2009-10-11 20:48:46 +00:00
|
|
|
/* DWARF2 EH unwinding support for SPARC Solaris.
|
2025-01-02 11:59:57 +01:00
|
|
|
Copyright (C) 2009-2025 Free Software Foundation, Inc.
|
2009-10-11 20:48:46 +00:00
|
|
|
|
|
|
|
This file is part of GCC.
|
|
|
|
|
|
|
|
GCC is free software; you can redistribute it and/or modify
|
|
|
|
it under the terms of the GNU General Public License as published by
|
|
|
|
the Free Software Foundation; either version 3, or (at your option)
|
|
|
|
any later version.
|
|
|
|
|
|
|
|
GCC 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 General Public License for more details.
|
|
|
|
|
|
|
|
Under Section 7 of GPL version 3, you are granted additional
|
|
|
|
permissions described in the GCC Runtime Library Exception, version
|
|
|
|
3.1, as published by the Free Software Foundation.
|
|
|
|
|
|
|
|
You should have received a copy of the GNU General Public License and
|
|
|
|
a copy of the GCC Runtime Library Exception along with this program;
|
|
|
|
see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
|
|
|
|
<http://www.gnu.org/licenses/>. */
|
|
|
|
|
|
|
|
/* Do code reading to identify a signal frame, and set the frame
|
|
|
|
state data appropriately. See unwind-dw2.c for the structs. */
|
|
|
|
|
|
|
|
#include <ucontext.h>
|
2011-02-14 12:32:11 +00:00
|
|
|
#include <sys/frame.h>
|
|
|
|
#include <sys/stack.h>
|
2009-10-11 20:48:46 +00:00
|
|
|
|
2011-12-11 11:23:15 +00:00
|
|
|
#ifdef __arch64__
|
2009-10-11 20:48:46 +00:00
|
|
|
|
2011-02-14 12:32:11 +00:00
|
|
|
#define IS_SIGHANDLER sparc64_is_sighandler
|
2009-10-11 20:48:46 +00:00
|
|
|
|
2011-02-14 12:32:11 +00:00
|
|
|
static int
|
2012-01-09 09:50:19 +00:00
|
|
|
sparc64_is_sighandler (unsigned int *pc, void *cfa, int *nframes)
|
2009-10-11 20:48:46 +00:00
|
|
|
{
|
2011-02-14 12:32:11 +00:00
|
|
|
if (/* Solaris 8+ - multi-threaded
|
|
|
|
----------------------------
|
|
|
|
<__sighndlr>: save %sp, -176, %sp
|
|
|
|
<__sighndlr+4>: mov %i0, %o0
|
|
|
|
<__sighndlr+8>: mov %i1, %o1
|
|
|
|
<__sighndlr+12>: call %i3
|
|
|
|
<__sighndlr+16>: mov %i2, %o2
|
|
|
|
<__sighndlr+20>: ret <--- PC
|
|
|
|
<__sighndlr+24>: restore */
|
|
|
|
pc[-5] == 0x9de3bf50
|
|
|
|
&& pc[-4] == 0x90100018
|
|
|
|
&& pc[-3] == 0x92100019
|
|
|
|
&& pc[-2] == 0x9fc6c000
|
|
|
|
&& pc[-1] == 0x9410001a
|
|
|
|
&& pc[ 0] == 0x81c7e008
|
|
|
|
&& pc[ 1] == 0x81e80000)
|
2009-10-11 20:48:46 +00:00
|
|
|
{
|
2012-01-09 09:50:19 +00:00
|
|
|
/* We have observed different calling frames among different
|
|
|
|
versions of the operating system, so that we need to
|
|
|
|
discriminate using the upper frame. We look for the return
|
|
|
|
address of the caller frame (there is an offset of 15 double
|
|
|
|
words between the frame address and the place where this return
|
|
|
|
address is stored) in order to do some more pattern matching. */
|
|
|
|
unsigned int cuh_pattern
|
|
|
|
= *(unsigned int *)(*(unsigned long *)(cfa + 15*8) - 4);
|
|
|
|
|
2012-02-15 08:13:09 +00:00
|
|
|
if (cuh_pattern == 0x92100019)
|
2014-04-22 12:30:59 +00:00
|
|
|
/* This matches the call_user_handler pattern in Solaris 11
|
|
|
|
libc.so.1:
|
|
|
|
|
|
|
|
<call_user_handler+864>: mov %i1, %o1
|
2019-05-28 17:27:51 +00:00
|
|
|
<call_user_handler+868>: call __sighndlr */
|
2012-01-09 09:50:19 +00:00
|
|
|
*nframes = 3;
|
2012-02-15 08:13:09 +00:00
|
|
|
|
2011-02-14 12:32:11 +00:00
|
|
|
return 1;
|
2009-10-11 20:48:46 +00:00
|
|
|
}
|
|
|
|
|
2011-02-14 12:32:11 +00:00
|
|
|
return 0;
|
2009-10-11 20:48:46 +00:00
|
|
|
}
|
|
|
|
|
2011-02-14 12:32:11 +00:00
|
|
|
#define MD_FALLBACK_FRAME_STATE_FOR sparc64_fallback_frame_state
|
|
|
|
|
2009-10-11 20:48:46 +00:00
|
|
|
#define MD_FROB_UPDATE_CONTEXT sparc64_frob_update_context
|
|
|
|
|
|
|
|
static void
|
|
|
|
sparc64_frob_update_context (struct _Unwind_Context *context,
|
|
|
|
_Unwind_FrameState *fs)
|
|
|
|
{
|
|
|
|
/* The column of %sp contains the old CFA, not the old value of %sp.
|
|
|
|
The CFA offset already comprises the stack bias so, when %sp is the
|
2013-05-22 10:39:03 +00:00
|
|
|
CFA register, we must avoid counting the stack bias twice. */
|
2009-10-11 20:48:46 +00:00
|
|
|
if (fs->regs.cfa_reg == __builtin_dwarf_sp_column ()
|
|
|
|
&& fs->regs.cfa_how == CFA_REG_OFFSET
|
2013-05-22 10:39:03 +00:00
|
|
|
&& fs->regs.cfa_offset != 0)
|
invoke.texi (SPARC options): Add -mflat.
gcc/
* doc/invoke.texi (SPARC options): Add -mflat.
* config/sparc/sparc.opt: Likewise.
* config/sparc/sparc-protos.h (sparc_expand_epilogue): Add parameter.
(sparc_flat_expand_prologue): Declare.
(sparc_flat_expand_epilogue): Likewise.
* config/sparc/sparc.h (CPP_CPU_SPEC): Do not handle -msoft-float.
(CPP_ENDIAN_SPEC): Replace with...
(CPP_OTHER_SPEC): ...this. Also handle -mflat and -msoft-float.
(CPP_SPEC): Adjust to above change.
(EXTRA_SPECS): Likewise.
(SPARC_INCOMING_INT_ARG_FIRST): Add TARGET_FLAT handling.
(INCOMING_REGNO): Likewise.
(OUTGOING_REGNO): Likewise.
(LOCAL_REGNO): Likewise.
(SETUP_FRAME_ADDRESSES): Likewise.
(FIXED_REGISTERS): Set 0 for %fp.
(CALL_USED_REGISTERS): Likewise.
(INITIAL_ELIMINATION_OFFSET): Pass current_function_is_leaf.
(EXIT_IGNORE_STACK): Define to 1 unconditionally.
(RETURN_ADDR_REGNUM): Define.
(RETURN_ADDR_RTX): Use it.
(INCOMING_RETURN_ADDR_REGNUM): Define.
(INCOMING_RETURN_ADDR_RTX): Use it.
(DWARF_FRAME_RETURN_COLUMN): Likewise.
(EH_RETURN_REGNUM): Define.
(EH_RETURN_STACKADJ_RTX): Use it.
(EH_RETURN_HANDLER_RTX): Delete.
(EPILOGUE_USES): Use them and add TARGET_FLAT handling.
* config/sparc/sparc.c (apparent_fsize, actual_fsize, num_gfregs):
Delete.
(struct machine_function): Add frame_size, apparent_frame_size,
frame_base_reg, frame_base_offset, n_global_fp_regs and
save_local_in_regs_p fields.
(sparc_frame_size, sparc_apparent_frame_size, sparc_frame_base_reg,
sparc_frame_base_offset, sparc_n_global_fp_regs,
sparc_save_local_in_regs_p): New macros.
(sparc_option_override): Error out if -fcall-saved-REG is specified
for Out registers.
(eligible_for_restore_insn): Fix formatting.
(eligible_for_return_delay): Likewise. Add TARGET_FLAT handling.
(eligible_for_sibcall_delay): Likewise.
(RTX_OK_FOR_OFFSET_P, RTX_OK_FOR_OLO10_P): Add MODE parameter.
(sparc_legitimate_address_p): Adjust to above change.
(save_global_or_fp_reg_p): New predicate.
(return_addr_reg_needed_p): Likewise.
(save_local_or_in_reg_p): Likewise.
(sparc_compute_frame_size): Use them. Add TARGET_FLAT handling.
(SORR_SAVE, SORR_RESTORE): Delete.
(sorr_pred_t): New typedef.
(sorr_act_t): New enum.
(save_or_restore_regs): Rename to...
(emit_save_or_restore_regs): ...this. Change type of LOW and HIGH
parameters, remove ACTION parameter, add LEAF_FUNCTION_P, SAVE_P,
ACTION_TRUE and ACTION_FALSE parameters. Implement more general
mechanism. Add CFI information for double-word saves in 32-bit mode.
(emit_adjust_base_to_offset): New function extracted from...
(emit_save_or_restore_regs): ...this. Rename the rest to...
(emit_save_or_restore_regs_global_fp_regs): ...this.
(emit_save_or_restore_regs_local_in_regs): New function.
(gen_create_flat_frame_[123]): New functions.
(sparc_expand_prologue): Use SIZE local variable. Adjust.
(sparc_flat_expand_prologue): New function.
(sparc_asm_function_prologue): Add TARGET_FLAT handling.
