dc6b9437a3
1256 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Arnaldo Carvalho de Melo
|
dc6b9437a3 |
emit: Handle structs with DW_AT_alignment=1 meaning __packed__
In the following struct the ceph_entity_addr entries all appear marked with a
__attribute__((__aligned__(8)), which, for the first two members of this type,
'peer_addr' and 'peer_addr_for_me', don't cause the regenerated struct to
differ in layout from the original layout put in place by the compiler as per
the original source code.
But the third member of this type, 'actual_peer_addr' ends up in a different
offset, even in a different cacheline, here is how it looks like in the code generated
from the original source code, at offset 568.
char in_banner[30]; /* 472 30 */
struct ceph_msg_connect out_connect; /* 502 33 */
/* --- cacheline 8 boundary (512 bytes) was 23 bytes ago --- */
struct ceph_msg_connect_reply in_reply; /* 535 26 */
struct ceph_entity_addr actual_peer_addr __attribute__((__aligned__(1))); /* 561 136 */
/* --- cacheline 10 boundary (640 bytes) was 57 bytes ago --- */
struct ceph_msg_header out_hdr; /* 697 53 */
/* XXX 2 bytes hole, try to pack */
/* --- cacheline 11 boundary (704 bytes) was 48 bytes ago --- */
And here is how it looks like when built from the regenerated source code, at
offset 568:
$ pfunct --compile /home/acme/git/build/v5.1-rc4+/fs/ceph/super.o > ceph.c
$ gcc -g -c ceph.c
$ pahole -C ceph_connection ceph.o | head -46
struct ceph_connection {
void * private; /* 0 8 */
const struct ceph_connection_operations * ops; /* 8 8 */
struct ceph_messenger * msgr; /* 16 8 */
atomic_t sock_state; /* 24 4 */
/* XXX 4 bytes hole, try to pack */
struct socket * sock; /* 32 8 */
struct ceph_entity_addr peer_addr __attribute__((__aligned__(8))); /* 40 136 */
/* --- cacheline 2 boundary (128 bytes) was 48 bytes ago --- */
struct ceph_entity_addr peer_addr_for_me __attribute__((__aligned__(8))); /* 176 136 */
/* --- cacheline 4 boundary (256 bytes) was 56 bytes ago --- */
long unsigned int flags; /* 312 8 */
/* --- cacheline 5 boundary (320 bytes) --- */
long unsigned int state; /* 320 8 */
const char * error_msg; /* 328 8 */
struct ceph_entity_name peer_name; /* 336 9 */
/* XXX 7 bytes hole, try to pack */
u64 peer_features; /* 352 8 */
u32 connect_seq; /* 360 4 */
u32 peer_global_seq; /* 364 4 */
struct ceph_auth_handshake * auth; /* 368 8 */
int auth_retry; /* 376 4 */
/* XXX 4 bytes hole, try to pack */
/* --- cacheline 6 boundary (384 bytes) --- */
struct mutex mutex; /* 384 32 */
struct list_head out_queue; /* 416 16 */
struct list_head out_sent; /* 432 16 */
/* --- cacheline 7 boundary (448 bytes) --- */
u64 out_seq; /* 448 8 */
u64 in_seq; /* 456 8 */
u64 in_seq_acked; /* 464 8 */
char in_banner[30]; /* 472 30 */
struct ceph_msg_connect out_connect; /* 502 33 */
/* --- cacheline 8 boundary (512 bytes) was 23 bytes ago --- */
struct ceph_msg_connect_reply in_reply; /* 535 26 */
/* XXX 7 bytes hole, try to pack */
struct ceph_entity_addr actual_peer_addr __attribute__((__aligned__(8))); /* 568 136 */
/* --- cacheline 11 boundary (704 bytes) --- */
$
That happens because 'struct ceph_entity_addr' has that __attribute__
((__aligned__(8)) in the regenerated source code, above, now look at how it
gets regenerated:
$ pahole -C ceph_entity_addr ceph.o
struct ceph_entity_addr {
__le32 type; /* 0 4 */
__le32 nonce; /* 4 4 */
struct __kernel_sockaddr_storage in_addr __attribute__((__aligned__(8))); /* 8 128 */
/* size: 136, cachelines: 3, members: 3 */
/* forced alignments: 1 */
/* last cacheline: 8 bytes */
} __attribute__((__aligned__(8)));
$
While when looking at the original DWARF:
$ pahole -C ceph_entity_addr /home/acme/git/build/v5.1-rc4+/fs/ceph/super.o
struct ceph_entity_addr {
__le32 type; /* 0 4 */
__le32 nonce; /* 4 4 */
struct __kernel_sockaddr_storage in_addr __attribute__((__aligned__(1))); /* 8 128 */
/* size: 136, cachelines: 3, members: 3 */
/* forced alignments: 1 */
/* last cacheline: 8 bytes */
} __attribute__((__aligned__(1)));
$
The confusion may further come from the fact that 'struct __kernel_sockaddr_storage' has,
in the regenerated source code, the __attribute__((__aligned__8)))
$ pahole -C __kernel_sockaddr_storage ceph.o
struct __kernel_sockaddr_storage {
__kernel_sa_family_t ss_family; /* 0 2 */
char __data[126]; /* 2 126 */
/* size: 128, cachelines: 2, members: 2 */
} __attribute__((__aligned__(8)));
$
Which is the same as in the original DWARF:
$ pahole -C __kernel_sockaddr_storage /home/acme/git/build/v5.1-rc4+/fs/ceph/super.o
struct __kernel_sockaddr_storage {
__kernel_sa_family_t ss_family; /* 0 2 */
char __data[126]; /* 2 126 */
/* size: 128, cachelines: 2, members: 2 */
} __attribute__((__aligned__(8)));
$
Looking at the original original source code for 'struct ceph_entity_addr'
helps here, as it reads:
include/linux/ceph/msgr.h, line 63:
/*
* entity_addr -- network address
*/
struct ceph_entity_addr {
__le32 type;
__le32 nonce; /* unique id for process (e.g. pid) */
struct sockaddr_storage in_addr;
} __attribute__ ((packed));
So the original code has no __attribute__((__aligned__(1))), so, lets look at
what the compiler generates for 'struct ceph_entity_addr':
$ readelf -wi /home/acme/git/build/v5.1-rc4+/fs/ceph/super.o | grep ceph_entity_addr -A7
<193a6> DW_AT_name : (indirect string, offset: 0x1586): ceph_entity_addr
<193aa> DW_AT_byte_size : 136
<193ab> DW_AT_alignment : 1
<193ac> DW_AT_decl_file : 296
<193ae> DW_AT_decl_line : 63
<193af> DW_AT_decl_column : 8
<193b0> DW_AT_sibling : <0x193e0>
<2><193b4>: Abbrev Number: 5 (DW_TAG_member)
$
So the natural alignment for 'struct ceph_entity_addr' ends up being the
natural alignment for 'struct __kernel_sockaddr_storage', which is 8, but
since 'struct ceph_entity_addr' was marked in the original source code as __packed__,
the compiler added the DW_AT_alignment: 1 to override that.
The heuristic in pahole, so far, took that __attribute__((__aligned__(1)))
literally:
$ pahole -C ceph_entity_addr /home/acme/git/build/v5.1-rc4+/fs/ceph/super.o
struct ceph_entity_addr {
__le32 type; /* 0 4 */
__le32 nonce; /* 4 4 */
struct __kernel_sockaddr_storage in_addr __attribute__((__aligned__(1))); /* 8 128 */
/* size: 136, cachelines: 3, members: 3 */
/* forced alignments: 1 */
/* last cacheline: 8 bytes */
} __attribute__((__aligned__(1)));
$
which ends up making the regenerated source code (with the __aligned__((1))),
generate a different layout, the __aligned__((8)) in one of its members
overrode that __aligned__((1)).
Take this into account and when faced with a structure which natural alignment
is not one and that has a DW_AT_alignment:1 to mean it really is __packed__.
