aboutsummaryrefslogtreecommitdiff
path: root/elf
AgeCommit message (Collapse)AuthorFilesLines
2010-02-10Add m68k TLS relocationsMaxim Kuvyrkov1-1/+22
2010-01-18Use correct type when casting d_tag.Andreas Schwab1-5/+10
2010-01-18Ignore negative dynamic entry types.Andreas Schwab1-1/+1
2010-01-15/390: Add runtime check for the highgprs kernel feature.Andreas Krebbel1-0/+6
2010-01-15ld.so: Adjust the auxv if ld.so is directly invoked.Andreas Krebbel3-20/+40
If a binary gets invoked by passing it as argument to ld.so the stack still holds the auxiliary vector of ld.so when entering the _start routine of the executable. So the invocation via ld.so is not fully transparent to the executable. This causes problems if the executable wants to scan the auxv itself.
2010-01-07Add new ELF constant PN_XNUM.Roland McGrath1-1/+7
2010-01-07Use NT_GNU_HWCAP instead of literal 2.Roland McGrath1-2/+3
2009-12-15Avoid ELF lookup race.Ulrich Drepper2-5/+10
On some architectures the update of the l_used field in the lookup functions races with setting the other bits in the bitfield. Simply avoid this and optimize use of l_used in general.
2009-11-06Fix up whitespaces.Ulrich Drepper1-1/+1
2009-11-06Properly handle STT_GNU_IFUNC symbols in do_sym.H.J. Lu2-2/+16
do_sym should use DL_FIXUP_VALUE_XXX macros to call IFUNC function. Otherwise, it won't compile on ia64. This patch fixes it and adds a test.
2009-10-30Add cast in tst-execstack to avoid warning.Ulrich Drepper1-1/+1
2009-10-30Uglify IFUNC tests for PPC.Alan Modra10-99/+24
2009-10-30Implement IFUNC for PPC.Alan Modra1-9/+14
2009-10-21Handle IFUNC symbols in dlsym.Andreas Schwab1-0/+5
2009-10-07Fix descriptor leak when calling dlopen with RTLD_NOLOADAndreas Schwab1-3/+6
2009-09-28Fix missing reloc dependencyUlrich Drepper1-6/+3
When DL_LOOKUP_ADD_DEPENDENCY was introduced a few callers of _dl_lookup_versioned_symbol were not properly adjusted. One of them survived until now.
2009-08-08Support mixed SSE/AVX audit and check AVX only once.H.J. Lu8-2/+758
This patch fixes mixed SSE/AVX audit and checks AVX only once in _dl_runtime_profile. When an AVX or SSE register value in pltenter is modified, we have to make sure that the SSE part value is the same in both lr_xmm and lr_vector fields so that pltexit will get the correct value from either lr_xmm or lr_vector fields. AVX-enabled pltenter should update both lr_xmm and lr_vector fields to support stacked AVX/SSE pltenter functions.
2009-08-07Fix expanding STB_GNU_UNIQUE hash table.Andreas Schwab1-1/+1
An invalid symbol name pointer was entered into the hash table while enlarging it.
2009-07-29Preserve SSE registers in runtime relocations on x86-64.Ulrich Drepper2-0/+21
SSE registers are used for passing parameters and must be preserved in runtime relocations. This is inside ld.so enforced through the tests in tst-xmmymm.sh. But the malloc routines used after startup come from libc.so and can be arbitrarily complex. It's overkill to save the SSE registers all the time because of that. These calls are rare. Instead we save them on demand. The new infrastructure put in place in this patch makes this possible and efficient.
2009-07-28Fix bookkeeping of static TLS block for TLS_TCB_AT_TP architectures.Ulrich Drepper1-1/+4
2009-07-27Break out _dl_scope_free into its own file.Ulrich Drepper3-36/+59
This reduces the coarse static callgraph that can be discovered by looking at the object files.
