* [RFD+PATCH] ISA bit treatment on the MIPS platform @ 2012-05-14 15:37 Maciej W. Rozycki 2012-05-18 22:01 ` Maciej W. Rozycki 2012-06-11 18:21 ` Joel Brobecker 0 siblings, 2 replies; 27+ messages in thread From: Maciej W. Rozycki @ 2012-05-14 15:37 UTC (permalink / raw) To: gdb-patches; +Cc: Rich Fuhler, Richard Sandiford, binutils, gcc Hello everybody, I am cc-ing GCC and binutils mailing lists because while my proposal is contained in GDB, it is GCC and binutils that produce DWARF-2 records that GDB has to deal with. Feedback from all the parties involved is very warmly welcome, be it enthusiastic or more critical. I will therefore start with a short explanation of the matters considered, so that those of us with good ideas and extensive knowledge of our tools, binary formats, respective standards, etc. but who do not specialise in the area of MIPS processors have a better understanding what the underlying implication are. 0. Introduction As some of you may have been aware, GDB has had issues with handling MIPS16 binaries related to the use of the ISA (Instruction Set Architecture) bit as the selector between the regular MIPS and the MIPS16 execution (ISA) mode. With the imminent addition of microMIPS support to GDB these issues will apply there as well as the microMIPS execution mode uses the ISA bit exactly as the MIPS16 mode does. 1. Background information The MIPS architecture, as originally designed and implemented in mid-1980s has a uniform instruction word size that is 4 bytes, naturally aligned. As such all MIPS instructions are located at addresses that have their bits #1 and #0 set to zeroes, and any attempt to execute an instruction from an address that has any of the two bits set to one causes an address error exception. This may for example happen when a jump-register instruction is executed whose register value used as the jump target has any of these bits set. Then in mid 1990s LSI sought a way to improve code density for their TinyRISC family of MIPS cores and invented an alternatively encoded instruction set in a joint effort with MIPS Technologies (then a subsidiary of SGI). The new instruction set has been named the MIPS16 ASE (Application-Specific Extension) and uses a variable instruction word size, which is 2 bytes (as the name of the ASE suggests) for most, but there are a couple of exceptions that take 4 bytes, and then most of the 2-byte instructions can be treated with a 2-byte extension prefix to expand the range of the immediate operands used. As a result instructions are no longer 4-byte aligned, instead they are aligned to a multiple of 2. That left the bit #0 still unused for code references, be it for the standard MIPS (i.e. as originally invented) or for the MIPS16 instruction set, and based on that observation a clever trick was invented that on one hand allowed the processor to be seamlessly switched between the two instruction sets at any time at the run time while on the other avoided the introduction of any special control register to do that. So it is the bit #0 of the instruction address that was chosen as the selector and named the ISA bit. Any instruction executed at an even address is interpreted as a standard MIPS instruction (the address still has to have its bit #1 clear), any instruction executed at an odd address is interpreted as a MIPS16 instruction. To switch between modes ordinary jump instructions are used, such as used for function calls and returns, specifically the bit #0 of the source register used in jump-register instructions selects the execution (ISA) mode for the following piece of code to be interpreted in. Additionally new jump-immediate instructions were added that flipped the ISA bit to select the opposite mode upon execution. They were considered necessary to avoid the need to make register jumps in all cases as the original jump-immediate instructions provided no way to change the bit #0 at all. This was all important for cases where standard MIPS and MIPS16 code had to be mixed, either for compatibility with the existing binary code base or to access resources not reachable from MIPS16 code (the MIPS16 instruction set only provides access to general-purpose registers, and not for example floating-point unit registers or privileged coprocessor 0 registers) -- pieces of code in the opposite mode can be executed as ordinary subroutine calls. A similar approach has been more recently adopted for the MIPS16 replacement instruction set defined as the so called microMIPS ASE. This is another instruction set encoding introduced to the MIPS architecture. Just like the MIPS16 ASE, the microMIPS instruction set uses a variable-length encoding, where each instruction takes a multiple of 2 bytes. The ISA bit has been reused and for microMIPS-capable processors selects between the standard MIPS and the microMIPS mode instead. All our tools have supported, to the various level of quality, the MIPS16 ASE instruction set and associated binary file structures for a long time now. Support for the microMIPS instruction set and binary file peculiarities has been recently added to binutils and is being reviewed for GDB; my understanding is the corresponding GCC bits are supposed to follow in a not so distant future. Anything considered here therefore equally applies to both MIPS16 and microMIPS binaries, although for the purpose of this consideration I will focus on the MIPS16 ASE, as well established in our code base now. 2. Statement of the problem To put it shortly, MIPS16 and microMIPS code pointers used by GDB are different to these observed at the run time. This results in the same expressions being evaluated producing different results in GDB and in the program being debugged. Obviously it's the results obtained at the run time that are correct (they define how the program behaves) and therefore by definition the results obtained in GDB are incorrect. A bit longer description will record that obviously at the run time the ISA bit has to be set correctly (refer to background information above if unsure why so) or the program will not run as expected. This is recorded in all the executable file structures used at the run time: the dynamic symbol table (but not always the static one!), the GOT, and obviously in all the addresses embedded in code or data of the program itself, calculated by applying the appropriate relocations at the static link time. While a program is being processed by GDB, the ISA bit is stripped off from any code addresses, presumably to make them the same as the respective raw memory byte address used by the processor to access the instruction in the instruction fetch access cycle. This stripping is actually performed outside GDB proper, in BFD, specifically _bfd_mips_elf_symbol_processing (elfxx-mips.c, see the piece of code at the very bottom of that function, starting with an: "If this is an odd-valued function symbol, assume it's a MIPS16 or microMIPS one." comment). This function is also responsible for symbol table dumps made by `objdump' too, so you'll never see the ISA bit reported there by that tool, you need to use `readelf'. This is however unlike what is ever done at the run time, the ISA bit once present is never stripped off, for example a cast like this: (short *) main will not strip the ISA bit off and if the resulting pointer is intended to be used to access instructions as data, for example for software instruction decoding (like for fault recovery or emulation in a signal handler) or for self-modifying code then the bit still has to be stripped off by an explicit AND operation. This is probably best illustrated with a simple real program example. Let's consider the following simple program: $ cat foobar.c int __attribute__ ((mips16)) foo (void) { return 1; } int __attribute__ ((mips16)) bar (void) { return 2; } int __attribute__ ((nomips16)) foo32 (void) { return 3; } int (*foo32p) (void) = foo32; int (*foop) (void) = foo; int fooi = (int) foo; int main (void) { return foop (); } $ This is plain C with no odd tricks, except from the instruction mode attributes. They are not necessary to trigger this problem, I just put them here so that the program can be contained in a single source file and to make it obvious which function is MIPS16 code and which is not. Let's try it with Linux, so that everyone can repeat this experiment: $ mips-linux-gnu-gcc -mips16 -g -O2 -o foobar foobar.c $ Let's have a look at some interesting symbols: $ mips-linux-gnu-readelf -s foobar | egrep 'table|foo|bar' Symbol table '.dynsym' contains 7 entries: Symbol table '.symtab' contains 95 entries: 55: 00000000 0 FILE LOCAL DEFAULT ABS foobar.c 66: 0040068c 4 FUNC GLOBAL DEFAULT [MIPS16] 12 bar 68: 00410848 4 OBJECT GLOBAL DEFAULT 21 foo32p 70: 00410844 4 OBJECT GLOBAL DEFAULT 21 foop 78: 00400684 8 FUNC GLOBAL DEFAULT 12 foo32 80: 00400680 4 FUNC GLOBAL DEFAULT [MIPS16] 12 foo 88: 00410840 4 OBJECT GLOBAL DEFAULT 21 fooi $ Hmm, no sight of the ISA bit, but notice how foo and bar (but not foo32!) have been marked as MIPS16 functions (ELF symbol structure's st_other field is used for that). So let's try to run and poke at this program with GDB. I'll be using a native system for simplicity (I'll be using ellipses here and there to remove unrelated clutter): $ ./foobar $ echo $? 1 $ So far, so good. $ gdb ./foobar [...] (gdb) break main Breakpoint 1 at 0x400490: file foobar.c, line 23. (gdb) run Starting program: .../foobar Breakpoint 1, main () at foobar.c:23 23 return foop (); (gdb) Yay, it worked! OK, so let's poke at it: (gdb) print main $1 = {int (void)} 0x400490 <main> (gdb) print foo32 $2 = {int (void)} 0x400684 <foo32> (gdb) print foo32p $3 = (int (*)(void)) 0x400684 <foo32> (gdb) print bar $4 = {int (void)} 0x40068c <bar> (gdb) print foo $5 = {int (void)} 0x400680 <foo> (gdb) print foop $6 = (int (*)(void)) 0x400681 <foo> (gdb) A-ha! Here's the difference and finally the ISA bit! (gdb) print /x fooi $7 = 0x400681 (gdb) p/x $pc p/x $pc $8 = 0x400491 (gdb) And here as well... (gdb) advance foo foo () at foobar.c:4 4 } (gdb) disassemble Dump of assembler code for function foo: 0x00400680 <+0>: jr ra 0x00400682 <+2>: li v0,1 End of assembler dump. (gdb) finish Run till exit from #0 foo () at foobar.c:4 main () at foobar.c:24 24 } Value returned is $9 = 1 (gdb) continue Continuing. [Inferior 1 (process 14103) exited with code 01] (gdb) So let's be a bit inquisitive... (gdb) run Starting program: .../foobar Breakpoint 1, main () at foobar.c:23 23 return foop (); (gdb) Actually we do not like to run foo here at all. Let's run bar instead! (gdb) set foop = bar (gdb) print foop $10 = (int (*)(void)) 0x40068c <bar> (gdb) Hmm, no ISA bit. Is it going to work? (gdb) advance bar bar () at foobar.c:9 9 } (gdb) p/x $pc $11 = 0x40068c (gdb) disassemble Dump of assembler code for function bar: => 0x0040068c <+0>: jr ra 0x0040068e <+2>: li v0,2 End of assembler dump. (gdb) finish Run till exit from #0 bar () at foobar.c:9 Program received signal SIGILL, Illegal instruction. bar () at foobar.c:9 9 } (gdb) Oops! (gdb) p/x $pc $12 = 0x40068c (gdb) We're still there! (gdb) continue Continuing. Program terminated with signal SIGILL, Illegal instruction. The program no longer exists. (gdb) So let's try something else: (gdb) run Starting program: .../foobar Breakpoint 1, main () at foobar.c:23 23 return foop (); (gdb) set foop = foo (gdb) advance foo foo () at foobar.c:4 4 } (gdb) disassemble Dump of assembler code for function foo: => 0x00400680 <+0>: jr ra 0x00400682 <+2>: li v0,1 End of assembler dump. (gdb) finish Run till exit from #0 foo () at foobar.c:4 Program received signal SIGILL, Illegal instruction. foo () at foobar.c:4 4 } (gdb) continue Continuing. Program terminated with signal SIGILL, Illegal instruction. The program no longer exists. (gdb) The same problem! (gdb) run Starting program: /net/build2-lucid-cs/scratch/macro/mips-linux-fsf-gcc/isa-bit/foobar Breakpoint 1, main () at foobar.c:23 23 return foop (); (gdb) set foop = foo32 (gdb) advance foo32 foo32 () at foobar.c:14 14 } (gdb) disassemble Dump of assembler code for function foo32: => 0x00400684 <+0>: jr ra 0x00400688 <+4>: li v0,3 End of assembler dump. (gdb) finish Run till exit from #0 foo32 () at foobar.c:14 main () at foobar.c:24 24 } Value returned is $14 = 3 (gdb) continue Continuing. [Inferior 1 (process 14113) exited with code 03] (gdb) That did work though, so it's the ISA bit only! (gdb) quit Enough! That's the tip of the iceberg only though. So let's rebuild the executable with some dynamic symbols: $ mips-linux-gnu-gcc -mips16 -Wl,--export-dynamic -g -O2 -o foobar-dyn foobar.c $ mips-linux-gnu-readelf -s foobar-dyn | egrep 'table|foo|bar' Symbol table '.dynsym' contains 32 entries: 6: 004009cd 4 FUNC GLOBAL DEFAULT 12 bar 8: 00410b88 4 OBJECT GLOBAL DEFAULT 21 foo32p 9: 00410b84 4 OBJECT GLOBAL DEFAULT 21 foop 15: 004009c4 8 FUNC GLOBAL DEFAULT 12 foo32 17: 004009c1 4 FUNC GLOBAL DEFAULT 12 foo 25: 00410b80 4 OBJECT GLOBAL DEFAULT 21 fooi Symbol table '.symtab' contains 95 entries: 55: 00000000 0 FILE LOCAL DEFAULT ABS foobar.c 69: 004009cd 4 FUNC GLOBAL DEFAULT 12 bar 71: 00410b88 4 OBJECT GLOBAL DEFAULT 21 foo32p 72: 00410b84 4 OBJECT GLOBAL DEFAULT 21 foop 79: 004009c4 8 FUNC GLOBAL DEFAULT 12 foo32 81: 004009c1 4 FUNC GLOBAL DEFAULT 12 foo 89: 00410b80 4 OBJECT GLOBAL DEFAULT 21 fooi $ OK, now the ISA bit is there for a change, but the MIPS16 st_other attribute gone, hmm... What does `objdump' do then: $ mips-linux-gnu-objdump -Tt foobar-dyn | egrep 'SYMBOL|foo|bar' foobar-dyn: file format elf32-tradbigmips SYMBOL TABLE: 00000000 l df *ABS* 00000000 foobar.c 004009cc g F .text 00000004 0xf0 bar 00410b88 g O .data 00000004 foo32p 00410b84 g O .data 00000004 foop 004009c4 g F .text 00000008 foo32 004009c0 g F .text 00000004 0xf0 foo 00410b80 g O .data 00000004 fooi DYNAMIC SYMBOL TABLE: 004009cc g DF .text 00000004 Base 0xf0 bar 00410b88 g DO .data 00000004 Base foo32p 00410b84 g DO .data 00000004 Base foop 004009c4 g DF .text 00000008 Base foo32 004009c0 g DF .text 00000004 Base 0xf0 foo 00410b80 g DO .data 00000004 Base fooi $ Hmm, the attribute (0xf0, printed raw) is back, and the ISA bit gone again. Let's have a look at some DWARF-2 records GDB uses (I'll be stripping off a lot here for brevity) -- debug info: $ mips-linux-gnu-readelf -wi foobar Contents of the .debug_info section: [...] Compilation Unit @ offset 0x88: Length: 0xbb (32-bit) Version: 4 Abbrev Offset: 62 Pointer Size: 4 <0><93>: Abbrev Number: 1 (DW_TAG_compile_unit) <94> DW_AT_producer : (indirect string, offset: 0x19e): GNU C 4.8.0 20120513 (experimental) -meb -mips16 -march=mips32r2 -mhard-float -mllsc -mplt -mno-synci -mno-shared -mabi=32 -g -O2 <98> DW_AT_language : 1 (ANSI C) <99> DW_AT_name : (indirect string, offset: 0x190): foobar.c <9d> DW_AT_comp_dir : (indirect string, offset: 0x225): [...] <a1> DW_AT_ranges : 0x0 <a5> DW_AT_low_pc : 0x0 <a9> DW_AT_stmt_list : 0x27 <1><ad>: Abbrev Number: 2 (DW_TAG_subprogram) <ae> DW_AT_external : 1 <ae> DW_AT_name : foo <b2> DW_AT_decl_file : 1 <b3> DW_AT_decl_line : 1 <b4> DW_AT_prototyped : 1 <b4> DW_AT_type : <0xc2> <b8> DW_AT_low_pc : 0x400680 <bc> DW_AT_high_pc : 0x400684 <c0> DW_AT_frame_base : 1 byte block: 9c (DW_OP_call_frame_cfa) <c2> DW_AT_GNU_all_call_sites: 1 <1><c2>: Abbrev Number: 3 (DW_TAG_base_type) <c3> DW_AT_byte_size : 4 <c4> DW_AT_encoding : 5 (signed) <c5> DW_AT_name : int <1><c9>: Abbrev Number: 4 (DW_TAG_subprogram) <ca> DW_AT_external : 1 <ca> DW_AT_name : (indirect string, offset: 0x18a): foo32 <ce> DW_AT_decl_file : 1 <cf> DW_AT_decl_line : 11 <d0> DW_AT_prototyped : 1 <d0> DW_AT_type : <0xc2> <d4> DW_AT_low_pc : 0x400684 <d8> DW_AT_high_pc : 0x40068c <dc> DW_AT_frame_base : 1 byte block: 9c (DW_OP_call_frame_cfa) <de> DW_AT_GNU_all_call_sites: 1 <1><de>: Abbrev Number: 2 (DW_TAG_subprogram) <df> DW_AT_external : 1 <df> DW_AT_name : bar <e3> DW_AT_decl_file : 1 <e4> DW_AT_decl_line : 6 <e5> DW_AT_prototyped : 1 <e5> DW_AT_type : <0xc2> <e9> DW_AT_low_pc : 0x40068c <ed> DW_AT_high_pc : 0x400690 <f1> DW_AT_frame_base : 1 byte block: 9c (DW_OP_call_frame_cfa) <f3> DW_AT_GNU_all_call_sites: 1 <1><f3>: Abbrev Number: 5 (DW_TAG_subprogram) <f4> DW_AT_external : 1 <f4> DW_AT_name : (indirect string, offset: 0x199): main <f8> DW_AT_decl_file : 1 <f9> DW_AT_decl_line : 21 <fa> DW_AT_prototyped : 1 <fa> DW_AT_type : <0xc2> <fe> DW_AT_low_pc : 0x400490 <102> DW_AT_high_pc : 0x4004a4 <106> DW_AT_frame_base : 1 byte block: 9c (DW_OP_call_frame_cfa) <108> DW_AT_GNU_all_tail_call_sites: 1 [...] $ -- no sign of the ISA bit anywhere -- frame info: $ mips-linux-gnu-readelf -wf foobar [...] Contents of the .debug_frame section: 00000000 0000000c ffffffff CIE Version: 1 Augmentation: "" Code alignment factor: 1 Data alignment factor: -4 Return address column: 31 DW_CFA_def_cfa_register: r29 DW_CFA_nop 00000010 0000000c 00000000 FDE cie=00000000 pc=00400680..00400684 00000020 0000000c 00000000 FDE cie=00000000 pc=00400684..0040068c 00000030 0000000c 00000000 FDE cie=00000000 pc=0040068c..00400690 00000040 00000018 00000000 FDE cie=00000000 pc=00400490..004004a4 DW_CFA_advance_loc: 6 to 00400496 DW_CFA_def_cfa_offset: 32 DW_CFA_offset: r31 at cfa-4 DW_CFA_advance_loc: 6 to 0040049c DW_CFA_restore: r31 DW_CFA_def_cfa_offset: 0 DW_CFA_nop DW_CFA_nop DW_CFA_nop [...] $ -- no sign of the ISA bit anywhere -- range info (GDB doesn't use arange): $ mips-linux-gnu-readelf -wR foobar Contents of the .debug_ranges section: Offset Begin End 00000000 00400680 00400690 00000000 00400490 004004a4 00000000 <End of list> $ -- no sign of the ISA bit anywhere -- line info: $ mips-linux-gnu-readelf -wl foobar Raw dump of debug contents of section .debug_line: [...] Offset: 0x27 Length: 78 DWARF Version: 2 Prologue Length: 31 Minimum Instruction Length: 1 Initial value of 'is_stmt': 1 Line Base: -5 Line Range: 14 Opcode Base: 13 Opcodes: Opcode 1 has 0 args Opcode 2 has 1 args Opcode 3 has 1 args Opcode 4 has 1 args Opcode 5 has 1 args Opcode 6 has 0 args Opcode 7 has 0 args Opcode 8 has 0 args Opcode 9 has 1 args Opcode 10 has 0 args Opcode 11 has 0 args Opcode 12 has 1 args The Directory Table is empty. The File Name Table: Entry Dir Time Size Name 1 0 0 0 foobar.c Line Number Statements: Extended opcode 2: set Address to 0x400681 Special opcode 6: advance Address by 0 to 0x400681 and Line by 1 to 2 Special opcode 7: advance Address by 0 to 0x400681 and Line by 2 to 4 Special opcode 55: advance Address by 3 to 0x400684 and Line by 8 to 12 Special opcode 7: advance Address by 0 to 0x400684 and Line by 2 to 14 Advance Line by -7 to 7 Special opcode 131: advance Address by 9 to 0x40068d and Line by 0 to 7 Special opcode 7: advance Address by 0 to 0x40068d and Line by 2 to 9 Advance PC by 3 to 0x400690 Extended opcode 1: End of Sequence Extended opcode 2: set Address to 0x400491 Advance Line by 21 to 22 Copy Special opcode 6: advance Address by 0 to 0x400491 and Line by 1 to 23 Special opcode 60: advance Address by 4 to 0x400495 and Line by -1 to 22 Special opcode 34: advance Address by 2 to 0x400497 and Line by 1 to 23 Special opcode 62: advance Address by 4 to 0x40049b and Line by 1 to 24 Special opcode 32: advance Address by 2 to 0x40049d and Line by -1 to 23 Special opcode 6: advance Address by 0 to 0x40049d and Line by 1 to 24 Advance PC by 7 to 0x4004a4 Extended opcode 1: End of Sequence [...] -- a-ha, the ISA bit is there! However it's not always right for some reason, I don't have a small test case to show it, but here's an excerpt from MIPS16 libc, a prologue of a function: 00019630 <__libc_init_first>: 19630: e8a0 jrc ra 19632: 6500 nop 00019634 <_init>: 19634: f000 6a11 li v0,17 19638: f7d8 0b08 la v1,15e00 <_DYNAMIC+0x15c54> 1963c: f400 3240 sll v0,16 19640: e269 addu v0,v1 19642: 659a move gp,v0 19644: 64f6 save 48,ra,s0-s1 19646: 671c move s0,gp 19648: d204 sw v0,16(sp) 1964a: f352 984c lw v0,-27828(s0) 1964e: 6724 move s1,a0 and the corresponding DWARF-2 line info: Line Number Statements: Extended opcode 2: set Address to 0x19631 Advance Line by 44 to 45 Copy Special opcode 8: advance Address by 0 to 0x19631 and Line by 3 to 48 Special opcode 66: advance Address by 4 to 0x19635 and Line by 5 to 53 Advance PC by constant 17 to 0x19646 Special opcode 25: advance Address by 1 to 0x19647 and Line by 6 to 59 Advance Line by -6 to 53 Special opcode 33: advance Address by 2 to 0x19649 and Line by 0 to 53 Special opcode 39: advance Address by 2 to 0x1964b and Line by 6 to 59 Advance Line by -6 to 53 Special opcode 61: advance Address by 4 to 0x1964f and Line by 0 to 53 -- see that "Advance PC by constant 17" there? It clears the ISA bit, however code at 0x19646 is not standard MIPS code at all. For some reason the constant is always 17, I've never seen DW_LNS_const_add_pc used with any other value -- is that a binutils bug or what? 3. Proposed solution: I think we should retain the value of the ISA bit in code references, that is effectively treat them as cookies as they indeed are (although trivially calculated) rather than raw memory byte addresses. In a perfect world both the static symbol table and the respective DWARF-2 records should be fixed to include the ISA bit in all the cases. I think however that this is infeasible. All the uses of _bfd_mips_elf_symbol_processing can not necessarily be tracked down. This function is used by elf_slurp_symbol_table that in turn is used by bfd_canonicalize_symtab and bfd_canonicalize_dynamic_symtab, which are public interfaces. Similarly DWARF-2 records are used outside GDB, one notable if a bit questionable is the exception unwinder (libgcc/unwind-dw2.c) -- I have identified at least bits in execute_cfa_program and uw_frame_state_for, both around the calls to _Unwind_IsSignalFrame, that would need an update as they effectively flip the ISA bit freely; see also the comment about MASK_RETURN_ADDR in gcc/config/mips/mips.h. But there may be more places. Any change in how DWARF-2 records are produced would require an update there and would cause compatibility problems with libgcc.a binaries already distributed; given that this is a static library a complex change involving function renames would likely be required. I propose therefore to accept the existing inconsistencies and deal with them entirely within GDB. I have figured out that the ISA bit lost in various places can still be recovered as long as we have symbol information -- that'll have the st_other attribute correctly set to one of standard MIPS/MIPS16/microMIPS. Here's the resulting change. It adds a couple of new gdbarch hooks, one to update symbol information with the ISA bit lost in _bfd_mips_elf_symbol_processing, and two other ones to adjust DWARF-2 records as they're processed. I have figured out only FDE addresses, line information and ranges need to be adjusted; CFA records in particular are FDE-relative and therefore work "automagically" after these adjustments. The ISA bit is set in each address handled according to information retrieved from the symbol table for the symbol spanning the address if any; limits are adjusted based on the address they point to related to the respective base address. Additionally minimal symbol information has to be adjusted accordingly in its gdbarch hook. With these changes in place some complications with ISA bit juggling in the PC that never fully worked can be removed from the MIPS backend. Conversely, the generic dynamic linker event special breakpoint symbol handler has to be updated to call the minimal symbol gdbarch hook to record that the symbol is a MIPS16 or microMIPS address if applicable or the breakpoint will be set at the wrong address and either fail to work or cause SIGTRAPs (this is because the symbol is handled early on and bypasses regular symbol processing). 4. Results obtained The change fixes the example above -- to repeat only the crucial steps: (gdb) break main Breakpoint 1 at 0x400491: file foobar.c, line 23. (gdb) run Starting program: .../foobar Breakpoint 1, main () at foobar.c:23 23 return foop (); (gdb) print foo $1 = {int (void)} 0x400681 <foo> (gdb) set foop = bar (gdb) advance bar bar () at foobar.c:9 9 } (gdb) disassemble Dump of assembler code for function bar: => 0x0040068d <+0>: jr ra 0x0040068f <+2>: li v0,2 End of assembler dump. (gdb) finish Run till exit from #0 bar () at foobar.c:9 main () at foobar.c:24 24 } Value returned is $2 = 2 (gdb) continue Continuing. [Inferior 1 (process 14128) exited with code 02] (gdb) -- excellent! The change removes about 90 failures per MIPS16 multilib in mips-sde-elf testing too, results for MIPS16 are now similar to that for standard MIPS; microMIPS results are a bit worse because of host-I/O problems in QEMU I use for microMIPS only: standard MIPS: === gdb Summary === # of expected passes 14299 # of unexpected failures 187 # of expected failures 56 # of known failures 58 # of unresolved testcases 11 # of untested testcases 52 # of unsupported tests 174 MIPS16: === gdb Summary === # of expected passes 14298 # of unexpected failures 187 # of unexpected successes 2 # of expected failures 54 # of known failures 58 # of unresolved testcases 12 # of untested testcases 52 # of unsupported tests 174 microMIPS: === gdb Summary === # of expected passes 14149 # of unexpected failures 201 # of unexpected successes 2 # of expected failures 54 # of known failures 58 # of unresolved testcases 7 # of untested testcases 53 # of unsupported tests 175 I'll be rerunning mips-linux-gnu testing as the results I have got turned out questionable, so I need to double-check. I'll post them when I have them ready. This change regresses five cases too: gdb.cp/cp-relocate.exp: print caller $3 = {int (void)} 0x1 <caller()> (gdb) FAIL: gdb.cp/cp-relocate.exp: get address of caller and gdb.cp/expand-psymtabs-cxx.exp: p 'method(long)' $1 = {void (long)} 0x1 <method(long)> (gdb) FAIL: gdb.cp/expand-psymtabs-cxx.exp: before expand p method $2 = {void (long)} 0x1 <method(long)> (gdb) FAIL: gdb.cp/expand-psymtabs-cxx.exp: force expand p 'method(long)' $3 = {void (long)} 0x1 <method(long)> (gdb) FAIL: gdb.cp/expand-psymtabs-cxx.exp: after expand -- the two above are clearly test case bugs -- they load relocatables into GDB and shouldn't really require unlinked symbol references to resolve as NULLs, they probably just need to be fixed to accept output produced above too. gdb.dwarf2/dw2-empty-pc-range.exp: No symbol "realrange" in current context. (gdb) FAIL: gdb.dwarf2/dw2-empty-pc-range.exp: valid range after CU load -- this test uses this source: pc_start: .byte 0 pc_end: which flips the ISA bit on pc_end as a side effect. This bit is (IMO) correctly cleared in processing with my change making the instruction range between pc_start and pc_end empty; I'd be just tempted to change .byte into .word or suchlike, the exact size does not matter for this test and I think the space requested above should not be smaller than the (smallest) instruction size on any target. gdb.dwarf2/dw2-skip-prologue.exp: continue Continuing. Program received signal SIGBUS, Bus error. 