include, libctf, ld: extend variable section to contain functions too

The CTF variable section is an optional (usually-not-present) section in
the CTF dict which contains name -> type mappings corresponding to data
symbols that are present in the linker input but not in the output
symbol table: the idea is that programs that use their own symbol-
resolution mechanisms can use this section to look up the types of
symbols they have found using their own mechanism.

Because these removed symbols (mostly static variables, functions, etc)
all have names that are unlikely to appear in the ELF symtab and because
very few programs have their own symbol-resolution mechanisms, a special
linker flag (--ctf-variables) is needed to emit this section.

Historically, we emitted only removed data symbols into the variable
section.  This seemed to make sense at the time, but in hindsight it
really doesn't: functions are symbols too, and a C program can look them
up just like any other type.  So extend the variable section so that it
contains all static function symbols too (if it is emitted at all), with
types of kind CTF_K_FUNCTION.

This is a little fiddly.  We relied on compiler assistance for data
symbols: the compiler simply emits all data symbols twice, once into the
symtypetab as an indexed symbol and once into the variable section.

Rather than wait for a suitably adjusted compiler that does the same for
function symbols, we can pluck unreported function symbols out of the
symtab and add them to the variable section ourselves.  While we're at
it, we do the same with data symbols: this is redundant right now
because the compiler does it, but it costs very little time and lets the
compiler drop this kludge and save a little space in .o files.

include/
	* ctf.h: Mention the new things we can see in the variable
	section.

ld/
	* testsuite/ld-ctf/data-func-conflicted-vars.d: New test.

libctf/
	* ctf-link.c (ctf_link_deduplicating_variables): Duplicate
	symbols into the variable section too.
	* ctf-serialize.c (symtypetab_delete_nonstatic_vars): Rename
	to...
	(symtypetab_delete_nonstatics): ... this.  Check the funchash
	when pruning redundant variables.
	(ctf_symtypetab_sect_sizes): Adjust accordingly.
	* NEWS: Describe this change.
This commit is contained in:
Nick Alcock 2022-03-16 15:29:25 +00:00
parent a12c988767
commit 203bfa2f6b
5 changed files with 130 additions and 16 deletions

View File

@ -89,13 +89,13 @@ extern "C"
entries and reorder them accordingly (dropping the indexes in the process).
Variable records (as distinct from data objects) provide a modicum of support
for non-ELF systems, mapping a variable name to a CTF type ID. The variable
names are sorted into ASCIIbetical order, permitting binary searching. We do
not define how the consumer maps these variable names to addresses or
for non-ELF systems, mapping a variable or function name to a CTF type ID.
The names are sorted into ASCIIbetical order, permitting binary searching.
We do not define how the consumer maps these variable names to addresses or
anything else, or indeed what these names represent: they might be names
looked up at runtime via dlsym() or names extracted at runtime by a debugger
or anything else the consumer likes. Variable records with identically-
named entries in the data object section are removed.
named entries in the data object or function index section are removed.
The data types section is a list of variable size records that represent each
type, in order by their ID. The types themselves form a directed graph,

