glibc/sysdeps/powerpc/powerpc64/__longjmp-common.S
Ulrich Weigand 8b8a692cfd PowerPC64 ELFv2 ABI 4/6: Stack frame layout changes
This updates glibc for the changes in the ELFv2 relating to the
stack frame layout.  These are described in more detail here:
http://gcc.gnu.org/ml/gcc-patches/2013-11/msg01149.html
http://gcc.gnu.org/ml/gcc-patches/2013-11/msg01146.html

Specifically, the "compiler and linker doublewords" were removed,
which has the effect that the save slot for the TOC register is
now at offset 24 rather than 40 to the stack pointer.

In addition, a function may now no longer necessarily assume that
its caller has set up a 64-byte register save area its use.

To address the first change, the patch goes through all assembler
files and replaces immediate offsets in instructions accessing the
ABI-defined stack slots by symbolic offsets.  Those already were
defined in ucontext_i.sym and used in some of the context routines,
but that doesn't really seem like the right place for those defines.

The patch instead defines those symbolic offsets in sysdeps.h,
in two variants for the old and new ABI, and uses them systematically
in all assembler files, not just the context routines.

The second change only affected a few assembler files that used
the save area to temporarily store some registers.  In those
cases where this happens within a leaf function, this patch
changes the code to store those registers to the "red zone"
below the stack pointer.  Otherwise, the functions already allocate
a stack frame, and the patch changes them to add extra space in
these frames as temporary space for the ELFv2 ABI.
2013-12-04 07:41:39 -06:00

