This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.
| Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
|---|---|---|
| Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
| Other format: | [Raw text] | |
On Sat, Jun 28, 2008 at 06:58:58PM +0100, Richard Sandiford wrote: > I suppose the good news is that we can pick the best bits of each > implementation as the official one. And that's what I've been working on. We had to pick some bits from the CodeSourcery implementation that I wouldn't describe as "best" - arbitrarily different or a bit less efficient - due to ABI constraints we've already discussed. But I was able to pick up clever bits from both sets as I combined them. I have attached patch for gdb and glibc. These are only slightly changed from my last posting. I've also attached quilt series for binutils and gcc; these are based on Richard's. I disabled a couple of patches that had already been applied to binutils HEAD, but otherwise did not change any of Richard's patches. I added a cs-plts.patch to the end of both series. This form is probably better for review and discussion than it would be for committing; the bits from cs-plts.patch belong in several earlier patches in each series, but I decided this would be an unworkable way to compare patches. I have tested this thoroughly on mips-linux and somewhat on N32 / N64 (binutils testsuites and GCC's builtins.exp). The testing was before some last minute fixes, though, and I've only retested individual tests since. I'll be rerunning a complete test cycle, but I wanted to get this posted and I'm 99% sure I didn't break anything (this time). All comments welcome - Richard, especially from you. How would you like to proceed? I think the first step should be to get your other binutils/gcc patches merged, including MIPS16 PIC; I used those as a base. But see a few of the notes for potential problems with those patches. My notes about the merge: - Richard's ld -r support is an addition to the ABI, but does not conflict with anything else, so I included it. I discovered two potential problems: - If a symbol with STO_MIPS_PIC is localized using objcopy, binutils will ignore the flag. I don't think this is presently worth implementing but it might be wise to add an error message. I haven't done that yet. - Superfluous la25 stubs are suppressed when a PIC2 file uses jal. This is an optimization performed by gcc -mno-shared. It will not work after ld -r into a non-PIC file; the jump will appear to come from a non-PIC object and be redirected to a new stub. This is only a minor performance pessimization and I do not plan to fix it. - Some mips16 call stubs called through $1 instead of $25. I believe I've fixed them all, though I am not set up to test mips16. I don't think this would have caused a problem, since as far as I can tell the indirect jumps are only used when the target is mips16; but better safe than sorry so we now follow the ABI. - It would be nice to generate, in some cases, both a .MIPS.stubs lazy binding stub and a PLT entry. However, I determined that considerable additional work would be required to do this; most likely we'd need two entries for the same symbol in the dynamic symbol table so that the GOT entry could be associated with the lazy loading stub. As things stand it is possible to link non-PIC and PIC code together, and if both call the same function and the non-PIC code takes its address, calls to the function from PIC code will be penalized. I do not expect this case to matter. Most applications will be predominantly PIC or non-PIC. - I've dropped support for a non-fixed $gp. This is a handy optimization, but it was getting in the way and it was the part of the GCC patch Richard had the most comments on. I can resubmit it after everything else is merged. - Richard's implementation had __PIC__ mean abicalls. Our patch changed __PIC__ to mean pic2 abicalls only. I've included that in this patch. My reasoning is that most non-pic non-abicalls code works properly even with pic0 abicalls; the only problem is indirect calls through a register other than $25. This lets glibc automatically use some more efficient sequences in static applications. - With Richard's patches I saw a crash building libstdc++ with symbol versioning. This was caused by not following indirect symbols in all cases during check_relocs; we would assign a GGA_NORMAL GOT slot to an indirect symbol, and trigger an assertion failure while removing it. This is sensitive to link order. - A small change in glibc fixes the canonical address of functions with PLT entries, when referenced from PIC code. - I added pointer_equality_needed support to binutils to suppress setting st_value to the PLT entry in most cases. - The GCC new-static-chain.patch causes nested-func-4.exe to fail. _mcount is called through a PLT entry, which clobbers $15. I believe we need to add this to MIPS_SAVE_REG_FOR_PROFILING_P. I didn't fix this yet. - no-fn-name-already-declared.patch removed the call to ASM_OUTPUT_TYPE_DIRECTIVE for Linux. .ent has similar effect, but is suppressed by flag_inhibit_size_directive. This caused glibc's _init to be STT_OBJECT, and not get a PIC call stub. I've changed GCC to emit .type for all platforms; Richard, if this should be restricted to the status quo (i.e. Linux) let me know. Also the STT_FUNC check in the linker was unnecessary now that we only use la25 stubs for jump relocations. -- Daniel Jacobowitz CodeSourcery
Attachment:
gcc-quilt.tar.gz
Description: Binary data
Attachment:
binutils-quilt.tar.gz
Description: Binary data
2008-03-07 Daniel Jacobowitz <dan@codesourcery.com>
gdb/
* mips-tdep.c (mips32_scan_prologue): Stop scanning at branches.
2008-07-24 Daniel Jacobowitz <dan@codesourcery.com>
gdb/
* mips-linux-tdep.c (mips_linux_in_dynsym_resolve_code): Update
comments.
(mips_linux_skip_resolver): Also use glibc_skip_solib_resolver.
(mips_linux_init_abi): Do not override skip_trampoline_code.
* configure.tgt (mips*-*-linux*): Add glibc-tdep.o.
* mips-tdep.c (mips_stub_frame_sniffer): Use the stub frame sniffer
for PIC stubs.
(mips_skip_mips16_trampoline_code): Rename from
mips_skip_trampoline_code.
(mips_skip_pic_trampoline_code, mips_skip_trampoline_code): New.
* infrun.c (handle_inferior_event): Do not check
IN_SOLIB_DYNSYM_RESOLVE_CODE. Do not pass zero to
in_solib_dynsym_resolve_code.
* Makefile.in (mips-linux-tdep.o): Update.
2008-07-24 Daniel Jacobowitz <dan@codesourcery.com>
gdb/gdbserver/
* configure.srv: Combine mips*64*-*-linux* and mips*-*-linux*.
Index: gdb/Makefile.in
===================================================================
RCS file: /scratch/gcc/repos/src/src/gdb/Makefile.in,v
retrieving revision 1.1027
diff -u -p -r1.1027 Makefile.in
--- gdb/Makefile.in 10 Jun 2008 10:23:53 -0000 1.1027
+++ gdb/Makefile.in 27 Jun 2008 14:28:26 -0000
@@ -2519,7 +2519,7 @@ mips-linux-tdep.o: mips-linux-tdep.c $(d
$(gdb_assert_h) $(frame_h) $(regcache_h) $(trad_frame_h) \
$(tramp_frame_h) $(gdbtypes_h) $(solib_h) $(symtab_h) \
$(mips_linux_tdep_h) $(solist_h) $(solib_svr4_h) \
- $(target_descriptions_h)
+ $(target_descriptions_h) $(glibc_tdep_h)
mipsnbsd-nat.o: mipsnbsd-nat.c $(defs_h) $(inferior_h) $(regcache_h) \
$(target_h) $(mips_tdep_h) $(mipsnbsd_tdep_h) $(inf_ptrace_h)
mipsnbsd-tdep.o: mipsnbsd-tdep.c $(defs_h) $(gdbcore_h) $(regcache_h) \
Index: gdb/configure.tgt
===================================================================
RCS file: /scratch/gcc/repos/src/src/gdb/configure.tgt,v
retrieving revision 1.203
diff -u -p -r1.203 configure.tgt
--- gdb/configure.tgt 1 May 2008 23:09:14 -0000 1.203
+++ gdb/configure.tgt 27 Jun 2008 14:28:26 -0000
@@ -297,7 +297,7 @@ mips*-sgi-irix6*)
;;
mips*-*-linux*)
# Target: Linux/MIPS
- gdb_target_obs="mips-tdep.o mips-linux-tdep.o \
+ gdb_target_obs="mips-tdep.o mips-linux-tdep.o glibc-tdep.o \
corelow.o solib.o solib-svr4.o symfile-mem.o"
gdb_sim=../sim/mips/libsim.a
build_gdbserver=yes
Index: gdb/infrun.c
===================================================================
RCS file: /scratch/gcc/repos/src/src/gdb/infrun.c,v
retrieving revision 1.282
diff -u -p -r1.282 infrun.c
--- gdb/infrun.c 24 Jun 2008 19:30:18 -0000 1.282
+++ gdb/infrun.c 27 Jun 2008 14:28:26 -0000
@@ -2863,12 +2863,7 @@ infrun: BPSTAT_WHAT_SET_LONGJMP_RESUME (
until we exit the run time loader code and reach the callee's
address. */
if (step_over_calls == STEP_OVER_UNDEBUGGABLE
-#ifdef IN_SOLIB_DYNSYM_RESOLVE_CODE
- && IN_SOLIB_DYNSYM_RESOLVE_CODE (stop_pc)
-#else
- && in_solib_dynsym_resolve_code (stop_pc)
-#endif
- )
+ && in_solib_dynsym_resolve_code (stop_pc))
{
CORE_ADDR pc_after_resolver =
gdbarch_skip_solib_resolver (current_gdbarch, stop_pc);
@@ -2961,13 +2956,7 @@ infrun: BPSTAT_WHAT_SET_LONGJMP_RESUME (
if (real_stop_pc != 0)
ecs->stop_func_start = real_stop_pc;
- if (
-#ifdef IN_SOLIB_DYNSYM_RESOLVE_CODE
- IN_SOLIB_DYNSYM_RESOLVE_CODE (ecs->stop_func_start)
-#else
- in_solib_dynsym_resolve_code (ecs->stop_func_start)
-#endif