Doing that makes the regenerated source code match the original structure
layouts, i.e. after the patch we get:
$ pahole -C ceph_entity_addr /home/acme/git/build/v5.1-rc4+/fs/ceph/super.o
struct ceph_entity_addr {
__le32 type; /* 0 4 */
__le32 nonce; /* 4 4 */
struct __kernel_sockaddr_storage in_addr __attribute__((__aligned__(1))); /* 8 128 */
/* size: 136, cachelines: 3, members: 3 */
/* forced alignments: 1 */
/* last cacheline: 8 bytes */
} __attribute__((__packed__));
$
And that member in 'struct ceph_connection', in the original, continues to read:
$ pahole -C ceph_connection /home/acme/git/build/v5.1-rc4+/fs/ceph/super.o | grep -w actual_peer_addr -B4 -A6
char in_banner[30]; /* 472 30 */
struct ceph_msg_connect out_connect; /* 502 33 */
/* --- cacheline 8 boundary (512 bytes) was 23 bytes ago --- */
struct ceph_msg_connect_reply in_reply; /* 535 26 */
struct ceph_entity_addr actual_peer_addr __attribute__((__aligned__(1))); /* 561 136 */
/* --- cacheline 10 boundary (640 bytes) was 57 bytes ago --- */
struct ceph_msg_header out_hdr; /* 697 53 */
/* XXX 2 bytes hole, try to pack */
/* --- cacheline 11 boundary (704 bytes) was 48 bytes ago --- */
$
While in the regenerated DWARF from the regenerated source code reads:
$ pfunct --compile /home/acme/git/build/v5.1-rc4+/fs/ceph/super.o > ceph.c
$ gcc -g -c ceph.c
$ pahole -C ceph_connection ceph.o | grep -w actual_peer_addr -B4 -A6
char in_banner[30]; /* 472 30 */
struct ceph_msg_connect out_connect; /* 502 33 */
/* --- cacheline 8 boundary (512 bytes) was 23 bytes ago --- */
struct ceph_msg_connect_reply in_reply; /* 535 26 */
struct ceph_entity_addr actual_peer_addr __attribute__((__aligned__(1))); /* 561 136 */
/* --- cacheline 10 boundary (640 bytes) was 57 bytes ago --- */
struct ceph_msg_header out_hdr; /* 697 53 */
/* XXX 2 bytes hole, try to pack */
/* --- cacheline 11 boundary (704 bytes) was 48 bytes ago --- */
$
I.e. it now matches.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
f78633cfb9 |
core: Infer __packed__ for union struct members
I.e. check that all the structs that are embedded in a union have their natural
alignment satisfied by the size of the array they are contained in,
Before this change we ended up not marking union struct members that didn't had
natural alignment violations as __packed__ even tho they had to be to be in a
struct that didn't satisfied their natural alignment requirements, which would
violate them when said union was in an array, i.e. the second entry would have
the non __packed__ union struct member in a bad position.
E.g. Before:
$ pahole -C ceph_osd_op /home/acme/git/build/v5.1-rc4+/net/ceph/osd_client.o
struct ceph_osd_op {
__le16 op; /* 0 2 */
__le32 flags; /* 2 4 */
union {
struct {
__le64 offset; /* 6 8 */
__le64 length; /* 14 8 */
__le64 truncate_size; /* 22 8 */
__le32 truncate_seq; /* 30 4 */
} __attribute__((__packed__)) extent; /* 6 28 */
struct {
__le32 name_len; /* 6 4 */
__le32 value_len; /* 10 4 */
__u8 cmp_op; /* 14 1 */
__u8 cmp_mode; /* 15 1 */
} __attribute__((__packed__)) xattr; /* 6 10 */
struct {
__u8 class_len; /* 6 1 */
__u8 method_len; /* 7 1 */
__u8 argc; /* 8 1 */
__le32 indata_len; /* 9 4 */
} __attribute__((__packed__)) cls; /* 6 7 */
struct {
__le64 cookie; /* 6 8 */
__le64 count; /* 14 8 */
} pgls; /* 6 16 */
struct {
__le64 snapid; /* 6 8 */
} snap; /* 6 8 */
struct {
__le64 cookie; /* 6 8 */
__le64 ver; /* 14 8 */
__u8 op; /* 22 1 */
__le32 gen; /* 23 4 */
} __attribute__((__packed__)) watch; /* 6 21 */
struct {
__le64 cookie; /* 6 8 */
} notify; /* 6 8 */
struct {
__le64 offset; /* 6 8 */
__le64 length; /* 14 8 */
__le64 src_offset; /* 22 8 */
} clonerange; /* 6 24 */
struct {
__le64 expected_object_size; /* 6 8 */
__le64 expected_write_size; /* 14 8 */
} alloc_hint; /* 6 16 */
struct {
__le64 snapid; /* 6 8 */
__le64 src_version; /* 14 8 */
__u8 flags; /* 22 1 */
__le32 src_fadvise_flags; /* 23 4 */
} __attribute__((__packed__)) copy_from; /* 6 21 */
}; /* 6 28 */
__le32 payload_len; /* 34 4 */
/* size: 38, cachelines: 1, members: 4 */
/* last cacheline: 38 bytes */
} __attribute__((__packed__));
After:
$ pahole -C ceph_osd_op /home/acme/git/build/v5.1-rc4+/net/ceph/osd_client.o
struct ceph_osd_op {
__le16 op; /* 0 2 */
__le32 flags; /* 2 4 */
union {
struct {
__le64 offset; /* 6 8 */
__le64 length; /* 14 8 */
__le64 truncate_size; /* 22 8 */
__le32 truncate_seq; /* 30 4 */
} __attribute__((__packed__)) extent; /* 6 28 */
struct {
__le32 name_len; /* 6 4 */
__le32 value_len; /* 10 4 */
__u8 cmp_op; /* 14 1 */
__u8 cmp_mode; /* 15 1 */
} __attribute__((__packed__)) xattr; /* 6 10 */
struct {
__u8 class_len; /* 6 1 */
__u8 method_len; /* 7 1 */
__u8 argc; /* 8 1 */
__le32 indata_len; /* 9 4 */
} __attribute__((__packed__)) cls; /* 6 7 */
struct {
__le64 cookie; /* 6 8 */
__le64 count; /* 14 8 */
} pgls; /* 6 16 */
struct {
__le64 snapid; /* 6 8 */
} snap; /* 6 8 */
struct {
__le64 cookie; /* 6 8 */
__le64 ver; /* 14 8 */
__u8 op; /* 22 1 */
__le32 gen; /* 23 4 */
} __attribute__((__packed__)) watch; /* 6 21 */
struct {
__le64 cookie; /* 6 8 */
} notify; /* 6 8 */
struct {
__le64 offset; /* 6 8 */
__le64 length; /* 14 8 */
__le64 src_offset; /* 22 8 */
} clonerange; /* 6 24 */
struct {
__le64 expected_object_size; /* 6 8 */
__le64 expected_write_size; /* 14 8 */
} alloc_hint; /* 6 16 */
struct {
__le64 snapid; /* 6 8 */
__le64 src_version; /* 14 8 */
__u8 flags; /* 22 1 */
__le32 src_fadvise_flags; /* 23 4 */
} __attribute__((__packed__)) copy_from; /* 6 21 */
}; /* 6 28 */
__le32 payload_len; /* 34 4 */
/* size: 38, cachelines: 1, members: 4 */
/* last cacheline: 38 bytes */
} __attribute__((__packed__));
$
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
75c52de9c6 |
core: Move packed_attribute_inferred from 'class' to 'type' class
Since we need to infer the attributes of union members too, so move there. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
1bb4527220 |
fprintf: Fixup const pointers
Before:
$ pahole -C nft_ctx /home/acme/git/build/v5.1-rc4+/net/netfilter/nft_set_rbtree.o
struct nft_ctx {
struct net * net; /* 0 8 */
struct nft_table * table; /* 8 8 */
struct nft_chain * chain; /* 16 8 */
const const struct nlattr * * nla; /* 24 8 */
u32 portid; /* 32 4 */
u32 seq; /* 36 4 */
u8 family; /* 40 1 */
u8 level; /* 41 1 */
bool report; /* 42 1 */
/* size: 48, cachelines: 1, members: 9 */
/* padding: 5 */
/* last cacheline: 48 bytes */
};
$
Original:
struct nft_ctx {
struct net *net;
struct nft_table *table;
struct nft_chain *chain;
const struct nlattr * const *nla;
u32 portid;
u32 seq;
u8 family;
u8 level;
bool report;
};
DWARF tags:
<1><12c8a>: Abbrev Number: 12 (DW_TAG_structure_type)
<12c8b> DW_AT_name : (indirect string, offset: 0xcc6f): nlattr
<12c8f> DW_AT_byte_size : 4
<12c93> DW_AT_sibling : <0x12cb2>
<1><12cb2>: Abbrev Number: 17 (DW_TAG_const_type)
<12cb3> DW_AT_type : <0x12c8a>
<1><12cf9>: Abbrev Number: 4 (DW_TAG_pointer_type)
<12cfa> DW_AT_byte_size : 8
<12cfb> DW_AT_type : <0x12cb2>
<1><12cff>: Abbrev Number: 17 (DW_TAG_const_type)
<12d00> DW_AT_type : <0x12cf9>
<1><1d54b>: Abbrev Number: 4 (DW_TAG_pointer_type)
<1d54c> DW_AT_byte_size : 8
<1d54d> DW_AT_type : <0x12cff>
<2><1e52e>: Abbrev Number: 14 (DW_TAG_member)
<1e52f> DW_AT_name : nla
<1e536> DW_AT_type : <0x1d54b>
<1e53a> DW_AT_data_member_location: 24
Fixed now:
$ pahole -C nft_ctx /home/acme/git/build/v5.1-rc4+/net/netfilter/nft_set_rbtree.o
struct nft_ctx {
struct net * net; /* 0 8 */
struct nft_table * table; /* 8 8 */
struct nft_chain * chain; /* 16 8 */
const struct nlattr * const * nla; /* 24 8 */
u32 portid; /* 32 4 */
u32 seq; /* 36 4 */
u8 family; /* 40 1 */
u8 level; /* 41 1 */
bool report; /* 42 1 */
/* size: 48, cachelines: 1, members: 9 */
/* padding: 5 */
/* last cacheline: 48 bytes */
};
$
So, one more full circled:
$ fullcircle /home/acme/git/build/v5.1-rc4+/net/netfilter/nft_set_rbtree.o
$
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
dc3d441961 |
core: Improve the natural alignment calculation
We need to take more than just arrays into account when figuring out the
natural alignment of struct members, looking recursively at types till
we get to basic types and pointers.