(sparc_expand_epilogue): Use SIZE local variable. Adjust.
(sparc_flat_expand_epilogue): New function.
(sparc_can_use_return_insn_p): Add TARGET_FLAT handling.
(output_return): Likewise.
(output_sibcall): Likewise.
(sparc_output_mi_thunk): Likewise.
(sparc_frame_pointer_required): Likewise.
(sparc_conditional_register_usage): If TARGET_FLAT, disable the leaf
function optimization.
* config/sparc/sparc.md (flat): New attribute.
(prologue): Add TARGET_FLAT handling.
(save_register_window): Disable if TARGET_FLAT.
(create_flat_frame_[123]): New patterns.
(epilogue): Add TARGET_FLAT handling.
(sibcall_epilogue): Likewise.
(eh_return): New expander.
(eh_return_internal): New insn and splitter.
(return_internal): Add TARGET_FLAT handling.
(untyped_return): Remove bogus test and use RETURN_ADDR_REGNUM.
(save_stack_nonlocal): Use RETURN_ADDR_REGNUM.
(nonlocal_goto): Add TARGET_FLAT handling.
* config/sparc/t-elf: Add -mflat multilib.
* config/sparc/t-leon: Likewise.
libgcc/
* config/sparc/linux-unwind.h (STACK_BIAS): Define.
(sparc64_fallback_frame_state): Use it.
(sparc64_frob_update_context): Further adjust context.
* config/sparc/sol2-unwind.h (sparc64_frob_update_context): Likewise.
* config/sparc/sol2-ci.S: Add TARGET_FLAT handling.
* config/sparc/sol2-cn.S: Likewise.
Co-Authored-By: Laurent Rougé <laurent.rouge@menta.fr>
From-SVN: r174897
2011-06-10 10:50:20 +00:00
|
|
|
{
|
|
|
|
long i;
|
|
|
|
|
|
|
|
context->cfa -= STACK_BIAS;
|
|
|
|
|
Use -fbuilding-libgcc for more target macros used in libgcc.
gcc/c-family:
* c-cppbuiltin.c (c_cpp_builtins): Also define
__LIBGCC_EH_TABLES_CAN_BE_READ_ONLY__,
__LIBGCC_EH_FRAME_SECTION_NAME__, __LIBGCC_JCR_SECTION_NAME__,
__LIBGCC_CTORS_SECTION_ASM_OP__, __LIBGCC_DTORS_SECTION_ASM_OP__,
__LIBGCC_TEXT_SECTION_ASM_OP__, __LIBGCC_INIT_SECTION_ASM_OP__,
__LIBGCC_INIT_ARRAY_SECTION_ASM_OP__,
__LIBGCC_STACK_GROWS_DOWNWARD__,
__LIBGCC_DONT_USE_BUILTIN_SETJMP__,
__LIBGCC_DWARF_ALT_FRAME_RETURN_COLUMN__,
__LIBGCC_DWARF_FRAME_REGISTERS__,
__LIBGCC_EH_RETURN_STACKADJ_RTX__, __LIBGCC_JMP_BUF_SIZE__,
__LIBGCC_STACK_POINTER_REGNUM__ and
__LIBGCC_VTABLE_USES_DESCRIPTORS__ for -fbuilding-libgcc.
(builtin_define_with_value): Handle backslash-escaping in string
macro values.
libgcc:
* Makefile.in (CRTSTUFF_CFLAGS): Add -fbuilding-libgcc.
* config/aarch64/linux-unwind.h (STACK_POINTER_REGNUM): Change all
uses to __LIBGCC_STACK_POINTER_REGNUM__.
(DWARF_ALT_FRAME_RETURN_COLUMN): Change all uses to
__LIBGCC_DWARF_ALT_FRAME_RETURN_COLUMN__.
* config/alpha/vms-unwind.h (DWARF_ALT_FRAME_RETURN_COLUMN):
Change use to __LIBGCC_DWARF_ALT_FRAME_RETURN_COLUMN__.
* config/cr16/unwind-cr16.c (STACK_GROWS_DOWNWARD): Change all
uses to __LIBGCC_STACK_GROWS_DOWNWARD__.
(DWARF_FRAME_REGISTERS): Change all uses to
__LIBGCC_DWARF_FRAME_REGISTERS__.
(EH_RETURN_STACKADJ_RTX): Change all uses to
__LIBGCC_EH_RETURN_STACKADJ_RTX__.
* config/cr16/unwind-dw2.h (DWARF_FRAME_REGISTERS): Change use to
__LIBGCC_DWARF_FRAME_REGISTERS__. Remove conditional definition.
* config/i386/cygming-crtbegin.c (EH_FRAME_SECTION_NAME): Change
use to __LIBGCC_EH_FRAME_SECTION_NAME__.
(JCR_SECTION_NAME): Change use to __LIBGCC_JCR_SECTION_NAME__.
* config/i386/cygming-crtend.c (EH_FRAME_SECTION_NAME): Change use
to __LIBGCC_EH_FRAME_SECTION_NAME__.
(JCR_SECTION_NAME): Change use to __LIBGCC_JCR_SECTION_NAME__
* config/mips/linux-unwind.h (STACK_POINTER_REGNUM): Change use to
__LIBGCC_STACK_POINTER_REGNUM__.
(DWARF_ALT_FRAME_RETURN_COLUMN): Change all uses to
__LIBGCC_DWARF_ALT_FRAME_RETURN_COLUMN__.
* config/nios2/linux-unwind.h (STACK_POINTER_REGNUM): Change use
to __LIBGCC_STACK_POINTER_REGNUM__.
* config/pa/hpux-unwind.h (DWARF_ALT_FRAME_RETURN_COLUMN): Change
all uses to __LIBGCC_DWARF_ALT_FRAME_RETURN_COLUMN__.
* config/pa/linux-unwind.h (DWARF_ALT_FRAME_RETURN_COLUMN): Change
all uses to __LIBGCC_DWARF_ALT_FRAME_RETURN_COLUMN__.
* config/rs6000/aix-unwind.h (DWARF_ALT_FRAME_RETURN_COLUMN):
Change all uses to __LIBGCC_DWARF_ALT_FRAME_RETURN_COLUMN__.
(STACK_POINTER_REGNUM): Change all uses to
__LIBGCC_STACK_POINTER_REGNUM__.
* config/rs6000/darwin-fallback.c (STACK_POINTER_REGNUM): Change
use to __LIBGCC_STACK_POINTER_REGNUM__.
* config/rs6000/linux-unwind.h (STACK_POINTER_REGNUM): Change all
uses to __LIBGCC_STACK_POINTER_REGNUM__.
* config/sparc/linux-unwind.h (DWARF_FRAME_REGISTERS): Change use
to __LIBGCC_DWARF_FRAME_REGISTERS__.
* config/sparc/sol2-unwind.h (DWARF_FRAME_REGISTERS): Change use
to __LIBGCC_DWARF_FRAME_REGISTERS__.
* config/tilepro/linux-unwind.h (STACK_POINTER_REGNUM): Change use
to __LIBGCC_STACK_POINTER_REGNUM__.
* config/xtensa/unwind-dw2-xtensa.h (DWARF_FRAME_REGISTERS):
Remove conditional definition.
* crtstuff.c (TEXT_SECTION_ASM_OP): Change all uses to
__LIBGCC_TEXT_SECTION_ASM_OP__.
(EH_FRAME_SECTION_NAME): Change all uses to
__LIBGCC_EH_FRAME_SECTION_NAME__.
(EH_TABLES_CAN_BE_READ_ONLY): Change all uses to
__LIBGCC_EH_TABLES_CAN_BE_READ_ONLY__.
(CTORS_SECTION_ASM_OP): Change all uses to
__LIBGCC_CTORS_SECTION_ASM_OP__.
(DTORS_SECTION_ASM_OP): Change all uses to
__LIBGCC_DTORS_SECTION_ASM_OP__.
(JCR_SECTION_NAME): Change all uses to
__LIBGCC_JCR_SECTION_NAME__.
(INIT_SECTION_ASM_OP): Change all uses to
__LIBGCC_INIT_SECTION_ASM_OP__.
(INIT_ARRAY_SECTION_ASM_OP): Change all uses to
__LIBGCC_INIT_ARRAY_SECTION_ASM_OP__.
* generic-morestack.c (STACK_GROWS_DOWNWARD): Change all uses to
__LIBGCC_STACK_GROWS_DOWNWARD__.
* libgcc2.c (INIT_SECTION_ASM_OP): Change all uses to
__LIBGCC_INIT_SECTION_ASM_OP__.
(INIT_ARRAY_SECTION_ASM_OP): Change all uses to
__LIBGCC_INIT_ARRAY_SECTION_ASM_OP__.
(EH_FRAME_SECTION_NAME): Change all uses to
__LIBGCC_EH_FRAME_SECTION_NAME__.
* libgcov-profiler.c (VTABLE_USES_DESCRIPTORS): Remove conditional
definitions. Change all uses to
__LIBGCC_VTABLE_USES_DESCRIPTORS__.
* unwind-dw2.c (STACK_GROWS_DOWNWARD): Change all uses to
__LIBGCC_STACK_GROWS_DOWNWARD__.
(DWARF_FRAME_REGISTERS): Change all uses to
__LIBGCC_DWARF_FRAME_REGISTERS__.
(EH_RETURN_STACKADJ_RTX): Change all uses to
__LIBGCC_EH_RETURN_STACKADJ_RTX__.
* unwind-dw2.h (DWARF_FRAME_REGISTERS): Remove conditional
definition. Change use to __LIBGCC_DWARF_FRAME_REGISTERS__.
* unwind-sjlj.c (DONT_USE_BUILTIN_SETJMP): Change all uses to
__LIBGCC_DONT_USE_BUILTIN_SETJMP__.
(JMP_BUF_SIZE): Change use to __LIBGCC_JMP_BUF_SIZE__.
