2017-03-17 14:18:50 +08:00
|
|
|
/* SPDX-License-Identifier: GPL-2.0 */
|
|
|
|
#ifndef _BCACHEFS_EXTENTS_H
|
|
|
|
#define _BCACHEFS_EXTENTS_H
|
|
|
|
|
|
|
|
#include "bcachefs.h"
|
|
|
|
#include "bkey.h"
|
|
|
|
#include "extents_types.h"
|
|
|
|
|
|
|
|
struct bch_fs;
|
2019-03-14 10:44:04 +08:00
|
|
|
struct btree_trans;
|
2024-05-09 06:40:42 +08:00
|
|
|
enum bch_validate_flags;
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
/* extent entries: */
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-07-26 01:52:14 +08:00
|
|
|
#define extent_entry_last(_e) \
|
|
|
|
((typeof(&(_e).v->start[0])) bkey_val_end(_e))
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define entry_to_ptr(_entry) \
|
|
|
|
({ \
|
|
|
|
EBUG_ON((_entry) && !extent_entry_is_ptr(_entry)); \
|
|
|
|
\
|
|
|
|
__builtin_choose_expr( \
|
|
|
|
type_is_exact(_entry, const union bch_extent_entry *), \
|
|
|
|
(const struct bch_extent_ptr *) (_entry), \
|
|
|
|
(struct bch_extent_ptr *) (_entry)); \
|
|
|
|
})
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
/* downcast, preserves const */
|
|
|
|
#define to_entry(_entry) \
|
|
|
|
({ \
|
|
|
|
BUILD_BUG_ON(!type_is(_entry, union bch_extent_crc *) && \
|
|
|
|
!type_is(_entry, struct bch_extent_ptr *) && \
|
|
|
|
!type_is(_entry, struct bch_extent_stripe_ptr *)); \
|
|
|
|
\
|
|
|
|
__builtin_choose_expr( \
|
|
|
|
(type_is_exact(_entry, const union bch_extent_crc *) || \
|
|
|
|
type_is_exact(_entry, const struct bch_extent_ptr *) ||\
|
|
|
|
type_is_exact(_entry, const struct bch_extent_stripe_ptr *)),\
|
|
|
|
(const union bch_extent_entry *) (_entry), \
|
|
|
|
(union bch_extent_entry *) (_entry)); \
|
|
|
|
})
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
#define extent_entry_next(_entry) \
|
|
|
|
((typeof(_entry)) ((void *) (_entry) + extent_entry_bytes(_entry)))
|
|
|
|
|
2024-03-09 04:25:27 +08:00
|
|
|
#define extent_entry_next_safe(_entry, _end) \
|
|
|
|
(likely(__extent_entry_type(_entry) < BCH_EXTENT_ENTRY_MAX) \
|
|
|
|
? extent_entry_next(_entry) \
|
|
|
|
: _end)
|
|
|
|
|
2017-03-17 14:18:50 +08:00
|
|
|
static inline unsigned
|
|
|
|
__extent_entry_type(const union bch_extent_entry *e)
|
|
|
|
{
|
|
|
|
return e->type ? __ffs(e->type) : BCH_EXTENT_ENTRY_MAX;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline enum bch_extent_entry_type
|
|
|
|
extent_entry_type(const union bch_extent_entry *e)
|
|
|
|
{
|
|
|
|
int ret = __ffs(e->type);
|
|
|
|
|
|
|
|
EBUG_ON(ret < 0 || ret >= BCH_EXTENT_ENTRY_MAX);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline size_t extent_entry_bytes(const union bch_extent_entry *entry)
|
|
|
|
{
|
|
|
|
switch (extent_entry_type(entry)) {
|
2018-10-01 06:39:20 +08:00
|
|
|
#define x(f, n) \
|
|
|
|
case BCH_EXTENT_ENTRY_##f: \
|
|
|
|
return sizeof(struct bch_extent_##f);
|
|
|
|
BCH_EXTENT_ENTRY_TYPES()
|
|
|
|
#undef x
|
2017-03-17 14:18:50 +08:00
|
|
|
default:
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline size_t extent_entry_u64s(const union bch_extent_entry *entry)
|
|
|
|
{
|
|
|
|
return extent_entry_bytes(entry) / sizeof(u64);
|
|
|
|
}
|
|
|
|
|
2023-02-23 06:57:59 +08:00
|
|
|
static inline void __extent_entry_insert(struct bkey_i *k,
|
|
|
|
union bch_extent_entry *dst,
|
|
|
|
union bch_extent_entry *new)
|
|
|
|
{
|
|
|
|
union bch_extent_entry *end = bkey_val_end(bkey_i_to_s(k));
|
|
|
|
|
|
|
|
memmove_u64s_up_small((u64 *) dst + extent_entry_u64s(new),
|
|
|
|
dst, (u64 *) end - (u64 *) dst);
|
|
|
|
k->k.u64s += extent_entry_u64s(new);
|
|
|
|
memcpy_u64s_small(dst, new, extent_entry_u64s(new));
|
|
|
|
}
|
|
|
|
|
bcachefs: rebalance_work
This adds a new btree, rebalance_work, to eliminate scanning required
for finding extents that need work done on them in the background - i.e.
for the background_target and background_compression options.
rebalance_work is a bitset btree, where a KEY_TYPE_set corresponds to an
extent in the extents or reflink btree at the same pos.
A new extent field is added, bch_extent_rebalance, which indicates that
this extent has work that needs to be done in the background - and which
options to use. This allows per-inode options to be propagated to
indirect extents - at least in some circumstances. In this patch,
changing IO options on a file will not propagate the new options to
indirect extents pointed to by that file.
Updating (setting/clearing) the rebalance_work btree is done by the
extent trigger, which looks at the bch_extent_rebalance field.
Scanning is still requrired after changing IO path options - either just
for a given inode, or for the whole filesystem. We indicate that
scanning is required by adding a KEY_TYPE_cookie key to the
rebalance_work btree: the cookie counter is so that we can detect that
scanning is still required when an option has been flipped mid-way
through an existing scan.
