haiku/src/system/kernel/team.cpp

3137 lines
74 KiB
C++
Raw Normal View History

/*
* Copyright 2002-2007, Axel Dörfler, axeld@pinc-software.de.
* Distributed under the terms of the MIT License.
*
* Copyright 2001-2002, Travis Geiselbrecht. All rights reserved.
* Distributed under the terms of the NewOS License.
*/
/*! Team functions */
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/wait.h>
#include <OS.h>
#include <AutoDeleter.h>
#include <elf.h>
#include <file_cache.h>
#include <int.h>
#include <kernel.h>
#include <kimage.h>
#include <kscheduler.h>
#include <ksignal.h>
#include <port.h>
#include <sem.h>
#include <syscall_process_info.h>
axeld + bonefish: * Implemented automatic syscall restarts: - A syscall can indicate that it has been interrupted and can be restarted by setting a respective bit in thread::flags. It can store parameters it wants to be preserved for the restart in thread::syscall_restart::parameters. Another thread::flags bit indicates whether it has been restarted. - handle_signals() clears the restart flag, if the handled signal has a handler function installed and SA_RESTART is not set. Another thread flag (THREAD_FLAGS_DONT_RESTART_SYSCALL) can prevent syscalls from being restarted, even if they could be (not used yet, but we might want to use it in resume_thread(), so that we stay behaviorally compatible with BeOS). - The architecture specific syscall handler restarts the syscall, if the restart flag is set. Implemented for x86 only. - Added some support functions in the private <syscall_restart.h> to simplify the syscall restart code in the syscalls. - Adjusted all syscalls that can potentially be restarted accordingly. - _user_ioctl() sets new thread flag THREAD_FLAGS_IOCTL_SYSCALL while calling the underlying FS's/driver's hook, so that syscall restarts can also be supported there. * thread_at_kernel_exit() invokes handle_signals() in a loop now, as long as the latter indicates that the thread shall be suspended, so that after waking up signals received in the meantime will be handled before the thread returns to userland. Adjusted handle_signals() accordingly -- when encountering a suspending signal we don't check for further signals. * Fixed sigsuspend(): Suspending the thread and rescheduling doesn't result in the correct behavior. Instead we employ a temporary condition variable and interruptably wait on it. The POSIX test suite test passes, now. * Made the switch_sem[_etc]() behavior on interruption consistent. Depending on when the signal arrived (before the call or when already waiting) the first semaphore would or wouldn't be released. Now we consistently release it. * Refactored _user_{read,write}[v]() syscalls. Use a common function for either pair. The iovec version doesn't fail anymore, if anything could be read/written at all. It also checks whether a complete vector could be read/written, so that we won't skip data, if the underlying FS/driver couldn't read/write more ATM. * Some refactoring in the x86 syscall handler: The int 99 and sysenter handlers use a common subroutine to avoid code duplication. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23983 a95241bf-73f2-0310-859d-f6bbb57e9c96
2008-02-17 18:48:30 +03:00
#include <syscall_restart.h>
#include <syscalls.h>
#include <team.h>
#include <tls.h>
#include <tracing.h>
#include <user_runtime.h>
#include <vfs.h>
#include <vm.h>
#include <vm_address_space.h>
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
#include <util/AutoLock.h>
#include <util/khash.h>
//#define TRACE_TEAM
#ifdef TRACE_TEAM
# define TRACE(x) dprintf x
#else
# define TRACE(x) ;
#endif
struct team_key {
team_id id;
};
struct team_arg {
uint32 arg_count;
char **args;
uint32 env_count;
char **env;
port_id error_port;
uint32 error_token;
};
struct fork_arg {
area_id user_stack_area;
addr_t user_stack_base;
size_t user_stack_size;
addr_t user_local_storage;
sigset_t sig_block_mask;
struct arch_fork_arg arch_info;
};
static hash_table *sTeamHash = NULL;
static hash_table *sGroupHash = NULL;
static struct team *sKernelTeam = NULL;
// some arbitrary chosen limits - should probably depend on the available
// memory (the limit is not yet enforced)
static int32 sMaxTeams = 2048;
static int32 sUsedTeams = 1;
spinlock team_spinlock = 0;
// #pragma mark - Tracing
#ifdef TEAM_TRACING
namespace TeamTracing {
class TeamForked : public AbstractTraceEntry {
public:
TeamForked(thread_id forkedThread)
:
fForkedThread(forkedThread)
{
Initialized();
}
virtual void AddDump(TraceOutput& out)
{
out.Print("team forked, new thread %ld", fForkedThread);
}
private:
thread_id fForkedThread;
};
class ExecTeam : public AbstractTraceEntry {
public:
ExecTeam(const char* path, int32 argCount, const char* const* args,
int32 envCount, const char* const* env)
:
fArgCount(argCount),
fArgs(NULL)
{
fPath = alloc_tracing_buffer_strcpy(path, B_PATH_NAME_LENGTH,
false);
// determine the buffer size we need for the args
size_t argBufferSize = 0;
for (int32 i = 0; i < argCount; i++)
argBufferSize += strlen(args[i]) + 1;
// allocate a buffer
fArgs = (char*)alloc_tracing_buffer(argBufferSize);
if (fArgs) {
char* buffer = fArgs;
for (int32 i = 0; i < argCount; i++) {
size_t argSize = strlen(args[i]) + 1;
memcpy(buffer, args[i], argSize);
buffer += argSize;
}
}
// ignore env for the time being
(void)envCount;
(void)env;
Initialized();
}
virtual void AddDump(TraceOutput& out)
{
out.Print("team exec, \"%p\", args:", fPath);
char* args = fArgs;
for (int32 i = 0; !out.IsFull() && i < fArgCount; i++) {
out.Print(" \"%s\"", args);
args += strlen(args) + 1;
}
}
private:
char* fPath;
int32 fArgCount;
char* fArgs;
};
static const char*
job_control_state_name(job_control_state state)
{
switch (state) {
case JOB_CONTROL_STATE_NONE:
return "none";
case JOB_CONTROL_STATE_STOPPED:
return "stopped";
case JOB_CONTROL_STATE_CONTINUED:
return "continued";
case JOB_CONTROL_STATE_DEAD:
return "dead";
default:
return "invalid";
}
}
class SetJobControlState : public AbstractTraceEntry {
public:
SetJobControlState(team_id team, job_control_state newState, int signal)
:
fTeam(team),
fNewState(newState),
fSignal(signal)
{
Initialized();
}
virtual void AddDump(TraceOutput& out)
{
out.Print("team set job control state, team %ld, "
"new state: %s, signal: %d",
fTeam, job_control_state_name(fNewState), fSignal);
}
private:
team_id fTeam;
job_control_state fNewState;
int fSignal;
};
class WaitForChild : public AbstractTraceEntry {
public:
WaitForChild(pid_t child, uint32 flags)
:
fChild(child),
fFlags(flags)
{
Initialized();
}
virtual void AddDump(TraceOutput& out)
{
out.Print("team wait for child, child: %ld, "
"flags: 0x%lx", fChild, fFlags);
}
private:
pid_t fChild;
uint32 fFlags;
};
class WaitForChildDone : public AbstractTraceEntry {
public:
WaitForChildDone(const job_control_entry& entry)
:
fState(entry.state),
fTeam(entry.thread),
fStatus(entry.status),
fReason(entry.reason),
fSignal(entry.signal)
{
Initialized();
}
WaitForChildDone(status_t error)
:
fTeam(error)
{
Initialized();
}
virtual void AddDump(TraceOutput& out)
{
if (fTeam >= 0) {
out.Print("team wait for child done, team: %ld, "
"state: %s, status: 0x%lx, reason: 0x%x, signal: %d\n",
fTeam, job_control_state_name(fState), fStatus, fReason,
fSignal);
} else {
out.Print("team wait for child failed, error: "
"0x%lx, ", fTeam);
}
}
private:
job_control_state fState;
team_id fTeam;
status_t fStatus;
uint16 fReason;
uint16 fSignal;
};
} // namespace TeamTracing
# define T(x) new(std::nothrow) TeamTracing::x;
#else
# define T(x) ;
#endif
// #pragma mark - Private functions
static void
_dump_team_info(struct team *team)
{
kprintf("TEAM: %p\n", team);
kprintf("id: %ld (%#lx)\n", team->id, team->id);
kprintf("name: '%s'\n", team->name);
kprintf("args: '%s'\n", team->args);
kprintf("next: %p\n", team->next);
kprintf("parent: %p", team->parent);
if (team->parent != NULL) {
kprintf(" (id = %ld)\n", team->parent->id);
} else
kprintf("\n");
kprintf("children: %p\n", team->children);
kprintf("num_threads: %d\n", team->num_threads);
kprintf("state: %d\n", team->state);
kprintf("flags: 0x%lx\n", team->flags);
kprintf("io_context: %p\n", team->io_context);
if (team->address_space)
kprintf("address_space: %p\n", team->address_space);
kprintf("main_thread: %p\n", team->main_thread);
kprintf("thread_list: %p\n", team->thread_list);
kprintf("group_id: %ld\n", team->group_id);
kprintf("session_id: %ld\n", team->session_id);
}
static int
dump_team_info(int argc, char **argv)
{
struct hash_iterator iterator;
struct team *team;
team_id id = -1;
bool found = false;
if (argc < 2) {
struct thread* thread = thread_get_current_thread();
if (thread != NULL && thread->team != NULL)
_dump_team_info(thread->team);
else
kprintf("No current team!\n");
return 0;
}
id = strtoul(argv[1], NULL, 0);
if (IS_KERNEL_ADDRESS(id)) {
// semi-hack
_dump_team_info((struct team *)id);
return 0;
}
// walk through the thread list, trying to match name or id
hash_open(sTeamHash, &iterator);
while ((team = (struct team*)hash_next(sTeamHash, &iterator)) != NULL) {
if ((team->name && strcmp(argv[1], team->name) == 0) || team->id == id) {
_dump_team_info(team);
found = true;
break;
}
}
hash_close(sTeamHash, &iterator, false);
if (!found)
kprintf("team \"%s\" (%ld) doesn't exist!\n", argv[1], id);
return 0;
}
static int
dump_teams(int argc, char **argv)
{
struct hash_iterator iterator;
struct team *team;
kprintf("team id parent name\n");
hash_open(sTeamHash, &iterator);
while ((team = (struct team*)hash_next(sTeamHash, &iterator)) != NULL) {
kprintf("%p%7ld %p %s\n", team, team->id, team->parent, team->name);
}
hash_close(sTeamHash, &iterator, false);
return 0;
}
/*! Frees an array of strings in kernel space.
\param strings strings array
\param count number of strings in array
*/
static void
free_strings_array(char **strings, int32 count)
{
int32 i;
if (strings == NULL)
return;
for (i = 0; i < count; i++)
free(strings[i]);
free(strings);
}
/*! Copy an array of strings in kernel space
\param strings strings array to be copied
\param count number of strings in array
\param kstrings pointer to the kernel copy
\return \c B_OK on success, or an appropriate error code on
failure.
*/
static status_t
kernel_copy_strings_array(char * const *in, int32 count, char ***_strings)
{
status_t status;
char **strings;
int32 i = 0;
strings = (char **)malloc((count + 1) * sizeof(char *));
if (strings == NULL)
return B_NO_MEMORY;
for (; i < count; i++) {
strings[i] = strdup(in[i]);
if (strings[i] == NULL) {
status = B_NO_MEMORY;
goto error;
}
}
strings[count] = NULL;
*_strings = strings;
return B_OK;
error:
free_strings_array(strings, i);
return status;
}
/*! Copy an array of strings from user space to kernel space
\param strings userspace strings array
\param count number of strings in array
\param kstrings pointer to the kernel copy
\return \c B_OK on success, or an appropriate error code on
failure.
