6885 CTF Everywhere Part 1

6886 Want ctfdiff
6887 ctfdump should be written in terms of libctf
6888 ctfmerge should be implemented in terms of libctf
6889 ctfconvert should be implemented in terms of libctf
6890 Want general workq
6891 Want general mergeq
6892 ctf_add_encoded assigns() incorrect byte size to types
6893 ctf_add_{struct,union,enum} can reuse forwards
6894 ctf_add_{struct,union,enum} occasionally forget to dirty the ctf_file_t
6895 ctf_add_member could better handle bitfields
6896 ctf_type_size() reports wrong size for forwards
6897 Want libctf ctf_kind_name() function
6898 Want libctf function to set struct/union size
Portions contributed by: John Levon <john.levon@joyent.com>
Portions contributed by: Richard Lowe <richlowe@richlowe.net>
Reviewed by: John Levon <john.levon@joyent.com>
Reviewed by: Andy Fiddaman <andy@omniosce.org>
Reviewed by: Gergő Doma <domag02@gmail.com>
Approved by: Dan McDonald <danmcd@joyent.com>

1 files changed, 311 insertions, 0 deletions

diff --git a/usr/src/lib/libdwarf/common/dwarf_util.h b/usr/src/lib/libdwarf/common/dwarf_util.h
new file mode 100644
index 0000000000..4046bb2478
--- /dev/null
+++ b/usr/src/lib/libdwarf/common/dwarf_util.h
@@ -0,0 +1,311 @@
+#ifndef DWARF_UTIL_H
+#define DWARF_UTIL_H
+/*
+
+  Copyright (C) 2000,2003,2004 Silicon Graphics, Inc.  All Rights Reserved.
+  Portions Copyright (C) 2007-2010 David Anderson. All Rights Reserved.
+
+  This program is free software; you can redistribute it and/or modify it
+  under the terms of version 2.1 of the GNU Lesser General Public License 
+  as published by the Free Software Foundation.
+
+  This program is distributed in the hope that it would be useful, but
+  WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  
+
+  Further, this software is distributed without any warranty that it is
+  free of the rightful claim of any third person regarding infringement 
+  or the like.  Any license provided herein, whether implied or 
+  otherwise, applies only to this software file.  Patent licenses, if
+  any, provided herein do not apply to combinations of this program with 
+  other software, or any other product whatsoever.  
+
+  You should have received a copy of the GNU Lesser General Public
+  License along with this program; if not, write the Free Software
+  Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston MA 02110-1301,
+  USA.
+
+  Contact information:  Silicon Graphics, Inc., 1500 Crittenden Lane,
+  Mountain View, CA 94043, or:
+
+  http://www.sgi.com
+
+  For further information regarding this notice, see:
+
+  http://oss.sgi.com/projects/GenInfo/NoticeExplan
+
+*/
+/* The address of the Free Software Foundation is
+   Free Software Foundation, Inc., 51 Franklin St, Fifth Floor, 
+   Boston, MA 02110-1301, USA.
+   SGI has moved from the Crittenden Lane address.
+*/
+
+
+
+
+/*
+    Decodes unsigned leb128 encoded numbers.
+    Make sure ptr is a pointer to a 1-byte type.  
+    In 2003 and earlier this was a hand-inlined
+    version of _dwarf_decode_u_leb128() which did
+    not work correctly if Dwarf_Word was 64 bits.
+*/
+#define DECODE_LEB128_UWORD(ptr, value) \
+    do { \
+       Dwarf_Word uleblen; \
+	value = _dwarf_decode_u_leb128(ptr,&uleblen); \
+        ptr += uleblen; \
+    } while (0)
+
+/*
+    Decodes signed leb128 encoded numbers.
+    Make sure ptr is a pointer to a 1-byte type.
+    In 2003 and earlier this was a hand-inlined
+    version of _dwarf_decode_s_leb128() which did
+    not work correctly if Dwarf_Word was 64 bits.
+
+*/
+#define DECODE_LEB128_SWORD(ptr, value) \
+    do { \
+       Dwarf_Word sleblen; \
+	value = _dwarf_decode_s_leb128(ptr,&sleblen); \
+        ptr += sleblen; \
+    } while(0)
+
+
+/*
+    Skips leb128_encoded numbers that are guaranteed 
+    to be no more than 4 bytes long.  Same for both
+    signed and unsigned numbers.
