273d2b19b4
Hm, the const isn't going to work.
2002-09-26 13:43:52 +00:00
00088c228c
Do not attempt to free(_rtld_objself.path) -- it's a pointer into our data
...
segment.
2002-09-26 13:42:05 +00:00
f4f740de3a
By default, compile without DEBUG, and with -fomit-frame-pointer.
2002-09-26 12:59:29 +00:00
15025431b4
Update a comment.
2002-09-26 03:25:29 +00:00
fd7e0d92f3
Add a rdbg() in _rtld_bind().
2002-09-26 02:52:05 +00:00
723a5f8a50
Remove a meaningless bit from a rdbg().
2002-09-26 02:51:27 +00:00
ee398b4d07
Remove vestigal comments about _rtld_bind_start_0() -- it works now.
2002-09-26 02:25:57 +00:00
26043b23b0
Make room for the argument save area in _rtld_start(). The PLT stub already
...
made room from _rtld_bind_start(), but add a comment explaining the instruction
it uses.
2002-09-26 02:14:10 +00:00
f1c792d648
Make the stack frame 96 bytes, like GCC does.
2002-09-26 02:05:41 +00:00
bde7a965da
Tweak the save instruction in the PLT stub rather than adding an instruction
...
to the binder entry stub...
2002-09-26 01:59:16 +00:00
d314d82b47
Allocate space for the argument space area in both _rtld_start() and
...
_rtld_bind_start(). Not needed normally, but if we compile ld.elf_so with
-g, it blows up without this.
2002-09-26 01:54:52 +00:00
6bff9ffc3d
For consistency, use Elf_Word* to point to code, not Elf_Addr*.
2002-09-25 22:33:54 +00:00
0079455251
Add a comment explaining some of the magick.
2002-09-25 22:29:12 +00:00
f6db390b5a
Fix a bug that caused PLT entries to effectively not be patched.
...
Also, fix more bugs in the large-PLT case -- my test program works now.
2002-09-25 22:25:11 +00:00
2997d64776
Make this work with PLTs larger than 8192 entries (64KB). Not tested
...
exhaustively, but it works in some simple cases.
2002-09-25 21:11:18 +00:00
d4e6f3ff08
Swap registers %o0 and %o1 in the binder stub to say one instruction.
...
Fix an obvious bug in the 64-bit PLT fixup: the SLLX was by 12 bits, when it
should be 32.
Fix what *appear* to be two bugs in the >32768 PLT entry stub:
* One division was wrong (/14 rather than /24).
* We need to subtract 1048576 (to make the offset relative to the beginning of
the upper section), not add it.
This path is still untested, and buggy.
2002-09-25 16:35:08 +00:00
9e31b44846
There is no need to save and restore a second register window in the binder
...
stub.
2002-09-25 14:38:51 +00:00
afb1082706
_rtld_bind_start_0_stub() is not actually used.
2002-09-25 14:36:37 +00:00
028dd5150a
Push a multiple from _rtld_bind_start_[01]() to _rtld_bind().
2002-09-25 14:35:39 +00:00
23b2275ee8
Partly fix this port -- still needs a self-reloc function.
2002-09-25 08:37:57 +00:00
a9c999925c
We don't need to save quite as many registers in the stub.
2002-09-25 08:00:26 +00:00
729925dfb9
Push the _rtld_bind() interface into MD code -- it's just a trivial wrapper
...
anyway.
2002-09-25 07:27:49 +00:00
305b0ffce5
#if defined -> #ifdef
2002-09-25 06:43:46 +00:00
896795d8a3
Minor cleanup, for clarity.
2002-09-25 03:57:15 +00:00
a3c903f7cb
Resolve the GOT before doing relocations. Then, when doing relocations, for
...
symbols in the global part of the symbol table, use the updated GOT entry
rather than doing a lookup. (This provides the same effect as `-z combreloc'
on other platforms -- at most one lookup is done per symbol.)
Unfortunately, it is necessary to turn off lazy binding on MIPS. As the
comment says:
* XXX DANGER WILL ROBINSON!
* You might think this is stupid, as it intentionally
* defeats lazy binding -- and you'd be right.
* Unfortunately, for lazy binding to work right, we
* need to a way to force the GOT slots used for
* function pointers to be resolved immediately. This
* is supposed to be done automatically by the linker,
* by not outputting a PLT slot and setting st_value
* to 0, but GNU ld does not do so reliably.
