CbC/CbC_gcc: gcc/config/alpha/osf5-unwind.h annotate

author	Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp>
date	Sun, 21 Aug 2011 07:07:55 +0900
parents
children

rev	line source
68 561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	1 /* DWARF2 EH unwinding support for Alpha Tru64.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	2 Copyright (C) 2010 Free Software Foundation, Inc.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	3
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	4 This file is part of GCC.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	5
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	6 GCC is free software; you can redistribute it and/or modify it under
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	7 the terms of the GNU General Public License as published by the Free
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	8 Software Foundation; either version 3, or (at your option) any later
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	9 version.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	10
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	11 GCC is distributed in the hope that it will be useful, but WITHOUT ANY
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	12 WARRANTY; without even the implied warranty of MERCHANTABILITY or
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	13 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	14 for more details.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	15
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	16 You should have received a copy of the GNU General Public License
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	17 along with GCC; see the file COPYING3. If not see
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	18 <http://www.gnu.org/licenses/>. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	19
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	20 /* This file implements the MD_FALLBACK_FRAME_STATE_FOR macro, triggered when
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	21 the GCC table based unwinding process hits a frame for which no unwind info
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	22 has been registered. This typically occurs when raising an exception from a
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	23 signal handler, because the handler is actually called from the OS kernel.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	24
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	25 The basic idea is to detect that we are indeed trying to unwind past a
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	26 signal handler and to fill out the GCC internal unwinding structures for
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	27 the OS kernel frame as if it had been directly called from the interrupted
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	28 context.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	29
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	30 This is all assuming that the code to set the handler asked the kernel to
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	31 pass a pointer to such context information. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	32
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	33 /* --------------------------------------------------------------------------
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	34 -- Basic principles of operation:
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	35 --------------------------------------------------------------------------
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	36
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	37 1/ We first need a way to detect if we are trying to unwind past a signal
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	38 handler.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	39
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	40 The typical method that is used on most platforms is to look at the code
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	41 around the return address we have and check if it matches the OS code
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	42 calling a handler. To determine what this code is expected to be, get a
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	43 breakpoint into a real signal handler and look at the code around the
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	44 return address. Depending on the library versions the pattern of the
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	45 signal handler is different; this is the reason why we check against more
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	46 than one pattern.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	47
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	48 On this target, the return address is right after the call and every
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	49 instruction is 4 bytes long. For the simple case of a null dereference in
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	50 a single-threaded app, it went like:
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	51
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	52 # Check that we indeed have something we expect: the instruction right
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	53 # before the return address is within a __sigtramp function and is a call.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	54
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	55 [... run gdb and break at the signal handler entry ...]
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	56
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	57 (gdb) x /i $ra-4
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	58 <__sigtramp+160>: jsr ra,(a3),0x3ff800d0ed4 <_fpdata+36468>
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	59
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	60 # Look at the code around that return address, and eventually observe a
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	61 # significantly large chunk of constant code right before the call:
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	62
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	63 (gdb) x /10i $ra-44
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	64 <__sigtramp+120>: lda gp,-27988(gp)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	65 <__sigtramp+124>: ldq at,-18968(gp)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	66 <__sigtramp+128>: lda t0,-1
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	67 <__sigtramp+132>: stq t0,0(at)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	68 <__sigtramp+136>: ldq at,-18960(gp)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	69 <__sigtramp+140>: ldl t1,8(at)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	70 <__sigtramp+144>: ldq at,-18960(gp)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	71 <__sigtramp+148>: stl t1,12(at)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	72 <__sigtramp+152>: ldq at,-18960(gp)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	73 <__sigtramp+156>: stl t0,8(at)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	74
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	75 # The hexadecimal equivalent that we will have to match is:
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	76
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	77 (gdb) x /10x $ra-44
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	78 <__sigtramp+120>: 0x23bd92ac 0xa79db5e8 0x203fffff 0xb43c0000
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	79 <__sigtramp+136>: 0xa79db5f0 0xa05c0008 0xa79db5f0 0xb05c000c
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	80 <__sigtramp+152>: 0xa79db5f0 0xb03c0008
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	81
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	82 The problem observed on this target with this approach is that although
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	83 we found a constant set of instruction patterns there were some
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	84 gp-related offsets that made the machine code to differ from one
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	85 installation to another. This problem could have been overcome by masking
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	86 these offsets, but we found that it would be simpler and more efficient to
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	87 check whether the return address was part of a signal handler, by comparing
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	88 it against some expected code offset from __sigtramp.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	89
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	90 # Check that we indeed have something we expect: the instruction
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	91 # right before the return address is within a __sigtramp
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	92 # function and is a call. We also need to obtain the offset
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	93 # between the return address and the start address of __sigtramp.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	94
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	95 [... run gdb and break at the signal handler entry ...]