179 lines
4.5 KiB
ArmAsm

/* longjmp for PowerPC64.
Copyright (C) 1995-2013 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, see
<http://www.gnu.org/licenses/>. */
#include <sysdep.h>
#define _ASM
#define _SETJMP_H
#ifdef __NO_VMX__
# include <novmxsetjmp.h>
#else
# include <jmpbuf-offsets.h>
#endif
#ifndef __NO_VMX__
.section ".toc","aw"
.LC__dl_hwcap:
# ifdef SHARED
# ifdef IS_IN_rtld
/* Inside ld.so we use the local alias to avoid runtime GOT
relocations. */
.tc _rtld_local_ro[TC],_rtld_local_ro
# else
.tc _rtld_global_ro[TC],_rtld_global_ro
# endif
# else
.tc _dl_hwcap[TC],_dl_hwcap
# endif
.section ".text"
#endif
.machine "altivec"
ENTRY (__longjmp)
CALL_MCOUNT 2
#ifndef __NO_VMX__
ld r5,.LC__dl_hwcap@toc(r2)
# ifdef SHARED
/* Load _rtld-global._dl_hwcap. */
ld r5,RTLD_GLOBAL_RO_DL_HWCAP_OFFSET(r5)
# else
ld r5,0(r5) /* Load extern _dl_hwcap. */
# endif
andis. r5,r5,(PPC_FEATURE_HAS_ALTIVEC >> 16)
beq L(no_vmx)
la r5,((JB_VRS)*8)(3)
andi. r6,r5,0xf
lwz r0,((JB_VRSAVE)*8)(3) /* 32-bit VRSAVE. */
mtspr VRSAVE,r0
beq+ L(aligned_restore_vmx)
addi r6,r5,16
lvsl v0,0,r5
lvx v1,0,r5
addi r5,r5,32
lvx v21,0,r6
vperm v20,v1,v21,v0
# define load_misaligned_vmx_lo_loaded(loadvr,lovr,shiftvr,loadgpr,addgpr) \
addi addgpr,addgpr,32; \
lvx lovr,0,loadgpr; \
vperm loadvr,loadvr,lovr,shiftvr;
load_misaligned_vmx_lo_loaded(v21,v22,v0,r5,r6)
load_misaligned_vmx_lo_loaded(v22,v23,v0,r6,r5)
load_misaligned_vmx_lo_loaded(v23,v24,v0,r5,r6)
load_misaligned_vmx_lo_loaded(v24,v25,v0,r6,r5)
load_misaligned_vmx_lo_loaded(v25,v26,v0,r5,r6)
load_misaligned_vmx_lo_loaded(v26,v27,v0,r6,r5)
load_misaligned_vmx_lo_loaded(v27,v28,v0,r5,r6)
load_misaligned_vmx_lo_loaded(v28,v29,v0,r6,r5)
load_misaligned_vmx_lo_loaded(v29,v30,v0,r5,r6)
load_misaligned_vmx_lo_loaded(v30,v31,v0,r6,r5)
lvx v1,0,r5
vperm v31,v31,v1,v0
b L(no_vmx)
L(aligned_restore_vmx):
addi r6,r5,16
lvx v20,0,r5
addi r5,r5,32
lvx v21,0,r6
addi r6,r6,32
lvx v22,0,r5
addi r5,r5,32
lvx v23,0,r6
addi r6,r6,32
lvx v24,0,r5
addi r5,r5,32
lvx v25,0,r6
addi r6,r6,32
lvx v26,0,r5
addi r5,r5,32
lvx v27,0,r6
addi r6,r6,32
lvx v28,0,r5
addi r5,r5,32
lvx v29,0,r6
addi r6,r6,32
lvx v30,0,r5
lvx v31,0,r6
L(no_vmx):
#endif
#if defined PTR_DEMANGLE || defined CHECK_SP
ld r22,(JB_GPR1*8)(r3)
#else
ld r1,(JB_GPR1*8)(r3)
#endif
#ifdef PTR_DEMANGLE
# ifdef CHECK_SP
PTR_DEMANGLE3 (r22, r22, r25)
# else
PTR_DEMANGLE3 (r1, r22, r25)
# endif
#endif
#ifdef CHECK_SP
CHECK_SP (r22)
mr r1,r22
#endif
ld r2,(JB_GPR2*8)(r3)
ld r0,(JB_LR*8)(r3)
ld r14,((JB_GPRS+0)*8)(r3)
lfd fp14,((JB_FPRS+0)*8)(r3)
#if defined SHARED && !defined IS_IN_rtld
std r2,FRAME_TOC_SAVE(r1) /* Restore the callers TOC save area. */
#endif
ld r15,((JB_GPRS+1)*8)(r3)
lfd fp15,((JB_FPRS+1)*8)(r3)
ld r16,((JB_GPRS+2)*8)(r3)
lfd fp16,((JB_FPRS+2)*8)(r3)
ld r17,((JB_GPRS+3)*8)(r3)
lfd fp17,((JB_FPRS+3)*8)(r3)
ld r18,((JB_GPRS+4)*8)(r3)
lfd fp18,((JB_FPRS+4)*8)(r3)
ld r19,((JB_GPRS+5)*8)(r3)
lfd fp19,((JB_FPRS+5)*8)(r3)
ld r20,((JB_GPRS+6)*8)(r3)
lfd fp20,((JB_FPRS+6)*8)(r3)
#ifdef PTR_DEMANGLE
PTR_DEMANGLE2 (r0, r25)
#endif
mtlr r0
/* std r2,FRAME_TOC_SAVE(r1) Restore the TOC save area. */
ld r21,((JB_GPRS+7)*8)(r3)
lfd fp21,((JB_FPRS+7)*8)(r3)
ld r22,((JB_GPRS+8)*8)(r3)
lfd fp22,((JB_FPRS+8)*8)(r3)
lwz r0,((JB_CR*8)+4)(r3) /* 32-bit CR. */
ld r23,((JB_GPRS+9)*8)(r3)
lfd fp23,((JB_FPRS+9)*8)(r3)
ld r24,((JB_GPRS+10)*8)(r3)
lfd fp24,((JB_FPRS+10)*8)(r3)
ld r25,((JB_GPRS+11)*8)(r3)
lfd fp25,((JB_FPRS+11)*8)(r3)
mtcrf 0xFF,r0
ld r26,((JB_GPRS+12)*8)(r3)
lfd fp26,((JB_FPRS+12)*8)(r3)
ld r27,((JB_GPRS+13)*8)(r3)
lfd fp27,((JB_FPRS+13)*8)(r3)
ld r28,((JB_GPRS+14)*8)(r3)
lfd fp28,((JB_FPRS+14)*8)(r3)
ld r29,((JB_GPRS+15)*8)(r3)
lfd fp29,((JB_FPRS+15)*8)(r3)
ld r30,((JB_GPRS+16)*8)(r3)
lfd fp30,((JB_FPRS+16)*8)(r3)
ld r31,((JB_GPRS+17)*8)(r3)
lfd fp31,((JB_FPRS+17)*8)(r3)
mr r3,r4
blr
END (__longjmp)