Before this patch the 'new' struct field in the 'v' union was considered
__packed__, when in fact it is not, as the natural alignment for the
'state_id' typedef is 4, so it can start at offset 36 (or 4 considering
just its container struct), see below:
$ pahole -IC nfsd4_lock /home/acme/git/build/v5.1-rc4+/fs/nfsd/nfs4xdr.o
/* Used at: /home/acme/git/linux/fs/nfsd/nfs4xdr.c */
/* <1717a> /home/acme/git/linux/fs/nfsd/xdr4.h:156 */
struct nfsd4_lock {
u32 lk_type; /* 0 4 */
u32 lk_reclaim; /* 4 4 */
u64 lk_offset; /* 8 8 */
u64 lk_length; /* 16 8 */
u32 lk_is_new; /* 24 4 */
/* XXX 4 bytes hole, try to pack */
union {
struct {
u32 open_seqid; /* 32 4 */
stateid_t open_stateid; /* 36 16 */
u32 lock_seqid; /* 52 4 */
clientid_t clientid; /* 56 8 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct xdr_netobj owner; /* 64 16 */
} __attribute__((__packed__)) new; /* 32 48 */
struct {
stateid_t lock_stateid; /* 32 16 */
u32 lock_seqid; /* 48 4 */
} __attribute__((__packed__)) old; /* 32 20 */
} v; /* 32 48 */
/* --- cacheline 1 boundary (64 bytes) was 16 bytes ago --- */
union {
struct {
stateid_t stateid; /* 80 16 */
} ok; /* 80 16 */
struct nfsd4_lock_denied denied; /* 80 48 */
} u; /* 80 48 */
/* size: 128, cachelines: 2, members: 7 */
/* sum members: 124, holes: 1, sum holes: 4 */
};
$
Asking for -rEIC, i.e. relative offsets, expand types we can see that
stateid_t opaque type:
struct {
/* typedef u32 -> __u32 */ unsigned int open_seqid; /* 0 4 */
/* typedef stateid_t */ struct {
/* typedef u32 -> __u32 */ unsigned int si_generation; /* 0 4 */
/* typedef stateid_opaque_t */ struct {
/* typedef clientid_t */ struct {
/* typedef u32 -> __u32 */ unsigned int cl_boot; /* 0 4 */
/* typedef u32 -> __u32 */ unsigned int cl_id; /* 4 4 */
} so_clid; /* 0 8 */
/* typedef u32 -> __u32 */ unsigned int so_id; /* 8 4 */
} si_opaque; /* 4 12 */
} open_stateid; /* 4 16 */
With the algorithm implemented in this patch we get it correctly as not
packed:
$ pahole -IC nfsd4_lock /home/acme/git/build/v5.1-rc4+/fs/nfsd/nfs4xdr.o
/* Used at: /home/acme/git/linux/fs/nfsd/nfs4xdr.c */
/* <1717a> /home/acme/git/linux/fs/nfsd/xdr4.h:156 */
struct nfsd4_lock {
u32 lk_type; /* 0 4 */
u32 lk_reclaim; /* 4 4 */
u64 lk_offset; /* 8 8 */
u64 lk_length; /* 16 8 */
u32 lk_is_new; /* 24 4 */
/* XXX 4 bytes hole, try to pack */
union {
struct {
u32 open_seqid; /* 32 4 */
stateid_t open_stateid; /* 36 16 */
u32 lock_seqid; /* 52 4 */
clientid_t clientid; /* 56 8 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct xdr_netobj owner; /* 64 16 */
} new; /* 32 48 */
struct {
stateid_t lock_stateid; /* 32 16 */
u32 lock_seqid; /* 48 4 */
} old; /* 32 20 */
} v; /* 32 48 */
/* --- cacheline 1 boundary (64 bytes) was 16 bytes ago --- */
union {
struct {
stateid_t stateid; /* 80 16 */
} ok; /* 80 16 */
struct nfsd4_lock_denied denied; /* 80 48 */
} u; /* 80 48 */
/* size: 128, cachelines: 2, members: 7 */
/* sum members: 124, holes: 1, sum holes: 4 */
};
Fixes:
|
||
|
Arnaldo Carvalho de Melo
|
ac32e5e908 |
codiff: Fix comparision of multi-cu against single-cu files
When the first arg, the old object file has multiple compile units, i.e.
multiple objects that were then linked into one, and the second just one
.o, or equivalent, i.e. a .BTF file, then codiff shouldn't try to
find the types in the single CU in each of the old CUs.
Think about a .BTF file generated from a multi-CU DWARF binary, it will
contain all the types in all of the DWARF CUs, so if we go on trying to
find all the BTF files in each of the CUs, we'll fail.
It only makes sense to go on the DWARF CUs looking for the type on the
.BTF section and then compare them.
Fixes:
|
||
|
Arnaldo Carvalho de Melo
|
f2641ce169 |
core: Take arrays into account when inferring if a struct is packed
Before:
$ pahole -C qrwlock /home/acme/git/build/v5.1-rc4+/fs/ceph/dir.o
struct qrwlock {
union {
atomic_t cnts; /* 0 4 */
struct {
u8 wlocked; /* 0 1 */
u8 __lstate[3]; /* 1 3 */
} __attribute__((__packed__)); /* 0 4 */
}; /* 0 4 */
arch_spinlock_t wait_lock; /* 4 4 */
/* size: 8, cachelines: 1, members: 2 */
/* last cacheline: 8 bytes */
};
I.e. __lstate's class_member->byte_size is 3, causing the misinference that that
struct was packed, it is naturally aligned, we need to look at the size of the
array's entries to figure out its natural alignment:
After:
$ pahole -C qrwlock /home/acme/git/build/v5.1-rc4+/fs/ceph/dir.o
struct qrwlock {
union {
atomic_t cnts; /* 0 4 */
struct {
u8 wlocked; /* 0 1 */
u8 __lstate[3]; /* 1 3 */
}; /* 0 4 */
}; /* 0 4 */
arch_spinlock_t wait_lock; /* 4 4 */
/* size: 8, cachelines: 1, members: 2 */
/* last cacheline: 8 bytes */
};
$
To further test:
$ cat packed_array_struct.c
struct sarray {
short array[3];
long long first;
} __attribute__((__packed__));
void foo(struct sarray *s) {}
$ gcc -g -c packed_array_struct.c
$ pahole packed_array_struct.o
struct sarray {
short int array[3]; /* 0 6 */
long long int first; /* 6 8 */
/* size: 14, cachelines: 1, members: 2 */
/* last cacheline: 14 bytes */
} __attribute__((__packed__));
$ cat packed_array_struct.c
struct sarray {
short array[3];
long long first;
};
void foo(struct sarray *s) {}
$ gcc -g -c packed_array_struct.c
$ pahole packed_array_struct.o
struct sarray {
short int array[3]; /* 0 6 */
/* XXX 2 bytes hole, try to pack */
long long int first; /* 8 8 */
/* size: 16, cachelines: 1, members: 2 */
/* sum members: 14, holes: 1, sum holes: 2 */
/* last cacheline: 16 bytes */
};
$
One more test:
$ cat packed_array_struct.c
struct sarray {
short a;
short array[3];
long long b;
};
void foo(struct sarray *s) {}
$
Before this patch:
$ gcc -g -c packed_array_struct.c
$ pahole packed_array_struct.o
struct sarray {
short int a; /* 0 2 */
short int array[3]; /* 2 6 */
long long int b; /* 8 8 */
/* size: 16, cachelines: 1, members: 3 */
/* last cacheline: 16 bytes */
} __attribute__((__packed__));
After:
$ pahole packed_array_struct.o
struct sarray {
short int a; /* 0 2 */
short int array[3]; /* 2 6 */
long long int b; /* 8 8 */
/* size: 16, cachelines: 1, members: 3 */
/* last cacheline: 16 bytes */
};
$
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
85c9936963 |
fprintf: Do not add explicit padding when struct has __aligned__ attr
Fixes:
|
||
|
Arnaldo Carvalho de Melo
|
b5e8fab596 |
emit: Cover void ** as a function parameter
Like in: static int cifs_setlease(struct file *file, long arg, struct file_lock **lease, void **priv) Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
28a3bc7add |
fprintf: Support packed enums
Check if the size is different than sizeof(int), which should be good enough for now for both 64-bit and 32-bit targets. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
f77a442f09 |
fprintf: Do not print the __aligned__ attribute if asked
I.e. honour conf_fprintf.suppress_aligned_attribute, noticed with
btfdiff, as BTF doesn't carries the alignment attribute, so can't
regenerate it, we need to suppress it so as to compare the output of
DWARF with that of the equivalent BTF.