+*/
+#define SKIP_LEB128_WORD(ptr) \
+    do{ if ((*(ptr++) & 0x80) != 0) { \
+        if ((*(ptr++) & 0x80) != 0) { \
+            if ((*(ptr++) & 0x80) != 0) { \
+	        if ((*(ptr++) & 0x80) != 0) { \
+	        } \
+	    } \
+        } \
+    } } while (0)
+
+
+#define CHECK_DIE(die, error_ret_value) \
+do {if (die == NULL) { \
+	_dwarf_error(NULL, error, DW_DLE_DIE_NULL); \
+	return(error_ret_value); \
+    } \
+    if (die->di_cu_context == NULL) { \
+	_dwarf_error(NULL, error, DW_DLE_DIE_NO_CU_CONTEXT); \
+	return(error_ret_value); \
+    } \
+    if (die->di_cu_context->cc_dbg == NULL) { \
+	_dwarf_error(NULL, error, DW_DLE_DBG_NULL); \
+	return(error_ret_value); \
+    }  \
+} while (0)
+
+
+/* 
+   Reads 'source' for 'length' bytes from unaligned addr.
+
+   Avoids any constant-in-conditional warnings and
+   avoids a test in the generated code (for non-const cases,
+	which are in the majority.)
+   Uses a temp to avoid the test.
+   The decl here should avoid any problem of size in the temp.
+   This code is ENDIAN DEPENDENT
+   The memcpy args are the endian issue.
+*/
+typedef Dwarf_Unsigned BIGGEST_UINT;
+
+#ifdef WORDS_BIGENDIAN
+#define READ_UNALIGNED(dbg,dest,desttype, source, length) \
+    do { \
+      BIGGEST_UINT _ltmp = 0;  \
+      dbg->de_copy_word( (((char *)(&_ltmp)) + sizeof(_ltmp) - length), \
+			source, length) ; \
+      dest = (desttype)_ltmp;  \
+    } while (0)
+
+
+/*
+    This macro sign-extends a variable depending on the length.
+    It fills the bytes between the size of the destination and
+    the length with appropriate padding.
+    This code is ENDIAN DEPENDENT but dependent only
+    on host endianness, not object file endianness.
+    The memcpy args are the issue.
+*/
+#define SIGN_EXTEND(dest, length) \
+    do {if (*(Dwarf_Sbyte *)((char *)&dest + sizeof(dest) - length) < 0) {\
+	memcpy((char *)&dest, "\xff\xff\xff\xff\xff\xff\xff\xff", \
+	    sizeof(dest) - length);  \
+        } \
+     } while (0)
+#else /* LITTLE ENDIAN */
+
+#define READ_UNALIGNED(dbg,dest,desttype, source, length) \
+    do  { \
+      BIGGEST_UINT _ltmp = 0;  \
+      dbg->de_copy_word( (char *)(&_ltmp) , \
+                        source, length) ; \
+      dest = (desttype)_ltmp;  \
+     } while (0)
+
+
+/*
+    This macro sign-extends a variable depending on the length.
+    It fills the bytes between the size of the destination and
+    the length with appropriate padding.
+    This code is ENDIAN DEPENDENT but dependent only
+    on host endianness, not object file endianness.
+    The memcpy args are the issue.
+*/
+#define SIGN_EXTEND(dest, length) \
+    do {if (*(Dwarf_Sbyte *)((char *)&dest + (length-1)) < 0) {\
+        memcpy((char *)&dest+length,    \
+                "\xff\xff\xff\xff\xff\xff\xff\xff", \
+            sizeof(dest) - length); \
+        }  \
+    } while (0)
+
+#endif /* ! LITTLE_ENDIAN */
+
+
+
+/*
+   READ_AREA LENGTH reads the length (the older way
+   of pure 32 or 64 bit
+   or the new proposed dwarfv2.1 64bit-extension way)
+
+   It reads the bits from where rw_src_data_p  points to 
+   and updates the rw_src_data_p to point past what was just read.
+
+   It updates w_length_size (to the size of an offset, either 4 or 8)
+   and w_exten_size (set 0 unless this frame has the DWARF3,4 64bit
+   extension, in which case w_exten_size is set to 4).
+
+   r_dbg is just the current dbg pointer.
+   w_target is the output length field.
+   r_targtype is the output type. Always Dwarf_Unsigned so far.
+  
+*/
+/* This one handles the v2.1 64bit extension  
+   and 32bit (and   MIPS fixed 64  bit via the
+	dwarf_init-set r_dbg->de_length_size)..
+   It does not recognize any but the one distingushed value
+   (the only one with defined meaning).
+   It assumes that no CU will have a length
+	0xffffffxx  (32bit length)
+	or
+	0xffffffxx xxxxxxxx (64bit length)
+   which makes possible auto-detection of the extension.
+
+   This depends on knowing that only a non-zero length
+   is legitimate (AFAICT), and for IRIX non-standard -64 
+   dwarf that the first 32 bits of the 64bit offset will be
+   zero (because the compiler could not handle a truly large 
+   value as of Jan 2003 and because no app has that much debug 
+   info anyway, at least not in the IRIX case).
+
+   At present not testing for '64bit elf' here as that
+   does not seem necessary (none of the 64bit length seems 
+   appropriate unless it's  ident[EI_CLASS] == ELFCLASS64).
