv1_begin) * sizeof(*ptr),
|
|
"(%s) ", umid);
|
|
ptr = &argv1_begin[strlen(argv1_begin)];
|
|
}
|
|
else ptr = argv1_begin;
|
|
|
|
snprintf(ptr, (argv1_end - ptr) * sizeof(*ptr), "[%s]", cmd);
|
|
memset(argv1_begin + strlen(argv1_begin), '\0',
|
|
argv1_end - argv1_begin - strlen(argv1_begin));
|
|
#endif
|
|
}
|
|
|
|
static char *usage_string =
|
|
"User Mode Linux v%s\n"
|
|
" available at http://user-mode-linux.sourceforge.net/\n\n";
|
|
|
|
static int __init uml_version_setup(char *line, int *add)
|
|
{
|
|
printf("%s\n", system_utsname.release);
|
|
exit(0);
|
|
|
|
return 0;
|
|
}
|
|
|
|
__uml_setup("--version", uml_version_setup,
|
|
"--version\n"
|
|
" Prints the version number of the kernel.\n\n"
|
|
);
|
|
|
|
static int __init uml_root_setup(char *line, int *add)
|
|
{
|
|
have_root = 1;
|
|
return 0;
|
|
}
|
|
|
|
__uml_setup("root=", uml_root_setup,
|
|
"root=<file containing the root fs>\n"
|
|
" This is actually used by the generic kernel in exactly the same\n"
|
|
" way as in any other kernel. If you configure a number of block\n"
|
|
" devices and want to boot off something other than ubd0, you \n"
|
|
" would use something like:\n"
|
|
" root=/dev/ubd5\n\n"
|
|
);
|
|
|
|
#ifndef CONFIG_MODE_TT
|
|
|
|
static int __init no_skas_debug_setup(char *line, int *add)
|
|
{
|
|
printf("'debug' is not necessary to gdb UML in skas mode - run \n");
|
|
printf("'gdb linux' and disable CONFIG_CMDLINE_ON_HOST if gdb \n");
|
|
printf("doesn't work as expected\n");
|
|
|
|
return 0;
|
|
}
|
|
|
|
__uml_setup("debug", no_skas_debug_setup,
|
|
"debug\n"
|
|
" this flag is not needed to run gdb on UML in skas mode\n\n"
|
|
);
|
|
|
|
#endif
|
|
|
|
#ifdef CONFIG_SMP
|
|
static int __init uml_ncpus_setup(char *line, int *add)
|
|
{
|
|
if (!sscanf(line, "%d", &ncpus)) {
|
|
printf("Couldn't parse [%s]\n", line);
|
|
return -1;
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
__uml_setup("ncpus=", uml_ncpus_setup,
|
|
"ncpus=<# of desired CPUs>\n"
|
|
" This tells an SMP kernel how many virtual processors to start.\n\n"
|
|
);
|
|
#endif
|
|
|
|
static int force_tt = 0;
|
|
|
|
#if defined(CONFIG_MODE_TT) && defined(CONFIG_MODE_SKAS)
|
|
#define DEFAULT_TT 0
|
|
|
|
static int __init mode_tt_setup(char *line, int *add)
|
|
{
|
|
force_tt = 1;
|
|
return(0);
|
|
}
|
|
|
|
#else
|
|
#ifdef CONFIG_MODE_SKAS
|
|
|
|
#define DEFAULT_TT 0
|
|
|
|
static int __init mode_tt_setup(char *line, int *add)
|
|
{
|
|
printf("CONFIG_MODE_TT disabled - 'mode=tt' ignored\n");
|
|
return(0);
|
|
}
|
|
|
|
#else
|
|
#ifdef CONFIG_MODE_TT
|
|
|
|
#define DEFAULT_TT 1
|
|
|
|
static int __init mode_tt_setup(char *line, int *add)
|
|
{
|
|
printf("CONFIG_MODE_SKAS disabled - 'mode=tt' redundant\n");
|
|
return(0);
|
|
}
|
|
|
|
#endif
|
|
#endif
|
|
#endif
|
|
|
|
__uml_setup("mode=tt", mode_tt_setup,
|
|
"mode=tt\n"
|
|
" When both CONFIG_MODE_TT and CONFIG_MODE_SKAS are enabled, this option\n"
|
|
" forces UML to run in tt (tracing thread) mode. It is not the default\n"