Future possible work:
- Propagate options to indirect extents when being changed
- Add other IO path options - nr_replicas, ec, to rebalance_work so
they can be applied in the background when they change
- Add a counter, for bcachefs fs usage output, showing the pending
amount of rebalance work: we'll probably want to do this after the
disk space accounting rewrite (moving it to a new btree)
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
2023-10-21 01:33:14 +08:00
|
|
|
static inline void extent_entry_drop(struct bkey_s k, union bch_extent_entry *entry)
|
|
|
|
{
|
|
|
|
union bch_extent_entry *next = extent_entry_next(entry);
|
|
|
|
|
|
|
|
/* stripes have ptrs, but their layout doesn't work with this code */
|
|
|
|
BUG_ON(k.k->type == KEY_TYPE_stripe);
|
|
|
|
|
|
|
|
memmove_u64s_down(entry, next,
|
|
|
|
(u64 *) bkey_val_end(k) - (u64 *) next);
|
|
|
|
k.k->u64s -= (u64 *) next - (u64 *) entry;
|
|
|
|
}
|
|
|
|
|
2017-03-17 14:18:50 +08:00
|
|
|
static inline bool extent_entry_is_ptr(const union bch_extent_entry *e)
|
|
|
|
{
|
2024-03-17 09:22:24 +08:00
|
|
|
return __extent_entry_type(e) == BCH_EXTENT_ENTRY_ptr;
|
2021-10-14 01:12:26 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline bool extent_entry_is_stripe_ptr(const union bch_extent_entry *e)
|
|
|
|
{
|
2024-03-17 09:22:24 +08:00
|
|
|
return __extent_entry_type(e) == BCH_EXTENT_ENTRY_stripe_ptr;
|
2017-03-17 14:18:50 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline bool extent_entry_is_crc(const union bch_extent_entry *e)
|
|
|
|
{
|
2024-03-17 09:22:24 +08:00
|
|
|
switch (__extent_entry_type(e)) {
|
2018-09-28 09:08:39 +08:00
|
|
|
case BCH_EXTENT_ENTRY_crc32:
|
|
|
|
case BCH_EXTENT_ENTRY_crc64:
|
|
|
|
case BCH_EXTENT_ENTRY_crc128:
|
|
|
|
return true;
|
|
|
|
default:
|
|
|
|
return false;
|
|
|
|
}
|
2017-03-17 14:18:50 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
union bch_extent_crc {
|
|
|
|
u8 type;
|
|
|
|
struct bch_extent_crc32 crc32;
|
|
|
|
struct bch_extent_crc64 crc64;
|
|
|
|
struct bch_extent_crc128 crc128;
|
|
|
|
};
|
|
|
|
|
|
|
|
#define __entry_to_crc(_entry) \
|
|
|
|
__builtin_choose_expr( \
|
|
|
|
type_is_exact(_entry, const union bch_extent_entry *), \
|
|
|
|
(const union bch_extent_crc *) (_entry), \
|
|
|
|
(union bch_extent_crc *) (_entry))
|
|
|
|
|
|
|
|
#define entry_to_crc(_entry) \
|
|
|
|
({ \
|
|
|
|
EBUG_ON((_entry) && !extent_entry_is_crc(_entry)); \
|
|
|
|
\
|
|
|
|
__entry_to_crc(_entry); \
|
|
|
|
})
|
|
|
|
|
|
|
|
static inline struct bch_extent_crc_unpacked
|
|
|
|
bch2_extent_crc_unpack(const struct bkey *k, const union bch_extent_crc *crc)
|
|
|
|
{
|
|
|
|
#define common_fields(_crc) \
|
|
|
|
.csum_type = _crc.csum_type, \
|
|
|
|
.compression_type = _crc.compression_type, \
|
|
|
|
.compressed_size = _crc._compressed_size + 1, \
|
|
|
|
.uncompressed_size = _crc._uncompressed_size + 1, \
|
|
|
|
.offset = _crc.offset, \
|
|
|
|
.live_size = k->size
|
|
|
|
|
2018-10-03 04:40:12 +08:00
|
|
|
if (!crc)
|
2017-03-17 14:18:50 +08:00
|
|
|
return (struct bch_extent_crc_unpacked) {
|
|
|
|
.compressed_size = k->size,
|
|
|
|
.uncompressed_size = k->size,
|
|
|
|
.live_size = k->size,
|
|
|
|
};
|
2018-10-03 04:40:12 +08:00
|
|
|
|
|
|
|
switch (extent_entry_type(to_entry(crc))) {
|
|
|
|
case BCH_EXTENT_ENTRY_crc32: {
|
2017-03-17 14:18:50 +08:00
|
|
|
struct bch_extent_crc_unpacked ret = (struct bch_extent_crc_unpacked) {
|
|
|
|
common_fields(crc->crc32),
|
|
|
|
};
|
|
|
|
|
2023-07-07 10:47:42 +08:00
|
|
|
*((__le32 *) &ret.csum.lo) = (__le32 __force) crc->crc32.csum;
|
2017-03-17 14:18:50 +08:00
|
|
|
return ret;
|
|
|
|
}
|
2018-10-03 04:40:12 +08:00
|
|
|
case BCH_EXTENT_ENTRY_crc64: {
|
2017-03-17 14:18:50 +08:00
|
|
|
struct bch_extent_crc_unpacked ret = (struct bch_extent_crc_unpacked) {
|
|
|
|
common_fields(crc->crc64),
|
|
|
|
.nonce = crc->crc64.nonce,
|
|
|
|
.csum.lo = (__force __le64) crc->crc64.csum_lo,
|
|
|
|
};
|
|
|
|
|
2023-07-07 10:47:42 +08:00
|
|
|
*((__le16 *) &ret.csum.hi) = (__le16 __force) crc->crc64.csum_hi;
|
2017-03-17 14:18:50 +08:00
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
2018-10-03 04:40:12 +08:00
|
|
|
case BCH_EXTENT_ENTRY_crc128: {
|
2017-03-17 14:18:50 +08:00
|
|
|
struct bch_extent_crc_unpacked ret = (struct bch_extent_crc_unpacked) {
|
|
|
|
common_fields(crc->crc128),
|
|
|
|
.nonce = crc->crc128.nonce,
|
|
|
|
.csum = crc->crc128.csum,
|
|
|
|
};
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
default:
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
#undef common_fields
|
|
|
|
}
|
|
|
|
|
2018-02-24 05:26:10 +08:00
|
|
|
static inline bool crc_is_compressed(struct bch_extent_crc_unpacked crc)
|
|
|
|
{
|
|
|
|
return (crc.compression_type != BCH_COMPRESSION_TYPE_none &&
|
|
|
|
crc.compression_type != BCH_COMPRESSION_TYPE_incompressible);
|
|
|
|
}
|
|
|
|
|
2023-10-22 23:33:02 +08:00
|
|
|
static inline bool crc_is_encoded(struct bch_extent_crc_unpacked crc)
|
|
|
|
{
|
|
|
|
return crc.csum_type != BCH_CSUM_none || crc_is_compressed(crc);
|
|
|
|
}
|
|
|
|
|
2024-06-29 04:25:39 +08:00
|
|
|
void bch2_extent_crc_unpacked_to_text(struct printbuf *, struct bch_extent_crc_unpacked *);
|
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
/* bkey_ptrs: generically over any key type that has ptrs */
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
struct bkey_ptrs_c {
|
|
|
|
const union bch_extent_entry *start;
|
|
|
|
const union bch_extent_entry *end;
|
|
|
|
};
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
struct bkey_ptrs {
|
|
|
|
union bch_extent_entry *start;
|
|
|
|
union bch_extent_entry *end;
|
|
|
|
};
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
static inline struct bkey_ptrs_c bch2_bkey_ptrs_c(struct bkey_s_c k)
|
|
|
|
{
|
|
|
|
switch (k.k->type) {
|
|
|
|
case KEY_TYPE_btree_ptr: {