From-SVN: r214954
2014-09-05 13:03:46 +01:00
|
|
|
for (i = 0; i < __LIBGCC_DWARF_FRAME_REGISTERS__ + 1; ++i)
|
libgcc: Decrease size of _Unwind_FrameState and even more size of cleared area in uw_frame_state_for
The following patch implements something that has Florian found as
low hanging fruit in our unwinder and has been discussed in the
https://gcc.gnu.org/wiki/cauldron2022#cauldron2022talks.inprocess_unwinding_bof
talk.
_Unwind_FrameState type seems to be (unlike the pre-GCC 3 frame_state
which has been part of ABI) private to unwind-dw2.c + unwind.inc it
includes, it is always defined on the stack of some entrypoints, initialized
by static uw_frame_state_for and the address of it is also passed to other
static functions or the static inlines handling machine dependent unwinding,
but it isn't fortunately passed to any callbacks or public functions, so I
think we can safely change it any time we want.
Florian mentioned that the structure is large even on x86_64, 384 bytes
there, starts with 328 bytes long element with frame_state_reg_info type
which then starts with an array with __LIBGCC_DWARF_FRAME_REGISTERS__ + 1
elements, each of them is 16 bytes long, on x86_64
__LIBGCC_DWARF_FRAME_REGISTERS__ is just 17 but even that is big, on say
riscv __LIBGCC_DWARF_FRAME_REGISTERS__ is I think 128, on powerpc 111,
on sh 153 etc. And, we memset to zero the whole fs variable with the
_Unwind_FrameState type at the start of the unwinding.
The reason why each element is 16 byte (on 64-bit arches) is that it
contains some pointer or pointer sized integer and then an enum (with just
7 different enumerators) + padding.
The following patch decreases it by moving the enum into a separate
array and using just one byte for each register in that second array.
We could compress it even more, say 4 bits per register, but I don't
want to uglify the code for it too much and make the accesses slower.
Furthermore, the clearing of the object can clear only thos how array
and members after it, because REG_UNSAVED enumerator (0) doesn't actually
need any pointer or pointer sized integer, it is just the other kinds
that need to have there something.
By doing this, on x86_64 the above numbers change to _Unwind_FrameState
type being now 264 bytes long, frame_state_reg_info 208 bytes and we
don't clear the first 144 bytes of the object, so the memset is 120 bytes,
so ~ 31% of the old clearing size. On riscv 64-bit assuming it has same
structure layout rules for the few types used there that would be
~ 2160 bytes of _Unwind_FrameState type before and ~ 1264 bytes after,
with the memset previously ~ 2160 bytes and after ~ 232 bytes after.
We've also talked about possibly adding a number of initially initialized
regs and initializing the rest lazily, but at least for x86_64 with
18 elements in the array that doesn't seem to be worth it anymore,
especially because return address column is 16 there and that is usually the
first thing to be touched. It might theory help with lots of registers if
they are usually untouched, but would uglify and complicate any stores to
how by having to check there for the not initialized yet cases and lazy
initialization, and similarly for all reads of how to do there if below
last initialized one, use how, otherwise imply REG_UNSAVED.
The disadvantage of the patch is that touching reg[x].loc and how[x]
now means 2 cachelines rather than one as before, and I admit beyond
bootstrap/regtest I haven't benchmarked it in any way.
2022-10-06 Jakub Jelinek <jakub@redhat.com>
* unwind-dw2.h (REG_UNSAVED, REG_SAVED_OFFSET, REG_SAVED_REG,
REG_SAVED_EXP, REG_SAVED_VAL_OFFSET, REG_SAVED_VAL_EXP,
REG_UNDEFINED): New anonymous enum, moved from inside of
struct frame_state_reg_info.
(struct frame_state_reg_info): Remove reg[].how element and the
anonymous enum there. Add how element.
* unwind-dw2.c: Include stddef.h.
(uw_frame_state_for): Don't clear first
offsetof (_Unwind_FrameState, regs.how[0]) bytes of *fs.
(execute_cfa_program, __frame_state_for, uw_update_context_1,
uw_update_context): Use fs->regs.how[X] instead of fs->regs.reg[X].how
or fs.regs.how[X] instead of fs.regs.reg[X].how.
* config/sh/linux-unwind.h (sh_fallback_frame_state): Likewise.
* config/bfin/linux-unwind.h (bfin_fallback_frame_state): Likewise.
* config/pa/linux-unwind.h (pa32_fallback_frame_state): Likewise.
* config/pa/hpux-unwind.h (UPDATE_FS_FOR_SAR, UPDATE_FS_FOR_GR,
UPDATE_FS_FOR_FR, UPDATE_FS_FOR_PC, pa_fallback_frame_state):
Likewise.
* config/alpha/vms-unwind.h (alpha_vms_fallback_frame_state):
Likewise.
* config/alpha/linux-unwind.h (alpha_fallback_frame_state): Likewise.
* config/arc/linux-unwind.h (arc_fallback_frame_state,
arc_frob_update_context): Likewise.
* config/riscv/linux-unwind.h (riscv_fallback_frame_state): Likewise.
* config/nios2/linux-unwind.h (NIOS2_REG): Likewise.
* config/nds32/linux-unwind.h (NDS32_PUT_FS_REG): Likewise.
* config/s390/tpf-unwind.h (s390_fallback_frame_state): Likewise.
* config/s390/linux-unwind.h (s390_fallback_frame_state): Likewise.
* config/sparc/sol2-unwind.h (sparc64_frob_update_context,
MD_FALLBACK_FRAME_STATE_FOR): Likewise.
* config/sparc/linux-unwind.h (sparc64_fallback_frame_state,
sparc64_frob_update_context, sparc_fallback_frame_state): Likewise.
* config/i386/sol2-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/w32-unwind.h (i386_w32_fallback_frame_state): Likewise.
* config/i386/linux-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/freebsd-unwind.h (x86_64_freebsd_fallback_frame_state):
Likewise.
* config/i386/dragonfly-unwind.h
(x86_64_dragonfly_fallback_frame_state): Likewise.
* config/i386/gnu-unwind.h (x86_gnu_fallback_frame_state): Likewise.
* config/csky/linux-unwind.h (csky_fallback_frame_state): Likewise.
* config/aarch64/linux-unwind.h (aarch64_fallback_frame_state):
Likewise.
* config/aarch64/freebsd-unwind.h
(aarch64_freebsd_fallback_frame_state): Likewise.
* config/aarch64/aarch64-unwind.h (aarch64_frob_update_context):
Likewise.
* config/or1k/linux-unwind.h (or1k_fallback_frame_state): Likewise.
* config/mips/linux-unwind.h (mips_fallback_frame_state): Likewise.
* config/loongarch/linux-unwind.h (loongarch_fallback_frame_state):
Likewise.
* config/m68k/linux-unwind.h (m68k_fallback_frame_state): Likewise.
* config/xtensa/linux-unwind.h (xtensa_fallback_frame_state):
Likewise.
* config/rs6000/darwin-fallback.c (set_offset): Likewise.
* config/rs6000/aix-unwind.h (MD_FROB_UPDATE_CONTEXT): Likewise.
* config/rs6000/linux-unwind.h (ppc_fallback_frame_state): Likewise.
* config/rs6000/freebsd-unwind.h (frob_update_context): Likewise.
2022-10-06 11:04:52 +02:00
|
|
|
if (fs->regs.how[i] == REG_SAVED_OFFSET)
|
invoke.texi (SPARC options): Add -mflat.
gcc/
* doc/invoke.texi (SPARC options): Add -mflat.
* config/sparc/sparc.opt: Likewise.
* config/sparc/sparc-protos.h (sparc_expand_epilogue): Add parameter.
(sparc_flat_expand_prologue): Declare.
(sparc_flat_expand_epilogue): Likewise.
* config/sparc/sparc.h (CPP_CPU_SPEC): Do not handle -msoft-float.
(CPP_ENDIAN_SPEC): Replace with...
(CPP_OTHER_SPEC): ...this. Also handle -mflat and -msoft-float.
(CPP_SPEC): Adjust to above change.
(EXTRA_SPECS): Likewise.
(SPARC_INCOMING_INT_ARG_FIRST): Add TARGET_FLAT handling.
(INCOMING_REGNO): Likewise.
(OUTGOING_REGNO): Likewise.
(LOCAL_REGNO): Likewise.
(SETUP_FRAME_ADDRESSES): Likewise.
(FIXED_REGISTERS): Set 0 for %fp.
(CALL_USED_REGISTERS): Likewise.
(INITIAL_ELIMINATION_OFFSET): Pass current_function_is_leaf.
(EXIT_IGNORE_STACK): Define to 1 unconditionally.
(RETURN_ADDR_REGNUM): Define.
(RETURN_ADDR_RTX): Use it.
(INCOMING_RETURN_ADDR_REGNUM): Define.
(INCOMING_RETURN_ADDR_RTX): Use it.
(DWARF_FRAME_RETURN_COLUMN): Likewise.
(EH_RETURN_REGNUM): Define.
(EH_RETURN_STACKADJ_RTX): Use it.
(EH_RETURN_HANDLER_RTX): Delete.
(EPILOGUE_USES): Use them and add TARGET_FLAT handling.
* config/sparc/sparc.c (apparent_fsize, actual_fsize, num_gfregs):
Delete.
(struct machine_function): Add frame_size, apparent_frame_size,
frame_base_reg, frame_base_offset, n_global_fp_regs and
save_local_in_regs_p fields.
(sparc_frame_size, sparc_apparent_frame_size, sparc_frame_base_reg,
sparc_frame_base_offset, sparc_n_global_fp_regs,
sparc_save_local_in_regs_p): New macros.
(sparc_option_override): Error out if -fcall-saved-REG is specified
for Out registers.
(eligible_for_restore_insn): Fix formatting.
(eligible_for_return_delay): Likewise. Add TARGET_FLAT handling.
(eligible_for_sibcall_delay): Likewise.
(RTX_OK_FOR_OFFSET_P, RTX_OK_FOR_OLO10_P): Add MODE parameter.
(sparc_legitimate_address_p): Adjust to above change.
(save_global_or_fp_reg_p): New predicate.
(return_addr_reg_needed_p): Likewise.
(save_local_or_in_reg_p): Likewise.