-)
+ if (real_stop_pc != 0 && in_solib_dynsym_resolve_code (real_stop_pc))
{
struct symtab_and_line sr_sal;
init_sal (&sr_sal);
Index: gdb/mips-linux-tdep.c
===================================================================
RCS file: /scratch/gcc/repos/src/src/gdb/mips-linux-tdep.c,v
retrieving revision 1.71
diff -u -p -r1.71 mips-linux-tdep.c
--- gdb/mips-linux-tdep.c 30 Apr 2008 21:25:16 -0000 1.71
+++ gdb/mips-linux-tdep.c 27 Jun 2008 14:28:26 -0000
@@ -37,6 +37,7 @@
#include "symtab.h"
#include "target-descriptions.h"
#include "mips-linux-tdep.h"
+#include "glibc-tdep.h"
static struct target_so_ops mips_svr4_so_ops;
@@ -666,13 +667,13 @@ mips_linux_in_dynsym_stub (CORE_ADDR pc,
}
/* Return non-zero iff PC belongs to the dynamic linker resolution
- code or to a stub. */
+ code, a PLT entry, or a lazy binding stub. */
static int
mips_linux_in_dynsym_resolve_code (CORE_ADDR pc)
{
/* Check whether PC is in the dynamic linker. This also checks
- whether it is in the .plt section, which MIPS does not use. */
+ whether it is in the .plt section, used by non-PIC executables. */
if (svr4_in_dynsym_resolve_code (pc))
return 1;
@@ -688,8 +689,8 @@ mips_linux_in_dynsym_resolve_code (CORE_
and glibc_skip_solib_resolver in glibc-tdep.c. The normal glibc
implementation of this triggers at "fixup" from the same objfile as
"_dl_runtime_resolve"; MIPS GNU/Linux can trigger at
- "__dl_runtime_resolve" directly. An unresolved PLT entry will
- point to _dl_runtime_resolve, which will first call
+ "__dl_runtime_resolve" directly. An unresolved lazy binding
+ stub will point to _dl_runtime_resolve, which will first call
__dl_runtime_resolve, and then pass control to the resolved
function. */
@@ -703,7 +704,7 @@ mips_linux_skip_resolver (struct gdbarch
if (resolver && SYMBOL_VALUE_ADDRESS (resolver) == pc)
return frame_pc_unwind (get_current_frame ());
- return 0;
+ return glibc_skip_solib_resolver (gdbarch, pc);
}
/* Signal trampoline support. There are four supported layouts for a
@@ -1151,7 +1152,6 @@ mips_linux_init_abi (struct gdbarch_info
break;
}
- set_gdbarch_skip_trampoline_code (gdbarch, find_solib_trampoline_target);
set_gdbarch_skip_solib_resolver (gdbarch, mips_linux_skip_resolver);
set_gdbarch_software_single_step (gdbarch, mips_software_single_step);
Index: gdb/mips-tdep.c
===================================================================
RCS file: /scratch/gcc/repos/src/src/gdb/mips-tdep.c,v
retrieving revision 1.475
diff -u -p -r1.475 mips-tdep.c
--- gdb/mips-tdep.c 3 Jun 2008 10:53:34 -0000 1.475
+++ gdb/mips-tdep.c 23 Jul 2008 19:28:33 -0000
@@ -1928,6 +1928,7 @@ mips32_scan_prologue (CORE_ADDR start_pc
CORE_ADDR end_prologue_addr = 0;
int seen_sp_adjust = 0;
int load_immediate_bytes = 0;
+ int in_delay_slot = 0;
struct gdbarch *gdbarch = get_frame_arch (this_frame);
int regsize_is_64_bits = (mips_abi_regsize (gdbarch) == 8);
@@ -2085,7 +2086,18 @@ restart:
instructions? */
if (end_prologue_addr == 0)
end_prologue_addr = cur_pc;
+
+ /* Check for branches and jumps. For now, only jump to
+ register are caught (i.e. returns). */
+ if ((itype_op (inst) & 0x07) == 0 && rtype_funct (inst) == 8)
+ in_delay_slot = 1;
}
+
+ /* If the previous instruction was a jump, we must have reached
+ the end of the prologue by now. Stop scanning so that we do
+ not go past the function return. */
+ if (in_delay_slot)
+ break;
}
if (this_cache != NULL)
@@ -2287,6 +2299,7 @@ mips_stub_frame_sniffer (const struct fr
gdb_byte dummy[4];
struct obj_section *s;
CORE_ADDR pc = get_frame_address_in_block (this_frame);
+ struct minimal_symbol *msym;
/* Use the stub unwinder for unreadable code. */
if (target_read_memory (get_frame_pc (this_frame), dummy, 4) != 0)
@@ -2303,6 +2316,14 @@ mips_stub_frame_sniffer (const struct fr
".MIPS.stubs") == 0)
return 1;
+ /* Calling a PIC function from a non-PIC function passes through a
+ stub. The stub for foo is named ".pic.foo". */
+ msym = lookup_minimal_symbol_by_pc (pc);
+ if (msym != NULL
+ && SYMBOL_LINKAGE_NAME (msym) != NULL
+ && strncmp (SYMBOL_LINKAGE_NAME (msym), ".pic.", 5) == 0)
+ return 1;
+
return 0;
}
@@ -5020,7 +5041,7 @@ mips_breakpoint_from_pc (struct gdbarch
gory details. */
static CORE_ADDR
-mips_skip_trampoline_code (struct frame_info *frame, CORE_ADDR pc)
+mips_skip_mips16_trampoline_code (struct frame_info *frame, CORE_ADDR pc)
{
char *name;
CORE_ADDR start_addr;
@@ -5099,6 +5120,80 @@ mips_skip_trampoline_code (struct frame_
return 0; /* not a stub */
}
+/* If the current PC is the start of a non-PIC-to-PIC stub, return the
+ PC of the stub target. The stub just loads $t9 and jumps to it,
+ so that $t9 has the correct value at function entry. */
+
+static CORE_ADDR
+mips_skip_pic_trampoline_code (struct frame_info *frame, CORE_ADDR pc)
+{
+ struct minimal_symbol *msym;
+ int i;
+ gdb_byte stub_code[16];
+ int32_t stub_words[4];
+
+ /* The stub for foo is named ".pic.foo", and is either two
+ instructions inserted before foo or a three instruction sequence
+ which jumps to foo. */
+ msym = lookup_minimal_symbol_by_pc (pc);
+ if (msym == NULL
+ || SYMBOL_VALUE_ADDRESS (msym) != pc
+ || SYMBOL_LINKAGE_NAME (msym) == NULL
+ || strncmp (SYMBOL_LINKAGE_NAME (msym), ".pic.", 5) != 0)
+ return 0;
+
+ /* A two-instruction header. */
+ if (MSYMBOL_SIZE (msym) == 8)
+ return pc + 8;
+
+ /* A three-instruction (plus delay slot) trampoline. */
+ if (MSYMBOL_SIZE (msym) == 16)
+ {
+ if (target_read_memory (pc, stub_code, 16) != 0)
+ return 0;
+ for (i = 0; i < 4; i++)
+ stub_words[i] = extract_unsigned_integer (stub_code + i * 4, 4);
+
+ /* A stub contains these instructions:
+ lui t9, %hi(target)
+ j target
+ addiu t9, t9, %lo(target)
+ nop
+
+ This works even for N64, since stubs are only generated with
+ -msym32. */
+ if ((stub_words[0] & 0xffff0000U) == 0x3c190000
+ && (stub_words[1] & 0xfc000000U) == 0x08000000
+ && (stub_words[2] & 0xffff0000U) == 0x27390000
+ && stub_words[3] == 0x00000000)
+ return (((stub_words[0] & 0x0000ffff) << 16)
+ + (stub_words[2] & 0x0000ffff));
+ }
+
+ /* Not a recognized stub. */
+ return 0;
+}
+
+static CORE_ADDR
+mips_skip_trampoline_code (struct frame_info *frame, CORE_ADDR pc)
+{
+ CORE_ADDR target_pc;
+
+ target_pc = mips_skip_mips16_trampoline_code (frame, pc);
+ if (target_pc)
+ return target_pc;
+
+ target_pc = find_solib_trampoline_target (frame, pc);
+ if (target_pc)
+ return target_pc;
+
+ target_pc = mips_skip_pic_trampoline_code (frame, pc);
+ if (target_pc)
+ return target_pc;
+
+ return 0;
+}
+
/* Convert a dbx stab register number (from `r' declaration) to a GDB
[1 * gdbarch_num_regs .. 2 * gdbarch_num_regs) REGNUM. */
Index: gdb/gdbserver/configure.srv
===================================================================
RCS file: /scratch/gcc/repos/src/src/gdb/gdbserver/configure.srv,v
retrieving revision 1.35
diff -u -p -r1.35 configure.srv
--- gdb/gdbserver/configure.srv 3 May 2008 17:16:43 -0000 1.35
+++ gdb/gdbserver/configure.srv 23 Jul 2008 14:11:54 -0000
@@ -86,22 +86,15 @@ case "${target}" in
srv_linux_regsets=yes
srv_linux_thread_db=yes
;;
- mips*64*-*-linux*) srv_regobj=mips64-linux.o
+ mips*-*-linux*) srv_regobj="mips-linux.o mips64-linux.o"
srv_tgtobj="linux-low.o linux-mips-low.o"
- srv_xmlfiles="mips64-linux.xml"
- srv_xmlfiles="${srv_xmlfiles} mips64-cpu.xml"
- srv_xmlfiles="${srv_xmlfiles} mips64-cp0.xml"
- srv_xmlfiles="${srv_xmlfiles} mips64-fpu.xml"
- srv_linux_regsets=yes
- srv_linux_usrregs=yes
- srv_linux_thread_db=yes
- ;;
- mips*-*-linux*) srv_regobj=mips-linux.o
- srv_tgtobj="linux-low.o linux-mips-low.o"
- srv_xmlfiles="mips-linux.xml"
+ srv_xmlfiles="mips-linux.xml mips64-linux.xml"
srv_xmlfiles="${srv_xmlfiles} mips-cpu.xml"
srv_xmlfiles="${srv_xmlfiles} mips-cp0.xml"
srv_xmlfiles="${srv_xmlfiles} mips-fpu.xml"
+ srv_xmlfiles="${srv_xmlfiles} mips64-cpu.xml"
+ srv_xmlfiles="${srv_xmlfiles} mips64-cp0.xml"
+ srv_xmlfiles="${srv_xmlfiles} mips64-fpu.xml"
srv_linux_regsets=yes
srv_linux_usrregs=yes
srv_linux_thread_db=yes
2008-07-24 Mark Shinwell <shinwell@codesourcery.com>
Daniel Jacobowitz <dan@codesourcery.com>
Richard Sandiford <rdsandiford@googlemail.com>
* sysdeps/mips/dl-lookup.c: New.
* sysdeps/mips/do-lookup.h: New.
* sysdeps/mips/dl-machine.h (ELF_MACHINE_NO_PLT): Remove
definition.
(ELF_MACHINE_JMP_SLOT): Alter definition and update comment.
(elf_machine_type_class): Likewise.
(ELF_MACHINE_PLT_REL): Define.
(elf_machine_fixup_plt): New.
(elf_machine_plt_value): New.
(elf_machine_reloc): Handle jump slot and copy relocations.
(elf_machine_lazy_rel): Point relocation place at PLT if
required.
(RESOLVE_GOTSYM): Take a relocation type argument.
(elf_machine_got_rel): Bind lazy stubs directly to their target if
!lazy. Skip lazy binding for PLT symbols.
(elf_machine_runtime_setup): Fill in .got.plt header.
* sysdeps/mips/dl-trampoline.c (IFNEWABI): New macro.
(_dl_runtime_pltresolve): New.
* sysdeps/mips/bits/linkmap.h: New file.
* sysdeps/mips/tls-macros.h: Load $gp as required. Merge 32-bit and
64-bit versions.