0x004006fb in main () (gdb) FAIL: gdb.dwarf2/dw2-skip-prologue.exp: continue to breakpoint: func info break $bpnum Num Type Disp Enb Address What 2 breakpoint keep y <MULTIPLE> 2.1 y 0x00400671 in func at main.c:5 2.2 y 0x004006ac <func+59> (gdb) PASS: gdb.dwarf2/dw2-skip-prologue.exp: 2 locations found p v $1 = 8 (gdb) FAIL: gdb.dwarf2/dw2-skip-prologue.exp: no statement got executed -- this is plain bogus, the test strips symbol information for fund that is a MIPS16 function at 0x004006ad, the data required is permanently lost and the incorrect software breakpoint instruction encoding used caused execution to go into the weed (this only regresses on simulators that pretend to use hardware breakpoints and disregard the ISA bit; for any other target the test didn't work anyway). I don't know how to fix the test and retain functionality covered; this has to be rethought and reimplemented differently I would say. 5. Conclusion As noted at the beginning I am looking forward to all feedback to this proposal. I'm working on a small proper test case for our test suite to cover the function pointer issues shown above. Maciej 2012-05-14 Maciej W. Rozycki <macro@codesourcery.com> Maciej W. Rozycki <macro@mips.com> Pedro Alves <pedro@codesourcery.com> * gdbarch.sh (make_symbol_special): New architecture method. (adjust_dwarf2_addr, adjust_dwarf2_line): Likewise. (objfile, symbol): New declarations. * arch-utils.h (default_make_symbol_special): New prototype. (default_adjust_dwarf2_addr): Likewise. (default_adjust_dwarf2_line): Likewise. * arch-utils.c (default_make_symbol_special): New function. * dwarf2-frame.c (decode_frame_entry_1): Call gdbarch_adjust_dwarf2_addr. * dwarf2loc.c (dwarf2_find_location_expression): Likewise. * dwarf2read.c (read_func_scope): Call gdbarch_make_symbol_special. (dwarf2_ranges_read): Call gdbarch_adjust_dwarf2_addr. (read_attribute_value): Likewise. (dwarf_decode_lines_1): Call gdbarch_adjust_dwarf2_line. * mips-tdep.c (mips_elf_make_msymbol_special): Set the ISA bit in the symbol's address appropriately. (mips_make_symbol_special): New function. (mips_pc_is_mips): Set the ISA bit before symbol lookup. (mips_pc_is_mips16): Likewise. (mips_pc_is_micromips): Likewise. (mips_pc_isa): Likewise. (mips_adjust_dwarf2_addr): New function. (mips_adjust_dwarf2_line): Likewise. (mips_read_pc, mips_unwind_pc): Keep the ISA bit. (mips_addr_bits_remove): Likewise. (mips_skip_trampoline_code): Likewise. (mips_write_pc): Don't set the ISA bit. (mips_eabi_push_dummy_call): Likewise. (mips_o64_push_dummy_call): Likewise. (mips_gdbarch_init): Install mips_make_symbol_special, mips_adjust_dwarf2_addr and mips_adjust_dwarf2_line gdbarch handlers. * solib.c (gdb_bfd_lookup_symbol_from_symtab): Get target-specific symbol address adjustments. * gdbarch.h: Regenerate. * gdbarch.c: Regenerate. gdb-mips16-isa-bit.diff Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.c 2012-05-14 16:00:33.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.c 2012-05-14 16:02:02.235560558 +0100 @@ -358,9 +358,15 @@ mips_elf_make_msymbol_special (asymbol * return; if (ELF_ST_IS_MICROMIPS (elfsym->internal_elf_sym.st_other)) - MSYMBOL_TARGET_FLAG_2 (msym) = 1; + { + MSYMBOL_TARGET_FLAG_2 (msym) = 1; + SYMBOL_VALUE_ADDRESS (msym) |= 1; + } else if (ELF_ST_IS_MIPS16 (elfsym->internal_elf_sym.st_other)) - MSYMBOL_TARGET_FLAG_1 (msym) = 1; + { + MSYMBOL_TARGET_FLAG_1 (msym) = 1; + SYMBOL_VALUE_ADDRESS (msym) |= 1; + } } /* Return one iff MSYM refers to standard ISA code. */ @@ -387,6 +393,24 @@ msymbol_is_micromips (struct minimal_sym return MSYMBOL_TARGET_FLAG_2 (msym); } +/* Set the ISA bit in the main symbol too, complementing the corresponding + minimal symbol setting and reflecting the run-time value of the symbol. */ + +static void +mips_make_symbol_special (struct symbol *sym, struct objfile *objfile) +{ + if (SYMBOL_CLASS (sym) == LOC_BLOCK) + { + CORE_ADDR compact_block_start; + struct minimal_symbol *msym; + + compact_block_start = BLOCK_START (SYMBOL_BLOCK_VALUE (sym)) | 1; + msym = lookup_minimal_symbol_by_pc (compact_block_start); + if (msym && !msymbol_is_mips (msym)) + BLOCK_START (SYMBOL_BLOCK_VALUE (sym)) = compact_block_start; + } +} + /* XFER a value from the big/little/left end of the register. Depending on the size of the value it might occupy the entire register or just part of it. Make an allowance for this, aligning @@ -1123,7 +1147,7 @@ mips_pc_is_mips (CORE_ADDR memaddr) stored by elfread.c in the high bit of the info field. Use this to decide if the function is standard MIPS. Otherwise if bit 0 of the address is clear, then this is a standard MIPS function. */ - sym = lookup_minimal_symbol_by_pc (memaddr); + sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); if (sym) return msymbol_is_mips (sym); else @@ -1141,7 +1165,7 @@ mips_pc_is_mips16 (struct gdbarch *gdbar elfread.c in the high bit of the info field. Use this to decide if the function is MIPS16. Otherwise if bit 0 of the address is set, then ELF file flags will tell if this is a MIPS16 function. */ - sym = lookup_minimal_symbol_by_pc (memaddr); + sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); if (sym) return msymbol_is_mips16 (sym); else @@ -1160,7 +1184,7 @@ mips_pc_is_micromips (struct gdbarch *gd if the function is microMIPS. Otherwise if bit 0 of the address is set, then ELF file flags will tell if this is a microMIPS function. */ - sym = lookup_minimal_symbol_by_pc (memaddr); + sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); if (sym) return msymbol_is_micromips (sym); else @@ -1180,7 +1204,7 @@ mips_pc_isa (struct gdbarch *gdbarch, CO this to decide if the function is MIPS16 or microMIPS or normal MIPS. Otherwise if bit 0 of the address is set, then ELF file flags will tell if this is a MIPS16 or a microMIPS function. */ - sym = lookup_minimal_symbol_by_pc (memaddr); + sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); if (sym) { if (msymbol_is_micromips (sym)) @@ -1201,6 +1225,32 @@ mips_pc_isa (struct gdbarch *gdbarch, CO } } +/* Set the ISA bit correctly in the PC, used by DWARF-2 machinery. */ + +static CORE_ADDR +mips_adjust_dwarf2_addr (CORE_ADDR pc) +{ + pc = unmake_compact_addr (pc); + return mips_pc_is_mips (pc) ? pc : make_compact_addr (pc); +} + +/* Recalculate the line record requested so that the resulting PC has the + ISA bit set correctly, used by DWARF-2 machinery. */ + +static CORE_ADDR +mips_adjust_dwarf2_line (CORE_ADDR addr, int rel) +{ + static CORE_ADDR adj_pc; + static CORE_ADDR pc; + CORE_ADDR isa_pc; + + pc = rel ? pc + addr : addr; + isa_pc = mips_adjust_dwarf2_addr (pc); + addr = rel ? isa_pc - adj_pc : isa_pc; + adj_pc = isa_pc; + return addr; +} + /* Various MIPS16 thunk (aka stub or trampoline) names. */ static const char mips_str_mips16_call_stub[] = "__mips16_call_stub_"; @@ -1249,8 +1299,6 @@ mips_read_pc (struct regcache *regcache) ULONGEST pc; int regnum = mips_regnum (get_regcache_arch (regcache))->pc; regcache_cooked_read_signed (regcache, regnum, &pc); - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); return pc; } @@ -1261,8 +1309,6 @@ mips_unwind_pc (struct gdbarch *gdbarch, pc = frame_unwind_register_signed (next_frame, gdbarch_num_regs (gdbarch) + mips_regnum (gdbarch)->pc); - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); /* macro/2012-04-20: This hack skips over MIPS16 call thunks as intermediate frames. In this case we can get the caller's address from $ra, or if $ra contains an address within a thunk as well, then @@ -1272,15 +1318,9 @@ mips_unwind_pc (struct gdbarch *gdbarch, { pc = frame_unwind_register_signed (next_frame, gdbarch_num_regs (gdbarch) + MIPS_RA_REGNUM); - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); if (mips_in_frame_stub (pc)) - { - pc = frame_unwind_register_signed - (next_frame, gdbarch_num_regs (gdbarch) + MIPS_S2_REGNUM); - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); - } + pc = frame_unwind_register_signed + (next_frame, gdbarch_num_regs (gdbarch) + MIPS_S2_REGNUM); } return pc; } @@ -1312,10 +1352,7 @@ mips_write_pc (struct regcache *regcache { int regnum = mips_regnum (get_regcache_arch (regcache))->pc; - if (mips_pc_is_mips (pc)) - regcache_cooked_write_unsigned (regcache, regnum, pc); - else - regcache_cooked_write_unsigned (regcache, regnum, make_compact_addr (pc)); + regcache_cooked_write_unsigned (regcache, regnum, pc); } /* Fetch and return instruction from the specified location. Handle @@ -3603,10 +3640,6 @@ static CORE_ADDR mips_addr_bits_remove (struct gdbarch *gdbarch, CORE_ADDR addr) { struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch); - - if (is_compact_addr (addr)) - addr = unmake_compact_addr (addr); - if (mips_mask_address_p (tdep) && (((ULONGEST) addr) >> 32 == 0xffffffffUL)) /* This hack is a work-around for existing boards using PMON, the simulator, and any other 64-bit targets that doesn't have true @@ -4298,25 +4331,9 @@ mips_eabi_push_dummy_call (struct gdbarc "mips_eabi_push_dummy_call: %d len=%d type=%d", argnum + 1, len, (int) typecode); - /* Function pointer arguments to mips16 code need to be made into - mips16 pointers. */ - if (typecode == TYPE_CODE_PTR - && TYPE_CODE (TYPE_TARGET_TYPE (arg_type)) == TYPE_CODE_FUNC) - { - CORE_ADDR addr = extract_signed_integer (value_contents (arg), - len, byte_order); - if (mips_pc_is_mips (addr)) - val = value_contents (arg); - else - { - store_signed_integer (valbuf, len, byte_order, - make_compact_addr (addr)); - val = valbuf; - } - } /* The EABI passes structures that do not fit in a register by reference. */ - else if (len > regsize + if (len > regsize && (typecode == TYPE_CODE_STRUCT || typecode == TYPE_CODE_UNION)) { store_unsigned_integer (valbuf, regsize, byte_order, @@ -5686,7 +5703,6 @@ mips_o64_push_dummy_call (struct gdbarch for (argnum = 0; argnum < nargs; argnum++) { const gdb_byte *val; - gdb_byte valbuf[MAX_REGISTER_SIZE]; struct value *arg = args[argnum]; struct type *arg_type = check_typedef (value_type (arg)); int len = TYPE_LENGTH (arg_type); @@ -5699,21 +5715,6 @@ mips_o64_push_dummy_call (struct gdbarch val = value_contents (arg); - /* Function pointer arguments to mips16 code need to be made into - mips16 pointers. */ - if (typecode == TYPE_CODE_PTR - && TYPE_CODE (TYPE_TARGET_TYPE (arg_type)) == TYPE_CODE_FUNC) - { - CORE_ADDR addr = extract_signed_integer (value_contents (arg), - len, byte_order); - if (!mips_pc_is_mips (addr)) - { - store_signed_integer (valbuf, len, byte_order, - make_compact_addr (addr)); - val = valbuf; - } - } - /* Floating point arguments passed in registers have to be treated specially. On 32-bit architectures, doubles are passed in register pairs; the even FP register gets the @@ -7637,27 +7638,15 @@ mips_skip_trampoline_code (struct frame_ new_pc = mips_skip_mips16_trampoline_code (frame, pc); if (new_pc) - { - pc = new_pc; - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); - } + pc = new_pc; new_pc = find_solib_trampoline_target (frame, pc); if (new_pc) - { - pc = new_pc; - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); - } + pc = new_pc; new_pc = mips_skip_pic_trampoline_code (frame, pc); if (new_pc) - { - pc = new_pc; - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); - } + pc = new_pc; } while (pc != target_pc); @@ -8315,6 +8304,9 @@ mips_gdbarch_init (struct gdbarch_info i set_gdbarch_elf_make_msymbol_special (gdbarch, mips_elf_make_msymbol_special); + set_gdbarch_make_symbol_special (gdbarch, mips_make_symbol_special); + set_gdbarch_adjust_dwarf2_addr (gdbarch, mips_adjust_dwarf2_addr); + set_gdbarch_adjust_dwarf2_line (gdbarch, mips_adjust_dwarf2_line); regnum = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct mips_regnum); *regnum = mips_regnum; Index: gdb-fsf-trunk-quilt/gdb/arch-utils.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/arch-utils.c 2012-05-14 15:56:47.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/arch-utils.c 2012-05-14 16:02:02.205560708 +0100 @@ -31,6 +31,7 @@ #include "osabi.h" #include "target-descriptions.h" #include "objfiles.h" +#include "symtab.h" #include "version.h" @@ -178,6 +179,24 @@ default_coff_make_msymbol_special (int v return; } +void +default_make_symbol_special (struct symbol *sym, struct objfile *objfile) +{ + return; +} + +CORE_ADDR +default_adjust_dwarf2_addr (CORE_ADDR pc) +{ + return pc; +} + +CORE_ADDR +default_adjust_dwarf2_line (CORE_ADDR addr, int rel) +{ + return addr; +} + int cannot_register_not (struct gdbarch *gdbarch, int regnum) { Index: gdb-fsf-trunk-quilt/gdb/arch-utils.h =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/arch-utils.h 2012-05-14 15:56:47.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/arch-utils.h 2012-05-14 16:02:02.235560558 +0100 @@ -81,6 +81,18 @@ void default_elf_make_msymbol_special (a void default_coff_make_msymbol_special (int val, struct minimal_symbol *msym); +/* Do nothing version of make_symbol_special. */ + +void default_make_symbol_special (struct symbol *sym, struct objfile *objfile); + +/* Do nothing version of adjust_dwarf2_addr. */ + +CORE_ADDR default_adjust_dwarf2_addr (CORE_ADDR pc); + +/* Do nothing version of adjust_dwarf2_line. */ + +CORE_ADDR default_adjust_dwarf2_line (CORE_ADDR addr, int rel); + /* Version of cannot_fetch_register() / cannot_store_register() that always fails. */ Index: gdb-fsf-trunk-quilt/gdb/dwarf2read.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/dwarf2read.c 2012-05-14 15:56:47.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/dwarf2read.c 2012-05-14 16:02:02.245560384 +0100 @@ -7328,6 +7328,7 @@ static void read_func_scope (struct die_info *die, struct dwarf2_cu *cu) { struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); struct context_stack *new; CORE_ADDR lowpc; CORE_ADDR highpc; @@ -7478,6 +7479,8 @@ read_func_scope (struct die_info *die, s /* If we have address ranges, record them. */ dwarf2_record_block_ranges (die, block, baseaddr, cu); + gdbarch_make_symbol_special (gdbarch, new->name, objfile); + /* Attach template arguments to function. */ if (! VEC_empty (symbolp, template_args)) { @@ -7827,6 +7830,7 @@ dwarf2_ranges_read (unsigned offset, COR struct partial_symtab *ranges_pst) { struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); struct comp_unit_head *cu_header = &cu->header; bfd *obfd = objfile->obfd; unsigned int addr_size = cu_header->addr_size; @@ -7922,6 +7926,9 @@ dwarf2_ranges_read (unsigned offset, COR range_beginning += base; range_end += base; + range_beginning = gdbarch_adjust_dwarf2_addr (gdbarch, range_beginning); + range_end = gdbarch_adjust_dwarf2_addr (gdbarch, range_end); + if (ranges_pst != NULL) addrmap_set_empty (objfile->psymtabs_addrmap, range_beginning + baseaddr, @@ -11552,6 +11559,8 @@ read_attribute_value (const struct die_r gdb_byte *info_ptr) { struct dwarf2_cu *cu = reader->cu; + struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); bfd *abfd = reader->abfd; struct comp_unit_head *cu_header = &cu->header; unsigned int bytes_read; @@ -11570,6 +11579,7 @@ read_attribute_value (const struct die_r break; case DW_FORM_addr: DW_ADDR (attr) = read_address (abfd, info_ptr, cu, &bytes_read); + DW_ADDR (attr) = gdbarch_adjust_dwarf2_addr (gdbarch, DW_ADDR (attr)); info_ptr += bytes_read; break; case DW_FORM_block2: @@ -12767,7 +12777,7 @@ dwarf_decode_lines_1 (struct line_header while (line_ptr < line_end) { /* state machine registers */ - CORE_ADDR address = 0; + CORE_ADDR address = gdbarch_adjust_dwarf2_line (gdbarch, 0, 0); unsigned int file = 1; unsigned int line = 1; unsigned int column = 0; @@ -12805,11 +12815,14 @@ dwarf_decode_lines_1 (struct line_header if (op_code >= lh->opcode_base) { + CORE_ADDR addr_adj; + /* Special operand. */ adj_opcode = op_code - lh->opcode_base; - address += (((op_index + (adj_opcode / lh->line_range)) + addr_adj = (((op_index + (adj_opcode / lh->line_range)) / lh->maximum_ops_per_instruction) * lh->minimum_instruction_length); + address += gdbarch_adjust_dwarf2_line (gdbarch, addr_adj, 1); op_index = ((op_index + (adj_opcode / lh->line_range)) % lh->maximum_ops_per_instruction); line += lh->line_base + (adj_opcode % lh->line_range); @@ -12872,6 +12885,7 @@ dwarf_decode_lines_1 (struct line_header op_index = 0; line_ptr += bytes_read; address += baseaddr; + address = gdbarch_adjust_dwarf2_line (gdbarch, address, 0); break; case DW_LNE_define_file: { @@ -12938,10 +12952,12 @@ dwarf_decode_lines_1 (struct line_header { CORE_ADDR adjust = read_unsigned_leb128 (abfd, line_ptr, &bytes_read); + CORE_ADDR addr_adj; - address += (((op_index + adjust) + addr_adj = (((op_index + adjust) / lh->maximum_ops_per_instruction) * lh->minimum_instruction_length); + address += gdbarch_adjust_dwarf2_line (gdbarch, addr_adj, 1); op_index = ((op_index + adjust) % lh->maximum_ops_per_instruction); line_ptr += bytes_read; @@ -12994,18 +13010,25 @@ dwarf_decode_lines_1 (struct line_header case DW_LNS_const_add_pc: { CORE_ADDR adjust = (255 - lh->opcode_base) / lh->line_range; + CORE_ADDR addr_adj; - address += (((op_index + adjust) + addr_adj = (((op_index + adjust) / lh->maximum_ops_per_instruction) * lh->minimum_instruction_length); + address += gdbarch_adjust_dwarf2_line (gdbarch, addr_adj, 1); op_index = ((op_index + adjust) % lh->maximum_ops_per_instruction); } break; case DW_LNS_fixed_advance_pc: - address += read_2_bytes (abfd, line_ptr); - op_index = 0; - line_ptr += 2; + { + CORE_ADDR addr_adj; + + addr_adj = read_2_bytes (abfd, line_ptr); + address += gdbarch_adjust_dwarf2_line (gdbarch, addr_adj, 1); + op_index = 0; + line_ptr += 2; + } break; default: { Index: gdb-fsf-trunk-quilt/gdb/gdbarch.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.c 2012-05-14 15:56:47.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/gdbarch.c 2012-05-14 16:02:02.205560708 +0100 @@ -230,6 +230,9 @@ struct gdbarch gdbarch_in_function_epilogue_p_ftype *in_function_epilogue_p; gdbarch_elf_make_msymbol_special_ftype *elf_make_msymbol_special; gdbarch_coff_make_msymbol_special_ftype *coff_make_msymbol_special; + gdbarch_make_symbol_special_ftype *make_symbol_special; + gdbarch_adjust_dwarf2_addr_ftype *adjust_dwarf2_addr; + gdbarch_adjust_dwarf2_line_ftype *adjust_dwarf2_line; int cannot_step_breakpoint; int have_nonsteppable_watchpoint; gdbarch_address_class_type_flags_ftype *address_class_type_flags; @@ -398,6 +401,9 @@ struct gdbarch startup_gdbarch = generic_in_function_epilogue_p, /* in_function_epilogue_p */ 0, /* elf_make_msymbol_special */ 0, /* coff_make_msymbol_special */ + 0, /* make_symbol_special */ + 0, /* adjust_dwarf2_addr */ + 0, /* adjust_dwarf2_line */ 0, /* cannot_step_breakpoint */ 0, /* have_nonsteppable_watchpoint */ 0, /* address_class_type_flags */ @@ -532,6 +538,9 @@ gdbarch_alloc (const struct gdbarch_info gdbarch->in_function_epilogue_p = generic_in_function_epilogue_p; gdbarch->elf_make_msymbol_special = default_elf_make_msymbol_special; gdbarch->coff_make_msymbol_special = default_coff_make_msymbol_special; + gdbarch->make_symbol_special = default_make_symbol_special; + gdbarch->adjust_dwarf2_addr = default_adjust_dwarf2_addr; + gdbarch->adjust_dwarf2_line = default_adjust_dwarf2_line; gdbarch->register_reggroup_p = default_register_reggroup_p; gdbarch->displaced_step_hw_singlestep = default_displaced_step_hw_singlestep; gdbarch->displaced_step_fixup = NULL; @@ -699,6 +708,9 @@ verify_gdbarch (struct gdbarch *gdbarch) /* Skip verify of in_function_epilogue_p, invalid_p == 0 */ /* Skip verify of elf_make_msymbol_special, invalid_p == 0 */ /* Skip verify of coff_make_msymbol_special, invalid_p == 0 */ + /* Skip verify of make_symbol_special, invalid_p == 0 */ + /* Skip verify of adjust_dwarf2_addr, invalid_p == 0 */ + /* Skip verify of adjust_dwarf2_line, invalid_p == 0 */ /* Skip verify of cannot_step_breakpoint, invalid_p == 0 */ /* Skip verify of have_nonsteppable_watchpoint, invalid_p == 0 */ /* Skip verify of address_class_type_flags, has predicate. */ @@ -811,6 +823,12 @@ gdbarch_dump (struct gdbarch *gdbarch, s "gdbarch_dump: adjust_breakpoint_address = <%s>\n", host_address_to_string (gdbarch->adjust_breakpoint_address)); fprintf_unfiltered (file, + "gdbarch_dump: adjust_dwarf2_addr = <%s>\n", + host_address_to_string (gdbarch->adjust_dwarf2_addr)); + fprintf_unfiltered (file, + "gdbarch_dump: adjust_dwarf2_line = <%s>\n", + host_address_to_string (gdbarch->adjust_dwarf2_line)); + fprintf_unfiltered (file, "gdbarch_dump: auto_charset = <%s>\n", host_address_to_string (gdbarch->auto_charset)); fprintf_unfiltered (file, @@ -1087,6 +1105,9 @@ gdbarch_dump (struct gdbarch *gdbarch, s "gdbarch_dump: make_corefile_notes = <%s>\n", host_address_to_string (gdbarch->make_corefile_notes)); fprintf_unfiltered (file, + "gdbarch_dump: make_symbol_special = <%s>\n", + host_address_to_string (gdbarch->make_symbol_special)); + fprintf_unfiltered (file, "gdbarch_dump: gdbarch_max_insn_length_p() = %d\n", gdbarch_max_insn_length_p (gdbarch)); fprintf_unfiltered (file, @@ -3141,6 +3162,57 @@ set_gdbarch_coff_make_msymbol_special (s gdbarch->coff_make_msymbol_special = coff_make_msymbol_special; } +void +gdbarch_make_symbol_special (struct gdbarch *gdbarch, struct symbol *sym, struct objfile *objfile) +{ + gdb_assert (gdbarch != NULL); + gdb_assert (gdbarch->make_symbol_special != NULL); + if (gdbarch_debug >= 2) + fprintf_unfiltered (gdb_stdlog, "gdbarch_make_symbol_special called\n"); + gdbarch->make_symbol_special (sym, objfile); +} + +void +set_gdbarch_make_symbol_special (struct gdbarch *gdbarch, + gdbarch_make_symbol_special_ftype make_symbol_special) +{ + gdbarch->make_symbol_special = make_symbol_special; +} + +CORE_ADDR +gdbarch_adjust_dwarf2_addr (struct gdbarch *gdbarch, CORE_ADDR pc) +{ + gdb_assert (gdbarch != NULL); + gdb_assert (gdbarch->adjust_dwarf2_addr != NULL); + if (gdbarch_debug >= 2) + fprintf_unfiltered (gdb_stdlog, "gdbarch_adjust_dwarf2_addr called\n"); + return gdbarch->adjust_dwarf2_addr (pc); +} + +void +set_gdbarch_adjust_dwarf2_addr (struct gdbarch *gdbarch, + gdbarch_adjust_dwarf2_addr_ftype adjust_dwarf2_addr) +{ + gdbarch->adjust_dwarf2_addr = adjust_dwarf2_addr; +} + +CORE_ADDR +gdbarch_adjust_dwarf2_line (struct gdbarch *gdbarch, CORE_ADDR addr, int rel) +{ + gdb_assert (gdbarch != NULL); + gdb_assert (gdbarch->adjust_dwarf2_line != NULL); + if (gdbarch_debug >= 2) + fprintf_unfiltered (gdb_stdlog, "gdbarch_adjust_dwarf2_line called\n"); + return gdbarch->adjust_dwarf2_line (addr, rel); +} + +void +set_gdbarch_adjust_dwarf2_line (struct gdbarch *gdbarch, + gdbarch_adjust_dwarf2_line_ftype adjust_dwarf2_line) +{ + gdbarch->adjust_dwarf2_line = adjust_dwarf2_line; +} + int gdbarch_cannot_step_breakpoint (struct gdbarch *gdbarch) { Index: gdb-fsf-trunk-quilt/gdb/gdbarch.h =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.h 2012-05-14 15:56:47.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/gdbarch.h 2012-05-14 16:02:02.205560708 +0100 @@ -50,6 +50,8 @@ struct target_ops; struct obstack; struct bp_target_info; struct target_desc; +struct objfile; +struct symbol; struct displaced_step_closure; struct core_regset_section; struct syscall; @@ -658,6 +660,18 @@ typedef void (gdbarch_coff_make_msymbol_ extern void gdbarch_coff_make_msymbol_special (struct gdbarch *gdbarch, int val, struct minimal_symbol *msym); extern void set_gdbarch_coff_make_msymbol_special (struct gdbarch *gdbarch, gdbarch_coff_make_msymbol_special_ftype *coff_make_msymbol_special); +typedef void (gdbarch_make_symbol_special_ftype) (struct symbol *sym, struct objfile *objfile); +extern void gdbarch_make_symbol_special (struct gdbarch *gdbarch, struct symbol *sym, struct objfile *objfile); +extern void set_gdbarch_make_symbol_special (struct gdbarch *gdbarch, gdbarch_make_symbol_special_ftype *make_symbol_special); + +typedef CORE_ADDR (gdbarch_adjust_dwarf2_addr_ftype) (CORE_ADDR pc); +extern CORE_ADDR gdbarch_adjust_dwarf2_addr (struct gdbarch *gdbarch, CORE_ADDR pc); +extern void set_gdbarch_adjust_dwarf2_addr (struct gdbarch *gdbarch, gdbarch_adjust_dwarf2_addr_ftype *adjust_dwarf2_addr); + +typedef CORE_ADDR (gdbarch_adjust_dwarf2_line_ftype) (CORE_ADDR addr, int rel); +extern CORE_ADDR gdbarch_adjust_dwarf2_line (struct gdbarch *gdbarch, CORE_ADDR addr, int rel); +extern void set_gdbarch_adjust_dwarf2_line (struct gdbarch *gdbarch, gdbarch_adjust_dwarf2_line_ftype *adjust_dwarf2_line); + extern int gdbarch_cannot_step_breakpoint (struct gdbarch *gdbarch); extern void set_gdbarch_cannot_step_breakpoint (struct gdbarch *gdbarch, int cannot_step_breakpoint); Index: gdb-fsf-trunk-quilt/gdb/gdbarch.sh =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.sh 2012-05-14 15:56:47.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/gdbarch.sh 2012-05-14 16:02:02.215560349 +0100 @@ -614,6 +614,9 @@ m:int:in_solib_return_trampoline:CORE_AD m:int:in_function_epilogue_p:CORE_ADDR addr:addr:0:generic_in_function_epilogue_p::0 f:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym::default_elf_make_msymbol_special::0 f:void:coff_make_msymbol_special:int val, struct minimal_symbol *msym:val, msym::default_coff_make_msymbol_special::0 +f:void:make_symbol_special:struct symbol *sym, struct objfile *objfile:sym, objfile::default_make_symbol_special::0 +f:CORE_ADDR:adjust_dwarf2_addr:CORE_ADDR pc:pc::default_adjust_dwarf2_addr::0 +f:CORE_ADDR:adjust_dwarf2_line:CORE_ADDR addr, int rel:addr, rel::default_adjust_dwarf2_line::0 v:int:cannot_step_breakpoint:::0:0::0 v:int:have_nonsteppable_watchpoint:::0:0::0 F:int:address_class_type_flags:int byte_size, int dwarf2_addr_class:byte_size, dwarf2_addr_class @@ -1044,6 +1047,8 @@ struct target_ops; struct obstack; struct bp_target_info; struct target_desc; +struct objfile; +struct symbol; struct displaced_step_closure; struct core_regset_section; struct syscall; Index: gdb-fsf-trunk-quilt/gdb/dwarf2-frame.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/dwarf2-frame.c 2012-05-14 15:56:43.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/dwarf2-frame.c 2012-05-14 16:02:02.255559958 +0100 @@ -2095,6 +2095,7 @@ decode_frame_entry_1 (struct comp_unit * { /* This is a FDE. */ struct dwarf2_fde *fde; + CORE_ADDR addr; /* Check that an FDE was expected. */ if ((entry_type & EH_FDE_TYPE_ID) == 0) @@ -2128,14 +2129,16 @@ decode_frame_entry_1 (struct comp_unit * gdb_assert (fde->cie != NULL); - fde->initial_location = - read_encoded_value (unit, fde->cie->encoding, fde->cie->ptr_size, - buf, &bytes_read, 0); + addr = read_encoded_value (unit, fde->cie->encoding, fde->cie->ptr_size, + buf, &bytes_read, 0); + fde->initial_location = gdbarch_adjust_dwarf2_addr (gdbarch, addr); buf += bytes_read; fde->address_range = read_encoded_value (unit, fde->cie->encoding & 0x0f, fde->cie->ptr_size, buf, &bytes_read, 0); + addr = gdbarch_adjust_dwarf2_addr (gdbarch, addr + fde->address_range); + fde->address_range = addr - fde->initial_location; buf += bytes_read; /* A 'z' augmentation in the CIE implies the presence of an Index: gdb-fsf-trunk-quilt/gdb/dwarf2loc.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/dwarf2loc.c 2012-05-14 15:56:43.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/dwarf2loc.c 2012-05-14 16:02:02.255559958 +0100 @@ -122,6 +122,9 @@ dwarf2_find_location_expression (struct low += base_address; high += base_address; + low = gdbarch_adjust_dwarf2_addr (gdbarch, low); + high = gdbarch_adjust_dwarf2_addr (gdbarch, high); + length = extract_unsigned_integer (loc_ptr, 2, byte_order); loc_ptr += 2; Index: gdb-fsf-trunk-quilt/gdb/solib.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/solib.c 2012-05-14 15:56:45.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/solib.c 2012-05-14 16:01:34.645558637 +0100 @@ -1384,8 +1384,27 @@ gdb_bfd_lookup_symbol_from_symtab (bfd * if (match_sym (sym, data)) { + symaddr = sym->value; + + /* macro/2012-04-20: Some ELF targets fiddle with addresses + of symbols they consider special. They use minimal symbols + to do that and this is needed for correct breakpoint + placement, but we do not have full data here to build a + complete minimal symbol, so just set the address and let the + targets cope with that. */ + if (bfd_get_flavour (abfd) == bfd_target_elf_flavour) + { + struct minimal_symbol msym; + + memset (&msym, 0, sizeof (msym)); + SYMBOL_VALUE_ADDRESS (&msym) = symaddr; + gdbarch_elf_make_msymbol_special (target_gdbarch, + sym, &msym); + symaddr = SYMBOL_VALUE_ADDRESS (&msym); + } + /* BFD symbols are section relative. */ - symaddr = sym->value + sym->section->vma; + symaddr += sym->section->vma; break; } } ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [RFD+PATCH] ISA bit treatment on the MIPS platform 2012-05-14 15:37 [RFD+PATCH] ISA bit treatment on the MIPS platform Maciej W. Rozycki @ 2012-05-18 22:01 ` Maciej W. Rozycki 2012-06-11 18:21 ` Joel Brobecker 1 sibling, 0 replies; 27+ messages in thread From: Maciej W. Rozycki @ 2012-05-18 22:01 UTC (permalink / raw) To: gdb-patches; +Cc: Rich Fuhler, Richard Sandiford, binutils, gcc Hello again, As promised I am providing Linux results now. With the recent fixes to gdbserver and mips-linux-tdep the number of Linux progressions matches, as far as the order of magnitude is concerned, that of bare-iron ones and is around 30 for MIPS16 and around 100 for microMIPS multilibs. The exact new numbers are as follows (as usually, these are indicative only because of intermittent failures, mainly in the thread tests): standard MIPS/o32: === gdb Summary === # of expected passes 15667 # of unexpected failures 77 # of unexpected successes 2 # of expected failures 45 # of known failures 62 # of unresolved testcases 1 # of untested testcases 45 # of unsupported tests 161 standard MIPS/n64: === gdb Summary === # of expected passes 15407 # of unexpected failures 158 # of unexpected successes 2 # of expected failures 47 # of known failures 62 # of unresolved testcases 5 # of untested testcases 46 # of unsupported tests 162 MIPS16/o32: === gdb Summary === # of expected passes 15569 # of unexpected failures 145 # of unexpected successes 2 # of expected failures 45 # of known failures 62 # of unresolved testcases 3 # of untested testcases 46 # of unsupported tests 161 microMIPS/o32: === gdb Summary === # of expected passes 15480 # of unexpected failures 97 # of unexpected successes 2 # of expected failures 45 # of known failures 62 # of unresolved testcases 1 # of untested testcases 46 # of unsupported tests 162 Also as promised here's a test case to explicitly cover this issue, similar in spirit to the examples used in the problem report. It fails for MIPS16 multilibs with our current code (and for microMIPS ones with my microMIPS support patch applied) and passes with my proposed ISA bit changes included. And last but not least, I've included an update to cover Linux microMIPS signal trampolines, complementing the change just sent as a followup to the original microMIPS support patch. 