Fixes:
|
||
|
Arnaldo Carvalho de Melo
|
ea583dac52 |
fprintf: Print zero sized flat arrays as [], not [0]
To match the case when we really have just one dimension, so the --flat-arrays should show for zero sized arrays, [], not [0]: Noticed with btfdiff. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
f909f13dd7 |
fprintf: Fixup handling of unnamed bitfields
We were only handling holes inside bitfields as a request to change the
byte_offset, which is not the case when instead of 'int foo:0;' we have
'int foo:6;' to ask for a explicit 6 bit hole inside a bitfield, like
in:
Before this patch:
$ pahole -F btf -C kvm_mmu_page_role /home/acme/git/build/v5.1-rc4+/arch/x86/kvm/hyperv.o
union kvm_mmu_page_role {
u32 word; /* 0 4 */
struct {
unsigned int level:4; /* 0: 0 4 */
unsigned int gpte_is_8_bytes:1; /* 0: 4 4 */
unsigned int quadrant:2; /* 0: 5 4 */
unsigned int direct:1; /* 0: 7 4 */
unsigned int access:3; /* 0: 8 4 */
unsigned int invalid:1; /* 0:11 4 */
unsigned int nxe:1; /* 0:12 4 */
unsigned int cr0_wp:1; /* 0:13 4 */
unsigned int smep_andnot_wp:1; /* 0:14 4 */
unsigned int smap_andnot_wp:1; /* 0:15 4 */
unsigned int ad_disabled:1; /* 0:16 4 */
unsigned int guest_mode:1; /* 0:17 4 */
/* XXX 6 bits hole, try to pack */
/* Force alignment to the next boundary: */
unsigned int :0;
unsigned int smm:8; /* 0:24 4 */
}; /* 0 4 */
};
$
After:
$ pahole -F btf -C kvm_mmu_page_role /home/acme/git/build/v5.1-rc4+/arch/x86/kvm/hyperv.o
union kvm_mmu_page_role {
u32 word; /* 0 4 */
struct {
unsigned int level:4; /* 0: 0 4 */
unsigned int gpte_is_8_bytes:1; /* 0: 4 4 */
unsigned int quadrant:2; /* 0: 5 4 */
unsigned int direct:1; /* 0: 7 4 */
unsigned int access:3; /* 0: 8 4 */
unsigned int invalid:1; /* 0:11 4 */
unsigned int nxe:1; /* 0:12 4 */
unsigned int cr0_wp:1; /* 0:13 4 */
unsigned int smep_andnot_wp:1; /* 0:14 4 */
unsigned int smap_andnot_wp:1; /* 0:15 4 */
unsigned int ad_disabled:1; /* 0:16 4 */
unsigned int guest_mode:1; /* 0:17 4 */
/* XXX 6 bits hole, try to pack */
unsigned int :6;
unsigned int smm:8; /* 0:24 4 */
}; /* 0 4 */
};
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Yonghong Song <yhs@fb.com>
Fixes:
|
||
|
Arnaldo Carvalho de Melo
|
3247a777dc |
core: Infer if a struct is packed by the offsets/natural alignments
As DWARF (nor BTF) provides explicit attributes, we need to look at the
natural alignments, a byte is always alignted, etc.
This probably fails with things like __attribute__(__aligned(power-of-two)),
but with it most of the kernel data structures are full circled, i.e.
'pfunct --compile' regenerates source code from debug info that when
compiled generats debug info that end up matching the original sources.
$ cat a.c
#define __packed __attribute__((__packed__))
struct filename {
const char * name;
const char * uptr;
int refcnt;
};
void m(struct filename *f) {}
$ gcc -g -c a.c
$ pahole a.o
struct filename {
const char * name; /* 0 8 */
const char * uptr; /* 8 8 */
int refcnt; /* 16 4 */
/* size: 24, cachelines: 1, members: 3 */
/* padding: 4 */
/* last cacheline: 24 bytes */
};
$ cat a.c
#define __packed __attribute__((__packed__))
struct filename {
const char * name;
const char * uptr;
int refcnt;
} __packed;
void m(struct filename *f) {}
$ gcc -g -c a.c
$ pahole a.o
struct filename {
const char * name; /* 0 8 */
const char * uptr; /* 8 8 */
int refcnt; /* 16 4 */
/* size: 20, cachelines: 1, members: 3 */
/* last cacheline: 20 bytes */
} __attribute__((__packed__));
$ cat a.c
#define __packed __attribute__((__packed__))
struct filename {
const char * name;
int refcnt;
const char * uptr;
};
void m(struct filename *f) {}
$ gcc -g -c a.c
$ pahole a.o
struct filename {
const char * name; /* 0 8 */
int refcnt; /* 8 4 */
/* XXX 4 bytes hole, try to pack */
const char * uptr; /* 16 8 */
/* size: 24, cachelines: 1, members: 3 */
/* sum members: 20, holes: 1, sum holes: 4 */
/* last cacheline: 24 bytes */
};
$ cat a.c
#define __packed __attribute__((__packed__))
struct filename {
const char * name;
int refcnt;
const char * uptr;
} __packed;
void m(struct filename *f) {}
$ gcc -g -c a.c
$ pahole a.o
struct filename {
const char * name; /* 0 8 */
int refcnt; /* 8 4 */
const char * uptr; /* 12 8 */
/* size: 20, cachelines: 1, members: 3 */
/* last cacheline: 20 bytes */
} __attribute__((__packed__));
$ cat a.c
#define __packed __attribute__((__packed__))
struct filename {
const char * name;
const char * uptr;
unsigned char refcnt;
};
void m(struct filename *f) {}
$ gcc -g -c a.c
$ pahole a.o
struct filename {
const char * name; /* 0 8 */
const char * uptr; /* 8 8 */
unsigned char refcnt; /* 16 1 */
/* size: 24, cachelines: 1, members: 3 */
/* padding: 7 */
/* last cacheline: 24 bytes */
};
$ cat a.c
#define __packed __attribute__((__packed__))
struct filename {
const char * name;
const char * uptr;
unsigned char refcnt;
} __packed;
void m(struct filename *f) {}
$ gcc -g -c a.c
$ pahole a.o
struct filename {
const char * name; /* 0 8 */
const char * uptr; /* 8 8 */
unsigned char refcnt; /* 16 1 */
/* size: 17, cachelines: 1, members: 3 */
/* last cacheline: 17 bytes */
} __attribute__((__packed__));
$ cat a.c
#define __packed __attribute__((__packed__))
struct filename {
const char * name;
unsigned char refcnt;
const char * uptr;
};
void m(struct filename *f) {}
$ gcc -g -c a.c
$ pahole a.o
struct filename {
const char * name; /* 0 8 */
unsigned char refcnt; /* 8 1 */
/* XXX 7 bytes hole, try to pack */
const char * uptr; /* 16 8 */
/* size: 24, cachelines: 1, members: 3 */
/* sum members: 17, holes: 1, sum holes: 7 */
/* last cacheline: 24 bytes */
};
$ cat a.c
#define __packed __attribute__((__packed__))
struct filename {
const char * name;
unsigned char refcnt;
const char * uptr;
} __packed;
void m(struct filename *f) {}
$ gcc -g -c a.c
$ pahole a.o
struct filename {
const char * name; /* 0 8 */
unsigned char refcnt; /* 8 1 */
const char * uptr; /* 9 8 */
/* size: 17, cachelines: 1, members: 3 */
/* last cacheline: 17 bytes */
} __attribute__((__packed__));
$
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
13e5b9fc00 |
fprintf: Add unnamed bitfield padding at the end to rebuild original type
Just like the 'struct timex' in the linux kernel UAPI, that is now
correctly reconstructed as:
$ pahole -IC timex /home/acme/git/build/v5.0-rc2+/kernel/time/posix-clock.o | tail -32
__kernel_long_t ppsfreq; /* 96 8 */
__kernel_long_t jitter; /* 104 8 */
int shift; /* 112 4 */
/* XXX 4 bytes hole, try to pack */
__kernel_long_t stabil; /* 120 8 */
/* --- cacheline 2 boundary (128 bytes) --- */
__kernel_long_t jitcnt; /* 128 8 */
__kernel_long_t calcnt; /* 136 8 */
__kernel_long_t errcnt; /* 144 8 */
__kernel_long_t stbcnt; /* 152 8 */
int tai; /* 160 4 */
/* Force padding: */
int :32;
int :32;
int :32;
int :32;
int :32;
int :32;
int :32;
int :32;
int :32;
int :32;
int :32;
/* size: 208, cachelines: 4, members: 20 */
/* sum members: 152, holes: 3, sum holes: 12 */
/* padding: 44 */
/* last cacheline: 16 bytes */
};
$
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
ccd67bdb20 |
fprintf: Print "const" for class members more early, in type__fprintf()
We want to reach array__fprintf() from here, with the class_member
name, as __tag__name() isn't handling arrays properly.