2002-09-25 03:52:06 +00:00
646628d7fa
DOH! Adjust all the stack slots in _rtld_bind_start() so that the saved $a0
...
is not in _rtld_bind_mips()'s $a3 argument save area.
2002-09-24 21:57:20 +00:00
fb3dfbb81b
Add a comment about the function pointer nonsense.
2002-09-24 20:27:07 +00:00
216a20fcc4
Put back the `in_plt' nonsense for now.
...
It turns out there is some deep-seated wackiness WRT function pointer
comparisons...
2002-09-24 20:23:11 +00:00
9d1d2c30c8
Reverse the previous patch, but make the dbg()s more informative.
2002-09-24 18:28:42 +00:00
8510028476
Touch up a little more -- some of the dbg()s never actually printed anything.
2002-09-24 18:25:54 +00:00
1ad20cb78f
Previous change was done wrong.
2002-09-24 18:23:14 +00:00
d0e76e42b6
G/c dbg()'s.
2002-09-24 18:03:46 +00:00
5afbf1cbce
Make sure that dbg() is disabled if LD_DEBUG is not defined. Now
...
debugging is easier with RTLD_DEBUG_RELOC.
2002-09-24 17:51:11 +00:00
64d3d6215a
Ignore JMP_SLOT relocs when doing non-PLT relocations. The relalim trimming
...
hack doesn't always work, because ld(1) sometimes puts the sections in the
wrong order.
2002-09-24 15:04:48 +00:00
bddb0d021d
Similarly, remove support for short arguments.
2002-09-24 14:09:43 +00:00
fa344ccc4b
Remove quad support. We don't need it, and it wastes a lot of space. (It
...
pulls in {,u}{div,mod}di3() and qdivrem().)
2002-09-24 14:05:33 +00:00
56288dc0e5
%ld -> %p
2002-09-24 13:59:31 +00:00
fa5df29420
Include xassert() only if DEBUG, like the real assert().
...
Also, xassert() calls abort(), so there's no need to do it in all of the
callers.
2002-09-24 13:06:07 +00:00
7f44855294
Remove the SVR4_LIBDIR madness.
2002-09-24 12:52:20 +00:00
3c16ffb1dd
Woops, unmap the first page in the `win' case, too.
...
Tidy up a bit.
2002-09-24 12:44:58 +00:00
72b0e512c3
Change the comment about l_offs.
2002-09-24 12:18:32 +00:00
9d8c69b929
Remove vestiges of VARPSZ.
2002-09-24 12:12:07 +00:00
175c20cbcf
Nuke PAGESIZE.
2002-09-24 09:45:59 +00:00
4b4a707c66
Use _rtld_pagesz instead of PAGESIZE.
2002-09-24 09:35:13 +00:00
d190613fbb
space -> tab.
2002-09-24 09:26:43 +00:00
93f2d9023a
Don't need to use union any longer.
2002-09-24 09:22:51 +00:00
272e959468
Remove '\n' at the end of dbg strings.
2002-09-24 08:56:30 +00:00
61e1361da6
END_SYM is no longer used.
2002-09-24 06:43:14 +00:00
97aed9957d
-I${.CURDIR} is done at the top-level, so there's no need to do it here.
2002-09-24 01:59:44 +00:00
b59c269424
Nuke reference to RTLD_RELOCATE_SELF.
2002-09-24 01:58:44 +00:00
c3ece2a45f
Use/require AT_PAGESZ on all platforms. (The kernel has exported it for ~3
...
years now.) Use _rtld_pagesz instead of getpagesize() to determine the page
size in our local malloc(). Saves a system call.
Also, since we're now relocated early, we don't need to be careful to avoid
globals, so most of the VARPSZ hacks are eliminated.
2002-09-24 01:24:44 +00:00
d1351c627c
Um, a little bit of sanity here:
...
l_addr is always supposed to be obj->relocbase -- or so says the GDB code that
uses it. So, set it to this on all platforms. It already was on VAX
explicitly, and on everything else except MIPS implicitly (because
mapbase==relocbase for all existing shlibs). For some silly/stupid reason, a
new field was created that the MIPS GDB currently uses.