(sparc_compute_frame_size): Use them. Add TARGET_FLAT handling.
(SORR_SAVE, SORR_RESTORE): Delete.
(sorr_pred_t): New typedef.
(sorr_act_t): New enum.
(save_or_restore_regs): Rename to...
(emit_save_or_restore_regs): ...this. Change type of LOW and HIGH
parameters, remove ACTION parameter, add LEAF_FUNCTION_P, SAVE_P,
ACTION_TRUE and ACTION_FALSE parameters. Implement more general
mechanism. Add CFI information for double-word saves in 32-bit mode.
(emit_adjust_base_to_offset): New function extracted from...
(emit_save_or_restore_regs): ...this. Rename the rest to...
(emit_save_or_restore_regs_global_fp_regs): ...this.
(emit_save_or_restore_regs_local_in_regs): New function.
(gen_create_flat_frame_[123]): New functions.
(sparc_expand_prologue): Use SIZE local variable. Adjust.
(sparc_flat_expand_prologue): New function.
(sparc_asm_function_prologue): Add TARGET_FLAT handling.
(sparc_expand_epilogue): Use SIZE local variable. Adjust.
(sparc_flat_expand_epilogue): New function.
(sparc_can_use_return_insn_p): Add TARGET_FLAT handling.
(output_return): Likewise.
(output_sibcall): Likewise.
(sparc_output_mi_thunk): Likewise.
(sparc_frame_pointer_required): Likewise.
(sparc_conditional_register_usage): If TARGET_FLAT, disable the leaf
function optimization.
* config/sparc/sparc.md (flat): New attribute.
(prologue): Add TARGET_FLAT handling.
(save_register_window): Disable if TARGET_FLAT.
(create_flat_frame_[123]): New patterns.
(epilogue): Add TARGET_FLAT handling.
(sibcall_epilogue): Likewise.
(eh_return): New expander.
(eh_return_internal): New insn and splitter.
(return_internal): Add TARGET_FLAT handling.
(untyped_return): Remove bogus test and use RETURN_ADDR_REGNUM.
(save_stack_nonlocal): Use RETURN_ADDR_REGNUM.
(nonlocal_goto): Add TARGET_FLAT handling.
* config/sparc/t-elf: Add -mflat multilib.
* config/sparc/t-leon: Likewise.
libgcc/
* config/sparc/linux-unwind.h (STACK_BIAS): Define.
(sparc64_fallback_frame_state): Use it.
(sparc64_frob_update_context): Further adjust context.
* config/sparc/sol2-unwind.h (sparc64_frob_update_context): Likewise.
* config/sparc/sol2-ci.S: Add TARGET_FLAT handling.
* config/sparc/sol2-cn.S: Likewise.
Co-Authored-By: Laurent Rougé <laurent.rouge@menta.fr>
From-SVN: r174897
2011-06-10 10:50:20 +00:00
|
|
|
_Unwind_SetGRPtr (context, i,
|
|
|
|
_Unwind_GetGRPtr (context, i) - STACK_BIAS);
|
|
|
|
}
|
2009-10-11 20:48:46 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
#else
|
|
|
|
|
2011-02-14 12:32:11 +00:00
|
|
|
#define IS_SIGHANDLER sparc_is_sighandler
|
|
|
|
|
|
|
|
static int
|
2012-01-09 09:50:19 +00:00
|
|
|
sparc_is_sighandler (unsigned int *pc, void *cfa, int *nframes)
|
2011-02-14 12:32:11 +00:00
|
|
|
{
|
|
|
|
if(/* Solaris 8+ - multi-threaded
|
|
|
|
----------------------------
|
|
|
|
<__sighndlr>: save %sp, -96, %sp
|
|
|
|
<__sighndlr+4>: mov %i0, %o0
|
|
|
|
<__sighndlr+8>: mov %i1, %o1
|
|
|
|
<__sighndlr+12>: call %i3
|
|
|
|
<__sighndlr+16>: mov %i2, %o2
|
|
|
|
<__sighndlr+20>: ret <--- PC
|
|
|
|
<__sighndlr+24>: restore */
|
|
|
|
pc[-5] == 0x9de3bfa0
|
|
|
|
&& pc[-4] == 0x90100018
|
|
|
|
&& pc[-3] == 0x92100019
|
|
|
|
&& pc[-2] == 0x9fc6c000
|
|
|
|
&& pc[-1] == 0x9410001a
|
|
|
|
&& pc[ 0] == 0x81c7e008
|
|
|
|
&& pc[ 1] == 0x81e80000)
|
|
|
|
{
|
2012-01-09 09:50:19 +00:00
|
|
|
/* We have observed different calling frames among different
|
|
|
|
versions of the operating system, so that we need to
|
|
|
|
discriminate using the upper frame. We look for the return
|
|
|
|
address of the caller frame (there is an offset of 15 words
|
|
|
|
between the frame address and the place where this return
|
|
|
|
address is stored) in order to do some more pattern matching. */
|
|
|
|
unsigned int cuh_pattern
|
|
|
|
= *(unsigned int *)(*(unsigned int *)(cfa + 15*4) - 4);
|
|
|
|
|
2012-02-15 08:13:09 +00:00
|
|
|
if (cuh_pattern == 0x92100019)
|
2014-04-22 12:30:59 +00:00
|
|
|
/* This matches the call_user_handler pattern in Solaris 11
|
|
|
|
libc.so.1:
|
|
|
|
|
|
|
|
<call_user_handler+876>: mov %i1, %o1
|
2019-05-28 17:27:51 +00:00
|
|
|
<call_user_handler+880>: call __sighndlr */
|
2012-01-09 09:50:19 +00:00
|
|
|
*nframes = 3;
|
2012-02-15 08:13:09 +00:00
|
|
|
|
2011-02-14 12:32:11 +00:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-10-11 20:48:46 +00:00
|
|
|
#define MD_FALLBACK_FRAME_STATE_FOR sparc_fallback_frame_state
|
|
|
|
|
2011-02-14 12:32:11 +00:00
|
|
|
#endif
|
|
|
|
|
2009-10-11 20:48:46 +00:00
|
|
|
static _Unwind_Reason_Code
|
2011-02-14 12:32:11 +00:00
|
|
|
MD_FALLBACK_FRAME_STATE_FOR (struct _Unwind_Context *context,
|
|
|
|
_Unwind_FrameState *fs)
|
2009-10-11 20:48:46 +00:00
|
|
|
{
|
|
|
|
void *pc = context->ra;
|
|
|
|
void *this_cfa = context->cfa;
|
2013-05-22 10:39:03 +00:00
|
|
|
int nframes = 0;
|
2011-02-14 12:32:11 +00:00
|
|
|
long new_cfa;
|
|
|
|
void *ra_location, *shifted_ra_location;
|
|
|
|
mcontext_t *mctx;
|
2009-10-11 20:48:46 +00:00
|
|
|
int i;
|
|
|
|
|
|
|
|
/* Deal with frame-less function from which a signal was raised. */
|
|
|
|
if (_Unwind_IsSignalFrame (context))
|
|
|
|
{
|
|
|
|
/* The CFA is by definition unmodified in this case. */
|
|
|
|
fs->regs.cfa_how = CFA_REG_OFFSET;
|
|
|
|
fs->regs.cfa_reg = __builtin_dwarf_sp_column ();
|
|
|
|
fs->regs.cfa_offset = 0;
|
|
|
|
|
|
|
|
/* This is the canonical RA column. */
|
|
|
|
fs->retaddr_column = 15;
|
|
|
|
|
|
|
|
return _URC_NO_REASON;
|
|
|
|
}
|
|
|
|
|
2013-05-22 10:39:03 +00:00
|
|
|
/* Do some pattern matching at the return address. */
|
2012-01-09 09:50:19 +00:00
|
|
|
if (IS_SIGHANDLER (pc, this_cfa, &nframes))
|
2009-10-11 20:48:46 +00:00
|
|
|
{
|
2013-05-22 10:39:03 +00:00
|
|
|
struct frame *fp = (struct frame *) this_cfa;
|
2011-02-14 12:32:11 +00:00
|
|
|
struct handler_args {
|
|
|
|
struct frame frwin;
|
|
|
|
ucontext_t ucontext;
|
|
|
|
} *handler_args;
|
|
|
|
ucontext_t *ucp;
|
|
|
|
|
2013-05-22 10:39:03 +00:00
|
|
|
/* this_cfa points into the frame after the saved frame pointer and
|
2011-02-14 12:32:11 +00:00
|
|
|
saved pc (struct frame).
|
|
|
|
|
|
|
|
The ucontext_t structure is in the kernel frame after a struct
|
|
|
|
frame. Since the frame sizes vary even within OS releases, we
|
|
|
|
need to walk the stack to get there. */
|
|
|
|
for (i = 0; i < nframes; i++)
|
|
|
|
fp = (struct frame *) ((char *)fp->fr_savfp + STACK_BIAS);
|
|
|
|
|
|
|
|
handler_args = (struct handler_args *) fp;
|
|
|
|
ucp = &handler_args->ucontext;
|
|
|
|
mctx = &ucp->uc_mcontext;
|
2009-10-11 20:48:46 +00:00
|
|
|
}
|
|
|
|
else
|
|
|
|
return _URC_END_OF_STACK;
|
|
|
|
|
2011-02-14 12:32:11 +00:00
|
|
|
/* The frame address is %sp + STACK_BIAS in 64-bit mode. */
|
2013-05-22 10:39:03 +00:00
|
|
|
new_cfa = mctx->gregs[REG_SP] + STACK_BIAS;
|
2011-02-14 12:32:11 +00:00
|
|
|
|
2009-10-11 20:48:46 +00:00
|
|
|
fs->regs.cfa_how = CFA_REG_OFFSET;
|
|
|
|
fs->regs.cfa_reg = __builtin_dwarf_sp_column ();
|
2013-05-22 10:39:03 +00:00
|
|
|
fs->regs.cfa_offset = new_cfa - (long) this_cfa + STACK_BIAS;
|
2009-10-11 20:48:46 +00:00
|
|
|
|
|
|
|
/* Restore global and out registers (in this order) from the
|
|
|
|
ucontext_t structure, uc_mcontext.gregs field. */
|
|
|
|
for (i = 1; i < 16; i++)
|
|
|
|
{
|
|
|
|
/* We never restore %sp as everything is purely CFA-based. */
|
|
|
|
if ((unsigned int) i == __builtin_dwarf_sp_column ())
|
|
|
|
continue;
|
|
|
|
|
2011-02-14 12:32:11 +00:00
|
|
|
/* First the global registers and then the out registers. */
|
libgcc: Decrease size of _Unwind_FrameState and even more size of cleared area in uw_frame_state_for
The following patch implements something that has Florian found as
low hanging fruit in our unwinder and has been discussed in the
https://gcc.gnu.org/wiki/cauldron2022#cauldron2022talks.inprocess_unwinding_bof
talk.