|
|
|
|
struct bkey_s_c_btree_ptr e = bkey_s_c_to_btree_ptr(k);
|
2022-10-23 03:59:53 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
return (struct bkey_ptrs_c) {
|
|
|
|
to_entry(&e.v->start[0]),
|
|
|
|
to_entry(extent_entry_last(e))
|
|
|
|
};
|
|
|
|
}
|
|
|
|
case KEY_TYPE_extent: {
|
|
|
|
struct bkey_s_c_extent e = bkey_s_c_to_extent(k);
|
2022-10-23 03:59:53 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
return (struct bkey_ptrs_c) {
|
|
|
|
e.v->start,
|
|
|
|
extent_entry_last(e)
|
|
|
|
};
|
|
|
|
}
|
|
|
|
case KEY_TYPE_stripe: {
|
|
|
|
struct bkey_s_c_stripe s = bkey_s_c_to_stripe(k);
|
2022-10-23 03:59:53 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
return (struct bkey_ptrs_c) {
|
|
|
|
to_entry(&s.v->ptrs[0]),
|
|
|
|
to_entry(&s.v->ptrs[s.v->nr_blocks]),
|
|
|
|
};
|
|
|
|
}
|
|
|
|
case KEY_TYPE_reflink_v: {
|
|
|
|
struct bkey_s_c_reflink_v r = bkey_s_c_to_reflink_v(k);
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
return (struct bkey_ptrs_c) {
|
|
|
|
r.v->start,
|
|
|
|
bkey_val_end(r),
|
|
|
|
};
|
|
|
|
}
|
2020-02-08 02:38:02 +08:00
|
|
|
case KEY_TYPE_btree_ptr_v2: {
|
|
|
|
struct bkey_s_c_btree_ptr_v2 e = bkey_s_c_to_btree_ptr_v2(k);
|
2022-10-23 03:59:53 +08:00
|
|
|
|
2020-02-08 02:38:02 +08:00
|
|
|
return (struct bkey_ptrs_c) {
|
|
|
|
to_entry(&e.v->start[0]),
|
|
|
|
to_entry(extent_entry_last(e))
|
|
|
|
};
|
|
|
|
}
|
2019-11-17 05:25:58 +08:00
|
|
|
default:
|
|
|
|
return (struct bkey_ptrs_c) { NULL, NULL };
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct bkey_ptrs bch2_bkey_ptrs(struct bkey_s k)
|
|
|
|
{
|
|
|
|
struct bkey_ptrs_c p = bch2_bkey_ptrs_c(k.s_c);
|
|
|
|
|
|
|
|
return (struct bkey_ptrs) {
|
|
|
|
(void *) p.start,
|
|
|
|
(void *) p.end
|
|
|
|
};
|
|
|
|
}
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define __bkey_extent_entry_for_each_from(_start, _end, _entry) \
|
|
|
|
for ((_entry) = (_start); \
|
|
|
|
(_entry) < (_end); \
|
2024-03-09 04:25:27 +08:00
|
|
|
(_entry) = extent_entry_next_safe(_entry, _end))
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define __bkey_ptr_next(_ptr, _end) \
|
2017-03-17 14:18:50 +08:00
|
|
|
({ \
|
2018-11-02 03:10:01 +08:00
|
|
|
typeof(_end) _entry; \
|
2017-03-17 14:18:50 +08:00
|
|
|
\
|
2018-11-02 03:10:01 +08:00
|
|
|
__bkey_extent_entry_for_each_from(to_entry(_ptr), _end, _entry) \
|
2018-09-28 09:08:39 +08:00
|
|
|
if (extent_entry_is_ptr(_entry)) \
|
|
|
|
break; \
|
2017-03-17 14:18:50 +08:00
|
|
|
\
|
2018-11-02 03:10:01 +08:00
|
|
|
_entry < (_end) ? entry_to_ptr(_entry) : NULL; \
|
2017-03-17 14:18:50 +08:00
|
|
|
})
|
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define bkey_extent_entry_for_each_from(_p, _entry, _start) \
|
|
|
|
__bkey_extent_entry_for_each_from(_start, (_p).end, _entry)
|
2018-09-28 09:08:39 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define bkey_extent_entry_for_each(_p, _entry) \
|
|
|
|
bkey_extent_entry_for_each_from(_p, _entry, _p.start)
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define __bkey_for_each_ptr(_start, _end, _ptr) \
|
2023-12-22 04:47:15 +08:00
|
|
|
for (typeof(_start) (_ptr) = (_start); \
|
2018-11-02 03:10:01 +08:00
|
|
|
((_ptr) = __bkey_ptr_next(_ptr, _end)); \
|
|
|
|
(_ptr)++)
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define bkey_ptr_next(_p, _ptr) \
|
|
|
|
__bkey_ptr_next(_ptr, (_p).end)
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define bkey_for_each_ptr(_p, _ptr) \
|
|
|
|
__bkey_for_each_ptr(&(_p).start->ptr, (_p).end, _ptr)
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define __bkey_ptr_next_decode(_k, _end, _ptr, _entry) \
|
2017-03-17 14:18:50 +08:00
|
|
|
({ \
|
|
|
|
__label__ out; \
|
|
|
|
\
|
2018-11-02 03:13:19 +08:00
|
|
|
(_ptr).idx = 0; \
|
2019-10-09 06:45:29 +08:00
|
|
|
(_ptr).has_ec = false; \
|
2018-11-02 03:13:19 +08:00
|
|
|
\
|
2018-11-02 03:10:01 +08:00
|
|
|
__bkey_extent_entry_for_each_from(_entry, _end, _entry) \
|
2024-03-09 04:25:27 +08:00
|
|
|
switch (__extent_entry_type(_entry)) { \
|
2018-09-28 09:08:39 +08:00
|
|
|
case BCH_EXTENT_ENTRY_ptr: \
|
|
|
|
(_ptr).ptr = _entry->ptr; \
|
2017-03-17 14:18:50 +08:00
|
|
|
goto out; \
|
2018-09-28 09:08:39 +08:00
|
|
|
case BCH_EXTENT_ENTRY_crc32: \
|
|
|
|
case BCH_EXTENT_ENTRY_crc64: \
|
|
|
|
case BCH_EXTENT_ENTRY_crc128: \
|
2018-11-02 03:10:01 +08:00
|
|
|
(_ptr).crc = bch2_extent_crc_unpack(_k, \
|
2018-09-28 09:08:39 +08:00
|
|
|
entry_to_crc(_entry)); \
|
|
|
|
break; \
|
2018-11-02 03:13:19 +08:00
|
|
|
case BCH_EXTENT_ENTRY_stripe_ptr: \
|
2019-10-09 06:45:29 +08:00
|
|
|
(_ptr).ec = _entry->stripe_ptr; \
|
|
|
|
(_ptr).has_ec = true; \
|
2018-11-02 03:13:19 +08:00
|
|
|
break; \
|
2023-06-28 07:02:17 +08:00
|
|
|
default: \
|
|
|
|
/* nothing */ \
|
|
|
|
break; \
|
2017-03-17 14:18:50 +08:00
|
|
|
} \
|
|
|
|
out: \
|
2018-11-02 03:10:01 +08:00
|
|
|
_entry < (_end); \
|
2017-03-17 14:18:50 +08:00
|
|
|
})
|
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define __bkey_for_each_ptr_decode(_k, _start, _end, _ptr, _entry) \
|
|
|
|
for ((_ptr).crc = bch2_extent_crc_unpack(_k, NULL), \
|
|
|
|
(_entry) = _start; \
|
|
|
|
__bkey_ptr_next_decode(_k, _end, _ptr, _entry); \
|
2024-03-09 04:25:27 +08:00
|
|
|
(_entry) = extent_entry_next_safe(_entry, _end))
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define bkey_for_each_ptr_decode(_k, _p, _ptr, _entry) \
|
|
|
|
__bkey_for_each_ptr_decode(_k, (_p).start, (_p).end, \
|
|
|
|
_ptr, _entry)
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-07-26 01:52:14 +08:00
|
|
|
#define bkey_crc_next(_k, _start, _end, _crc, _iter) \
|
|
|
|
({ \
|
|
|
|
__bkey_extent_entry_for_each_from(_iter, _end, _iter) \
|
|
|
|
if (extent_entry_is_crc(_iter)) { \
|
|
|
|
(_crc) = bch2_extent_crc_unpack(_k, \
|
|
|
|
entry_to_crc(_iter)); \
|
|
|
|