*/
static status_t
user_copy_strings_array(char * const *userStrings, int32 count, char ***_strings)
{
char *buffer;
char **strings;
status_t err;
int32 i = 0;
if (!IS_USER_ADDRESS(userStrings))
return B_BAD_ADDRESS;
// buffer for safely accessing the user string
// TODO: maybe have a user_strdup() instead?
buffer = (char *)malloc(4 * B_PAGE_SIZE);
if (buffer == NULL)
return B_NO_MEMORY;
strings = (char **)malloc((count + 1) * sizeof(char *));
if (strings == NULL) {
err = B_NO_MEMORY;
goto error;
}
if ((err = user_memcpy(strings, userStrings, count * sizeof(char *))) < B_OK)
goto error;
// scan all strings and copy to kernel space
for (; i < count; i++) {
err = user_strlcpy(buffer, strings[i], 4 * B_PAGE_SIZE);
if (err < B_OK)
goto error;
strings[i] = strdup(buffer);
if (strings[i] == NULL) {
err = B_NO_MEMORY;
goto error;
}
}
strings[count] = NULL;
*_strings = strings;
free(buffer);
return B_OK;
error:
free_strings_array(strings, i);
free(buffer);
TRACE(("user_copy_strings_array failed %ld\n", err));
return err;
}
static status_t
copy_strings_array(char * const *strings, int32 count, char ***_strings,
bool kernel)
{
if (kernel)
return kernel_copy_strings_array(strings, count, _strings);
return user_copy_strings_array(strings, count, _strings);
}
static int
team_struct_compare(void *_p, const void *_key)
{
struct team *p = (struct team*)_p;
const struct team_key *key = (const struct team_key*)_key;
if (p->id == key->id)
return 0;
return 1;
}
static uint32
team_struct_hash(void *_p, const void *_key, uint32 range)
{
struct team *p = (struct team*)_p;
const struct team_key *key = (const struct team_key*)_key;
if (p != NULL)
return p->id % range;
return (uint32)key->id % range;
}
static int
process_group_compare(void *_group, const void *_key)
{
struct process_group *group = (struct process_group*)_group;
const struct team_key *key = (const struct team_key*)_key;
if (group->id == key->id)
return 0;
return 1;
}
static uint32
process_group_hash(void *_group, const void *_key, uint32 range)
{
struct process_group *group = (struct process_group*)_group;
const struct team_key *key = (const struct team_key*)_key;
if (group != NULL)
return group->id % range;
return (uint32)key->id % range;
}
static void
insert_team_into_parent(struct team *parent, struct team *team)
{
ASSERT(parent != NULL);
team->siblings_next = parent->children;
parent->children = team;
team->parent = parent;
}
/*! Note: must have team lock held */
static void
remove_team_from_parent(struct team *parent, struct team *team)
{
struct team *child, *last = NULL;
for (child = parent->children; child != NULL; child = child->siblings_next) {
if (child == team) {
if (last == NULL)
parent->children = child->siblings_next;
else
last->siblings_next = child->siblings_next;
team->parent = NULL;
break;
}
last = child;
}
}
/*! Reparent each of our children
Note: must have team lock held
*/
static void
reparent_children(struct team *team)
{
struct team *child;
while ((child = team->children) != NULL) {
// remove the child from the current proc and add to the parent
remove_team_from_parent(team, child);
insert_team_into_parent(sKernelTeam, child);
}
// move job control entries too
sKernelTeam->stopped_children->entries.MoveFrom(
&team->stopped_children->entries);
sKernelTeam->continued_children->entries.MoveFrom(
&team->continued_children->entries);
// Note, we don't move the dead children entries. Those will be deleted
// when the team structure is deleted.
}
static bool
is_session_leader(struct team *team)
{
return team->session_id == team->id;
}
static bool
is_process_group_leader(struct team *team)
{
return team->group_id == team->id;
}
/*! You must hold the team lock when calling this function. */
static void
insert_group_into_session(struct process_session *session, struct process_group *group)
{
if (group == NULL)
return;
group->session = session;
hash_insert(sGroupHash, group);
session->group_count++;
}
/*! You must hold the team lock when calling this function. */
static void
insert_team_into_group(struct process_group *group, struct team *team)
{
team->group = group;
team->group_id = group->id;
team->session_id = group->session->id;
team->group_next = group->teams;
group->teams = team;
}
/*! Removes a group from a session, and puts the session object
back into the session cache, if it's not used anymore.
You must hold the team lock when calling this function.
*/
static void
remove_group_from_session(struct process_group *group)
{
struct process_session *session = group->session;
// the group must be in any session to let this function have any effect
if (session == NULL)
return;
hash_remove(sGroupHash, group);
// we cannot free the resource here, so we're keeping the group link
// around - this way it'll be freed by free_process_group()
if (--session->group_count > 0)
group->session = NULL;
}
/*! Removes the team from the group. If that group becomes therefore
unused, it will set \a _freeGroup to point to the group - otherwise
it will be \c NULL.
It cannot be freed here because this function has to be called
with having the team lock held.
\param team the team that'll be removed from it's group
\param _freeGroup points to the group to be freed or NULL
*/
static void
remove_team_from_group(struct team *team, struct process_group **_freeGroup)
{
struct process_group *group = team->group;
struct team *current, *last = NULL;
*_freeGroup = NULL;
// the team must be in any team to let this function have any effect
if (group == NULL)
return;
for (current = group->teams; current != NULL; current = current->group_next) {
if (current == team) {
if (last == NULL)
group->teams = current->group_next;
else
last->group_next = current->group_next;
team->group = NULL;
break;
}
last = current;
}
team->group = NULL;
team->group_next = NULL;
if (group->teams != NULL)
return;
// we can remove this group as it is no longer used
remove_group_from_session(group);
*_freeGroup = group;
}
static struct process_group *
create_process_group(pid_t id)
{
struct process_group *group = (struct process_group *)malloc(sizeof(struct process_group));
if (group == NULL)
return NULL;
group->id = id;
group->session = NULL;
group->teams = NULL;
group->orphaned = true;
return group;
}
static struct process_session *
create_process_session(pid_t id)
{
struct process_session *session
= (struct process_session *)malloc(sizeof(struct process_session));
if (session == NULL)
return NULL;
session->id = id;
session->group_count = 0;
session->controlling_tty = -1;
session->foreground_group = -1;
return session;
}
static void
set_team_name(struct team* team, const char* name)
{
if (const char* lastSlash = strrchr(name, '/'))
name = lastSlash + 1;
strlcpy(team->name, name, B_OS_NAME_LENGTH);
}
static struct team *
create_team_struct(const char *name, bool kernel)
{
struct team *team = (struct team *)malloc(sizeof(struct team));
if (team == NULL)
return NULL;
MemoryDeleter teamDeleter(team);
team->next = team->siblings_next = team->children = team->parent = NULL;
team->id = allocate_thread_id();
set_team_name(team, name);
team->args[0] = '\0';
team->num_threads = 0;
team->io_context = NULL;
team->address_space = NULL;
team->thread_list = NULL;
team->main_thread = NULL;
team->loading_info = NULL;
team->state = TEAM_STATE_BIRTH;
team->flags = 0;
team->death_sem = -1;
team->dead_threads_kernel_time = 0;
team->dead_threads_user_time = 0;
// dead threads
list_init(&team->dead_threads);
team->dead_threads_count = 0;
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
// dead children
team->dead_children = new(nothrow) team_dead_children;
if (team->dead_children == NULL)
return NULL;
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
ObjectDeleter<team_dead_children> deadChildrenDeleter(team->dead_children);
team->dead_children->count = 0;
team->dead_children->kernel_time = 0;
team->dead_children->user_time = 0;
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
// stopped children
team->stopped_children = new(nothrow) team_job_control_children;
if (team->stopped_children == NULL)
return NULL;
ObjectDeleter<team_job_control_children> stoppedChildrenDeleter(
team->stopped_children);
// continued children
team->continued_children = new(nothrow) team_job_control_children;
if (team->continued_children == NULL)
return NULL;
ObjectDeleter<team_job_control_children> continuedChildrenDeleter(
team->continued_children);
// job control entry
team->job_control_entry = new(nothrow) job_control_entry;
if (team->job_control_entry == NULL)
return NULL;
ObjectDeleter<job_control_entry> jobControlEntryDeleter(
team->job_control_entry);
team->job_control_entry->state = JOB_CONTROL_STATE_NONE;
team->job_control_entry->thread = team->id;
team->job_control_entry->team = team;
list_init(&team->image_list);
list_init(&team->watcher_list);
clear_team_debug_info(&team->debug_info, true);
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
if (arch_team_init_team_struct(team, kernel) < 0)
return NULL;
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
// publish dead/stopped/continued children condition vars
team->dead_children->condition_variable.Publish(team->dead_children,
"dead children");
team->stopped_children->condition_variable.Publish(team->stopped_children,
"stopped children");
team->continued_children->condition_variable.Publish(
team->continued_children, "continued children");
// keep all allocated structures
jobControlEntryDeleter.Detach();
continuedChildrenDeleter.Detach();
stoppedChildrenDeleter.Detach();
deadChildrenDeleter.Detach();
teamDeleter.Detach();
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
return team;
}
static void
delete_team_struct(struct team *team)
{
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
team->stopped_children->condition_variable.Unpublish();
team->continued_children->condition_variable.Unpublish();
team->dead_children->condition_variable.Unpublish();
while (death_entry* threadDeathEntry = (death_entry*)list_remove_head_item(
&team->dead_threads)) {
free(threadDeathEntry);
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
while (job_control_entry* entry = team->dead_children->entries.RemoveHead())
delete entry;
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
delete team->job_control_entry;
// usually already NULL and transferred to the parent
delete team->continued_children;
delete team->stopped_children;
delete team->dead_children;
free(team);
}
static uint32
get_arguments_data_size(char **args, int32 argc)
{
uint32 size = 0;
int32 count;
for (count = 0; count < argc; count++)
size += strlen(args[count]) + 1;
return size + (argc + 1) * sizeof(char *) + sizeof(struct user_space_program_args);
}
static void
free_team_arg(struct team_arg *teamArg)
{
free_strings_array(teamArg->args, teamArg->arg_count);
free_strings_array(teamArg->env, teamArg->env_count);
free(teamArg);
}
static status_t
create_team_arg(struct team_arg **_teamArg, int32 argCount, char * const *args,
int32 envCount, char * const *env, port_id port, uint32 token, bool kernel)
{
status_t status;
char **argsCopy;
char **envCopy;
struct team_arg *teamArg = (struct team_arg *)malloc(sizeof(struct team_arg));
if (teamArg == NULL)
return B_NO_MEMORY;
// copy the args over
status = copy_strings_array(args, argCount, &argsCopy, kernel);
if (status != B_OK)
return status;
status = copy_strings_array(env, envCount, &envCopy, kernel);
if (status != B_OK) {
free_strings_array(argsCopy, argCount);
return status;
}
teamArg->arg_count = argCount;
teamArg->args = argsCopy;
teamArg->env_count = envCount;
teamArg->env = envCopy;
teamArg->error_port = port;
teamArg->error_token = token;
*_teamArg = teamArg;
return B_OK;
}
static int32
team_create_thread_start(void *args)
{
status_t err;
struct thread *t;
struct team *team;
struct team_arg *teamArgs = (struct team_arg*)args;
const char *path;
addr_t entry;
char ustack_name[128];
uint32 sizeLeft;
char **userArgs;
char **userEnv;
char *userDest;
struct user_space_program_args *programArgs;
uint32 argCount, envCount, i;
t = thread_get_current_thread();
team = t->team;
cache_node_launched(teamArgs->arg_count, teamArgs->args);
TRACE(("team_create_thread_start: entry thread %ld\n", t->id));
// create an initial primary stack area
// Main stack area layout is currently as follows (starting from 0):
//
// size | usage
// -----------------------------+--------------------------------
// USER_MAIN_THREAD_STACK_SIZE | actual stack
// TLS_SIZE | TLS data
// ENV_SIZE | environment variables
// arguments size | arguments passed to the team
// ToDo: ENV_SIZE is a) limited, and b) not used after libroot copied it to the heap
// ToDo: we could reserve the whole USER_STACK_REGION upfront...