I.e. to print an array when we have its name we can't use __tag__name().
This also stops printing 0 for zero sized arrays and trows away the
extra DW_TAG_const_type that comes with zero sized arrays, where we
have:
class_member type: DW_TAG_const_type 1
DW_TAG_const_type 1: DW_TAG_array_type 2
DW_TAG_array_type 2: 0 entries, type: DW_TAG_const_type 3
DW_TAG_const_type 3: real type of the zero sized array
For instance, after this patch we get a sane reconstruction of this
type:
$ pahole -C filename /home/acme/git/build/v5.0-rc2+/ipc/mqueue.o
struct filename {
const char * name; /* 0 8 */
const char * uptr; /* 8 8 */
int refcnt; /* 16 4 */
/* XXX 4 bytes hole, try to pack */
struct audit_names * aname; /* 24 8 */
const char iname[]; /* 32 0 */
/* size: 32, cachelines: 1, members: 5 */
/* sum members: 28, holes: 1, sum holes: 4 */
/* last cacheline: 32 bytes */
};
$
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
b42d77b0bb |
fprintf: Print __attribute__((__aligned__(N))) for structs/classes
For instance:
$ pahole -C kern_ipc_perm /home/acme/git/build/v5.0-rc2+/ipc/util.o
struct kern_ipc_perm {
spinlock_t lock; /* 0 4 */
bool deleted; /* 4 1 */
/* XXX 3 bytes hole, try to pack */
int id; /* 8 4 */
key_t key; /* 12 4 */
kuid_t uid; /* 16 4 */
kgid_t gid; /* 20 4 */
kuid_t cuid; /* 24 4 */
kgid_t cgid; /* 28 4 */
umode_t mode; /* 32 2 */
/* XXX 6 bytes hole, try to pack */
long unsigned int seq; /* 40 8 */
void * security; /* 48 8 */
struct rhash_head khtnode; /* 56 8 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct callback_head rcu __attribute__((__aligned__(8))); /* 64 16 */
refcount_t refcount; /* 80 4 */
/* size: 128, cachelines: 2, members: 14 */
/* sum members: 75, holes: 2, sum holes: 9 */
/* padding: 44 */
/* forced alignments: 1 */
} __attribute__((__aligned__(64)));
$
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
1c9c1d6bbd |
dwarf_loader: Store DW_AT_alignment if available in DW_TAG_{structure,union,class}_type
That is not just for DW_TAG_class_member :-) Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
41c55858da |
codiff: Add --quiet option
To avoid printing anything when there are no differences to show. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
a104eb1ea1 |
fprintf: Notice explicit bitfield alignment modifications
I.e. when we find that the last member has a bit_hole, i.e. it is part
of a bitfield, and the current field has a bitfield_size, i.e. it _also_
is part of a bitfield, the only explanation is that they were
artificially put in different base types, i.e. like in these fields
in the linux kernel 'struct task_struct', here reconstructed by pahole:
$ pahole -C task_struct ~/git/build/v5.1-rc2+/kernel/sched/core.o | grep :0 -B9 -A12
unsigned int personality; /* 1128 4 */
unsigned int sched_reset_on_fork:1; /* 1132: 0 4 */
unsigned int sched_contributes_to_load:1; /* 1132: 1 4 */
unsigned int sched_migrated:1; /* 1132: 2 4 */
unsigned int sched_remote_wakeup:1; /* 1132: 3 4 */
/* XXX 28 bits hole, try to pack */
/* Force alignment to the next boundary: */
unsigned int :0;
unsigned int in_execve:1; /* 1136: 0 4 */
unsigned int in_iowait:1; /* 1136: 1 4 */
unsigned int restore_sigmask:1; /* 1136: 2 4 */
unsigned int in_user_fault:1; /* 1136: 3 4 */
unsigned int no_cgroup_migration:1; /* 1136: 4 4 */
unsigned int use_memdelay:1; /* 1136: 5 4 */
/* XXX 26 bits hole, try to pack */
/* XXX 4 bytes hole, try to pack */
long unsigned int atomic_flags; /* 1144 8 */
$
This matches the original definition in the original kernel sources, and
further more, the following sequence proves that with this and DW_AT_alignment,
we can go full circle, i.e.:
1. from an object file reconstruct the source code for all the types that
appears in function signatures, if pointers, them they will be fully defined,
not just forward declared:
$ pfunct --compile=sched_change_group ~/git/build/v5.1-rc2+/kernel/sched/core.o | egrep -w 'sched_change_group|task_struct {' -B10 -A5
/* --- cacheline 3 boundary (192 bytes) --- */
struct fpu fpu __attribute__((__aligned__(64))); /* 192 4160 */
/* size: 4352, cachelines: 68, members: 21 */
/* sum members: 4316, holes: 2, sum holes: 32 */
/* sum bitfield members: 2 bits, bit holes: 1, sum bit holes: 30 bits */
/* forced alignments: 1, forced holes: 1, sum forced holes: 28 */
};
struct task_struct {
struct thread_info thread_info; /* 0 16 */
/* XXX last struct has 4 bytes of padding */
volatile long int state; /* 16 8 */
--
/* --- cacheline 104 boundary (6656 bytes) --- */
struct thread_struct thread __attribute__((__aligned__(64))); /* 6656 4352 */
/* size: 11008, cachelines: 172, members: 207 */
/* sum members: 10902, holes: 16, sum holes: 98 */
/* sum bitfield members: 10 bits, bit holes: 2, sum bit holes: 54 bits */
/* paddings: 3, sum paddings: 14 */
/* forced alignments: 6, forced holes: 1, sum forced holes: 40 */
};
void sched_change_group(struct task_struct * tsk, int type)
{
}
$
2. Build the regenerated skeleton function + its types:
$ pfunct --compile=sched_change_group ~/git/build/v5.1-rc2+/kernel/sched/core.o > sched_change_group.c
$ gcc -g -c sched_change_group.c
$ file sched_change_group.o
sched_change_group.o: ELF 64-bit LSB relocatable, x86-64, version 1 (SYSV), with debug_info, not stripped
$
3. Now lets see if the original 'struct task_struct' printed by pahole, matches
the the output printed by pahole for the DWARF info generated for the regenerated
'struct task_struct' source code in sched_change_group.c:
$ pahole -C task_struct sched_change_group.o | tail
/* --- cacheline 104 boundary (6656 bytes) --- */
struct thread_struct thread __attribute__((__aligned__(64))); /* 6656 4352 */
/* size: 11008, cachelines: 172, members: 207 */
/* sum members: 10902, holes: 16, sum holes: 98 */
/* sum bitfield members: 10 bits, bit holes: 2, sum bit holes: 54 bits */
/* paddings: 3, sum paddings: 14 */
/* forced alignments: 6, forced holes: 1, sum forced holes: 40 */
};
$ pahole -C task_struct ~/git/build/v5.1-rc2+/kernel/sched/core.o | tail
/* --- cacheline 104 boundary (6656 bytes) --- */
struct thread_struct thread __attribute__((__aligned__(64))); /* 6656 4352 */
/* size: 11008, cachelines: 172, members: 207 */
/* sum members: 10902, holes: 16, sum holes: 98 */
/* sum bitfield members: 10 bits, bit holes: 2, sum bit holes: 54 bits */
/* paddings: 3, sum paddings: 14 */
/* forced alignments: 6, forced holes: 1, sum forced holes: 40 */
};
$
Furthermore:
$ pahole -C task_struct ~/git/build/v5.1-rc2+/kernel/sched/core.o > /tmp/original
$ pahole -C task_struct sched_change_group.o > /tmp/regenerated
$ diff -u /tmp/original /tmp/regenerated
$
So one of the most complex data structures in the Linux kernel seems to be under control,
and it uses zero sized unnamed bitfields and __attribute__((aligned(N))), a DWARF5 goodie,
time to go tag v1.13!
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Yonghong Song <yhs@fb.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
75f32a24c7 |
codiff: Improve the comparision of anonymous struct members
I.e. 'union {};', 'struct {};' members were always appearing as having
been removed, as we normally do lookup by member name, to find out if
its offset, size, type, etc changed.
For unnamed members, try a different heuristic, i.e. look for the nth
anonymous member, this way we're just trying to compare the first
unnamed member of, say, struct OLD with the first unnamed member of
struct NEW, etc.
For OLD == NEW, this works well, for OLD != NEW because some non
anonymous field got added, removed or moved around, ditto, and when the
number of unnamed fields gets decreased, then we can mix things up, and
compare the previously first in A with the previously first in B.