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	96
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	97 (gdb) x /2i $ra-4
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	98 <__sigtramp+160>: jsr ra,(a3),0x3ff800d0ed4 <_fpdata+36468>
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	99 <__sigtramp+164>: ldah gp,16381(ra)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	100
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	101 (gdb) p (long)$ra - (long)&__sigtramp
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	102 $2 = 164
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	103
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	104 --------------------------------------------------------------------------
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	105
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	106 2/ Once we know we are going through a signal handler, we need a way to
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	107 retrieve information about the interrupted run-time context.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	108
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	109 On this platform, the third handler's argument is a pointer to a structure
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	110 describing this context (struct sigcontext *). We unfortunately have no
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	111 direct way to transfer this value here, so a couple of tricks are required
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	112 to compute it.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	113
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	114 As documented at least in some header files (e.g. sys/machine/context.h),
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	115 the structure the handler gets a pointer to is located on the stack. As of
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	116 today, while writing this macro, we have unfortunately not been able to
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	117 find a detailed description of the full stack layout at handler entry time,
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	118 so we'll have to resort to empirism :)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	119
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	120 When unwinding here, we have the handler's CFA at hand, as part of the
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	121 current unwinding context which is one of our arguments. We presume that
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	122 for each call to a signal handler by the same kernel routine, the context's
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	123 structure location on the stack is always at the same offset from the
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	124 handler's CFA, and we compute that offset from bare observation:
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	125
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	126 For the simple case of a bare null dereference in a single-threaded app,
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	127 computing the offset was done using GNAT like this:
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	128
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	129 # Break on the first handler's instruction, before the prologue to have the
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	130 # CFA in $sp, and get there:
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	131
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	132 (gdb) b *&__gnat_error_handler
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	133 Breakpoint 1 at 0x120016090: file init.c, line 378.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	134
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	135 (gdb) r
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	136 Program received signal SIGSEGV, Segmentation fault.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	137
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	138 (gdb) c
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	139 Breakpoint 1, __gnat_error_handler (sig=..., sip=..., context=...)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	140
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	141 # The displayed argument value are meaningless because we stopped before
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	142 # their final "homing". We know they are passed through $a0, $a1 and $a2
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	143 # from the ABI, though, so ...