* sysdeps/unix/sysv/linux/mips/mips32/sysdep.h (SYSCALL_ERROR_LABEL):
Delete definition.
* sysdeps/unix/sysv/linux/mips/nptl/sysdep-cancel.h (PSEUDO_CPLOAD,
PSEUDO_ERRJMP, PSEUDO_SAVEGP, PSEUDO_LOADGP): Define.
(PSEUDO): Use them. Move outside __PIC__.
(PSEUDO_JMP): New.
(CENABLE, CDISABLE): Use it.
Index: sysdeps/unix/sysv/linux/mips/mips32/sysdep.h
===================================================================
--- sysdeps/unix/sysv/linux/mips/mips32/sysdep.h (revision 213009)
+++ sysdeps/unix/sysv/linux/mips/mips32/sysdep.h (working copy)
@@ -35,15 +35,7 @@
# define SYS_ify(syscall_name) __NR_/**/syscall_name
#endif
-#ifdef __ASSEMBLER__
-
-/* We don't want the label for the error handler to be visible in the symbol
- table when we define it here. */
-#ifdef __PIC__
-# define SYSCALL_ERROR_LABEL 99b
-#endif
-
-#else /* ! __ASSEMBLER__ */
+#ifndef __ASSEMBLER__
/* Define a macro which expands into the inline wrapper code for a system
call. */
Index: sysdeps/unix/sysv/linux/mips/nptl/sysdep-cancel.h
===================================================================
--- sysdeps/unix/sysv/linux/mips/nptl/sysdep-cancel.h (revision 213009)
+++ sysdeps/unix/sysv/linux/mips/nptl/sysdep-cancel.h (working copy)
@@ -25,28 +25,38 @@
#if !defined NOT_IN_libc || defined IS_IN_libpthread || defined IS_IN_librt
-#ifdef __PIC__
+# ifdef __PIC__
+# define PSEUDO_CPLOAD .cpload t9;
+# define PSEUDO_ERRJMP la t9, __syscall_error; jr t9;
+# define PSEUDO_SAVEGP sw gp, 32(sp); cfi_rel_offset (gp, 32);
+# define PSEUDO_LOADGP lw gp, 32(sp);
+# else
+# define PSEUDO_CPLOAD
+# define PSEUDO_ERRJMP j __syscall_error;
+# define PSEUDO_SAVEGP
+# define PSEUDO_LOADGP
+# endif
+
# undef PSEUDO
# define PSEUDO(name, syscall_name, args) \
.align 2; \
L(pseudo_start): \
cfi_startproc; \
- 99: la t9,__syscall_error; \
- jr t9; \
+ 99: PSEUDO_ERRJMP \
.type __##syscall_name##_nocancel, @function; \
.globl __##syscall_name##_nocancel; \
__##syscall_name##_nocancel: \
.set noreorder; \
- .cpload t9; \
+ PSEUDO_CPLOAD \
li v0, SYS_ify(syscall_name); \
syscall; \
.set reorder; \
- bne a3, zero, SYSCALL_ERROR_LABEL; \
+ bne a3, zero, 99b; \
ret; \
.size __##syscall_name##_nocancel,.-__##syscall_name##_nocancel; \
ENTRY (name) \
.set noreorder; \
- .cpload t9; \
+ PSEUDO_CPLOAD \
.set reorder; \
SINGLE_THREAD_P(v1); \
bne zero, v1, L(pseudo_cancel); \
@@ -54,17 +64,16 @@
li v0, SYS_ify(syscall_name); \
syscall; \
.set reorder; \
- bne a3, zero, SYSCALL_ERROR_LABEL; \
+ bne a3, zero, 99b; \
ret; \
L(pseudo_cancel): \
SAVESTK_##args; \
sw ra, 28(sp); \
cfi_rel_offset (ra, 28); \
- sw gp, 32(sp); \
- cfi_rel_offset (gp, 32); \
+ PSEUDO_SAVEGP \
PUSHARGS_##args; /* save syscall args */ \
CENABLE; \
- lw gp, 32(sp); \
+ PSEUDO_LOADGP \
sw v0, 44(sp); /* save mask */ \
POPARGS_##args; /* restore syscall args */ \
.set noreorder; \
@@ -75,12 +84,12 @@
sw a3, 40(sp); /* save syscall error flag */ \
lw a0, 44(sp); /* pass mask as arg1 */ \
CDISABLE; \
- lw gp, 32(sp); \
+ PSEUDO_LOADGP \
lw v0, 36(sp); /* restore syscall result */ \
lw a3, 40(sp); /* restore syscall error flag */ \
lw ra, 28(sp); /* restore return address */ \
.set noreorder; \
- bne a3, zero, SYSCALL_ERROR_LABEL; \
+ bne a3, zero, 99b; \
RESTORESTK; \
L(pseudo_end): \
.set reorder;
@@ -88,8 +97,6 @@
# undef PSEUDO_END
# define PSEUDO_END(sym) cfi_endproc; .end sym; .size sym,.-sym
-#endif
-
# define PUSHARGS_0 /* nothing to do */
# define PUSHARGS_1 PUSHARGS_0 sw a0, 0(sp); cfi_rel_offset (a0, 0);
# define PUSHARGS_2 PUSHARGS_1 sw a1, 4(sp); cfi_rel_offset (a1, 4);
@@ -136,19 +143,25 @@
# define RESTORESTK addu sp, STKSPACE; cfi_adjust_cfa_offset(-STKSPACE)
+# ifdef __PIC__
/* We use jalr rather than jal. This means that the assembler will not
automatically restore $gp (in case libc has multiple GOTs) so we must
do it manually - which we have to do anyway since we don't use .cprestore.
It also shuts up the assembler warning about not using .cprestore. */
+# define PSEUDO_JMP(sym) la t9, sym; jalr t9;
+# else
+# define PSEUDO_JMP(sym) jal sym;
+# endif
+
# ifdef IS_IN_libpthread
-# define CENABLE la t9, __pthread_enable_asynccancel; jalr t9;
-# define CDISABLE la t9, __pthread_disable_asynccancel; jalr t9;
+# define CENABLE PSEUDO_JMP (__pthread_enable_asynccancel)
+# define CDISABLE PSEUDO_JMP (__pthread_disable_asynccancel)
# elif defined IS_IN_librt
-# define CENABLE la t9, __librt_enable_asynccancel; jalr t9;
-# define CDISABLE la t9, __librt_disable_asynccancel; jalr t9;
+# define CENABLE PSEUDO_JMP (__librt_enable_asynccancel)
+# define CDISABLE PSEUDO_JMP (__librt_disable_asynccancel)
# else
-# define CENABLE la t9, __libc_enable_asynccancel; jalr t9;
-# define CDISABLE la t9, __libc_disable_asynccancel; jalr t9;
+# define CENABLE PSEUDO_JMP (__libc_enable_asynccancel)
+# define CDISABLE PSEUDO_JMP (__libc_disable_asynccancel)
# endif
# ifndef __ASSEMBLER__
Index: sysdeps/mips/do-lookup.h
===================================================================
--- sysdeps/mips/do-lookup.h (revision 0)
+++ sysdeps/mips/do-lookup.h (revision 0)
@@ -0,0 +1,37 @@
+/* MIPS-specific veneer to GLIBC's do-lookup.h.
+ Copyright (C) 2008 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+/* The semantics of zero/non-zero values of undefined symbols differs
+ depending on whether the non-PIC ABI is in use. Under the non-PIC ABI,
+ a non-zero value indicates that there is an address reference to the
+ symbol and thus it must always be resolved (except when resolving a jump
+ slot relocation) to the PLT entry whose address is provided as the
+ symbol's value; a zero value indicates that this canonical-address
+ behaviour is not required. Yet under the classic MIPS psABI, a zero value
+ indicates that there is an address reference to the function and the
+ dynamic linker must resolve the symbol immediately upon loading. To
+ avoid conflict, symbols for which the dynamic linker must assume the
+ non-PIC ABI semantics are marked with the STO_MIPS_PLT flag. The
+ following ugly hack causes the code in the platform-independent
+ do-lookup.h file to check this flag correctly. */
+#define st_value st_shndx == SHN_UNDEF && !(sym->st_other & STO_MIPS_PLT)) \
+ || (sym->st_value
+#include_next "do-lookup.h"
+#undef st_value
+
Index: sysdeps/mips/tls-macros.h
===================================================================
--- sysdeps/mips/tls-macros.h (revision 213009)
+++ sysdeps/mips/tls-macros.h (working copy)
@@ -1,44 +1,56 @@
/* Macros to support TLS testing in times of missing compiler support. */
-#if _MIPS_SIM != _ABI64
+#include <sys/cdefs.h>
+#include <sys/asm.h>
-/* These versions are for o32 and n32. */
-
-# define TLS_GD(x) \
- ({ void *__result; \
- extern void *__tls_get_addr (void *); \
- asm ("addiu %0, $28, %%tlsgd(" #x ")" \
- : "=r" (__result)); \
- (int *)__tls_get_addr (__result); })
+#define __STRING2(X) __STRING(X)
+#define ADDU __STRING2(PTR_ADDU)
+#define ADDIU __STRING2(PTR_ADDIU)
+#define LW __STRING2(PTR_L)
+
+/* Load the GOT pointer, which may not be in $28 in a non-PIC
+ (abicalls pic0) function. */
+#ifndef __PIC__
+# if _MIPS_SIM != _ABI64
+# define LOAD_GP "move %[tmp], $28\n\tla $28, __gnu_local_gp\n\t"
+# else
+# define LOAD_GP "move %[tmp], $28\n\tdla $28, __gnu_local_gp\n\t"
+# endif
+# define UNLOAD_GP "\n\tmove $28, %[tmp]"
#else
+# define LOAD_GP
+# define UNLOAD_GP
+#endif
+
# define TLS_GD(x) \
- ({ void *__result; \
+ ({ void *__result, *__tmp; \
extern void *__tls_get_addr (void *); \
- asm ("daddiu %0, $28, %%tlsgd(" #x ")" \
- : "=r" (__result)); \
+ asm (LOAD_GP ADDIU " %0, $28, %%tlsgd(" #x ")" \
+ UNLOAD_GP \
+ : "=r" (__result), [tmp] "=&r" (__tmp)); \
(int *)__tls_get_addr (__result); })
-#endif
-
-#if _MIPS_SIM != _ABI64
# define TLS_LD(x) \
- ({ void *__result; \
+ ({ void *__result, *__tmp; \
extern void *__tls_get_addr (void *); \
- asm ("addiu %0, $28, %%tlsldm(" #x ")" \
- : "=r" (__result)); \
+ asm (LOAD_GP ADDIU " %0, $28, %%tlsldm(" #x ")" \
+ UNLOAD_GP \
+ : "=r" (__result), [tmp] "=&r" (__tmp)); \
__result = __tls_get_addr (__result); \
asm ("lui $3,%%dtprel_hi(" #x ")\n\t" \
"addiu $3,$3,%%dtprel_lo(" #x ")\n\t" \
- "addu %0,%0,$3" \
+ ADDU " %0,%0,$3" \
: "+r" (__result) : : "$3"); \
__result; })
# define TLS_IE(x) \
- ({ void *__result; \
+ ({ void *__result, *__tmp; \
asm (".set push\n\t.set mips32r2\n\t" \
"rdhwr\t%0,$29\n\t.set pop" \
: "=v" (__result)); \
- asm ("lw $3,%%gottprel(" #x ")($28)\n\t" \
- "addu %0,%0,$3" \
- : "+r" (__result) : : "$3"); \
+ asm (LOAD_GP LW " $3,%%gottprel(" #x ")($28)\n\t" \
+ ADDU " %0,%0,$3" \