_Unwind_FrameState type seems to be (unlike the pre-GCC 3 frame_state
which has been part of ABI) private to unwind-dw2.c + unwind.inc it
includes, it is always defined on the stack of some entrypoints, initialized
by static uw_frame_state_for and the address of it is also passed to other
static functions or the static inlines handling machine dependent unwinding,
but it isn't fortunately passed to any callbacks or public functions, so I
think we can safely change it any time we want.
Florian mentioned that the structure is large even on x86_64, 384 bytes
there, starts with 328 bytes long element with frame_state_reg_info type
which then starts with an array with __LIBGCC_DWARF_FRAME_REGISTERS__ + 1
elements, each of them is 16 bytes long, on x86_64
__LIBGCC_DWARF_FRAME_REGISTERS__ is just 17 but even that is big, on say
riscv __LIBGCC_DWARF_FRAME_REGISTERS__ is I think 128, on powerpc 111,
on sh 153 etc. And, we memset to zero the whole fs variable with the
_Unwind_FrameState type at the start of the unwinding.
The reason why each element is 16 byte (on 64-bit arches) is that it
contains some pointer or pointer sized integer and then an enum (with just
7 different enumerators) + padding.
The following patch decreases it by moving the enum into a separate
array and using just one byte for each register in that second array.
We could compress it even more, say 4 bits per register, but I don't
want to uglify the code for it too much and make the accesses slower.
Furthermore, the clearing of the object can clear only thos how array
and members after it, because REG_UNSAVED enumerator (0) doesn't actually
need any pointer or pointer sized integer, it is just the other kinds
that need to have there something.
By doing this, on x86_64 the above numbers change to _Unwind_FrameState
type being now 264 bytes long, frame_state_reg_info 208 bytes and we
don't clear the first 144 bytes of the object, so the memset is 120 bytes,
so ~ 31% of the old clearing size. On riscv 64-bit assuming it has same
structure layout rules for the few types used there that would be
~ 2160 bytes of _Unwind_FrameState type before and ~ 1264 bytes after,
with the memset previously ~ 2160 bytes and after ~ 232 bytes after.
We've also talked about possibly adding a number of initially initialized
regs and initializing the rest lazily, but at least for x86_64 with
18 elements in the array that doesn't seem to be worth it anymore,
especially because return address column is 16 there and that is usually the
first thing to be touched. It might theory help with lots of registers if
they are usually untouched, but would uglify and complicate any stores to
how by having to check there for the not initialized yet cases and lazy
initialization, and similarly for all reads of how to do there if below
last initialized one, use how, otherwise imply REG_UNSAVED.
The disadvantage of the patch is that touching reg[x].loc and how[x]
now means 2 cachelines rather than one as before, and I admit beyond
bootstrap/regtest I haven't benchmarked it in any way.
2022-10-06 Jakub Jelinek <jakub@redhat.com>
* unwind-dw2.h (REG_UNSAVED, REG_SAVED_OFFSET, REG_SAVED_REG,
REG_SAVED_EXP, REG_SAVED_VAL_OFFSET, REG_SAVED_VAL_EXP,
REG_UNDEFINED): New anonymous enum, moved from inside of
struct frame_state_reg_info.
(struct frame_state_reg_info): Remove reg[].how element and the
anonymous enum there. Add how element.
* unwind-dw2.c: Include stddef.h.
(uw_frame_state_for): Don't clear first
offsetof (_Unwind_FrameState, regs.how[0]) bytes of *fs.
(execute_cfa_program, __frame_state_for, uw_update_context_1,
uw_update_context): Use fs->regs.how[X] instead of fs->regs.reg[X].how
or fs.regs.how[X] instead of fs.regs.reg[X].how.
* config/sh/linux-unwind.h (sh_fallback_frame_state): Likewise.
* config/bfin/linux-unwind.h (bfin_fallback_frame_state): Likewise.
* config/pa/linux-unwind.h (pa32_fallback_frame_state): Likewise.
* config/pa/hpux-unwind.h (UPDATE_FS_FOR_SAR, UPDATE_FS_FOR_GR,
UPDATE_FS_FOR_FR, UPDATE_FS_FOR_PC, pa_fallback_frame_state):
Likewise.
* config/alpha/vms-unwind.h (alpha_vms_fallback_frame_state):
Likewise.
* config/alpha/linux-unwind.h (alpha_fallback_frame_state): Likewise.
* config/arc/linux-unwind.h (arc_fallback_frame_state,
arc_frob_update_context): Likewise.
* config/riscv/linux-unwind.h (riscv_fallback_frame_state): Likewise.
* config/nios2/linux-unwind.h (NIOS2_REG): Likewise.
* config/nds32/linux-unwind.h (NDS32_PUT_FS_REG): Likewise.
* config/s390/tpf-unwind.h (s390_fallback_frame_state): Likewise.
* config/s390/linux-unwind.h (s390_fallback_frame_state): Likewise.
* config/sparc/sol2-unwind.h (sparc64_frob_update_context,
MD_FALLBACK_FRAME_STATE_FOR): Likewise.
* config/sparc/linux-unwind.h (sparc64_fallback_frame_state,
sparc64_frob_update_context, sparc_fallback_frame_state): Likewise.
* config/i386/sol2-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/w32-unwind.h (i386_w32_fallback_frame_state): Likewise.
* config/i386/linux-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/freebsd-unwind.h (x86_64_freebsd_fallback_frame_state):
Likewise.
* config/i386/dragonfly-unwind.h
(x86_64_dragonfly_fallback_frame_state): Likewise.
* config/i386/gnu-unwind.h (x86_gnu_fallback_frame_state): Likewise.
* config/csky/linux-unwind.h (csky_fallback_frame_state): Likewise.
* config/aarch64/linux-unwind.h (aarch64_fallback_frame_state):
Likewise.
* config/aarch64/freebsd-unwind.h
(aarch64_freebsd_fallback_frame_state): Likewise.
* config/aarch64/aarch64-unwind.h (aarch64_frob_update_context):
Likewise.
* config/or1k/linux-unwind.h (or1k_fallback_frame_state): Likewise.
* config/mips/linux-unwind.h (mips_fallback_frame_state): Likewise.
* config/loongarch/linux-unwind.h (loongarch_fallback_frame_state):
Likewise.
* config/m68k/linux-unwind.h (m68k_fallback_frame_state): Likewise.
* config/xtensa/linux-unwind.h (xtensa_fallback_frame_state):
Likewise.
* config/rs6000/darwin-fallback.c (set_offset): Likewise.
* config/rs6000/aix-unwind.h (MD_FROB_UPDATE_CONTEXT): Likewise.
* config/rs6000/linux-unwind.h (ppc_fallback_frame_state): Likewise.
* config/rs6000/freebsd-unwind.h (frob_update_context): Likewise.
2022-10-06 11:04:52 +02:00
|
|
|
fs->regs.how[i] = REG_SAVED_OFFSET;
|
2011-02-14 12:32:11 +00:00
|
|
|
fs->regs.reg[i].loc.offset = (long)&mctx->gregs[REG_Y + i] - new_cfa;
|
2009-10-11 20:48:46 +00:00
|
|
|
}
|
|
|
|
|
2011-02-14 12:32:11 +00:00
|
|
|
/* Just above the stack pointer there are 16 extended words in which
|
|
|
|
the register window (in and local registers) was saved. */
|
2009-10-11 20:48:46 +00:00
|
|
|
for (i = 0; i < 16; i++)
|
|
|
|
{
|
libgcc: Decrease size of _Unwind_FrameState and even more size of cleared area in uw_frame_state_for
The following patch implements something that has Florian found as
low hanging fruit in our unwinder and has been discussed in the
https://gcc.gnu.org/wiki/cauldron2022#cauldron2022talks.inprocess_unwinding_bof
talk.
_Unwind_FrameState type seems to be (unlike the pre-GCC 3 frame_state
which has been part of ABI) private to unwind-dw2.c + unwind.inc it
includes, it is always defined on the stack of some entrypoints, initialized
by static uw_frame_state_for and the address of it is also passed to other
static functions or the static inlines handling machine dependent unwinding,
but it isn't fortunately passed to any callbacks or public functions, so I
think we can safely change it any time we want.
Florian mentioned that the structure is large even on x86_64, 384 bytes
there, starts with 328 bytes long element with frame_state_reg_info type
which then starts with an array with __LIBGCC_DWARF_FRAME_REGISTERS__ + 1
elements, each of them is 16 bytes long, on x86_64
__LIBGCC_DWARF_FRAME_REGISTERS__ is just 17 but even that is big, on say
riscv __LIBGCC_DWARF_FRAME_REGISTERS__ is I think 128, on powerpc 111,
on sh 153 etc. And, we memset to zero the whole fs variable with the
_Unwind_FrameState type at the start of the unwinding.
The reason why each element is 16 byte (on 64-bit arches) is that it
contains some pointer or pointer sized integer and then an enum (with just
7 different enumerators) + padding.
The following patch decreases it by moving the enum into a separate
array and using just one byte for each register in that second array.
We could compress it even more, say 4 bits per register, but I don't
want to uglify the code for it too much and make the accesses slower.