2009-07-27Fix STB_GNU_UNIQUE handling for > 30 unique symbols.Jakub Jelinek1-2/+3
There were several issues when the initial 31 entries hashtab filled up. size * 3 <= tab->n_elements is always false, table can't have more elements than its size. I assume from libiberty/hashtab.c this meant to be check for 3/4 full. Even after fixing that, _dl_higher_prime_number (31) apparently returns 31, only _dl_higher_prime_number (32) returns 61. And, size variable wasn't updated during reallocation, which means during reallocation the insertion of the new entry was done into a wrong spot. All this lead to a hang in ld.so, because a search with n_elements 31 size 31 wouldn't ever terminate.
2009-07-26Define NT_GNU_GOLD_VERSION.Ulrich Drepper1-0/+3
2009-07-23Minor optimization of unloading prevention patch.Ulrich Drepper1-5/+5
References to unique symbols from copy relocations can only come from executables which cannot be unloaded anyway. Optimize the code to set the unload flag a bit.
2009-07-23Fix comment.Ulrich Drepper1-1/+1
2009-07-23Some more fixes for the unique symbol handling.Ulrich Drepper1-5/+4
The hash value can be zero, so we need a different test for empty slots. And unify the way we prevent a DSO from being unloaded.
2009-07-23Handle unloading of DSO with a unique symbol definition.Ulrich Drepper2-5/+48
2009-07-16Remove warning and little optimization.Ulrich Drepper1-0/+1
The prototype for _dl_higher_prime_number was missing. While at it, the function is now marked with internal_function.
2009-07-15Fix build issue with modules for audit test on machines != x86-64.Ulrich Drepper1-3/+5
2009-07-10Add AVX support to ld.so auditing for x86-64.H.J. Lu7-2/+554
2009-07-10Remove do-lookup.h.Ulrich Drepper3-388/+366
It is not necessary/useful anymore to have the content separate from dl-lookup.c.
2009-07-09Implement STB_GNU_UNIQUE handling.Ulrich Drepper11-11/+291
Some symbols have to be identified process-wide by their name. This is particularly important for some C++ features (e.g., class local static data and static variables in inline functions). This cannot completely be implemented with ELF functionality so far. The STB_GNU_UNIQUE binding helps by ensuring the dynamic linker will always use the same definition for all symbols with the same name and this binding.
2009-07-07Define STB_GNU_UNIQUE.Ulrich Drepper1-0/+1
2009-07-07Clean up code for hash table handling in ld.so.Ulrich Drepper1-0/+64
2009-07-06Optimize test for valid ELF symbol types in lookup function.ibm/masterUlrich Drepper1-7/+7
2009-07-02Fix whitespace in last patch.Ulrich Drepper2-2/+2
2009-07-02Fix handling of xmm6 in ld.so audit hooks on x86-64.H.J. Lu4-1/+210
2009-06-22More whitespace cleanups.Ulrich Drepper2-2/+2
2009-06-22Clean up whitespaces in last patch.Ulrich Drepper3-6/+6
2009-06-22Add more IFUNC tests.H.J. Lu18-5/+399
Mostly tests around not-exported IFUNC functions, IFUNC in statically linked binaries and PIEs, etc.
2009-06-17Use +link-pie Makefile rule to link PIE ifunc tests.H.J. Lu1-16/+2
2009-06-16Don't treat bug reporting message as a format string.Andreas Schwab1-2/+2
2009-06-16Consolidate PIE linking Makefile rules.H.J. Lu1-8/+1
2009-06-15libc part of prelink IFUNC support.Jakub Jelinek1-6/+11
The dl-lookup.c changes are needed for prelink (support in prelink checked into SVN, tested for both i?86 and x86-64), dl-irel.h just something I discovered by code inspection.
2009-06-03Test for ELF IFUNC functionality.H.J. Lu24-1/+627
2009-06-03Terminate correct buffer.Ulrich Drepper1-1/+1
Little typo could cause additional characters to be printed.
2009-05-31Finish IFUNC support for x86 and x86-64.Ulrich Drepper1-2/+4
Add support for the IRELAIVE relocation and IFUNC in static executables.
2009-05-16Remove redundant .gitignore files.Andreas Schwab1-6/+0
2009-05-15rename each .cvsignore file to .gitignoreJim Meyering1-0/+0
2009-05-15Test DSOs for executable stack.Ulrich Drepper2-2/+170
Add a text program, built to run on the host, to check all newly built DSOs for executable stacks and fail if the stack information is missing or indicates executable stacks.