2012-05-18 Maciej W. Rozycki <macro@codesourcery.com> gdb/testsuite/ * gdb.base/func-ptrs.c: New file. * gdb.base/func-ptrs.exp: New file. 2012-05-18 Maciej W. Rozycki <macro@codesourcery.com> gdb/ * mips-tdep.h (mips_unmake_compact_addr): New prototype. * mips-tdep.c (unmake_compact_addr): Rename to... (mips_unmake_compact_addr): ... this. Make external. (make_compact_addr): Rename to... (mips_make_compact_addr): ... this. (mips_pc_is_mips): Update accordingly. (mips_pc_is_mips16): Likewise. (mips_pc_is_micromips): Likewise. (mips_pc_isa): Likewise. (mips_adjust_dwarf2_addr): Likewise. (mips_fetch_instruction): Likewise. (mips_breakpoint_from_pc): Likewise. (mips_remote_breakpoint_from_pc): Likewise. (mips_adjust_breakpoint_address): Likewise. * mips-linux-tdep.c (micromips_linux_sigframe_validate): Strip off the ISA bit from the address returned. Maciej gdb-test-func-ptrs.diff Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.base/func-ptrs.c =================================================================== --- /dev/null 1970-01-01 00:00:00.000000000 +0000 +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.base/func-ptrs.c 2012-05-14 22:46:46.185606207 +0100 @@ -0,0 +1,50 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright 2012 Free Software Foundation, Inc. + + This program 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 of the License, or + (at your option) any later version. + + This program 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. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ + +void +sentinel (void) +{ + return; +} + +int +incr (int i) +{ + sentinel (); + return i + 1; +} + +int +decr (int i) +{ + sentinel (); + return i - 1; +} + +int (*calc) (int) = incr; + +int +main (void) +{ + int i = -1; + + i = calc (i); + i = calc (i); + i = calc (i); + + return i; +} Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.base/func-ptrs.exp =================================================================== --- /dev/null 1970-01-01 00:00:00.000000000 +0000 +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.base/func-ptrs.exp 2012-05-14 23:17:51.115558348 +0100 @@ -0,0 +1,95 @@ +# Copyright 2012 Free Software Foundation, Inc. + +# This program 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 of the License, or +# (at your option) any later version. +# +# This program 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. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. + +set testname func-ptrs +set srcfile ${testname}.c +if { [prepare_for_testing ${testname}.exp ${testname} ${srcfile}] } { + return -1 +} + +if { ![runto_main] } { + untested ${testname}.exp + return -1 +} + + +# First set our breakpoints. + +set breakpoint_re \ + "Breakpoint $decimal at $hex: file .*${srcfile}, line $decimal\\." +gdb_test "break sentinel if calc == decr" \ + "${breakpoint_re}" \ + "breakpoint at sentinel" +gdb_test "break incr" \ + "${breakpoint_re}" \ + "breakpoint at incr" +gdb_test "break decr" \ + "${breakpoint_re}" \ + "breakpoint at decr" + + +# Check if we run through to the breakpoint in incr. + +gdb_test "continue" \ + "Breakpoint $decimal, incr \\(i=-1\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*" \ + "continue to incr, first time" + + +# Go back up, make sure the return value is 0. + +gdb_test "finish" \ + "Run till exit from #0 +incr \\(i=-1\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+($hex in )?main \\(\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*Value returned is \\$$decimal = 0" \ + "go back to main from incr, first time" + + +# Redirect calc and see if we run to the breakpoint in decr instead. + +gdb_test_no_output "set calc = decr" "set calc to decr" +gdb_test "continue" \ + "Breakpoint $decimal, decr \\(i=0\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*" \ + "continue to decr" + + +# Go back up, check if we stop in sentinel instead. + +gdb_test "finish" \ + "Run till exit from #0 +decr \\(i=0\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+Breakpoint $decimal, sentinel \\(\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*" \ + "stop in sentinel" + + +# Go back all the way up to main, make sure the return value is -1. + +gdb_test_no_output "up-silently" "move up to decr" +gdb_test "finish" \ + "Run till exit from #1 +($hex in )?decr \\(i=0\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+($hex in )?main \\(\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*Value returned is \\$$decimal = -1" \ + "go back to main from decr" + + +# Reset calc and see if we run to the breakpoint in incr again. + +gdb_test_no_output "set calc = incr" "set calc to incr" +gdb_test "continue" \ + "Breakpoint $decimal, incr \\(i=-1\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*" \ + "continue to incr, second time" + + +# Go back up again, make sure the return value is 0. + +gdb_test "finish" \ + "Run till exit from #0 +incr \\(i=-1\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+($hex in )?main \\(\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*Value returned is \\$$decimal = 0" \ + "go back to main from incr, second time" + + +# All done! gdb-mips16-isa-bit-sigtramp.diff Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.c 2012-05-18 20:56:25.545667440 +0100 +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.c 2012-05-18 20:46:31.615619743 +0100 @@ -321,8 +321,8 @@ is_micromips_addr (struct gdbarch *gdbar /* Strip the ISA (compression) bit off from ADDR. */ -static CORE_ADDR -unmake_compact_addr (CORE_ADDR addr) +CORE_ADDR +mips_unmake_compact_addr (CORE_ADDR addr) { return ((addr) & ~(CORE_ADDR) 1); } @@ -330,7 +330,7 @@ unmake_compact_addr (CORE_ADDR addr) /* Add the ISA (compression) bit to ADDR. */ static CORE_ADDR -make_compact_addr (CORE_ADDR addr) +mips_make_compact_addr (CORE_ADDR addr) { return ((addr) | (CORE_ADDR) 1); } @@ -1147,7 +1147,7 @@ mips_pc_is_mips (CORE_ADDR memaddr) stored by elfread.c in the high bit of the info field. Use this to decide if the function is standard MIPS. Otherwise if bit 0 of the address is clear, then this is a standard MIPS function. */ - sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); + sym = lookup_minimal_symbol_by_pc (mips_make_compact_addr (memaddr)); if (sym) return msymbol_is_mips (sym); else @@ -1165,7 +1165,7 @@ mips_pc_is_mips16 (struct gdbarch *gdbar elfread.c in the high bit of the info field. Use this to decide if the function is MIPS16. Otherwise if bit 0 of the address is set, then ELF file flags will tell if this is a MIPS16 function. */ - sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); + sym = lookup_minimal_symbol_by_pc (mips_make_compact_addr (memaddr)); if (sym) return msymbol_is_mips16 (sym); else @@ -1184,7 +1184,7 @@ mips_pc_is_micromips (struct gdbarch *gd if the function is microMIPS. Otherwise if bit 0 of the address is set, then ELF file flags will tell if this is a microMIPS function. */ - sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); + sym = lookup_minimal_symbol_by_pc (mips_make_compact_addr (memaddr)); if (sym) return msymbol_is_micromips (sym); else @@ -1204,7 +1204,7 @@ mips_pc_isa (struct gdbarch *gdbarch, CO this to decide if the function is MIPS16 or microMIPS or normal MIPS. Otherwise if bit 0 of the address is set, then ELF file flags will tell if this is a MIPS16 or a microMIPS function. */ - sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); + sym = lookup_minimal_symbol_by_pc (mips_make_compact_addr (memaddr)); if (sym) { if (msymbol_is_micromips (sym)) @@ -1230,8 +1230,8 @@ mips_pc_isa (struct gdbarch *gdbarch, CO static CORE_ADDR mips_adjust_dwarf2_addr (CORE_ADDR pc) { - pc = unmake_compact_addr (pc); - return mips_pc_is_mips (pc) ? pc : make_compact_addr (pc); + pc = mips_unmake_compact_addr (pc); + return mips_pc_is_mips (pc) ? pc : mips_make_compact_addr (pc); } /* Recalculate the line record requested so that the resulting PC has the @@ -1374,7 +1374,7 @@ mips_fetch_instruction (struct gdbarch * case ISA_MICROMIPS: case ISA_MIPS16: instlen = MIPS_INSN16_SIZE; - addr = unmake_compact_addr (addr); + addr = mips_unmake_compact_addr (addr); break; case ISA_MIPS: instlen = MIPS_INSN32_SIZE; @@ -6798,7 +6798,7 @@ mips_breakpoint_from_pc (struct gdbarch if (mips_pc_is_mips16 (gdbarch, pc)) { static gdb_byte mips16_big_breakpoint[] = { 0xe8, 0xa5 }; - *pcptr = unmake_compact_addr (pc); + *pcptr = mips_unmake_compact_addr (pc); *lenptr = sizeof (mips16_big_breakpoint); return mips16_big_breakpoint; } @@ -6813,7 +6813,7 @@ mips_breakpoint_from_pc (struct gdbarch insn = mips_fetch_instruction (gdbarch, ISA_MICROMIPS, pc, &status); size = status ? 2 : mips_insn_size (ISA_MICROMIPS, insn) == 2 ? 2 : 4; - *pcptr = unmake_compact_addr (pc); + *pcptr = mips_unmake_compact_addr (pc); *lenptr = size; return (size == 2) ? micromips16_big_breakpoint : micromips32_big_breakpoint; @@ -6849,7 +6849,7 @@ mips_breakpoint_from_pc (struct gdbarch if (mips_pc_is_mips16 (gdbarch, pc)) { static gdb_byte mips16_little_breakpoint[] = { 0xa5, 0xe8 }; - *pcptr = unmake_compact_addr (pc); + *pcptr = mips_unmake_compact_addr (pc); *lenptr = sizeof (mips16_little_breakpoint); return mips16_little_breakpoint; } @@ -6864,7 +6864,7 @@ mips_breakpoint_from_pc (struct gdbarch insn = mips_fetch_instruction (gdbarch, ISA_MICROMIPS, pc, &status); size = status ? 2 : mips_insn_size (ISA_MICROMIPS, insn) == 2 ? 2 : 4; - *pcptr = unmake_compact_addr (pc); + *pcptr = mips_unmake_compact_addr (pc); *lenptr = size; return (size == 2) ? micromips16_little_breakpoint : micromips32_little_breakpoint; @@ -6908,7 +6908,7 @@ mips_remote_breakpoint_from_pc (struct g if (mips_pc_is_mips16 (gdbarch, pc)) { - *pcptr = unmake_compact_addr (pc); + *pcptr = mips_unmake_compact_addr (pc); *kindptr = 2; } else if (mips_pc_is_micromips (gdbarch, pc)) @@ -6919,7 +6919,7 @@ mips_remote_breakpoint_from_pc (struct g insn = mips_fetch_instruction (gdbarch, ISA_MICROMIPS, pc, &status); size = status ? 2 : mips_insn_size (ISA_MICROMIPS, insn) == 2 ? 2 : 4; - *pcptr = unmake_compact_addr (pc); + *pcptr = mips_unmake_compact_addr (pc); *kindptr = size | 1; } else @@ -7169,7 +7169,7 @@ mips_adjust_breakpoint_address (struct g CORE_ADDR addr, jmpaddr; int i; - boundary = unmake_compact_addr (boundary); + boundary = mips_unmake_compact_addr (boundary); /* The only MIPS16 instructions with delay slots are JAL, JALX, JALR and JR. An absolute JAL/JALX is always 4 bytes long, @@ -7187,7 +7187,7 @@ mips_adjust_breakpoint_address (struct g addr = bpaddr; for (i = 1; i < 4; i++) { - if (unmake_compact_addr (addr) == boundary) + if (mips_unmake_compact_addr (addr) == boundary) break; addr -= MIPS_INSN16_SIZE; if (i == 1 && instruction_has_delay_slot (gdbarch, addr, 0)) Index: gdb-fsf-trunk-quilt/gdb/mips-linux-tdep.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/mips-linux-tdep.c 2012-05-18 20:47:17.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/mips-linux-tdep.c 2012-05-18 20:56:18.315621662 +0100 @@ -1237,7 +1237,10 @@ micromips_linux_sigframe_validate (const struct frame_info *this_frame, CORE_ADDR pc) { - return mips_pc_is_micromips (get_frame_arch (this_frame), pc) ? pc : 0; + if (mips_pc_is_micromips (get_frame_arch (this_frame), pc)) + return mips_unmake_compact_addr (pc); + else + return 0; } /* Implement the "write_pc" gdbarch method. */ Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.h =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.h 2012-05-18 20:40:32.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.h 2012-05-18 20:43:04.005621216 +0100 @@ -161,6 +161,9 @@ enum /* Single step based on where the current instruction will take us. */ extern int mips_software_single_step (struct frame_info *frame); +/* Strip the ISA (compression) bit off from ADDR. */ +extern CORE_ADDR mips_unmake_compact_addr (CORE_ADDR addr); + /* Tell if the program counter value in MEMADDR is in a standard MIPS function. */ extern int mips_pc_is_mips (bfd_vma memaddr); ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [RFD+PATCH] ISA bit treatment on the MIPS platform 2012-05-14 15:37 [RFD+PATCH] ISA bit treatment on the MIPS platform Maciej W. Rozycki 2012-05-18 22:01 ` Maciej W. Rozycki @ 2012-06-11 18:21 ` Joel Brobecker 2012-06-12 14:05 ` Pedro Alves 2014-10-06 0:42 ` [PATCH v2 1/2] " Maciej W. Rozycki 1 sibling, 2 replies; 27+ messages in thread From: Joel Brobecker @ 2012-06-11 18:21 UTC (permalink / raw) To: Maciej W. Rozycki Cc: gdb-patches, Rich Fuhler, Richard Sandiford, binutils, gcc > I propose therefore to accept the existing inconsistencies and deal > with them entirely within GDB. I have figured out that the ISA bit > lost in various places can still be recovered as long as we have > symbol information -- that'll have the st_other attribute correctly > set to one of standard MIPS/MIPS16/microMIPS. That seems reasonable to me. > 2012-05-14 Maciej W. Rozycki <macro@codesourcery.com> > Maciej W. Rozycki <macro@mips.com> > Pedro Alves <pedro@codesourcery.com> > > * gdbarch.sh (make_symbol_special): New architecture method. > (adjust_dwarf2_addr, adjust_dwarf2_line): Likewise. > (objfile, symbol): New declarations. > * arch-utils.h (default_make_symbol_special): New prototype. > (default_adjust_dwarf2_addr): Likewise. > (default_adjust_dwarf2_line): Likewise. > * arch-utils.c (default_make_symbol_special): New function. > * dwarf2-frame.c (decode_frame_entry_1): Call > gdbarch_adjust_dwarf2_addr. > * dwarf2loc.c (dwarf2_find_location_expression): Likewise. > * dwarf2read.c (read_func_scope): Call > gdbarch_make_symbol_special. > (dwarf2_ranges_read): Call gdbarch_adjust_dwarf2_addr. > (read_attribute_value): Likewise. > (dwarf_decode_lines_1): Call gdbarch_adjust_dwarf2_line. > * mips-tdep.c (mips_elf_make_msymbol_special): Set the ISA bit > in the symbol's address appropriately. > (mips_make_symbol_special): New function. > (mips_pc_is_mips): Set the ISA bit before symbol lookup. > (mips_pc_is_mips16): Likewise. > (mips_pc_is_micromips): Likewise. > (mips_pc_isa): Likewise. > (mips_adjust_dwarf2_addr): New function. > (mips_adjust_dwarf2_line): Likewise. > (mips_read_pc, mips_unwind_pc): Keep the ISA bit. > (mips_addr_bits_remove): Likewise. > (mips_skip_trampoline_code): Likewise. > (mips_write_pc): Don't set the ISA bit. > (mips_eabi_push_dummy_call): Likewise. > (mips_o64_push_dummy_call): Likewise. > (mips_gdbarch_init): Install mips_make_symbol_special, > mips_adjust_dwarf2_addr and mips_adjust_dwarf2_line gdbarch > handlers. > * solib.c (gdb_bfd_lookup_symbol_from_symtab): Get > target-specific symbol address adjustments. > * gdbarch.h: Regenerate. > * gdbarch.c: Regenerate. From the ChangeLog entry, it seems like Pedro was involved in the making of that patch, so perhaps he could be a good reviewer? There are also changes in dwarf2*, so perhaps people such as Doug and Tom could also be involved.... Just some minor comments below: > gdb-mips16-isa-bit.diff > Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.c > =================================================================== > --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.c 2012-05-14 16:00:33.000000000 +0100 > +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.c 2012-05-14 16:02:02.235560558 +0100 > @@ -358,9 +358,15 @@ mips_elf_make_msymbol_special (asymbol * > return; > > if (ELF_ST_IS_MICROMIPS (elfsym->internal_elf_sym.st_other)) > - MSYMBOL_TARGET_FLAG_2 (msym) = 1; > + { > + MSYMBOL_TARGET_FLAG_2 (msym) = 1; > + SYMBOL_VALUE_ADDRESS (msym) |= 1; > + } > else if (ELF_ST_IS_MIPS16 (elfsym->internal_elf_sym.st_other)) > - MSYMBOL_TARGET_FLAG_1 (msym) = 1; > + { > + MSYMBOL_TARGET_FLAG_1 (msym) = 1; > + SYMBOL_VALUE_ADDRESS (msym) |= 1; > + } This remark is not to be considered as part of this patch's review, since this is already an established practice, but I think we could do better than using magic numbers for those flags. I understand they have to be that way in the common code, but perhaps mips-tdep could define aliases? Something like MSYMBOL_MIPS_TARGET_FLAG_BLAH? > Index: gdb-fsf-trunk-quilt/gdb/arch-utils.c [...] > +void > +default_make_symbol_special (struct symbol *sym, struct objfile *objfile) Can you put a comment at the start of all these new functions to say that the documentation is in the header file? For instance, I think we typically say: /* See arch-utils.h. */ This is to help someone reviewing the code to know that there is in fact documentation. > +/* Do nothing version of make_symbol_special. */ > + > +void default_make_symbol_special (struct symbol *sym, struct objfile *objfile); > + > +/* Do nothing version of adjust_dwarf2_addr. */ > + > +CORE_ADDR default_adjust_dwarf2_addr (CORE_ADDR pc); > + > +/* Do nothing version of adjust_dwarf2_line. */ > + > +CORE_ADDR default_adjust_dwarf2_line (CORE_ADDR addr, int rel); I would definitely mention gdbarch in the function comment. Prior art uses: /* Default implementation of gdbarch_displaced_hw_singlestep. */ Should we remain consistent with that? It can be a mix of both, saying what method it implements, and then what it actually does, Eg: /* Default version of gdbarch_make_symbol_special (does nothing). */ > Index: gdb-fsf-trunk-quilt/gdb/solib.c > =================================================================== > --- gdb-fsf-trunk-quilt.orig/gdb/solib.c 2012-05-14 15:56:45.000000000 +0100 > +++ gdb-fsf-trunk-quilt/gdb/solib.c 2012-05-14 16:01:34.645558637 +0100 > @@ -1384,8 +1384,27 @@ gdb_bfd_lookup_symbol_from_symtab (bfd * > > if (match_sym (sym, data)) > { > + symaddr = sym->value; > + > + /* macro/2012-04-20: Some ELF targets fiddle with addresses > + of symbols they consider special. They use minimal symbols > + to do that and this is needed for correct breakpoint > + placement, but we do not have full data here to build a > + complete minimal symbol, so just set the address and let the > + targets cope with that. */ > + if (bfd_get_flavour (abfd) == bfd_target_elf_flavour) > + { > + struct minimal_symbol msym; > + > + memset (&msym, 0, sizeof (msym)); > + SYMBOL_VALUE_ADDRESS (&msym) = symaddr; > + gdbarch_elf_make_msymbol_special (target_gdbarch, > + sym, &msym); > + symaddr = SYMBOL_VALUE_ADDRESS (&msym); > + } Is there a way we could avoid that block if gdbarch_elf_make_msymbol_special points to the default implementation? It just seems sad to be doing all this work of creating a temporary symbol for every solib symbol on non-mips targets... One possible way would have been to provide no default value for gdbarch_elf_make_msymbol_special, but then it makes the calling of that method a little trickier. That could be worked around by changing the default value of gdbarch_elf_make_msymbol_special to NULL, and then having an extra wrapper function that call the gdbarch method if not NULL or else does nothing. -- Joel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [RFD+PATCH] ISA bit treatment on the MIPS platform 2012-06-11 18:21 ` Joel Brobecker @ 2012-06-12 14:05 ` Pedro Alves 2014-10-06 0:42 ` [PATCH v2 1/2] " Maciej W. Rozycki 1 sibling, 0 replies; 27+ messages in thread From: Pedro Alves @ 2012-06-12 14:05 UTC (permalink / raw) To: Joel Brobecker Cc: Maciej W. Rozycki, gdb-patches, Rich Fuhler, Richard Sandiford, binutils, gcc On 06/11/2012 07:20 PM, Joel Brobecker wrote: > From the ChangeLog entry, it seems like Pedro was involved in the making > of that patch, so perhaps he could be a good reviewer? All involvement I recall was updating a couple lines to new interfaces in the context of a merge from upstream. All else is pretty much as good as new to me. :-) -- Pedro Alves ^ permalink raw reply [flat|nested] 27+ messages in thread
* [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2012-06-11 18:21 ` Joel Brobecker 2012-06-12 14:05 ` Pedro Alves @ 2014-10-06 0:42 ` Maciej W. Rozycki 2014-10-06 0:42 ` [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests Maciej W. Rozycki ` (5 more replies) 1 sibling, 6 replies; 27+ messages in thread From: Maciej W. Rozycki @ 2014-10-06 0:42 UTC (permalink / raw) To: Joel Brobecker; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford Joel, I'd like to get back to this review, this change is important for reliable debugging compressed code, as test results quoted below indicate. On Mon, 11 Jun 2012, Joel Brobecker wrote: > > I propose therefore to accept the existing inconsistencies and deal > > with them entirely within GDB. I have figured out that the ISA bit > > lost in various places can still be recovered as long as we have > > symbol information -- that'll have the st_other attribute correctly > > set to one of standard MIPS/MIPS16/microMIPS. > > That seems reasonable to me. Great! > Just some minor comments below: > > > gdb-mips16-isa-bit.diff > > Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.c > > =================================================================== > > --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.c 2012-05-14 16:00:33.000000000 +0100 > > +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.c 2012-05-14 16:02:02.235560558 +0100 > > @@ -358,9 +358,15 @@ mips_elf_make_msymbol_special (asymbol * > > return; > > > > if (ELF_ST_IS_MICROMIPS (elfsym->internal_elf_sym.st_other)) > > - MSYMBOL_TARGET_FLAG_2 (msym) = 1; > > + { > > + MSYMBOL_TARGET_FLAG_2 (msym) = 1; > > + SYMBOL_VALUE_ADDRESS (msym) |= 1; > > + } > > else if (ELF_ST_IS_MIPS16 (elfsym->internal_elf_sym.st_other)) > > - MSYMBOL_TARGET_FLAG_1 (msym) = 1; > > + { > > + MSYMBOL_TARGET_FLAG_1 (msym) = 1; > > + SYMBOL_VALUE_ADDRESS (msym) |= 1; > > + } > > This remark is not to be considered as part of this patch's review, > since this is already an established practice, but I think we could do > better than using magic numbers for those flags. I understand they have > to be that way in the common code, but perhaps mips-tdep could define > aliases? Something like MSYMBOL_MIPS_TARGET_FLAG_BLAH? Good point, I'll push such a change as a followup; I have it ready now. > > Index: gdb-fsf-trunk-quilt/gdb/arch-utils.c > [...] > > +void > > +default_make_symbol_special (struct symbol *sym, struct objfile *objfile) > > Can you put a comment at the start of all these new functions to say > that the documentation is in the header file? For instance, I think > we typically say: > > /* See arch-utils.h. */ > > This is to help someone reviewing the code to know that there is > in fact documentation. Done now; please note there's no prior example of such commentary in arch-utils.c, so it's easy to fall in the trap of following surrounding code. > > +/* Do nothing version of make_symbol_special. */ > > + > > +void default_make_symbol_special (struct symbol *sym, struct objfile *objfile); > > + > > +/* Do nothing version of adjust_dwarf2_addr. */ > > + > > +CORE_ADDR default_adjust_dwarf2_addr (CORE_ADDR pc); > > + > > +/* Do nothing version of adjust_dwarf2_line. */ > > + > > +CORE_ADDR default_adjust_dwarf2_line (CORE_ADDR addr, int rel); > > I would definitely mention gdbarch in the function comment. > Prior art uses: > > /* Default implementation of gdbarch_displaced_hw_singlestep. */ > > Should we remain consistent with that? It can be a mix of both, > saying what method it implements, and then what it actually does, Eg: > > /* Default version of gdbarch_make_symbol_special (does nothing). */ Like above, e.g.: /* Do nothing version of coff_make_msymbol_special. */ but again I find your suggestion very reasonable. I decided upon wording like: /* Do nothing default implementation of gdbarch_make_symbol_special. */ > > Index: gdb-fsf-trunk-quilt/gdb/solib.c > > =================================================================== > > --- gdb-fsf-trunk-quilt.orig/gdb/solib.c 2012-05-14 15:56:45.000000000 +0100 > > +++ gdb-fsf-trunk-quilt/gdb/solib.c 2012-05-14 16:01:34.645558637 +0100 > > @@ -1384,8 +1384,27 @@ gdb_bfd_lookup_symbol_from_symtab (bfd * > > > > if (match_sym (sym, data)) > > { > > + symaddr = sym->value; > > + > > + /* macro/2012-04-20: Some ELF targets fiddle with addresses > > + of symbols they consider special. They use minimal symbols > > + to do that and this is needed for correct breakpoint > > + placement, but we do not have full data here to build a > > + complete minimal symbol, so just set the address and let the > > + targets cope with that. */ > > + if (bfd_get_flavour (abfd) == bfd_target_elf_flavour) > > + { > > + struct minimal_symbol msym; > > + > > + memset (&msym, 0, sizeof (msym)); > > + SYMBOL_VALUE_ADDRESS (&msym) = symaddr; > > + gdbarch_elf_make_msymbol_special (target_gdbarch, > > + sym, &msym); > > + symaddr = SYMBOL_VALUE_ADDRESS (&msym); > > + } > > Is there a way we could avoid that block if > gdbarch_elf_make_msymbol_special points to the default implementation? > It just seems sad to be doing all this work of creating a temporary symbol > for every solib symbol on non-mips targets... > > One possible way would have been to provide no default value for > gdbarch_elf_make_msymbol_special, but then it makes the calling > of that method a little trickier. That could be worked around by > changing the default value of gdbarch_elf_make_msymbol_special to NULL, > and then having an extra wrapper function that call the gdbarch > method if not NULL or else does nothing. The only other call sites are in `elf_symtab_read' in gdb/elfread.c, so while I agree the block in gdb/solib.c will best be avoided, I'd call your proposal to make a wrapper an overkill and actually a pessimisation. Instead I propose the change I included below, where `gdbarch_elf_make_msymbol_special_p' is called once per `elf_symtab_read' invocation only, avoiding all the `gdbarch_elf_make_msymbol_special_p' or `gdbarch_elf_make_msymbol_special' indirection being called for nothing where the latter function doesn't do anything. This is actually an improvement over our current trunk as it is. We can think of something fancier if a need arises in the future. Here's an updated change. Beside the changes above I have made an update to `micromips_linux_sigframe_validate' in gdb/mips-linux-tdep.c, to take the change to support microMIPS signal frames posted separately into account. Also I have updated numerous places in gdb/dwarf2read.c so that the ISA bit is applied to the PC consistently. This cured some regressions the original change caused, so that there are no regressions after this change and the test suite fix sent as a complement. Finally a test case is now also included that covers scenarios this fix addresses; the test case fails for MIPS16 and microMIPS multilibs with our current trunk as it is. It is quite possible that the test case will reveal issues on other targets too. For those who have already forgotten what it is all about, the original description is available here: http://sourceware.org/ml/gdb-patches/2012-05/msg00515.html I have regression tested this updated version with the mips-linux-gnu target and the following multilibs: -EB -EB -msoft-float -EB -mips16 -EB -mips16 -msoft-float -EB -mmicromips -EB -mmicromips -msoft-float -EB -mabi=n32 -EB -mabi=n32 -msoft-float -EB -mabi=64 -EB -mabi=64 -msoft-float and the -EL variants of same, reducing the total number of failures counted across all multilibs from ~8500 to ~3750. Numbers are indicative because of intermittent errors; gdb.threads/schedlock.exp is particularly notorious with fluctuations in the range of 50 total across all multilibs being normal. These numbers include fixes made with the other change. An indicative test summary change for a single multilib, -EB -mmicromips, before: === gdb Summary === # of expected passes 16585 # of unexpected failures 1272 # of unexpected successes 1 # of expected failures 29 # of known failures 54 # of unresolved testcases 5 # of untested testcases 50 # of unsupported tests 257 and after: === gdb Summary === # of expected passes 17717 # of unexpected failures 175 # of unexpected successes 1 # of expected failures 29 # of known failures 56 # of unresolved testcases 5 # of untested testcases 50 # of unsupported tests 268 This change does cause regressions, across all multilibs, in some test cases that use artificial DWARF-2 records. These include: gdb.cp/nsalias.exp gdb.dwarf2/dw2-canonicalize-type.exp gdb.dwarf2/dw2-empty-pc-range.exp gdb.dwarf2/pr11465.exp It also causes regressions across compressed ISA multilibs only in: gdb.cp/expand-psymtabs-cxx.exp gdb.dwarf2/dw2-case-insensitive.exp These tests just make wrong assumptions that can't stand for the MIPS target, e.g. that a function can be 1-byte long (wrong, the smallest MIPS instruction takes 2 bytes). The other patch sent as a complement addresses these issues, which are all test case deficiencies rather than problems with this change. OK to apply? 