Another MD #ifdef bites it.
2002-09-24 00:33:39 +00:00
f1184d192e
Very minor cleanup of the previous.
2002-09-24 00:02:46 +00:00
86103e2f0e
Several small changes that shave 7-8% off the simple-exec-loop test:
...
* Rename _rtld_find_library() to _rtld_load_library(). It now calls
_rtld_load_object() if necessary to actually load the object, rather
than having the caller do it. To do this, it also takes the `mode'
argument that gets passed to _rtld_load_object().
* On a related note, remove _rtld_check_library(), and instead call
_rtld_load_object() to instead try actually loading the object. We
save two extra namei's and a bunch of redundant work (almost
literally the same code) this way.
* In _rtld_map_object(), mmap(2) the first page read-only, rather than
read(2)ing it.
* In _rtld_symlook_obj(), compare the *second* character of the symbol
name before calling strcmp(). (This first character is too
frequently `_', and turns out to not be helpful, in libc.)
* Also in _rtld_symlook_obj(), remove the bogus STT_FUNC special case
-- this also allows removing the `in_plt' argument to
_rtld_symlook_list() and _rtld_symlook_obj().
Also:
* In _rtld_obj_from_addr(), rather than trying to look up `_end' in
the each object, instead use obj->mapsize as the upper bound.
2002-09-23 23:56:46 +00:00
c3a2d8b7ba
Add combreloc code here too.
2002-09-21 17:51:44 +00:00
930998f892
Print the name of dynamic linker actually loaded.
2002-09-21 05:24:17 +00:00
48910dcfc1
Cosmetic changes (long lines).
2002-09-20 03:39:22 +00:00
18139870a5
Cosmetic changes.
2002-09-19 14:05:37 +00:00
6d0f752a65
don't core dump setting progname if argv[0] == NULL. From xs@kittenz.org
2002-09-19 13:59:46 +00:00
a5390eb121
_rtld_setup_pltgot() belongs to reloc.c.
2002-09-18 07:35:40 +00:00
9754fa4d8e
Make immediate binding work again by actually implementing
...
_rtld_relocate_plt_objects(). Note that this is for i386 only;
any efforts to make it work on other platforms are left to
those who are using them...
2002-09-17 07:29:46 +00:00
7e6d181789
No PC24 relocs will ever be unaligned. Pointed out by mycroft.
2002-09-15 01:08:04 +00:00
672853f217
It is possible for reloc entries to point to unaligned locations; handle
...
this.
2002-09-15 00:52:08 +00:00
a2130a3a09
It is possible for reloc entries to point to unaligned locations; handle
...
this.
2002-09-14 23:53:21 +00:00
b01748bb26
It is possible for reloc entries to point to unaligned locations; handle
...
this.
2002-09-14 23:21:13 +00:00
a95d62a38e
In _rtld_bind_mips(), if the symbol is undefined, die.
2002-09-13 20:32:24 +00:00
712b838f43
Kill the extra newline in assert messages.
2002-09-13 19:50:00 +00:00
cc2b1e9844
Add an explanatory comment.
2002-09-13 19:27:19 +00:00
e3db70741f
Turn DEBUG back on by default. It doesn't take much space.
2002-09-13 18:28:58 +00:00
60330f4a69
The sense of the SHN_ABS test has always been reversed. Fortunately nobody
...
ever tried to reference stuff like _DYNAMIC_LINK and _gp_disp in C code...
2002-09-13 17:36:00 +00:00
8f7fa0cfde
Put back the lazy binding. For some bizarre reason, the linker occasionally
...
does not create a PLT entry for a symbol, so we test st_value!=0 to see
whether we can do a PLT fixup.
2002-09-13 17:07:12 +00:00
a8dd8878cf
Er, turn off lazy binding for the moment. It doesn't seem to be fully working.
2002-09-13 16:54:14 +00:00
c1cf8538ec
Simplify the GOT relocation code somewhat.
2002-09-13 16:31:28 +00:00
e8b2fe1d2a
Actually do lazy binding on this platform.
2002-09-13 16:03:20 +00:00
89a39bb54e
The textrel special case is only needed for MIPS (due to a bug in BFD).
2002-09-13 15:27:30 +00:00
7a962f974a
Kill the sparc64 argc conversion hack.