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	144
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	145 # Observe that $sp and the context pointer are in the same (stack) area,
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	146 # and compute the offset:
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	147
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	148 (gdb) p /x $sp
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	149 $2 = 0x11fffbc80
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	150
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	151 (gdb) p /x $a2
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	152 $3 = 0x11fffbcf8
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	153
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	154 (gdb) p /x (long)$a2 - (long)$sp
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	155 $4 = 0x78
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	156
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	157 --------------------------------------------------------------------------
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	158
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	159 3/ Once we know we are unwinding through a signal handler and have the
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	160 address of the structure describing the interrupted context at hand, we
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	161 have to fill the internal frame-state/unwind-context structures properly
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	162 to allow the unwinding process to proceed.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	163
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	164 Roughly, we are provided with an unwinding CONTEXT, describing the state
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	165 of some point P in the call chain we are unwinding through. The macro we
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	166 implement has to fill a "frame state" structure FS that describe the P's
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	167 caller state, by way of rules to compute its CFA, return address, and
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	168 saved registers locations.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	169
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	170 For the case we are going to deal with, the caller is some kernel code
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	171 calling a signal handler, and:
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	172
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	173 o The saved registers are all in the interrupted run-time context,
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	174
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	175 o The CFA is the stack pointer value when the kernel code is entered, that
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	176 is, the stack pointer value at the interruption point, also part of the
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	177 interrupted run-time context.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	178
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	179 o We want the return address to appear as the address of the active
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	180 instruction at the interruption point, so that the unwinder proceeds as
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	181 if the interruption had been a regular call. This address is also part
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	182 of the interrupted run-time context.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	183
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	184 --
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	185
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	186 Also, note that there is an important difference between the return address
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	187 we need to claim for the kernel frame and the value of the return address
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	188 register at the interruption point.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	189
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	190 The latter might be required to be able to unwind past the interrupted
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	191 routine, for instance if it is interrupted before saving the incoming
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	192 register value in its own frame, which may typically happen during stack
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	193 probes for stack-checking purposes.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	194
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	195 It is then essential that the rules stated to locate the kernel frame
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	196 return address don't clobber the rules describing where is saved the return
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	197 address register at the interruption point, so some scratch register state
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	198 entry should be used for the former. We have DWARF_ALT_FRAME_RETURN_COLUMN
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	199 at hand exactly for that purpose.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	200
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	201 --------------------------------------------------------------------------
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	202
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	203 4/ Depending on the context (single-threaded or multi-threaded app, ...),
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	204 the code calling the handler and the handler-cfa to interrupted-context
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	205 offset might change, so we use a simple generic data structure to track
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	206 the possible variants. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	207
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	208 /* This is the structure to wrap information about each possible sighandler
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	209 caller we may have to identify. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	210
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	211 typedef struct {
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	212 /* Expected return address when being called from a sighandler. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	213 void *ra_value;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	214
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	215 /* Offset to get to the sigcontext structure from the handler's CFA
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	216 when the pattern matches. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	217 int cfa_to_context_offset;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	218
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	219 } sighandler_call_t;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	220
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	221 /* Helper macro for MD_FALLBACK_FRAME_STATE_FOR below.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	222
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	223 Look at RA to see if it matches within a sighandler caller.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	224 Set SIGCTX to the corresponding sigcontext structure (computed from
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	225 CFA) if it does, or to 0 otherwise. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	226