Furthermore, the clearing of the object can clear only thos how array
and members after it, because REG_UNSAVED enumerator (0) doesn't actually
need any pointer or pointer sized integer, it is just the other kinds
that need to have there something.
By doing this, on x86_64 the above numbers change to _Unwind_FrameState
type being now 264 bytes long, frame_state_reg_info 208 bytes and we
don't clear the first 144 bytes of the object, so the memset is 120 bytes,
so ~ 31% of the old clearing size. On riscv 64-bit assuming it has same
structure layout rules for the few types used there that would be
~ 2160 bytes of _Unwind_FrameState type before and ~ 1264 bytes after,
with the memset previously ~ 2160 bytes and after ~ 232 bytes after.
We've also talked about possibly adding a number of initially initialized
regs and initializing the rest lazily, but at least for x86_64 with
18 elements in the array that doesn't seem to be worth it anymore,
especially because return address column is 16 there and that is usually the
first thing to be touched. It might theory help with lots of registers if
they are usually untouched, but would uglify and complicate any stores to
how by having to check there for the not initialized yet cases and lazy
initialization, and similarly for all reads of how to do there if below
last initialized one, use how, otherwise imply REG_UNSAVED.
The disadvantage of the patch is that touching reg[x].loc and how[x]
now means 2 cachelines rather than one as before, and I admit beyond
bootstrap/regtest I haven't benchmarked it in any way.
2022-10-06 Jakub Jelinek <jakub@redhat.com>
* unwind-dw2.h (REG_UNSAVED, REG_SAVED_OFFSET, REG_SAVED_REG,
REG_SAVED_EXP, REG_SAVED_VAL_OFFSET, REG_SAVED_VAL_EXP,
REG_UNDEFINED): New anonymous enum, moved from inside of
struct frame_state_reg_info.
(struct frame_state_reg_info): Remove reg[].how element and the
anonymous enum there. Add how element.
* unwind-dw2.c: Include stddef.h.
(uw_frame_state_for): Don't clear first
offsetof (_Unwind_FrameState, regs.how[0]) bytes of *fs.
(execute_cfa_program, __frame_state_for, uw_update_context_1,
uw_update_context): Use fs->regs.how[X] instead of fs->regs.reg[X].how
or fs.regs.how[X] instead of fs.regs.reg[X].how.
* config/sh/linux-unwind.h (sh_fallback_frame_state): Likewise.
* config/bfin/linux-unwind.h (bfin_fallback_frame_state): Likewise.
* config/pa/linux-unwind.h (pa32_fallback_frame_state): Likewise.
* config/pa/hpux-unwind.h (UPDATE_FS_FOR_SAR, UPDATE_FS_FOR_GR,
UPDATE_FS_FOR_FR, UPDATE_FS_FOR_PC, pa_fallback_frame_state):
Likewise.
* config/alpha/vms-unwind.h (alpha_vms_fallback_frame_state):
Likewise.
* config/alpha/linux-unwind.h (alpha_fallback_frame_state): Likewise.
* config/arc/linux-unwind.h (arc_fallback_frame_state,
arc_frob_update_context): Likewise.
* config/riscv/linux-unwind.h (riscv_fallback_frame_state): Likewise.
* config/nios2/linux-unwind.h (NIOS2_REG): Likewise.
* config/nds32/linux-unwind.h (NDS32_PUT_FS_REG): Likewise.
* config/s390/tpf-unwind.h (s390_fallback_frame_state): Likewise.
* config/s390/linux-unwind.h (s390_fallback_frame_state): Likewise.
* config/sparc/sol2-unwind.h (sparc64_frob_update_context,
MD_FALLBACK_FRAME_STATE_FOR): Likewise.
* config/sparc/linux-unwind.h (sparc64_fallback_frame_state,
sparc64_frob_update_context, sparc_fallback_frame_state): Likewise.
* config/i386/sol2-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/w32-unwind.h (i386_w32_fallback_frame_state): Likewise.
* config/i386/linux-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/freebsd-unwind.h (x86_64_freebsd_fallback_frame_state):
Likewise.
* config/i386/dragonfly-unwind.h
(x86_64_dragonfly_fallback_frame_state): Likewise.
* config/i386/gnu-unwind.h (x86_gnu_fallback_frame_state): Likewise.
* config/csky/linux-unwind.h (csky_fallback_frame_state): Likewise.
* config/aarch64/linux-unwind.h (aarch64_fallback_frame_state):
Likewise.
* config/aarch64/freebsd-unwind.h
(aarch64_freebsd_fallback_frame_state): Likewise.
* config/aarch64/aarch64-unwind.h (aarch64_frob_update_context):
Likewise.
* config/or1k/linux-unwind.h (or1k_fallback_frame_state): Likewise.
* config/mips/linux-unwind.h (mips_fallback_frame_state): Likewise.
* config/loongarch/linux-unwind.h (loongarch_fallback_frame_state):
Likewise.
* config/m68k/linux-unwind.h (m68k_fallback_frame_state): Likewise.
* config/xtensa/linux-unwind.h (xtensa_fallback_frame_state):
Likewise.
* config/rs6000/darwin-fallback.c (set_offset): Likewise.
* config/rs6000/aix-unwind.h (MD_FROB_UPDATE_CONTEXT): Likewise.
* config/rs6000/linux-unwind.h (ppc_fallback_frame_state): Likewise.
* config/rs6000/freebsd-unwind.h (frob_update_context): Likewise.
2022-10-06 11:04:52 +02:00
|
|
|
fs->regs.how[i + 16] = REG_SAVED_OFFSET;
|
2013-05-22 10:39:03 +00:00
|
|
|
fs->regs.reg[i + 16].loc.offset = i * sizeof(long);
|
2009-10-11 20:48:46 +00:00
|
|
|
}
|
|
|
|
|
2010-10-27 20:34:36 +00:00
|
|
|
/* Check whether we need to restore FPU registers. */
|
2011-02-14 12:32:11 +00:00
|
|
|
if (mctx->fpregs.fpu_qcnt)
|
2009-10-11 20:48:46 +00:00
|
|
|
{
|
|
|
|
for (i = 0; i < 32; i++)
|
|
|
|
{
|
libgcc: Decrease size of _Unwind_FrameState and even more size of cleared area in uw_frame_state_for
The following patch implements something that has Florian found as
low hanging fruit in our unwinder and has been discussed in the
https://gcc.gnu.org/wiki/cauldron2022#cauldron2022talks.inprocess_unwinding_bof
talk.
_Unwind_FrameState type seems to be (unlike the pre-GCC 3 frame_state
which has been part of ABI) private to unwind-dw2.c + unwind.inc it
includes, it is always defined on the stack of some entrypoints, initialized
by static uw_frame_state_for and the address of it is also passed to other
static functions or the static inlines handling machine dependent unwinding,
but it isn't fortunately passed to any callbacks or public functions, so I
think we can safely change it any time we want.
Florian mentioned that the structure is large even on x86_64, 384 bytes
there, starts with 328 bytes long element with frame_state_reg_info type
which then starts with an array with __LIBGCC_DWARF_FRAME_REGISTERS__ + 1
elements, each of them is 16 bytes long, on x86_64
__LIBGCC_DWARF_FRAME_REGISTERS__ is just 17 but even that is big, on say
riscv __LIBGCC_DWARF_FRAME_REGISTERS__ is I think 128, on powerpc 111,
on sh 153 etc. And, we memset to zero the whole fs variable with the
_Unwind_FrameState type at the start of the unwinding.
The reason why each element is 16 byte (on 64-bit arches) is that it
contains some pointer or pointer sized integer and then an enum (with just
7 different enumerators) + padding.
The following patch decreases it by moving the enum into a separate
array and using just one byte for each register in that second array.
We could compress it even more, say 4 bits per register, but I don't
want to uglify the code for it too much and make the accesses slower.
Furthermore, the clearing of the object can clear only thos how array
and members after it, because REG_UNSAVED enumerator (0) doesn't actually
need any pointer or pointer sized integer, it is just the other kinds
that need to have there something.
By doing this, on x86_64 the above numbers change to _Unwind_FrameState
type being now 264 bytes long, frame_state_reg_info 208 bytes and we
don't clear the first 144 bytes of the object, so the memset is 120 bytes,
so ~ 31% of the old clearing size. On riscv 64-bit assuming it has same
structure layout rules for the few types used there that would be
~ 2160 bytes of _Unwind_FrameState type before and ~ 1264 bytes after,
with the memset previously ~ 2160 bytes and after ~ 232 bytes after.
We've also talked about possibly adding a number of initially initialized
regs and initializing the rest lazily, but at least for x86_64 with
18 elements in the array that doesn't seem to be worth it anymore,
especially because return address column is 16 there and that is usually the
first thing to be touched. It might theory help with lots of registers if
they are usually untouched, but would uglify and complicate any stores to
how by having to check there for the not initialized yet cases and lazy
initialization, and similarly for all reads of how to do there if below
last initialized one, use how, otherwise imply REG_UNSAVED.
The disadvantage of the patch is that touching reg[x].loc and how[x]
now means 2 cachelines rather than one as before, and I admit beyond
bootstrap/regtest I haven't benchmarked it in any way.
2022-10-06 Jakub Jelinek <jakub@redhat.com>
* unwind-dw2.h (REG_UNSAVED, REG_SAVED_OFFSET, REG_SAVED_REG,
REG_SAVED_EXP, REG_SAVED_VAL_OFFSET, REG_SAVED_VAL_EXP,
REG_UNDEFINED): New anonymous enum, moved from inside of
struct frame_state_reg_info.
(struct frame_state_reg_info): Remove reg[].how element and the
anonymous enum there. Add how element.
* unwind-dw2.c: Include stddef.h.
(uw_frame_state_for): Don't clear first
offsetof (_Unwind_FrameState, regs.how[0]) bytes of *fs.
(execute_cfa_program, __frame_state_for, uw_update_context_1,
uw_update_context): Use fs->regs.how[X] instead of fs->regs.reg[X].how
or fs.regs.how[X] instead of fs.regs.reg[X].how.
* config/sh/linux-unwind.h (sh_fallback_frame_state): Likewise.
