bpf: Add uptr support in the map_value of the task local storage.

This patch adds uptr support in the map_value of the task local storage.

struct map_value {
struct user_data __uptr *uptr;
};

struct {
__uint(type, BPF_MAP_TYPE_TASK_STORAGE);
__uint(map_flags, BPF_F_NO_PREALLOC);
__type(key, int);
__type(value, struct value_type);
} datamap SEC(".maps");

A new bpf_obj_pin_uptrs() is added to pin the user page and
also stores the kernel address back to the uptr for the
bpf prog to use later. It currently does not support
the uptr pointing to a user struct across two pages.
It also excludes PageHighMem support to keep it simple.
As of now, the 32bit bpf jit is missing other more crucial bpf
features. For example, many important bpf features depend on
bpf kfunc now but so far only one arch (x86-32) supports it
which was added by me as an example when kfunc was first
introduced to bpf.

The uptr can only be stored to the task local storage by the
syscall update_elem. Meaning the uptr will not be considered
if it is provided by the bpf prog through
bpf_task_storage_get(BPF_LOCAL_STORAGE_GET_F_CREATE).
This is enforced by only calling
bpf_local_storage_update(swap_uptrs==true) in
bpf_pid_task_storage_update_elem. Everywhere else will
have swap_uptrs==false.

This will pump down to bpf_selem_alloc(swap_uptrs==true). It is
the only case that bpf_selem_alloc() will take the uptr value when
updating the newly allocated selem. bpf_obj_swap_uptrs() is added
to swap the uptr between the SDATA(selem)->data and the user provided
map_value in "void *value". bpf_obj_swap_uptrs() makes the
SDATA(selem)->data takes the ownership of the uptr and the user space
provided map_value will have NULL in the uptr.

The bpf_obj_unpin_uptrs() is called after map->ops->map_update_elem()
returning error. If the map->ops->map_update_elem has reached
a state that the local storage has taken the uptr ownership,
the bpf_obj_unpin_uptrs() will be a no op because the uptr
is NULL. A "__"bpf_obj_unpin_uptrs is added to make this
error path unpin easier such that it does not have to check
the map->record is NULL or not.

BPF_F_LOCK is not supported when the map_value has uptr.
This can be revisited later if there is a use case. A similar
swap_uptrs idea can be considered.

The final bit is to do unpin_user_page in the bpf_obj_free_fields().
The earlier patch has ensured that the bpf_obj_free_fields() has
gone through the rcu gp when needed.

Cc: [e-mail]
Cc: Shakeel Butt <[e-mail]>
Signed-off-by: Martin KaFai Lau <[e-mail]>
Acked-by: Shakeel Butt <[e-mail]>
Link: https://lore.kernel.org/r/[e-mail]
Signed-off-by: Alexei Starovoitov <[e-mail]>

include/linux/bpf.h

···
 	case BPF_KPTR_UNREF:
 	case BPF_KPTR_REF:
 	case BPF_KPTR_PERCPU:
+	case BPF_UPTR:
 		break;
 	default:
 		WARN_ON_ONCE(1);
···
 static inline void copy_map_value_long(struct bpf_map *map, void *dst, void *src)
 {
 	bpf_obj_memcpy(map->record, dst, src, map->value_size, true);
+}
+
+static inline void bpf_obj_swap_uptrs(const struct btf_record *rec, void *dst, void *src)
+{
+	unsigned long *src_uptr, *dst_uptr;
+	const struct btf_field *field;
+	int i;
+
+	if (!btf_record_has_field(rec, BPF_UPTR))
+		return;
+
+	for (i = 0, field = rec->fields; i < rec->cnt; i++, field++) {
+		if (field->type != BPF_UPTR)
+			continue;
+
+		src_uptr = src + field->offset;
+		dst_uptr = dst + field->offset;
+		swap(*src_uptr, *dst_uptr);
+	}
 }
 
 static inline void bpf_obj_memzero(struct btf_record *rec, void *dst, u32 size)

kernel/bpf/bpf_local_storage.c

···
 	}
 
 	if (selem) {
-		if (value)
+		if (value) {
+			/* No need to call check_and_init_map_value as memory is zero init */
 			copy_map_value(&smap->map, SDATA(selem)->data, value);
-		/* No need to call check_and_init_map_value as memory is zero init */
+			if (swap_uptrs)
+				bpf_obj_swap_uptrs(smap->map.record, SDATA(selem)->data, value);
+		}
 		return selem;
 	}
 

kernel/bpf/bpf_task_storage.c

···
 	struct pid *pid;
 	int fd, err;
 