View File

@ -0,0 +1,69 @@
#as:
#source: data-func-1.c
#source: data-func-2.c
#objdump: --ctf
#ld: -shared -s --ctf-variables
#name: Conflicted data syms, partially indexed, stripped, with variables
.*: +file format .*
Contents of CTF section \.ctf:
Header:
Magic number: 0xdff2
Version: 4 \(CTF_VERSION_3\)
#...
Data object section: .* \(0x[1-9a-f][0-9a-f]* bytes\)
Function info section: .* \(0x[1-9a-f][0-9a-f]* bytes\)
Object index section: .* \(0xc bytes\)
Variable section: .* \(0x10 bytes\)
Type section: .* \(0x118 bytes\)
String section: .*
#...
Data objects:
bar -> 0x[0-9a-f]*: \(kind 6\) struct var_3 \(size 0x[0-9a-f]*\) \(aligned at 0x[0-9a-f]*\)
var_1 -> 0x[0-9a-f]*: \(kind 10\) foo_t \(size 0x[0-9a-f]*\) \(aligned at 0x[0-9a-f]*\) -> .*
var_666 -> 0x[0-9a-f]*: \(kind 3\) foo_t \* \(size 0x[0-9a-f]*\) \(aligned at 0x[0-9a-f]*\) -> .*
Function objects:
func_[0-9]* -> 0x[0-9a-f]*: \(kind 5\) void \*\(\*\) \(const char \*restrict, int \(\*\)\(\*\) \(const char \*\)\) \(aligned at 0x[0-9a-f]*\)
#...
Variables:
funcs -> .*
other_func -> .*
#...
Types:
#...
.*: \(kind 6\) struct var_3 .*
#...
CTF archive member: .*/data-func-1\.c:
Header:
Magic number: 0xdff2
Version: 4 \(CTF_VERSION_3\)
#...
Parent name: \.ctf
Compilation unit name: .*/data-func-1\.c
Data object section: .* \(0x[1-9a-f][0-9a-f]* bytes\)
Type section: .* \(0xc bytes\)
String section: .*
Labels:
Data objects:
var_[0-9]* -> 0x80000001*: \(kind 10\) foo_t \(size 0x[0-9a-f]*\) \(aligned at 0x[0-9a-f]*\) -> .*
var_[0-9]* -> 0x80000001*: \(kind 10\) foo_t \(size 0x[0-9a-f]*\) \(aligned at 0x[0-9a-f]*\) -> .*
var_[0-9]* -> 0x80000001*: \(kind 10\) foo_t \(size 0x[0-9a-f]*\) \(aligned at 0x[0-9a-f]*\) -> .*
var_[0-9]* -> 0x80000001*: \(kind 10\) foo_t \(size 0x[0-9a-f]*\) \(aligned at 0x[0-9a-f]*\) -> .*
var_[0-9]* -> 0x80000001*: \(kind 10\) foo_t \(size 0x[0-9a-f]*\) \(aligned at 0x[0-9a-f]*\) -> .*
var_[0-9]* -> 0x80000001*: \(kind 10\) foo_t \(size 0x[0-9a-f]*\) \(aligned at 0x[0-9a-f]*\) -> .*
var_[0-9]* -> 0x80000001*: \(kind 10\) foo_t \(size 0x[0-9a-f]*\) \(aligned at 0x[0-9a-f]*\) -> .*
var_[0-9]* -> 0x80000001*: \(kind 10\) foo_t \(size 0x[0-9a-f]*\) \(aligned at 0x[0-9a-f]*\) -> .*
#...
Function objects:
Variables:
Types:
0x80000001: \(kind 10\) foo_t .* -> .* int .*
#...

View File

@ -1,5 +1,14 @@
-*- text -*-
Changes in 2.39:
* New features
** The CTF variable section (if generated via ld --ctf-variables) now contains
entries for static functions, hidden functions, and other functions with
no associated symbol. The associated type is of kind CTF_K_FUNCTION.
(No change if --ctf-variables is not specified, which is the default.)
Changes in 2.37:
* New features