2002-09-13 14:23:07 +00:00
a18c524897
Figure out obj->pltrel(a) after the dynamic section is parsed, so the relative
...
order of DT_PLTREL and DT_JMPREL is irrelevant. Removes the need for yet
another weird #ifdef.
Also, be slightly more careful with the rel(a)lim trimming.
2002-09-13 13:28:43 +00:00
c74fdd7069
Use the addend on GLOB_DAT relocs (same as ABS32).
2002-09-13 05:45:46 +00:00
c65330320a
Do not use the implicit addend on REFQUAD relocs (same as GLOB_DAT).
2002-09-13 05:45:13 +00:00
daf8118731
Remove an old comment.
2002-09-13 04:09:49 +00:00
2f59687514
Almost all of the items in here are done.
2002-09-13 04:05:00 +00:00
777a2b8bb0
If we match a symbol name, but it's undefined, do not continue searching the
...
hash chain.
2002-09-13 03:40:40 +00:00
bbe45a0623
Remove a dumb #ifdef -- if VARPSZ is not defined on sparc, the build will
...
bomb anyway.
2002-09-13 03:22:08 +00:00
4de09f873b
Nuke an unneeded hack for the Alpha.
2002-09-13 03:17:21 +00:00
807a6ace45
Nuke two frivolous assert()s.
2002-09-13 03:12:40 +00:00
4b55da479c
_rtld_bind() is not used on MIPS. Should probably move it to MD code.
2002-09-13 03:09:38 +00:00
f64e457cee
Remove one more `dodebug' argument.
2002-09-12 23:11:37 +00:00
a3b892d178
Nuke -DRTLD_RELOCATE_SELF and `dodebug' from orbit.
2002-09-12 22:56:28 +00:00
f13bdd9850
Add some consts.
2002-09-12 21:06:38 +00:00
5ad776fbd2
Move another declaration that doesn't belong in MI source.
2002-09-12 20:27:34 +00:00
de4565ee93
Prototype different functions.
2002-09-12 20:24:03 +00:00
547bc13b97
_rtld_bind_start() is not used in MI code, so declare it in the MD code.
2002-09-12 20:20:58 +00:00
bf04efbfda
ld.elf_so has no textrels on arm, i386 or vax -- and shouldn't on any platform,
...
but that's beside the point.
2002-09-12 19:43:03 +00:00
6530c89689
Slight tweak.
2002-09-12 18:43:17 +00:00
326f270bfb
The GOT is Elf_Addr[], not Elf_Word[]...
2002-09-12 18:36:43 +00:00
6f0bda739f
This ld script is now unnecessary.
2002-09-12 18:33:39 +00:00
1165ce6f38
Clean up a little.
2002-09-12 18:28:53 +00:00
2346aaad04
Move GOT relocation code from _rtld_setup_pltgot() to
...
_rtld_relocate_nonplt_objects(). Add GOT self-relocation in
_rtld_relocate_nonplt_self().
MIPS ld.elf_so can now relocate itself.
2002-09-12 18:21:18 +00:00
2b2534febf
Add a `relocbase' argument to _rtld() and _rtld_init(), which is passed in
...
from the glue code in _rtld_start(). This is used to set objself.relocbase,
rather than assuming that it's the same as objself.mapbase (or 0 on MIPS).
Now -- with a bug fix to the kernel -- ld.elf_so can be linked at any VMA.
2002-09-12 17:45:41 +00:00
d04429c6a7
Figure out the relocation offset and pass it to _rtld().
2002-09-12 17:43:30 +00:00
ebf50cdfbc
Update comments.
2002-09-12 17:18:37 +00:00
ac81e3266b
Snapshot of work to make MIPS ld.elf_so relocatable.
2002-09-12 17:08:32 +00:00
5ee0e5e1c4
Pass the calculated relocation offset down to _rtld().
2002-09-12 17:07:25 +00:00
3cc5ae01bc
Correct somewhere where, in rev 1.3, two new sentences had been dropped
...
into the middle of an existing sentence, and put them after it instead.
2002-09-11 23:11:41 +00:00
1da1ca49f2
Do the _rtld_relocate_nonplt_self() thing here, and nuke the ld script.
2002-09-11 22:28:32 +00:00
fb69b8bec4
Implement _rtld_relocate_nonplt_self() here.