+*/
+#   define    READ_AREA_LENGTH(r_dbg,w_target,r_targtype,         \
+	rw_src_data_p,w_length_size,w_exten_size)                 \
+do {    READ_UNALIGNED(r_dbg,w_target,r_targtype,                 \
+                rw_src_data_p, ORIGINAL_DWARF_OFFSET_SIZE);       \
+    if(w_target == DISTINGUISHED_VALUE) {                         \
+	     /* dwarf3 64bit extension */                         \
+             w_length_size  = DISTINGUISHED_VALUE_OFFSET_SIZE;    \
+             rw_src_data_p += ORIGINAL_DWARF_OFFSET_SIZE;         \
+             w_exten_size   = ORIGINAL_DWARF_OFFSET_SIZE;         \
+             READ_UNALIGNED(r_dbg,w_target,r_targtype,            \
+                  rw_src_data_p, DISTINGUISHED_VALUE_OFFSET_SIZE);\
+             rw_src_data_p += DISTINGUISHED_VALUE_OFFSET_SIZE;    \
+    } else {                                                      \
+	if(w_target == 0 && r_dbg->de_big_endian_object) {        \
+	     /* IRIX 64 bit, big endian.  This test */            \
+	     /* is not a truly precise test, a precise test */    \
+             /* would check if the target was IRIX.  */           \
+             READ_UNALIGNED(r_dbg,w_target,r_targtype,            \
+                rw_src_data_p, DISTINGUISHED_VALUE_OFFSET_SIZE);  \
+	     w_length_size  = DISTINGUISHED_VALUE_OFFSET_SIZE;    \
+	     rw_src_data_p += DISTINGUISHED_VALUE_OFFSET_SIZE;    \
+	     w_exten_size = 0;                                    \
+	} else {                                                  \
+	     /* standard 32 bit dwarf2/dwarf3 */                  \
+	     w_exten_size   = 0;                                  \
+             w_length_size  = ORIGINAL_DWARF_OFFSET_SIZE;         \
+             rw_src_data_p += w_length_size;                      \
+	}                                                         \
+    } } while(0)
+
+Dwarf_Unsigned
+_dwarf_decode_u_leb128(Dwarf_Small * leb128,
+		       Dwarf_Word * leb128_length);
+
+Dwarf_Signed
+_dwarf_decode_s_leb128(Dwarf_Small * leb128,
+		       Dwarf_Word * leb128_length);
+
+Dwarf_Unsigned
+_dwarf_get_size_of_val(Dwarf_Debug dbg,
+    Dwarf_Unsigned form,
+    Dwarf_Half address_size,
+    Dwarf_Small * val_ptr, 
+    int v_length_size);
+
+struct Dwarf_Hash_Table_Entry_s;
+/* This single struct is the base for the hash table.
+   The intent is that once the total_abbrev_count across
+   all the entries is greater than  10*current_table_entry_count
+   one should build a new Dwarf_Hash_Table_Base_s, rehash
+   all the existing entries, and delete the old table and entries. 
+   (10 is a heuristic, nothing magic about it, but once the
+   count gets to 30 or 40 times current_table_entry_count
+   things really slow down a lot. One (500MB) application had
+   127000 abbreviations in one compilation unit)
+   The incoming 'code' is an abbrev number and those simply
+   increase linearly so the hashing is perfect always.
+*/
+struct Dwarf_Hash_Table_s {
+      unsigned long       tb_table_entry_count;
+      unsigned long       tb_total_abbrev_count;
+      /* Each table entry is a list of abbreviations. */
+      struct  Dwarf_Hash_Table_Entry_s *tb_entries;
+};
+
+/*
+    This struct is used to build a hash table for the
+    abbreviation codes for a compile-unit.  
+*/
+struct Dwarf_Hash_Table_Entry_s {
+    Dwarf_Abbrev_List at_head;
+};
+
+
+
+Dwarf_Abbrev_List
+_dwarf_get_abbrev_for_code(Dwarf_CU_Context cu_context,
+			   Dwarf_Unsigned code);
+
+
+/* return 1 if string ends before 'endptr' else
+** return 0 meaning string is not properly terminated.
+** Presumption is the 'endptr' pts to end of some dwarf section data.
+*/
+int _dwarf_string_valid(void *startptr, void *endptr);
+
+Dwarf_Unsigned _dwarf_length_of_cu_header(Dwarf_Debug,
+					  Dwarf_Unsigned offset);
+Dwarf_Unsigned _dwarf_length_of_cu_header_simple(Dwarf_Debug);
+
+int  _dwarf_load_debug_info(Dwarf_Debug dbg, Dwarf_Error *error);
+void _dwarf_free_abbrev_hash_table_contents(Dwarf_Debug dbg,
+    struct Dwarf_Hash_Table_s* hash_table);
+int _dwarf_get_address_size(Dwarf_Debug dbg, Dwarf_Die die);
+
+#endif /* DWARF_UTIL_H */