For the current intended use case of:
1) compile a .c file into a .o file with debugging info, say FILE.o
2) use 'pfunct --compile FILE.o > regenerated-FILE.c'
3) compile regenerated-FILE.c into regenerated-FILE.o with debugging info
4) codiff --struct FILE.o regenerated-FILE.o and find out if they match
This gets us moving forward as we'll spot differences with this algo.
For the future we can use a few more heuristics or stop using search by
name members, instead traversing both structs in tandem, spotting the
differences by comparing the fields that way.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
6b1e43f2c1 |
codiff: When comparing against a file with just one CU don't bother finding by name
I.e. when we have two object files with debugging info, and one of them jas just one CU, then compare all the CUs in the other file to this unique CU. Case in hand: encode BTF in a file, then the BTF info has everything in just one "compile unit", so when looking at the types in the DWARF originals, we should just compare its types to what is in the single BTF "compile unit". Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
15a754f224 |
core: Add nr_entries member to 'struct cus'
Will be used when considering comparing multiple CU entries in a struct cus to the sole compile unit in a second file, like when comparing the types in a multi-CU DWARF file like vmlinux against the combined, deduplicated entries in that vmlinux .BTF section. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
99750f244c |
pfunct: Generate a valid return type for the --compile bodies
Before:
$ pfunct --compile examples/tcp.o > tcp.pahole.c
$ clang --target=bpf -c -g tcp.pahole.c |& tail
tcp.pahole.c:6154:1: warning: control reaches end of non-void function [-Wreturn-type]
}
^
tcp.pahole.c:6158:1: warning: control reaches end of non-void function [-Wreturn-type]
}
^
tcp.pahole.c:6170:1: warning: control reaches end of non-void function [-Wreturn-type]
}
^
192 warnings generated.
$ head -6170 tcp.pahole.c | tail -3
inline int arch_atomic_read(const atomic_t * v)
{
}
$
After:
$ pfunct --compile examples/tcp.o > tcp.pahole.c
$ clang --target=bpf -c -g tcp.pahole.c
$ grep -A3 -w arch_atomic_read tcp.pahole.c
inline int arch_atomic_read(const atomic_t * v)
{
return 0;
}
$
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
881aabd6fc |
reorganize: Introduce class__for_each_member_from_safe()
Reducing boilerplate, keeping consistent with the other member traversal helpers. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
1b2e3389f3 |
reorganize: Introduce class__for_each_member_reverse()
Reducing boilerplate, keeping consistent with the other member traversal helpers. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
10fef2916d |
reorganize: Introduce class__for_each_member_continue()
To get a lot of boilerplate behind a nice helper. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
e7a56ee8cc |
reorganize: Introduce class__for_each_member_from()
To get a lot of boilerplate behind a nice helper. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
9a79bb6ced |
tag: Introduce tag__is_pointer_to()
To shorten the check if a tag is a pointer to a particular type. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
45ad545944 |
tag: Introduce tag__is_pointer()
For the usual idiom to ask if a tag is a pointer, removing a bit of DWARFism and shortening the operation. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
89ce57a02e |
pdwtags: Find holes in structs
The pdwtags prints all tags, so call class__find_holes() for structs so that we don't print BFAs. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
ce6f393bc9 |
fprintf: Fixup the printing of const parameters
The last problem with 'pfunct --compile' at least for tcp.o:
Before:
$ pfunct --compile examples/tcp.o > tcp.pahole.c
$ gcc -c tcp.pahole.c -g
tcp.pahole.c:1808:48: error: unknown type name ‘u8const’; did you mean ‘const’?
inline void tcp_set_ca_state(struct sock * sk, u8const ca_state)
^~~~~~~
const
tcp.pahole.c:5346:56: error: unknown type name ‘intconst’; did you mean ‘const’?
inline void skb_set_tail_pointer(struct sk_buff * skb, intconst offset)
^~~~~~~~
const
tcp.pahole.c:5914:37: error: unknown type name ‘gfp_tconst’; did you mean ‘gfp_t’?
inline bool gfpflags_allow_blocking(gfp_tconst gfp_flags)
^~~~~~~~~~
gfp_t
tcp.pahole.c:5926:24: error: unknown type name ‘ktime_tconst’; did you mean ‘ktime_t’?
inline s64 ktime_to_ns(ktime_tconst kt)
^~~~~~~~~~~~
ktime_t
tcp.pahole.c:5939:54: warning: ‘struct timespec64const’ declared inside parameter list will not be visible outside of this definition or declaration
inline struct timespec timespec64_to_timespec(struct timespec64const ts64)
^~~~~~~~~~~~~~~
tcp.pahole.c:5939:70: error: parameter 1 (‘ts64’) has incomplete type
inline struct timespec timespec64_to_timespec(struct timespec64const ts64)
~~~~~~~~~~~~~~~~~~~~~~~^~~~
$
After:
$ pfunct --compile examples/tcp.o > tcp.pahole.c
$ gcc -c tcp.pahole.c -g
Because:
$ grep -A2 tcp_set_ca_state tcp.pahole.c
inline void tcp_set_ca_state(struct sock * sk, const u8 ca_state)
{
}
$
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
7aec7dd6c2 |
pfunct: Do not reconstruct external functions
I.e. those with DW_AT_external set, to avoid regenerating multiple times
things like __compiletime_assert_1504:
<5><2fc41>: Abbrev Number: 100 (DW_TAG_subprogram)
<2fc42> DW_AT_external : 1
<2fc42> DW_AT_name : (indirect string, offset: 0x1751f): __compiletime_assert_1504
<2fc46> DW_AT_decl_file : 1
<2fc47> DW_AT_decl_line : 1504
<2fc49> DW_AT_decl_column : 2
<2fc4a> DW_AT_prototyped : 1
<2fc4a> DW_AT_declaration : 1
<5><2fc4a>: Abbrev Number: 0
<4><2fc4b>: Abbrev Number: 0
<3><2fc4c>: Abbrev Number: 0
<2><2fc4d>: Abbrev Number: 34 (DW_TAG_lexical_block)
<3><2fc4e>: Abbrev Number: 12 (DW_TAG_variable)
<2fc4f> DW_AT_name : (indirect string, offset: 0xbcc6): ____ptr
<2fc53> DW_AT_decl_file : 1
<2fc54> DW_AT_decl_line : 1504
<2fc56> DW_AT_decl_column : 2
<2fc57> DW_AT_type : <0x6441>
<3><2fc5b>: Abbrev Number: 34 (DW_TAG_lexical_block)
<4><2fc5c>: Abbrev Number: 12 (DW_TAG_variable)
<2fc5d> DW_AT_name : (indirect string, offset: 0xeb74): __mptr
<2fc61> DW_AT_decl_file : 1
<2fc62> DW_AT_decl_line : 1504
<2fc64> DW_AT_decl_column : 2
<2fc65> DW_AT_type : <0x5a2>
<4><2fc69>: Abbrev Number: 34 (DW_TAG_lexical_block)
<5><2fc6a>: Abbrev Number: 100 (DW_TAG_subprogram)
<2fc6b> DW_AT_external : 1
<2fc6b> DW_AT_name : (indirect string, offset: 0x1751f): __compiletime_assert_1504
<2fc6f> DW_AT_decl_file : 1
<2fc70> DW_AT_decl_line : 1504
<2fc72> DW_AT_decl_column : 2
<2fc73> DW_AT_prototyped : 1
<2fc73> DW_AT_declaration : 1
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
163b873f81 |
pfunct: Do not reconstruct inline expansions of functions
I.e. those that point back to the inline function via
DW_AT_abstract_origin.