+ UNLOAD_GP \
+ : "+r" (__result), [tmp] "=&r" (__tmp) \
+ : : "$3"); \
__result; })
# define TLS_LE(x) \
({ void *__result; \
@@ -47,42 +59,6 @@
: "=v" (__result)); \
asm ("lui $3,%%tprel_hi(" #x ")\n\t" \
"addiu $3,$3,%%tprel_lo(" #x ")\n\t" \
- "addu %0,%0,$3" \
- : "+r" (__result) : : "$3"); \
- __result; })
-
-#else
-
-/* These versions are for n64. */
-
-# define TLS_LD(x) \
- ({ void *__result; \
- extern void *__tls_get_addr (void *); \
- asm ("daddiu %0, $28, %%tlsldm(" #x ")" \
- : "=r" (__result)); \
- __result = __tls_get_addr (__result); \
- asm ("lui $3,%%dtprel_hi(" #x ")\n\t" \
- "daddiu $3,$3,%%dtprel_lo(" #x ")\n\t" \
- "daddu %0,%0,$3" \
- : "+r" (__result) : : "$3"); \
- __result; })
-# define TLS_IE(x) \
- ({ void *__result; \
- asm (".set push\n\t.set mips32r2\n\t" \
- "rdhwr\t%0,$29\n\t.set pop" \
- : "=v" (__result)); \
- asm ("ld $3,%%gottprel(" #x ")($28)\n\t" \
- "daddu %0,%0,$3" \
- : "+r" (__result) : : "$3"); \
- __result; })
-# define TLS_LE(x) \
- ({ void *__result; \
- asm (".set push\n\t.set mips32r2\n\t" \
- "rdhwr\t%0,$29\n\t.set pop" \
- : "=v" (__result)); \
- asm ("lui $3,%%tprel_hi(" #x ")\n\t" \
- "daddiu $3,$3,%%tprel_lo(" #x ")\n\t" \
- "daddu %0,%0,$3" \
+ ADDU " %0,%0,$3" \
: "+r" (__result) : : "$3"); \
__result; })
-#endif
Index: sysdeps/mips/dl-machine.h
===================================================================
--- sysdeps/mips/dl-machine.h (revision 213009)
+++ sysdeps/mips/dl-machine.h (working copy)
@@ -25,8 +25,6 @@
#define ELF_MACHINE_NAME "MIPS"
-#define ELF_MACHINE_NO_PLT
-
#include <entry.h>
#ifndef ENTRY_POINT
@@ -56,10 +54,14 @@
#endif
/* A reloc type used for ld.so cmdline arg lookups to reject PLT entries.
- This makes no sense on MIPS but we have to define this to R_MIPS_REL32
- to avoid the asserts in dl-lookup.c from blowing. */
-#define ELF_MACHINE_JMP_SLOT R_MIPS_REL32
-#define elf_machine_type_class(type) ELF_RTYPE_CLASS_PLT
+ This only makes sense on MIPS when using PLTs, so choose the
+ PLT relocation (not encountered when not using PLTs). */
+#define ELF_MACHINE_JMP_SLOT R_MIPS_JUMP_SLOT
+#define elf_machine_type_class(type) \
+ ((((type) == ELF_MACHINE_JMP_SLOT) * ELF_RTYPE_CLASS_PLT) \
+ | (((type) == R_MIPS_COPY) * ELF_RTYPE_CLASS_COPY))
+
+#define ELF_MACHINE_PLT_REL 1
/* Translate a processor specific dynamic tag to the index
in l_info array. */
@@ -73,6 +75,15 @@ do { if ((l)->l_info[DT_MIPS (RLD_MAP)])
(ElfW(Addr)) (r); \
} while (0)
+/* Allow ABIVERSION == 1, meaning PLTs and copy relocations are
+ required. */
+#define VALID_ELF_ABIVERSION(ver) (ver == 0 || ver == 2)
+#define VALID_ELF_OSABI(osabi) (osabi == ELFOSABI_SYSV)
+#define VALID_ELF_HEADER(hdr,exp,size) \
+ memcmp (hdr,exp,size-2) == 0 \
+ && VALID_ELF_OSABI (hdr[EI_OSABI]) \
+ && VALID_ELF_ABIVERSION (hdr[EI_ABIVERSION])
+
/* Return nonzero iff ELF header is compatible with the running host. */
static inline int __attribute_used__
elf_machine_matches_host (const ElfW(Ehdr) *ehdr)
@@ -294,6 +305,24 @@ do { \
# define ARCH_LA_PLTEXIT mips_n64_gnu_pltexit
# endif
+/* For a non-writable PLT, rewrite the .got.plt entry at RELOC_ADDR to
+ point at the symbol with address VALUE. For a writable PLT, rewrite
+ the corresponding PLT entry instead. */
+static inline ElfW(Addr)
+elf_machine_fixup_plt (struct link_map *map, lookup_t t,
+ const ElfW(Rel) *reloc,
+ ElfW(Addr) *reloc_addr, ElfW(Addr) value)
+{
+ return *reloc_addr = value;
+}
+
+static inline ElfW(Addr)
+elf_machine_plt_value (struct link_map *map, const ElfW(Rel) *reloc,
+ ElfW(Addr) value)
+{
+ return value;
+}
+
#endif /* !dl_machine_h */
#ifdef RESOLVE_MAP
@@ -461,6 +490,57 @@ elf_machine_reloc (struct link_map *map,
#endif
case R_MIPS_NONE: /* Alright, Wilbur. */
break;
+
+ case R_MIPS_JUMP_SLOT:
+ {
+ struct link_map *sym_map;
+ ElfW(Addr) value;
+
+ /* The addend for a jump slot relocation must always be zero:
+ calls via the PLT always branch to the symbol's address and
+ not to the address plus a non-zero offset. */
+ if (r_addend != 0)
+ _dl_signal_error (0, map->l_name, NULL,
+ "found jump slot relocation with non-zero addend");
+
+ sym_map = RESOLVE_MAP (&sym, version, r_type);
+ value = sym_map == NULL ? 0 : sym_map->l_addr + sym->st_value;
+ *addr_field = value;
+
+ break;
+ }
+
+ case R_MIPS_COPY:
+ {
+ const ElfW(Sym) *const refsym = sym;
+ struct link_map *sym_map;
+ ElfW(Addr) value;
+
+ /* Calculate the address of the symbol. */
+ sym_map = RESOLVE_MAP (&sym, version, r_type);
+ value = sym_map == NULL ? 0 : sym_map->l_addr + sym->st_value;
+
+ if (__builtin_expect (sym == NULL, 0))
+ /* This can happen in trace mode if an object could not be
+ found. */
+ break;
+ if (__builtin_expect (sym->st_size > refsym->st_size, 0)
+ || (__builtin_expect (sym->st_size < refsym->st_size, 0)
+ && GLRO(dl_verbose)))
+ {
+ const char *strtab;
+
+ strtab = (const void *) D_PTR (map, l_info[DT_STRTAB]);
+ _dl_error_printf ("\
+ %s: Symbol `%s' has different size in shared object, consider re-linking\n",
+ rtld_progname ?: "<program name unknown>",
+ strtab + refsym->st_name);
+ }
+ memcpy (reloc_addr, (void *) value,
+ MIN (sym->st_size, refsym->st_size));
+ break;
+ }
+
#if _MIPS_SIM == _ABI64
case R_MIPS_64:
/* For full compliance with the ELF64 ABI, one must precede the
@@ -505,9 +585,23 @@ elf_machine_rel_relative (ElfW(Addr) l_a
auto inline void
__attribute__((always_inline))
elf_machine_lazy_rel (struct link_map *map,
- ElfW(Addr) l_addr, const ElfW(Rela) *reloc)
+ ElfW(Addr) l_addr, const ElfW(Rel) *reloc)
{
- /* Do nothing. */
+ ElfW(Addr) *const reloc_addr = (void *) (l_addr + reloc->r_offset);
+ const unsigned int r_type = ELFW(R_TYPE) (reloc->r_info);
+ /* Check for unexpected PLT reloc type. */
+ if (__builtin_expect (r_type == R_MIPS_JUMP_SLOT, 1))
+ {
+ if (__builtin_expect (map->l_mach.plt, 0) == 0)
+ {
+ /* Nothing is required here since we only support lazy
+ relocation in executables. */
+ }
+ else
+ *reloc_addr = map->l_mach.plt;
+ }
+ else
+ _dl_reloc_bad_type (map, r_type, 1);
}
auto inline void
@@ -538,13 +632,13 @@ elf_machine_got_rel (struct link_map *ma
const ElfW(Half) *vernum;
int i, n, symidx;
-#define RESOLVE_GOTSYM(sym,vernum,sym_index) \
+#define RESOLVE_GOTSYM(sym,vernum,sym_index,reloc) \
({ \
const ElfW(Sym) *ref = sym; \
const struct r_found_version *version \
= vernum ? &map->l_versions[vernum[sym_index] & 0x7fff] : NULL; \
struct link_map *sym_map; \
- sym_map = RESOLVE_MAP (&ref, version, R_MIPS_REL32); \
+ sym_map = RESOLVE_MAP (&ref, version, reloc); \
ref ? sym_map->l_addr + ref->st_value : 0; \
})
@@ -585,25 +679,38 @@ elf_machine_got_rel (struct link_map *ma
{
if (sym->st_shndx == SHN_UNDEF)
{
- if (ELFW(ST_TYPE) (sym->st_info) == STT_FUNC
- && sym->st_value && lazy)
- *got = sym->st_value + map->l_addr;
+ if (ELFW(ST_TYPE) (sym->st_info) == STT_FUNC && sym->st_value
+ && !(sym->st_other & STO_MIPS_PLT))
+ {
+ if (lazy)
+ *got = sym->st_value + map->l_addr;
+ else
+ /* This is a lazy-binding stub, so we don't need the
+ canonical address. */
+ *got = RESOLVE_GOTSYM (sym, vernum, symidx, R_MIPS_JUMP_SLOT);
+ }
else
- *got = RESOLVE_GOTSYM (sym, vernum, symidx);
+ *got = RESOLVE_GOTSYM (sym, vernum, symidx, R_MIPS_32);
}
else if (sym->st_shndx == SHN_COMMON)
- *got = RESOLVE_GOTSYM (sym, vernum, symidx);
+ *got = RESOLVE_GOTSYM (sym, vernum, symidx, R_MIPS_32);
else if (ELFW(ST_TYPE) (sym->st_info) == STT_FUNC
- && *got != sym->st_value
- && lazy)
- *got += map->l_addr;
+ && *got != sym->st_value)
+ {
+ if (lazy)
+ *got += map->l_addr;
+ else
+ /* This is a lazy-binding stub, so we don't need the
+ canonical address. */
+ *got = RESOLVE_GOTSYM (sym, vernum, symidx, R_MIPS_JUMP_SLOT);
+ }
else if (ELFW(ST_TYPE) (sym->st_info) == STT_SECTION)
{
if (sym->st_other == 0)
*got += map->l_addr;
}
else
- *got = RESOLVE_GOTSYM (sym, vernum, symidx);
+ *got = RESOLVE_GOTSYM (sym, vernum, symidx, R_MIPS_32);
++got;
++sym;
@@ -624,6 +731,7 @@ elf_machine_runtime_setup (struct link_m
# ifndef RTLD_BOOTSTRAP
ElfW(Addr) *got;
extern void _dl_runtime_resolve (ElfW(Word));
+ extern void _dl_runtime_pltresolve (void);
extern int _dl_mips_gnu_objects;
if (lazy)
@@ -650,6 +758,20 @@ elf_machine_runtime_setup (struct link_m
/* Relocate global offset table. */
elf_machine_got_rel (l, lazy);
+ /* If using PLTs, fill in the first two entries of .got.plt. */
+ if (l->l_info[DT_JMPREL] && lazy)
+ {
+ ElfW(Addr) *gotplt;
+ gotplt = (ElfW(Addr) *) D_PTR (l, l_info[DT_MIPS (PLTGOT)]);
+ /* If a library is prelinked but we have to relocate anyway,
+ we have to be able to undo the prelinking of .got.plt.
