Implemented team_get_address_space() which gets you a pointer to the

team's address space - for internal use, anyway.
Added a ToDo: item to exec_team() to remind me of alarms and signals.
Started implementing fork_team() based on team_create_team() - it's not
completed (or even works), though.


git-svn-id: file:///srv/svn/repos/haiku/trunk/current@9279 a95241bf-73f2-0310-859d-f6bbb57e9c96
This commit is contained in:
Axel Dörfler 2004-10-10 17:30:42 +00:00
parent f3c2082a3d
commit 71eeb427c3

View File

@ -361,6 +361,40 @@ team_get_current_team_id(void)
} }
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)
atomic_add(&kernel_team->kaspace->ref_count, 1);
*_addressSpace = kernel_team->kaspace;
return B_OK;
}
state = disable_interrupts();
GRAB_TEAM_LOCK();
team = team_get_team_struct_locked(id);
if (team != NULL) {
atomic_add(&team->aspace->ref_count, 1);
*_addressSpace = team->aspace;
status = B_OK;
} else
status = B_BAD_VALUE;
RELEASE_TEAM_LOCK();
restore_interrupts(state);
return status;
}
static struct team * static struct team *
create_team_struct(const char *name, bool kernel) create_team_struct(const char *name, bool kernel)
{ {
@ -453,8 +487,7 @@ team_delete_team(struct team *team)
// free team resources // free team resources
ASSERT(team->aspace->ref_count == 1); vm_delete_aspace(team->aspace);
vm_put_aspace(team->aspace);
delete_owned_ports(team->id); delete_owned_ports(team->id);
sem_delete_owned_sems(team->id); sem_delete_owned_sems(team->id);
remove_images(team); remove_images(team);
@ -660,7 +693,7 @@ team_create_team(const char *path, const char *name, char **args, int argc, char
} }
// create a new io_context for this team // create a new io_context for this team
team->io_context = vfs_new_io_context(thread_get_current_thread()->team->io_context); team->io_context = vfs_new_io_context(parent->io_context);
if (!team->io_context) { if (!team->io_context) {
err = B_NO_MEMORY; err = B_NO_MEMORY;
goto err2; goto err2;
@ -718,7 +751,7 @@ exec_team(const char *path, int32 argCount, char **args, int32 envCount, char **
struct team_arg *teamArgs; struct team_arg *teamArgs;
status_t status; status_t status;
TRACE(("exec_team(path = \"%s\", argc = %ld, envCount = %ld\n", path, argCount, envCount)); TRACE(("exec_team(path = \"%s\", argc = %ld, envCount = %ld)\n", path, argCount, envCount));
// switching the kernel at run time is probably not a good idea :) // switching the kernel at run time is probably not a good idea :)
if (team == team_get_kernel_team()) if (team == team_get_kernel_team())
@ -739,6 +772,7 @@ exec_team(const char *path, int32 argCount, char **args, int32 envCount, char **
return B_NO_MEMORY; return B_NO_MEMORY;
// ToDo: remove team resources if there are any left // ToDo: remove team resources if there are any left
// alarm, signals
vm_delete_areas(team->aspace); vm_delete_areas(team->aspace);
delete_owned_ports(team->id); delete_owned_ports(team->id);
@ -757,6 +791,111 @@ exec_team(const char *path, int32 argCount, char **args, int32 envCount, char **
} }
static thread_id
fork_team(void)
{
struct team *forkedTeam = thread_get_current_thread()->team, *team;
struct thread *forkedThread = thread_get_current_thread(), *thread;
struct area_info info;
cpu_status state;
status_t status;
int32 cookie;
team_id pid;
TRACE(("fork_team()\n"));
if (forkedTeam == team_get_kernel_team())
return B_NOT_ALLOWED;
dprintf("fork() is not yet implemented!\n");
// create a new team
// ToDo: this is very similar to team_create_team() - maybe we can do something about it :)
team = create_team_struct(forkedTeam->name, false);
if (team == NULL)
return B_NO_MEMORY;
pid = team->id;
state = disable_interrupts();
GRAB_TEAM_LOCK();
hash_insert(team_hash, team);
insert_team_into_parent(forkedTeam, team);
RELEASE_TEAM_LOCK();
restore_interrupts(state);
// create a new io_context for this team
team->io_context = vfs_new_io_context(forkedTeam->io_context);
if (!team->io_context) {
status = B_NO_MEMORY;
goto err2;
}
// create an address space for this team
status = vm_create_aspace(team->name, USER_BASE, USER_SIZE, false, &team->aspace);
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;
status = vm_copy_area(team->aspace->id, info.name, &address, B_CLONE_ADDRESS,
info.protection, info.area);
if (status < B_OK)
break;
}
if (status < B_OK)
goto err4;
/*
// cut the path from the main thread name
threadName = strrchr(name, '/');
if (threadName != NULL)
threadName++;
else
threadName = name;
*/
/*
// create a kernel thread, but under the context of the new team
tid = spawn_kernel_thread_etc(team_create_team2, threadName, B_NORMAL_PRIORITY, teamArgs, team->id);
if (tid < 0) {
err = tid;
goto err4;
}
resume_thread(tid);
*/
return pid;
err4:
vm_put_aspace(team->aspace);
err3:
vfs_free_io_context(team->io_context);
err2:
//free_team_arg(teamArgs);
// remove the team structure from the team hash table and delete the team structure
state = disable_interrupts();
GRAB_TEAM_LOCK();
remove_team_from_parent(forkedTeam, team);
hash_remove(team_hash, team);
RELEASE_TEAM_LOCK();
restore_interrupts(state);
delete_team_struct(team);
return status;
}
/** This is the kernel backend for waitpid(). It is a bit more powerful when it comes /** 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. * to the reason why a thread has died than waitpid() can be.
*/ */
@ -1092,8 +1231,7 @@ _user_exec(const char *userPath, int32 argCount, char * const *userArgs,
thread_id thread_id
_user_fork(void) _user_fork(void)
{ {
dprintf("fork() is not yet implemented!\n"); return fork_team();
return B_ERROR;
} }