2014-10-06 Maciej W. Rozycki <macro@codesourcery.com> Maciej W. Rozycki <macro@mips.com> Pedro Alves <pedro@codesourcery.com> * gdbarch.sh (elf_make_msymbol_special): Change type to `F', remove `predefault' and `invalid_p' initializers. (make_symbol_special): New architecture method. (adjust_dwarf2_addr, adjust_dwarf2_line): Likewise. (objfile, symbol): New declarations. * arch-utils.h (default_elf_make_msymbol_special): Remove prototype. (default_make_symbol_special): New prototype. (default_adjust_dwarf2_addr): Likewise. (default_adjust_dwarf2_line): Likewise. * mips-tdep.h (mips_unmake_compact_addr): New prototype. * arch-utils.c (default_elf_make_msymbol_special): Remove function. (default_make_symbol_special): New function. (default_adjust_dwarf2_addr): Likewise. (default_adjust_dwarf2_line): Likewise. * dwarf2-frame.c (decode_frame_entry_1): Call `gdbarch_adjust_dwarf2_addr'. * dwarf2loc.c (dwarf2_find_location_expression): Likewise. * dwarf2read.c (create_addrmap_from_index): Likewise. (process_psymtab_comp_unit_reader): Likewise. (add_partial_symbol): Likewise. (add_partial_subprogram): Likewise. (process_full_comp_unit): Likewise. (read_file_scope): Likewise. (read_func_scope): Likewise. Call `gdbarch_make_symbol_special'. (read_lexical_block_scope): Call `gdbarch_adjust_dwarf2_addr'. (read_call_site_scope): Likewise. (dwarf2_ranges_read): Likewise. (dwarf2_record_block_ranges): Likewise. (read_attribute_value): Likewise. (dwarf_decode_lines_1): Call `gdbarch_adjust_dwarf2_line'. (new_symbol_full): Call `gdbarch_adjust_dwarf2_addr'. * elfread.c (elf_symtab_read): Don't call `gdbarch_elf_make_msymbol_special' if unset. * mips-linux-tdep.c (micromips_linux_sigframe_validate): Strip the ISA bit from the PC. * mips-tdep.c (mips_unmake_compact_addr): New function. (mips_elf_make_msymbol_special): Set the ISA bit in the symbol's address appropriately. (mips_make_symbol_special): New function. (mips_pc_is_mips): Set the ISA bit before symbol lookup. (mips_pc_is_mips16): Likewise. (mips_pc_is_micromips): Likewise. (mips_pc_isa): Likewise. (mips_adjust_dwarf2_addr): New function. (mips_adjust_dwarf2_line): Likewise. (mips_read_pc, mips_unwind_pc): Keep the ISA bit. (mips_addr_bits_remove): Likewise. (mips_skip_trampoline_code): Likewise. (mips_write_pc): Don't set the ISA bit. (mips_eabi_push_dummy_call): Likewise. (mips_o64_push_dummy_call): Likewise. (mips_gdbarch_init): Install `mips_make_symbol_special', `mips_adjust_dwarf2_addr' and `mips_adjust_dwarf2_line' gdbarch handlers. * solib.c (gdb_bfd_lookup_symbol_from_symtab): Get target-specific symbol address adjustments. * gdbarch.h: Regenerate. * gdbarch.c: Regenerate. 2014-10-06 Maciej W. Rozycki <macro@codesourcery.com> gdb/testsuite/ * gdb.base/func-ptrs.c: New file. * gdb.base/func-ptrs.exp: New file. Maciej gdb-mips16-isa-bit.diff Index: gdb-fsf-trunk-quilt/gdb/arch-utils.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/arch-utils.c 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/arch-utils.c 2014-10-03 14:50:26.000000000 +0100 @@ -31,6 +31,7 @@ #include "target-descriptions.h" #include "objfiles.h" #include "language.h" +#include "symtab.h" #include "version.h" @@ -167,17 +168,35 @@ no_op_reg_to_regnum (struct gdbarch *gdb } void -default_elf_make_msymbol_special (asymbol *sym, struct minimal_symbol *msym) +default_coff_make_msymbol_special (int val, struct minimal_symbol *msym) { return; } +/* See arch-utils.h. */ + void -default_coff_make_msymbol_special (int val, struct minimal_symbol *msym) +default_make_symbol_special (struct symbol *sym, struct objfile *objfile) { return; } +/* See arch-utils.h. */ + +CORE_ADDR +default_adjust_dwarf2_addr (CORE_ADDR pc) +{ + return pc; +} + +/* See arch-utils.h. */ + +CORE_ADDR +default_adjust_dwarf2_line (CORE_ADDR addr, int rel) +{ + return addr; +} + int cannot_register_not (struct gdbarch *gdbarch, int regnum) { Index: gdb-fsf-trunk-quilt/gdb/arch-utils.h =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/arch-utils.h 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/arch-utils.h 2014-10-03 14:50:26.000000000 +0100 @@ -68,15 +68,22 @@ extern gdbarch_convert_from_func_ptr_add extern int no_op_reg_to_regnum (struct gdbarch *gdbarch, int reg); -/* Do nothing version of elf_make_msymbol_special. */ - -void default_elf_make_msymbol_special (asymbol *sym, - struct minimal_symbol *msym); - /* Do nothing version of coff_make_msymbol_special. */ void default_coff_make_msymbol_special (int val, struct minimal_symbol *msym); +/* Do nothing default implementation of gdbarch_make_symbol_special. */ + +void default_make_symbol_special (struct symbol *sym, struct objfile *objfile); + +/* Do nothing default implementation of gdbarch_adjust_dwarf2_addr. */ + +CORE_ADDR default_adjust_dwarf2_addr (CORE_ADDR pc); + +/* Do nothing default implementation of gdbarch_adjust_dwarf2_line. */ + +CORE_ADDR default_adjust_dwarf2_line (CORE_ADDR addr, int rel); + /* Version of cannot_fetch_register() / cannot_store_register() that always fails. */ Index: gdb-fsf-trunk-quilt/gdb/dwarf2-frame.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/dwarf2-frame.c 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/dwarf2-frame.c 2014-10-03 14:50:26.398945943 +0100 @@ -2067,6 +2067,7 @@ decode_frame_entry_1 (struct comp_unit * { /* This is a FDE. */ struct dwarf2_fde *fde; + CORE_ADDR addr; /* Check that an FDE was expected. */ if ((entry_type & EH_FDE_TYPE_ID) == 0) @@ -2100,14 +2101,16 @@ decode_frame_entry_1 (struct comp_unit * gdb_assert (fde->cie != NULL); - fde->initial_location = - read_encoded_value (unit, fde->cie->encoding, fde->cie->ptr_size, - buf, &bytes_read, 0); + addr = read_encoded_value (unit, fde->cie->encoding, fde->cie->ptr_size, + buf, &bytes_read, 0); + fde->initial_location = gdbarch_adjust_dwarf2_addr (gdbarch, addr); buf += bytes_read; fde->address_range = read_encoded_value (unit, fde->cie->encoding & 0x0f, fde->cie->ptr_size, buf, &bytes_read, 0); + addr = gdbarch_adjust_dwarf2_addr (gdbarch, addr + fde->address_range); + fde->address_range = addr - fde->initial_location; buf += bytes_read; /* A 'z' augmentation in the CIE implies the presence of an Index: gdb-fsf-trunk-quilt/gdb/dwarf2loc.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/dwarf2loc.c 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/dwarf2loc.c 2014-10-03 14:50:26.398945943 +0100 @@ -4294,6 +4294,9 @@ loclist_describe_location (struct symbol low += base_address; high += base_address; + low = gdbarch_adjust_dwarf2_addr (gdbarch, low); + high = gdbarch_adjust_dwarf2_addr (gdbarch, high); + length = extract_unsigned_integer (loc_ptr, 2, byte_order); loc_ptr += 2; Index: gdb-fsf-trunk-quilt/gdb/dwarf2read.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/dwarf2read.c 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/dwarf2read.c 2014-10-06 00:13:56.039006986 +0100 @@ -2824,6 +2824,7 @@ create_signatured_type_table_from_index static void create_addrmap_from_index (struct objfile *objfile, struct mapped_index *index) { + struct gdbarch *gdbarch = get_objfile_arch (objfile); const gdb_byte *iter, *end; struct obstack temp_obstack; struct addrmap *mutable_map; @@ -2865,8 +2866,9 @@ create_addrmap_from_index (struct objfil continue; } - addrmap_set_empty (mutable_map, lo + baseaddr, hi + baseaddr - 1, - dw2_get_cutu (cu_index)); + lo = gdbarch_adjust_dwarf2_addr (gdbarch, lo + baseaddr); + hi = gdbarch_adjust_dwarf2_addr (gdbarch, hi + baseaddr); + addrmap_set_empty (mutable_map, lo, hi - 1, dw2_get_cutu (cu_index)); } objfile->psymtabs_addrmap = addrmap_create_fixed (mutable_map, @@ -5848,6 +5850,7 @@ process_psymtab_comp_unit_reader (const { struct dwarf2_cu *cu = reader->cu; struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); struct dwarf2_per_cu_data *per_cu = cu->per_cu; struct attribute *attr; CORE_ADDR baseaddr; @@ -5892,8 +5895,11 @@ process_psymtab_comp_unit_reader (const /* Store the contiguous range if it is not empty; it can be empty for CUs with no code. */ addrmap_set_empty (objfile->psymtabs_addrmap, - best_lowpc + baseaddr, - best_highpc + baseaddr - 1, pst); + gdbarch_adjust_dwarf2_addr (gdbarch, + best_lowpc + baseaddr), + gdbarch_adjust_dwarf2_addr (gdbarch, + best_highpc + baseaddr) - 1, + pst); /* Check if comp unit has_children. If so, read the rest of the partial symbols from this comp unit. @@ -5924,8 +5930,8 @@ process_psymtab_comp_unit_reader (const best_highpc = highpc; } } - pst->textlow = best_lowpc + baseaddr; - pst->texthigh = best_highpc + baseaddr; + pst->textlow = gdbarch_adjust_dwarf2_addr (gdbarch, best_lowpc + baseaddr); + pst->texthigh = gdbarch_adjust_dwarf2_addr (gdbarch, best_highpc + baseaddr); pst->n_global_syms = objfile->global_psymbols.next - (objfile->global_psymbols.list + pst->globals_offset); @@ -6788,6 +6794,7 @@ static void add_partial_symbol (struct partial_die_info *pdi, struct dwarf2_cu *cu) { struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); CORE_ADDR addr = 0; const char *actual_name = NULL; CORE_ADDR baseaddr; @@ -6805,31 +6812,30 @@ add_partial_symbol (struct partial_die_i switch (pdi->tag) { case DW_TAG_subprogram: + addr = gdbarch_adjust_dwarf2_addr (gdbarch, pdi->lowpc + baseaddr); if (pdi->is_external || cu->language == language_ada) { /* brobecker/2007-12-26: Normally, only "external" DIEs are part of the global scope. But in Ada, we want to be able to access nested procedures globally. So all Ada subprograms are stored in the global scope. */ - /* prim_record_minimal_symbol (actual_name, pdi->lowpc + baseaddr, - mst_text, objfile); */ + /* prim_record_minimal_symbol (actual_name, addr, mst_text, + objfile); */ add_psymbol_to_list (actual_name, strlen (actual_name), built_actual_name != NULL, VAR_DOMAIN, LOC_BLOCK, &objfile->global_psymbols, - 0, pdi->lowpc + baseaddr, - cu->language, objfile); + 0, addr, cu->language, objfile); } else { - /* prim_record_minimal_symbol (actual_name, pdi->lowpc + baseaddr, - mst_file_text, objfile); */ + /* prim_record_minimal_symbol (actual_name, addr, mst_file_text, + objfile); */ add_psymbol_to_list (actual_name, strlen (actual_name), built_actual_name != NULL, VAR_DOMAIN, LOC_BLOCK, &objfile->static_psymbols, - 0, pdi->lowpc + baseaddr, - cu->language, objfile); + 0, addr, cu->language, objfile); } break; case DW_TAG_constant: @@ -7020,6 +7026,9 @@ add_partial_subprogram (struct partial_d CORE_ADDR *lowpc, CORE_ADDR *highpc, int set_addrmap, struct dwarf2_cu *cu) { + struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); + if (pdi->tag == DW_TAG_subprogram) { if (pdi->has_pc_info) @@ -7030,14 +7039,18 @@ add_partial_subprogram (struct partial_d *highpc = pdi->highpc; if (set_addrmap) { - CORE_ADDR baseaddr; struct objfile *objfile = cu->objfile; + CORE_ADDR baseaddr; + CORE_ADDR highpc; + CORE_ADDR lowpc; baseaddr = ANOFFSET (objfile->section_offsets, SECT_OFF_TEXT (objfile)); - addrmap_set_empty (objfile->psymtabs_addrmap, - pdi->lowpc + baseaddr, - pdi->highpc - 1 + baseaddr, + lowpc = gdbarch_adjust_dwarf2_addr (gdbarch, + pdi->lowpc + baseaddr); + highpc = gdbarch_adjust_dwarf2_addr (gdbarch, + pdi->highpc + baseaddr); + addrmap_set_empty (objfile->psymtabs_addrmap, lowpc, highpc - 1, cu->per_cu->v.psymtab); } } @@ -7921,11 +7934,13 @@ process_full_comp_unit (struct dwarf2_pe { struct dwarf2_cu *cu = per_cu->cu; struct objfile *objfile = per_cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); CORE_ADDR lowpc, highpc; struct symtab *symtab; struct cleanup *back_to, *delayed_list_cleanup; CORE_ADDR baseaddr; struct block *static_block; + CORE_ADDR addr; baseaddr = ANOFFSET (objfile->section_offsets, SECT_OFF_TEXT (objfile)); @@ -7956,8 +7971,8 @@ process_full_comp_unit (struct dwarf2_pe it, by scanning the DIE's below the compilation unit. */ get_scope_pc_bounds (cu->dies, &lowpc, &highpc, cu); - static_block - = end_symtab_get_static_block (highpc + baseaddr, objfile, 0, 1); + addr = gdbarch_adjust_dwarf2_addr (gdbarch, highpc + baseaddr); + static_block = end_symtab_get_static_block (addr, objfile, 0, 1); /* If the comp unit has DW_AT_ranges, it may have discontiguous ranges. Also, DW_AT_ranges may record ranges not belonging to any child DIEs @@ -9001,6 +9016,7 @@ static void read_file_scope (struct die_info *die, struct dwarf2_cu *cu) { struct objfile *objfile = dwarf2_per_objfile->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); struct cleanup *back_to = make_cleanup (null_cleanup, 0); CORE_ADDR lowpc = ((CORE_ADDR) -1); CORE_ADDR highpc = ((CORE_ADDR) 0); @@ -9019,8 +9035,7 @@ read_file_scope (struct die_info *die, s from finish_block. */ if (lowpc == ((CORE_ADDR) -1)) lowpc = highpc; - lowpc += baseaddr; - highpc += baseaddr; + lowpc = gdbarch_adjust_dwarf2_addr (gdbarch, lowpc + baseaddr); find_file_and_directory (die, cu, &name, &comp_dir); @@ -11131,6 +11146,7 @@ static void read_func_scope (struct die_info *die, struct dwarf2_cu *cu) { struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); struct context_stack *new; CORE_ADDR lowpc; CORE_ADDR highpc; @@ -11183,8 +11199,8 @@ read_func_scope (struct die_info *die, s return; } - lowpc += baseaddr; - highpc += baseaddr; + lowpc = gdbarch_adjust_dwarf2_addr (gdbarch, lowpc + baseaddr); + highpc = gdbarch_adjust_dwarf2_addr (gdbarch, highpc + baseaddr); /* If we have any template arguments, then we must allocate a different sort of symbol. */ @@ -11271,6 +11287,8 @@ read_func_scope (struct die_info *die, s /* If we have address ranges, record them. */ dwarf2_record_block_ranges (die, block, baseaddr, cu); + gdbarch_make_symbol_special (gdbarch, new->name, objfile); + /* Attach template arguments to function. */ if (! VEC_empty (symbolp, template_args)) { @@ -11307,6 +11325,7 @@ static void read_lexical_block_scope (struct die_info *die, struct dwarf2_cu *cu) { struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); struct context_stack *new; CORE_ADDR lowpc, highpc; struct die_info *child_die; @@ -11321,8 +11340,8 @@ read_lexical_block_scope (struct die_inf describe ranges. */ if (!dwarf2_get_pc_bounds (die, &lowpc, &highpc, cu, NULL)) return; - lowpc += baseaddr; - highpc += baseaddr; + lowpc = gdbarch_adjust_dwarf2_addr (gdbarch, lowpc + baseaddr); + highpc = gdbarch_adjust_dwarf2_addr (gdbarch, highpc + baseaddr); push_context (0, lowpc); if (die->child != NULL) @@ -11384,6 +11403,7 @@ read_call_site_scope (struct die_info *d return; } pc = attr_value_as_address (attr) + baseaddr; + pc = gdbarch_adjust_dwarf2_addr (gdbarch, pc); if (cu->call_site_htab == NULL) cu->call_site_htab = htab_create_alloc_ex (16, core_addr_hash, core_addr_eq, @@ -11532,7 +11552,10 @@ read_call_site_scope (struct die_info *d "low pc, for referencing DIE 0x%x [in module %s]"), die->offset.sect_off, objfile_name (objfile)); else - SET_FIELD_PHYSADDR (call_site->target, lowpc + baseaddr); + { + lowpc = gdbarch_adjust_dwarf2_addr (gdbarch, lowpc + baseaddr); + SET_FIELD_PHYSADDR (call_site->target, lowpc); + } } } else @@ -11660,6 +11683,7 @@ dwarf2_ranges_read (unsigned offset, COR struct partial_symtab *ranges_pst) { struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); struct comp_unit_head *cu_header = &cu->header; bfd *obfd = objfile->obfd; unsigned int addr_size = cu_header->addr_size; @@ -11767,10 +11791,17 @@ dwarf2_ranges_read (unsigned offset, COR } if (ranges_pst != NULL) - addrmap_set_empty (objfile->psymtabs_addrmap, - range_beginning + baseaddr, - range_end - 1 + baseaddr, - ranges_pst); + { + CORE_ADDR lowpc; + CORE_ADDR highpc; + + lowpc = gdbarch_adjust_dwarf2_addr (gdbarch, + range_beginning + baseaddr); + highpc = gdbarch_adjust_dwarf2_addr (gdbarch, + range_end + baseaddr); + addrmap_set_empty (objfile->psymtabs_addrmap, lowpc, highpc - 1, + ranges_pst); + } /* FIXME: This is recording everything as a low-high segment of consecutive addresses. We should have a @@ -11984,6 +12015,7 @@ dwarf2_record_block_ranges (struct die_i CORE_ADDR baseaddr, struct dwarf2_cu *cu) { struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); struct attribute *attr; struct attribute *attr_high; @@ -11999,7 +12031,9 @@ dwarf2_record_block_ranges (struct die_i if (cu->header.version >= 4 && attr_form_is_constant (attr_high)) high += low; - record_block_range (block, baseaddr + low, baseaddr + high - 1); + low = gdbarch_adjust_dwarf2_addr (gdbarch, low + baseaddr); + high = gdbarch_adjust_dwarf2_addr (gdbarch, high + baseaddr); + record_block_range (block, low, high - 1); } } @@ -12103,6 +12137,8 @@ dwarf2_record_block_ranges (struct die_i continue; } + start = gdbarch_adjust_dwarf2_addr (gdbarch, start); + end = gdbarch_adjust_dwarf2_addr (gdbarch, end); record_block_range (block, start, end - 1); } } @@ -15926,6 +15962,8 @@ read_attribute_value (const struct die_r const gdb_byte *info_ptr) { struct dwarf2_cu *cu = reader->cu; + struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); bfd *abfd = reader->abfd; struct comp_unit_head *cu_header = &cu->header; unsigned int bytes_read; @@ -15948,6 +15986,7 @@ read_attribute_value (const struct die_r break; case DW_FORM_addr: DW_ADDR (attr) = read_address (abfd, info_ptr, cu, &bytes_read); + DW_ADDR (attr) = gdbarch_adjust_dwarf2_addr (gdbarch, DW_ADDR (attr)); info_ptr += bytes_read; break; case DW_FORM_block2: @@ -17278,7 +17317,7 @@ dwarf_decode_lines_1 (struct line_header while (line_ptr < line_end) { /* state machine registers */ - CORE_ADDR address = 0; + CORE_ADDR address = gdbarch_adjust_dwarf2_line (gdbarch, 0, 0); unsigned int file = 1; unsigned int line = 1; int is_stmt = lh->default_is_stmt; @@ -17321,12 +17360,14 @@ dwarf_decode_lines_1 (struct line_header { /* Special opcode. */ unsigned char adj_opcode; + CORE_ADDR addr_adj; int line_delta; adj_opcode = op_code - lh->opcode_base; - address += (((op_index + (adj_opcode / lh->line_range)) + addr_adj = (((op_index + (adj_opcode / lh->line_range)) / lh->maximum_ops_per_instruction) * lh->minimum_instruction_length); + address += gdbarch_adjust_dwarf2_line (gdbarch, addr_adj, 1); op_index = ((op_index + (adj_opcode / lh->line_range)) % lh->maximum_ops_per_instruction); line_delta = lh->line_base + (adj_opcode % lh->line_range); @@ -17403,6 +17444,7 @@ dwarf_decode_lines_1 (struct line_header op_index = 0; line_ptr += bytes_read; address += baseaddr; + address = gdbarch_adjust_dwarf2_line (gdbarch, address, 0); break; case DW_LNE_define_file: { @@ -17480,10 +17522,12 @@ dwarf_decode_lines_1 (struct line_header { CORE_ADDR adjust = read_unsigned_leb128 (abfd, line_ptr, &bytes_read); + CORE_ADDR addr_adj; - address += (((op_index + adjust) + addr_adj = (((op_index + adjust) / lh->maximum_ops_per_instruction) * lh->minimum_instruction_length); + address += gdbarch_adjust_dwarf2_line (gdbarch, addr_adj, 1); op_index = ((op_index + adjust) % lh->maximum_ops_per_instruction); line_ptr += bytes_read; @@ -17543,18 +17587,25 @@ dwarf_decode_lines_1 (struct line_header case DW_LNS_const_add_pc: { CORE_ADDR adjust = (255 - lh->opcode_base) / lh->line_range; + CORE_ADDR addr_adj; - address += (((op_index + adjust) + addr_adj = (((op_index + adjust) / lh->maximum_ops_per_instruction) * lh->minimum_instruction_length); + address += gdbarch_adjust_dwarf2_line (gdbarch, addr_adj, 1); op_index = ((op_index + adjust) % lh->maximum_ops_per_instruction); } break; case DW_LNS_fixed_advance_pc: - address += read_2_bytes (abfd, line_ptr); - op_index = 0; - line_ptr += 2; + { + CORE_ADDR addr_adj; + + addr_adj = read_2_bytes (abfd, line_ptr); + address += gdbarch_adjust_dwarf2_line (gdbarch, addr_adj, 1); + op_index = 0; + line_ptr += 2; + } break; default: { @@ -17811,6 +17862,7 @@ new_symbol_full (struct die_info *die, s struct symbol *space) { struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); struct symbol *sym = NULL; const char *name; struct attribute *attr = NULL; @@ -17888,8 +17940,13 @@ new_symbol_full (struct die_info *die, s case DW_TAG_label: attr = dwarf2_attr (die, DW_AT_low_pc, cu); if (attr) - SYMBOL_VALUE_ADDRESS (sym) - = attr_value_as_address (attr) + baseaddr; + { + CORE_ADDR addr; + + addr = attr_value_as_address (attr); + addr = gdbarch_adjust_dwarf2_addr (gdbarch, addr + baseaddr); + SYMBOL_VALUE_ADDRESS (sym) = addr; + } SYMBOL_TYPE (sym) = objfile_type (objfile)->builtin_core_addr; SYMBOL_DOMAIN (sym) = LABEL_DOMAIN; SYMBOL_ACLASS_INDEX (sym) = LOC_LABEL; Index: gdb-fsf-trunk-quilt/gdb/elfread.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/elfread.c 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/elfread.c 2014-10-03 14:50:26.398945943 +0100 @@ -247,6 +247,8 @@ elf_symtab_read (struct objfile *objfile const char *filesymname = ""; struct dbx_symfile_info *dbx = DBX_SYMFILE_INFO (objfile); int stripped = (bfd_get_symcount (objfile->obfd) == 0); + int elf_make_msymbol_special_p + = gdbarch_elf_make_msymbol_special_p (gdbarch); for (i = 0; i < number_of_symbols; i++) { @@ -330,7 +332,8 @@ elf_symtab_read (struct objfile *objfile if (msym != NULL) { msym->filename = filesymname; - gdbarch_elf_make_msymbol_special (gdbarch, sym, msym); + if (elf_make_msymbol_special_p) + gdbarch_elf_make_msymbol_special (gdbarch, sym, msym); } continue; } @@ -558,7 +561,8 @@ elf_symtab_read (struct objfile *objfile } msym->filename = filesymname; - gdbarch_elf_make_msymbol_special (gdbarch, sym, msym); + if (elf_make_msymbol_special_p) + gdbarch_elf_make_msymbol_special (gdbarch, sym, msym); } /* If we see a default versioned symbol, install it under @@ -597,7 +601,9 @@ elf_symtab_read (struct objfile *objfile SET_MSYMBOL_SIZE (mtramp, MSYMBOL_SIZE (msym)); mtramp->created_by_gdb = 1; mtramp->filename = filesymname; - gdbarch_elf_make_msymbol_special (gdbarch, sym, mtramp); + if (elf_make_msymbol_special_p) + gdbarch_elf_make_msymbol_special (gdbarch, + sym, mtramp); } } } Index: gdb-fsf-trunk-quilt/gdb/gdbarch.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.c 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/gdbarch.c 2014-10-03 14:50:26.000000000 +0100 @@ -252,6 +252,9 @@ struct gdbarch gdbarch_in_function_epilogue_p_ftype *in_function_epilogue_p; gdbarch_elf_make_msymbol_special_ftype *elf_make_msymbol_special; gdbarch_coff_make_msymbol_special_ftype *coff_make_msymbol_special; + gdbarch_make_symbol_special_ftype *make_symbol_special; + gdbarch_adjust_dwarf2_addr_ftype *adjust_dwarf2_addr; + gdbarch_adjust_dwarf2_line_ftype *adjust_dwarf2_line; int cannot_step_breakpoint; int have_nonsteppable_watchpoint; gdbarch_address_class_type_flags_ftype *address_class_type_flags; @@ -390,8 +393,10 @@ gdbarch_alloc (const struct gdbarch_info gdbarch->skip_solib_resolver = generic_skip_solib_resolver; gdbarch->in_solib_return_trampoline = generic_in_solib_return_trampoline; gdbarch->in_function_epilogue_p = generic_in_function_epilogue_p; - gdbarch->elf_make_msymbol_special = default_elf_make_msymbol_special; gdbarch->coff_make_msymbol_special = default_coff_make_msymbol_special; + gdbarch->make_symbol_special = default_make_symbol_special; + gdbarch->adjust_dwarf2_addr = default_adjust_dwarf2_addr; + gdbarch->adjust_dwarf2_line = default_adjust_dwarf2_line; gdbarch->register_reggroup_p = default_register_reggroup_p; gdbarch->displaced_step_hw_singlestep = default_displaced_step_hw_singlestep; gdbarch->displaced_step_fixup = NULL; @@ -561,8 +566,11 @@ verify_gdbarch (struct gdbarch *gdbarch) /* Skip verify of skip_solib_resolver, invalid_p == 0 */ /* Skip verify of in_solib_return_trampoline, invalid_p == 0 */ /* Skip verify of in_function_epilogue_p, invalid_p == 0 */ - /* Skip verify of elf_make_msymbol_special, invalid_p == 0 */ + /* Skip verify of elf_make_msymbol_special, has predicate. */ /* Skip verify of coff_make_msymbol_special, invalid_p == 0 */ + /* Skip verify of make_symbol_special, invalid_p == 0 */ + /* Skip verify of adjust_dwarf2_addr, invalid_p == 0 */ + /* Skip verify of adjust_dwarf2_line, invalid_p == 0 */ /* Skip verify of cannot_step_breakpoint, invalid_p == 0 */ /* Skip verify of have_nonsteppable_watchpoint, invalid_p == 0 */ /* Skip verify of address_class_type_flags, has predicate. */ @@ -684,6 +692,12 @@ gdbarch_dump (struct gdbarch *gdbarch, s "gdbarch_dump: adjust_breakpoint_address = <%s>\n", host_address_to_string (gdbarch->adjust_breakpoint_address)); fprintf_unfiltered (file, + "gdbarch_dump: adjust_dwarf2_addr = <%s>\n", + host_address_to_string (gdbarch->adjust_dwarf2_addr)); + fprintf_unfiltered (file, + "gdbarch_dump: adjust_dwarf2_line = <%s>\n", + host_address_to_string (gdbarch->adjust_dwarf2_line)); + fprintf_unfiltered (file, "gdbarch_dump: auto_charset = <%s>\n", host_address_to_string (gdbarch->auto_charset)); fprintf_unfiltered (file, @@ -834,6 +848,9 @@ gdbarch_dump (struct gdbarch *gdbarch, s "gdbarch_dump: ecoff_reg_to_regnum = <%s>\n", host_address_to_string (gdbarch->ecoff_reg_to_regnum)); fprintf_unfiltered (file, + "gdbarch_dump: gdbarch_elf_make_msymbol_special_p() = %d\n", + gdbarch_elf_make_msymbol_special_p (gdbarch)); + fprintf_unfiltered (file, "gdbarch_dump: elf_make_msymbol_special = <%s>\n", host_address_to_string (gdbarch->elf_make_msymbol_special)); fprintf_unfiltered (file, @@ -1008,6 +1025,9 @@ gdbarch_dump (struct gdbarch *gdbarch, s "gdbarch_dump: make_corefile_notes = <%s>\n", host_address_to_string (gdbarch->make_corefile_notes)); fprintf_unfiltered (file, + "gdbarch_dump: make_symbol_special = <%s>\n", + host_address_to_string (gdbarch->make_symbol_special)); + fprintf_unfiltered (file, "gdbarch_dump: gdbarch_max_insn_length_p() = %d\n", gdbarch_max_insn_length_p (gdbarch)); fprintf_unfiltered (file, @@ -3055,6 +3075,13 @@ set_gdbarch_in_function_epilogue_p (stru gdbarch->in_function_epilogue_p = in_function_epilogue_p; } +int +gdbarch_elf_make_msymbol_special_p (struct gdbarch *gdbarch) +{ + gdb_assert (gdbarch != NULL); + return gdbarch->elf_make_msymbol_special != NULL; +} + void gdbarch_elf_make_msymbol_special (struct gdbarch *gdbarch, asymbol *sym, struct minimal_symbol *msym) { @@ -3089,6 +3116,57 @@ set_gdbarch_coff_make_msymbol_special (s gdbarch->coff_make_msymbol_special = coff_make_msymbol_special; } +void +gdbarch_make_symbol_special (struct gdbarch *gdbarch, struct symbol *sym, struct objfile *objfile) +{ + gdb_assert (gdbarch != NULL); + gdb_assert (gdbarch->make_symbol_special != NULL); + if (gdbarch_debug >= 2) + fprintf_unfiltered (gdb_stdlog, "gdbarch_make_symbol_special called\n"); + gdbarch->make_symbol_special (sym, objfile); +} + +void +set_gdbarch_make_symbol_special (struct gdbarch *gdbarch, + gdbarch_make_symbol_special_ftype make_symbol_special) +{ + gdbarch->make_symbol_special = make_symbol_special; +} + +CORE_ADDR +gdbarch_adjust_dwarf2_addr (struct gdbarch *gdbarch, CORE_ADDR pc) +{ + gdb_assert (gdbarch != NULL); + gdb_assert (gdbarch->adjust_dwarf2_addr != NULL); + if (gdbarch_debug >= 2) + fprintf_unfiltered (gdb_stdlog, "gdbarch_adjust_dwarf2_addr called\n"); + return gdbarch->adjust_dwarf2_addr (pc); +} + +void +set_gdbarch_adjust_dwarf2_addr (struct gdbarch *gdbarch, + gdbarch_adjust_dwarf2_addr_ftype adjust_dwarf2_addr) +{ + gdbarch->adjust_dwarf2_addr = adjust_dwarf2_addr; +} + +CORE_ADDR +gdbarch_adjust_dwarf2_line (struct gdbarch *gdbarch, CORE_ADDR addr, int rel) +{ + gdb_assert (gdbarch != NULL); + gdb_assert (gdbarch->adjust_dwarf2_line != NULL); + if (gdbarch_debug >= 2) + fprintf_unfiltered (gdb_stdlog, "gdbarch_adjust_dwarf2_line called\n"); + return gdbarch->adjust_dwarf2_line (addr, rel); +} + +void +set_gdbarch_adjust_dwarf2_line (struct gdbarch *gdbarch, + gdbarch_adjust_dwarf2_line_ftype adjust_dwarf2_line) +{ + gdbarch->adjust_dwarf2_line = adjust_dwarf2_line; +} + int gdbarch_cannot_step_breakpoint (struct gdbarch *gdbarch) { Index: gdb-fsf-trunk-quilt/gdb/gdbarch.h =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.h 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/gdbarch.h 2014-10-03 14:50:26.000000000 +0100 @@ -51,6 +51,8 @@ struct target_ops; struct obstack; struct bp_target_info; struct target_desc; +struct objfile; +struct symbol; struct displaced_step_closure; struct core_regset_section; struct syscall; @@ -682,6 +684,8 @@ typedef int (gdbarch_in_function_epilogu extern int gdbarch_in_function_epilogue_p (struct gdbarch *gdbarch, CORE_ADDR addr); extern void set_gdbarch_in_function_epilogue_p (struct gdbarch *gdbarch, gdbarch_in_function_epilogue_p_ftype *in_function_epilogue_p); +extern int gdbarch_elf_make_msymbol_special_p (struct gdbarch *gdbarch); + typedef void (gdbarch_elf_make_msymbol_special_ftype) (asymbol *sym, struct minimal_symbol *msym); extern void gdbarch_elf_make_msymbol_special (struct gdbarch *gdbarch, asymbol *sym, struct minimal_symbol *msym); extern void set_gdbarch_elf_make_msymbol_special (struct gdbarch *gdbarch, gdbarch_elf_make_msymbol_special_ftype *elf_make_msymbol_special); @@ -690,6 +694,18 @@ typedef void (gdbarch_coff_make_msymbol_ extern void gdbarch_coff_make_msymbol_special (struct gdbarch *gdbarch, int val, struct minimal_symbol *msym); extern void set_gdbarch_coff_make_msymbol_special (struct gdbarch *gdbarch, gdbarch_coff_make_msymbol_special_ftype *coff_make_msymbol_special); +typedef void (gdbarch_make_symbol_special_ftype) (struct symbol *sym, struct objfile *objfile); +extern void gdbarch_make_symbol_special (struct gdbarch *gdbarch, struct symbol *sym, struct objfile *objfile); +extern void set_gdbarch_make_symbol_special (struct gdbarch *gdbarch, gdbarch_make_symbol_special_ftype *make_symbol_special); + +typedef CORE_ADDR (gdbarch_adjust_dwarf2_addr_ftype) (CORE_ADDR pc); +extern CORE_ADDR gdbarch_adjust_dwarf2_addr (struct gdbarch *gdbarch, CORE_ADDR pc); +extern void set_gdbarch_adjust_dwarf2_addr (struct gdbarch *gdbarch, gdbarch_adjust_dwarf2_addr_ftype *adjust_dwarf2_addr); + +typedef CORE_ADDR (gdbarch_adjust_dwarf2_line_ftype) (CORE_ADDR addr, int rel); +extern CORE_ADDR gdbarch_adjust_dwarf2_line (struct gdbarch *gdbarch, CORE_ADDR addr, int rel); +extern void set_gdbarch_adjust_dwarf2_line (struct gdbarch *gdbarch, gdbarch_adjust_dwarf2_line_ftype *adjust_dwarf2_line); + extern int gdbarch_cannot_step_breakpoint (struct gdbarch *gdbarch); extern void set_gdbarch_cannot_step_breakpoint (struct gdbarch *gdbarch, int cannot_step_breakpoint); Index: gdb-fsf-trunk-quilt/gdb/gdbarch.sh =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.sh 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/gdbarch.sh 2014-10-03 14:50:26.000000000 +0100 @@ -635,8 +635,11 @@ m:int:in_solib_return_trampoline:CORE_AD # which don't suffer from that problem could just let this functionality # untouched. m:int:in_function_epilogue_p:CORE_ADDR addr:addr:0:generic_in_function_epilogue_p::0 -f:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym::default_elf_make_msymbol_special::0 +F:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym f:void:coff_make_msymbol_special:int val, struct minimal_symbol *msym:val, msym::default_coff_make_msymbol_special::0 +f:void:make_symbol_special:struct symbol *sym, struct objfile *objfile:sym, objfile::default_make_symbol_special::0 +f:CORE_ADDR:adjust_dwarf2_addr:CORE_ADDR pc:pc::default_adjust_dwarf2_addr::0 +f:CORE_ADDR:adjust_dwarf2_line:CORE_ADDR addr, int rel:addr, rel::default_adjust_dwarf2_line::0 v:int:cannot_step_breakpoint:::0:0::0 v:int:have_nonsteppable_watchpoint:::0:0::0 F:int:address_class_type_flags:int byte_size, int dwarf2_addr_class:byte_size, dwarf2_addr_class @@ -1140,6 +1143,8 @@ struct target_ops; struct obstack; struct bp_target_info; struct target_desc; +struct objfile; +struct symbol; struct displaced_step_closure; struct core_regset_section; struct syscall; Index: gdb-fsf-trunk-quilt/gdb/mips-linux-tdep.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/mips-linux-tdep.c 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/mips-linux-tdep.c 2014-10-03 14:50:26.398945943 +0100 @@ -1372,7 +1372,13 @@ micromips_linux_sigframe_validate (const struct frame_info *this_frame, CORE_ADDR *pc) { - return mips_pc_is_micromips (get_frame_arch (this_frame), *pc); + if (mips_pc_is_micromips (get_frame_arch (this_frame), *pc)) + { + *pc = mips_unmake_compact_addr (*pc); + return 1; + } + else + return 0; } /* Implement the "write_pc" gdbarch method. */ Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.c 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.c 2014-10-06 00:11:55.028611249 +0100 @@ -333,6 +333,15 @@ make_compact_addr (CORE_ADDR addr) return ((addr) | (CORE_ADDR) 1); } +/* Extern version of unmake_compact_addr; we use a separate function + so that unmake_compact_addr can be inlined throughout this file. */ + +CORE_ADDR +mips_unmake_compact_addr (CORE_ADDR addr) +{ + return unmake_compact_addr (addr); +} + /* Functions for setting and testing a bit in a minimal symbol that marks it as MIPS16 or microMIPS function. The MSB of the minimal symbol's "info" field is used for this purpose. @@ -362,9 +371,15 @@ mips_elf_make_msymbol_special (asymbol * return; if (ELF_ST_IS_MICROMIPS (st_other)) - MSYMBOL_TARGET_FLAG_2 (msym) = 1; + { + MSYMBOL_TARGET_FLAG_2 (msym) = 1; + SET_MSYMBOL_VALUE_ADDRESS (msym, MSYMBOL_VALUE_RAW_ADDRESS (msym) | 1); + } else if (ELF_ST_IS_MIPS16 (st_other)) - MSYMBOL_TARGET_FLAG_1 (msym) = 1; + { + MSYMBOL_TARGET_FLAG_1 (msym) = 1; + SET_MSYMBOL_VALUE_ADDRESS (msym, MSYMBOL_VALUE_RAW_ADDRESS (msym) | 1); + } } /* Return one iff MSYM refers to standard ISA code. */ @@ -391,6 +406,29 @@ msymbol_is_micromips (struct minimal_sym return MSYMBOL_TARGET_FLAG_2 (msym); } +/* Set the ISA bit in the main symbol too, complementing the corresponding + minimal symbol setting and reflecting the run-time value of the symbol. */ + +static void +mips_make_symbol_special (struct symbol *sym, struct objfile *objfile) +{ + if (SYMBOL_CLASS (sym) == LOC_BLOCK) + { + CORE_ADDR compact_block_start; + struct bound_minimal_symbol msym; + + compact_block_start = BLOCK_START (SYMBOL_BLOCK_VALUE (sym)) | 1; + msym = lookup_minimal_symbol_by_pc (compact_block_start); + if (msym.minsym && !msymbol_is_mips (msym.minsym)) + { + /* We are in symbol reading so it is OK to cast away constness. */ + struct block *block = (struct block *) SYMBOL_BLOCK_VALUE (sym); + + BLOCK_START (block) = compact_block_start; + } + } +} + /* XFER a value from the big/little/left end of the register. Depending on the size of the value it might occupy the entire register or just part of it. Make an allowance for this, aligning @@ -1125,7 +1163,7 @@ mips_pc_is_mips (CORE_ADDR memaddr) stored by elfread.c in the high bit of the info field. Use this to decide if the function is standard MIPS. Otherwise if bit 0 of the address is clear, then this is a standard MIPS function. */ - sym = lookup_minimal_symbol_by_pc (memaddr); + sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); if (sym.minsym) return msymbol_is_mips (sym.minsym); else @@ -1143,7 +1181,7 @@ mips_pc_is_mips16 (struct gdbarch *gdbar elfread.c in the high bit of the info field. Use this to decide if the function is MIPS16. Otherwise if bit 0 of the address is set, then ELF file flags will tell if this is a MIPS16 function. */ - sym = lookup_minimal_symbol_by_pc (memaddr); + sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); if (sym.minsym) return msymbol_is_mips16 (sym.minsym); else @@ -1162,7 +1200,7 @@ mips_pc_is_micromips (struct gdbarch *gd if the function is microMIPS. Otherwise if bit 0 of the address is set, then ELF file flags will tell if this is a microMIPS function. */ - sym = lookup_minimal_symbol_by_pc (memaddr); + sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); if (sym.minsym) return msymbol_is_micromips (sym.minsym); else @@ -1182,7 +1220,7 @@ mips_pc_isa (struct gdbarch *gdbarch, CO this to decide if the function is MIPS16 or microMIPS or normal MIPS. Otherwise if bit 0 of the address is set, then ELF file flags will tell if this is a MIPS16 or a microMIPS function. */ - sym = lookup_minimal_symbol_by_pc (memaddr); + sym = lookup_minimal_symbol_by_pc (make_compact_addr (memaddr)); if (sym.minsym) { if (msymbol_is_micromips (sym.minsym)) @@ -1203,6 +1241,32 @@ mips_pc_isa (struct gdbarch *gdbarch, CO } } +/* Set the ISA bit correctly in the PC, used by DWARF-2 machinery. */ + +static CORE_ADDR +mips_adjust_dwarf2_addr (CORE_ADDR pc) +{ + pc = unmake_compact_addr (pc); + return mips_pc_is_mips (pc) ? pc : make_compact_addr (pc); +} + +/* Recalculate the line record requested so that the resulting PC has the + ISA bit set correctly, used by DWARF-2 machinery. */ + +static CORE_ADDR +mips_adjust_dwarf2_line (CORE_ADDR addr, int rel) +{ + static CORE_ADDR adj_pc; + static CORE_ADDR pc; + CORE_ADDR isa_pc; + + pc = rel ? pc + addr : addr; + isa_pc = mips_adjust_dwarf2_addr (pc); + addr = rel ? isa_pc - adj_pc : isa_pc; + adj_pc = isa_pc; + return addr; +} + /* Various MIPS16 thunk (aka stub or trampoline) names. */ static const char mips_str_mips16_call_stub[] = "__mips16_call_stub_"; @@ -1252,8 +1316,6 @@ mips_read_pc (struct regcache *regcache) LONGEST pc; regcache_cooked_read_signed (regcache, regnum, &pc); - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); return pc; } @@ -1263,8 +1325,6 @@ mips_unwind_pc (struct gdbarch *gdbarch, CORE_ADDR pc; pc = frame_unwind_register_signed (next_frame, gdbarch_pc_regnum (gdbarch)); - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); /* macro/2012-04-20: This hack skips over MIPS16 call thunks as intermediate frames. In this case we can get the caller's address from $ra, or if $ra contains an address within a thunk as well, then @@ -1274,15 +1334,9 @@ mips_unwind_pc (struct gdbarch *gdbarch, { pc = frame_unwind_register_signed (next_frame, gdbarch_num_regs (gdbarch) + MIPS_RA_REGNUM); - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); if (mips_in_frame_stub (pc)) - { - pc = frame_unwind_register_signed - (next_frame, gdbarch_num_regs (gdbarch) + MIPS_S2_REGNUM); - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); - } + pc = frame_unwind_register_signed + (next_frame, gdbarch_num_regs (gdbarch) + MIPS_S2_REGNUM); } return pc; } @@ -1316,10 +1370,7 @@ mips_write_pc (struct regcache *regcache { int regnum = gdbarch_pc_regnum (get_regcache_arch (regcache)); - if (mips_pc_is_mips (pc)) - regcache_cooked_write_unsigned (regcache, regnum, pc); - else - regcache_cooked_write_unsigned (regcache, regnum, make_compact_addr (pc)); + regcache_cooked_write_unsigned (regcache, regnum, pc); } /* Fetch and return instruction from the specified location. Handle @@ -3650,9 +3701,6 @@ mips_addr_bits_remove (struct gdbarch *g { struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch); - if (is_compact_addr (addr)) - addr = unmake_compact_addr (addr); - if (mips_mask_address_p (tdep) && (((ULONGEST) addr) >> 32 == 0xffffffffUL)) /* This hack is a work-around for existing boards using PMON, the simulator, and any other 64-bit targets that doesn't have true @@ -4351,25 +4399,9 @@ mips_eabi_push_dummy_call (struct gdbarc "mips_eabi_push_dummy_call: %d len=%d type=%d", argnum + 1, len, (int) typecode); - /* Function pointer arguments to mips16 code need to be made into - mips16 pointers. */ - if (typecode == TYPE_CODE_PTR - && TYPE_CODE (TYPE_TARGET_TYPE (arg_type)) == TYPE_CODE_FUNC) - { - CORE_ADDR addr = extract_signed_integer (value_contents (arg), - len, byte_order); - if (mips_pc_is_mips (addr)) - val = value_contents (arg); - else - { - store_signed_integer (valbuf, len, byte_order, - make_compact_addr (addr)); - val = valbuf; - } - } /* The EABI passes structures that do not fit in a register by reference. */ - else if (len > regsize + if (len > regsize && (typecode == TYPE_CODE_STRUCT || typecode == TYPE_CODE_UNION)) { store_unsigned_integer (valbuf, regsize, byte_order, @@ -5734,7 +5766,6 @@ mips_o64_push_dummy_call (struct gdbarch for (argnum = 0; argnum < nargs; argnum++) { const gdb_byte *val; - gdb_byte valbuf[MAX_REGISTER_SIZE]; struct value *arg = args[argnum]; struct type *arg_type = check_typedef (value_type (arg)); int len = TYPE_LENGTH (arg_type); @@ -5747,21 +5778,6 @@ mips_o64_push_dummy_call (struct gdbarch val = value_contents (arg); - /* Function pointer arguments to mips16 code need to be made into - mips16 pointers. */ - if (typecode == TYPE_CODE_PTR - && TYPE_CODE (TYPE_TARGET_TYPE (arg_type)) == TYPE_CODE_FUNC) - { - CORE_ADDR addr = extract_signed_integer (value_contents (arg), - len, byte_order); - if (!mips_pc_is_mips (addr)) - { - store_signed_integer (valbuf, len, byte_order, - make_compact_addr (addr)); - val = valbuf; - } - } - /* Floating point arguments passed in registers have to be treated specially. On 32-bit architectures, doubles are passed in register pairs; the even FP register gets the @@ -7678,27 +7694,15 @@ mips_skip_trampoline_code (struct frame_ new_pc = mips_skip_mips16_trampoline_code (frame, pc); if (new_pc) - { - pc = new_pc; - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); - } + pc = new_pc; new_pc = find_solib_trampoline_target (frame, pc); if (new_pc) - { - pc = new_pc; - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); - } + pc = new_pc; new_pc = mips_skip_pic_trampoline_code (frame, pc); if (new_pc) - { - pc = new_pc; - if (is_compact_addr (pc)) - pc = unmake_compact_addr (pc); - } + pc = new_pc; } while (pc != target_pc); @@ -8354,6 +8358,9 @@ mips_gdbarch_init (struct gdbarch_info i set_gdbarch_elf_make_msymbol_special (gdbarch, mips_elf_make_msymbol_special); + set_gdbarch_make_symbol_special (gdbarch, mips_make_symbol_special); + set_gdbarch_adjust_dwarf2_addr (gdbarch, mips_adjust_dwarf2_addr); + set_gdbarch_adjust_dwarf2_line (gdbarch, mips_adjust_dwarf2_line); regnum = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct mips_regnum); *regnum = mips_regnum; Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.h =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.h 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.h 2014-10-05 23:03:19.088958183 +0100 @@ -155,6 +155,9 @@ enum /* Single step based on where the current instruction will take us. */ extern int mips_software_single_step (struct frame_info *frame); +/* Strip the ISA (compression) bit off from ADDR. */ +extern CORE_ADDR mips_unmake_compact_addr (CORE_ADDR addr); + /* Tell if the program counter value in MEMADDR is in a standard MIPS function. */ extern int mips_pc_is_mips (bfd_vma memaddr); Index: gdb-fsf-trunk-quilt/gdb/solib.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/solib.c 2014-10-03 13:52:46.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/solib.c 2014-10-03 14:50:26.398945943 +0100 @@ -1444,8 +1444,28 @@ gdb_bfd_lookup_symbol_from_symtab (bfd * if (match_sym (sym, data)) { + struct gdbarch *gdbarch = target_gdbarch (); + symaddr = sym->value; + + /* Some ELF targets fiddle with addresses of symbols they + consider special. They use minimal symbols to do that + and this is needed for correct breakpoint placement, + but we do not have full data here to build a complete + minimal symbol, so just set the address and let the + targets cope with that. */ + if (bfd_get_flavour (abfd) == bfd_target_elf_flavour + && gdbarch_elf_make_msymbol_special_p (gdbarch)) + { + struct minimal_symbol msym; + + memset (&msym, 0, sizeof (msym)); + SET_MSYMBOL_VALUE_ADDRESS (&msym, symaddr); + gdbarch_elf_make_msymbol_special (gdbarch, sym, &msym); + symaddr = MSYMBOL_VALUE_RAW_ADDRESS (&msym); + } + /* BFD symbols are section relative. */ - symaddr = sym->value + sym->section->vma; + symaddr += sym->section->vma; break; } } Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.base/func-ptrs.c =================================================================== --- /dev/null 1970-01-01 00:00:00.000000000 +0000 +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.base/func-ptrs.c 2014-10-03 14:50:26.398945943 +0100 @@ -0,0 +1,50 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright 2014 Free Software Foundation, Inc. + + This program 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 of the License, or + (at your option) any later version. + + This program 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. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ + +void +sentinel (void) +{ + return; +} + +int +incr (int i) +{ + sentinel (); + return i + 1; +} + +int +decr (int i) +{ + sentinel (); + return i - 1; +} + +int (*calc) (int) = incr; + +int +main (void) +{ + int i = -1; + + i = calc (i); + i = calc (i); + i = calc (i); + + return i; +} Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.base/func-ptrs.exp =================================================================== --- /dev/null 1970-01-01 00:00:00.000000000 +0000 +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.base/func-ptrs.exp 2014-10-03 14:50:26.398945943 +0100 @@ -0,0 +1,95 @@ +# Copyright 2014 Free Software Foundation, Inc. + +# This program 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 of the License, or +# (at your option) any later version. +# +# This program 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. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. + +set testname func-ptrs +set srcfile ${testname}.c +if { [prepare_for_testing ${testname}.exp ${testname} ${srcfile}] } { + return -1 +} + +if { ![runto_main] } { + untested ${testname}.exp + return -1 +} + + +# First set our breakpoints. + +set fp_breakpoint_re \ + "Breakpoint $decimal at $hex: file .*${srcfile}, line $decimal\\." +gdb_test "break sentinel if calc == decr" \ + "${fp_breakpoint_re}" \ + "breakpoint at sentinel" +gdb_test "break incr" \ + "${fp_breakpoint_re}" \ + "breakpoint at incr" +gdb_test "break decr" \ + "${fp_breakpoint_re}" \ + "breakpoint at decr" + + +# Check if we run through to the breakpoint in incr. + +gdb_test "continue" \ + "Breakpoint $decimal, incr \\(i=-1\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*" \ + "continue to incr, first time" + + +# Go back up, make sure the return value is 0. + +gdb_test "finish" \ + "Run till exit from #0 +incr \\(i=-1\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+($hex in )?main \\(\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*Value returned is \\$$decimal = 0" \ + "go back to main from incr, first time" + + +# Redirect calc and see if we run to the breakpoint in decr instead. + +gdb_test_no_output "set calc = decr" "set calc to decr" +gdb_test "continue" \ + "Breakpoint $decimal, decr \\(i=0\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*" \ + "continue to decr" + + +# Go back up, check if we stop in sentinel instead. + +gdb_test "finish" \ + "Run till exit from #0 +decr \\(i=0\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+Breakpoint $decimal, sentinel \\(\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*" \ + "stop in sentinel" + + +# Go back all the way up to main, make sure the return value is -1. + +gdb_test_no_output "up-silently" "move up to decr" +gdb_test "finish" \ + "Run till exit from #1 +($hex in )?decr \\(i=0\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+($hex in )?main \\(\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*Value returned is \\$$decimal = -1" \ + "go back to main from decr" + + +# Reset calc and see if we run to the breakpoint in incr again. + +gdb_test_no_output "set calc = incr" "set calc to incr" +gdb_test "continue" \ + "Breakpoint $decimal, incr \\(i=-1\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*" \ + "continue to incr, second time" + + +# Go back up again, make sure the return value is 0. + +gdb_test "finish" \ + "Run till exit from #0 +incr \\(i=-1\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+($hex in )?main \\(\\)\[ \r\n\]+at .*${srcfile}:$decimal\[\r\n\]+.*Value returned is \\$$decimal = 0" \ + "go back to main from incr, second time" + + +# All done! ^ permalink raw reply [flat|nested] 27+ messages in thread
* [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests 2014-10-06 0:42 ` [PATCH v2 1/2] " Maciej W. Rozycki @ 2014-10-06 0:42 ` Maciej W. Rozycki 2014-11-16 11:09 ` Joel Brobecker 2014-11-16 22:28 ` Doug Evans 2014-10-06 14:10 ` [PATCH v2 1/2] ISA bit treatment on the MIPS platform Joel Brobecker ` (4 subsequent siblings) 5 siblings, 2 replies; 27+ messages in thread From: Maciej W. Rozycki @ 2014-10-06 0:42 UTC (permalink / raw) To: Joel Brobecker; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford Joel, Here's the second change, to cover issues triggered by the MIPS ISA bit handling change, usually in tests that make artificial DWARF-2 records. Here are fixes for the problems, listed individually: * gdb.cp/expand-psymtabs-cxx.exp -- this test is debugging an object file and assuming addresses will be 0; with the ISA bit set code addresses are 1 instead: (gdb) PASS: gdb.cp/expand-psymtabs-cxx.exp: set language c++ p 'method(long)' $1 = {void (long)} 0x1 <method(long)> (gdb) FAIL: gdb.cp/expand-psymtabs-cxx.exp: before expand p method $2 = {void (long)} 0x1 <method(long)> (gdb) FAIL: gdb.cp/expand-psymtabs-cxx.exp: force expand p 'method(long)' $3 = {void (long)} 0x1 <method(long)> (gdb) FAIL: gdb.cp/expand-psymtabs-cxx.exp: after expand Fixed by matching any hex number, there's no value AFAICT for the test in matching 0 exactly, and I suppose the method's offset within section can be non-zero for some other reasons on other targets too. * gdb.cp/nsalias.exp -- this assumes instructions can be aligned arbitrarily and places code labels at odd addreses, setting the ISA bit and wreaking havoc: (gdb) PASS: gdb.cp/nsalias.exp: print outer::inner::innermost::x list outer::inner::innermost::foo Function "outer::inner::innermost::foo" not defined. (gdb) FAIL: gdb.cp/nsalias.exp: list outer::inner::innermost::foo break *outer::inner::innermost::foo No symbol "foo" in namespace "outer::inner::innermost". (gdb) FAIL: gdb.cp/nsalias.exp: setting breakpoint at *outer::inner::innermost::foo delete $bpnum No breakpoint number 6. (gdb) FAIL: gdb.cp/nsalias.exp: (outer::inner::innermost): delete $bpnum -- etc., etc... Fixed by aligning labels to 4; required by many processors. * gdb.dwarf2/dw2-canonicalize-type.exp, gdb.dwarf2/dw2-empty-pc-range.exp, gdb.dwarf2/pr11465.exp -- these assume an instruction and consequently a function can take as little as 1 byte, which makes it impossible to look up a code symbol by an address with the ISA bit set as the address is already beyond the end of the function: (gdb) ptype f No symbol "f" in current context. (gdb) FAIL: gdb.dwarf2/dw2-canonicalize-type.exp: ptype f (gdb) PASS: gdb.dwarf2/dw2-empty-pc-range.exp: empty range before CU load ptype realrange No symbol "realrange" in current context. (gdb) FAIL: gdb.dwarf2/dw2-empty-pc-range.exp: valid range after CU load (gdb) p N::c.C Cannot take address of method C. (gdb) FAIL: gdb.dwarf2/pr11465.exp: p N::c.C -- fixed by increasing the size of the function to 4 (perhaps code in gdb/mips-tdep.c could look up code symbols up to twice, with and failing that without the ISA bit set, but it seems wrong to me to implement specific handling for invalid code just to satisfy test cases that assume too much about the target). * gdb.dwarf2/dw2-case-insensitive.exp -- an artificial code label is created, but does not work because data (a `.align' pseudo-op in this case) follows and as a result the label has no MIPS16 or microMIPS annotation in the symbol table: (gdb) PASS: gdb.dwarf2/dw2-case-insensitive.exp: set case-sensitive off info functions fUnC_lang All functions matching regular expression "fUnC_lang": File file1.txt: foo FUNC_lang(void); Non-debugging symbols: 0x004006e0 FUNC_lang_start (gdb) FAIL: gdb.dwarf2/dw2-case-insensitive.exp: regexp case-sensitive off -- fixed by adding a `.insn' pseudo-op on MIPS targets; the pseudo-op marks data as instructions. * gdb.dwarf2/dw2-stack-boundary.exp -- the test case enables complaints and assumes none will be issued beyond ones explicitly arranged by the test case, however overlapping sections are noticed because now minimal symbols are looked up by `mips_adjust_dwarf2_addr' in DWARF-2 record processing: (gdb) set complaints 100 (gdb) PASS: gdb.dwarf2/dw2-stack-boundary.exp: set complaints 100 file ./dw2-stack-boundary Reading symbols from ./dw2-stack-boundary...location description stack underflow...location description stack overflow...unexpected overlap between: (A) section `.reginfo' from `.../gdb.dwarf2/dw2-stack-boundary' [0x0, 0x18) (B) section `*COM*' from `.../gdb.dwarf2/dw2-stack-boundary' [0x0, 0x0). Will ignore section B...unexpected overlap between: (A) section `.reginfo' from `.../gdb.dwarf2/dw2-stack-boundary' [0x0, 0x18) (B) section `*UND*' from `.../gdb.dwarf2/dw2-stack-boundary' [0x0, 0x0). Will ignore section B...unexpected overlap between: (A) section `.reginfo' from `.../gdb.dwarf2/dw2-stack-boundary' [0x0, 0x18) (B) section `*ABS*' from `.../gdb.dwarf2/dw2-stack-boundary' [0x0, 0x0). Will ignore section B...done. (gdb) FAIL: gdb.dwarf2/dw2-stack-boundary.exp: check partial symtab errors -- fixed by ignoring any extra noise as long as what we look for is found. OK to apply? 2014-10-06 Maciej W. Rozycki <macro@codesourcery.com> gdb/testsuite/ * gdb.cp/expand-psymtabs-cxx.exp: Accept any address of `method(long)', not just 0x0. * gdb.cp/nsalias.exp: Align code labels to 4. * gdb.dwarf2/dw2-canonicalize-type.S (main): Expand to 4-bytes. * gdb.dwarf2/dw2-empty-pc-range.S (main): Likewise. * gdb.dwarf2/pr11465.S (_ZN1N1cE): Likewise. * gdb.dwarf2/dw2-case-insensitive.c (START_INSNS): New macro. (cu_text_start, FUNC_lang_start): Use `START_INSNS'. * gdb.dwarf2/dw2-stack-boundary.exp: Accept noise in complaints. Maciej gdb-mips16-isa-bit-test-fix.diff Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.cp/expand-psymtabs-cxx.exp =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.cp/expand-psymtabs-cxx.exp 2014-10-02 07:56:23.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.cp/expand-psymtabs-cxx.exp 2014-10-02 07:58:10.978958268 +0100 @@ -30,9 +30,9 @@ gdb_test_no_output "set language c++" # FAIL was: # $1 = {<text variable, no debug info>} 0 -gdb_test "p 'method(long)'" { = {void \(long\)} 0x0 <method.long.>} \ +gdb_test "p 'method(long)'" " = {void \\(long\\)} $hex <method.long.>" \ "before expand" -gdb_test "p method" { = {void \(long\)} 0x0 <method.long.>} \ +gdb_test "p method" " = {void \\(long\\)} $hex <method.long.>" \ "force expand" -gdb_test "p 'method(long)'" { = {void \(long\)} 0x0 <method.long.>} \ +gdb_test "p 'method(long)'" " = {void \\(long\\)} $hex <method.long.>" \ "after expand" Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.cp/nsalias.exp =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.cp/nsalias.exp 2014-10-02 07:56:23.