2002-09-11 21:14:08 +00:00
a57ae642a3
We don't use _GLOBAL_OFFSET_TABLE_ any more, so don't .global it.
2002-09-11 20:56:09 +00:00
fa0b5c37b2
This is a gratuitous null commit to force rtld.o to be rebuilt.
2002-09-11 20:48:21 +00:00
f1725abea5
Add a _rtld_relocate_nonplt_self() here too.
2002-09-11 20:45:52 +00:00
403e42f619
Do self-relocation similar to sparc.
2002-09-11 19:46:40 +00:00
e46eb805f1
This ld script is no longer necessary.
2002-09-11 19:46:04 +00:00
aa1330ce71
Do the _rtld_relocate_nonplt_self() thing for SPARC.
2002-09-11 19:11:05 +00:00
82148aec74
The ld script is no longer necessary.
2002-09-11 18:20:47 +00:00
d603bed369
Implement _rtld_relocate_nonplt_self() on Alpha.
2002-09-11 18:18:37 +00:00
735f9f38c0
Split the `self' case out of _rtld_relocate_nonplt_objects, into a new function
...
_rtld_relocate_nonplt_self(), which is called from _rtld_start.
Now we're completely relocated before main() is called.
We also no longer need _GOT_END_, so junk the ld script.
This code assumes that ld.elf_so only contains RELATIVE relocs, but that's
supposed to be the case for -Bsymbolic anyway.
2002-09-11 17:23:23 +00:00
b240241362
Add $NetBSD$.
2002-09-11 14:19:30 +00:00
db439ed878
Only one linker script is needed.
2002-09-10 14:19:18 +00:00
289d4ac1a5
Remove unneeded cruft from _rtld_start.
2002-09-10 02:52:33 +00:00
d048e19122
Shorten the startup sequence by a few instructions.
2002-09-10 02:24:51 +00:00
9f0e145106
Nuke RTLD_RELOCATE_SELF on ARM.
2002-09-09 19:28:36 +00:00
d22977bb61
Indent branch delay instructions.
2002-09-09 19:01:52 +00:00
c127ff08fb
Nuke RTLD_RELOCATE_SELF for SPARC.
2002-09-09 18:10:20 +00:00
89520a5f6e
Sync with current binutils.
2002-09-09 14:41:25 +00:00
ae0c8dfbc2
Get rid of RTLD_RELOCATE_SELF on i386.
...
* Add a ld.so.script that exports _GOT_END_.
* Prebind the GOT in _rtld_start.
* Skip over GOT relocs in _rtld_relocate_nonplt_objects().
This makes debugging work better at least.
2002-09-09 14:08:07 +00:00
221d594f18
Mention that -DDEBUG compilation is needed to get LD_DEBUG support.
...
Noted by Julio Merino in misc/18231.
2002-09-08 23:44:09 +00:00
6a37882ec7
Fix a serious performance problem for large programs on the Alpha.
...
Large programs need multiple GOTs. The lazy binding stub in the PLT
can be reached from any of these GOTs, but the dynamic linker only
has enough information to fix up the first GOT entry. Thus, calls
through the other GOTs went through the time-consuming lazy binding
process on every call.
This fix rewrites the PLT entries themselves to bypass the lazy binding
for those GOT entries that the dynamic linker can't fixup.
Fix from FreeBSD.
Note that now that we patch up the PLT, we need to put back the "imb"
that was removed from the binder exit path.
2002-09-08 02:48:28 +00:00
492895a9f9
Expand the comment about the R_TYPE(RELATIVE) special-case, as it's not quite
...
obvious from the code. Also, just do it all the time, as an optimization.
2002-09-06 15:51:23 +00:00
3eee01c5b9
Make the reloc tables const.
2002-09-06 15:32:56 +00:00
6f0fac5efa
Pass down an additional flag to _rtld_relocate_nonplt_objects() which
...
indicates whether we're relocating ld.elf_so itself. Use this in some places
rather than hackish tests on `dodebug'. (The Alpha and HPPA `dodebug' tests
were actually noops, because RTLD_RELOCATE_SELF is not set, and therefore
dodebug is always true.)
2002-09-06 15:17:53 +00:00
3ffe941c3a
Remove typedef which was never used. Meaningless comment as well.