break; \
|
|
|
|
} \
|
|
|
|
\
|
|
|
|
(_iter) < (_end); \
|
|
|
|
})
|
|
|
|
|
|
|
|
#define __bkey_for_each_crc(_k, _start, _end, _crc, _iter) \
|
|
|
|
for ((_crc) = bch2_extent_crc_unpack(_k, NULL), \
|
|
|
|
(_iter) = (_start); \
|
|
|
|
bkey_crc_next(_k, _start, _end, _crc, _iter); \
|
|
|
|
(_iter) = extent_entry_next(_iter))
|
|
|
|
|
|
|
|
#define bkey_for_each_crc(_k, _p, _crc, _iter) \
|
|
|
|
__bkey_for_each_crc(_k, (_p).start, (_p).end, _crc, _iter)
|
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
/* Iterate over pointers in KEY_TYPE_extent: */
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
#define extent_for_each_entry_from(_e, _entry, _start) \
|
|
|
|
__bkey_extent_entry_for_each_from(_start, \
|
2022-10-23 03:59:53 +08:00
|
|
|
extent_entry_last(_e), _entry)
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
#define extent_for_each_entry(_e, _entry) \
|
|
|
|
extent_for_each_entry_from(_e, _entry, (_e).v->start)
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
#define extent_ptr_next(_e, _ptr) \
|
|
|
|
__bkey_ptr_next(_ptr, extent_entry_last(_e))
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
#define extent_for_each_ptr(_e, _ptr) \
|
|
|
|
__bkey_for_each_ptr(&(_e).v->start->ptr, extent_entry_last(_e), _ptr)
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
#define extent_for_each_ptr_decode(_e, _ptr, _entry) \
|
|
|
|
__bkey_for_each_ptr_decode((_e).k, (_e).v->start, \
|
|
|
|
extent_entry_last(_e), _ptr, _entry)
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
/* utility code common to all keys with pointers: */
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2024-06-29 01:28:30 +08:00
|
|
|
struct bch_dev_io_failures *bch2_dev_io_failures(struct bch_io_failures *,
|
|
|
|
unsigned);
|
2018-11-02 03:10:01 +08:00
|
|
|
void bch2_mark_io_failure(struct bch_io_failures *,
|
|
|
|
struct extent_ptr_decoded *);
|
|
|
|
int bch2_bkey_pick_read_device(struct bch_fs *, struct bkey_s_c,
|
|
|
|
struct bch_io_failures *,
|
|
|
|
struct extent_ptr_decoded *);
|
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
/* KEY_TYPE_btree_ptr: */
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2024-08-13 09:31:25 +08:00
|
|
|
int bch2_btree_ptr_validate(struct bch_fs *, struct bkey_s_c,
|
|
|
|
enum bch_validate_flags);
|
2018-11-02 03:10:01 +08:00
|
|
|
void bch2_btree_ptr_to_text(struct printbuf *, struct bch_fs *,
|
|
|
|
struct bkey_s_c);
|
2020-04-01 04:25:30 +08:00
|
|
|
|
2024-08-13 09:31:25 +08:00
|
|
|
int bch2_btree_ptr_v2_validate(struct bch_fs *, struct bkey_s_c,
|
|
|
|
enum bch_validate_flags);
|
2022-04-04 05:50:01 +08:00
|
|
|
void bch2_btree_ptr_v2_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
|
2020-01-08 02:29:32 +08:00
|
|
|
void bch2_btree_ptr_v2_compat(enum btree_id, unsigned, unsigned,
|
|
|
|
int, struct bkey_s);
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2022-10-23 03:59:53 +08:00
|
|
|
#define bch2_bkey_ops_btree_ptr ((struct bkey_ops) { \
|
2024-08-13 09:31:25 +08:00
|
|
|
.key_validate = bch2_btree_ptr_validate, \
|
2018-11-02 03:10:01 +08:00
|
|
|
.val_to_text = bch2_btree_ptr_to_text, \
|
|
|
|
.swab = bch2_ptr_swab, \
|
2024-01-01 10:01:06 +08:00
|
|
|
.trigger = bch2_trigger_extent, \
|
2022-10-23 03:59:53 +08:00
|
|
|
})
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2022-10-23 03:59:53 +08:00
|
|
|
#define bch2_bkey_ops_btree_ptr_v2 ((struct bkey_ops) { \
|
2024-08-13 09:31:25 +08:00
|
|
|
.key_validate = bch2_btree_ptr_v2_validate, \
|
2020-04-01 04:25:30 +08:00
|
|
|
.val_to_text = bch2_btree_ptr_v2_to_text, \
|
2020-02-08 02:38:02 +08:00
|
|
|
.swab = bch2_ptr_swab, \
|
2020-01-08 02:29:32 +08:00
|
|
|
.compat = bch2_btree_ptr_v2_compat, \
|
2024-01-01 10:01:06 +08:00
|
|
|
.trigger = bch2_trigger_extent, \
|
2023-04-30 01:24:18 +08:00
|
|
|
.min_val_size = 40, \
|
2022-10-23 03:59:53 +08:00
|
|
|
})
|
2020-02-08 02:38:02 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
/* KEY_TYPE_extent: */
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2021-04-29 11:49:30 +08:00
|
|
|
bool bch2_extent_merge(struct bch_fs *, struct bkey_s, struct bkey_s_c);
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2022-10-23 03:59:53 +08:00
|
|
|
#define bch2_bkey_ops_extent ((struct bkey_ops) { \
|
2024-08-13 09:31:25 +08:00
|
|
|
.key_validate = bch2_bkey_ptrs_validate, \
|
2022-04-04 05:50:01 +08:00
|
|
|
.val_to_text = bch2_bkey_ptrs_to_text, \
|
2018-11-02 03:10:01 +08:00
|
|
|
.swab = bch2_ptr_swab, \
|
|
|
|
.key_normalize = bch2_extent_normalize, \
|
|
|
|
.key_merge = bch2_extent_merge, \
|
2024-01-01 10:01:06 +08:00
|
|
|
.trigger = bch2_trigger_extent, \
|
2022-10-23 03:59:53 +08:00
|
|
|
})
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
/* KEY_TYPE_reservation: */
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2024-08-13 09:31:25 +08:00
|
|
|
int bch2_reservation_validate(struct bch_fs *, struct bkey_s_c,
|
|
|
|
enum bch_validate_flags);
|
2018-11-02 03:10:01 +08:00
|
|
|
void bch2_reservation_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c);
|
2021-04-29 11:49:30 +08:00
|
|
|
bool bch2_reservation_merge(struct bch_fs *, struct bkey_s, struct bkey_s_c);
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2022-10-23 03:59:53 +08:00
|
|
|
#define bch2_bkey_ops_reservation ((struct bkey_ops) { \
|
2024-08-13 09:31:25 +08:00
|
|
|
.key_validate = bch2_reservation_validate, \
|
2018-11-02 03:10:01 +08:00
|
|
|
.val_to_text = bch2_reservation_to_text, \
|
|
|
|
.key_merge = bch2_reservation_merge, \
|
2024-01-01 10:01:06 +08:00
|
|
|
.trigger = bch2_trigger_reservation, \
|
2023-04-30 01:24:18 +08:00
|
|
|
.min_val_size = 8, \
|
2022-10-23 03:59:53 +08:00
|
|
|
})
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
/* Extent checksum entries: */
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