sizeLeft = PAGE_ALIGN(USER_MAIN_THREAD_STACK_SIZE + TLS_SIZE + ENV_SIZE +
get_arguments_data_size(teamArgs->args, teamArgs->arg_count));
t->user_stack_base = USER_STACK_REGION + USER_STACK_REGION_SIZE - sizeLeft;
t->user_stack_size = USER_MAIN_THREAD_STACK_SIZE;
// the exact location at the end of the user stack area
sprintf(ustack_name, "%s_main_stack", team->name);
t->user_stack_area = create_area_etc(team, ustack_name, (void **)&t->user_stack_base,
B_EXACT_ADDRESS, sizeLeft, B_NO_LOCK, B_READ_AREA | B_WRITE_AREA | B_STACK_AREA);
if (t->user_stack_area < 0) {
dprintf("team_create_thread_start: could not create default user stack region\n");
free_team_arg(teamArgs);
return t->user_stack_area;
}
// now that the TLS area is allocated, initialize TLS
arch_thread_init_tls(t);
argCount = teamArgs->arg_count;
envCount = teamArgs->env_count;
programArgs = (struct user_space_program_args *)(t->user_stack_base
+ t->user_stack_size + TLS_SIZE + ENV_SIZE);
userArgs = (char **)(programArgs + 1);
userDest = (char *)(userArgs + argCount + 1);
TRACE(("addr: stack base = 0x%lx, userArgs = %p, userDest = %p, sizeLeft = %lu\n",
t->user_stack_base, userArgs, userDest, sizeLeft));
sizeLeft = t->user_stack_base + sizeLeft - (addr_t)userDest;
for (i = 0; i < argCount; i++) {
ssize_t length = user_strlcpy(userDest, teamArgs->args[i], sizeLeft);
if (length < B_OK) {
argCount = 0;
break;
}
userArgs[i] = userDest;
userDest += ++length;
sizeLeft -= length;
}
userArgs[argCount] = NULL;
userEnv = (char **)(t->user_stack_base + t->user_stack_size + TLS_SIZE);
sizeLeft = ENV_SIZE;
userDest = (char *)userEnv + ENV_SIZE - 1;
// the environment variables are copied from back to front
TRACE(("team_create_thread_start: envc: %ld, env: %p\n",
teamArgs->env_count, (void *)teamArgs->env));
for (i = 0; i < envCount; i++) {
ssize_t length = strlen(teamArgs->env[i]) + 1;
userDest -= length;
if (userDest < (char *)&userEnv[envCount]) {
envCount = i;
break;
}
userEnv[i] = userDest;
if (user_memcpy(userDest, teamArgs->env[i], length) < B_OK) {
envCount = 0;
break;
}
sizeLeft -= length;
}
userEnv[envCount] = NULL;
path = teamArgs->args[0];
if (user_memcpy(programArgs->program_path, path,
sizeof(programArgs->program_path)) < B_OK
|| user_memcpy(&programArgs->arg_count, &argCount, sizeof(int32)) < B_OK
|| user_memcpy(&programArgs->args, &userArgs, sizeof(char **)) < B_OK
|| user_memcpy(&programArgs->env_count, &envCount, sizeof(int32)) < B_OK
|| user_memcpy(&programArgs->env, &userEnv, sizeof(char **)) < B_OK
|| user_memcpy(&programArgs->error_port, &teamArgs->error_port,
sizeof(port_id)) < B_OK
|| user_memcpy(&programArgs->error_token, &teamArgs->error_token,
sizeof(uint32)) < B_OK) {
// the team deletion process will clean this mess
return B_BAD_ADDRESS;
}
TRACE(("team_create_thread_start: loading elf binary '%s'\n", path));
// add args to info member
team->args[0] = 0;
strlcpy(team->args, path, sizeof(team->args));
for (i = 1; i < argCount; i++) {
strlcat(team->args, " ", sizeof(team->args));
strlcat(team->args, teamArgs->args[i], sizeof(team->args));
}
free_team_arg(teamArgs);
// the arguments are already on the user stack, we no longer need them in this form
// ToDo: don't use fixed paths!
err = elf_load_user_image("/boot/beos/system/runtime_loader", team, 0, &entry);
if (err < B_OK) {
// Luckily, we don't have to clean up the mess we created - that's
// done for us by the normal team deletion process
TRACE(("team_create_thread_start: error when elf_load_user_image() %s\n", strerror(err)));
return err;
}
TRACE(("team_create_thread_start: loaded elf. entry = %#lx\n", entry));
team->state = TEAM_STATE_NORMAL;
// jump to the entry point in user space
return arch_thread_enter_userspace(t, entry, programArgs, NULL);
// only returns in case of error
}
/*! The BeOS kernel exports a function with this name, but most probably with
different parameters; we should not make it public.
*/
static thread_id
load_image_etc(int32 argCount, char * const *args, int32 envCount,
char * const *env, int32 priority, uint32 flags,
port_id errorPort, uint32 errorToken, bool kernel)
{
struct process_group *group;
struct team *team, *parent;
const char *threadName;
thread_id thread;
status_t status;
cpu_status state;
struct team_arg *teamArgs;
struct team_loading_info loadingInfo;
if (args == NULL || argCount == 0)
return B_BAD_VALUE;
TRACE(("load_image_etc: name '%s', args = %p, argCount = %ld\n",
args[0], args, argCount));
team = create_team_struct(args[0], false);
if (team == NULL)
return B_NO_MEMORY;
parent = thread_get_current_thread()->team;
if (flags & B_WAIT_TILL_LOADED) {
loadingInfo.thread = thread_get_current_thread();
loadingInfo.result = B_ERROR;
loadingInfo.done = false;
team->loading_info = &loadingInfo;
}
state = disable_interrupts();
GRAB_TEAM_LOCK();
hash_insert(sTeamHash, team);
insert_team_into_parent(parent, team);
insert_team_into_group(parent->group, team);
sUsedTeams++;
RELEASE_TEAM_LOCK();
restore_interrupts(state);
status = create_team_arg(&teamArgs, argCount, args, envCount, env,
errorPort, errorToken, kernel);
if (status != B_OK)
goto err1;
// create a new io_context for this team
team->io_context = vfs_new_io_context(parent->io_context);
if (!team->io_context) {
status = B_NO_MEMORY;
goto err2;
}
// remove any fds that have the CLOEXEC flag set (emulating BeOS behaviour)
vfs_exec_io_context(team->io_context);
// create an address space for this team
status = vm_create_address_space(team->id, USER_BASE, USER_SIZE, false,
&team->address_space);
if (status < B_OK)
goto err3;
// cut the path from the main thread name
threadName = strrchr(args[0], '/');
if (threadName != NULL)
threadName++;
else
threadName = args[0];
// Create a kernel thread, but under the context of the new team
// The new thread will take over ownership of teamArgs
thread = spawn_kernel_thread_etc(team_create_thread_start, threadName,
B_NORMAL_PRIORITY, teamArgs, team->id, team->id);
if (thread < 0) {
status = thread;
goto err4;
}
// wait for the loader of the new team to finish its work
if (flags & B_WAIT_TILL_LOADED) {
struct thread *mainThread;
state = disable_interrupts();
GRAB_THREAD_LOCK();
mainThread = thread_get_thread_struct_locked(thread);
if (mainThread) {
// resume the team's main thread
if (mainThread->state == B_THREAD_SUSPENDED)
scheduler_enqueue_in_run_queue(mainThread);
// Now suspend ourselves until loading is finished.
// We will be woken either by the thread, when it finished or
// aborted loading, or when the team is going to die (e.g. is
// killed). In either case the one setting `loadingInfo.done' is
// responsible for removing the info from the team structure.
while (!loadingInfo.done) {
thread_get_current_thread()->next_state = B_THREAD_SUSPENDED;
scheduler_reschedule();
}
} else {
// Impressive! Someone managed to kill the thread in this short
// time.
}
RELEASE_THREAD_LOCK();
restore_interrupts(state);
if (loadingInfo.result < B_OK)
return loadingInfo.result;
}
// notify the debugger
user_debug_team_created(team->id);
return thread;
err4:
vm_put_address_space(team->address_space);
err3:
vfs_free_io_context(team->io_context);
err2:
free_team_arg(teamArgs);
err1:
// remove the team structure from the team hash table and delete the team structure
state = disable_interrupts();
GRAB_TEAM_LOCK();
remove_team_from_group(team, &group);
remove_team_from_parent(parent, team);
hash_remove(sTeamHash, team);
RELEASE_TEAM_LOCK();
restore_interrupts(state);
team_delete_process_group(group);
delete_team_struct(team);
return status;
}
/*! Almost shuts down the current team and loads a new image into it.
If successful, this function does not return and will takeover ownership of
the arguments provided.
This function may only be called from user space.
*/
static status_t
exec_team(const char *path, int32 argCount, char * const *args,
int32 envCount, char * const *env)
{
struct team *team = thread_get_current_thread()->team;
struct team_arg *teamArgs;
const char *threadName;
status_t status = B_OK;
cpu_status state;
struct thread *thread;
thread_id nubThreadID = -1;
TRACE(("exec_team(path = \"%s\", argc = %ld, envCount = %ld): team %ld\n",
args[0], argCount, envCount, team->id));
T(ExecTeam(path, argCount, args, envCount, env));
// switching the kernel at run time is probably not a good idea :)
if (team == team_get_kernel_team())
return B_NOT_ALLOWED;
// we currently need to be single threaded here
// ToDo: maybe we should just kill all other threads and
// make the current thread the team's main thread?
if (team->main_thread != thread_get_current_thread())
return B_NOT_ALLOWED;
// The debug nub thread, a pure kernel thread, is allowed to survive.