* config/bfin/linux-unwind.h (bfin_fallback_frame_state): Likewise.
* config/pa/linux-unwind.h (pa32_fallback_frame_state): Likewise.
* config/pa/hpux-unwind.h (UPDATE_FS_FOR_SAR, UPDATE_FS_FOR_GR,
UPDATE_FS_FOR_FR, UPDATE_FS_FOR_PC, pa_fallback_frame_state):
Likewise.
* config/alpha/vms-unwind.h (alpha_vms_fallback_frame_state):
Likewise.
* config/alpha/linux-unwind.h (alpha_fallback_frame_state): Likewise.
* config/arc/linux-unwind.h (arc_fallback_frame_state,
arc_frob_update_context): Likewise.
* config/riscv/linux-unwind.h (riscv_fallback_frame_state): Likewise.
* config/nios2/linux-unwind.h (NIOS2_REG): Likewise.
* config/nds32/linux-unwind.h (NDS32_PUT_FS_REG): Likewise.
* config/s390/tpf-unwind.h (s390_fallback_frame_state): Likewise.
* config/s390/linux-unwind.h (s390_fallback_frame_state): Likewise.
* config/sparc/sol2-unwind.h (sparc64_frob_update_context,
MD_FALLBACK_FRAME_STATE_FOR): Likewise.
* config/sparc/linux-unwind.h (sparc64_fallback_frame_state,
sparc64_frob_update_context, sparc_fallback_frame_state): Likewise.
* config/i386/sol2-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/w32-unwind.h (i386_w32_fallback_frame_state): Likewise.
* config/i386/linux-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/freebsd-unwind.h (x86_64_freebsd_fallback_frame_state):
Likewise.
* config/i386/dragonfly-unwind.h
(x86_64_dragonfly_fallback_frame_state): Likewise.
* config/i386/gnu-unwind.h (x86_gnu_fallback_frame_state): Likewise.
* config/csky/linux-unwind.h (csky_fallback_frame_state): Likewise.
* config/aarch64/linux-unwind.h (aarch64_fallback_frame_state):
Likewise.
* config/aarch64/freebsd-unwind.h
(aarch64_freebsd_fallback_frame_state): Likewise.
* config/aarch64/aarch64-unwind.h (aarch64_frob_update_context):
Likewise.
* config/or1k/linux-unwind.h (or1k_fallback_frame_state): Likewise.
* config/mips/linux-unwind.h (mips_fallback_frame_state): Likewise.
* config/loongarch/linux-unwind.h (loongarch_fallback_frame_state):
Likewise.
* config/m68k/linux-unwind.h (m68k_fallback_frame_state): Likewise.
* config/xtensa/linux-unwind.h (xtensa_fallback_frame_state):
Likewise.
* config/rs6000/darwin-fallback.c (set_offset): Likewise.
* config/rs6000/aix-unwind.h (MD_FROB_UPDATE_CONTEXT): Likewise.
* config/rs6000/linux-unwind.h (ppc_fallback_frame_state): Likewise.
* config/rs6000/freebsd-unwind.h (frob_update_context): Likewise.
2022-10-06 11:04:52 +02:00
|
|
|
fs->regs.how[i + 32] = REG_SAVED_OFFSET;
|
2009-10-11 20:48:46 +00:00
|
|
|
fs->regs.reg[i + 32].loc.offset
|
2011-02-14 12:32:11 +00:00
|
|
|
= (long)&mctx->fpregs.fpu_fr.fpu_regs[i] - new_cfa;
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef __arch64__
|
|
|
|
/* For 64-bit, fpu_fr.fpu_dregs contains 32 instead of 16 doubles. */
|
|
|
|
for (i = 32; i < 64; i++)
|
|
|
|
{
|
|
|
|
if (i > 32 && (i & 1))
|
|
|
|
continue;
|
|
|
|
|
libgcc: Decrease size of _Unwind_FrameState and even more size of cleared area in uw_frame_state_for
The following patch implements something that has Florian found as
low hanging fruit in our unwinder and has been discussed in the
https://gcc.gnu.org/wiki/cauldron2022#cauldron2022talks.inprocess_unwinding_bof
talk.
_Unwind_FrameState type seems to be (unlike the pre-GCC 3 frame_state
which has been part of ABI) private to unwind-dw2.c + unwind.inc it
includes, it is always defined on the stack of some entrypoints, initialized
by static uw_frame_state_for and the address of it is also passed to other
static functions or the static inlines handling machine dependent unwinding,
but it isn't fortunately passed to any callbacks or public functions, so I
think we can safely change it any time we want.
Florian mentioned that the structure is large even on x86_64, 384 bytes
there, starts with 328 bytes long element with frame_state_reg_info type
which then starts with an array with __LIBGCC_DWARF_FRAME_REGISTERS__ + 1
elements, each of them is 16 bytes long, on x86_64
__LIBGCC_DWARF_FRAME_REGISTERS__ is just 17 but even that is big, on say
riscv __LIBGCC_DWARF_FRAME_REGISTERS__ is I think 128, on powerpc 111,
on sh 153 etc. And, we memset to zero the whole fs variable with the
_Unwind_FrameState type at the start of the unwinding.
The reason why each element is 16 byte (on 64-bit arches) is that it
contains some pointer or pointer sized integer and then an enum (with just
7 different enumerators) + padding.
The following patch decreases it by moving the enum into a separate
array and using just one byte for each register in that second array.
We could compress it even more, say 4 bits per register, but I don't
want to uglify the code for it too much and make the accesses slower.
Furthermore, the clearing of the object can clear only thos how array
and members after it, because REG_UNSAVED enumerator (0) doesn't actually
need any pointer or pointer sized integer, it is just the other kinds
that need to have there something.
By doing this, on x86_64 the above numbers change to _Unwind_FrameState
type being now 264 bytes long, frame_state_reg_info 208 bytes and we
don't clear the first 144 bytes of the object, so the memset is 120 bytes,
so ~ 31% of the old clearing size. On riscv 64-bit assuming it has same
structure layout rules for the few types used there that would be
~ 2160 bytes of _Unwind_FrameState type before and ~ 1264 bytes after,
with the memset previously ~ 2160 bytes and after ~ 232 bytes after.
We've also talked about possibly adding a number of initially initialized
regs and initializing the rest lazily, but at least for x86_64 with
18 elements in the array that doesn't seem to be worth it anymore,
especially because return address column is 16 there and that is usually the
first thing to be touched. It might theory help with lots of registers if
they are usually untouched, but would uglify and complicate any stores to
how by having to check there for the not initialized yet cases and lazy
initialization, and similarly for all reads of how to do there if below
last initialized one, use how, otherwise imply REG_UNSAVED.
The disadvantage of the patch is that touching reg[x].loc and how[x]
now means 2 cachelines rather than one as before, and I admit beyond
bootstrap/regtest I haven't benchmarked it in any way.
2022-10-06 Jakub Jelinek <jakub@redhat.com>
* unwind-dw2.h (REG_UNSAVED, REG_SAVED_OFFSET, REG_SAVED_REG,
REG_SAVED_EXP, REG_SAVED_VAL_OFFSET, REG_SAVED_VAL_EXP,
REG_UNDEFINED): New anonymous enum, moved from inside of
struct frame_state_reg_info.
(struct frame_state_reg_info): Remove reg[].how element and the
anonymous enum there. Add how element.
* unwind-dw2.c: Include stddef.h.
(uw_frame_state_for): Don't clear first
offsetof (_Unwind_FrameState, regs.how[0]) bytes of *fs.
(execute_cfa_program, __frame_state_for, uw_update_context_1,
uw_update_context): Use fs->regs.how[X] instead of fs->regs.reg[X].how
or fs.regs.how[X] instead of fs.regs.reg[X].how.
* config/sh/linux-unwind.h (sh_fallback_frame_state): Likewise.
* config/bfin/linux-unwind.h (bfin_fallback_frame_state): Likewise.
* config/pa/linux-unwind.h (pa32_fallback_frame_state): Likewise.
* config/pa/hpux-unwind.h (UPDATE_FS_FOR_SAR, UPDATE_FS_FOR_GR,
UPDATE_FS_FOR_FR, UPDATE_FS_FOR_PC, pa_fallback_frame_state):
Likewise.
* config/alpha/vms-unwind.h (alpha_vms_fallback_frame_state):
Likewise.
* config/alpha/linux-unwind.h (alpha_fallback_frame_state): Likewise.
* config/arc/linux-unwind.h (arc_fallback_frame_state,
arc_frob_update_context): Likewise.
* config/riscv/linux-unwind.h (riscv_fallback_frame_state): Likewise.
* config/nios2/linux-unwind.h (NIOS2_REG): Likewise.
* config/nds32/linux-unwind.h (NDS32_PUT_FS_REG): Likewise.
* config/s390/tpf-unwind.h (s390_fallback_frame_state): Likewise.
* config/s390/linux-unwind.h (s390_fallback_frame_state): Likewise.
* config/sparc/sol2-unwind.h (sparc64_frob_update_context,
MD_FALLBACK_FRAME_STATE_FOR): Likewise.
* config/sparc/linux-unwind.h (sparc64_fallback_frame_state,
sparc64_frob_update_context, sparc_fallback_frame_state): Likewise.
* config/i386/sol2-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/w32-unwind.h (i386_w32_fallback_frame_state): Likewise.
* config/i386/linux-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/freebsd-unwind.h (x86_64_freebsd_fallback_frame_state):
Likewise.
* config/i386/dragonfly-unwind.h
(x86_64_dragonfly_fallback_frame_state): Likewise.
* config/i386/gnu-unwind.h (x86_gnu_fallback_frame_state): Likewise.
* config/csky/linux-unwind.h (csky_fallback_frame_state): Likewise.
* config/aarch64/linux-unwind.h (aarch64_fallback_frame_state):
Likewise.
* config/aarch64/freebsd-unwind.h
(aarch64_freebsd_fallback_frame_state): Likewise.
* config/aarch64/aarch64-unwind.h (aarch64_frob_update_context):
Likewise.