bool bch2_can_narrow_extent_crcs(struct bkey_s_c,
|
|
|
|
struct bch_extent_crc_unpacked);
|
|
|
|
bool bch2_bkey_narrow_crcs(struct bkey_i *, struct bch_extent_crc_unpacked);
|
|
|
|
void bch2_extent_crc_append(struct bkey_i *,
|
|
|
|
struct bch_extent_crc_unpacked);
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
/* Generic code for keys with pointers: */
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2021-06-11 09:44:27 +08:00
|
|
|
static inline bool bkey_is_btree_ptr(const struct bkey *k)
|
|
|
|
{
|
|
|
|
switch (k->type) {
|
|
|
|
case KEY_TYPE_btree_ptr:
|
|
|
|
case KEY_TYPE_btree_ptr_v2:
|
|
|
|
return true;
|
|
|
|
default:
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-09-08 04:13:20 +08:00
|
|
|
static inline bool bkey_extent_is_direct_data(const struct bkey *k)
|
2018-11-02 03:10:01 +08:00
|
|
|
{
|
|
|
|
switch (k->type) {
|
|
|
|
case KEY_TYPE_btree_ptr:
|
2020-02-08 02:38:02 +08:00
|
|
|
case KEY_TYPE_btree_ptr_v2:
|
2018-11-02 03:10:01 +08:00
|
|
|
case KEY_TYPE_extent:
|
2019-08-16 21:59:56 +08:00
|
|
|
case KEY_TYPE_reflink_v:
|
2018-11-02 03:10:01 +08:00
|
|
|
return true;
|
2017-03-17 14:18:50 +08:00
|
|
|
default:
|
2018-11-02 03:10:01 +08:00
|
|
|
return false;
|
2017-03-17 14:18:50 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2020-10-25 07:51:34 +08:00
|
|
|
static inline bool bkey_extent_is_inline_data(const struct bkey *k)
|
|
|
|
{
|
|
|
|
return k->type == KEY_TYPE_inline_data ||
|
|
|
|
k->type == KEY_TYPE_indirect_inline_data;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline unsigned bkey_inline_data_offset(const struct bkey *k)
|
|
|
|
{
|
|
|
|
switch (k->type) {
|
|
|
|
case KEY_TYPE_inline_data:
|
|
|
|
return sizeof(struct bch_inline_data);
|
|
|
|
case KEY_TYPE_indirect_inline_data:
|
|
|
|
return sizeof(struct bch_indirect_inline_data);
|
|
|
|
default:
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline unsigned bkey_inline_data_bytes(const struct bkey *k)
|
|
|
|
{
|
|
|
|
return bkey_val_bytes(k) - bkey_inline_data_offset(k);
|
|
|
|
}
|
|
|
|
|
|
|
|
#define bkey_inline_data_p(_k) (((void *) (_k).v) + bkey_inline_data_offset((_k).k))
|
|
|
|
|
2019-09-08 04:13:20 +08:00
|
|
|
static inline bool bkey_extent_is_data(const struct bkey *k)
|
|
|
|
{
|
2020-10-25 07:51:34 +08:00
|
|
|
return bkey_extent_is_direct_data(k) ||
|
|
|
|
bkey_extent_is_inline_data(k) ||
|
2019-09-08 04:13:20 +08:00
|
|
|
k->type == KEY_TYPE_reflink_p;
|
|
|
|
}
|
|
|
|
|
2019-08-16 21:59:56 +08:00
|
|
|
/*
|
|
|
|
* Should extent be counted under inode->i_sectors?
|
|
|
|
*/
|
2018-11-02 03:10:01 +08:00
|
|
|
static inline bool bkey_extent_is_allocation(const struct bkey *k)
|
2017-03-17 14:18:50 +08:00
|
|
|
{
|
2018-11-02 03:10:01 +08:00
|
|
|
switch (k->type) {
|
|
|
|
case KEY_TYPE_extent:
|
|
|
|
case KEY_TYPE_reservation:
|
2019-08-16 21:59:56 +08:00
|
|
|
case KEY_TYPE_reflink_p:
|
|
|
|
case KEY_TYPE_reflink_v:
|
2019-11-10 05:43:16 +08:00
|
|
|
case KEY_TYPE_inline_data:
|
2020-10-25 07:51:34 +08:00
|
|
|
case KEY_TYPE_indirect_inline_data:
|
2023-09-30 03:49:16 +08:00
|
|
|
case KEY_TYPE_error:
|
2018-11-02 03:10:01 +08:00
|
|
|
return true;
|
2017-03-17 14:18:50 +08:00
|
|
|
default:
|
2018-11-02 03:10:01 +08:00
|
|
|
return false;
|
2017-03-17 14:18:50 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2022-11-14 07:59:01 +08:00
|
|
|
static inline bool bkey_extent_is_unwritten(struct bkey_s_c k)
|
|
|
|
{
|
|
|
|
struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
|
|
|
|
|
|
|
|
bkey_for_each_ptr(ptrs, ptr)
|
|
|
|
if (ptr->unwritten)
|
|
|
|
return true;
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline bool bkey_extent_is_reservation(struct bkey_s_c k)
|
|
|
|
{
|
|
|
|
return k.k->type == KEY_TYPE_reservation ||
|
|
|
|
bkey_extent_is_unwritten(k);
|
|
|
|
}
|
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
static inline struct bch_devs_list bch2_bkey_devs(struct bkey_s_c k)
|
|
|
|
{
|
|
|
|
struct bch_devs_list ret = (struct bch_devs_list) { 0 };
|
|
|
|
struct bkey_ptrs_c p = bch2_bkey_ptrs_c(k);
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
bkey_for_each_ptr(p, ptr)
|
2023-12-24 10:02:45 +08:00
|
|
|
ret.data[ret.nr++] = ptr->dev;
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
return ret;
|
|
|
|
}
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
static inline struct bch_devs_list bch2_bkey_dirty_devs(struct bkey_s_c k)
|
|
|
|
{
|
|
|
|
struct bch_devs_list ret = (struct bch_devs_list) { 0 };
|
|
|
|
struct bkey_ptrs_c p = bch2_bkey_ptrs_c(k);
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
bkey_for_each_ptr(p, ptr)
|
|
|
|
if (!ptr->cached)
|
2023-12-24 10:02:45 +08:00
|
|
|
ret.data[ret.nr++] = ptr->dev;
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
return ret;
|
|
|
|
}
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
static inline struct bch_devs_list bch2_bkey_cached_devs(struct bkey_s_c k)
|
|
|
|
{
|
|
|
|
struct bch_devs_list ret = (struct bch_devs_list) { 0 };
|
|
|
|
struct bkey_ptrs_c p = bch2_bkey_ptrs_c(k);
|
2018-11-02 03:10:01 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
bkey_for_each_ptr(p, ptr)
|
|
|
|
if (ptr->cached)
|
2023-12-24 10:02:45 +08:00
|
|
|
ret.data[ret.nr++] = ptr->dev;
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
unsigned bch2_bkey_nr_ptrs(struct bkey_s_c);
|
|
|
|
unsigned bch2_bkey_nr_ptrs_allocated(struct bkey_s_c);
|
|
|
|
unsigned bch2_bkey_nr_ptrs_fully_allocated(struct bkey_s_c);
|
2018-02-24 05:26:10 +08:00
|
|
|
bool bch2_bkey_is_incompressible(struct bkey_s_c);
|
2019-11-17 05:25:58 +08:00
|
|
|
unsigned bch2_bkey_sectors_compressed(struct bkey_s_c);
|
bcachefs: Change when we allow overwrites
Originally, we'd check for -ENOSPC when getting a disk reservation
whenever the new extent took up more space on disk than the old extent.