// We iterate through the thread list to make sure that there's no other
// thread.
state = disable_interrupts();
GRAB_TEAM_LOCK();
GRAB_TEAM_DEBUG_INFO_LOCK(team->debug_info);
if (team->debug_info.flags & B_TEAM_DEBUG_DEBUGGER_INSTALLED)
nubThreadID = team->debug_info.nub_thread;
RELEASE_TEAM_DEBUG_INFO_LOCK(team->debug_info);
for (thread = team->thread_list; thread; thread = thread->team_next) {
if (thread != team->main_thread && thread->id != nubThreadID) {
status = B_NOT_ALLOWED;
break;
}
}
RELEASE_TEAM_LOCK();
restore_interrupts(state);
if (status != B_OK)
return status;
status = create_team_arg(&teamArgs, argCount, args, envCount, env,
-1, 0, false);
if (status != B_OK)
return status;
// replace args[0] with the path argument, just to be on the safe side
free(teamArgs->args[0]);
teamArgs->args[0] = strdup(path);
// ToDo: remove team resources if there are any left
// thread_atkernel_exit() might not be called at all
thread_reset_for_exec();
user_debug_prepare_for_exec();
vm_delete_areas(team->address_space);
delete_owned_ports(team->id);
sem_delete_owned_sems(team->id);
remove_images(team);
vfs_exec_io_context(team->io_context);
user_debug_finish_after_exec();
// rename the team
set_team_name(team, path);
// cut the path from the team name and rename the main thread, too
threadName = strrchr(path, '/');
if (threadName != NULL)
threadName++;
else
threadName = path;
rename_thread(thread_get_current_thread_id(), threadName);
atomic_or(&team->flags, TEAM_FLAG_EXEC_DONE);
status = team_create_thread_start(teamArgs);
// this one usually doesn't return...
// sorry, we have to kill us, there is no way out anymore
// (without any areas left and all that)
exit_thread(status);
// we return a status here since the signal that is sent by the
// call above is not immediately handled
return B_ERROR;
}
/*! This is the first function to be called from the newly created
main child thread.
It will fill in everything what's left to do from fork_arg, and
return from the parent's fork() syscall to the child.
*/
static int32
fork_team_thread_start(void *_args)
{
struct thread *thread = thread_get_current_thread();
struct fork_arg *forkArgs = (struct fork_arg *)_args;
struct arch_fork_arg archArgs = forkArgs->arch_info;
// we need a local copy of the arch dependent part
thread->user_stack_area = forkArgs->user_stack_area;
thread->user_stack_base = forkArgs->user_stack_base;
thread->user_stack_size = forkArgs->user_stack_size;
thread->user_local_storage = forkArgs->user_local_storage;
thread->sig_block_mask = forkArgs->sig_block_mask;
arch_thread_init_tls(thread);
free(forkArgs);
// set frame of the parent thread to this one, too
arch_restore_fork_frame(&archArgs);
// This one won't return here
return 0;
}
static thread_id
fork_team(void)
{
struct thread *parentThread = thread_get_current_thread();
struct team *parentTeam = parentThread->team, *team;
struct process_group *group = NULL;
struct fork_arg *forkArgs;
struct area_info info;
thread_id threadID;
cpu_status state;
status_t status;
int32 cookie;
TRACE(("fork_team(): team %ld\n", parentTeam->id));
if (parentTeam == team_get_kernel_team())
return B_NOT_ALLOWED;
// create a new team
// ToDo: this is very similar to team_create_team() - maybe we can do something about it :)
team = create_team_struct(parentTeam->name, false);
if (team == NULL)
return B_NO_MEMORY;
strlcpy(team->args, parentTeam->args, sizeof(team->args));
state = disable_interrupts();
GRAB_TEAM_LOCK();
hash_insert(sTeamHash, team);
insert_team_into_parent(parentTeam, team);
insert_team_into_group(parentTeam->group, team);
sUsedTeams++;
RELEASE_TEAM_LOCK();
restore_interrupts(state);
forkArgs = (struct fork_arg *)malloc(sizeof(struct fork_arg));
if (forkArgs == NULL) {
status = B_NO_MEMORY;
goto err1;
}
// create a new io_context for this team
team->io_context = vfs_new_io_context(parentTeam->io_context);
if (!team->io_context) {
status = B_NO_MEMORY;
goto err2;
}
// create an address space for this team
status = vm_create_address_space(team->id, USER_BASE, USER_SIZE, false,
&team->address_space);
if (status < B_OK)
goto err3;
// copy all areas of the team
// ToDo: should be able to handle stack areas differently (ie. don't have them copy-on-write)
// ToDo: all stacks of other threads than the current one could be left out
cookie = 0;
while (get_next_area_info(B_CURRENT_TEAM, &cookie, &info) == B_OK) {
void *address;
area_id area = vm_copy_area(team->address_space->id, info.name,
&address, B_CLONE_ADDRESS, info.protection, info.area);
if (area < B_OK) {
status = area;
break;
}
if (info.area == parentThread->user_stack_area)
forkArgs->user_stack_area = area;
}
if (status < B_OK)
goto err4;
forkArgs->user_stack_base = parentThread->user_stack_base;
forkArgs->user_stack_size = parentThread->user_stack_size;
forkArgs->user_local_storage = parentThread->user_local_storage;
forkArgs->sig_block_mask = parentThread->sig_block_mask;
arch_store_fork_frame(&forkArgs->arch_info);
// ToDo: copy image list
// create a kernel thread under the context of the new team
threadID = spawn_kernel_thread_etc(fork_team_thread_start,
parentThread->name, parentThread->priority, forkArgs,
team->id, team->id);
if (threadID < 0) {
status = threadID;
goto err4;
}
// notify the debugger
user_debug_team_created(team->id);
T(TeamForked(threadID));
resume_thread(threadID);
return threadID;
err4:
vm_delete_address_space(team->address_space);
err3:
vfs_free_io_context(team->io_context);
err2:
free(forkArgs);
err1:
// remove the team structure from the team hash table and delete the team structure
state = disable_interrupts();
GRAB_TEAM_LOCK();
remove_team_from_group(team, &group);
remove_team_from_parent(parentTeam, team);
hash_remove(sTeamHash, team);
RELEASE_TEAM_LOCK();
restore_interrupts(state);
team_delete_process_group(group);
delete_team_struct(team);
return status;
}
/*! Returns if the specified \a team has any children belonging to the
specified \a group.
Must be called with the team lock held.
*/
static bool
has_children_in_group(struct team *parent, pid_t groupID)
{
struct team *team;
struct process_group *group = team_get_process_group_locked(
parent->group->session, groupID);
if (group == NULL)
return false;
for (team = group->teams; team; team = team->group_next) {
if (team->parent == parent)
return true;
}
return false;
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
static job_control_entry*
get_job_control_entry(team_job_control_children* children, pid_t id)
{
for (JobControlEntryList::Iterator it = children->entries.GetIterator();
job_control_entry* entry = it.Next();) {
if (id > 0) {
if (entry->thread == id)
return entry;
} else if (id == -1) {
return entry;
} else {
pid_t processGroup
= (entry->team ? entry->team->group_id : entry->group_id);
if (processGroup == -id)
return entry;
}
}
return NULL;
}
static job_control_entry*
get_job_control_entry(struct team* team, pid_t id, uint32 flags)
{
job_control_entry* entry = get_job_control_entry(team->dead_children, id);
if (entry == NULL && (flags & WCONTINUED) != 0)
entry = get_job_control_entry(team->continued_children, id);
if (entry == NULL && (flags & WUNTRACED) != 0)
entry = get_job_control_entry(team->stopped_children, id);
return entry;
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
/*! This is the kernel backend for waitpid(). It is a bit more powerful when it
comes to the reason why a thread has died than waitpid() can be.
*/
static thread_id
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
wait_for_child(pid_t child, uint32 flags, int32 *_reason,
status_t *_returnCode)
{
struct thread* thread = thread_get_current_thread();
struct team* team = thread->team;
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
struct job_control_entry foundEntry;
struct job_control_entry* freeDeathEntry = NULL;
status_t status = B_OK;
TRACE(("wait_for_child(child = %ld, flags = %ld)\n", child, flags));
T(WaitForChild(child, flags));
if (child == 0) {
// wait for all children in the process group of the calling team
child = -team->group_id;
}
bool ignoreFoundEntries = false;
bool ignoreFoundEntriesChecked = false;
while (true) {
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
InterruptsSpinLocker locker(team_spinlock);
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
// check whether any condition holds
job_control_entry* entry = get_job_control_entry(team, child, flags);
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
// If we don't have an entry yet, check whether there are any children
// complying to the process group specification at all.
if (entry == NULL) {
// No success yet -- check whether there are any children we could
// wait for.
bool childrenExist = false;
if (child == -1) {
childrenExist = team->children != NULL;
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
} else if (child < -1) {
childrenExist = has_children_in_group(team, -child);
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
} else {
if (struct team* childTeam = team_get_team_struct_locked(child))
childrenExist = childTeam->parent == team;
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
}
if (!childrenExist) {
// there is no child we could wait for
status = ECHILD;
} else {
// the children we're waiting for are still running
status = B_WOULD_BLOCK;
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
} else {
// got something
foundEntry = *entry;
if (entry->state == JOB_CONTROL_STATE_DEAD) {
// The child is dead. Reap its death entry.
freeDeathEntry = entry;
team->dead_children->entries.Remove(entry);
team->dead_children->count--;
} else {
// The child is well. Reset its job control state.
team_set_job_control_state(entry->team,
JOB_CONTROL_STATE_NONE, 0, false);
}
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
// If we haven't got anything yet, add prepare for waiting for the
// condition variables.
ConditionVariableEntry<team_dead_children> deadWaitEntry;
ConditionVariableEntry<team_job_control_children> continuedWaitEntry;
ConditionVariableEntry<team_job_control_children> stoppedWaitEntry;
if (status == B_WOULD_BLOCK && (flags & WNOHANG) == 0) {
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
deadWaitEntry.Add(team->dead_children);
if ((flags & WCONTINUED) != 0) {
continuedWaitEntry.Add(team->continued_children,
&deadWaitEntry);
}
if ((flags & WUNTRACED) != 0)
stoppedWaitEntry.Add(team->stopped_children, &deadWaitEntry);
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
locker.Unlock();
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
// we got our entry and can return to our caller
if (status == B_OK) {
if (ignoreFoundEntries) {
// ... unless we shall ignore found entries
delete freeDeathEntry;
freeDeathEntry = NULL;
continue;
}
break;
}
if (status != B_WOULD_BLOCK || (flags & WNOHANG) != 0) {
T(WaitForChildDone(status));
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
return status;
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
status = deadWaitEntry.Wait(B_CAN_INTERRUPT);
if (status == B_INTERRUPTED) {
T(WaitForChildDone(status));
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
return status;
}
// If SA_NOCLDWAIT is set or SIGCHLD is ignored, we shall wait until
// all our children are dead and fail with ECHILD. We check the
// condition at this point.
if (!ignoreFoundEntriesChecked) {
struct sigaction& handler = thread->sig_action[SIGCHLD - 1];
if ((handler.sa_flags & SA_NOCLDWAIT) != 0
|| handler.sa_handler == SIG_IGN) {
ignoreFoundEntries = true;
}
ignoreFoundEntriesChecked = true;
}
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
delete freeDeathEntry;
// when we got here, we have a valid death entry, and
// already got unregistered from the team or group
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
int reason = 0;
switch (foundEntry.state) {
case JOB_CONTROL_STATE_DEAD:
reason = foundEntry.reason;
break;
case JOB_CONTROL_STATE_STOPPED:
reason = THREAD_STOPPED;
break;
case JOB_CONTROL_STATE_CONTINUED:
reason = THREAD_CONTINUED;
break;
case JOB_CONTROL_STATE_NONE:
// can't happen
break;
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
*_returnCode = foundEntry.status;
*_reason = (foundEntry.signal << 16) | reason;
// If SIGCHLD is blocked, we shall clear pending SIGCHLDs, if no other child
// status is available.
if (is_signal_blocked(SIGCHLD)) {
InterruptsSpinLocker locker(team_spinlock);
if (get_job_control_entry(team, child, flags) == NULL)
atomic_and(&thread->sig_pending, ~SIGNAL_TO_MASK(SIGCHLD));
}
T(WaitForChildDone(foundEntry));
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
return foundEntry.thread;
}
/*! Fills the team_info structure with information from the specified
team.