For instance:
<1><34b65>: Abbrev Number: 156 (DW_TAG_subprogram)
<34b67> DW_AT_external : 1
<34b67> DW_AT_name : (indirect string, offset: 0x2404): tcp_enter_memory_pressure
<34b6b> DW_AT_decl_file : 1
<34b6c> DW_AT_decl_line : 324
<34b6e> DW_AT_decl_column : 6
<34b6f> DW_AT_prototyped : 1
<34b6f> DW_AT_inline : 1 (inlined)
<SNIP>
<1><37f45>: Abbrev Number: 149 (DW_TAG_subprogram)
<37f47> DW_AT_abstract_origin: <0x34b65>
<37f4b> DW_AT_low_pc : 0x1000
<37f53> DW_AT_high_pc : 0x48
<37f5b> DW_AT_frame_base : 1 byte block: 9c (DW_OP_call_frame_cfa)
<37f5d> DW_AT_GNU_all_call_sites: 1
<37f5d> DW_AT_sibling : <0x38032>
Generated by:
Compilation Unit @ offset 0x0:
Length: 0x3b40b (32-bit)
Version: 4
Abbrev Offset: 0x0
Pointer Size: 8
<0><b>: Abbrev Number: 215 (DW_TAG_compile_unit)
<d> DW_AT_producer : (indirect string, offset: 0xb0bc): GNU C89 8.2.1 20181215 (Red Hat 8.2.1-6) -mno-sse -mno-mmx -mno-sse2 -mno-3dnow -mno-avx -m64 -mno-80387 -mno-fp-ret-in-387 -mpreferred-stack-boundary=3 -mskip-rax-setup -mtune=generic -mno-red-zone -mcmodel=kernel -mindirect-branch=thunk-extern -mindirect-branch-register -mrecord-mcount -mfentry -march=x86-64 -g -O2 -std=gnu90 -p -fno-strict-aliasing -fno-common -fshort-wchar -fno-PIE -falign-jumps=1 -falign-loops=1 -fno-asynchronous-unwind-tables -fno-delete-null-pointer-checks -fstack-protector-strong -fno-var-tracking-assignments -fno-strict-overflow -fno-merge-all-constants -fmerge-constants -fstack-check=no -fconserve-stack --param allow-store-data-races=0
<11> DW_AT_language : 1 (ANSI C)
<12> DW_AT_name : (indirect string, offset: 0x10daa): /home/acme/git/linux/net/ipv4/tcp.c
<16> DW_AT_comp_dir : (indirect string, offset: 0x1d8c5): /home/acme/git/build/v5.0+
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
ea83b780ec |
pfunct: Handle unnamed struct typedefs
If we first reconstruct all the types needed for a typedef to then
reconstruct it, we may end up with the unnamed struct correctly
reconstructed and then this:
tcp.pahole.c:1987:17: warning: useless storage class specifier in empty declaration
typedef struct read_descriptor_t;
I.e.:
typedef struct {
size_t written; /* 0 8 */
size_t count; /* 8 8 */
union {
char * buf; /* 16 8 */
void * data; /* 16 8 */
} arg; /* 16 8 */
int error; /* 24 4 */
/* size: 32, cachelines: 1, members: 4 */
/* padding: 4 */
/* last cacheline: 32 bytes */
} read_descriptor_t;
typedef struct read_descriptor_t;
So special case it.
XXX
I'll revisit this, looks suboptimal, we manage to get this right
when reconstructing nameless struct typedefs found as the types for
members of structs or unions...
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
e7ebc05d12 |
emit: Unwind the definitions for typedefs in type__emit_definitions()
For instance, with the existing code we ended up with:
typedef __kernel_size_t size_t;
size_t tcp_opt_stats_get_size(void)
{
}
Which lacks unwinding what a __kernel_size_t is, i.e.
type__emit_definitions() was only emitting definitions for the members
of structs and unions, do it for typedefs too, and then we end up what
we need, which is:
typedef long unsigned int __kernel_ulong_t;
typedef __kernel_ulong_t __kernel_size_t;
typedef __kernel_size_t size_t;
size_t tcp_opt_stats_get_size(void)
{
}
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
093135b0bf |
pfunct: Do not emit a type multiple times
We need to check if tag__type(type)->definition_emitted is set before asking for that type to be emitted, otherwise we get type redefinition errors when trying to compile the output from pahole --expand_types. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
3ce2c52166 |
pfunct: Ask for generating compilable output that generates DWARF for types
I.e. --compile is similar to --expand_types but also makes sure the function have empty function bodies, which ends up making at least gcc to generate the DWARF info for the types referenced by the function arguments and return types. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
e7a786540d |
pfunct: Make --expand_types/-b without -f expand types for all functions
Previously we wouldn't expand types if the user didn't provide a function name, make it expand the types for all function arguments if no function name is provided. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
9b2eadf97b |
pfunct: Follow const, restrict, volatile in --expand_types
Take: int tcp_md5_hash_key(struct tcp_md5sig_pool *hp, const struct tcp_md5sig_key *key) We were not generating the 'struct tcp_md5sig_key' forward decl or struct definition, stopping at 'const', which made this uncompilable. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
f3f86f2f89 |
pfunct: Reconstruct function return types for --expand_types
Not just for the function arguments, so that we are able to get something closer to buildable. So far we got it buildable when the return types were reconstructed because they also appeared in one of the function arguments or in structs used by them. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
a7d9c58cb8 |
fprintf: Add missing closing parens to the align attribute
Noticed while trying to use pfunct's -b option, that will show a
function prototype + the types it uses in its function signature, i.e.:
$ pfunct -b -f tcp_sendmsg tcp.o
typedef long long unsigned int __u64;
typedef __u64 __addrpair;
typedef unsigned int __u32;
typedef __u32 __be32;
typedef short unsigned int __u16;
typedef __u32 __portpair;
typedef __u16 __be16;
struct hlist_node {
struct hlist_node * next; /* 0 8 */
struct hlist_node * * pprev; /* 8 8 */
/* size: 16, cachelines: 1, members: 2 */
/* last cacheline: 16 bytes */
};
<SNIP tons of types>
struct sock {
struct sock_common __sk_common; /* 0 136 */
/* --- cacheline 2 boundary (128 bytes) was 8 bytes ago --- */
socket_lock_t sk_lock; /* 136 32 */
atomic_t sk_drops; /* 168 4 */
/* --- cacheline 10 boundary (640 bytes) --- */
<SNIP the rest of the 'struct sock' members>
struct sock_cgroup_data sk_cgrp_data; /* 640 8 */
struct mem_cgroup * sk_memcg; /* 648 8 */
void (*sk_state_change)(struct sock *); /* 656 8 */
void (*sk_data_ready)(struct sock *); /* 664 8 */
void (*sk_write_space)(struct sock *); /* 672 8 */
void (*sk_error_report)(struct sock *); /* 680 8 */
int (*sk_backlog_rcv)(struct sock *, struct sk_buff *); /* 688 8 */
void (*sk_destruct)(struct sock *); /* 696 8 */
/* --- cacheline 11 boundary (704 bytes) --- */
struct sock_reuseport * sk_reuseport_cb; /* 704 8 */
struct callback_head sk_rcu __attribute__((__aligned__(8))); /* 712 16 */
/* size: 728, cachelines: 12, members: 84 */
/* sum members: 715, holes: 4, sum holes: 8 */
/* sum bitfield members: 40 bits (5 bytes) */
/* paddings: 1, sum paddings: 4 */
/* forced alignments: 1 */
/* last cacheline: 24 bytes */
};
<SNIP some more types>
struct kiocb;
struct msghdr {
void * msg_name; /* 0 8 */
int msg_namelen; /* 8 4 */
/* XXX 4 bytes hole, try to pack */
struct iov_iter msg_iter; /* 16 40 */
void * msg_control; /* 56 8 */
/* --- cacheline 1 boundary (64 bytes) --- */
__kernel_size_t msg_controllen; /* 64 8 */
unsigned int msg_flags; /* 72 4 */
/* XXX 4 bytes hole, try to pack */
struct kiocb * msg_iocb; /* 80 8 */
/* size: 88, cachelines: 2, members: 7 */
/* sum members: 80, holes: 2, sum holes: 8 */
/* last cacheline: 24 bytes */
};
typedef __kernel_size_t size_t;
int tcp_sendmsg(struct sock * sk, struct msghdr * msg, size_t size);
$
So if we then redirect the output to a file and if we make it a empty
function instead of a prototype, i.e. if we make the last line above to
become this:
int tcp_sendmsg(struct sock * sk, struct msghdr * msg, size_t size) {}
then build with gcc -g to have it build as a .o with DWARF info, then we
should be able to see if the struct rebuilt from DWARF matches the
original struct used to generate the DWARF, going full circle:
$ pfunct -b -f tcp_sendmsg tcp.o > tcp_sendmsg_types.c
$ gcc -c tcp_sendmsg_types.c -g
$ file tcp_sendmsg_types.o
tcp_sendmsg_types.o: ELF 64-bit LSB relocatable, x86-64, version 1 (SYSV), with debug_info, not stripped
$ pahole -E -C sock tcp_sendmsg_types.o > tcp_sendmsg_types.o.pahole
$ pahole -E -C sock tcp.o > tcp.o.pahole
$ diff -u tcp_sendmsg_types.o.pahole tcp.o.pahole
$ wc -l tcp_sendmsg_types.o.pahole
420 tcp_sendmsg_types.o.pahole
$
So all the types that come from sock are expanded and all its details
are reconstructed in the same way for both cases.
$ pahole -C sock tcp.o | tail
struct sock_reuseport * sk_reuseport_cb; /* 704 8 */
struct callback_head sk_rcu __attribute__((__aligned__(8))); /* 712 16 */
/* size: 728, cachelines: 12, members: 84 */
/* sum members: 715, holes: 4, sum holes: 8 */
/* sum bitfield members: 40 bits (5 bytes) */
/* paddings: 1, sum paddings: 4 */
/* forced alignments: 1 */
/* last cacheline: 24 bytes */
};
$ pahole -C sock tcp_sendmsg_types.o | tail
struct sock_reuseport * sk_reuseport_cb; /* 704 8 */
struct callback_head sk_rcu __attribute__((__aligned__(8))); /* 712 16 */
/* size: 728, cachelines: 12, members: 84 */
/* sum members: 715, holes: 4, sum holes: 8 */
/* sum bitfield members: 40 bits (5 bytes) */
/* paddings: 1, sum paddings: 4 */
/* forced alignments: 1 */
/* last cacheline: 24 bytes */
};
$
Reported-by: Jiri Olsa <jolsa@kernel.org>
Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Yonghong Song <yhs@fb.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
d83d9f578f |
dwarf_loader: Handle DW_TAG_label in inline expansions
Just add it to the current lexblock.