+ The prelinker saved the address of .plt for us here. */
+ if (gotplt[1])
+ l->l_mach.plt = gotplt[1] + l->l_addr;
+ gotplt[0] = (ElfW(Addr)) &_dl_runtime_pltresolve;
+ gotplt[1] = (ElfW(Addr)) l;
+ }
+
# endif
return lazy;
}
Index: sysdeps/mips/dl-lookup.c
===================================================================
--- sysdeps/mips/dl-lookup.c (revision 0)
+++ sysdeps/mips/dl-lookup.c (revision 0)
@@ -0,0 +1,581 @@
+/* Look up a symbol in the loaded objects.
+ MIPS/Linux version - this is identical to the common version, but
+ because it is in sysdeps/mips, it gets sysdeps/mips/do-lookup.h.
+ Using <do-lookup.h> instead of "do-lookup.h" would work too.
+
+ Copyright (C) 1995-2005, 2006, 2007 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <alloca.h>
+#include <libintl.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <ldsodefs.h>
+#include <dl-hash.h>
+#include <dl-machine.h>
+#include <sysdep-cancel.h>
+#include <bits/libc-lock.h>
+#include <tls.h>
+
+#include <assert.h>
+
+#define VERSTAG(tag) (DT_NUM + DT_THISPROCNUM + DT_VERSIONTAGIDX (tag))
+
+/* We need this string more than once. */
+static const char undefined_msg[] = "undefined symbol: ";
+
+
+struct sym_val
+ {
+ const ElfW(Sym) *s;
+ struct link_map *m;
+ };
+
+
+#define make_string(string, rest...) \
+ ({ \
+ const char *all[] = { string, ## rest }; \
+ size_t len, cnt; \
+ char *result, *cp; \
+ \
+ len = 1; \
+ for (cnt = 0; cnt < sizeof (all) / sizeof (all[0]); ++cnt) \
+ len += strlen (all[cnt]); \
+ \
+ cp = result = alloca (len); \
+ for (cnt = 0; cnt < sizeof (all) / sizeof (all[0]); ++cnt) \
+ cp = __stpcpy (cp, all[cnt]); \
+ \
+ result; \
+ })
+
+/* Statistics function. */
+#ifdef SHARED
+# define bump_num_relocations() ++GL(dl_num_relocations)
+#else
+# define bump_num_relocations() ((void) 0)
+#endif
+
+
+/* The actual lookup code. */
+#include "do-lookup.h"
+
+
+static uint_fast32_t
+dl_new_hash (const char *s)
+{
+ uint_fast32_t h = 5381;
+ for (unsigned char c = *s; c != '\0'; c = *++s)
+ h = h * 33 + c;
+ return h & 0xffffffff;
+}
+
+
+/* Add extra dependency on MAP to UNDEF_MAP. */
+static int
+internal_function
+add_dependency (struct link_map *undef_map, struct link_map *map, int flags)
+{
+ struct link_map *runp;
+ unsigned int i;
+ int result = 0;
+
+ /* Avoid self-references and references to objects which cannot be
+ unloaded anyway. */
+ if (undef_map == map)
+ return 0;
+
+ /* Avoid references to objects which cannot be unloaded anyway. */
+ assert (map->l_type == lt_loaded);
+ if ((map->l_flags_1 & DF_1_NODELETE) != 0)
+ return 0;
+
+ struct link_map_reldeps *l_reldeps
+ = atomic_forced_read (undef_map->l_reldeps);
+
+ /* Make sure l_reldeps is read before l_initfini. */
+ atomic_read_barrier ();
+
+ /* Determine whether UNDEF_MAP already has a reference to MAP. First
+ look in the normal dependencies. */
+ struct link_map **l_initfini = atomic_forced_read (undef_map->l_initfini);
+ if (l_initfini != NULL)
+ {
+ for (i = 0; l_initfini[i] != NULL; ++i)
+ if (l_initfini[i] == map)
+ return 0;
+ }
+
+ /* No normal dependency. See whether we already had to add it
+ to the special list of dynamic dependencies. */
+ unsigned int l_reldepsact = 0;
+ if (l_reldeps != NULL)
+ {
+ struct link_map **list = &l_reldeps->list[0];
+ l_reldepsact = l_reldeps->act;
+ for (i = 0; i < l_reldepsact; ++i)
+ if (list[i] == map)
+ return 0;
+ }
+
+ /* Save serial number of the target MAP. */
+ unsigned long long serial = map->l_serial;
+
+ /* Make sure nobody can unload the object while we are at it. */
+ if (__builtin_expect (flags & DL_LOOKUP_GSCOPE_LOCK, 0))
+ {
+ /* We can't just call __rtld_lock_lock_recursive (GL(dl_load_lock))
+ here, that can result in ABBA deadlock. */
+ THREAD_GSCOPE_RESET_FLAG ();
+ __rtld_lock_lock_recursive (GL(dl_load_lock));
+ /* While MAP value won't change, after THREAD_GSCOPE_RESET_FLAG ()
+ it can e.g. point to unallocated memory. So avoid the optimizer
+ treating the above read from MAP->l_serial as ensurance it
+ can safely dereference it. */
+ map = atomic_forced_read (map);
+
+ /* From this point on it is unsafe to dereference MAP, until it
+ has been found in one of the lists. */
+
+ /* Redo the l_initfini check in case undef_map's l_initfini
+ changed in the mean time. */
+ if (undef_map->l_initfini != l_initfini
+ && undef_map->l_initfini != NULL)
+ {
+ l_initfini = undef_map->l_initfini;
+ for (i = 0; l_initfini[i] != NULL; ++i)
+ if (l_initfini[i] == map)
+ goto out_check;
+ }
+
+ /* Redo the l_reldeps check if undef_map's l_reldeps changed in
+ the mean time. */
+ if (undef_map->l_reldeps != NULL)
+ {
+ if (undef_map->l_reldeps != l_reldeps)
+ {
+ struct link_map **list = &undef_map->l_reldeps->list[0];
+ l_reldepsact = undef_map->l_reldeps->act;
+ for (i = 0; i < l_reldepsact; ++i)
+ if (list[i] == map)
+ goto out_check;
+ }
+ else if (undef_map->l_reldeps->act > l_reldepsact)
+ {
+ struct link_map **list
+ = &undef_map->l_reldeps->list[0];
+ i = l_reldepsact;
+ l_reldepsact = undef_map->l_reldeps->act;
+ for (; i < l_reldepsact; ++i)
+ if (list[i] == map)
+ goto out_check;
+ }
+ }
+ }
+ else
+ __rtld_lock_lock_recursive (GL(dl_load_lock));
+
+ /* The object is not yet in the dependency list. Before we add
+ it make sure just one more time the object we are about to
+ reference is still available. There is a brief period in
+ which the object could have been removed since we found the
+ definition. */
+ runp = GL(dl_ns)[undef_map->l_ns]._ns_loaded;
+ while (runp != NULL && runp != map)
+ runp = runp->l_next;
+
+ if (runp != NULL)
+ {
+ /* The object is still available. */
+
+ /* MAP could have been dlclosed, freed and then some other dlopened
+ library could have the same link_map pointer. */
+ if (map->l_serial != serial)
+ goto out_check;
+
+ /* Redo the NODELETE check, as when dl_load_lock wasn't held
+ yet this could have changed. */
+ if ((map->l_flags_1 & DF_1_NODELETE) != 0)
+ goto out;
+
+ /* If the object with the undefined reference cannot be removed ever
+ just make sure the same is true for the object which contains the
+ definition. */
+ if (undef_map->l_type != lt_loaded
+ || (undef_map->l_flags_1 & DF_1_NODELETE) != 0)
+ {
+ map->l_flags_1 |= DF_1_NODELETE;
+ goto out;
+ }
+
+ /* Add the reference now. */
+ if (__builtin_expect (l_reldepsact >= undef_map->l_reldepsmax, 0))
+ {
+ /* Allocate more memory for the dependency list. Since this
+ can never happen during the startup phase we can use
+ `realloc'. */
+ struct link_map_reldeps *newp;
+ unsigned int max
+ = undef_map->l_reldepsmax ? undef_map->l_reldepsmax * 2 : 10;
+
+ newp = malloc (sizeof (*newp) + max * sizeof (struct link_map *));
+ if (newp == NULL)
+ {
+ /* If we didn't manage to allocate memory for the list this is
+ no fatal problem. We simply make sure the referenced object
+ cannot be unloaded. This is semantically the correct
+ behavior. */
+ map->l_flags_1 |= DF_1_NODELETE;
+ goto out;
+ }
+ else
+ {
+ if (l_reldepsact)
+ memcpy (&newp->list[0], &undef_map->l_reldeps->list[0],
+ l_reldepsact * sizeof (struct link_map *));
+ newp->list[l_reldepsact] = map;
+ newp->act = l_reldepsact + 1;
+ atomic_write_barrier ();
+ void *old = undef_map->l_reldeps;
+ undef_map->l_reldeps = newp;
+ undef_map->l_reldepsmax = max;
+ if (old)
+ _dl_scope_free (old);
+ }
+ }
+ else
+ {
+ undef_map->l_reldeps->list[l_reldepsact] = map;
+ atomic_write_barrier ();
+ undef_map->l_reldeps->act = l_reldepsact + 1;
+ }
+
+ /* Display information if we are debugging. */
+ if (__builtin_expect (GLRO(dl_debug_mask) & DL_DEBUG_FILES, 0))
+ _dl_debug_printf ("\
+\nfile=%s [%lu]; needed by %s [%lu] (relocation dependency)\n\n",
+ map->l_name[0] ? map->l_name : rtld_progname,
+ map->l_ns,
+ undef_map->l_name[0]
+ ? undef_map->l_name : rtld_progname,
+ undef_map->l_ns);
+ }
+ else
+ /* Whoa, that was bad luck. We have to search again. */
+ result = -1;
+
+ out:
+ /* Release the lock. */
+ __rtld_lock_unlock_recursive (GL(dl_load_lock));
+
+ if (__builtin_expect (flags & DL_LOOKUP_GSCOPE_LOCK, 0))
+ THREAD_GSCOPE_SET_FLAG ();
+
+ return result;
+
+ out_check:
+ if (map->l_serial != serial)
+ result = -1;
+ goto out;
+}
+
+static void
+internal_function
+_dl_debug_bindings (const char *undef_name, struct link_map *undef_map,
+ const ElfW(Sym) **ref, struct sym_val *value,
+ const struct r_found_version *version, int type_class,
+ int protected);
+
+
+/* Search loaded objects' symbol tables for a definition of the symbol
+ UNDEF_NAME, perhaps with a requested version for the symbol.
