kernel/team: Do not inherit anything from the kernel's IO context.

We already didn't inherit FDs, which meant that the only thing we
did meaningfully inherit was the table size. That meant that basically
no applications actually had a table size of the default 256, but all
were at the kernel's 4096 (except Tracker and anything started by it,
as Tracker resets it to 512), and also that basically all applications
had FD tables allocated with the raw allocator instead of the block
allocator, which isn't very efficient.

Since this reduces the default FD table size, some applications
might encounter problems. However, build systems and other such
tools should already increase this by default as needed, and it's
easy enough to patch in calls to setrlimit() if it turns out
some applications needed a higher default after all.

Also remove a redundant call to vfs_exec_io_context. Calling
vfs_new_io_context with the second argument set to "true"
already skips cloning CLOEXEC FDs.

headers/private/kernel/vfs.h

@@ -59,7 +59,6 @@ typedef struct io_context {
 	struct list node_monitors;
 	uint32		num_monitors;
 	uint32		max_monitors;
-	bool		inherit_fds;
 } io_context;
 
 

src/system/kernel/fs/vfs.cpp

@@ -4968,35 +4968,34 @@ vfs_new_io_context(const io_context* parentContext, bool purgeCloseOnExec)
 	if (parentContext != NULL) {
 		mutex_lock(&sIOContextRootLock);
 		context->root = parentContext->root;
-		if (context->root)
+		if (context->root != NULL)
 			inc_vnode_ref_count(context->root);
 		mutex_unlock(&sIOContextRootLock);
 
 		context->cwd = parentContext->cwd;
-		if (context->cwd)
+		if (context->cwd != NULL)
 			inc_vnode_ref_count(context->cwd);
 
-		if (parentContext->inherit_fds) {
-			for (size_t i = 0; i < tableSize; i++) {
-				struct file_descriptor* descriptor = parentContext->fds[i];
-
-				if (descriptor != NULL
-						&& (descriptor->open_mode & O_DISCONNECTED) == 0) {
-					const bool closeOnExec = fd_close_on_exec(parentContext, i);
-					if (closeOnExec && purgeCloseOnExec)
-						continue;
-
-					TFD(InheritFD(context, i, descriptor, parentContext));
-
-					context->fds[i] = descriptor;
-					context->num_used_fds++;
-					atomic_add(&descriptor->ref_count, 1);
-					atomic_add(&descriptor->open_count, 1);
-
-					if (closeOnExec)
-						fd_set_close_on_exec(context, i, true);
-				}
+		for (size_t i = 0; i < tableSize; i++) {
+			struct file_descriptor* descriptor = parentContext->fds[i];
+			if (descriptor == NULL
+					|| (descriptor->open_mode & O_DISCONNECTED) != 0) {
+				continue;
 			}
+
+			const bool closeOnExec = fd_close_on_exec(parentContext, i);
+			if (closeOnExec && purgeCloseOnExec)
+				continue;
+
+			TFD(InheritFD(context, i, descriptor, parentContext));
+
+			context->fds[i] = descriptor;
+			context->num_used_fds++;
+			atomic_add(&descriptor->ref_count, 1);
+			atomic_add(&descriptor->open_count, 1);
+
+			if (closeOnExec)
+				fd_set_close_on_exec(context, i, true);
 		}
 
 		parentLocker.Unlock();
@@ -5012,7 +5011,6 @@ vfs_new_io_context(const io_context* parentContext, bool purgeCloseOnExec)
 	}
 
 	context->table_size = tableSize;
-	context->inherit_fds = parentContext != NULL;
 
 	list_init(&context->node_monitors);
 	context->max_monitors = DEFAULT_NODE_MONITORS;

src/system/kernel/team.cpp

@@ -1788,8 +1788,9 @@ load_image_internal(char**& _flatArgs, size_t flatArgsSize, int32 argCount,
 	inherit_parent_user_and_group(team, parent);
 
 	// get a reference to the parent's I/O context -- we need it to create ours
-	parentIOContext = parent->io_context;
-	vfs_get_io_context(parentIOContext);
+	parentIOContext = (parent->id == B_SYSTEM_TEAM) ? NULL : parent->io_context;
+	if (parentIOContext != NULL)
+		vfs_get_io_context(parentIOContext);
 
 	team->Unlock();
 	parent->UnlockTeamAndProcessGroup();
@@ -1808,18 +1809,18 @@ load_image_internal(char**& _flatArgs, size_t flatArgsSize, int32 argCount,
 	team->SetArgs(path, teamArgs->flat_args + 1, argCount - 1);
 
 	// create a new io_context for this team
+	// remove any fds that have the CLOEXEC flag set (emulating BeOS behaviour)
 	team->io_context = vfs_new_io_context(parentIOContext, true);
-	if (!team->io_context) {
+	if (team->io_context == NULL) {
 		status = B_NO_MEMORY;
 		goto err2;
 	}
 
-	// We don't need the parent's I/O context any longer.
-	vfs_put_io_context(parentIOContext);
-	parentIOContext = NULL;
-
-	// remove any fds that have the CLOEXEC flag set (emulating BeOS behaviour)
-	vfs_exec_io_context(team->io_context);
+	if (parentIOContext != NULL) {
+		// We don't need the parent's I/O context any longer.
+		vfs_put_io_context(parentIOContext);
+		parentIOContext = NULL;
+	}
 
 	// create an address space for this team
 	status = VMAddressSpace::Create(team->id, USER_BASE, USER_SIZE, false,