The team lock must be held when called.
*/
static status_t
fill_team_info(struct team *team, team_info *info, size_t size)
{
if (size != sizeof(team_info))
return B_BAD_VALUE;
// ToDo: Set more informations for team_info
memset(info, 0, size);
info->team = team->id;
info->thread_count = team->num_threads;
info->image_count = count_images(team);
//info->area_count =
info->debugger_nub_thread = team->debug_info.nub_thread;
info->debugger_nub_port = team->debug_info.nub_port;
//info->uid =
//info->gid =
strlcpy(info->args, team->args, sizeof(info->args));
info->argc = 1;
return B_OK;
}
/*! Updates the \c orphaned field of a process_group and returns its new value.
Interrupts must be disabled and team lock be held.
*/
static bool
update_orphaned_process_group(process_group* group, pid_t dyingProcess)
{
// Orphaned Process Group: "A process group in which the parent of every
// member is either itself a member of the group or is not a member of the
// group's session." (Open Group Base Specs Issue 6)
// once orphaned, things won't change (exception: cf. setpgid())
if (group->orphaned)
return true;
struct team* team = group->teams;
while (team != NULL) {
struct team* parent = team->parent;
if (team->id != dyingProcess && parent->id != dyingProcess
&& parent->group_id != group->id
&& parent->session_id == group->session->id) {
return false;
}
team = team->group_next;
}
group->orphaned = true;
return true;
}
/*! Returns whether the process group contains stopped processes.
Interrupts must be disabled and team lock be held.
*/
static bool
process_group_has_stopped_processes(process_group* group)
{
SpinLocker _(thread_spinlock);
struct team* team = group->teams;
while (team != NULL) {
if (team->main_thread->state == B_THREAD_SUSPENDED)
return true;
team = team->group_next;
}
return false;
}
// #pragma mark - Private kernel API
status_t
team_init(kernel_args *args)
{
struct process_session *session;
struct process_group *group;
// create the team hash table
sTeamHash = hash_init(16, offsetof(struct team, next),
&team_struct_compare, &team_struct_hash);
sGroupHash = hash_init(16, offsetof(struct process_group, next),
&process_group_compare, &process_group_hash);
// create initial session and process groups
session = create_process_session(1);
if (session == NULL)
panic("Could not create initial session.\n");
group = create_process_group(1);
if (group == NULL)
panic("Could not create initial process group.\n");
insert_group_into_session(session, group);
// create the kernel team
sKernelTeam = create_team_struct("kernel_team", true);
if (sKernelTeam == NULL)
panic("could not create kernel team!\n");
strcpy(sKernelTeam->args, sKernelTeam->name);
sKernelTeam->state = TEAM_STATE_NORMAL;
insert_team_into_group(group, sKernelTeam);
sKernelTeam->io_context = vfs_new_io_context(NULL);
if (sKernelTeam->io_context == NULL)
panic("could not create io_context for kernel team!\n");
// stick it in the team hash
hash_insert(sTeamHash, sKernelTeam);
add_debugger_command_etc("team", &dump_team_info,
"Dump info about a particular team",
"[ <id> | <address> | <name> ]\n"
"Prints information about the specified team. If no argument is given\n"
"the current team is selected.\n"
" <id> - The ID of the team.\n"
" <address> - The address of the team structure.\n"
" <name> - The team's name.\n", 0);
add_debugger_command_etc("teams", &dump_teams, "List all teams",
"\n"
"Prints a list of all existing teams.\n", 0);
return 0;
}
int32
team_max_teams(void)
{
return sMaxTeams;
}
int32
team_used_teams(void)
{
return sUsedTeams;
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
/*! Fills the provided death entry if it's in the team.
You need to have the team lock held when calling this function.
*/
job_control_entry*
team_get_death_entry(struct team *team, thread_id child, bool* _deleteEntry)
{
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
if (child <= 0)
return NULL;
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
job_control_entry* entry = get_job_control_entry(team->dead_children,
child);
if (entry) {
// remove the entry only, if the caller is the parent of the found team
if (team_get_current_team_id() == entry->thread) {
team->dead_children->entries.Remove(entry);
team->dead_children->count--;
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
*_deleteEntry = true;
} else {
*_deleteEntry = false;
}
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
return entry;
}
/*! Quick check to see if we have a valid team ID. */
bool
team_is_valid(team_id id)
{
struct team *team;
cpu_status state;
if (id <= 0)
return false;
state = disable_interrupts();
GRAB_TEAM_LOCK();
team = team_get_team_struct_locked(id);
RELEASE_TEAM_LOCK();
restore_interrupts(state);
return team != NULL;
}
struct team *
team_get_team_struct_locked(team_id id)
{
struct team_key key;
key.id = id;
return (struct team*)hash_lookup(sTeamHash, &key);
}
/*! This searches the session of the team for the specified group ID.
You must hold the team lock when you call this function.
*/
struct process_group *
team_get_process_group_locked(struct process_session *session, pid_t id)
{
struct process_group *group;
struct team_key key;
key.id = id;
group = (struct process_group *)hash_lookup(sGroupHash, &key);
if (group != NULL && (session == NULL || session == group->session))
return group;
return NULL;
}
void
team_delete_process_group(struct process_group *group)
{
if (group == NULL)
return;
TRACE(("team_delete_process_group(id = %ld)\n", group->id));
// remove_group_from_session() keeps this pointer around
// only if the session can be freed as well
if (group->session) {
TRACE(("team_delete_process_group(): frees session %ld\n", group->session->id));
free(group->session);
}
free(group);
}
void
team_set_controlling_tty(int32 ttyIndex)
{
struct team* team = thread_get_current_thread()->team;
InterruptsSpinLocker _(team_spinlock);
team->group->session->controlling_tty = ttyIndex;
team->group->session->foreground_group = -1;
}
int32
team_get_controlling_tty()
{
struct team* team = thread_get_current_thread()->team;
InterruptsSpinLocker _(team_spinlock);
return team->group->session->controlling_tty;
}
status_t
team_set_foreground_process_group(int32 ttyIndex, pid_t processGroupID)
{
struct thread* thread = thread_get_current_thread();
struct team* team = thread->team;
InterruptsSpinLocker locker(team_spinlock);
process_session* session = team->group->session;
// must be the controlling tty of the calling process
if (session->controlling_tty != ttyIndex)
return ENOTTY;
// check process group -- must belong to our session
process_group* group = team_get_process_group_locked(session,
processGroupID);
if (group == NULL)
return B_BAD_VALUE;
// If we are a background group, we can't do that unharmed, only if we
// ignore or block SIGTTOU. Otherwise the group gets a SIGTTOU.
if (session->foreground_group != -1
&& session->foreground_group != team->group_id
&& thread->sig_action[SIGTTOU - 1].sa_handler != SIG_IGN
&& !is_signal_blocked(SIGTTOU)) {
pid_t groupID = team->group->id;
locker.Unlock();
send_signal(-groupID, SIGTTOU);
return B_INTERRUPTED;
}
team->group->session->foreground_group = processGroupID;
return B_OK;
}
/*! Removes the specified team from the global team hash, and from its parent.
It also moves all of its children up to the parent.
You must hold the team lock when you call this function.
If \a _freeGroup is set to a value other than \c NULL, it must be freed
from the calling function.
*/
void
team_remove_team(struct team *team, struct process_group **_freeGroup)
{
struct team *parent = team->parent;
// remember how long this team lasted
parent->dead_children->kernel_time += team->dead_threads_kernel_time
+ team->dead_children->kernel_time;
parent->dead_children->user_time += team->dead_threads_user_time
+ team->dead_children->user_time;
hash_remove(sTeamHash, team);
sUsedTeams--;
team->state = TEAM_STATE_DEATH;
// If we're a controlling process (i.e. a session leader with controlling
// terminal), there's a bit of signalling we have to do.
if (team->session_id == team->id
&& team->group->session->controlling_tty >= 0) {
process_session* session = team->group->session;
session->controlling_tty = -1;
// send SIGHUP to the foreground
if (session->foreground_group >= 0) {
send_signal_etc(-session->foreground_group, SIGHUP,
SIGNAL_FLAG_TEAMS_LOCKED);
}
// send SIGHUP + SIGCONT to all newly-orphaned process groups with
// stopped processes
struct team* child = team->children;
while (child != NULL) {
process_group* childGroup = child->group;
if (!childGroup->orphaned
&& update_orphaned_process_group(childGroup, team->id)
&& process_group_has_stopped_processes(childGroup)) {
send_signal_etc(-childGroup->id, SIGHUP,
SIGNAL_FLAG_TEAMS_LOCKED);
send_signal_etc(-childGroup->id, SIGCONT,
SIGNAL_FLAG_TEAMS_LOCKED);
}
child = child->siblings_next;
}
} else {
// update "orphaned" flags of all children's process groups
struct team* child = team->children;
while (child != NULL) {
process_group* childGroup = child->group;
if (!childGroup->orphaned)
update_orphaned_process_group(childGroup, team->id);
child = child->siblings_next;
}
// update "orphaned" flag of this team's process group
update_orphaned_process_group(team->group, team->id);
}
// reparent each of the team's children
reparent_children(team);
// remove us from our process group
remove_team_from_group(team, _freeGroup);
// remove us from our parent
remove_team_from_parent(parent, team);
}
void
team_delete_team(struct team *team)
{
team_id teamID = team->id;
port_id debuggerPort = -1;
cpu_status state;
if (team->num_threads > 0) {
// there are other threads still in this team,
// cycle through and signal kill on each of the threads
// ToDo: this can be optimized. There's got to be a better solution.
struct thread *temp_thread;
char death_sem_name[B_OS_NAME_LENGTH];
sem_id deathSem;
int32 threadCount;
sprintf(death_sem_name, "team %ld death sem", teamID);
deathSem = create_sem(0, death_sem_name);
if (deathSem < 0)
panic("team_delete_team: cannot init death sem for team %ld\n", teamID);
state = disable_interrupts();
GRAB_TEAM_LOCK();
team->death_sem = deathSem;
threadCount = team->num_threads;
// If the team was being debugged, that will stop with the termination
// of the nub thread. The team structure has already been removed from
// the team hash table at this point, so noone can install a debugger
// anymore. We fetch the debugger's port to send it a message at the
// bitter end.