+
+ We must never have calls to the audit functions inside this function
+ or in any function which gets called. If this would happen the audit
+ code might create a thread which can throw off all the scope locking. */
+lookup_t
+internal_function
+_dl_lookup_symbol_x (const char *undef_name, struct link_map *undef_map,
+ const ElfW(Sym) **ref,
+ struct r_scope_elem *symbol_scope[],
+ const struct r_found_version *version,
+ int type_class, int flags, struct link_map *skip_map)
+{
+ const uint_fast32_t new_hash = dl_new_hash (undef_name);
+ unsigned long int old_hash = 0xffffffff;
+ struct sym_val current_value = { NULL, NULL };
+ struct r_scope_elem **scope = symbol_scope;
+
+ bump_num_relocations ();
+
+ /* No other flag than DL_LOOKUP_ADD_DEPENDENCY or DL_LOOKUP_GSCOPE_LOCK
+ is allowed if we look up a versioned symbol. */
+ assert (version == NULL
+ || (flags & ~(DL_LOOKUP_ADD_DEPENDENCY | DL_LOOKUP_GSCOPE_LOCK))
+ == 0);
+
+ size_t i = 0;
+ if (__builtin_expect (skip_map != NULL, 0))
+ /* Search the relevant loaded objects for a definition. */
+ while ((*scope)->r_list[i] != skip_map)
+ ++i;
+
+ /* Search the relevant loaded objects for a definition. */
+ for (size_t start = i; *scope != NULL; start = 0, ++scope)
+ {
+ int res = do_lookup_x (undef_name, new_hash, &old_hash, *ref,
+ ¤t_value, *scope, start, version, flags,
+ skip_map, type_class);
+ if (res > 0)
+ break;
+
+ if (__builtin_expect (res, 0) < 0 && skip_map == NULL)
+ {
+ /* Oh, oh. The file named in the relocation entry does not
+ contain the needed symbol. This code is never reached
+ for unversioned lookups. */
+ assert (version != NULL);
+ const char *reference_name = undef_map ? undef_map->l_name : NULL;
+
+ /* XXX We cannot translate the message. */
+ _dl_signal_cerror (0, (reference_name[0]
+ ? reference_name
+ : (rtld_progname ?: "<main program>")),
+ N_("relocation error"),
+ make_string ("symbol ", undef_name, ", version ",
+ version->name,
+ " not defined in file ",
+ version->filename,
+ " with link time reference",
+ res == -2
+ ? " (no version symbols)" : ""));
+ *ref = NULL;
+ return 0;
+ }
+ }
+
+ if (__builtin_expect (current_value.s == NULL, 0))
+ {
+ if ((*ref == NULL || ELFW(ST_BIND) ((*ref)->st_info) != STB_WEAK)
+ && skip_map == NULL)
+ {
+ /* We could find no value for a strong reference. */
+ const char *reference_name = undef_map ? undef_map->l_name : "";
+ const char *versionstr = version ? ", version " : "";
+ const char *versionname = (version && version->name
+ ? version->name : "");
+
+ /* XXX We cannot translate the message. */
+ _dl_signal_cerror (0, (reference_name[0]
+ ? reference_name
+ : (rtld_progname ?: "<main program>")),
+ N_("symbol lookup error"),
+ make_string (undefined_msg, undef_name,
+ versionstr, versionname));
+ }
+ *ref = NULL;
+ return 0;
+ }
+
+ int protected = (*ref
+ && ELFW(ST_VISIBILITY) ((*ref)->st_other) == STV_PROTECTED);
+ if (__builtin_expect (protected != 0, 0))
+ {
+ /* It is very tricky. We need to figure out what value to
+ return for the protected symbol. */
+ if (type_class == ELF_RTYPE_CLASS_PLT)
+ {
+ if (current_value.s != NULL && current_value.m != undef_map)
+ {
+ current_value.s = *ref;
+ current_value.m = undef_map;
+ }
+ }
+ else
+ {
+ struct sym_val protected_value = { NULL, NULL };
+
+ for (scope = symbol_scope; *scope != NULL; i = 0, ++scope)
+ if (do_lookup_x (undef_name, new_hash, &old_hash, *ref,
+ &protected_value, *scope, i, version, flags,
+ skip_map, ELF_RTYPE_CLASS_PLT) != 0)
+ break;
+
+ if (protected_value.s != NULL && protected_value.m != undef_map)
+ {
+ current_value.s = *ref;
+ current_value.m = undef_map;
+ }
+ }
+ }
+
+ /* We have to check whether this would bind UNDEF_MAP to an object
+ in the global scope which was dynamically loaded. In this case
+ we have to prevent the latter from being unloaded unless the
+ UNDEF_MAP object is also unloaded. */
+ if (__builtin_expect (current_value.m->l_type == lt_loaded, 0)
+ /* Don't do this for explicit lookups as opposed to implicit
+ runtime lookups. */
+ && (flags & DL_LOOKUP_ADD_DEPENDENCY) != 0
+ /* Add UNDEF_MAP to the dependencies. */
+ && add_dependency (undef_map, current_value.m, flags) < 0)
+ /* Something went wrong. Perhaps the object we tried to reference
+ was just removed. Try finding another definition. */
+ return _dl_lookup_symbol_x (undef_name, undef_map, ref,
+ (flags & DL_LOOKUP_GSCOPE_LOCK)
+ ? undef_map->l_scope : symbol_scope,
+ version, type_class, flags, skip_map);
+
+ /* The object is used. */
+ current_value.m->l_used = 1;
+
+ if (__builtin_expect (GLRO(dl_debug_mask)
+ & (DL_DEBUG_BINDINGS|DL_DEBUG_PRELINK), 0))
+ _dl_debug_bindings (undef_name, undef_map, ref,
+ ¤t_value, version, type_class, protected);
+
+ *ref = current_value.s;
+ return LOOKUP_VALUE (current_value.m);
+}
+
+
+/* Cache the location of MAP's hash table. */
+
+void
+internal_function
+_dl_setup_hash (struct link_map *map)
+{
+ Elf_Symndx *hash;
+ Elf_Symndx nchain;
+
+ if (__builtin_expect (map->l_info[DT_ADDRTAGIDX (DT_GNU_HASH) + DT_NUM
+ + DT_THISPROCNUM + DT_VERSIONTAGNUM
+ + DT_EXTRANUM + DT_VALNUM] != NULL, 1))
+ {
+ Elf32_Word *hash32
+ = (void *) D_PTR (map, l_info[DT_ADDRTAGIDX (DT_GNU_HASH) + DT_NUM
+ + DT_THISPROCNUM + DT_VERSIONTAGNUM
+ + DT_EXTRANUM + DT_VALNUM]);
+ map->l_nbuckets = *hash32++;
+ Elf32_Word symbias = *hash32++;
+ Elf32_Word bitmask_nwords = *hash32++;
+ /* Must be a power of two. */
+ assert ((bitmask_nwords & (bitmask_nwords - 1)) == 0);
+ map->l_gnu_bitmask_idxbits = bitmask_nwords - 1;
+ map->l_gnu_shift = *hash32++;
+
+ map->l_gnu_bitmask = (ElfW(Addr) *) hash32;
+ hash32 += __ELF_NATIVE_CLASS / 32 * bitmask_nwords;
+
+ map->l_gnu_buckets = hash32;
+ hash32 += map->l_nbuckets;
+ map->l_gnu_chain_zero = hash32 - symbias;
+ return;
+ }
+
+ if (!map->l_info[DT_HASH])
+ return;
+ hash = (void *) D_PTR (map, l_info[DT_HASH]);
+
+ map->l_nbuckets = *hash++;
+ nchain = *hash++;
+ map->l_buckets = hash;
+ hash += map->l_nbuckets;
+ map->l_chain = hash;
+}
+
+
+static void
+internal_function
+_dl_debug_bindings (const char *undef_name, struct link_map *undef_map,
+ const ElfW(Sym) **ref, struct sym_val *value,
+ const struct r_found_version *version, int type_class,
+ int protected)
+{
+ const char *reference_name = undef_map->l_name;
+
+ if (GLRO(dl_debug_mask) & DL_DEBUG_BINDINGS)