Erasure coding screwed this up, because with erasure coding writes are
initially replicated, and then in the background the extra replicas are
dropped when the stripe is created. This means that with erasure coding
enabled, writes will always take up more space on disk than the data
they're overwriting - but, according to posix, overwrites aren't
supposed to return ENOSPC.
So, in this patch we fudge things: if the new extent has more replicas
than the _effective_ replicas of the old extent, or if the old extent is
compressed and the new one isn't, we check for ENOSPC when getting the
disk reservation - otherwise, we don't.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
2020-12-15 10:59:33 +08:00
|
|
|
|
|
|
|
unsigned bch2_bkey_replicas(struct bch_fs *, struct bkey_s_c);
|
2023-06-14 03:12:04 +08:00
|
|
|
unsigned bch2_extent_ptr_desired_durability(struct bch_fs *, struct extent_ptr_decoded *);
|
bcachefs: Nocow support
This adds support for nocow mode, where we do writes in-place when
possible. Patch components:
- New boolean filesystem and inode option, nocow: note that when nocow
is enabled, data checksumming and compression are implicitly disabled
- To prevent in-place writes from racing with data moves
(data_update.c) or bucket reuse (i.e. a bucket being reused and
re-allocated while a nocow write is in flight, we have a new locking
mechanism.
Buckets can be locked for either data update or data move, using a
fixed size hash table of two_state_shared locks. We don't have any
chaining, meaning updates and moves to different buckets that hash to
the same lock will wait unnecessarily - we'll want to watch for this
becoming an issue.
- The allocator path also needs to check for in-place writes in flight
to a given bucket before giving it out: thus we add another counter
to bucket_alloc_state so we can track this.
- Fsync now may need to issue cache flushes to block devices instead of
flushing the journal. We add a device bitmask to bch_inode_info,
ei_devs_need_flush, which tracks devices that need to have flushes
issued - note that this will lead to unnecessary flushes when other
codepaths have already issued flushes, we may want to replace this with
a sequence number.
- New nocow write path: look up extents, and if they're writable write
to them - otherwise fall back to the normal COW write path.
XXX: switch to sequence numbers instead of bitmask for devs needing
journal flush
XXX: ei_quota_lock being a mutex means bch2_nocow_write_done() needs to
run in process context - see if we can improve this
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
2022-11-03 05:12:00 +08:00
|
|
|
unsigned bch2_extent_ptr_durability(struct bch_fs *, struct extent_ptr_decoded *);
|
2019-11-17 05:25:58 +08:00
|
|
|
unsigned bch2_bkey_durability(struct bch_fs *, struct bkey_s_c);
|
|
|
|
|
2022-11-25 07:03:55 +08:00
|
|
|
void bch2_bkey_drop_device(struct bkey_s, unsigned);
|
|
|
|
void bch2_bkey_drop_device_noerror(struct bkey_s, unsigned);
|
2023-03-11 05:28:37 +08:00
|
|
|
|
|
|
|
const struct bch_extent_ptr *bch2_bkey_has_device_c(struct bkey_s_c, unsigned);
|
|
|
|
|
|
|
|
static inline struct bch_extent_ptr *bch2_bkey_has_device(struct bkey_s k, unsigned dev)
|
|
|
|
{
|
|
|
|
return (void *) bch2_bkey_has_device_c(k.s_c, dev);
|
|
|
|
}
|
|
|
|
|
2022-11-25 07:03:55 +08:00
|
|
|
bool bch2_bkey_has_target(struct bch_fs *, struct bkey_s_c, unsigned);
|
|
|
|
|
2021-02-18 02:37:22 +08:00
|
|
|
void bch2_bkey_extent_entry_drop(struct bkey_i *, union bch_extent_entry *);
|
2022-11-25 07:03:55 +08:00
|
|
|
|
|
|
|
static inline void bch2_bkey_append_ptr(struct bkey_i *k, struct bch_extent_ptr ptr)
|
|
|
|
{
|
bcachefs: Refactor memcpy into direct assignment
The memcpy() in bch2_bkey_append_ptr() is operating on an embedded fake
flexible array which looks to the compiler like it has 0 size. This
causes W=1 builds to emit warnings due to -Wstringop-overflow:
In file included from include/linux/string.h:254,
from include/linux/bitmap.h:11,
from include/linux/cpumask.h:12,
from include/linux/smp.h:13,
from include/linux/lockdep.h:14,
from include/linux/radix-tree.h:14,
from include/linux/backing-dev-defs.h:6,
from fs/bcachefs/bcachefs.h:182:
fs/bcachefs/extents.c: In function 'bch2_bkey_append_ptr':
include/linux/fortify-string.h:57:33: warning: writing 8 bytes into a region of size 0 [-Wstringop-overflow=]
57 | #define __underlying_memcpy __builtin_memcpy
| ^
include/linux/fortify-string.h:648:9: note: in expansion of macro '__underlying_memcpy'
648 | __underlying_##op(p, q, __fortify_size); \
| ^~~~~~~~~~~~~
include/linux/fortify-string.h:693:26: note: in expansion of macro '__fortify_memcpy_chk'
693 | #define memcpy(p, q, s) __fortify_memcpy_chk(p, q, s, \
| ^~~~~~~~~~~~~~~~~~~~
fs/bcachefs/extents.c:235:17: note: in expansion of macro 'memcpy'
235 | memcpy((void *) &k->v + bkey_val_bytes(&k->k),
| ^~~~~~
fs/bcachefs/bcachefs_format.h:287:33: note: destination object 'v' of size 0
287 | struct bch_val v;
| ^
Avoid making any structure changes and just replace the u64 copy into a
direct assignment, side-stepping the entire problem.