GRAB_TEAM_DEBUG_INFO_LOCK(team->debug_info);
if (team->debug_info.flags & B_TEAM_DEBUG_DEBUGGER_INSTALLED)
debuggerPort = team->debug_info.debugger_port;
RELEASE_TEAM_DEBUG_INFO_LOCK(team->debug_info);
// we can safely walk the list because of the lock. no new threads can be created
// because of the TEAM_STATE_DEATH flag on the team
temp_thread = team->thread_list;
while (temp_thread) {
struct thread *next = temp_thread->team_next;
send_signal_etc(temp_thread->id, SIGKILLTHR, B_DO_NOT_RESCHEDULE);
temp_thread = next;
}
RELEASE_TEAM_LOCK();
restore_interrupts(state);
// wait until all threads in team are dead.
acquire_sem_etc(team->death_sem, threadCount, 0, 0);
delete_sem(team->death_sem);
}
// If someone is waiting for this team to be loaded, but it dies
// unexpectedly before being done, we need to notify the waiting
// thread now.
state = disable_interrupts();
GRAB_TEAM_LOCK();
if (team->loading_info) {
// there's indeed someone waiting
struct team_loading_info *loadingInfo = team->loading_info;
team->loading_info = NULL;
loadingInfo->result = B_ERROR;
loadingInfo->done = true;
GRAB_THREAD_LOCK();
// wake up the waiting thread
if (loadingInfo->thread->state == B_THREAD_SUSPENDED)
scheduler_enqueue_in_run_queue(loadingInfo->thread);
RELEASE_THREAD_LOCK();
}
RELEASE_TEAM_LOCK();
restore_interrupts(state);
// notify team watchers
{
// we're not reachable from anyone anymore at this point, so we
// can safely access the list without any locking
struct team_watcher *watcher;
while ((watcher = (struct team_watcher*)list_remove_head_item(
&team->watcher_list)) != NULL) {
watcher->hook(teamID, watcher->data);
free(watcher);
}
}
// free team resources
vfs_free_io_context(team->io_context);
delete_owned_ports(teamID);
sem_delete_owned_sems(teamID);
remove_images(team);
vm_delete_address_space(team->address_space);
delete_team_struct(team);
// notify the debugger, that the team is gone
user_debug_team_deleted(teamID, debuggerPort);
}
struct team *
team_get_kernel_team(void)
{
return sKernelTeam;
}
team_id
team_get_kernel_team_id(void)
{
if (!sKernelTeam)
return 0;
return sKernelTeam->id;
}
team_id
team_get_current_team_id(void)
{
return thread_get_current_thread()->team->id;
}
status_t
team_get_address_space(team_id id, vm_address_space **_addressSpace)
{
cpu_status state;
struct team *team;
status_t status;
// ToDo: we need to do something about B_SYSTEM_TEAM vs. its real ID (1)
if (id == 1) {
// we're the kernel team, so we don't have to go through all
// the hassle (locking and hash lookup)
*_addressSpace = vm_get_kernel_address_space();
return B_OK;
}
state = disable_interrupts();
GRAB_TEAM_LOCK();
team = team_get_team_struct_locked(id);
if (team != NULL) {
atomic_add(&team->address_space->ref_count, 1);
*_addressSpace = team->address_space;
status = B_OK;
} else
status = B_BAD_VALUE;
RELEASE_TEAM_LOCK();
restore_interrupts(state);
return status;
}
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
/*! Sets the team's job control state.
Interrupts must be disabled and the team lock be held.
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
\a threadsLocked indicates whether the thread lock is being held, too.
*/
void
team_set_job_control_state(struct team* team, job_control_state newState,
int signal, bool threadsLocked)
{
if (team == NULL || team->job_control_entry == NULL)
return;
// don't touch anything, if the state stays the same or the team is already
// dead
job_control_entry* entry = team->job_control_entry;
if (entry->state == newState || entry->state == JOB_CONTROL_STATE_DEAD)
return;
T(SetJobControlState(team->id, newState, signal));
* Introduced new job_control_entry structure which, among other things, is used instead of death_entry for team::dead_children. * Added team::{stopped,continued}_children, which, analoguously to dead_children, are used to track the state of stopped/continued children. * A team does have a job_control_entry, which is allocated at team creation time. It will be inserted into the parent's {stopped,continued}_children lists as the team's main thread is stopped/continued and removed when waitpid() retrieves the child state. When the team dies the entry is detached from the team and goes into the parent's dead_children list. * Removed the wait_for_any field from team_dead_children. It was solely used to avoid deletion of the contained entries in certain situations. wait_for_child() (the waitpid() backend) always deletes an entry now, regardless of whether other threads are waiting; that's in accordance with the waidpid() specification. wait_for_thread() removes the entry only, if the caller is the parent of the respective team. * Introduced team_set_job_control_state() which performes the job control entry transitions between the respective lists and wakes up threads waiting in wait_for_child(). It is invoked on team death and when the team's main thread receives job control signals. * Reorganized wait_for_child(). It handles WCONTINUED and WUNTRACED now, too. Removed a block that interpreted the supplied ID as thread ID. * Added missing parts in waitpid(). Job control starts to work, though it seems to have some glitches. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22088 a95241bf-73f2-0310-859d-f6bbb57e9c96
2007-08-28 07:29:14 +04:00
// remove from the old list
switch (entry->state) {
case JOB_CONTROL_STATE_NONE:
// entry is in no list ATM
break;
case JOB_CONTROL_STATE_DEAD:
// can't get here
break;
case JOB_CONTROL_STATE_STOPPED:
team->parent->stopped_children->entries.Remove(entry);
break;
case JOB_CONTROL_STATE_CONTINUED:
team->parent->continued_children->entries.Remove(entry);
break;
}
entry->state = newState;
entry->signal = signal;
// add to new list
team_job_control_children* childList = NULL;
switch (entry->state) {
case JOB_CONTROL_STATE_NONE:
// entry doesn't get into any list
break;
case JOB_CONTROL_STATE_DEAD:
childList = team->parent->dead_children;
team->parent->dead_children->count++;
// When a child dies, we need to notify all lists, since that might
// have been the last of the parent's children, and a waiting
// parent thread wouldn't wake up otherwise.
team->parent->stopped_children->condition_variable.NotifyAll(
threadsLocked);
team->parent->continued_children->condition_variable.NotifyAll(
threadsLocked);
break;
case JOB_CONTROL_STATE_STOPPED:
childList = team->parent->stopped_children;
break;
case JOB_CONTROL_STATE_CONTINUED:
childList = team->parent->continued_children;
break;
}
if (childList != NULL) {
childList->entries.Add(entry);
childList->condition_variable.NotifyAll(threadsLocked);
}
}
/*! Adds a hook to the team that is called as soon as this
team goes away.
This call might get public in the future.
*/
status_t
start_watching_team(team_id teamID, void (*hook)(team_id, void *), void *data)
{
struct team_watcher *watcher;
struct team *team;
cpu_status state;
if (hook == NULL || teamID < B_OK)
return B_BAD_VALUE;
watcher = (struct team_watcher*)malloc(sizeof(struct team_watcher));
if (watcher == NULL)
return B_NO_MEMORY;
watcher->hook = hook;
watcher->data = data;
// find team and add watcher
state = disable_interrupts();
GRAB_TEAM_LOCK();
team = team_get_team_struct_locked(teamID);
if (team != NULL)
list_add_item(&team->watcher_list, watcher);
RELEASE_TEAM_LOCK();
restore_interrupts(state);
if (team == NULL) {
free(watcher);
return B_BAD_TEAM_ID;
}
return B_OK;
}
status_t
stop_watching_team(team_id teamID, void (*hook)(team_id, void *), void *data)
{
struct team_watcher *watcher = NULL;
struct team *team;
cpu_status state;
if (hook == NULL || teamID < B_OK)
return B_BAD_VALUE;
// find team and remove watcher (if present)
state = disable_interrupts();
GRAB_TEAM_LOCK();
team = team_get_team_struct_locked(teamID);
if (team != NULL) {
// search for watcher
while ((watcher = (struct team_watcher*)list_get_next_item(
&team->watcher_list, watcher)) != NULL) {
if (watcher->hook == hook && watcher->data == data) {
// got it!
list_remove_item(&team->watcher_list, watcher);
break;
}
}
}
RELEASE_TEAM_LOCK();
restore_interrupts(state);
if (watcher == NULL)
return B_ENTRY_NOT_FOUND;
free(watcher);
return B_OK;
}
// #pragma mark - Public kernel API
thread_id
load_image(int32 argCount, const char **args, const char **env)
{
int32 envCount = 0;
// count env variables
while (env && env[envCount] != NULL)
envCount++;
return load_image_etc(argCount, (char * const *)args, envCount,
(char * const *)env, B_NORMAL_PRIORITY, B_WAIT_TILL_LOADED,
-1, 0, true);
}
status_t
wait_for_team(team_id id, status_t *_returnCode)
{
struct team *team;
thread_id thread;
cpu_status state;
// find main thread and wait for that
state = disable_interrupts();
GRAB_TEAM_LOCK();
team = team_get_team_struct_locked(id);
if (team != NULL && team->main_thread != NULL)
thread = team->main_thread->id;
else
thread = B_BAD_THREAD_ID;
RELEASE_TEAM_LOCK();
restore_interrupts(state);
if (thread < 0)
return thread;
return wait_for_thread(thread, _returnCode);
}
status_t
kill_team(team_id id)
{
status_t status = B_OK;
thread_id threadID = -1;
struct team *team;
cpu_status state;
state = disable_interrupts();
GRAB_TEAM_LOCK();
team = team_get_team_struct_locked(id);
if (team != NULL) {
if (team != sKernelTeam) {
threadID = team->id;
// the team ID is the same as the ID of its main thread
} else
status = B_NOT_ALLOWED;
} else
status = B_BAD_THREAD_ID;
RELEASE_TEAM_LOCK();
restore_interrupts(state);
if (status < B_OK)
return status;
// just kill the main thread in the team. The cleanup code there will
// take care of the team
return kill_thread(threadID);
}
status_t
_get_team_info(team_id id, team_info *info, size_t size)
{
cpu_status state;
status_t status = B_OK;
struct team *team;
state = disable_interrupts();
GRAB_TEAM_LOCK();
if (id == B_CURRENT_TEAM)
team = thread_get_current_thread()->team;
else
team = team_get_team_struct_locked(id);
if (team == NULL) {
status = B_BAD_TEAM_ID;
goto err;
}
status = fill_team_info(team, info, size);
err:
RELEASE_TEAM_LOCK();
restore_interrupts(state);
return status;
}
status_t
_get_next_team_info(int32 *cookie, team_info *info, size_t size)
{
status_t status = B_BAD_TEAM_ID;
struct team *team = NULL;
int32 slot = *cookie;
team_id lastTeamID;
cpu_status state;
if (slot < 1)
slot = 1;
state = disable_interrupts();
GRAB_TEAM_LOCK();
lastTeamID = peek_next_thread_id();
if (slot >= lastTeamID)
goto err;
// get next valid team
while (slot < lastTeamID && !(team = team_get_team_struct_locked(slot)))
slot++;
if (team) {
status = fill_team_info(team, info, size);
*cookie = ++slot;
}
err:
RELEASE_TEAM_LOCK();
restore_interrupts(state);
return status;
}
status_t
_get_team_usage_info(team_id id, int32 who, team_usage_info *info, size_t size)
{
bigtime_t kernelTime = 0, userTime = 0;
status_t status = B_OK;
struct team *team;
cpu_status state;
if (size != sizeof(team_usage_info)
|| (who != B_TEAM_USAGE_SELF && who != B_TEAM_USAGE_CHILDREN))
return B_BAD_VALUE;
state = disable_interrupts();
GRAB_TEAM_LOCK();
if (id == B_CURRENT_TEAM)
team = thread_get_current_thread()->team;
else
team = team_get_team_struct_locked(id);
if (team == NULL) {
status = B_BAD_TEAM_ID;
goto out;
}
switch (who) {
case B_TEAM_USAGE_SELF:
{
struct thread *thread = team->thread_list;
for (; thread != NULL; thread = thread->team_next) {
kernelTime += thread->kernel_time;
userTime += thread->user_time;
}
kernelTime += team->dead_threads_kernel_time;
userTime += team->dead_threads_user_time;
break;
}
case B_TEAM_USAGE_CHILDREN:
{
struct team *child = team->children;
for (; child != NULL; child = child->siblings_next) {
struct thread *thread = team->thread_list;
for (; thread != NULL; thread = thread->team_next) {
kernelTime += thread->kernel_time;
userTime += thread->user_time;