+ {
+ _dl_debug_printf ("binding file %s [%lu] to %s [%lu]: %s symbol `%s'",
+ (reference_name[0]
+ ? reference_name
+ : (rtld_progname ?: "<main program>")),
+ undef_map->l_ns,
+ value->m->l_name[0] ? value->m->l_name : rtld_progname,
+ value->m->l_ns,
+ protected ? "protected" : "normal", undef_name);
+ if (version)
+ _dl_debug_printf_c (" [%s]\n", version->name);
+ else
+ _dl_debug_printf_c ("\n");
+ }
+#ifdef SHARED
+ if (GLRO(dl_debug_mask) & DL_DEBUG_PRELINK)
+ {
+ int conflict = 0;
+ struct sym_val val = { NULL, NULL };
+
+ if ((GLRO(dl_trace_prelink_map) == NULL
+ || GLRO(dl_trace_prelink_map) == GL(dl_ns)[LM_ID_BASE]._ns_loaded)
+ && undef_map != GL(dl_ns)[LM_ID_BASE]._ns_loaded)
+ {
+ const uint_fast32_t new_hash = dl_new_hash (undef_name);
+ unsigned long int old_hash = 0xffffffff;
+
+ do_lookup_x (undef_name, new_hash, &old_hash, *ref, &val,
+ undef_map->l_local_scope[0], 0, version, 0, NULL,
+ type_class);
+
+ if (val.s != value->s || val.m != value->m)
+ conflict = 1;
+ }
+
+ if (value->s
+ && (__builtin_expect (ELFW(ST_TYPE) (value->s->st_info)
+ == STT_TLS, 0)))
+ type_class = 4;
+
+ if (conflict
+ || GLRO(dl_trace_prelink_map) == undef_map
+ || GLRO(dl_trace_prelink_map) == NULL
+ || type_class == 4)
+ {
+ _dl_printf ("%s 0x%0*Zx 0x%0*Zx -> 0x%0*Zx 0x%0*Zx ",
+ conflict ? "conflict" : "lookup",
+ (int) sizeof (ElfW(Addr)) * 2,
+ (size_t) undef_map->l_map_start,
+ (int) sizeof (ElfW(Addr)) * 2,
+ (size_t) (((ElfW(Addr)) *ref) - undef_map->l_map_start),
+ (int) sizeof (ElfW(Addr)) * 2,
+ (size_t) (value->s ? value->m->l_map_start : 0),
+ (int) sizeof (ElfW(Addr)) * 2,
+ (size_t) (value->s ? value->s->st_value : 0));
+
+ if (conflict)
+ _dl_printf ("x 0x%0*Zx 0x%0*Zx ",
+ (int) sizeof (ElfW(Addr)) * 2,
+ (size_t) (val.s ? val.m->l_map_start : 0),
+ (int) sizeof (ElfW(Addr)) * 2,
+ (size_t) (val.s ? val.s->st_value : 0));
+
+ _dl_printf ("/%x %s\n", type_class, undef_name);
+ }
+ }
+#endif
+}
Index: sysdeps/mips/dl-trampoline.c
===================================================================
--- sysdeps/mips/dl-trampoline.c (revision 213009)
+++ sysdeps/mips/dl-trampoline.c (working copy)
@@ -201,6 +201,7 @@ __dl_runtime_resolve (ElfW(Word) sym_ind
"
#define IFABIO32(X) X
+#define IFNEWABI(X)
#else /* _MIPS_SIM == _ABIN32 || _MIPS_SIM == _ABI64 */
@@ -231,6 +232,7 @@ __dl_runtime_resolve (ElfW(Word) sym_ind
"
#define IFABIO32(X)
+#define IFNEWABI(X) X
#endif
@@ -270,3 +272,57 @@ _dl_runtime_resolve:\n\
.end _dl_runtime_resolve\n\
.previous\n\
");
+
+/* Assembler veneer called from the PLT header code when using PLTs.
+
+ Code in each PLT entry and the PLT header fills in the arguments to
+ this function:
+
+ - $15 (o32 t7, n32/n64 t3) - caller's return address
+ - $24 (t8) - PLT entry index
+ - $25 (t9) - address of _dl_runtime_pltresolve
+ - o32 $28 (gp), n32/n64 $14 (t2) - address of .got.plt
+
+ Different registers are used for .got.plt because the ABI was
+ originally designed for o32, where gp was available (call
+ clobbered). On n32/n64 gp is call saved.
+
+ _dl_fixup needs:
+
+ - $4 (a0) - link map address
+ - $5 (a1) - .rel.plt offset (== PLT entry index * 8) */
+
+asm ("\n\
+ .text\n\
+ .align 2\n\
+ .globl _dl_runtime_pltresolve\n\
+ .type _dl_runtime_pltresolve,@function\n\
+ .ent _dl_runtime_pltresolve\n\
+_dl_runtime_pltresolve:\n\
+ .frame $29, " STRINGXP(ELF_DL_FRAME_SIZE) ", $31\n\
+ .set noreorder\n\
+ # Save arguments and sp value in stack.\n\
+ " STRINGXP(PTR_SUBIU) " $29, " STRINGXP(ELF_DL_FRAME_SIZE) "\n\
+ " IFABIO32(STRINGXP(PTR_L) " $10, " STRINGXP(PTRSIZE) "($28)") "\n\
+ " IFNEWABI(STRINGXP(PTR_L) " $10, " STRINGXP(PTRSIZE) "($14)") "\n\
+ # Modify t9 ($25) so as to point .cpload instruction.\n\
+ " IFABIO32(STRINGXP(PTR_ADDIU) " $25, 12\n") "\
+ # Compute GP.\n\
+ " STRINGXP(SETUP_GP) "\n\
+ " STRINGXV(SETUP_GP64 (0, _dl_runtime_pltresolve)) "\n\
+ .set reorder\n\
+ " IFABIO32(STRINGXP(CPRESTORE(32))) "\n\
+ " ELF_DL_SAVE_ARG_REGS "\
+ move $4, $10\n\
+ " IFABIO32("sll $5, $24, 3") "\n\
+ " IFNEWABI("sll $5, $24, 4") "\n\
+ jal _dl_fixup\n\
+ " ELF_DL_RESTORE_ARG_REGS "\
+ " STRINGXP(RESTORE_GP64) "\n\
+ " STRINGXP(PTR_ADDIU) " $29, " STRINGXP(ELF_DL_FRAME_SIZE) "\n\
+ move $25, $2\n\
+ jr $25\n\
+ .end _dl_runtime_pltresolve\n\
+ .previous\n\
+");
+
Index: sysdeps/mips/bits/linkmap.h
===================================================================
--- sysdeps/mips/bits/linkmap.h (revision 0)
+++ sysdeps/mips/bits/linkmap.h (revision 0)
@@ -0,0 +1,4 @@
+struct link_map_machine
+ {
+ ElfW(Addr) plt; /* Address of .plt */
+ };
2008-03-14 Mark Shinwell <shinwell@codesourcery.com> * elf/elf.h (STO_MIPS_PLT): New. (R_MIPS_COPY): New. (R_MIPS_JUMP_SLOT): New. (R_MIPS_NUM): Redefine to 128. (DT_MIPS_PLTGOT): New. (DT_MIPS_RWPLT): New. (DT_MIPS_NUM): Redefine to 0x35. Index: elf/elf.h =================================================================== --- elf/elf.h (revision 213009) +++ elf/elf.h (working copy) @@ -1395,6 +1395,7 @@ typedef struct #define STO_MIPS_INTERNAL 0x1 #define STO_MIPS_HIDDEN 0x2 #define STO_MIPS_PROTECTED 0x3 +#define STO_MIPS_PLT 0x8 #define STO_MIPS_SC_ALIGN_UNUSED 0xff /* MIPS specific values for `st_info'. */ @@ -1541,8 +1542,10 @@ typedef struct #define R_MIPS_TLS_TPREL_HI16 49 /* TP-relative offset, high 16 bits */ #define R_MIPS_TLS_TPREL_LO16 50 /* TP-relative offset, low 16 bits */ #define R_MIPS_GLOB_DAT 51 +#define R_MIPS_COPY 126 +#define R_MIPS_JUMP_SLOT 127 /* Keep this the last entry. */ -#define R_MIPS_NUM 52 +#define R_MIPS_NUM 128 /* Legal values for p_type field of Elf32_Phdr. */ @@ -1608,7 +1611,13 @@ typedef struct #define DT_MIPS_COMPACT_SIZE 0x7000002f /* (O32)Size of compact rel section. */ #define DT_MIPS_GP_VALUE 0x70000030 /* GP value for aux GOTs. */ #define DT_MIPS_AUX_DYNAMIC 0x70000031 /* Address of aux .dynamic. */ -#define DT_MIPS_NUM 0x32 +/* The address of .got.plt in an executable using the new non-PIC ABI. */ +#define DT_MIPS_PLTGOT 0x70000032 +/* The base of the PLT in an executable using the new non-PIC ABI if that + PLT is writable. For a non-writable PLT, this is omitted or has a zero + value. */ +#define DT_MIPS_RWPLT 0x70000034 +#define DT_MIPS_NUM 0x35 /* Legal values for DT_MIPS_FLAGS Elf32_Dyn entry. */
2008-06-27 Daniel Jacobowitz <dan@codesourcery.com>
* longlong.h: Update from GCC.