* config/or1k/linux-unwind.h (or1k_fallback_frame_state): Likewise.
* config/mips/linux-unwind.h (mips_fallback_frame_state): Likewise.
* config/loongarch/linux-unwind.h (loongarch_fallback_frame_state):
Likewise.
* config/m68k/linux-unwind.h (m68k_fallback_frame_state): Likewise.
* config/xtensa/linux-unwind.h (xtensa_fallback_frame_state):
Likewise.
* config/rs6000/darwin-fallback.c (set_offset): Likewise.
* config/rs6000/aix-unwind.h (MD_FROB_UPDATE_CONTEXT): Likewise.
* config/rs6000/linux-unwind.h (ppc_fallback_frame_state): Likewise.
* config/rs6000/freebsd-unwind.h (frob_update_context): Likewise.
2022-10-06 11:04:52 +02:00
|
|
|
fs->regs.how[i + 32] = REG_SAVED_OFFSET;
|
2011-02-14 12:32:11 +00:00
|
|
|
fs->regs.reg[i + 32].loc.offset
|
|
|
|
= (long)&mctx->fpregs.fpu_fr.fpu_dregs[i/2] - new_cfa;
|
2009-10-11 20:48:46 +00:00
|
|
|
}
|
2011-02-14 12:32:11 +00:00
|
|
|
#endif
|
2009-10-11 20:48:46 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* State the rules to find the kernel's code "return address", which is
|
|
|
|
the address of the active instruction when the signal was caught.
|
|
|
|
On the SPARC, since RETURN_ADDR_OFFSET (essentially 8) is defined, we
|
|
|
|
need to preventively subtract it from the purported return address. */
|
2011-02-14 12:32:11 +00:00
|
|
|
ra_location = &mctx->gregs[REG_PC];
|
|
|
|
shifted_ra_location = &mctx->gregs[REG_Y];
|
2009-10-11 20:48:46 +00:00
|
|
|
*(void **)shifted_ra_location = *(void **)ra_location - 8;
|
|
|
|
fs->retaddr_column = 0;
|
libgcc: Decrease size of _Unwind_FrameState and even more size of cleared area in uw_frame_state_for
The following patch implements something that has Florian found as
low hanging fruit in our unwinder and has been discussed in the
https://gcc.gnu.org/wiki/cauldron2022#cauldron2022talks.inprocess_unwinding_bof
talk.
_Unwind_FrameState type seems to be (unlike the pre-GCC 3 frame_state
which has been part of ABI) private to unwind-dw2.c + unwind.inc it
includes, it is always defined on the stack of some entrypoints, initialized
by static uw_frame_state_for and the address of it is also passed to other
static functions or the static inlines handling machine dependent unwinding,
but it isn't fortunately passed to any callbacks or public functions, so I
think we can safely change it any time we want.
Florian mentioned that the structure is large even on x86_64, 384 bytes
there, starts with 328 bytes long element with frame_state_reg_info type
which then starts with an array with __LIBGCC_DWARF_FRAME_REGISTERS__ + 1
elements, each of them is 16 bytes long, on x86_64
__LIBGCC_DWARF_FRAME_REGISTERS__ is just 17 but even that is big, on say
riscv __LIBGCC_DWARF_FRAME_REGISTERS__ is I think 128, on powerpc 111,
on sh 153 etc. And, we memset to zero the whole fs variable with the
_Unwind_FrameState type at the start of the unwinding.
The reason why each element is 16 byte (on 64-bit arches) is that it
contains some pointer or pointer sized integer and then an enum (with just
7 different enumerators) + padding.
The following patch decreases it by moving the enum into a separate
array and using just one byte for each register in that second array.
We could compress it even more, say 4 bits per register, but I don't
want to uglify the code for it too much and make the accesses slower.
Furthermore, the clearing of the object can clear only thos how array
and members after it, because REG_UNSAVED enumerator (0) doesn't actually
need any pointer or pointer sized integer, it is just the other kinds
that need to have there something.
By doing this, on x86_64 the above numbers change to _Unwind_FrameState
type being now 264 bytes long, frame_state_reg_info 208 bytes and we
don't clear the first 144 bytes of the object, so the memset is 120 bytes,
so ~ 31% of the old clearing size. On riscv 64-bit assuming it has same
structure layout rules for the few types used there that would be
~ 2160 bytes of _Unwind_FrameState type before and ~ 1264 bytes after,
with the memset previously ~ 2160 bytes and after ~ 232 bytes after.
We've also talked about possibly adding a number of initially initialized
regs and initializing the rest lazily, but at least for x86_64 with
18 elements in the array that doesn't seem to be worth it anymore,
especially because return address column is 16 there and that is usually the
first thing to be touched. It might theory help with lots of registers if
they are usually untouched, but would uglify and complicate any stores to
how by having to check there for the not initialized yet cases and lazy
initialization, and similarly for all reads of how to do there if below
last initialized one, use how, otherwise imply REG_UNSAVED.
The disadvantage of the patch is that touching reg[x].loc and how[x]
now means 2 cachelines rather than one as before, and I admit beyond
bootstrap/regtest I haven't benchmarked it in any way.
2022-10-06 Jakub Jelinek <jakub@redhat.com>
* unwind-dw2.h (REG_UNSAVED, REG_SAVED_OFFSET, REG_SAVED_REG,
REG_SAVED_EXP, REG_SAVED_VAL_OFFSET, REG_SAVED_VAL_EXP,
REG_UNDEFINED): New anonymous enum, moved from inside of
struct frame_state_reg_info.
(struct frame_state_reg_info): Remove reg[].how element and the
anonymous enum there. Add how element.
* unwind-dw2.c: Include stddef.h.
(uw_frame_state_for): Don't clear first
offsetof (_Unwind_FrameState, regs.how[0]) bytes of *fs.
(execute_cfa_program, __frame_state_for, uw_update_context_1,
uw_update_context): Use fs->regs.how[X] instead of fs->regs.reg[X].how
or fs.regs.how[X] instead of fs.regs.reg[X].how.
* config/sh/linux-unwind.h (sh_fallback_frame_state): Likewise.
* config/bfin/linux-unwind.h (bfin_fallback_frame_state): Likewise.
* config/pa/linux-unwind.h (pa32_fallback_frame_state): Likewise.
* config/pa/hpux-unwind.h (UPDATE_FS_FOR_SAR, UPDATE_FS_FOR_GR,
UPDATE_FS_FOR_FR, UPDATE_FS_FOR_PC, pa_fallback_frame_state):
Likewise.
* config/alpha/vms-unwind.h (alpha_vms_fallback_frame_state):
Likewise.
* config/alpha/linux-unwind.h (alpha_fallback_frame_state): Likewise.
* config/arc/linux-unwind.h (arc_fallback_frame_state,
arc_frob_update_context): Likewise.
* config/riscv/linux-unwind.h (riscv_fallback_frame_state): Likewise.
* config/nios2/linux-unwind.h (NIOS2_REG): Likewise.
* config/nds32/linux-unwind.h (NDS32_PUT_FS_REG): Likewise.
* config/s390/tpf-unwind.h (s390_fallback_frame_state): Likewise.
* config/s390/linux-unwind.h (s390_fallback_frame_state): Likewise.
* config/sparc/sol2-unwind.h (sparc64_frob_update_context,
MD_FALLBACK_FRAME_STATE_FOR): Likewise.
* config/sparc/linux-unwind.h (sparc64_fallback_frame_state,
sparc64_frob_update_context, sparc_fallback_frame_state): Likewise.
* config/i386/sol2-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/w32-unwind.h (i386_w32_fallback_frame_state): Likewise.
* config/i386/linux-unwind.h (x86_64_fallback_frame_state,
x86_fallback_frame_state): Likewise.
* config/i386/freebsd-unwind.h (x86_64_freebsd_fallback_frame_state):
Likewise.
* config/i386/dragonfly-unwind.h
(x86_64_dragonfly_fallback_frame_state): Likewise.
* config/i386/gnu-unwind.h (x86_gnu_fallback_frame_state): Likewise.
* config/csky/linux-unwind.h (csky_fallback_frame_state): Likewise.
* config/aarch64/linux-unwind.h (aarch64_fallback_frame_state):
Likewise.
* config/aarch64/freebsd-unwind.h
(aarch64_freebsd_fallback_frame_state): Likewise.
* config/aarch64/aarch64-unwind.h (aarch64_frob_update_context):
Likewise.
* config/or1k/linux-unwind.h (or1k_fallback_frame_state): Likewise.
* config/mips/linux-unwind.h (mips_fallback_frame_state): Likewise.
* config/loongarch/linux-unwind.h (loongarch_fallback_frame_state):
Likewise.
* config/m68k/linux-unwind.h (m68k_fallback_frame_state): Likewise.
* config/xtensa/linux-unwind.h (xtensa_fallback_frame_state):
Likewise.
* config/rs6000/darwin-fallback.c (set_offset): Likewise.
* config/rs6000/aix-unwind.h (MD_FROB_UPDATE_CONTEXT): Likewise.
* config/rs6000/linux-unwind.h (ppc_fallback_frame_state): Likewise.
* config/rs6000/freebsd-unwind.h (frob_update_context): Likewise.
2022-10-06 11:04:52 +02:00
|
|
|
fs->regs.how[0] = REG_SAVED_OFFSET;
|
2011-02-14 12:32:11 +00:00
|
|
|
fs->regs.reg[0].loc.offset = (long)shifted_ra_location - new_cfa;
|
2013-05-28 13:52:50 +00:00
|
|
|
|
|
|
|
/* SIGFPE for IEEE-754 exceptions is delivered after the faulting insn
|
|
|
|
rather than before it, so don't set fs->signal_frame in that case.
|
|
|
|
We test whether the cexc field of the FSR is zero. */
|
|
|
|
if ((mctx->fpregs.fpu_fsr & 0x1f) == 0)
|
|
|
|
fs->signal_frame = 1;
|
2009-10-11 20:48:46 +00:00
|
|
|
|
|
|
|
return _URC_NO_REASON;
|
2011-02-14 12:32:11 +00:00
|
|
|
}
|