}
kernelTime += child->dead_threads_kernel_time;
userTime += child->dead_threads_user_time;
}
kernelTime += team->dead_children->kernel_time;
userTime += team->dead_children->user_time;
break;
}
}
out:
RELEASE_TEAM_LOCK();
restore_interrupts(state);
if (status == B_OK) {
info->kernel_time = kernelTime;
info->user_time = userTime;
}
return status;
}
pid_t
getpid(void)
{
return thread_get_current_thread()->team->id;
}
pid_t
getppid(void)
{
struct team *team = thread_get_current_thread()->team;
cpu_status state;
pid_t parent;
state = disable_interrupts();
GRAB_TEAM_LOCK();
parent = team->parent->id;
RELEASE_TEAM_LOCK();
restore_interrupts(state);
return parent;
}
pid_t
getpgid(pid_t process)
{
struct thread *thread;
pid_t result = -1;
cpu_status state;
if (process == 0)
process = thread_get_current_thread()->team->id;
state = disable_interrupts();
GRAB_THREAD_LOCK();
thread = thread_get_thread_struct_locked(process);
if (thread != NULL)
result = thread->team->group_id;
RELEASE_THREAD_LOCK();
restore_interrupts(state);
return thread != NULL ? result : B_BAD_VALUE;
}
pid_t
getsid(pid_t process)
{
struct thread *thread;
pid_t result = -1;
cpu_status state;
if (process == 0)
process = thread_get_current_thread()->team->id;
state = disable_interrupts();
GRAB_THREAD_LOCK();
thread = thread_get_thread_struct_locked(process);
if (thread != NULL)
result = thread->team->session_id;
RELEASE_THREAD_LOCK();
restore_interrupts(state);
return thread != NULL ? result : B_BAD_VALUE;
}
// #pragma mark - User syscalls
status_t
_user_exec(const char *userPath, int32 argCount, char * const *userArgs,
int32 envCount, char * const *userEnvironment)
{
char path[B_PATH_NAME_LENGTH];
if (argCount < 1)
return B_BAD_VALUE;
if (!IS_USER_ADDRESS(userPath) || !IS_USER_ADDRESS(userArgs)
|| !IS_USER_ADDRESS(userEnvironment)
|| user_strlcpy(path, userPath, sizeof(path)) < B_OK)
return B_BAD_ADDRESS;
return exec_team(path, argCount, userArgs, envCount, userEnvironment);
// this one only returns in case of error
}
thread_id
_user_fork(void)
{
return fork_team();
}
thread_id
_user_wait_for_child(thread_id child, uint32 flags, int32 *_userReason, status_t *_userReturnCode)
{
status_t returnCode;
int32 reason;
thread_id deadChild;
if ((_userReason != NULL && !IS_USER_ADDRESS(_userReason))
|| (_userReturnCode != NULL && !IS_USER_ADDRESS(_userReturnCode)))
return B_BAD_ADDRESS;
deadChild = wait_for_child(child, flags, &reason, &returnCode);
if (deadChild >= B_OK) {
// copy result data on successful completion
axeld + bonefish: * Implemented automatic syscall restarts: - A syscall can indicate that it has been interrupted and can be restarted by setting a respective bit in thread::flags. It can store parameters it wants to be preserved for the restart in thread::syscall_restart::parameters. Another thread::flags bit indicates whether it has been restarted. - handle_signals() clears the restart flag, if the handled signal has a handler function installed and SA_RESTART is not set. Another thread flag (THREAD_FLAGS_DONT_RESTART_SYSCALL) can prevent syscalls from being restarted, even if they could be (not used yet, but we might want to use it in resume_thread(), so that we stay behaviorally compatible with BeOS). - The architecture specific syscall handler restarts the syscall, if the restart flag is set. Implemented for x86 only. - Added some support functions in the private <syscall_restart.h> to simplify the syscall restart code in the syscalls. - Adjusted all syscalls that can potentially be restarted accordingly. - _user_ioctl() sets new thread flag THREAD_FLAGS_IOCTL_SYSCALL while calling the underlying FS's/driver's hook, so that syscall restarts can also be supported there. * thread_at_kernel_exit() invokes handle_signals() in a loop now, as long as the latter indicates that the thread shall be suspended, so that after waking up signals received in the meantime will be handled before the thread returns to userland. Adjusted handle_signals() accordingly -- when encountering a suspending signal we don't check for further signals. * Fixed sigsuspend(): Suspending the thread and rescheduling doesn't result in the correct behavior. Instead we employ a temporary condition variable and interruptably wait on it. The POSIX test suite test passes, now. * Made the switch_sem[_etc]() behavior on interruption consistent. Depending on when the signal arrived (before the call or when already waiting) the first semaphore would or wouldn't be released. Now we consistently release it. * Refactored _user_{read,write}[v]() syscalls. Use a common function for either pair. The iovec version doesn't fail anymore, if anything could be read/written at all. It also checks whether a complete vector could be read/written, so that we won't skip data, if the underlying FS/driver couldn't read/write more ATM. * Some refactoring in the x86 syscall handler: The int 99 and sysenter handlers use a common subroutine to avoid code duplication. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23983 a95241bf-73f2-0310-859d-f6bbb57e9c96
2008-02-17 18:48:30 +03:00
if ((_userReason != NULL
&& user_memcpy(_userReason, &reason, sizeof(int32)) < B_OK)
|| (_userReturnCode != NULL
&& user_memcpy(_userReturnCode, &returnCode, sizeof(status_t))
< B_OK)) {
return B_BAD_ADDRESS;
axeld + bonefish: * Implemented automatic syscall restarts: - A syscall can indicate that it has been interrupted and can be restarted by setting a respective bit in thread::flags. It can store parameters it wants to be preserved for the restart in thread::syscall_restart::parameters. Another thread::flags bit indicates whether it has been restarted. - handle_signals() clears the restart flag, if the handled signal has a handler function installed and SA_RESTART is not set. Another thread flag (THREAD_FLAGS_DONT_RESTART_SYSCALL) can prevent syscalls from being restarted, even if they could be (not used yet, but we might want to use it in resume_thread(), so that we stay behaviorally compatible with BeOS). - The architecture specific syscall handler restarts the syscall, if the restart flag is set. Implemented for x86 only. - Added some support functions in the private <syscall_restart.h> to simplify the syscall restart code in the syscalls. - Adjusted all syscalls that can potentially be restarted accordingly. - _user_ioctl() sets new thread flag THREAD_FLAGS_IOCTL_SYSCALL while calling the underlying FS's/driver's hook, so that syscall restarts can also be supported there. * thread_at_kernel_exit() invokes handle_signals() in a loop now, as long as the latter indicates that the thread shall be suspended, so that after waking up signals received in the meantime will be handled before the thread returns to userland. Adjusted handle_signals() accordingly -- when encountering a suspending signal we don't check for further signals. * Fixed sigsuspend(): Suspending the thread and rescheduling doesn't result in the correct behavior. Instead we employ a temporary condition variable and interruptably wait on it. The POSIX test suite test passes, now. * Made the switch_sem[_etc]() behavior on interruption consistent. Depending on when the signal arrived (before the call or when already waiting) the first semaphore would or wouldn't be released. Now we consistently release it. * Refactored _user_{read,write}[v]() syscalls. Use a common function for either pair. The iovec version doesn't fail anymore, if anything could be read/written at all. It also checks whether a complete vector could be read/written, so that we won't skip data, if the underlying FS/driver couldn't read/write more ATM. * Some refactoring in the x86 syscall handler: The int 99 and sysenter handlers use a common subroutine to avoid code duplication. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23983 a95241bf-73f2-0310-859d-f6bbb57e9c96
2008-02-17 18:48:30 +03:00
}
return deadChild;
}
axeld + bonefish: * Implemented automatic syscall restarts: - A syscall can indicate that it has been interrupted and can be restarted by setting a respective bit in thread::flags. It can store parameters it wants to be preserved for the restart in thread::syscall_restart::parameters. Another thread::flags bit indicates whether it has been restarted. - handle_signals() clears the restart flag, if the handled signal has a handler function installed and SA_RESTART is not set. Another thread flag (THREAD_FLAGS_DONT_RESTART_SYSCALL) can prevent syscalls from being restarted, even if they could be (not used yet, but we might want to use it in resume_thread(), so that we stay behaviorally compatible with BeOS). - The architecture specific syscall handler restarts the syscall, if the restart flag is set. Implemented for x86 only. - Added some support functions in the private <syscall_restart.h> to simplify the syscall restart code in the syscalls. - Adjusted all syscalls that can potentially be restarted accordingly. - _user_ioctl() sets new thread flag THREAD_FLAGS_IOCTL_SYSCALL while calling the underlying FS's/driver's hook, so that syscall restarts can also be supported there. * thread_at_kernel_exit() invokes handle_signals() in a loop now, as long as the latter indicates that the thread shall be suspended, so that after waking up signals received in the meantime will be handled before the thread returns to userland. Adjusted handle_signals() accordingly -- when encountering a suspending signal we don't check for further signals. * Fixed sigsuspend(): Suspending the thread and rescheduling doesn't result in the correct behavior. Instead we employ a temporary condition variable and interruptably wait on it. The POSIX test suite test passes, now. * Made the switch_sem[_etc]() behavior on interruption consistent. Depending on when the signal arrived (before the call or when already waiting) the first semaphore would or wouldn't be released. Now we consistently release it. * Refactored _user_{read,write}[v]() syscalls. Use a common function for either pair. The iovec version doesn't fail anymore, if anything could be read/written at all. It also checks whether a complete vector could be read/written, so that we won't skip data, if the underlying FS/driver couldn't read/write more ATM. * Some refactoring in the x86 syscall handler: The int 99 and sysenter handlers use a common subroutine to avoid code duplication. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23983 a95241bf-73f2-0310-859d-f6bbb57e9c96
2008-02-17 18:48:30 +03:00
return syscall_restart_handle_post(deadChild);
}
pid_t
_user_process_info(pid_t process, int32 which)
{
// we only allow to return the parent of the current process
if (which == PARENT_ID
&& process != 0 && process != thread_get_current_thread()->team->id)
return B_BAD_VALUE;
switch (which) {
case SESSION_ID:
return getsid(process);
case GROUP_ID:
return getpgid(process);
case PARENT_ID:
return getppid();
}
return B_BAD_VALUE;
}
pid_t
_user_setpgid(pid_t processID, pid_t groupID)
{
struct thread *thread = thread_get_current_thread();
struct team *currentTeam = thread->team;
struct team *team;
if (groupID < 0)
return B_BAD_VALUE;
if (processID == 0)
processID = currentTeam->id;
// if the group ID is not specified, use the target process' ID
if (groupID == 0)
groupID = processID;
if (processID == currentTeam->id) {
// we set our own group
// we must not change our process group ID if we're a session leader
if (is_session_leader(currentTeam))
return B_NOT_ALLOWED;
} else {
// another team is the target of the call -- check it out
InterruptsSpinLocker _(team_spinlock);
team = team_get_team_struct_locked(processID);
if (team == NULL)
return ESRCH;
// The team must be a child of the calling team and in the same session.