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Brian Foster <bfoster@redhat.com>
Cc: linux-bcachefs@vger.kernel.org
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202309192314.VBsjiIm5-lkp@intel.com/
Link: https://lore.kernel.org/r/20231010235609.work.594-kees@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
2023-10-19 07:07:32 +08:00
|
|
|
struct bch_extent_ptr *dest;
|
|
|
|
|
2023-03-11 05:28:37 +08:00
|
|
|
EBUG_ON(bch2_bkey_has_device(bkey_i_to_s(k), ptr.dev));
|
2022-11-25 07:03:55 +08:00
|
|
|
|
|
|
|
switch (k->k.type) {
|
|
|
|
case KEY_TYPE_btree_ptr:
|
|
|
|
case KEY_TYPE_btree_ptr_v2:
|
|
|
|
case KEY_TYPE_extent:
|
|
|
|
EBUG_ON(bkey_val_u64s(&k->k) >= BKEY_EXTENT_VAL_U64s_MAX);
|
|
|
|
|
|
|
|
ptr.type = 1 << BCH_EXTENT_ENTRY_ptr;
|
bcachefs: Refactor memcpy into direct assignment
The memcpy() in bch2_bkey_append_ptr() is operating on an embedded fake
flexible array which looks to the compiler like it has 0 size. This
causes W=1 builds to emit warnings due to -Wstringop-overflow:
In file included from include/linux/string.h:254,
from include/linux/bitmap.h:11,
from include/linux/cpumask.h:12,
from include/linux/smp.h:13,
from include/linux/lockdep.h:14,
from include/linux/radix-tree.h:14,
from include/linux/backing-dev-defs.h:6,
from fs/bcachefs/bcachefs.h:182:
fs/bcachefs/extents.c: In function 'bch2_bkey_append_ptr':
include/linux/fortify-string.h:57:33: warning: writing 8 bytes into a region of size 0 [-Wstringop-overflow=]
57 | #define __underlying_memcpy __builtin_memcpy
| ^
include/linux/fortify-string.h:648:9: note: in expansion of macro '__underlying_memcpy'
648 | __underlying_##op(p, q, __fortify_size); \
| ^~~~~~~~~~~~~
include/linux/fortify-string.h:693:26: note: in expansion of macro '__fortify_memcpy_chk'
693 | #define memcpy(p, q, s) __fortify_memcpy_chk(p, q, s, \
| ^~~~~~~~~~~~~~~~~~~~
fs/bcachefs/extents.c:235:17: note: in expansion of macro 'memcpy'
235 | memcpy((void *) &k->v + bkey_val_bytes(&k->k),
| ^~~~~~
fs/bcachefs/bcachefs_format.h:287:33: note: destination object 'v' of size 0
287 | struct bch_val v;
| ^
Avoid making any structure changes and just replace the u64 copy into a
direct assignment, side-stepping the entire problem.
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Brian Foster <bfoster@redhat.com>
Cc: linux-bcachefs@vger.kernel.org
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202309192314.VBsjiIm5-lkp@intel.com/
Link: https://lore.kernel.org/r/20231010235609.work.594-kees@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
2023-10-19 07:07:32 +08:00
|
|
|
dest = (struct bch_extent_ptr *)((void *) &k->v + bkey_val_bytes(&k->k));
|
|
|
|
*dest = ptr;
|
2023-03-05 12:05:55 +08:00
|
|
|
k->k.u64s++;
|
2022-11-25 07:03:55 +08:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
void bch2_extent_ptr_decoded_append(struct bkey_i *,
|
|
|
|
struct extent_ptr_decoded *);
|
2024-09-05 05:49:20 +08:00
|
|
|
void bch2_bkey_drop_ptr_noerror(struct bkey_s, struct bch_extent_ptr *);
|
|
|
|
void bch2_bkey_drop_ptr(struct bkey_s, struct bch_extent_ptr *);
|
2018-10-01 06:28:23 +08:00
|
|
|
|
2018-11-02 03:10:01 +08:00
|
|
|
#define bch2_bkey_drop_ptrs(_k, _ptr, _cond) \
|
2018-10-01 06:28:23 +08:00
|
|
|
do { \
|
2024-09-05 05:49:20 +08:00
|
|
|
__label__ _again; \
|
|
|
|
struct bkey_ptrs _ptrs; \
|
|
|
|
_again: \
|
|
|
|
_ptrs = bch2_bkey_ptrs(_k); \
|
2018-11-02 03:10:01 +08:00
|
|
|
\
|
2024-09-05 05:49:20 +08:00
|
|
|
bkey_for_each_ptr(_ptrs, _ptr) \
|
2018-10-01 06:28:23 +08:00
|
|
|
if (_cond) { \
|
2024-09-05 05:49:20 +08:00
|
|
|
bch2_bkey_drop_ptr(_k, _ptr); \
|
|
|
|
goto _again; \
|
2018-10-01 06:28:23 +08:00
|
|
|
} \
|
|
|
|
} while (0)
|
2017-03-17 14:18:50 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
bool bch2_bkey_matches_ptr(struct bch_fs *, struct bkey_s_c,
|
|
|
|
struct bch_extent_ptr, u64);
|
2022-06-14 07:17:45 +08:00
|
|
|
bool bch2_extents_match(struct bkey_s_c, struct bkey_s_c);
|
2023-03-11 05:28:37 +08:00
|
|
|
struct bch_extent_ptr *
|
|
|
|
bch2_extent_has_ptr(struct bkey_s_c, struct extent_ptr_decoded, struct bkey_s);
|
2019-11-17 05:25:58 +08:00
|
|
|
|
2023-02-22 08:22:44 +08:00
|
|
|
void bch2_extent_ptr_set_cached(struct bkey_s, struct bch_extent_ptr *);
|
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
bool bch2_extent_normalize(struct bch_fs *, struct bkey_s);
|
2024-03-12 11:11:46 +08:00
|
|
|
void bch2_extent_ptr_to_text(struct printbuf *out, struct bch_fs *, const struct bch_extent_ptr *);
|
2019-11-17 05:25:58 +08:00
|
|
|
void bch2_bkey_ptrs_to_text(struct printbuf *, struct bch_fs *,
|
|
|
|
struct bkey_s_c);
|
2024-08-13 09:31:25 +08:00
|
|
|
int bch2_bkey_ptrs_validate(struct bch_fs *, struct bkey_s_c,
|
|
|
|
enum bch_validate_flags);
|
2019-11-17 05:25:58 +08:00
|
|
|
|
2020-02-07 09:15:15 +08:00
|
|
|
void bch2_ptr_swab(struct bkey_s);
|
2019-11-17 05:25:58 +08:00
|
|
|
|
bcachefs: rebalance_work
This adds a new btree, rebalance_work, to eliminate scanning required
for finding extents that need work done on them in the background - i.e.
for the background_target and background_compression options.
rebalance_work is a bitset btree, where a KEY_TYPE_set corresponds to an
extent in the extents or reflink btree at the same pos.
A new extent field is added, bch_extent_rebalance, which indicates that
this extent has work that needs to be done in the background - and which
options to use. This allows per-inode options to be propagated to
indirect extents - at least in some circumstances. In this patch,
changing IO options on a file will not propagate the new options to
indirect extents pointed to by that file.
Updating (setting/clearing) the rebalance_work btree is done by the
extent trigger, which looks at the bch_extent_rebalance field.