+	if ((map_flags & BPF_F_LOCK) && btf_record_has_field(map->record, BPF_UPTR))
+		return -EOPNOTSUPP;
+
 	fd = *(int *)key;
 	pid = pidfd_get_pid(fd, &f_flags);
 	if (IS_ERR(pid))
···
 	bpf_task_storage_lock();
 	sdata = bpf_local_storage_update(
 		task, (struct bpf_local_storage_map *)map, value, map_flags,
-		false, GFP_ATOMIC);
+		true, GFP_ATOMIC);
 	bpf_task_storage_unlock();
 
 	err = PTR_ERR_OR_ZERO(sdata);

kernel/bpf/syscall.c

···
 		synchronize_rcu();
 }
 
+static void unpin_uptr_kaddr(void *kaddr)
+{
+	if (kaddr)
+		unpin_user_page(virt_to_page(kaddr));
+}
+
+static void __bpf_obj_unpin_uptrs(struct btf_record *rec, u32 cnt, void *obj)
+{
+	const struct btf_field *field;
+	void **uptr_addr;
+	int i;
+
+	for (i = 0, field = rec->fields; i < cnt; i++, field++) {
+		if (field->type != BPF_UPTR)
+			continue;
+
+		uptr_addr = obj + field->offset;
+		unpin_uptr_kaddr(*uptr_addr);
+	}
+}
+
+static void bpf_obj_unpin_uptrs(struct btf_record *rec, void *obj)
+{
+	if (!btf_record_has_field(rec, BPF_UPTR))
+		return;
+
+	__bpf_obj_unpin_uptrs(rec, rec->cnt, obj);
+}
+
+static int bpf_obj_pin_uptrs(struct btf_record *rec, void *obj)
+{
+	const struct btf_field *field;
+	const struct btf_type *t;
+	unsigned long start, end;
+	struct page *page;
+	void **uptr_addr;
+	int i, err;
+
+	if (!btf_record_has_field(rec, BPF_UPTR))
+		return 0;
+
+	for (i = 0, field = rec->fields; i < rec->cnt; i++, field++) {
+		if (field->type != BPF_UPTR)
+			continue;
+
+		uptr_addr = obj + field->offset;
+		start = *(unsigned long *)uptr_addr;
+		if (!start)
+			continue;
+
+		t = btf_type_by_id(field->kptr.btf, field->kptr.btf_id);
+		/* t->size was checked for zero before */
+		if (check_add_overflow(start, t->size - 1, &end)) {
+			err = -EFAULT;
+			goto unpin_all;
+		}
+
+		/* The uptr's struct cannot span across two pages */
+		if ((start & PAGE_MASK) != (end & PAGE_MASK)) {
+			err = -EOPNOTSUPP;
+			goto unpin_all;
+		}
+
+		err = pin_user_pages_fast(start, 1, FOLL_LONGTERM | FOLL_WRITE, &page);
+		if (err != 1)
+			goto unpin_all;
+
+		if (PageHighMem(page)) {
+			err = -EOPNOTSUPP;
+			unpin_user_page(page);
+			goto unpin_all;
+		}
+
+		*uptr_addr = page_address(page) + offset_in_page(start);
+	}
+
+	return 0;
+
+unpin_all:
+	__bpf_obj_unpin_uptrs(rec, i, obj);
+	return err;
+}
+
 static int bpf_map_update_value(struct bpf_map *map, struct file *map_file,
 				void *key, void *value, __u64 flags)
 {
···
 		   map->map_type == BPF_MAP_TYPE_BLOOM_FILTER) {
 		err = map->ops->map_push_elem(map, value, flags);
 	} else {
-		rcu_read_lock();
-		err = map->ops->map_update_elem(map, key, value, flags);
-		rcu_read_unlock();
+		err = bpf_obj_pin_uptrs(map->record, value);
+		if (!err) {
+			rcu_read_lock();
+			err = map->ops->map_update_elem(map, key, value, flags);
+			rcu_read_unlock();
+			if (err)
+				bpf_obj_unpin_uptrs(map->record, value);
+		}
 	}
 	bpf_enable_instrumentation();
 
···
 				field->kptr.dtor(xchgd_field);
 			}
 			break;
+		case BPF_UPTR:
+			/* The caller ensured that no one is using the uptr */
+			unpin_uptr_kaddr(*(void **)field_ptr);
+			break;
 		case BPF_LIST_HEAD:
 			if (WARN_ON_ONCE(rec->spin_lock_off < 0))
 				continue;
···
 
 	map->record = btf_parse_fields(btf, value_type,
 				       BPF_SPIN_LOCK | BPF_TIMER | BPF_KPTR | BPF_LIST_HEAD |
-				       BPF_RB_ROOT | BPF_REFCOUNT | BPF_WORKQUEUE,
+				       BPF_RB_ROOT | BPF_REFCOUNT | BPF_WORKQUEUE | BPF_UPTR,
 				       map->value_size);
 	if (!IS_ERR_OR_NULL(map->record)) {
 		int i;
···
 				    map->map_type != BPF_MAP_TYPE_INODE_STORAGE &&
 				    map->map_type != BPF_MAP_TYPE_TASK_STORAGE &&
 				    map->map_type != BPF_MAP_TYPE_CGRP_STORAGE) {
+					ret = -EOPNOTSUPP;
+					goto free_map_tab;
+				}
+				break;
+			case BPF_UPTR:
+				if (map->map_type != BPF_MAP_TYPE_TASK_STORAGE) {
 					ret = -EOPNOTSUPP;
 					goto free_map_tab;
 				}