|
|
" because it's slower and less secure than skas mode.\n\n"
|
|
);
|
|
|
|
int mode_tt = DEFAULT_TT;
|
|
|
|
static int __init Usage(char *line, int *add)
|
|
{
|
|
const char **p;
|
|
|
|
printf(usage_string, system_utsname.release);
|
|
p = &__uml_help_start;
|
|
while (p < &__uml_help_end) {
|
|
printf("%s", *p);
|
|
p++;
|
|
}
|
|
exit(0);
|
|
|
|
return 0;
|
|
}
|
|
|
|
__uml_setup("--help", Usage,
|
|
"--help\n"
|
|
" Prints this message.\n\n"
|
|
);
|
|
|
|
static int __init uml_checksetup(char *line, int *add)
|
|
{
|
|
struct uml_param *p;
|
|
|
|
p = &__uml_setup_start;
|
|
while(p < &__uml_setup_end) {
|
|
int n;
|
|
|
|
n = strlen(p->str);
|
|
if(!strncmp(line, p->str, n)){
|
|
if (p->setup_func(line + n, add)) return 1;
|
|
}
|
|
p++;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
static void __init uml_postsetup(void)
|
|
{
|
|
initcall_t *p;
|
|
|
|
p = &__uml_postsetup_start;
|
|
while(p < &__uml_postsetup_end){
|
|
(*p)();
|
|
p++;
|
|
}
|
|
return;
|
|
}
|
|
|
|
/* Set during early boot */
|
|
unsigned long brk_start;
|
|
unsigned long end_iomem;
|
|
EXPORT_SYMBOL(end_iomem);
|
|
|
|
#define MIN_VMALLOC (32 * 1024 * 1024)
|
|
|
|
int linux_main(int argc, char **argv)
|
|
{
|
|
unsigned long avail, diff;
|
|
unsigned long virtmem_size, max_physmem;
|
|
unsigned int i, add;
|
|
char * mode;
|
|
|
|
for (i = 1; i < argc; i++){
|
|
if((i == 1) && (argv[i][0] == ' ')) continue;
|
|
add = 1;
|
|
uml_checksetup(argv[i], &add);
|
|
if (add)
|
|
add_arg(argv[i]);
|
|
}
|
|
if(have_root == 0)
|
|
add_arg(DEFAULT_COMMAND_LINE);
|
|
|
|
os_early_checks();
|
|
if (force_tt)
|
|
clear_can_do_skas();
|
|
mode_tt = force_tt ? 1 : !can_do_skas();
|
|
#ifndef CONFIG_MODE_TT
|
|
if (mode_tt) {
|
|
/*Since CONFIG_MODE_TT is #undef'ed, force_tt cannot be 1. So,
|
|
* can_do_skas() returned 0, and the message is correct. */
|
|
printf("Support for TT mode is disabled, and no SKAS support is present on the host.\n");
|
|
exit(1);
|
|
}
|
|
#endif
|
|
|
|
#ifndef CONFIG_MODE_SKAS
|
|
mode = "TT";
|
|
#else
|
|
/* Show to the user the result of selection */
|
|
if (mode_tt)
|
|
mode = "TT";
|
|
else if (proc_mm && ptrace_faultinfo)
|
|
mode = "SKAS3";
|
|
else
|
|
mode = "SKAS0";
|
|
#endif
|
|
|
|
printf("UML running in %s mode\n", mode);
|
|
|
|
uml_start = CHOOSE_MODE_PROC(set_task_sizes_tt, set_task_sizes_skas, 0,
|
|
&host_task_size, &task_size);
|
|
|
|
/*
|
|
* Setting up handlers to 'sig_info' struct
|
|
*/
|
|
os_fill_handlinfo(handlinfo_kern);
|
|
|
|
brk_start = (unsigned long) sbrk(0);
|
|
CHOOSE_MODE_PROC(before_mem_tt, before_mem_skas, brk_start);
|
|
/* Increase physical memory size for exec-shield users
|
|
so they actually get what they asked for. This should
|
|
add zero for non-exec shield users */
|
|
|
|
diff = UML_ROUND_UP(brk_start) - UML_ROUND_UP(&_end);
|
|
if(diff > 1024 * 1024){
|
|
printf("Adding %ld bytes to physical memory to account for "
|
|
"exec-shield gap\n", diff);
|
|
physmem_size += UML_ROUND_UP(brk_start) - UML_ROUND_UP(&_end);