Scanning is still requrired after changing IO path options - either just
for a given inode, or for the whole filesystem. We indicate that
scanning is required by adding a KEY_TYPE_cookie key to the
rebalance_work btree: the cookie counter is so that we can detect that
scanning is still required when an option has been flipped mid-way
through an existing scan.
Future possible work:
- Propagate options to indirect extents when being changed
- Add other IO path options - nr_replicas, ec, to rebalance_work so
they can be applied in the background when they change
- Add a counter, for bcachefs fs usage output, showing the pending
amount of rebalance work: we'll probably want to do this after the
disk space accounting rewrite (moving it to a new btree)
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
2023-10-21 01:33:14 +08:00
|
|
|
const struct bch_extent_rebalance *bch2_bkey_rebalance_opts(struct bkey_s_c);
|
|
|
|
unsigned bch2_bkey_ptrs_need_rebalance(struct bch_fs *, struct bkey_s_c,
|
|
|
|
unsigned, unsigned);
|
|
|
|
bool bch2_bkey_needs_rebalance(struct bch_fs *, struct bkey_s_c);
|
2024-08-24 03:35:22 +08:00
|
|
|
u64 bch2_bkey_sectors_need_rebalance(struct bch_fs *, struct bkey_s_c);
|
bcachefs: rebalance_work
This adds a new btree, rebalance_work, to eliminate scanning required
for finding extents that need work done on them in the background - i.e.
for the background_target and background_compression options.
rebalance_work is a bitset btree, where a KEY_TYPE_set corresponds to an
extent in the extents or reflink btree at the same pos.
A new extent field is added, bch_extent_rebalance, which indicates that
this extent has work that needs to be done in the background - and which
options to use. This allows per-inode options to be propagated to
indirect extents - at least in some circumstances. In this patch,
changing IO options on a file will not propagate the new options to
indirect extents pointed to by that file.
Updating (setting/clearing) the rebalance_work btree is done by the
extent trigger, which looks at the bch_extent_rebalance field.
Scanning is still requrired after changing IO path options - either just
for a given inode, or for the whole filesystem. We indicate that
scanning is required by adding a KEY_TYPE_cookie key to the
rebalance_work btree: the cookie counter is so that we can detect that
scanning is still required when an option has been flipped mid-way
through an existing scan.
Future possible work:
- Propagate options to indirect extents when being changed
- Add other IO path options - nr_replicas, ec, to rebalance_work so
they can be applied in the background when they change
- Add a counter, for bcachefs fs usage output, showing the pending
amount of rebalance work: we'll probably want to do this after the
disk space accounting rewrite (moving it to a new btree)
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
2023-10-21 01:33:14 +08:00
|
|
|
|
|
|
|
int bch2_bkey_set_needs_rebalance(struct bch_fs *, struct bkey_i *,
|
2024-01-17 05:20:21 +08:00
|
|
|
struct bch_io_opts *);
|
bcachefs: rebalance_work
This adds a new btree, rebalance_work, to eliminate scanning required
for finding extents that need work done on them in the background - i.e.
for the background_target and background_compression options.
rebalance_work is a bitset btree, where a KEY_TYPE_set corresponds to an
extent in the extents or reflink btree at the same pos.
A new extent field is added, bch_extent_rebalance, which indicates that
this extent has work that needs to be done in the background - and which
options to use. This allows per-inode options to be propagated to
indirect extents - at least in some circumstances. In this patch,
changing IO options on a file will not propagate the new options to
indirect extents pointed to by that file.
Updating (setting/clearing) the rebalance_work btree is done by the
extent trigger, which looks at the bch_extent_rebalance field.
Scanning is still requrired after changing IO path options - either just
for a given inode, or for the whole filesystem. We indicate that
scanning is required by adding a KEY_TYPE_cookie key to the
rebalance_work btree: the cookie counter is so that we can detect that
scanning is still required when an option has been flipped mid-way
through an existing scan.
Future possible work:
- Propagate options to indirect extents when being changed
- Add other IO path options - nr_replicas, ec, to rebalance_work so
they can be applied in the background when they change
- Add a counter, for bcachefs fs usage output, showing the pending
amount of rebalance work: we'll probably want to do this after the
disk space accounting rewrite (moving it to a new btree)
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
2023-10-21 01:33:14 +08:00
|
|
|
|
2019-11-17 05:25:58 +08:00
|
|
|
/* Generic extent code: */
|
|
|
|
|
2021-03-05 05:20:16 +08:00
|
|
|
enum bch_extent_overlap {
|
|
|
|
BCH_EXTENT_OVERLAP_ALL = 0,
|
|
|
|
BCH_EXTENT_OVERLAP_BACK = 1,
|
|
|
|
BCH_EXTENT_OVERLAP_FRONT = 2,
|
|
|
|
BCH_EXTENT_OVERLAP_MIDDLE = 3,
|
|
|
|
};
|
|
|
|
|
|
|
|
/* Returns how k overlaps with m */
|
|
|
|
static inline enum bch_extent_overlap bch2_extent_overlap(const struct bkey *k,
|
|
|
|
const struct bkey *m)
|
|
|
|
{
|
2022-11-24 16:12:22 +08:00
|
|
|
int cmp1 = bkey_lt(k->p, m->p);
|
|
|
|
int cmp2 = bkey_gt(bkey_start_pos(k), bkey_start_pos(m));
|
2021-03-05 05:20:16 +08:00
|
|
|
|
|
|
|
return (cmp1 << 1) + cmp2;
|
|
|
|
}
|
|
|
|
|
2019-11-10 08:02:48 +08:00
|
|
|
int bch2_cut_front_s(struct bpos, struct bkey_s);
|
|
|
|
int bch2_cut_back_s(struct bpos, struct bkey_s);
|
2018-11-28 07:30:56 +08:00
|
|
|
|
2019-07-19 05:21:21 +08:00
|
|
|
static inline void bch2_cut_front(struct bpos where, struct bkey_i *k)
|
2018-11-28 07:30:56 +08:00
|
|
|
{
|
2019-11-10 08:02:48 +08:00
|
|
|
bch2_cut_front_s(where, bkey_i_to_s(k));
|
2018-11-28 07:30:56 +08:00
|
|
|
}
|
|
|
|
|
2019-11-10 08:02:48 +08:00
|
|
|
static inline void bch2_cut_back(struct bpos where, struct bkey_i *k)
|
|
|
|
{
|
|
|
|
bch2_cut_back_s(where, bkey_i_to_s(k));
|
|
|
|
}
|
2019-08-23 04:07:37 +08:00
|
|
|
|
|
|
|
/**
|
|
|
|
* bch_key_resize - adjust size of @k
|
|
|
|
*
|
|
|
|
* bkey_start_offset(k) will be preserved, modifies where the extent ends
|
|
|
|
*/
|
|
|
|
static inline void bch2_key_resize(struct bkey *k, unsigned new_size)
|
|
|
|
{
|
|
|
|
k->p.offset -= k->size;
|
|
|
|
k->p.offset += new_size;
|
|
|
|
k->size = new_size;
|
|
|
|
}
|
2017-03-17 14:18:50 +08:00
|
|
|
|
|
|
|
#endif /* _BCACHEFS_EXTENTS_H */
|