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.cp/nsalias.exp 2014-10-02 07:58:10.978958268 +0100 @@ -153,20 +153,20 @@ Dwarf::assemble $asm_file { subprogram { {specification :$im_foo_label} - {low_pc 0x1 DW_FORM_addr} - {high_pc 0x2 DW_FORM_addr} + {low_pc 0x4 DW_FORM_addr} + {high_pc 0x7 DW_FORM_addr} } subprogram { {specification :$i_foo_label} - {low_pc 0x3 DW_FORM_addr} - {high_pc 0x4 DW_FORM_addr} + {low_pc 0x8 DW_FORM_addr} + {high_pc 0xb DW_FORM_addr} } subprogram { {specification :$o_foo_label} - {low_pc 0x5 DW_FORM_addr} - {high_pc 0x6 DW_FORM_addr} + {low_pc 0xc DW_FORM_addr} + {high_pc 0xf DW_FORM_addr} } } } Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S 2014-10-02 07:56:23.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S 2014-10-02 07:58:10.978958268 +0100 @@ -15,7 +15,8 @@ .text .globl main -main: .byte 0 +main: + .dc.l 0 .Lmain_end: .section .debug_info debug_start: Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-case-insensitive.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/dw2-case-insensitive.c 2014-10-02 07:56:23.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-case-insensitive.c 2014-10-02 07:58:10.978958268 +0100 @@ -15,13 +15,22 @@ You should have received a copy of the GNU General Public License along with this program. If not, see <http://www.gnu.org/licenses/>. */ +/* Target-specific way of forcing an instruction label. */ +#ifdef __mips__ +#define START_INSNS asm (".insn"); +#else +#define START_INSNS +#endif + /* Use DW_LANG_Fortran90 for case insensitive DWARF. */ asm (".globl cu_text_start"); asm ("cu_text_start:"); +START_INSNS asm (".globl FUNC_lang_start"); asm (".p2align 4"); asm ("FUNC_lang_start:"); +START_INSNS void FUNC_lang (void) Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-empty-pc-range.S =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/dw2-empty-pc-range.S 2014-10-02 07:56:23.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-empty-pc-range.S 2014-10-02 07:58:10.978958268 +0100 @@ -15,7 +15,7 @@ .text pc_start: - .byte 0 + .dc.l 0 pc_end: .section .debug_info Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-stack-boundary.exp =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/dw2-stack-boundary.exp 2014-10-02 07:56:23.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-stack-boundary.exp 2014-10-02 07:58:10.978958268 +0100 @@ -38,7 +38,7 @@ if [is_remote host] { } } gdb_test_no_output "set complaints 100" -gdb_test "file $binfile" {Reading symbols from .*\.\.\.location description stack underflow\.\.\.location description stack overflow\.\.\.done\.} "check partial symtab errors" +gdb_test "file $binfile" {Reading symbols from .*\.\.\.location description stack underflow\.\.\.location description stack overflow\.\.\..*done\.} "check partial symtab errors" gdb_test "p underflow" {Asked for position 0 of stack, stack only has 0 elements on it\.} gdb_test "p overflow" " = 2" Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/pr11465.S =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/pr11465.S 2014-10-02 07:56:23.000000000 +0100 +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/pr11465.S 2014-10-02 07:58:10.978958268 +0100 @@ -39,7 +39,7 @@ text_start: _ZN1N1cE: /* Valid function must have non-empty PC range. */ - .byte 0 + .dc.l 0 text_end: .section .debug_info @@ -117,7 +117,8 @@ dieaf: .uleb128 0xe /* DW_TAG_const_typ dieb4: .uleb128 0xf /* DW_TAG_subprogram */ .4byte die95-d /* DW_AT_abstract_origin */ .4byte _ZN1N1cE /* DW_AT_low_pc */ - .4byte _ZN1N1cE + 1 /* DW_AT_high_pc */ + .4byte _ZN1N1cE + (text_end - text_start) + /* DW_AT_high_pc */ diec9: .uleb128 0x10 /* DW_TAG_subprogram */ .4byte die9f-d /* DW_AT_abstract_origin */ .byte 2f-1f /* DW_AT_location */ @@ -136,7 +137,8 @@ dieda: .uleb128 0x11 /* DW_TAG_subprogr .4byte .LASF8 /* DW_AT_name */ .4byte dief2-d /* DW_AT_type */ .4byte _ZN1N1cE /* DW_AT_low_pc */ - .4byte _ZN1N1cE + 1 /* DW_AT_high_pc */ + .4byte _ZN1N1cE + (text_end - text_start) + /* DW_AT_high_pc */ dief2: .uleb128 0x12 /* DW_TAG_base_type */ .byte 0x4 /* DW_AT_byte_size */ .byte 0x5 /* DW_AT_encoding */ ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests 2014-10-06 0:42 ` [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests Maciej W. Rozycki @ 2014-11-16 11:09 ` Joel Brobecker 2014-11-16 18:32 ` Doug Evans 2014-11-16 22:28 ` Doug Evans 1 sibling, 1 reply; 27+ messages in thread From: Joel Brobecker @ 2014-11-16 11:09 UTC (permalink / raw) To: Maciej W. Rozycki; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford > 2014-10-06 Maciej W. Rozycki <macro@codesourcery.com> > > gdb/testsuite/ > * gdb.cp/expand-psymtabs-cxx.exp: Accept any address of > `method(long)', not just 0x0. > * gdb.cp/nsalias.exp: Align code labels to 4. > * gdb.dwarf2/dw2-canonicalize-type.S (main): Expand to 4-bytes. > * gdb.dwarf2/dw2-empty-pc-range.S (main): Likewise. > * gdb.dwarf2/pr11465.S (_ZN1N1cE): Likewise. > * gdb.dwarf2/dw2-case-insensitive.c (START_INSNS): New macro. > (cu_text_start, FUNC_lang_start): Use `START_INSNS'. > * gdb.dwarf2/dw2-stack-boundary.exp: Accept noise in complaints. Overall, everything looks very reasonable. Small nit: There is a space at the end of the first CL line (after "address of"). One question (pre-approved if my suggestion works): > Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S > =================================================================== > --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S 2014-10-02 07:56:23.000000000 +0100 > +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S 2014-10-02 07:58:10.978958268 +0100 > @@ -15,7 +15,8 @@ > > .text > .globl main > -main: .byte 0 > +main: > + .dc.l 0 We've never used .dl.l before, and I don't know how widely available it is (in fact, I couldn't find it in the GAS manual). How about using .word or .4byte? It would also be consistent with what we've usually be using. Would that work? > Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-empty-pc-range.S > =================================================================== > --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/dw2-empty-pc-range.S 2014-10-02 07:56:23.000000000 +0100 > +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-empty-pc-range.S 2014-10-02 07:58:10.978958268 +0100 > @@ -15,7 +15,7 @@ > > .text > pc_start: > - .byte 0 > + .dc.l 0 > pc_end: Same here... > Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/pr11465.S > =================================================================== > --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/pr11465.S 2014-10-02 07:56:23.000000000 +0100 > +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/pr11465.S 2014-10-02 07:58:10.978958268 +0100 > @@ -39,7 +39,7 @@ > text_start: > _ZN1N1cE: > /* Valid function must have non-empty PC range. */ > - .byte 0 > + .dc.l 0 ... and here. -- Joel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests 2014-11-16 11:09 ` Joel Brobecker @ 2014-11-16 18:32 ` Doug Evans 2014-11-16 19:49 ` Doug Evans 0 siblings, 1 reply; 27+ messages in thread From: Doug Evans @ 2014-11-16 18:32 UTC (permalink / raw) To: Joel Brobecker Cc: Maciej W. Rozycki, gdb-patches, Rich Fuhler, Richard Sandiford On Sun, Nov 16, 2014 at 3:09 AM, Joel Brobecker <brobecker@adacore.com> wrote: >> Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S >> =================================================================== >> --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S 2014-10-02 07:56:23.000000000 +0100 >> +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S 2014-10-02 07:58:10.978958268 +0100 >> @@ -15,7 +15,8 @@ >> >> .text >> .globl main >> -main: .byte 0 >> +main: >> + .dc.l 0 > > We've never used .dl.l before, and I don't know how widely available > it is (in fact, I couldn't find it in the GAS manual). How about using > .word or .4byte? It would also be consistent with what we've usually be > using. Would that work? Agreed. ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests 2014-11-16 18:32 ` Doug Evans @ 2014-11-16 19:49 ` Doug Evans 2014-11-16 20:05 ` Maciej W. Rozycki 0 siblings, 1 reply; 27+ messages in thread From: Doug Evans @ 2014-11-16 19:49 UTC (permalink / raw) To: Joel Brobecker Cc: Maciej W. Rozycki, gdb-patches, Rich Fuhler, Richard Sandiford Doug Evans <xdje42@gmail.com> writes: > On Sun, Nov 16, 2014 at 3:09 AM, Joel Brobecker <brobecker@adacore.com> wrote: >>> Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S >>> =================================================================== >>> --- >>> gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S >>> 2014-10-02 07:56:23.000000000 +0100 >>> +++ >>> gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S >>> 2014-10-02 07:58:10.978958268 +0100 >>> @@ -15,7 +15,8 @@ >>> >>> .text >>> .globl main >>> -main: .byte 0 >>> +main: >>> + .dc.l 0 >> >> We've never used .dl.l before, and I don't know how widely available >> it is (in fact, I couldn't find it in the GAS manual). How about using >> .word or .4byte? It would also be consistent with what we've usually be >> using. Would that work? > > Agreed. Sorry for the followup, but FAOD, .4byte. .word may be a different size on some platforms. ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests 2014-11-16 19:49 ` Doug Evans @ 2014-11-16 20:05 ` Maciej W. Rozycki 2014-11-16 21:52 ` Doug Evans 0 siblings, 1 reply; 27+ messages in thread From: Maciej W. Rozycki @ 2014-11-16 20:05 UTC (permalink / raw) To: Doug Evans; +Cc: Joel Brobecker, gdb-patches, Rich Fuhler, Richard Sandiford On Sun, 16 Nov 2014, Doug Evans wrote: > >>> Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S > >>> =================================================================== > >>> --- > >>> gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S > >>> 2014-10-02 07:56:23.000000000 +0100 > >>> +++ > >>> gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S > >>> 2014-10-02 07:58:10.978958268 +0100 > >>> @@ -15,7 +15,8 @@ > >>> > >>> .text > >>> .globl main > >>> -main: .byte 0 > >>> +main: > >>> + .dc.l 0 > >> > >> We've never used .dl.l before, and I don't know how widely available > >> it is (in fact, I couldn't find it in the GAS manual). How about using > >> .word or .4byte? It would also be consistent with what we've usually be > >> using. Would that work? > > > > Agreed. > > Sorry for the followup, but FAOD, .4byte. > .word may be a different size on some platforms. If anything, that would have to be `.4byte'. I chose `.dc.l' because it is the only fully portable GAS pseudo-op to produce 32-bit data output. Testing portability was the actual reason to add all the `.dc.*' pseudo-ops to GAS. As you've already observed `.word' is unportable, and neither is `.4byte' as the latter is only supported for ELF targets. However in DWARF-2 testing we're probably on an ELF target anyway. So I'll update the tests to use `.4byte'. Thanks for your review. Maciej ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests 2014-11-16 20:05 ` Maciej W. Rozycki @ 2014-11-16 21:52 ` Doug Evans 2014-12-04 0:24 ` Maciej W. Rozycki 0 siblings, 1 reply; 27+ messages in thread From: Doug Evans @ 2014-11-16 21:52 UTC (permalink / raw) To: Maciej W. Rozycki Cc: Joel Brobecker, gdb-patches, Rich Fuhler, Richard Sandiford On Sun, Nov 16, 2014 at 12:05 PM, Maciej W. Rozycki <macro@codesourcery.com> wrote: > On Sun, 16 Nov 2014, Doug Evans wrote: > >> >>> Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S >> >>> =================================================================== >> >>> --- >> >>> gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S >> >>> 2014-10-02 07:56:23.000000000 +0100 >> >>> +++ >> >>> gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-canonicalize-type.S >> >>> 2014-10-02 07:58:10.978958268 +0100 >> >>> @@ -15,7 +15,8 @@ >> >>> >> >>> .text >> >>> .globl main >> >>> -main: .byte 0 >> >>> +main: >> >>> + .dc.l 0 >> >> >> >> We've never used .dl.l before, and I don't know how widely available >> >> it is (in fact, I couldn't find it in the GAS manual). How about using >> >> .word or .4byte? It would also be consistent with what we've usually be >> >> using. Would that work? >> > >> > Agreed. >> >> Sorry for the followup, but FAOD, .4byte. >> .word may be a different size on some platforms. > > If anything, that would have to be `.4byte'. > > I chose `.dc.l' because it is the only fully portable GAS pseudo-op to > produce 32-bit data output. Testing portability was the actual reason > to add all the `.dc.*' pseudo-ops to GAS. As you've already observed > `.word' is unportable, and neither is `.4byte' as the latter is only > supported for ELF targets. However in DWARF-2 testing we're probably on > an ELF target anyway. Yeah. We've been using .4byte for portability in gdb.dwarf2 as long as I can remember. I didn't know about .dc.l, but there's nothing in the name that screams "portability" to me. In fact, I can imagine it causing at least a few people to have to spend time looking it up just to verify it's OK to use (whereas they'd be less inclined to with ".4byte"). Not good. > So I'll update the tests to use `.4byte'. Thanks for your review. Cool. I've got one more review coming. ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests 2014-11-16 21:52 ` Doug Evans @ 2014-12-04 0:24 ` Maciej W. Rozycki 0 siblings, 0 replies; 27+ messages in thread From: Maciej W. Rozycki @ 2014-12-04 0:24 UTC (permalink / raw) To: Doug Evans; +Cc: Joel Brobecker, gdb-patches, Rich Fuhler, Richard Sandiford On Sun, 16 Nov 2014, Doug Evans wrote: > >> Sorry for the followup, but FAOD, .4byte. > >> .word may be a different size on some platforms. > > > > If anything, that would have to be `.4byte'. > > > > I chose `.dc.l' because it is the only fully portable GAS pseudo-op to > > produce 32-bit data output. Testing portability was the actual reason > > to add all the `.dc.*' pseudo-ops to GAS. As you've already observed > > `.word' is unportable, and neither is `.4byte' as the latter is only > > supported for ELF targets. However in DWARF-2 testing we're probably on > > an ELF target anyway. > > Yeah. We've been using .4byte for portability in gdb.dwarf2 as long > as I can remember. > > I didn't know about .dc.l, but there's nothing in the name that screams > "portability" to me. In fact, I can imagine it causing at least a few > people to have > to spend time looking it up just to verify it's OK to use (whereas > they'd be less > inclined to with ".4byte"). Not good. I stuck to the wrong assumption GAS is going to be used with test suite runs, as if it was a part of GDB just as it is a part of binutils. It is not, despite sharing the repository. While maybe less common these days another assembler can be used by the compiler run in testing and `.dc.l' can only be considered portable among GAS's targets and not different assemblers. So we can't really use the pseudo-op, unlike `.4byte' that we already rely on. > > So I'll update the tests to use `.4byte'. Thanks for your review. > > Cool. I've got one more review coming. Applied now, with the update requested and having regression-tested it. Thanks again for the review. Maciej ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests 2014-10-06 0:42 ` [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests Maciej W. Rozycki 2014-11-16 11:09 ` Joel Brobecker @ 2014-11-16 22:28 ` Doug Evans 1 sibling, 0 replies; 27+ messages in thread From: Doug Evans @ 2014-11-16 22:28 UTC (permalink / raw) To: Maciej W. Rozycki Cc: Joel Brobecker, gdb-patches, Rich Fuhler, Richard Sandiford "Maciej W. Rozycki" <macro@codesourcery.com> writes: > Joel, > > Here's the second change, to cover issues triggered by the MIPS ISA bit > handling change, usually in tests that make artificial DWARF-2 records. > > [...] > 2014-10-06 Maciej W. Rozycki <macro@codesourcery.com> > > gdb/testsuite/ > * gdb.cp/expand-psymtabs-cxx.exp: Accept any address of > `method(long)', not just 0x0. > * gdb.cp/nsalias.exp: Align code labels to 4. > * gdb.dwarf2/dw2-canonicalize-type.S (main): Expand to 4-bytes. > * gdb.dwarf2/dw2-empty-pc-range.S (main): Likewise. > * gdb.dwarf2/pr11465.S (_ZN1N1cE): Likewise. > * gdb.dwarf2/dw2-case-insensitive.c (START_INSNS): New macro. > (cu_text_start, FUNC_lang_start): Use `START_INSNS'. > * gdb.dwarf2/dw2-stack-boundary.exp: Accept noise in complaints. > [...] > Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-case-insensitive.c > =================================================================== > --- gdb-fsf-trunk-quilt.orig/gdb/testsuite/gdb.dwarf2/dw2-case-insensitive.c 2014-10-02 07:56:23.000000000 +0100 > +++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.dwarf2/dw2-case-insensitive.c 2014-10-02 07:58:10.978958268 +0100 > @@ -15,13 +15,22 @@ > You should have received a copy of the GNU General Public License > along with this program. If not, see <http://www.gnu.org/licenses/>. */ > > +/* Target-specific way of forcing an instruction label. */ > +#ifdef __mips__ > +#define START_INSNS asm (".insn"); > +#else > +#define START_INSNS > +#endif > + > /* Use DW_LANG_Fortran90 for case insensitive DWARF. */ > asm (".globl cu_text_start"); > asm ("cu_text_start:"); > +START_INSNS > > asm (".globl FUNC_lang_start"); > asm (".p2align 4"); > asm ("FUNC_lang_start:"); > +START_INSNS > > void > FUNC_lang (void) Hi. Just a note to say this will probably not work with clang, but this test is already clang unfriendly. We can leave fixing that for another day. I have no other comments on the part of the patch set, so ok by me (with the .4byte change). [Still working on part 1.] ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-10-06 0:42 ` [PATCH v2 1/2] " Maciej W. Rozycki 2014-10-06 0:42 ` [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests Maciej W. Rozycki @ 2014-10-06 14:10 ` Joel Brobecker 2014-10-14 20:45 ` Maciej W. Rozycki 2014-11-16 19:23 ` [PATCH " Doug Evans 2014-10-06 15:43 ` Maciej W. Rozycki ` (3 subsequent siblings) 5 siblings, 2 replies; 27+ messages in thread From: Joel Brobecker @ 2014-10-06 14:10 UTC (permalink / raw) To: Maciej W. Rozycki; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford Hello everyone, On Mon, Oct 06, 2014 at 01:41:50AM +0100, Maciej W. Rozycki wrote: > Joel, > > I'd like to get back to this review, this change is important for > reliable debugging compressed code, as test results quoted below indicate. I am going to need help to get this patch reviewed. I don't think I'll have the availability to give it the attention it deserves for at least a week or two :-(. Any takers? Otherwise, I'll keep it at the top of my gdb-patches list. Thanks! -- Joel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-10-06 14:10 ` [PATCH v2 1/2] ISA bit treatment on the MIPS platform Joel Brobecker @ 2014-10-14 20:45 ` Maciej W. Rozycki 2014-10-20 17:10 ` [PING][PATCH " Maciej W. Rozycki 2014-11-16 19:23 ` [PATCH " Doug Evans 1 sibling, 1 reply; 27+ messages in thread From: Maciej W. Rozycki @ 2014-10-14 20:45 UTC (permalink / raw) To: Joel Brobecker; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford On Mon, 6 Oct 2014, Joel Brobecker wrote: > > I'd like to get back to this review, this change is important for > > reliable debugging compressed code, as test results quoted below indicate. > > I am going to need help to get this patch reviewed. I don't think > I'll have the availability to give it the attention it deserves > for at least a week or two :-(. Any takers? Otherwise, I'll keep it > at the top of my gdb-patches list. Thanks; please note that this change relies on http://sourceware.org/ml/gdb-patches/2014-09/msg00805.html to have been applied first. Although I can swap the order of the patches I suppose, the dependency is purely syntactical. Also I take back the last-minute update I sent as a follow-up. While there's a loop, the call to `gdbarch_elf_make_msymbol_special_p' will be made only once, even if placed within. This situation just tells me to avoid last-minute updates, sorry about the confusion. Maciej ^ permalink raw reply [flat|nested] 27+ messages in thread
* [PING][PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-10-14 20:45 ` Maciej W. Rozycki @ 2014-10-20 17:10 ` Maciej W. Rozycki 2014-11-03 16:13 ` [PING^2][PATCH " Maciej W. Rozycki 0 siblings, 1 reply; 27+ messages in thread From: Maciej W. Rozycki @ 2014-10-20 17:10 UTC (permalink / raw) To: Joel Brobecker; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford Joel, > > > I'd like to get back to this review, this change is important for > > > reliable debugging compressed code, as test results quoted below indicate. > > > > I am going to need help to get this patch reviewed. I don't think > > I'll have the availability to give it the attention it deserves > > for at least a week or two :-(. Any takers? Otherwise, I'll keep it > > at the top of my gdb-patches list. > > Thanks; please note that this change relies on > http://sourceware.org/ml/gdb-patches/2014-09/msg00805.html to have been > applied first. Although I can swap the order of the patches I suppose, > the dependency is purely syntactical. Can you please update me with your ETA to review this change? Thanks, Maciej ^ permalink raw reply [flat|nested] 27+ messages in thread
* [PING^2][PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-10-20 17:10 ` [PING][PATCH " Maciej W. Rozycki @ 2014-11-03 16:13 ` Maciej W. Rozycki 0 siblings, 0 replies; 27+ messages in thread From: Maciej W. Rozycki @ 2014-11-03 16:13 UTC (permalink / raw) To: Joel Brobecker; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford Joel, > > > > I'd like to get back to this review, this change is important for > > > > reliable debugging compressed code, as test results quoted below indicate. > > > > > > I am going to need help to get this patch reviewed. I don't think > > > I'll have the availability to give it the attention it deserves > > > for at least a week or two :-(. Any takers? Otherwise, I'll keep it > > > at the top of my gdb-patches list. > > > > Thanks; please note that this change relies on > > http://sourceware.org/ml/gdb-patches/2014-09/msg00805.html to have been > > applied first. Although I can swap the order of the patches I suppose, > > the dependency is purely syntactical. > > Can you please update me with your ETA to review this change? With 7.8.1 out of the door will you be able to get at it soon? I will appreciate that as I have more MIPS stuff to follow that touches generic code, in particular bits for PR gdb/7518 that are not necessarily straightforward, and I don't want it all to pile up (as if it hasn't already, sigh...). Thanks, Maciej ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-10-06 14:10 ` [PATCH v2 1/2] ISA bit treatment on the MIPS platform Joel Brobecker 2014-10-14 20:45 ` Maciej W. Rozycki @ 2014-11-16 19:23 ` Doug Evans 1 sibling, 0 replies; 27+ messages in thread From: Doug Evans @ 2014-11-16 19:23 UTC (permalink / raw) To: Joel Brobecker Cc: Maciej W. Rozycki, gdb-patches, Rich Fuhler, Richard Sandiford Joel Brobecker <brobecker@adacore.com> writes: > Hello everyone, > > On Mon, Oct 06, 2014 at 01:41:50AM +0100, Maciej W. Rozycki wrote: >> Joel, >> >> I'd like to get back to this review, this change is important for >> reliable debugging compressed code, as test results quoted below indicate. > > I am going to need help to get this patch reviewed. I don't think > I'll have the availability to give it the attention it deserves > for at least a week or two :-(. Any takers? Otherwise, I'll keep it > at the top of my gdb-patches list. For future reference, I didn't see this until now. Please feel free to include me directly in the To list for such requests. Especially *anything* affecting symbols and the dwarf reader. Cheers. ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-10-06 0:42 ` [PATCH v2 1/2] " Maciej W. Rozycki 2014-10-06 0:42 ` [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests Maciej W. Rozycki 2014-10-06 14:10 ` [PATCH v2 1/2] ISA bit treatment on the MIPS platform Joel Brobecker @ 2014-10-06 15:43 ` Maciej W. Rozycki 2014-11-16 10:37 ` Joel Brobecker ` (2 subsequent siblings) 5 siblings, 0 replies; 27+ messages in thread From: Maciej W. Rozycki @ 2014-10-06 15:43 UTC (permalink / raw) To: Joel Brobecker; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford On Mon, 6 Oct 2014, Maciej W. Rozycki wrote: > Index: gdb-fsf-trunk-quilt/gdb/solib.c > =================================================================== > --- gdb-fsf-trunk-quilt.orig/gdb/solib.c 2014-10-03 13:52:46.000000000 +0100 > +++ gdb-fsf-trunk-quilt/gdb/solib.c 2014-10-03 14:50:26.398945943 +0100 > @@ -1444,8 +1444,28 @@ gdb_bfd_lookup_symbol_from_symtab (bfd * > > if (match_sym (sym, data)) > { > + struct gdbarch *gdbarch = target_gdbarch (); > + symaddr = sym->value; > + > + /* Some ELF targets fiddle with addresses of symbols they > + consider special. They use minimal symbols to do that > + and this is needed for correct breakpoint placement, > + but we do not have full data here to build a complete > + minimal symbol, so just set the address and let the > + targets cope with that. */ > + if (bfd_get_flavour (abfd) == bfd_target_elf_flavour > + && gdbarch_elf_make_msymbol_special_p (gdbarch)) > + { > + struct minimal_symbol msym; > + > + memset (&msym, 0, sizeof (msym)); > + SET_MSYMBOL_VALUE_ADDRESS (&msym, symaddr); > + gdbarch_elf_make_msymbol_special (gdbarch, sym, &msym); > + symaddr = MSYMBOL_VALUE_RAW_ADDRESS (&msym); > + } > + > /* BFD symbols are section relative. */ > - symaddr = sym->value + sym->section->vma; > + symaddr += sym->section->vma; > break; > } > } I realised this part can be optimised similarly to `elf_symtab_read' in gdb/elfread.c so that `gdbarch_elf_make_msymbol_special_p' is called once per function call rather than once per symbol and I have now applied this update that I intend to consider a part of the final change. Maciej Index: gdb-fsf-trunk-quilt/gdb/solib.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/solib.c 2014-10-06 16:36:47.867795479 +0100 +++ gdb-fsf-trunk-quilt/gdb/solib.c 2014-10-06 16:36:25.368036677 +0100 @@ -1437,6 +1437,9 @@ gdb_bfd_lookup_symbol_from_symtab (bfd * struct cleanup *back_to = make_cleanup (xfree, symbol_table); unsigned int number_of_symbols = bfd_canonicalize_symtab (abfd, symbol_table); + struct gdbarch *gdbarch = target_gdbarch (); + int elf_make_msymbol_special_p + = gdbarch_elf_make_msymbol_special_p (gdbarch); for (i = 0; i < number_of_symbols; i++) { @@ -1444,7 +1447,6 @@ gdb_bfd_lookup_symbol_from_symtab (bfd * if (match_sym (sym, data)) { - struct gdbarch *gdbarch = target_gdbarch (); symaddr = sym->value; /* Some ELF targets fiddle with addresses of symbols they @@ -1454,7 +1456,7 @@ gdb_bfd_lookup_symbol_from_symtab (bfd * minimal symbol, so just set the address and let the targets cope with that. */ if (bfd_get_flavour (abfd) == bfd_target_elf_flavour - && gdbarch_elf_make_msymbol_special_p (gdbarch)) + && elf_make_msymbol_special_p) { struct minimal_symbol msym; ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-10-06 0:42 ` [PATCH v2 1/2] " Maciej W. Rozycki ` (2 preceding siblings ...) 