// (If that's the case it isn't a session leader either.)
if (team->parent != currentTeam
|| team->session_id != currentTeam->session_id) {
return B_NOT_ALLOWED;
}
if (team->group_id == groupID)
return groupID;
// The call is also supposed to fail on a child, when the child already
// has executed exec*() [EACCES].
if ((team->flags & TEAM_FLAG_EXEC_DONE) != 0)
return EACCES;
}
struct process_group *group = NULL;
if (groupID == processID) {
// A new process group might be needed.
group = create_process_group(groupID);
if (group == NULL)
return B_NO_MEMORY;
// Assume orphaned. We consider the situation of the team's parent
// below.
group->orphaned = true;
}
status_t status = B_OK;
struct process_group *freeGroup = NULL;
struct process_group *freeGroup2 = NULL;
InterruptsSpinLocker locker(team_spinlock);
team = team_get_team_struct_locked(processID);
if (team != NULL) {
// check the conditions again -- they might have changed in the meantime
if (is_session_leader(team)
|| team->session_id != currentTeam->session_id) {
status = B_NOT_ALLOWED;
} else if (team != currentTeam
&& (team->flags & TEAM_FLAG_EXEC_DONE) != 0) {
status = EACCES;
} else if (team->group_id == groupID) {
// the team is already in the desired process group
freeGroup = group;
} else {
// Check if a process group with the requested ID already exists.
struct process_group *targetGroup
= team_get_process_group_locked(team->group->session, groupID);
if (targetGroup != NULL) {
// In case of processID == groupID we have to free the
// allocated group.
freeGroup2 = group;
} else if (processID == groupID) {
// We created a new process group, let us insert it into the
// team's session.
insert_group_into_session(team->group->session, group);
targetGroup = group;
}
if (targetGroup != NULL) {
// we got a group, let's move the team there
process_group* oldGroup = team->group;
remove_team_from_group(team, &freeGroup);
insert_team_into_group(targetGroup, team);
// Update the "orphaned" flag of all potentially affected
// groups.
// the team's old group
if (oldGroup->teams != NULL) {
oldGroup->orphaned = false;
update_orphaned_process_group(oldGroup, -1);
}
// the team's new group
struct team* parent = team->parent;
targetGroup->orphaned &= parent == NULL
|| parent->group == targetGroup
|| team->parent->session_id != team->session_id;
// children's groups
struct team* child = team->children;
while (child != NULL) {
child->group->orphaned = false;
update_orphaned_process_group(child->group, -1);
child = child->siblings_next;
}
} else
status = B_NOT_ALLOWED;
}
} else
status = B_NOT_ALLOWED;
// Changing the process group might have changed the situation for a parent
// waiting in wait_for_child(). Hence we notify it.
if (status == B_OK) {
team->parent->dead_children->condition_variable.NotifyAll(false);
team->parent->stopped_children->condition_variable.NotifyAll(false);
team->parent->continued_children->condition_variable.NotifyAll(false);
}
locker.Unlock();
if (status != B_OK) {
// in case of error, the group hasn't been added into the hash
team_delete_process_group(group);
}
team_delete_process_group(freeGroup);
team_delete_process_group(freeGroup2);
return status == B_OK ? groupID : status;
}
pid_t
_user_setsid(void)
{
struct team *team = thread_get_current_thread()->team;
struct process_session *session;
struct process_group *group, *freeGroup = NULL;
cpu_status state;
bool failed = false;
// the team must not already be a process group leader
if (is_process_group_leader(team))
return B_NOT_ALLOWED;
group = create_process_group(team->id);
if (group == NULL)
return B_NO_MEMORY;
session = create_process_session(group->id);
if (session == NULL) {
team_delete_process_group(group);
return B_NO_MEMORY;
}
state = disable_interrupts();
GRAB_TEAM_LOCK();
// this may have changed since the check above
if (!is_process_group_leader(team)) {
remove_team_from_group(team, &freeGroup);
insert_group_into_session(session, group);
insert_team_into_group(group, team);
} else
failed = true;
RELEASE_TEAM_LOCK();
restore_interrupts(state);
if (failed) {
team_delete_process_group(group);
free(session);
return B_NOT_ALLOWED;
} else
team_delete_process_group(freeGroup);
return team->group_id;
}
status_t
_user_wait_for_team(team_id id, status_t *_userReturnCode)
{
status_t returnCode;
status_t status;
if (_userReturnCode != NULL && !IS_USER_ADDRESS(_userReturnCode))
return B_BAD_ADDRESS;
status = wait_for_team(id, &returnCode);
if (status >= B_OK && _userReturnCode != NULL) {
if (user_memcpy(_userReturnCode, &returnCode, sizeof(returnCode)) < B_OK)
return B_BAD_ADDRESS;
axeld + bonefish: * Implemented automatic syscall restarts: - A syscall can indicate that it has been interrupted and can be restarted by setting a respective bit in thread::flags. It can store parameters it wants to be preserved for the restart in thread::syscall_restart::parameters. Another thread::flags bit indicates whether it has been restarted. - handle_signals() clears the restart flag, if the handled signal has a handler function installed and SA_RESTART is not set. Another thread flag (THREAD_FLAGS_DONT_RESTART_SYSCALL) can prevent syscalls from being restarted, even if they could be (not used yet, but we might want to use it in resume_thread(), so that we stay behaviorally compatible with BeOS). - The architecture specific syscall handler restarts the syscall, if the restart flag is set. Implemented for x86 only. - Added some support functions in the private <syscall_restart.h> to simplify the syscall restart code in the syscalls. - Adjusted all syscalls that can potentially be restarted accordingly. - _user_ioctl() sets new thread flag THREAD_FLAGS_IOCTL_SYSCALL while calling the underlying FS's/driver's hook, so that syscall restarts can also be supported there. * thread_at_kernel_exit() invokes handle_signals() in a loop now, as long as the latter indicates that the thread shall be suspended, so that after waking up signals received in the meantime will be handled before the thread returns to userland. Adjusted handle_signals() accordingly -- when encountering a suspending signal we don't check for further signals. * Fixed sigsuspend(): Suspending the thread and rescheduling doesn't result in the correct behavior. Instead we employ a temporary condition variable and interruptably wait on it. The POSIX test suite test passes, now. * Made the switch_sem[_etc]() behavior on interruption consistent. Depending on when the signal arrived (before the call or when already waiting) the first semaphore would or wouldn't be released. Now we consistently release it. * Refactored _user_{read,write}[v]() syscalls. Use a common function for either pair. The iovec version doesn't fail anymore, if anything could be read/written at all. It also checks whether a complete vector could be read/written, so that we won't skip data, if the underlying FS/driver couldn't read/write more ATM. * Some refactoring in the x86 syscall handler: The int 99 and sysenter handlers use a common subroutine to avoid code duplication. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23983 a95241bf-73f2-0310-859d-f6bbb57e9c96
2008-02-17 18:48:30 +03:00
return B_OK;
}
axeld + bonefish: * Implemented automatic syscall restarts: - A syscall can indicate that it has been interrupted and can be restarted by setting a respective bit in thread::flags. It can store parameters it wants to be preserved for the restart in thread::syscall_restart::parameters. Another thread::flags bit indicates whether it has been restarted. - handle_signals() clears the restart flag, if the handled signal has a handler function installed and SA_RESTART is not set. Another thread flag (THREAD_FLAGS_DONT_RESTART_SYSCALL) can prevent syscalls from being restarted, even if they could be (not used yet, but we might want to use it in resume_thread(), so that we stay behaviorally compatible with BeOS). - The architecture specific syscall handler restarts the syscall, if the restart flag is set. Implemented for x86 only. - Added some support functions in the private <syscall_restart.h> to simplify the syscall restart code in the syscalls. - Adjusted all syscalls that can potentially be restarted accordingly. - _user_ioctl() sets new thread flag THREAD_FLAGS_IOCTL_SYSCALL while calling the underlying FS's/driver's hook, so that syscall restarts can also be supported there. * thread_at_kernel_exit() invokes handle_signals() in a loop now, as long as the latter indicates that the thread shall be suspended, so that after waking up signals received in the meantime will be handled before the thread returns to userland. Adjusted handle_signals() accordingly -- when encountering a suspending signal we don't check for further signals. * Fixed sigsuspend(): Suspending the thread and rescheduling doesn't result in the correct behavior. Instead we employ a temporary condition variable and interruptably wait on it. The POSIX test suite test passes, now. * Made the switch_sem[_etc]() behavior on interruption consistent. Depending on when the signal arrived (before the call or when already waiting) the first semaphore would or wouldn't be released. Now we consistently release it. * Refactored _user_{read,write}[v]() syscalls. Use a common function for either pair. The iovec version doesn't fail anymore, if anything could be read/written at all. It also checks whether a complete vector could be read/written, so that we won't skip data, if the underlying FS/driver couldn't read/write more ATM. * Some refactoring in the x86 syscall handler: The int 99 and sysenter handlers use a common subroutine to avoid code duplication. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23983 a95241bf-73f2-0310-859d-f6bbb57e9c96
2008-02-17 18:48:30 +03:00
return syscall_restart_handle_post(status);
}
team_id
_user_load_image(int32 argCount, const char **userArgs, int32 envCount,
const char **userEnv, int32 priority, uint32 flags, port_id errorPort,
uint32 errorToken)
{
TRACE(("_user_load_image_etc: argc = %ld\n", argCount));
if (argCount < 1 || userArgs == NULL || userEnv == NULL)
return B_BAD_VALUE;
if (!IS_USER_ADDRESS(userArgs) || !IS_USER_ADDRESS(userEnv))
return B_BAD_ADDRESS;
return load_image_etc(argCount, (char * const *)userArgs,
envCount, (char * const *)userEnv, priority, flags, errorPort,
errorToken, false);
}
void
_user_exit_team(status_t returnValue)
{
struct thread *thread = thread_get_current_thread();
thread->exit.status = returnValue;
thread->exit.reason = THREAD_RETURN_EXIT;
send_signal(thread->id, SIGKILL);
}
status_t
_user_kill_team(team_id team)
{
return kill_team(team);
}
status_t
_user_get_team_info(team_id id, team_info *userInfo)
{
status_t status;
team_info info;
if (!IS_USER_ADDRESS(userInfo))
return B_BAD_ADDRESS;
status = _get_team_info(id, &info, sizeof(team_info));
if (status == B_OK) {
if (user_memcpy(userInfo, &info, sizeof(team_info)) < B_OK)
return B_BAD_ADDRESS;
}
return status;
}
status_t
_user_get_next_team_info(int32 *userCookie, team_info *userInfo)
{
status_t status;
team_info info;
int32 cookie;
if (!IS_USER_ADDRESS(userCookie)
|| !IS_USER_ADDRESS(userInfo)
|| user_memcpy(&cookie, userCookie, sizeof(int32)) < B_OK)
return B_BAD_ADDRESS;
status = _get_next_team_info(&cookie, &info, sizeof(team_info));
if (status != B_OK)
return status;
if (user_memcpy(userCookie, &cookie, sizeof(int32)) < B_OK
|| user_memcpy(userInfo, &info, sizeof(team_info)) < B_OK)
return B_BAD_ADDRESS;
return status;
}
team_id
_user_get_current_team(void)
{
return team_get_current_team_id();
}
status_t
_user_get_team_usage_info(team_id team, int32 who, team_usage_info *userInfo, size_t size)
{
team_usage_info info;
status_t status;
if (!IS_USER_ADDRESS(userInfo))
return B_BAD_ADDRESS;
status = _get_team_usage_info(team, who, &info, size);
if (status != B_OK)
return status;
if (user_memcpy(userInfo, &info, size) < B_OK)
return B_BAD_ADDRESS;
return status;
}