Compare commits

...

1 Commits

Author SHA1 Message Date
Nick Alcock
bf69a386a7 libctf: link: rejig lazy opening to not need weak symbols
The ctf_link_add_ctf API function has a 'lazy opening' feature whereby,
if you pass in the file but not a CTF archive, the archive is opened
as late as possible during links.  This is valuable mostly in
cu-mapped links (a feature not accessible via GNU ld), where it
ensures that, rather than eventually needing memory for the original
link inputs, the smushed-together cu-mapped intermediate outputs,
*and* the final output, we only need enough memory for the smushed-
together outputs, the final output, and one input, since the inputs
can be closed immediately after they are smushed together.

(In GNU ld, the feature is useless because it loads all sections into
memory anyway.)

The lazy-opening feature uses libctf's ctf_open function, which uses
BFD: so it is not available in libctf-nobfd -- except that I thought I
had a cunning trick, and used a weak symbol so that if you linked
libctf-nobfd into your program and then also linked in bfd, the feature
stayed enabled.

This is silly -- if your program is licensed such that you can link in
BFD, you can just link in libctf.so and not bother with libctf-nobfd.so
in the first place.  Worse, the weak symbol usage broke MacOS builds,
since MacOS's system compiler uses a different means of introducing weak
symbols.  We could test for and use it, but this is the only place in
libctf to use weak symbols at all, and the feature of lazy-opening with
libctf-nobfd is so marginal we might as well drop it: it's almost
certain there are zero users, certainly fewer users than users of MacOS
with the system compiler.

While we're at it, simplify things by deleting the never-implemented
feature (not exposed in the API) to allow linking together raw buffers
of CTF data.  If we need it we can bring it back, but all it's doing
right now is complicating the code to no end at all.

libctf/
	* ctf-link.c (ctf_open): Delete weak pragma.
	(ctf_link_add): Fuse with...
	(ctf_link_add_ctf): ... this function.  Drop BUF, N args
	and corresponding unimplemented feature warnings.  Only check
	NOBFD to see whether lazy loading is available, not PIC as
	well.
	(ctf_link_lazy_open): Likewise.
2025-07-23 14:20:26 +01:00

View File

@@ -20,10 +20,6 @@
#include <ctf-impl.h>
#include <string.h>
#if defined (PIC)
#pragma weak ctf_open
#endif
/* CTF linking consists of adding CTF archives full of content to be merged into
this one to the current file (which must be writable) by calling
ctf_link_add_ctf. Once this is done, a call to ctf_link will merge the type
@@ -145,56 +141,33 @@ ctf_link_add_ctf_internal (ctf_dict_t *fp, ctf_archive_t *ctf,
return ctf_set_errno (fp, ENOMEM);
}
/* Add a file, memory buffer, or unopened file (by name) to a link.
/* Add an opened CTF archive or unopened file (by name) to a link.
If CTF is NULL and NAME is non-null, an unopened file is meant:
otherwise, the specified archive is assumed to have the given NAME.
You can call this with:
If CTF is NULL, the NAME is only opened when needed, and is closed when no
longer needed, so that large cu-mapped links will only use memory for their
cu-mapped inputs briefly (compensating for the memory usage of the
smushed-together cu-mapped verion).
CTF and NAME: link the passed ctf_archive_t, with the given NAME.
NAME alone: open NAME as a CTF file when needed.
BUF and NAME: open the BUF (of length N) as CTF, with the given NAME. (Not
yet implemented.)
Passed in CTF args are owned by the dictionary and will be freed by it.
Passed in CTF args are owned by the dictionary and will be freed by it.
The BUF arg is *not* owned by the dictionary, and the user should not free
its referent until the link is done.
The order of calls to this function influences the order of types in the
final link output, but otherwise is not important.
The order of calls to this function influences the order of types in the
final link output, but otherwise is not important.
Repeated additions of the same NAME have no effect; repeated additions of
different dicts with the same NAME add all the dicts with unique NAMEs
derived from NAME. */
Repeated additions of the same NAME have no effect; repeated additions of
different dicts with the same NAME add all the dicts with unique NAMEs
derived from NAME.
Private for now, but may in time become public once support for BUF is
implemented. */
static int
ctf_link_add (ctf_dict_t *fp, ctf_archive_t *ctf, const char *name,
void *buf _libctf_unused_, size_t n _libctf_unused_)
int
ctf_link_add_ctf (ctf_dict_t *fp, ctf_archive_t *ctf, const char *name)
{
if (buf)
return (ctf_set_errno (fp, ECTF_NOTYET));
if (!((ctf && name && !buf)
|| (name && !buf && !ctf)
|| (buf && name && !ctf)))
if (!name)
return (ctf_set_errno (fp, EINVAL));
/* We can only lazily open files if libctf.so is in use rather than
libctf-nobfd.so. This is a little tricky: in shared libraries, we can use
a weak symbol so that -lctf -lctf-nobfd works, but in static libraries we
must distinguish between the two libraries explicitly. */
#if defined (PIC)
if (!buf && !ctf && name && !ctf_open)
return (ctf_set_errno (fp, ECTF_NEEDSBFD));
#elif NOBFD
if (!buf && !ctf && name)
return (ctf_set_errno (fp, ECTF_NEEDSBFD));
#endif
if (fp->ctf_link_outputs)
return (ctf_set_errno (fp, ECTF_LINKADDEDLATE));
if (fp->ctf_link_inputs == NULL)
fp->ctf_link_inputs = ctf_dynhash_create (ctf_hash_string,
ctf_hash_eq_string, free,
@@ -203,24 +176,17 @@ ctf_link_add (ctf_dict_t *fp, ctf_archive_t *ctf, const char *name,
if (fp->ctf_link_inputs == NULL)
return (ctf_set_errno (fp, ENOMEM));
/* We can only lazily open files if libctf.so is in use rather than
libctf-nobfd.so. */
#if NOBFD
if (!ctf)
return (ctf_set_errno (fp, ECTF_NEEDSBFD));
#endif
return ctf_link_add_ctf_internal (fp, ctf, NULL, name);
}
/* Add an opened CTF archive or unopened file (by name) to a link.
If CTF is NULL and NAME is non-null, an unopened file is meant:
otherwise, the specified archive is assumed to have the given NAME.
Passed in CTF args are owned by the dictionary and will be freed by it.
The order of calls to this function influences the order of types in the
final link output, but otherwise is not important. */
int
ctf_link_add_ctf (ctf_dict_t *fp, ctf_archive_t *ctf, const char *name)
{
return ctf_link_add (fp, ctf, name, NULL, 0);
}
/* Lazily open a CTF archive for linking, if not already open.
Returns the number of files contained within the opened archive (0 for none),
@@ -238,12 +204,12 @@ ctf_link_lazy_open (ctf_dict_t *fp, ctf_link_input_t *input)
return 1;
/* See ctf_link_add_ctf. */
#if defined (PIC) || !NOBFD
input->clin_arc = ctf_open (input->clin_filename, NULL, &err);
#else
#if NOBFD
ctf_err_warn (fp, 0, ECTF_NEEDSBFD, _("cannot open %s lazily"),
input->clin_filename);
return ctf_set_errno (fp, ECTF_NEEDSBFD);
#else
input->clin_arc = ctf_open (input->clin_filename, NULL, &err);
#endif
/* Having no CTF sections is not an error. We just don't need to do