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	227 #define COMPUTE_SIGCONTEXT_FOR(RA,CFA,SIGCTX) \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	228 do { \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	229 /* Define and register the applicable patterns. */ \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	230 extern void __sigtramp (void); \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	231 \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	232 sighandler_call_t sighandler_calls [] = { \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	233 {__sigtramp + 164, 0x78} \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	234 }; \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	235 \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	236 int n_patterns_to_match \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	237 = sizeof (sighandler_calls) / sizeof (sighandler_call_t); \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	238 \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	239 int pn; /* pattern number */ \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	240 \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	241 int match = 0; /* Did last pattern match ? */ \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	242 \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	243 /* Try to match each pattern in turn. */ \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	244 for (pn = 0; !match && pn < n_patterns_to_match; pn ++) \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	245 match = ((RA) == sighandler_calls[pn].ra_value); \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	246 \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	247 (SIGCTX) = (struct sigcontext *) \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	248 (match ? ((CFA) + sighandler_calls[pn - 1].cfa_to_context_offset) : 0); \
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	249 } while (0);
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	250
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	251 #include <sys/context_t.h>
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	252
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	253 #define REG_SP 30 /* hard reg for stack pointer */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	254 #define REG_RA 26 /* hard reg for return address */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	255
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	256 #define MD_FALLBACK_FRAME_STATE_FOR alpha_fallback_frame_state
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	257
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	258 static _Unwind_Reason_Code
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	259 alpha_fallback_frame_state (struct _Unwind_Context *context,
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	260 _Unwind_FrameState *fs)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	261 {
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	262 /* Return address and CFA of the frame we're attempting to unwind through,
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	263 possibly a signal handler. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	264 void ctx_ra = (void )context->ra;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	265 void ctx_cfa = (void )context->cfa;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	266
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	267 /* CFA of the intermediate abstract kernel frame between the interrupted
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	268 code and the signal handler, if we're indeed unwinding through a signal
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	269 handler. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	270 void *k_cfa;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	271
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	272 /* Pointer to the sigcontext structure pushed by the kernel when we're
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	273 unwinding through a signal handler. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	274 struct sigcontext *sigctx;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	275 int i;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	276
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	277 COMPUTE_SIGCONTEXT_FOR (ctx_ra, ctx_cfa, sigctx);
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	278
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	279 if (sigctx == 0)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	280 return _URC_END_OF_STACK;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	281
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	282 /* The kernel frame's CFA is exactly the stack pointer value at the
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	283 interruption point. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	284 k_cfa = (void *) sigctx->sc_regs [REG_SP];
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	285
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	286 /* State the rules to compute the CFA we have the value of: use the
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	287 previous CFA and offset by the difference between the two. See
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	288 uw_update_context_1 for the supporting details. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	289 fs->regs.cfa_how = CFA_REG_OFFSET;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	290 fs->regs.cfa_reg = __builtin_dwarf_sp_column ();
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	291 fs->regs.cfa_offset = k_cfa - ctx_cfa;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	292
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	293 /* Fill the internal frame_state structure with information stating
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	294 where each register of interest in the saved context can be found
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	295 from the CFA. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	296
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	297 /* The general registers are in sigctx->sc_regs. Leave out r31, which
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	298 is read-as-zero. It makes no sense restoring it, and we are going to
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	299 use the state entry for the kernel return address rule below.
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	300
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	301 This loop must cover at least all the callee-saved registers, and
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	302 we just don't bother specializing the set here. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	303 for (i = 0; i <= 30; i ++)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	304 {
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	305 fs->regs.reg[i].how = REG_SAVED_OFFSET;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	306 fs->regs.reg[i].loc.offset
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	307 = (void ) &sigctx->sc_regs[i] - (void ) k_cfa;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	308 }
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	309
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	310 /* Ditto for the floating point registers in sigctx->sc_fpregs. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	311 for (i = 0; i <= 31; i ++)
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	312 {
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	313 fs->regs.reg[32+i].how = REG_SAVED_OFFSET;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	314 fs->regs.reg[32+i].loc.offset
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	315 = (void ) &sigctx->sc_fpregs[i] - (void ) k_cfa;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	316 }
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	317
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	318 /* State the rules to find the kernel's code "return address", which
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	319 is the address of the active instruction when the signal was caught,
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	320 in sigctx->sc_pc. Use DWARF_ALT_FRAME_RETURN_COLUMN since the return
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	321 address register is a general register and should be left alone. */
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	322 fs->retaddr_column = DWARF_ALT_FRAME_RETURN_COLUMN;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	323 fs->regs.reg[DWARF_ALT_FRAME_RETURN_COLUMN].how = REG_SAVED_OFFSET;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	324 fs->regs.reg[DWARF_ALT_FRAME_RETURN_COLUMN].loc.offset
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	325 = (void ) &sigctx->sc_pc - (void ) k_cfa;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	326 fs->signal_frame = 1;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	327
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	328 return _URC_NO_REASON;
561a7518be6b update gcc-4.6 Nobuyasu Oshiro <dimolto@cr.ie.u-ryukyu.ac.jp> parents: diff changeset	329 }

68

561a7518be6b update gcc-4.6