|
|
}
|
|
|
|
uml_physmem = uml_start;
|
|
|
|
/* Reserve up to 4M after the current brk */
|
|
uml_reserved = ROUND_4M(brk_start) + (1 << 22);
|
|
|
|
setup_machinename(system_utsname.machine);
|
|
|
|
#ifdef CONFIG_CMDLINE_ON_HOST
|
|
argv1_begin = argv[1];
|
|
argv1_end = &argv[1][strlen(argv[1])];
|
|
#endif
|
|
|
|
highmem = 0;
|
|
iomem_size = (iomem_size + PAGE_SIZE - 1) & PAGE_MASK;
|
|
max_physmem = get_kmem_end() - uml_physmem - iomem_size - MIN_VMALLOC;
|
|
|
|
/* Zones have to begin on a 1 << MAX_ORDER page boundary,
|
|
* so this makes sure that's true for highmem
|
|
*/
|
|
max_physmem &= ~((1 << (PAGE_SHIFT + MAX_ORDER)) - 1);
|
|
if(physmem_size + iomem_size > max_physmem){
|
|
highmem = physmem_size + iomem_size - max_physmem;
|
|
physmem_size -= highmem;
|
|
#ifndef CONFIG_HIGHMEM
|
|
highmem = 0;
|
|
printf("CONFIG_HIGHMEM not enabled - physical memory shrunk "
|
|
"to %Lu bytes\n", physmem_size);
|
|
#endif
|
|
}
|
|
|
|
high_physmem = uml_physmem + physmem_size;
|
|
end_iomem = high_physmem + iomem_size;
|
|
high_memory = (void *) end_iomem;
|
|
|
|
start_vm = VMALLOC_START;
|
|
|
|
setup_physmem(uml_physmem, uml_reserved, physmem_size, highmem);
|
|
if(init_maps(physmem_size, iomem_size, highmem)){
|
|
printf("Failed to allocate mem_map for %Lu bytes of physical "
|
|
"memory and %Lu bytes of highmem\n", physmem_size,
|
|
highmem);
|
|
exit(1);
|
|
}
|
|
|
|
virtmem_size = physmem_size;
|
|
avail = get_kmem_end() - start_vm;
|
|
if(physmem_size > avail) virtmem_size = avail;
|
|
end_vm = start_vm + virtmem_size;
|
|
|
|
if(virtmem_size < physmem_size)
|
|
printf("Kernel virtual memory size shrunk to %lu bytes\n",
|
|
virtmem_size);
|
|
|
|
uml_postsetup();
|
|
|
|
task_protections((unsigned long) &init_thread_info);
|
|
os_flush_stdout();
|
|
|
|
return(CHOOSE_MODE(start_uml_tt(), start_uml_skas()));
|
|
}
|
|
|
|
extern int uml_exitcode;
|
|
|
|
static int panic_exit(struct notifier_block *self, unsigned long unused1,
|
|
void *unused2)
|
|
{
|
|
bust_spinlocks(1);
|
|
show_regs(&(current->thread.regs));
|
|
bust_spinlocks(0);
|
|
uml_exitcode = 1;
|
|
machine_halt();
|
|
return(0);
|
|
}
|
|
|
|
static struct notifier_block panic_exit_notifier = {
|
|
.notifier_call = panic_exit,
|
|
.next = NULL,
|
|
.priority = 0
|
|
};
|
|
|
|
void __init setup_arch(char **cmdline_p)
|
|
{
|
|
atomic_notifier_chain_register(&panic_notifier_list,
|
|
&panic_exit_notifier);
|
|
paging_init();
|
|
strlcpy(saved_command_line, command_line, COMMAND_LINE_SIZE);
|
|
*cmdline_p = command_line;
|
|
setup_hostinfo();
|
|
}
|
|
|
|
void __init check_bugs(void)
|
|
{
|
|
arch_check_bugs();
|
|
os_check_bugs();
|
|
}
|
|
|
|
void apply_alternatives(struct alt_instr *start, struct alt_instr *end)
|
|
{
|
|
}
|
|
|
|
void alternatives_smp_module_add(struct module *mod, char *name,
|
|
void *locks, void *locks_end,
|
|
void *text, void *text_end)
|
|
{
|
|
}
|
|
|
|
void alternatives_smp_module_del(struct module *mod)
|
|
{
|
|
}
|