This removes the warnings that started to appear with the fix in the
previous patch, i.e. these:
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3aefafe> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3aeff8a> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af02e0> not handled!
Reported-by: Jiri Olsa <jolsa@kernel.org>
Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Yonghong Song <yhs@fb.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
73e545b144 |
dwarf_loader: Handle unsupported_tag in die__process_inline_expansion
It calls die__process_tag() that can return &unsupported_tag, for
instance for DW_TAG_label, as reported by Jiri, so handle that, which
ends up causing these warnings to appear in pahole:
struct userfaultfd_wake_range {
long unsigned int start; /* 0 8 */
long unsigned int len; /* 8 8 */
/* size: 16, cachelines: 1, members: 2 */
/* last cacheline: 16 bytes */
};
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3aefafe> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3aeff8a> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af02e0> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af1903> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af19bf> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af358d> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af3e87> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af4268> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af46ec> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af4bd3> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af4f8c> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af551e> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af5815> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af5c15> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af5cad> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af79e0> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af7b34> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af7df3> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af806e> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3af812c> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3afa8da> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3afaba5> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3afb22e> not handled!
die__process_inline_expansion: DW_TAG_label (0xa) @ <0x3afb45e> not handled!
struct kioctx_table {
struct callback_head rcu __attribute__((__aligned__(8)); /* 0 16 */
unsigned int nr; /* 16 4 */
/* XXX 4 bytes hole, try to pack */
struct kioctx * table[0]; /* 24 0 */
/* size: 24, cachelines: 1, members: 3 */
/* sum members: 20, holes: 1, sum holes: 4 */
/* forced alignments: 1 */
/* last cacheline: 24 bytes */
};
^C
[acme@quaco pahole]$
But at least no segfault takes place. Next csets should take care of it
more properly.
Reported-by: Jiri Olsa <jolsa@kernel.org>
Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Yonghong Song <yhs@fb.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Andrii Nakryiko
|
fe590758cb |
class__find_holes: Zero out bit_hole/hole on member
pahole --reorganize is calling class__find_holes() multiple times on the
same struct to re-calculate holes. If it so happens that after reorg
last struct's member had hole previously, we are not going to clear it
out, which will lead to weird output and BFA, like this:
$ pahole -F btf --reorganize -C netns_frags ~/tmp/vmlinux-default
struct netns_frags {
long int high_thresh; /* 0 8 */
long int low_thresh; /* 8 8 */
int timeout; /* 16 4 */
int max_dist; /* 20 4 */
struct inet_frags * f; /* 24 8 */
atomic_long_t mem; /* 32 8 */
/* XXX 24 bytes hole, try to pack */
/* --- cacheline 1 boundary (64 bytes) --- */
struct rhashtable rhashtable; /* 64 136 */
/* XXX 56 bytes hole, try to pack */
/* size: 200, cachelines: 4, members: 7 */
/* sum members: 176, holes: 1, sum holes: 80 */
/* last cacheline: 8 bytes */
/* BRAIN FART ALERT! 200 bytes != 176 (member bytes) + 0 (member bits) + 80 (byte holes) + 0 (bit holes), diff = -448 bits */
}; /* saved 120 bytes and 1 cacheline! */
After this change:
$ pahole -F btf --reorganize -C netns_frags ~/tmp/vmlinux-defaultstruct netns_frags {
long int high_thresh; /* 0 8 */
long int low_thresh; /* 8 8 */
int timeout; /* 16 4 */
int max_dist; /* 20 4 */
struct inet_frags * f; /* 24 8 */
atomic_long_t mem; /* 32 8 */
/* XXX 24 bytes hole, try to pack */
/* --- cacheline 1 boundary (64 bytes) --- */
struct rhashtable rhashtable; /* 64 136 */
/* size: 200, cachelines: 4, members: 7 */
/* sum members: 176, holes: 1, sum holes: 24 */
/* last cacheline: 8 bytes */
}; /* saved 120 bytes and 1 cacheline! */
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Yonghong Song <yhs@fb.com>
Cc: dwarves@vger.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
863c2af6e9 |
reorganize: Disable the bitfield coalescing/moving steps
We need to fix some bugs introduced recently, till then, disable steps that try to demote the base type of bitfields and those that move/combine bitfields to save space. We'll revisit those later, bringing them back to the reorg codebase. Acked-by: Andrii Nakryiko <andriin@fb.com> Cc: Alexei Starovoitov <ast@fb.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Mark Wielaard <mark@klomp.org> Cc: Yonghong Song <yhs@fb.com> Cc: Daniel Borkmann <daniel@iogearbox.net># Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
b95961db69 |
fprintf: Show statistics about holes due to forced alignments
$ pahole -C task_struct | tail
/* --- cacheline 104 boundary (6656 bytes) --- */
struct thread_struct thread __attribute__((__aligned__(64)); /* 6656 4352 */
/* size: 11008, cachelines: 172, members: 207 */
/* sum members: 10902, holes: 16, sum holes: 98 */
/* sum bitfield members: 10 bits, bit holes: 2, sum bit holes: 54 bits */
/* paddings: 3, sum paddings: 14 */
/* forced alignments: 6, forced holes: 1, sum forced holes: 40 */
};
$ pahole -C inet_timewait_death_row
struct inet_timewait_death_row {
atomic_t tw_count; /* 0 4 */
/* XXX 60 bytes hole, try to pack */
/* --- cacheline 1 boundary (64 bytes) --- */
struct inet_hashinfo * hashinfo __attribute__((__aligned__(64)); /* 64 8 */
int sysctl_max_tw_buckets; /* 72 4 */
/* size: 128, cachelines: 2, members: 3 */
/* sum members: 16, holes: 1, sum holes: 60 */
/* padding: 52 */
/* forced alignments: 1, forced holes: 1, sum forced holes: 60 */
};
$
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Mark Wielaard <mark@klomp.org>
Cc: Yonghong Song <yhs@fb.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
ec772f21f6 |
fprintf: Show the number of forced alignments in a class
For instance, for task_struct: $ pahole -C task_struct | tail /* --- cacheline 104 boundary (6656 bytes) --- */ struct thread_struct thread __attribute__((__aligned__(64)); /* 6656 4352 */ /* size: 11008, cachelines: 172, members: 207 */ /* sum members: 10902, holes: 16, sum holes: 98 */ /* sum bitfield members: 10 bits, bit holes: 2, sum bit holes: 54 bits */ /* paddings: 3, sum paddings: 14 */ /* forced alignments: 6 */ }; $ Cc: Alexei Starovoitov <ast@fb.com> Cc: Andrii Nakryiko <andriin@fb.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Mark Wielaard <mark@klomp.org> Cc: Yonghong Song <yhs@fb.com> Cc: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |
||
|
Arnaldo Carvalho de Melo
|
52d1c75ea4 |
btfdiff: Use --suppress_aligned_attribute with -F dwarf
Now that we collect DWARF5's DW_AT_alignent, btdiff shows, for instance:
$ btfdiff examples/tcp.o
<SNIP>
@@ -13450,7 +13450,7 @@ struct ip6_flowlabel {
struct in6_addr dst; /* 16 16 */
struct ipv6_txoptions * opt; /* 32 8 */
long unsigned int linger; /* 40 8 */
- struct callback_head rcu __attribute__((__aligned__(8)); /* 48 16 */
+ struct callback_head rcu; /* 48 16 */
/* --- cacheline 1 boundary (64 bytes) --- */
u8 share; /* 64 1 */
@@ -13616,7 +13616,7 @@ struct fib6_node {
__u16 fn_flags; /* 42 2 */
int fn_sernum; /* 44 4 */
struct fib6_info * rr_ptr; /* 48 8 */
- struct callback_head rcu __attribute__((__aligned__(8)); /* 56 16 */
+ struct callback_head rcu; /* 56 16 */
/* size: 72, cachelines: 2, members: 10 */
/* last cacheline: 8 bytes */
$
So ask for those attributes to be suppressed when comparing BTF and
DWARF output.
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Mark Wielaard <mark@klomp.org>
Cc: Yonghong Song <yhs@fb.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
||
|
Arnaldo Carvalho de Melo
|
6cd6a6bd87 |
dwarves_fprintf: Allow suppressing the __attribute__((__aligned__(N))
So that we can use it in things like btfdiff. Cc: Alexei Starovoitov <ast@fb.com> Cc: Andrii Nakryiko <andriin@fb.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Mark Wielaard <mark@klomp.org> Cc: Yonghong Song <yhs@fb.com> Cc: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> |