2002-09-06 13:27:48 +00:00
0df8d18100
Introduce a new flag, `isdynamic', which is used to remember whether the
...
executable was of type ET_DYN. Use this instead of `mainprog' to determine
whether we need to do base-relative fixups of the PLT. (This allows loading
non-relocatable objects, should we desire to do that at some point...)
2002-09-06 13:20:29 +00:00
69b9a24021
More const poisoning. Now the entire _rtld_bind() path useds a const Obj_Entry.
2002-09-06 12:00:40 +00:00
b97e4047b7
Re-poison a lot of consts, now that the mark shite is gone.
2002-09-06 03:12:04 +00:00
a04012e731
Split _rtld_relocate_plt_object() into two MD functions:
...
* _rtld_relocate_plt_lazy() fixes up all the relocs pointing to the PLT. (On
most platforms it just does a simple base-relative fixup; on SPARC it does
nothing.)
* _rtld_relocate_plt_object() does immediate binding for a PLT entry.
The basic gist is that this saves a bit of time on SPARC (where the iteration
through the pltrela table was gratuitous), and a little less time on all other
platforms. A whole lot of #ifdef'ed crap is moved out of reloc.c, too.
NOT tested on: hppa sh x86_64
2002-09-06 03:05:35 +00:00
5f016a11ef
Repair a slight biff I made earlier -- do not exit the loop in
...
_rtld_relocate_nonplt_objects() early.
2002-09-06 02:01:39 +00:00
c59ffb0cc4
Remove all of the `mark' code. This is responsible for the `undefined PLT
...
symbol' errors, probably because the increment gets interrupted occasionally by
a signal. In general, _rtld_bind() should not modify ANY internal state.
2002-09-05 21:57:09 +00:00
ac191986f3
Partially fix up some debug printf()s that don't need to use defobj.
2002-09-05 21:31:30 +00:00
bdc2ac1520
A few things:
...
* Pass a symbol number to _rtld_find_symdef(), not a r_info.
* Don't try to do a symbol lookup when we find an unsupported relocation;
instead get the symbol name from the referencing object's strtab.
* Add preliminary support for `-z combreloc'-style startup optimization on
i386, `#ifdef COMBRELOC'.
2002-09-05 21:21:06 +00:00
1fa950d044
Fix one indentation error.
2002-09-05 20:12:43 +00:00
74444a2ddb
Rename _rtld_relocate_nonplt_object() to _rtld_relocate_nonplt_objects(),
...
and push the outer loop into it. This actually shaves a couple % off startup
time at least on PCs.
2002-09-05 20:08:14 +00:00
00c3ad1f56
Split _rtld_relocate_nonplt_object() into separate MD files.
2002-09-05 18:25:45 +00:00
b4a8cd2dbf
Split MIPS relocation goo into mips_reloc.c.
2002-09-05 17:58:02 +00:00
ae16986458
Simplify a couple of tests to !__mips__.
2002-09-05 17:06:11 +00:00
9e3442dc47
Combine the code that skips COPY relocs on most platforms (other than MIPS).
...
Remove the code that skips JMP_SLOT relocs (thus making ld.elf_so bomb if it
hits one in the non-PLT relocation table).
2002-09-05 17:01:13 +00:00
cf2d41da85
If REL(A)SZ extends into .rel(a).plt, trim rel(a)lim. This is actually due to
...
what is arguably a bug in BFD on some platforms (at least PPC and SPARC), but
I can't fix old binaries.
2002-09-05 16:58:16 +00:00
515622e6db
Remove unused arguments of _rtld_find_symdef().
2002-09-05 16:33:57 +00:00
1ea7e8253b
Remove support for PowerPC static relocations, which are not present in shared
...
objects, and in any case are implemented WRONG. (They're symbol-relative, not
base-relative.)
2002-09-05 15:47:52 +00:00
f4cbd3d6e2
Remove unused prototypes (including one for a non-existant function!).
2002-09-05 15:45:54 +00:00
2635aa544a
Make _rtld_setup_pltgot() a consistent interface on all platforms.
2002-09-05 15:38:22 +00:00
1e06edb63b
There is no need to do an imb after each call to the binder, because we only
...
patch the GOT, not the PLT. However, do an imb when setting up the PLT thunk,
just in case.
2002-09-04 19:13:00 +00:00
mycroft