View File

@ -807,7 +807,12 @@ ctf_link_deduplicating_close_inputs (ctf_dict_t *fp, ctf_dynhash_t *cu_names,
return 0;
}
/* Do a deduplicating link of all variables in the inputs. */
/* Do a deduplicating link of all variables in the inputs.
Also, if we are not omitting the variable section, integrate all symbols from
the symtypetabs into the variable section too. (Duplication with the
symtypetab section in the output will be eliminated at serialization time.) */
static int
ctf_link_deduplicating_variables (ctf_dict_t *fp, ctf_dict_t **inputs,
size_t ninputs, int cu_mapped)
@ -820,6 +825,8 @@ ctf_link_deduplicating_variables (ctf_dict_t *fp, ctf_dict_t **inputs,
ctf_id_t type;
const char *name;
/* First the variables on the inputs. */
while ((type = ctf_variable_next (inputs[i], &it, &name)) != CTF_ERR)
{
if (ctf_link_one_variable (fp, inputs[i], name, type, cu_mapped) < 0)
@ -830,6 +837,34 @@ ctf_link_deduplicating_variables (ctf_dict_t *fp, ctf_dict_t **inputs,
}
if (ctf_errno (inputs[i]) != ECTF_NEXT_END)
return ctf_set_errno (fp, ctf_errno (inputs[i]));
/* Next the symbols. We integrate data symbols even though the compiler
is currently doing the same, to allow the compiler to stop in
future. */
while ((type = ctf_symbol_next (inputs[i], &it, &name, 0)) != CTF_ERR)
{
if (ctf_link_one_variable (fp, inputs[i], name, type, 1) < 0)
{
ctf_next_destroy (it);
return -1; /* errno is set for us. */
}
}
if (ctf_errno (inputs[i]) != ECTF_NEXT_END)
return ctf_set_errno (fp, ctf_errno (inputs[i]));
/* Finally the function symbols. */
while ((type = ctf_symbol_next (inputs[i], &it, &name, 1)) != CTF_ERR)
{
if (ctf_link_one_variable (fp, inputs[i], name, type, 1) < 0)
{
ctf_next_destroy (it);
return -1; /* errno is set for us. */
}
}
if (ctf_errno (inputs[i]) != ECTF_NEXT_END)
return ctf_set_errno (fp, ctf_errno (inputs[i]));
}
return 0;
}

View File

@ -431,12 +431,12 @@ emit_symtypetab_index (ctf_dict_t *fp, ctf_dict_t *symfp, uint32_t *dp,
return 0;
}
/* Delete data symbols that have been assigned names from the variable section.
Must be called from within ctf_serialize, because that is the only place
you can safely delete variables without messing up ctf_rollback. */
/* Delete symbols that have been assigned names from the variable section. Must
be called from within ctf_serialize, because that is the only place you can
safely delete variables without messing up ctf_rollback. */
static int
symtypetab_delete_nonstatic_vars (ctf_dict_t *fp, ctf_dict_t *symfp)
symtypetab_delete_nonstatics (ctf_dict_t *fp, ctf_dict_t *symfp)
{
ctf_dvdef_t *dvd, *nvd;
ctf_id_t type;
@ -445,8 +445,10 @@ symtypetab_delete_nonstatic_vars (ctf_dict_t *fp, ctf_dict_t *symfp)
{
nvd = ctf_list_next (dvd);
if (((type = (ctf_id_t) (uintptr_t)
ctf_dynhash_lookup (fp->ctf_objthash, dvd->dvd_name)) > 0)
if ((((type = (ctf_id_t) (uintptr_t)
ctf_dynhash_lookup (fp->ctf_objthash, dvd->dvd_name)) > 0)
|| (type = (ctf_id_t) (uintptr_t)
ctf_dynhash_lookup (fp->ctf_funchash, dvd->dvd_name)) > 0)
&& ctf_dynhash_lookup (symfp->ctf_dynsyms, dvd->dvd_name) != NULL
&& type == dvd->dvd_type)
ctf_dvd_delete (fp, dvd);
@ -560,13 +562,12 @@ ctf_symtypetab_sect_sizes (ctf_dict_t *fp, emit_symtypetab_state_t *s,
/* If we are filtering symbols out, those symbols that the linker has not
reported have now been removed from the ctf_objthash and ctf_funchash.
Delete entries from the variable section that duplicate newly-added data
symbols. There's no need to migrate new ones in, because the compiler
always emits both a variable and a data symbol simultaneously, and
filtering only happens at final link time. */
Delete entries from the variable section that duplicate newly-added
symbols. There's no need to migrate new ones in: we do that (if necessary)
in ctf_link_deduplicating_variables. */
if (s->filter_syms && s->symfp->ctf_dynsyms &&
symtypetab_delete_nonstatic_vars (fp, s->symfp) < 0)
symtypetab_delete_nonstatics (fp, s->symfp) < 0)
return -1;
return 0;