2014-10-06 15:43 ` Maciej W. Rozycki @ 2014-11-16 10:37 ` Joel Brobecker 2014-11-16 19:27 ` Doug Evans 2014-12-04 23:14 ` Maciej W. Rozycki 2014-11-17 1:17 ` Doug Evans 2014-12-12 16:38 ` [committed] MIPS: Define aliases for MSYMBOL_TARGET_FLAG macros Maciej W. Rozycki 5 siblings, 2 replies; 27+ messages in thread From: Joel Brobecker @ 2014-11-16 10:37 UTC (permalink / raw) To: Maciej W. Rozycki; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford > 2014-10-06 Maciej W. Rozycki <macro@codesourcery.com> > Maciej W. Rozycki <macro@mips.com> > Pedro Alves <pedro@codesourcery.com> > > * gdbarch.sh (elf_make_msymbol_special): Change type to `F', > remove `predefault' and `invalid_p' initializers. > (make_symbol_special): New architecture method. > (adjust_dwarf2_addr, adjust_dwarf2_line): Likewise. > (objfile, symbol): New declarations. > * arch-utils.h (default_elf_make_msymbol_special): Remove > prototype. > (default_make_symbol_special): New prototype. > (default_adjust_dwarf2_addr): Likewise. > (default_adjust_dwarf2_line): Likewise. > * mips-tdep.h (mips_unmake_compact_addr): New prototype. > * arch-utils.c (default_elf_make_msymbol_special): Remove > function. > (default_make_symbol_special): New function. > (default_adjust_dwarf2_addr): Likewise. > (default_adjust_dwarf2_line): Likewise. > * dwarf2-frame.c (decode_frame_entry_1): Call > `gdbarch_adjust_dwarf2_addr'. > * dwarf2loc.c (dwarf2_find_location_expression): Likewise. > * dwarf2read.c (create_addrmap_from_index): Likewise. > (process_psymtab_comp_unit_reader): Likewise. > (add_partial_symbol): Likewise. > (add_partial_subprogram): Likewise. > (process_full_comp_unit): Likewise. > (read_file_scope): Likewise. > (read_func_scope): Likewise. Call `gdbarch_make_symbol_special'. > (read_lexical_block_scope): Call `gdbarch_adjust_dwarf2_addr'. > (read_call_site_scope): Likewise. > (dwarf2_ranges_read): Likewise. > (dwarf2_record_block_ranges): Likewise. > (read_attribute_value): Likewise. > (dwarf_decode_lines_1): Call `gdbarch_adjust_dwarf2_line'. > (new_symbol_full): Call `gdbarch_adjust_dwarf2_addr'. > * elfread.c (elf_symtab_read): Don't call > `gdbarch_elf_make_msymbol_special' if unset. > * mips-linux-tdep.c (micromips_linux_sigframe_validate): Strip > the ISA bit from the PC. > * mips-tdep.c (mips_unmake_compact_addr): New function. > (mips_elf_make_msymbol_special): Set the ISA bit in the symbol's > address appropriately. > (mips_make_symbol_special): New function. > (mips_pc_is_mips): Set the ISA bit before symbol lookup. > (mips_pc_is_mips16): Likewise. > (mips_pc_is_micromips): Likewise. > (mips_pc_isa): Likewise. > (mips_adjust_dwarf2_addr): New function. > (mips_adjust_dwarf2_line): Likewise. > (mips_read_pc, mips_unwind_pc): Keep the ISA bit. > (mips_addr_bits_remove): Likewise. > (mips_skip_trampoline_code): Likewise. > (mips_write_pc): Don't set the ISA bit. > (mips_eabi_push_dummy_call): Likewise. > (mips_o64_push_dummy_call): Likewise. > (mips_gdbarch_init): Install `mips_make_symbol_special', > `mips_adjust_dwarf2_addr' and `mips_adjust_dwarf2_line' gdbarch > handlers. > * solib.c (gdb_bfd_lookup_symbol_from_symtab): Get > target-specific symbol address adjustments. > * gdbarch.h: Regenerate. > * gdbarch.c: Regenerate. > > 2014-10-06 Maciej W. Rozycki <macro@codesourcery.com> > > gdb/testsuite/ > * gdb.base/func-ptrs.c: New file. > * gdb.base/func-ptrs.exp: New file. Pre-approved, with the on minor documentation request below. > Index: gdb-fsf-trunk-quilt/gdb/gdbarch.sh > =================================================================== > --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.sh 2014-10-03 13:52:46.000000000 +0100 > +++ gdb-fsf-trunk-quilt/gdb/gdbarch.sh 2014-10-03 14:50:26.000000000 +0100 > @@ -635,8 +635,11 @@ m:int:in_solib_return_trampoline:CORE_AD > # which don't suffer from that problem could just let this functionality > # untouched. > m:int:in_function_epilogue_p:CORE_ADDR addr:addr:0:generic_in_function_epilogue_p::0 > -f:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym::default_elf_make_msymbol_special::0 > +F:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym > f:void:coff_make_msymbol_special:int val, struct minimal_symbol *msym:val, msym::default_coff_make_msymbol_special::0 > +f:void:make_symbol_special:struct symbol *sym, struct objfile *objfile:sym, objfile::default_make_symbol_special::0 > +f:CORE_ADDR:adjust_dwarf2_addr:CORE_ADDR pc:pc::default_adjust_dwarf2_addr::0 > +f:CORE_ADDR:adjust_dwarf2_line:CORE_ADDR addr, int rel:addr, rel::default_adjust_dwarf2_line::0 I know we haven't been all that good in the past at documenting gdbarch methods, but new entries should be fully documented. That way, all arch-specific implementations can then document with a reference to the documentation in gdbarch.sh/gdbarch.h. Would you mind documenting the ones you're adding? This is not a requirement for your patch, but if you happen to be able to quickly document elf_make_msymbol_special as well, that would be a very welcome and appreciated change. > +/* Recalculate the line record requested so that the resulting PC has the > + ISA bit set correctly, used by DWARF-2 machinery. */ > + > +static CORE_ADDR > +mips_adjust_dwarf2_line (CORE_ADDR addr, int rel) > +{ > + static CORE_ADDR adj_pc; > + static CORE_ADDR pc; > + CORE_ADDR isa_pc; > + > + pc = rel ? pc + addr : addr; > + isa_pc = mips_adjust_dwarf2_addr (pc); > + addr = rel ? isa_pc - adj_pc : isa_pc; > + adj_pc = isa_pc; > + return addr; > +} I got to stare that his code for quite a while, because of the use of the private static variables, which means that the result of a call to that function depends on the previous call. I am guessing that this is the reason behind the following hunk earlier in the patch: - CORE_ADDR address = 0; + CORE_ADDR address = gdbarch_adjust_dwarf2_line (gdbarch, 0, 0); (to initialize the context). I think we all know that if makes the implementation a little harder to maintain in the long run, but after thinking over it for a while, I tend to like the fact that this helps making the chhanges to dwarf2read.c a little simpler, with negative impact being only on the MIPS target. So while I wish things were perfect, I'm have this feeling that the above is a "good deal" for core GDB ;-). > +++ gdb-fsf-trunk-quilt/gdb/solib.c 2014-10-03 14:50:26.398945943 +0100 > @@ -1444,8 +1444,28 @@ gdb_bfd_lookup_symbol_from_symtab (bfd * > > if (match_sym (sym, data)) > { > + struct gdbarch *gdbarch = target_gdbarch (); > + symaddr = sym->value; > + > + /* Some ELF targets fiddle with addresses of symbols they > + consider special. They use minimal symbols to do that > + and this is needed for correct breakpoint placement, > + but we do not have full data here to build a complete > + minimal symbol, so just set the address and let the > + targets cope with that. */ > + if (bfd_get_flavour (abfd) == bfd_target_elf_flavour > + && gdbarch_elf_make_msymbol_special_p (gdbarch)) > + { > + struct minimal_symbol msym; > + > + memset (&msym, 0, sizeof (msym)); > + SET_MSYMBOL_VALUE_ADDRESS (&msym, symaddr); > + gdbarch_elf_make_msymbol_special (gdbarch, sym, &msym); > + symaddr = MSYMBOL_VALUE_RAW_ADDRESS (&msym); > + } > + > /* BFD symbols are section relative. */ > - symaddr = sym->value + sym->section->vma; > + symaddr += sym->section->vma; > break; FTR: I think I had some reservations about this part of the code, and you improved it to only do the work if gdbarch_elf_make_msymbol_special_p, which is a nice improvement anyway. But looking at it more closely, I probably missed at the time the fact that the minimal_symbol has a local scope. (duh!) Thank you, -- Joel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-11-16 10:37 ` Joel Brobecker @ 2014-11-16 19:27 ` Doug Evans 2014-12-04 23:14 ` Maciej W. Rozycki 1 sibling, 0 replies; 27+ messages in thread From: Doug Evans @ 2014-11-16 19:27 UTC (permalink / raw) To: Joel Brobecker, Maciej W. Rozycki Cc: gdb-patches, Rich Fuhler, Richard Sandiford Joel Brobecker <brobecker@adacore.com> writes: > Pre-approved, with the on minor documentation request below. For reference sake, I've got some comments too. Please hold off checking this in until I can finish reading the patch (which I'm in the middle of and should finish today). ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-11-16 10:37 ` Joel Brobecker 2014-11-16 19:27 ` Doug Evans @ 2014-12-04 23:14 ` Maciej W. Rozycki 2014-12-12 14:00 ` Maciej W. Rozycki 1 sibling, 1 reply; 27+ messages in thread From: Maciej W. Rozycki @ 2014-12-04 23:14 UTC (permalink / raw) To: Joel Brobecker, Doug Evans; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford Joel, Doug -- On Sun, 16 Nov 2014, Joel Brobecker wrote: > > Index: gdb-fsf-trunk-quilt/gdb/gdbarch.sh > > =================================================================== > > --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.sh 2014-10-03 13:52:46.000000000 +0100 > > +++ gdb-fsf-trunk-quilt/gdb/gdbarch.sh 2014-10-03 14:50:26.000000000 +0100 > > @@ -635,8 +635,11 @@ m:int:in_solib_return_trampoline:CORE_AD > > # which don't suffer from that problem could just let this functionality > > # untouched. > > m:int:in_function_epilogue_p:CORE_ADDR addr:addr:0:generic_in_function_epilogue_p::0 > > -f:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym::default_elf_make_msymbol_special::0 > > +F:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym > > f:void:coff_make_msymbol_special:int val, struct minimal_symbol *msym:val, msym::default_coff_make_msymbol_special::0 > > +f:void:make_symbol_special:struct symbol *sym, struct objfile *objfile:sym, objfile::default_make_symbol_special::0 > > +f:CORE_ADDR:adjust_dwarf2_addr:CORE_ADDR pc:pc::default_adjust_dwarf2_addr::0 > > +f:CORE_ADDR:adjust_dwarf2_line:CORE_ADDR addr, int rel:addr, rel::default_adjust_dwarf2_line::0 > > I know we haven't been all that good in the past at documenting > gdbarch methods, but new entries should be fully documented. That way, > all arch-specific implementations can then document with a reference > to the documentation in gdbarch.sh/gdbarch.h. Would you mind documenting > the ones you're adding? > > This is not a requirement for your patch, but if you happen to be able > to quickly document elf_make_msymbol_special as well, that would be > a very welcome and appreciated change. I have now added these descriptions, with some bias towards the MIPS specifics as they are what I am most familiar with. Please let me know if you think they might be further improved. > > +/* Recalculate the line record requested so that the resulting PC has the > > + ISA bit set correctly, used by DWARF-2 machinery. */ > > + > > +static CORE_ADDR > > +mips_adjust_dwarf2_line (CORE_ADDR addr, int rel) > > +{ > > + static CORE_ADDR adj_pc; > > + static CORE_ADDR pc; > > + CORE_ADDR isa_pc; > > + > > + pc = rel ? pc + addr : addr; > > + isa_pc = mips_adjust_dwarf2_addr (pc); > > + addr = rel ? isa_pc - adj_pc : isa_pc; > > + adj_pc = isa_pc; > > + return addr; > > +} > > I got to stare that his code for quite a while, because of the use > of the private static variables, which means that the result of > a call to that function depends on the previous call. I am guessing > that this is the reason behind the following hunk earlier in the patch: > > - CORE_ADDR address = 0; > + CORE_ADDR address = gdbarch_adjust_dwarf2_line (gdbarch, 0, 0); > > (to initialize the context). Correct, this simulates an implicit line entry at the start of processing so that a possible initial relative address adjustment line information entry works correctly. > I think we all know that if makes the implementation a little harder > to maintain in the long run, but after thinking over it for a while, > I tend to like the fact that this helps making the chhanges to > dwarf2read.c a little simpler, with negative impact being only on > the MIPS target. So while I wish things were perfect, I'm have this > feeling that the above is a "good deal" for core GDB ;-). Yes, as I already noted in my other reply today made to a similar concern Doug made, I concluded making it as simple as possible and putting the most burden on the MIPS backend will incur the least amount of long maintenance effort. Indeed, having persistent state kept in local static variables is a bit obscure, but I think this is a "good deal", as you say. > > +++ gdb-fsf-trunk-quilt/gdb/solib.c 2014-10-03 14:50:26.398945943 +0100 > > @@ -1444,8 +1444,28 @@ gdb_bfd_lookup_symbol_from_symtab (bfd * > > > > if (match_sym (sym, data)) > > { > > + struct gdbarch *gdbarch = target_gdbarch (); > > + symaddr = sym->value; > > + > > + /* Some ELF targets fiddle with addresses of symbols they > > + consider special. They use minimal symbols to do that > > + and this is needed for correct breakpoint placement, > > + but we do not have full data here to build a complete > > + minimal symbol, so just set the address and let the > > + targets cope with that. */ > > + if (bfd_get_flavour (abfd) == bfd_target_elf_flavour > > + && gdbarch_elf_make_msymbol_special_p (gdbarch)) > > + { > > + struct minimal_symbol msym; > > + > > + memset (&msym, 0, sizeof (msym)); > > + SET_MSYMBOL_VALUE_ADDRESS (&msym, symaddr); > > + gdbarch_elf_make_msymbol_special (gdbarch, sym, &msym); > > + symaddr = MSYMBOL_VALUE_RAW_ADDRESS (&msym); > > + } > > + > > /* BFD symbols are section relative. */ > > - symaddr = sym->value + sym->section->vma; > > + symaddr += sym->section->vma; > > break; > > FTR: I think I had some reservations about this part of the code, > and you improved it to only do the work if > gdbarch_elf_make_msymbol_special_p, which is a nice improvement > anyway. But looking at it more closely, I probably missed at the time > the fact that the minimal_symbol has a local scope. (duh!) I thought you actually were (quite validly IMO) concerned about all the preparatory steps made across all targets to set up `msym', just to do nothing about it then. So here are the changes I made to address the concerns you both expressed, as an incremental patch for easier review. If this looks good to you, then I'll fold it into the original change and commit (the comment addition for `elf_make_msymbol_special' may better go in as a separate change though). Thanks for your input. Maciej gdb-mips16-isa-bit-update.diff Index: gdb-fsf-trunk-quilt/gdb/dwarf2read.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/dwarf2read.c 2014-12-04 13:55:33.000000000 +0000 +++ gdb-fsf-trunk-quilt/gdb/dwarf2read.c 2014-12-04 18:22:26.428924108 +0000 @@ -17323,7 +17323,11 @@ dwarf_decode_lines_1 (struct line_header /* Read the statement sequences until there's nothing left. */ while (line_ptr < line_end) { - /* state machine registers */ + /* State machine registers. Call `gdbarch_adjust_dwarf2_line' + on the initial 0 address as if there was a line entry for it + so that the backend has a chance to adjust it and also record + it in case it needs it. This is currently used by MIPS code, + cf. `mips_adjust_dwarf2_line'. */ CORE_ADDR address = gdbarch_adjust_dwarf2_line (gdbarch, 0, 0); unsigned int file = 1; unsigned int line = 1; Index: gdb-fsf-trunk-quilt/gdb/gdbarch.h =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.h 2014-12-04 13:55:33.000000000 +0000 +++ gdb-fsf-trunk-quilt/gdb/gdbarch.h 2014-12-04 21:34:44.678927696 +0000 @@ -690,6 +690,14 @@ typedef int (gdbarch_in_function_epilogu extern int gdbarch_in_function_epilogue_p (struct gdbarch *gdbarch, CORE_ADDR addr); extern void set_gdbarch_in_function_epilogue_p (struct gdbarch *gdbarch, gdbarch_in_function_epilogue_p_ftype *in_function_epilogue_p); +/* Process an ELF symbol in the minimal symbol table in a backend-specific + way. Normally this hook is supposed to do nothing, however if required, + then this hook can be used to apply tranformations to symbols that are + considered special in some way. For example the MIPS backend uses it + to interpret `st_other' information to mark compressed code symbols so + that they can be treated in the appropriate manner in the processing of + the main symbol table and DWARF-2 records. */ + extern int gdbarch_elf_make_msymbol_special_p (struct gdbarch *gdbarch); typedef void (gdbarch_elf_make_msymbol_special_ftype) (asymbol *sym, struct minimal_symbol *msym); @@ -700,14 +708,41 @@ typedef void (gdbarch_coff_make_msymbol_ extern void gdbarch_coff_make_msymbol_special (struct gdbarch *gdbarch, int val, struct minimal_symbol *msym); extern void set_gdbarch_coff_make_msymbol_special (struct gdbarch *gdbarch, gdbarch_coff_make_msymbol_special_ftype *coff_make_msymbol_special); +/* Process a symbol in the main symbol table in a backend-specific way. + Normally this hook is supposed to do nothing, however if required, + then this hook can be used to apply tranformations to symbols that + are considered special in some way. This is currently used by the + MIPS backend to make sure compressed code symbols have the ISA bit + set. This in turn is needed for symbol values seen in GDB to match + the values used at the runtime by the program itself, for function + and label references. */ + typedef void (gdbarch_make_symbol_special_ftype) (struct symbol *sym, struct objfile *objfile); extern void gdbarch_make_symbol_special (struct gdbarch *gdbarch, struct symbol *sym, struct objfile *objfile); extern void set_gdbarch_make_symbol_special (struct gdbarch *gdbarch, gdbarch_make_symbol_special_ftype *make_symbol_special); +/* Adjust the address retrieved from a DWARF-2 record other than a line + entry in a backend-specific way. Normally this hook is supposed to + return the address passed unchanged, however if that is incorrect for + any reason, then this hook can be used to fix the address up in the + required manner. This is currently used by the MIPS backend to make + sure addresses in FDE, range records, etc. referring to compressed + code have the ISA bit set, matching line information and the symbol + table. */ + typedef CORE_ADDR (gdbarch_adjust_dwarf2_addr_ftype) (CORE_ADDR pc); extern CORE_ADDR gdbarch_adjust_dwarf2_addr (struct gdbarch *gdbarch, CORE_ADDR pc); extern void set_gdbarch_adjust_dwarf2_addr (struct gdbarch *gdbarch, gdbarch_adjust_dwarf2_addr_ftype *adjust_dwarf2_addr); +/* Adjust the address updated by a line entry in a backend-specific way. + Normally this hook is supposed to return the address passed unchanged, + however in the case of inconsistencies in these records, this hook can + be used to fix them up in the required manner. This is currently used + by the MIPS backend to make sure all line addresses in compressed code + are presented with the ISA bit set, which is not always the case. This + in turn ensures breakpoint addresses are correctly matched against the + stop PC. */ + typedef CORE_ADDR (gdbarch_adjust_dwarf2_line_ftype) (CORE_ADDR addr, int rel); extern CORE_ADDR gdbarch_adjust_dwarf2_line (struct gdbarch *gdbarch, CORE_ADDR addr, int rel); extern void set_gdbarch_adjust_dwarf2_line (struct gdbarch *gdbarch, gdbarch_adjust_dwarf2_line_ftype *adjust_dwarf2_line); Index: gdb-fsf-trunk-quilt/gdb/gdbarch.sh =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.sh 2014-12-04 13:55:33.000000000 +0000 +++ gdb-fsf-trunk-quilt/gdb/gdbarch.sh 2014-12-04 21:34:21.168924205 +0000 @@ -635,10 +635,41 @@ m:int:in_solib_return_trampoline:CORE_AD # which don't suffer from that problem could just let this functionality # untouched. m:int:in_function_epilogue_p:CORE_ADDR addr:addr:0:generic_in_function_epilogue_p::0 +# Process an ELF symbol in the minimal symbol table in a backend-specific +# way. Normally this hook is supposed to do nothing, however if required, +# then this hook can be used to apply tranformations to symbols that are +# considered special in some way. For example the MIPS backend uses it +# to interpret \`st_other' information to mark compressed code symbols so +# that they can be treated in the appropriate manner in the processing of +# the main symbol table and DWARF-2 records. F:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym f:void:coff_make_msymbol_special:int val, struct minimal_symbol *msym:val, msym::default_coff_make_msymbol_special::0 +# Process a symbol in the main symbol table in a backend-specific way. +# Normally this hook is supposed to do nothing, however if required, +# then this hook can be used to apply tranformations to symbols that +# are considered special in some way. This is currently used by the +# MIPS backend to make sure compressed code symbols have the ISA bit +# set. This in turn is needed for symbol values seen in GDB to match +# the values used at the runtime by the program itself, for function +# and label references. f:void:make_symbol_special:struct symbol *sym, struct objfile *objfile:sym, objfile::default_make_symbol_special::0 +# Adjust the address retrieved from a DWARF-2 record other than a line +# entry in a backend-specific way. Normally this hook is supposed to +# return the address passed unchanged, however if that is incorrect for +# any reason, then this hook can be used to fix the address up in the +# required manner. This is currently used by the MIPS backend to make +# sure addresses in FDE, range records, etc. referring to compressed +# code have the ISA bit set, matching line information and the symbol +# table. f:CORE_ADDR:adjust_dwarf2_addr:CORE_ADDR pc:pc::default_adjust_dwarf2_addr::0 +# Adjust the address updated by a line entry in a backend-specific way. +# Normally this hook is supposed to return the address passed unchanged, +# however in the case of inconsistencies in these records, this hook can +# be used to fix them up in the required manner. This is currently used +# by the MIPS backend to make sure all line addresses in compressed code +# are presented with the ISA bit set, which is not always the case. This +# in turn ensures breakpoint addresses are correctly matched against the +# stop PC. f:CORE_ADDR:adjust_dwarf2_line:CORE_ADDR addr, int rel:addr, rel::default_adjust_dwarf2_line::0 v:int:cannot_step_breakpoint:::0:0::0 v:int:have_nonsteppable_watchpoint:::0:0::0 Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.c 2014-12-04 13:55:33.000000000 +0000 +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.c 2014-12-04 22:25:22.498924888 +0000 @@ -414,23 +414,29 @@ msymbol_is_micromips (struct minimal_sym } /* Set the ISA bit in the main symbol too, complementing the corresponding - minimal symbol setting and reflecting the run-time value of the symbol. */ + minimal symbol setting and reflecting the run-time value of the symbol. + The need for comes from the ISA bit having been cleared as code in + `_bfd_mips_elf_symbol_processing' separated it into the ELF symbol's + `st_other' STO_MIPS16 or STO_MICROMIPS annotation, making the values + of symbols referring to compressed code different in GDB to the values + used by actual code. That in turn makes them evaluate incorrectly in + expressions, producing results different to what the same expressions + yield when compiled into the program being debugged. */ static void mips_make_symbol_special (struct symbol *sym, struct objfile *objfile) { if (SYMBOL_CLASS (sym) == LOC_BLOCK) { + /* We are in symbol reading so it is OK to cast away constness. */ + struct block *block = (struct block *) SYMBOL_BLOCK_VALUE (sym); CORE_ADDR compact_block_start; struct bound_minimal_symbol msym; - compact_block_start = BLOCK_START (SYMBOL_BLOCK_VALUE (sym)) | 1; + compact_block_start = BLOCK_START (block) | 1; msym = lookup_minimal_symbol_by_pc (compact_block_start); if (msym.minsym && !msymbol_is_mips (msym.minsym)) { - /* We are in symbol reading so it is OK to cast away constness. */ - struct block *block = (struct block *) SYMBOL_BLOCK_VALUE (sym); - BLOCK_START (block) = compact_block_start; } } @@ -1248,7 +1254,11 @@ mips_pc_isa (struct gdbarch *gdbarch, CO } } -/* Set the ISA bit correctly in the PC, used by DWARF-2 machinery. */ +/* Set the ISA bit correctly in the PC, used by DWARF-2 machinery. + The need for comes from the ISA bit having been cleared, making + addresses in FDE, range records, etc. referring to compressed code + different to those in line information, the symbol table and finally + the PC register. That in turn confuses many operations. */ static CORE_ADDR mips_adjust_dwarf2_addr (CORE_ADDR pc) @@ -1257,8 +1267,39 @@ mips_adjust_dwarf2_addr (CORE_ADDR pc) return mips_pc_is_mips (pc) ? pc : make_compact_addr (pc); } -/* Recalculate the line record requested so that the resulting PC has the - ISA bit set correctly, used by DWARF-2 machinery. */ +/* Recalculate the line record requested so that the resulting PC has + the ISA bit set correctly, used by DWARF-2 machinery. The need for + this adjustment comes from some records associated with compressed + code having the ISA bit cleared, most notably at function prologue + ends. The ISA bit is in this context retrieved from the minimal + symbol covering the address requested, which in turn has been + constructed from the binary's symbol table rather than DWARF-2 + information. The correct setting of the ISA bit is required for + breakpoint addresses to correctly match against the stop PC. + + As line entries can specify relative address adjustments we need to + keep track of the absolute value of the last line address recorded + in line information, so that we can calculate the actual address to + apply the ISA bit adjustment to. We use PC for this tracking and + keep the original address there. + + As such relative address adjustments can be odd within compressed + code we need to keep track of the last line address with the ISA + bit adjustment applied too, as the original address may or may not + have had the ISA bit set. We use ADJ_PC for this tracking and keep + the adjusted address there. + + For relative address adjustments we then use these variables to + calculate the address intended by line information, which will be + PC-relative, and return an updated adjustment carrying ISA bit + information, which will be ADJ_PC-relative. For absolute address + adjustments we just return the same address that we store in ADJ_PC + too. + + As the first line entry can be relative to an implied address value + of 0 we need to have the initial address set up that we store in PC + and ADJ_PC. This is arranged with a call from `dwarf_decode_lines_1' + that sets PC to 0 and ADJ_PC accordingly, usually 0 as well. */ static CORE_ADDR mips_adjust_dwarf2_line (CORE_ADDR addr, int rel) ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-12-04 23:14 ` Maciej W. Rozycki @ 2014-12-12 14:00 ` Maciej W. Rozycki 2014-12-12 17:22 ` Doug Evans 0 siblings, 1 reply; 27+ messages in thread From: Maciej W. Rozycki @ 2014-12-12 14:00 UTC (permalink / raw) To: Joel Brobecker, Doug Evans; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford Joel, Doug -- On Thu, 4 Dec 2014, Maciej W. Rozycki wrote: > > This is not a requirement for your patch, but if you happen to be able > > to quickly document elf_make_msymbol_special as well, that would be > > a very welcome and appreciated change. > > I have now added these descriptions, with some bias towards the MIPS > specifics as they are what I am most familiar with. Please let me know > if you think they might be further improved. I haven't heard back from you, so I have applied the change now. While reviewing the final ChangeLog entry I discovered I created an unneeded shadow local variable definition, so in addition to the changes already posted in this thread I have included the following cleanup, as obvious, in the final commit. Thanks for your review. Maciej gdb-mips16-isa-bit-cleanup.diff Index: gdb-fsf-trunk-quilt/gdb/dwarf2read.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/dwarf2read.c 2014-12-11 23:54:46.000000000 +0000 +++ gdb-fsf-trunk-quilt/gdb/dwarf2read.c 2014-12-12 00:21:46.768928206 +0000 @@ -7027,9 +7027,6 @@ add_partial_subprogram (struct partial_d CORE_ADDR *lowpc, CORE_ADDR *highpc, int set_addrmap, struct dwarf2_cu *cu) { - struct objfile *objfile = cu->objfile; - struct gdbarch *gdbarch = get_objfile_arch (objfile); - if (pdi->tag == DW_TAG_subprogram) { if (pdi->has_pc_info) @@ -7041,6 +7038,7 @@ add_partial_subprogram (struct partial_d if (set_addrmap) { struct objfile *objfile = cu->objfile; + struct gdbarch *gdbarch = get_objfile_arch (objfile); CORE_ADDR baseaddr; CORE_ADDR highpc; CORE_ADDR lowpc; ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-12-12 14:00 ` Maciej W. Rozycki @ 2014-12-12 17:22 ` Doug Evans 0 siblings, 0 replies; 27+ messages in thread From: Doug Evans @ 2014-12-12 17:22 UTC (permalink / raw) To: Maciej W. Rozycki Cc: Joel Brobecker, gdb-patches, Rich Fuhler, Richard Sandiford On Fri, Dec 12, 2014 at 6:00 AM, Maciej W. Rozycki <macro@codesourcery.com> wrote: > Joel, Doug -- > > On Thu, 4 Dec 2014, Maciej W. Rozycki wrote: > >> > This is not a requirement for your patch, but if you happen to be able >> > to quickly document elf_make_msymbol_special as well, that would be >> > a very welcome and appreciated change. >> >> I have now added these descriptions, with some bias towards the MIPS >> specifics as they are what I am most familiar with. Please let me know >> if you think they might be further improved. > > I haven't heard back from you, so I have applied the change now. > > While reviewing the final ChangeLog entry I discovered I created an > unneeded shadow local variable definition, so in addition to the changes > already posted in this thread I have included the following cleanup, as > obvious, in the final commit. Hi. Thanks very much for the added comments. ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-10-06 0:42 ` [PATCH v2 1/2] " Maciej W. Rozycki ` (3 preceding siblings ...) 2014-11-16 10:37 ` Joel Brobecker @ 2014-11-17 1:17 ` Doug Evans 2014-12-04 15:31 ` Maciej W. Rozycki 2014-12-12 16:38 ` [committed] MIPS: Define aliases for MSYMBOL_TARGET_FLAG macros Maciej W. Rozycki 5 siblings, 1 reply; 27+ messages in thread From: Doug Evans @ 2014-11-17 1:17 UTC (permalink / raw) To: Maciej W. Rozycki Cc: Joel Brobecker, gdb-patches, Rich Fuhler, Richard Sandiford "Maciej W. Rozycki" <macro@codesourcery.com> writes: > [...] > 2014-10-06 Maciej W. Rozycki <macro@codesourcery.com> > Maciej W. Rozycki <macro@mips.com> > Pedro Alves <pedro@codesourcery.com> > > * gdbarch.sh (elf_make_msymbol_special): Change type to `F', > remove `predefault' and `invalid_p' initializers. > (make_symbol_special): New architecture method. > (adjust_dwarf2_addr, adjust_dwarf2_line): Likewise. > (objfile, symbol): New declarations. > * arch-utils.h (default_elf_make_msymbol_special): Remove > prototype. > (default_make_symbol_special): New prototype. > (default_adjust_dwarf2_addr): Likewise. > (default_adjust_dwarf2_line): Likewise. > * mips-tdep.h (mips_unmake_compact_addr): New prototype. > * arch-utils.c (default_elf_make_msymbol_special): Remove > function. > (default_make_symbol_special): New function. > (default_adjust_dwarf2_addr): Likewise. > (default_adjust_dwarf2_line): Likewise. > * dwarf2-frame.c (decode_frame_entry_1): Call > `gdbarch_adjust_dwarf2_addr'. > * dwarf2loc.c (dwarf2_find_location_expression): Likewise. > * dwarf2read.c (create_addrmap_from_index): Likewise. > (process_psymtab_comp_unit_reader): Likewise. > (add_partial_symbol): Likewise. > (add_partial_subprogram): Likewise. > (process_full_comp_unit): Likewise. > (read_file_scope): Likewise. > (read_func_scope): Likewise. Call `gdbarch_make_symbol_special'. > (read_lexical_block_scope): Call `gdbarch_adjust_dwarf2_addr'. > (read_call_site_scope): Likewise. > (dwarf2_ranges_read): Likewise. > (dwarf2_record_block_ranges): Likewise. > (read_attribute_value): Likewise. > (dwarf_decode_lines_1): Call `gdbarch_adjust_dwarf2_line'. > (new_symbol_full): Call `gdbarch_adjust_dwarf2_addr'. > * elfread.c (elf_symtab_read): Don't call > `gdbarch_elf_make_msymbol_special' if unset. > * mips-linux-tdep.c (micromips_linux_sigframe_validate): Strip > the ISA bit from the PC. > * mips-tdep.c (mips_unmake_compact_addr): New function. > (mips_elf_make_msymbol_special): Set the ISA bit in the symbol's > address appropriately. > (mips_make_symbol_special): New function. > (mips_pc_is_mips): Set the ISA bit before symbol lookup. > (mips_pc_is_mips16): Likewise. > (mips_pc_is_micromips): Likewise. > (mips_pc_isa): Likewise. > (mips_adjust_dwarf2_addr): New function. > (mips_adjust_dwarf2_line): Likewise. > (mips_read_pc, mips_unwind_pc): Keep the ISA bit. > (mips_addr_bits_remove): Likewise. > (mips_skip_trampoline_code): Likewise. > (mips_write_pc): Don't set the ISA bit. > (mips_eabi_push_dummy_call): Likewise. > (mips_o64_push_dummy_call): Likewise. > (mips_gdbarch_init): Install `mips_make_symbol_special', > `mips_adjust_dwarf2_addr' and `mips_adjust_dwarf2_line' gdbarch > handlers. > * solib.c (gdb_bfd_lookup_symbol_from_symtab): Get > target-specific symbol address adjustments. > * gdbarch.h: Regenerate. > * gdbarch.c: Regenerate. > > 2014-10-06 Maciej W. Rozycki <macro@codesourcery.com> > > gdb/testsuite/ > * gdb.base/func-ptrs.c: New file. > * gdb.base/func-ptrs.exp: New file. Ummm, bleah. Going forward dwarf2*.c are going to be measurably more of a pain to maintain than they already are. I notice a fair bit of mips code gets a lot simpler. Such is life I guess. Just a few comments. >[...] > Index: gdb-fsf-trunk-quilt/gdb/gdbarch.sh > =================================================================== > --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.sh 2014-10-03 13:52:46.000000000 +0100 > +++ gdb-fsf-trunk-quilt/gdb/gdbarch.sh 2014-10-03 14:50:26.000000000 +0100 > @@ -635,8 +635,11 @@ m:int:in_solib_return_trampoline:CORE_AD > # which don't suffer from that problem could just let this functionality > # untouched. > m:int:in_function_epilogue_p:CORE_ADDR addr:addr:0:generic_in_function_epilogue_p::0 > -f:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym::default_elf_make_msymbol_special::0 > +F:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym > f:void:coff_make_msymbol_special:int val, struct minimal_symbol *msym:val, msym::default_coff_make_msymbol_special::0 > +f:void:make_symbol_special:struct symbol *sym, struct objfile *objfile:sym, objfile::default_make_symbol_special::0 > +f:CORE_ADDR:adjust_dwarf2_addr:CORE_ADDR pc:pc::default_adjust_dwarf2_addr::0 > +f:CORE_ADDR:adjust_dwarf2_line:CORE_ADDR addr, int rel:addr, rel::default_adjust_dwarf2_line::0 I don't know where the best place to put this is, but the adjust_dwarf2 functions need a lot more documentation. Why do they exist? What problem do they solve? Something along those lines. References to existing documentation is fine, as long as they're sufficient. I realize I can find that out with some manual labor, e.g., finding which arches provide an implementation of these functions, and reading them, but even when I get to mips-tdep.c I'm left wondering *why* the mips port needs this (e.g., why can't the problem be solved differently?). I'm more familiar with arm/thumb, and of course the first thing that comes to mind is that arm/thumb doesn't need this (maybe it does, but we've gotten this far without it). It would help to educate the reader of this code as to why mips is different. Maybe this patch is the best solution. If so, and I'm going to assume that it is, then let's get this documented in the code. Even just a reference to existing docs is fine. mips-tdep.c is a big file though - I don't want to have to read all of it just to understand why adjust_dwarf2_line exists. > v:int:cannot_step_breakpoint:::0:0::0 > v:int:have_nonsteppable_watchpoint:::0:0::0 > F:int:address_class_type_flags:int byte_size, int dwarf2_addr_class:byte_size, dwarf2_addr_class > @@ -1140,6 +1143,8 @@ struct target_ops; > struct obstack; > struct bp_target_info; > struct target_desc; > +struct objfile; > +struct symbol; > struct displaced_step_closure; > struct core_regset_section; > struct syscall; >[...] > Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.c > =================================================================== > --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.c 2014-10-03 13:52:46.000000000 +0100 > +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.c 2014-10-06 00:11:55.028611249 +0100 >[...] > @@ -391,6 +406,29 @@ msymbol_is_micromips (struct minimal_sym > return MSYMBOL_TARGET_FLAG_2 (msym); > } > > +/* Set the ISA bit in the main symbol too, complementing the corresponding > + minimal symbol setting and reflecting the run-time value of the symbol. */ > + > +static void > +mips_make_symbol_special (struct symbol *sym, struct objfile *objfile) > +{ > + if (SYMBOL_CLASS (sym) == LOC_BLOCK) > + { > + CORE_ADDR compact_block_start; > + struct bound_minimal_symbol msym; > + > + compact_block_start = BLOCK_START (SYMBOL_BLOCK_VALUE (sym)) | 1; > + msym = lookup_minimal_symbol_by_pc (compact_block_start); > + if (msym.minsym && !msymbol_is_mips (msym.minsym)) > + { > + /* We are in symbol reading so it is OK to cast away constness. */ > + struct block *block = (struct block *) SYMBOL_BLOCK_VALUE (sym); > + > + BLOCK_START (block) = compact_block_start; > + } > + } > +} This function would be easier to read if the assignment to block was moved up and its value used in both invocations of BLOCK_START. >[...] > +/* Recalculate the line record requested so that the resulting PC has the > + ISA bit set correctly, used by DWARF-2 machinery. */ > + > +static CORE_ADDR > +mips_adjust_dwarf2_line (CORE_ADDR addr, int rel) > +{ > + static CORE_ADDR adj_pc; > + static CORE_ADDR pc; > + CORE_ADDR isa_pc; > + > + pc = rel ? pc + addr : addr; > + isa_pc = mips_adjust_dwarf2_addr (pc); > + addr = rel ? isa_pc - adj_pc : isa_pc; > + adj_pc = isa_pc; > + return addr; > +} If this function was a "method", we could remove the "static" vars, but this is ok I guess. There's magic going on here, carrying the value of pc,adj_pc from one call to the next. There's a rule here that this will be called with rel == 0, before any calls with rel != 0. It's kinda obvious now, but it would have helped this reader to see it written down. Ok with those changes. ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH v2 1/2] ISA bit treatment on the MIPS platform 2014-11-17 1:17 ` Doug Evans @ 2014-12-04 15:31 ` Maciej W. Rozycki 0 siblings, 0 replies; 27+ messages in thread From: Maciej W. Rozycki @ 2014-12-04 15:31 UTC (permalink / raw) To: Doug Evans; +Cc: Joel Brobecker, gdb-patches, Rich Fuhler, Richard Sandiford On Mon, 17 Nov 2014, Doug Evans wrote: > "Maciej W. Rozycki" <macro@codesourcery.com> writes: > > [...] > > 2014-10-06 Maciej W. Rozycki <macro@codesourcery.com> > > Maciej W. Rozycki <macro@mips.com> > > Pedro Alves <pedro@codesourcery.com> > > > > * gdbarch.sh (elf_make_msymbol_special): Change type to `F', > > remove `predefault' and `invalid_p' initializers. > > (make_symbol_special): New architecture method. > > (adjust_dwarf2_addr, adjust_dwarf2_line): Likewise. > > (objfile, symbol): New declarations. > > * arch-utils.h (default_elf_make_msymbol_special): Remove > > prototype. > > (default_make_symbol_special): New prototype. > > (default_adjust_dwarf2_addr): Likewise. > > (default_adjust_dwarf2_line): Likewise. > > * mips-tdep.h (mips_unmake_compact_addr): New prototype. > > * arch-utils.c (default_elf_make_msymbol_special): Remove > > function. > > (default_make_symbol_special): New function. > > (default_adjust_dwarf2_addr): Likewise. > > (default_adjust_dwarf2_line): Likewise. > > * dwarf2-frame.c (decode_frame_entry_1): Call > > `gdbarch_adjust_dwarf2_addr'. > > * dwarf2loc.c (dwarf2_find_location_expression): Likewise. > > * dwarf2read.c (create_addrmap_from_index): Likewise. > > (process_psymtab_comp_unit_reader): Likewise. > > (add_partial_symbol): Likewise. > > (add_partial_subprogram): Likewise. > > (process_full_comp_unit): Likewise. > > (read_file_scope): Likewise. > > (read_func_scope): Likewise. Call `gdbarch_make_symbol_special'. > > (read_lexical_block_scope): Call `gdbarch_adjust_dwarf2_addr'. > > (read_call_site_scope): Likewise. > > (dwarf2_ranges_read): Likewise. > > (dwarf2_record_block_ranges): Likewise. > > (read_attribute_value): Likewise. > > (dwarf_decode_lines_1): Call `gdbarch_adjust_dwarf2_line'. > > (new_symbol_full): Call `gdbarch_adjust_dwarf2_addr'. > > * elfread.c (elf_symtab_read): Don't call > > `gdbarch_elf_make_msymbol_special' if unset. > > * mips-linux-tdep.c (micromips_linux_sigframe_validate): Strip > > the ISA bit from the PC. > > * mips-tdep.c (mips_unmake_compact_addr): New function. > > (mips_elf_make_msymbol_special): Set the ISA bit in the symbol's > > address appropriately. > > (mips_make_symbol_special): New function. > > (mips_pc_is_mips): Set the ISA bit before symbol lookup. > > (mips_pc_is_mips16): Likewise. > > (mips_pc_is_micromips): Likewise. > > (mips_pc_isa): Likewise. > > (mips_adjust_dwarf2_addr): New function. > > (mips_adjust_dwarf2_line): Likewise. > > (mips_read_pc, mips_unwind_pc): Keep the ISA bit. > > (mips_addr_bits_remove): Likewise. > > (mips_skip_trampoline_code): Likewise. > > (mips_write_pc): Don't set the ISA bit. > > (mips_eabi_push_dummy_call): Likewise. > > (mips_o64_push_dummy_call): Likewise. > > (mips_gdbarch_init): Install `mips_make_symbol_special', > > `mips_adjust_dwarf2_addr' and `mips_adjust_dwarf2_line' gdbarch > > handlers. > > * solib.c (gdb_bfd_lookup_symbol_from_symtab): Get > > target-specific symbol address adjustments. > > * gdbarch.h: Regenerate. > > * gdbarch.c: Regenerate. > > > > 2014-10-06 Maciej W. Rozycki <macro@codesourcery.com> > > > > gdb/testsuite/ > > * gdb.base/func-ptrs.c: New file. > > * gdb.base/func-ptrs.exp: New file. > > Ummm, bleah. Yeah, that's a consequence of a bad decision made (or the lack of one and the resulting semantics that spontaneously happened) many years ago as MIPS16 support was implemented. References to compressed code are made in an inconsistent way across DWARF records with information lost from some of them. That information has to be recovered from the symbol table. The missing information could be stored separately in DWARF records I suppose, but that won't help with many binaries out there; MIPS16 support has been there for well over a decade -- our first record is from Jan 1997 and support in GCC/GAS must have been there since around that time too. > Going forward dwarf2*.c are going to be measurably more of a pain > to maintain than they already are. > I notice a fair bit of mips code gets a lot simpler. And some stuff actually starts working, that never did before. I posted the full story along the original patch submission: http://sourceware.org/ml/gdb-patches/2012-05/msg00515.html > Such is life I guess. I tried hard figuring out a way that would avoid making changes to generic DWARF code and could not find one. Once the DWARF records have been processed there doesn't appear to me to be a way for places across GDB's generic code to gather the pieces of information required without turning upside down those places for a change. > Just a few comments. > > >[...] > > Index: gdb-fsf-trunk-quilt/gdb/gdbarch.sh > > =================================================================== > > --- gdb-fsf-trunk-quilt.orig/gdb/gdbarch.sh 2014-10-03 13:52:46.000000000 +0100 > > +++ gdb-fsf-trunk-quilt/gdb/gdbarch.sh 2014-10-03 14:50:26.000000000 +0100 > > @@ -635,8 +635,11 @@ m:int:in_solib_return_trampoline:CORE_AD > > # which don't suffer from that problem could just let this functionality > > # untouched. > > m:int:in_function_epilogue_p:CORE_ADDR addr:addr:0:generic_in_function_epilogue_p::0 > > -f:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym::default_elf_make_msymbol_special::0 > > +F:void:elf_make_msymbol_special:asymbol *sym, struct minimal_symbol *msym:sym, msym > > f:void:coff_make_msymbol_special:int val, struct minimal_symbol *msym:val, msym::default_coff_make_msymbol_special::0 > > +f:void:make_symbol_special:struct symbol *sym, struct objfile *objfile:sym, objfile::default_make_symbol_special::0 > > +f:CORE_ADDR:adjust_dwarf2_addr:CORE_ADDR pc:pc::default_adjust_dwarf2_addr::0 > > +f:CORE_ADDR:adjust_dwarf2_line:CORE_ADDR addr, int rel:addr, rel::default_adjust_dwarf2_line::0 > > I don't know where the best place to put this is, but the adjust_dwarf2 > functions need a lot more documentation. Why do they exist? > What problem do they solve? Something along those lines. > References to existing documentation is fine, as long as they're sufficient. The source code has so far been the documentation I am afraid. Which I think also contributes to the mess. Implied bits are buried in GCC in gcc/config/mips/mips.h and libgcc/unwind-dw2.c for example, see the comment around `MASK_RETURN_ADDR' in the former and the comment at the beginning of `execute_cfa_program' in the latter. The assumptions made in these both places make issues nasty, adding 1 to the return address in the latter place is incompatible with how the ISA bit is interpreted everywhere else. This function post-dates the introduction of the MIPS16 instruction set I believe and therefore I think it should have been designed with the issues around it in mind, e.g. I used the patch below in the early stage of investigation of possible solutions for the problem addressed here. > I realize I can find that out with some manual labor, e.g., finding > which arches provide an implementation of these functions, and reading them, > but even when I get to mips-tdep.c I'm left wondering *why* > the mips port needs this (e.g., why can't the problem be solved > differently?). I'm more familiar with arm/thumb, and of course > the first thing that comes to mind is that arm/thumb doesn't need this > (maybe it does, but we've gotten this far without it). I suspect that it does, but it can be found out easily as this is why I included gdb.base/func-ptrs.exp with the change. Any target that fails the test cases it contains will need addressing the problem, possibly just by implementing the equivalent of the MIPS bits posted here. > It would help > to educate the reader of this code as to why mips is different. > Maybe this patch is the best solution. If so, and I'm going to assume that > it is, then let's get this documented in the code. > Even just a reference to existing docs is fine. > mips-tdep.c is a big file though - I don't want to have to read all > of it just to understand why adjust_dwarf2_line exists. I'll try summarising the key points for the new hooks in gdbarch.sh (that is propagated by the script to gdbarch.h) and for the MIPS implementation in mips-tdep.c. I'll post the final version for you (and Joel) to comment on before I push it, in the next few days. > > Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.c > > =================================================================== > > --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.c 2014-10-03 13:52:46.000000000 +0100 > > +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.c 2014-10-06 00:11:55.028611249 +0100 > >[...] > > @@ -391,6 +406,29 @@ msymbol_is_micromips (struct minimal_sym > > return MSYMBOL_TARGET_FLAG_2 (msym); > > } > > > > +/* Set the ISA bit in the main symbol too, complementing the corresponding > > + minimal symbol setting and reflecting the run-time value of the symbol. */ > > + > > +static void > > +mips_make_symbol_special (struct symbol *sym, struct objfile *objfile) > > +{ > > + if (SYMBOL_CLASS (sym) == LOC_BLOCK) > > + { > > + CORE_ADDR compact_block_start; > > + struct bound_minimal_symbol msym; > > + > > + compact_block_start = BLOCK_START (SYMBOL_BLOCK_VALUE (sym)) | 1; > > + msym = lookup_minimal_symbol_by_pc (compact_block_start); > > + if (msym.minsym && !msymbol_is_mips (msym.minsym)) > > + { > > + /* We are in symbol reading so it is OK to cast away constness. */ > > + struct block *block = (struct block *) SYMBOL_BLOCK_VALUE (sym); > > + > > + BLOCK_START (block) = compact_block_start; > > + } > > + } > > +} > > This function would be easier to read if the assignment to block > was moved up and its value used in both invocations of BLOCK_START. Fixed. It looks to me like an outcome from the long history of this function (dating back to Jan 2008) and the number of adjustments it gathered as the structures it uses evolved. In the process I missed the possibility for this simplification. > >[...] > > +/* Recalculate the line record requested so that the resulting PC has the > > + ISA bit set correctly, used by DWARF-2 machinery. */ > > + > > +static CORE_ADDR > > +mips_adjust_dwarf2_line (CORE_ADDR addr, int rel) > > +{ > > + static CORE_ADDR adj_pc; > > + static CORE_ADDR pc; > > + CORE_ADDR isa_pc; > > + > > + pc = rel ? pc + addr : addr; > > + isa_pc = mips_adjust_dwarf2_addr (pc); > > + addr = rel ? isa_pc - adj_pc : isa_pc; > > + adj_pc = isa_pc; > > + return addr; > > +} > > If this function was a "method", we could remove the "static" vars, > but this is ok I guess. > There's magic going on here, carrying the value of pc,adj_pc > from one call to the next. There's a rule here that this will > be called with rel == 0, before any calls with rel != 0. > It's kinda obvious now, but it would have helped this reader to see > it written down. Indeed, already noticed by Joel. This piece gave me a bit of a headache and after considering available alternatives I decided to put the least burden on our DWARF machinery and contain all the stuff within the MIPS backend. This way people outside the MIPS world do not have to be bothered with it, and I think code is also relatively simple, which is going to help with long-term maintenance. The only drawback is backend's state has to be initialised and consequently an initial call with `rel == 0' is required. I'll document it at the call site and here as well. And here's the experimental patch for GCC EH unwinder I referred to above, published for posterity. Maciej gcc-mips-unwind-keep-isa-bit.diff Index: gcc-fsf-trunk-quilt/gcc/config/mips/mips.h =================================================================== --- gcc-fsf-trunk-quilt.orig/gcc/config/mips/mips.h 2012-05-01 20:44:04.000000000 +0100 +++ gcc-fsf-trunk-quilt/gcc/config/mips/mips.h 2012-05-02 23:44:40.375576966 +0100 @@ -2049,6 +2049,7 @@ enum reg_class #define RETURN_ADDR_RTX mips_return_addr +#if 0 /* Mask off the MIPS16 ISA bit in unwind addresses. The reason for this is a little subtle. When unwinding a call, @@ -2076,6 +2077,9 @@ enum reg_class Masking off the ISA bit means that the target-independent code will search for "(RA & -2) - 1", which is guaranteed to be odd. */ #define MASK_RETURN_ADDR GEN_INT (-2) +#else +#define MIN_INSN_SIZE 2 +#endif /* Similarly, don't use the least-significant bit to tell pointers to Index: gcc-fsf-trunk-quilt/libgcc/unwind-dw2.c =================================================================== --- gcc-fsf-trunk-quilt.orig/libgcc/unwind-dw2.c 2012-05-01 20:41:24.275429140 +0100 +++ gcc-fsf-trunk-quilt/libgcc/unwind-dw2.c 2012-05-02 23:44:40.385574339 +0100 @@ -927,7 +927,8 @@ execute_cfa_program (const unsigned char In signal frames, return address is after last completed instruction, so we add 1 to return address to make the comparison <=. */ while (insn_ptr < insn_end - && fs->pc < context->ra + _Unwind_IsSignalFrame (context)) + && fs->pc <= context->ra - (_Unwind_IsSignalFrame (context) + ? 0 : MIN_INSN_SIZE)) { unsigned char insn = *insn_ptr++; _uleb128_t reg, utmp; @@ -1177,7 +1178,8 @@ uw_frame_state_for (struct _Unwind_Conte if (context->ra == 0) return _URC_END_OF_STACK; - fde = _Unwind_Find_FDE (context->ra + _Unwind_IsSignalFrame (context) - 1, + fde = _Unwind_Find_FDE (context->ra - (_Unwind_IsSignalFrame (context) + ? 0 : MIN_INSN_SIZE), &context->bases); if (fde == NULL) { ^ permalink raw reply [flat|nested] 27+ messages in thread
* [committed] MIPS: Define aliases for MSYMBOL_TARGET_FLAG macros 2014-10-06 0:42 ` [PATCH v2 1/2] " Maciej W. Rozycki ` (4 preceding siblings ...) 2014-11-17 1:17 ` Doug Evans @ 2014-12-12 16:38 ` Maciej W. Rozycki 5 siblings, 0 replies; 27+ messages in thread From: Maciej W. Rozycki @ 2014-12-12 16:38 UTC (permalink / raw) To: Joel Brobecker; +Cc: gdb-patches, Rich Fuhler, Richard Sandiford On Mon, 6 Oct 2014, Maciej W. Rozycki wrote: > > Just some minor comments below: > > > > > gdb-mips16-isa-bit.diff > > > Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.c > > > =================================================================== > > > --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.c 2012-05-14 16:00:33.000000000 +0100 > > > +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.c 2012-05-14 16:02:02.235560558 +0100 > > > @@ -358,9 +358,15 @@ mips_elf_make_msymbol_special (asymbol * > > > return; > > > > > > if (ELF_ST_IS_MICROMIPS (elfsym->internal_elf_sym.st_other)) > > > - MSYMBOL_TARGET_FLAG_2 (msym) = 1; > > > + { > > > + MSYMBOL_TARGET_FLAG_2 (msym) = 1; > > > + SYMBOL_VALUE_ADDRESS (msym) |= 1; > > > + } > > > else if (ELF_ST_IS_MIPS16 (elfsym->internal_elf_sym.st_other)) > > > - MSYMBOL_TARGET_FLAG_1 (msym) = 1; > > > + { > > > + MSYMBOL_TARGET_FLAG_1 (msym) = 1; > > > + SYMBOL_VALUE_ADDRESS (msym) |= 1; > > > + } > > > > This remark is not to be considered as part of this patch's review, > > since this is already an established practice, but I think we could do > > better than using magic numbers for those flags. I understand they have > > to be that way in the common code, but perhaps mips-tdep could define > > aliases? Something like MSYMBOL_MIPS_TARGET_FLAG_BLAH? > > Good point, I'll push such a change as a followup; I have it ready now. I have committed this change now. 2014-12-12 Maciej W. Rozycki <macro@codesourcery.com> gdb/ * mips-tdep.h (MSYMBOL_TARGET_FLAG_MIPS16): New macro. (MSYMBOL_TARGET_FLAG_MICROMIPS): Likewise. * mips-tdep.c (mips_elf_make_msymbol_special): Use the new macros. (msymbol_is_mips, msymbol_is_mips16, msymbol_is_micromips): Likewise. Maciej gdb-mips-msymbol-target-flag.diff Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.c =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.c 2014-10-03 14:50:26.398945943 +0100 +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.c 2014-10-03 14:50:26.398945943 +0100 @@ -379,12 +379,12 @@ mips_elf_make_msymbol_special (asymbol * if (ELF_ST_IS_MICROMIPS (st_other)) { - MSYMBOL_TARGET_FLAG_2 (msym) = 1; + MSYMBOL_TARGET_FLAG_MICROMIPS (msym) = 1; SET_MSYMBOL_VALUE_ADDRESS (msym, MSYMBOL_VALUE_RAW_ADDRESS (msym) | 1); } else if (ELF_ST_IS_MIPS16 (st_other)) { - MSYMBOL_TARGET_FLAG_1 (msym) = 1; + MSYMBOL_TARGET_FLAG_MIPS16 (msym) = 1; SET_MSYMBOL_VALUE_ADDRESS (msym, MSYMBOL_VALUE_RAW_ADDRESS (msym) | 1); } } @@ -394,7 +394,8 @@ mips_elf_make_msymbol_special (asymbol * static int msymbol_is_mips (struct minimal_symbol *msym) { - return !(MSYMBOL_TARGET_FLAG_1 (msym) | MSYMBOL_TARGET_FLAG_2 (msym)); + return !(MSYMBOL_TARGET_FLAG_MIPS16 (msym) + | MSYMBOL_TARGET_FLAG_MICROMIPS (msym)); } /* Return one iff MSYM refers to MIPS16 code. */ @@ -402,7 +403,7 @@ msymbol_is_mips (struct minimal_symbol * static int msymbol_is_mips16 (struct minimal_symbol *msym) { - return MSYMBOL_TARGET_FLAG_1 (msym); + return MSYMBOL_TARGET_FLAG_MIPS16 (msym); } /* Return one iff MSYM refers to microMIPS code. */ @@ -410,7 +411,7 @@ msymbol_is_mips16 (struct minimal_symbol static int msymbol_is_micromips (struct minimal_symbol *msym) { - return MSYMBOL_TARGET_FLAG_2 (msym); + return MSYMBOL_TARGET_FLAG_MICROMIPS (msym); } /* Set the ISA bit in the main symbol too, complementing the corresponding Index: gdb-fsf-trunk-quilt/gdb/mips-tdep.h =================================================================== --- gdb-fsf-trunk-quilt.orig/gdb/mips-tdep.h 2014-10-03 14:50:26.398945943 +0100 +++ gdb-fsf-trunk-quilt/gdb/mips-tdep.h 2014-10-03 14:50:26.398945943 +0100 @@ -48,6 +48,10 @@ enum mips_isa ISA_MICROMIPS }; +/* Corresponding MSYMBOL_TARGET_FLAG aliases. */ +#define MSYMBOL_TARGET_FLAG_MIPS16 MSYMBOL_TARGET_FLAG_1 +#define MSYMBOL_TARGET_FLAG_MICROMIPS MSYMBOL_TARGET_FLAG_2 + /* Return the MIPS ISA's register size. Just a short cut to the BFD architecture's word size. */ extern int mips_isa_regsize (struct gdbarch *gdbarch); ^ permalink raw reply [flat|nested] 27+ messages in thread
end of thread, other threads:[~2014-12-12 17:22 UTC | newest] Thread overview: 27+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2012-05-14 15:37 [RFD+PATCH] ISA bit treatment on the MIPS platform Maciej W. Rozycki 2012-05-18 22:01 ` Maciej W. Rozycki 2012-06-11 18:21 ` Joel Brobecker 2012-06-12 14:05 ` Pedro Alves 2014-10-06 0:42 ` [PATCH v2 1/2] " Maciej W. Rozycki 2014-10-06 0:42 ` [PATCH v2 2/2] Correct invalid assumptions made by (mostly) DWARF-2 tests Maciej W. Rozycki 2014-11-16 11:09 ` Joel Brobecker 2014-11-16 18:32 ` Doug Evans 2014-11-16 19:49 ` Doug Evans 2014-11-16 20:05 ` Maciej W. Rozycki 2014-11-16 21:52 ` Doug Evans 2014-12-04 0:24 ` Maciej W. Rozycki 2014-11-16 22:28 ` Doug Evans 2014-10-06 14:10 ` [PATCH v2 1/2] ISA bit treatment on the MIPS platform Joel Brobecker 2014-10-14 20:45 ` Maciej W. Rozycki 2014-10-20 17:10 ` [PING][PATCH " Maciej W. Rozycki 2014-11-03 16:13 ` [PING^2][PATCH " Maciej W. Rozycki 2014-11-16 19:23 ` [PATCH " Doug Evans 2014-10-06 15:43 ` Maciej W. Rozycki 2014-11-16 10:37 ` Joel Brobecker 2014-11-16 19:27 ` Doug Evans 2014-12-04 23:14 ` Maciej W. Rozycki 2014-12-12 14:00 ` Maciej W. Rozycki 2014-12-12 17:22 ` Doug Evans 2014-11-17 1:17 ` Doug Evans 2014-12-04 15:31 ` Maciej W. Rozycki 2014-12-12 16:38 ` [committed] MIPS: Define aliases for MSYMBOL_TARGET_FLAG macros Maciej W. Rozycki
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).