Index: stdlib/longlong.h
===================================================================
RCS file: /cvs/glibc/libc/stdlib/longlong.h,v
retrieving revision 1.31
diff -u -p -r1.31 longlong.h
--- stdlib/longlong.h 15 May 2006 20:25:33 -0000 1.31
+++ stdlib/longlong.h 27 Jun 2008 13:05:23 -0000
@@ -229,6 +229,19 @@ UDItype __umulsidi3 (USItype, USItype);
#define UDIV_TIME 100
#endif /* __arm__ */
+#if defined(__arm__)
+/* Let gcc decide how best to implement count_leading_zeros. */
+#define count_leading_zeros(COUNT,X) ((COUNT) = __builtin_clz (X))
+#define COUNT_LEADING_ZEROS_0 32
+#endif
+
+#if defined (__CRIS__) && __CRIS_arch_version >= 3
+#define count_leading_zeros(COUNT, X) ((COUNT) = __builtin_clz (X))
+#if __CRIS_arch_version >= 8
+#define count_trailing_zeros(COUNT, X) ((COUNT) = __builtin_ctz (X))
+#endif
+#endif /* __CRIS__ */
+
#if defined (__hppa) && W_TYPE_SIZE == 32
#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
__asm__ ("add %4,%5,%1\n\taddc %2,%3,%0" \
@@ -315,7 +328,7 @@ UDItype __umulsidi3 (USItype, USItype);
#if (defined (__i386__) || defined (__i486__)) && W_TYPE_SIZE == 32
#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
- __asm__ ("addl %5,%1\n\tadcl %3,%0" \
+ __asm__ ("add{l} {%5,%1|%1,%5}\n\tadc{l} {%3,%0|%0,%3}" \
: "=r" ((USItype) (sh)), \
"=&r" ((USItype) (sl)) \
: "%0" ((USItype) (ah)), \
@@ -323,7 +336,7 @@ UDItype __umulsidi3 (USItype, USItype);
"%1" ((USItype) (al)), \
"g" ((USItype) (bl)))
#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
- __asm__ ("subl %5,%1\n\tsbbl %3,%0" \
+ __asm__ ("sub{l} {%5,%1|%1,%5}\n\tsbb{l} {%3,%0|%0,%3}" \
: "=r" ((USItype) (sh)), \
"=&r" ((USItype) (sl)) \
: "0" ((USItype) (ah)), \
@@ -331,31 +344,60 @@ UDItype __umulsidi3 (USItype, USItype);
"1" ((USItype) (al)), \
"g" ((USItype) (bl)))
#define umul_ppmm(w1, w0, u, v) \
- __asm__ ("mull %3" \
+ __asm__ ("mul{l} %3" \
: "=a" ((USItype) (w0)), \
"=d" ((USItype) (w1)) \
: "%0" ((USItype) (u)), \
"rm" ((USItype) (v)))
#define udiv_qrnnd(q, r, n1, n0, dv) \
- __asm__ ("divl %4" \
+ __asm__ ("div{l} %4" \
: "=a" ((USItype) (q)), \
"=d" ((USItype) (r)) \
: "0" ((USItype) (n0)), \
"1" ((USItype) (n1)), \
"rm" ((USItype) (dv)))
-#define count_leading_zeros(count, x) \
- do { \
- USItype __cbtmp; \
- __asm__ ("bsrl %1,%0" \
- : "=r" (__cbtmp) : "rm" ((USItype) (x))); \
- (count) = __cbtmp ^ 31; \
- } while (0)
-#define count_trailing_zeros(count, x) \
- __asm__ ("bsfl %1,%0" : "=r" (count) : "rm" ((USItype)(x)))
+#define count_leading_zeros(count, x) ((count) = __builtin_clz (x))
+#define count_trailing_zeros(count, x) ((count) = __builtin_ctz (x))
#define UMUL_TIME 40
#define UDIV_TIME 40
#endif /* 80x86 */
+#if (defined (__x86_64__) || defined (__i386__)) && W_TYPE_SIZE == 64
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+ __asm__ ("add{q} {%5,%1|%1,%5}\n\tadc{q} {%3,%0|%0,%3}" \
+ : "=r" ((UDItype) (sh)), \
+ "=&r" ((UDItype) (sl)) \
+ : "%0" ((UDItype) (ah)), \
+ "rme" ((UDItype) (bh)), \
+ "%1" ((UDItype) (al)), \
+ "rme" ((UDItype) (bl)))
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+ __asm__ ("sub{q} {%5,%1|%1,%5}\n\tsbb{q} {%3,%0|%0,%3}" \
+ : "=r" ((UDItype) (sh)), \
+ "=&r" ((UDItype) (sl)) \
+ : "0" ((UDItype) (ah)), \
+ "rme" ((UDItype) (bh)), \
+ "1" ((UDItype) (al)), \
+ "rme" ((UDItype) (bl)))
+#define umul_ppmm(w1, w0, u, v) \
+ __asm__ ("mul{q} %3" \
+ : "=a" ((UDItype) (w0)), \
+ "=d" ((UDItype) (w1)) \
+ : "%0" ((UDItype) (u)), \
+ "rm" ((UDItype) (v)))
+#define udiv_qrnnd(q, r, n1, n0, dv) \
+ __asm__ ("div{q} %4" \
+ : "=a" ((UDItype) (q)), \
+ "=d" ((UDItype) (r)) \
+ : "0" ((UDItype) (n0)), \
+ "1" ((UDItype) (n1)), \
+ "rm" ((UDItype) (dv)))
+#define count_leading_zeros(count, x) ((count) = __builtin_clzl (x))
+#define count_trailing_zeros(count, x) ((count) = __builtin_ctzl (x))
+#define UMUL_TIME 40
+#define UDIV_TIME 40
+#endif /* x86_64 */
+
#if defined (__i960__) && W_TYPE_SIZE == 32
#define umul_ppmm(w1, w0, u, v) \
({union {UDItype __ll; \
@@ -523,6 +565,11 @@ UDItype __umulsidi3 (USItype, USItype);
__asm__ ("bfffo %1{%b2:%b2},%0" \
: "=d" ((USItype) (count)) \
: "od" ((USItype) (x)), "n" (0))
+/* Some ColdFire architectures have a ff1 instruction supported via
+ __builtin_clz. */
+#elif defined (__mcfisaaplus__) || defined (__mcfisac__)
+#define count_leading_zeros(count,x) ((count) = __builtin_clz (x))
+#define COUNT_LEADING_ZEROS_0 32
#endif
#endif /* mc68000 */
@@ -585,14 +632,19 @@ UDItype __umulsidi3 (USItype, USItype);
#endif /* __m88000__ */
#if defined (__mips__) && W_TYPE_SIZE == 32
-#define umul_ppmm(w1, w0, u, v) \
- __asm__ ("multu %2,%3" \
- : "=l" ((USItype) (w0)), \
- "=h" ((USItype) (w1)) \
- : "d" ((USItype) (u)), \
- "d" ((USItype) (v)))
+#define umul_ppmm(w1, w0, u, v) \
+ do { \
+ UDItype __x = (UDItype) (USItype) (u) * (USItype) (v); \
+ (w1) = (USItype) (__x >> 32); \
+ (w0) = (USItype) (__x); \
+ } while (0)
#define UMUL_TIME 10
#define UDIV_TIME 100
+
+#if (__mips == 32 || __mips == 64) && ! __mips16
+#define count_leading_zeros(COUNT,X) ((COUNT) = __builtin_clz (X))
+#define COUNT_LEADING_ZEROS_0 32
+#endif
#endif /* __mips__ */
#if defined (__ns32000__) && W_TYPE_SIZE == 32
@@ -641,7 +693,9 @@ UDItype __umulsidi3 (USItype, USItype);
|| defined (__powerpc__) /* gcc */ \
|| defined (__POWERPC__) /* BEOS */ \
|| defined (__ppc__) /* Darwin */ \
- || defined (PPC) /* GNU/Linux, SysV */ \
+ || (defined (PPC) && ! defined (CPU_FAMILY)) /* gcc 2.7.x GNU&SysV */ \
+ || (defined (PPC) && defined (CPU_FAMILY) /* VxWorks */ \
+ && CPU_FAMILY == PPC) \
) && W_TYPE_SIZE == 32
#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
do { \
@@ -679,7 +733,10 @@ UDItype __umulsidi3 (USItype, USItype);
__asm__ ("{cntlz|cntlzw} %0,%1" : "=r" (count) : "r" (x))
#define COUNT_LEADING_ZEROS_0 32
#if defined (_ARCH_PPC) || defined (__powerpc__) || defined (__POWERPC__) \
- || defined (__ppc__) || defined (PPC)
+ || defined (__ppc__) \
+ || (defined (PPC) && ! defined (CPU_FAMILY)) /* gcc 2.7.x GNU&SysV */ \
+ || (defined (PPC) && defined (CPU_FAMILY) /* VxWorks */ \
+ && CPU_FAMILY == PPC)
#define umul_ppmm(ph, pl, m0, m1) \
do { \
USItype __m0 = (m0), __m1 = (m1); \
@@ -828,18 +885,51 @@ UDItype __umulsidi3 (USItype, USItype);
} while (0)
#endif
-#if defined (__sh2__) && W_TYPE_SIZE == 32
+#if defined(__sh__) && !__SHMEDIA__ && W_TYPE_SIZE == 32
+#ifndef __sh1__
#define umul_ppmm(w1, w0, u, v) \
__asm__ ( \
- "dmulu.l %2,%3\n\tsts macl,%1\n\tsts mach,%0" \
- : "=r" ((USItype)(w1)), \
- "=r" ((USItype)(w0)) \
+ "dmulu.l %2,%3\n\tsts%M1 macl,%1\n\tsts%M0 mach,%0" \
+ : "=r<" ((USItype)(w1)), \
+ "=r<" ((USItype)(w0)) \
: "r" ((USItype)(u)), \
"r" ((USItype)(v)) \
: "macl", "mach")
#define UMUL_TIME 5
#endif
+/* This is the same algorithm as __udiv_qrnnd_c. */
+#define UDIV_NEEDS_NORMALIZATION 1
+
+#define udiv_qrnnd(q, r, n1, n0, d) \
+ do { \
+ extern UWtype __udiv_qrnnd_16 (UWtype, UWtype) \
+ __attribute__ ((visibility ("hidden"))); \
+ /* r0: rn r1: qn */ /* r0: n1 r4: n0 r5: d r6: d1 */ /* r2: __m */ \
+ __asm__ ( \
+ "mov%M4 %4,r5\n" \
+" swap.w %3,r4\n" \
+" swap.w r5,r6\n" \
+" jsr @%5\n" \
+" shll16 r6\n" \
+" swap.w r4,r4\n" \
+" jsr @%5\n" \
+" swap.w r1,%0\n" \
+" or r1,%0" \
+ : "=r" (q), "=&z" (r) \
+ : "1" (n1), "r" (n0), "rm" (d), "r" (&__udiv_qrnnd_16) \
+ : "r1", "r2", "r4", "r5", "r6", "pr"); \
+ } while (0)
+
+#define UDIV_TIME 80
+
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+ __asm__ ("clrt;subc %5,%1; subc %4,%0" \
+ : "=r" (sh), "=r" (sl) \
+ : "0" (ah), "1" (al), "r" (bh), "r" (bl))
+
+#endif /* __sh__ */
+
#if defined (__SH5__) && __SHMEDIA__ && W_TYPE_SIZE == 32
#define __umulsidi3(u,v) ((UDItype)(USItype)u*(USItype)v)
#define count_leading_zeros(count, x) \
@@ -1153,6 +1243,23 @@ UDItype __umulsidi3 (USItype, USItype);
} while (0)
#endif /* __vax__ */
+#if defined (__xtensa__) && W_TYPE_SIZE == 32
+/* This code is not Xtensa-configuration-specific, so rely on the compiler
+ to expand builtin functions depending on what configuration features
+ are available. This avoids library calls when the operation can be
+ performed in-line. */
+#define umul_ppmm(w1, w0, u, v) \
+ do { \
+ DWunion __w; \
+ __w.ll = __builtin_umulsidi3 (u, v); \
+ w1 = __w.s.high; \
+ w0 = __w.s.low; \
+ } while (0)
+#define __umulsidi3(u, v) __builtin_umulsidi3 (u, v)
+#define count_leading_zeros(COUNT, X) ((COUNT) = __builtin_clz (X))
+#define count_trailing_zeros(COUNT, X) ((COUNT) = __builtin_ctz (X))
+#endif /* __xtensa__ */
+
#if defined (__z8000__) && W_TYPE_SIZE == 16
#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
__asm__ ("add %H1,%H5\n\tadc %H0,%H3" \
| Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
|---|---|---|
| Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |