Commit Graph

13 Commits

Author SHA1 Message Date
Yao Qi f68f11b76d Support catch syscall on aarch64 linux
Hi,
This patch is to support catch syscall on aarch64 linux.  We
implement gdbarch method get_syscall_number for aarch64-linux,
and add aarch64-linux.xml file, which looks straightforward, however
the changes to test case doesn't.

First of all, we enable catch-syscall.exp on aarch64-linux target,
but skip the multi_arch testing on current stage.  I plan to touch
multi arch debugging on aarch64-linux later.

Then, when I run catch-syscall.exp on aarch64-linux, gcc errors that
SYS_pipe isn't defined.  We find that aarch64 kernel only has pipe2
syscall and libc already convert pipe to pipe2.  As a result, I change
catch-syscall.c to use SYS_pipe if it is defined, otherwise use
SYS_pipe2 instead.  The vector all_syscalls in catch-syscall.exp can't
be pre-determined, so I add a new proc setup_all_syscalls to fill it,
according to the availability of SYS_pipe.

Regression tested on {x86_64, aarch64}-linux x {native, gdbserver}.

gdb:

2015-03-18  Yao Qi  <yao.qi@linaro.org>

	PR tdep/18107
	* aarch64-linux-tdep.c: Include xml-syscall.h
	(aarch64_linux_get_syscall_number): New function.
	(aarch64_linux_init_abi): Call
	set_gdbarch_get_syscall_number.
	* syscalls/aarch64-linux.xml: New file.

gdb/testsuite:

2015-03-18  Yao Qi  <yao.qi@linaro.org>

	PR tdep/18107
	* gdb.base/catch-syscall.c [!SYS_pipe] (pipe2_syscall): New
	variable.
	* gdb.base/catch-syscall.exp: Don't skip it on
	aarch64*-*-linux* target.  Remove elements in all_syscalls.
	(test_catch_syscall_multi_arch): Skip it on aarch64*-linux*
	target.
	(setup_all_syscalls): New proc.
2015-03-18 10:47:45 +00:00
Joel Brobecker 32d0add0a6 Update year range in copyright notice of all files owned by the GDB project.
gdb/ChangeLog:

        Update year range in copyright notice of all files.
2015-01-01 13:32:14 +04:00
Andreas Arnez 4108500a2a AARCH64: Migrate from 'regset_from_core_section' to 'iterate_over_regset_sections'
Don't define the 'regset_from_core_section' method, but the iterator
method instead.  This slightly reduces the code and enables non-native
use of the 'generate-core-file' command.

Also, when all instances of 'regset_from_core_section' are replaced,
it can be dropped from the gdbarch interface.

gdb/ChangeLog:

	* aarch64-linux-tdep.c (aarch64_linux_regset_from_core_section):
	Remove.
	(aarch64_linux_iterate_over_regset_sections): New.
	(aarch64_linux_init_abi): Adjust gdbarch initialization.
2014-09-30 09:14:33 +02:00
Andreas Arnez d4d793bfeb AARCH64 Linux: Fill 'collect_regset' in regset structures.
In order to provide 'collect_regset' support, the generic function
regcache_collect_regset is exploited.  Since this requires writing
appropriate register maps, these can be used for supply_regset as
well.
2014-08-07 17:31:49 +02:00
Andreas Arnez a069a2bdab AARCH64: Replace regset_alloc() invocations by static regset structures.
After removal of the regset_alloc invocations, the tdep fields
'gregset' and 'fpregset' become obsolete and are thus removed.
2014-05-21 17:49:16 +02:00
Hui Zhu f2205de008 Fix issue #15778: GDB Aarch64 signal frame unwinder issue
The root cause of this issue is unwinder of "#3  <signal handler called>"
doesn't supply right values of registers.
When GDB want to get the previous frame of "#3  <signal handler called>",
it will call cache init function of unwinder "aarch64_linux_sigframe_init".
The address or the value of the registers is get from this function.
So the bug is inside thie function.

I check the asm code of "#3  <signal handler called>":
(gdb) frame 3
(gdb) p $pc
$1 = (void (*)()) 0x7f931fa4d0
(gdb) disassemble $pc, +10
Dump of assembler code from 0x7f931fa4d0 to 0x7f931fa4da:
=> 0x0000007f931fa4d0:	mov	x8, #0x8b                  	// #139
   0x0000007f931fa4d4:	svc	#0x0
   0x0000007f931fa4d8:	nop

This is the syscall sys_rt_sigreturn, Linux kernel function "restore_sigframe"
will set the frame:
	for (i = 0; i < 31; i++)
		__get_user_error(regs->regs[i], &sf->uc.uc_mcontext.regs[i],
				 err);
	__get_user_error(regs->sp, &sf->uc.uc_mcontext.sp, err);
	__get_user_error(regs->pc, &sf->uc.uc_mcontext.pc, err);
The struct of uc_mcontext is:
struct sigcontext {
	__u64 fault_address;
	/* AArch64 registers */
	__u64 regs[31];
	__u64 sp;
	__u64 pc;
	__u64 pstate;
	/* 4K reserved for FP/SIMD state and future expansion */
	__u8 __reserved[4096] __attribute__((__aligned__(16)));
};

But in GDB function "aarch64_linux_sigframe_init", the code the get address
of registers is:
  for (i = 0; i < 31; i++)
    {
      trad_frame_set_reg_addr (this_cache,
			       AARCH64_X0_REGNUM + i,
			       sigcontext_addr + AARCH64_SIGCONTEXT_XO_OFFSET
			       + i * AARCH64_SIGCONTEXT_REG_SIZE);
    }

  trad_frame_set_reg_addr (this_cache, AARCH64_FP_REGNUM, fp);
  trad_frame_set_reg_addr (this_cache, AARCH64_LR_REGNUM, fp + 8);
  trad_frame_set_reg_addr (this_cache, AARCH64_PC_REGNUM, fp + 8);

The code that get pc and sp is not right, so I change the code according
to Linux kernel code:
  trad_frame_set_reg_addr (this_cache, AARCH64_SP_REGNUM,
			   sigcontext_addr + AARCH64_SIGCONTEXT_XO_OFFSET
			     + 31 * AARCH64_SIGCONTEXT_REG_SIZE);
  trad_frame_set_reg_addr (this_cache, AARCH64_PC_REGNUM,
			   sigcontext_addr + AARCH64_SIGCONTEXT_XO_OFFSET
			     + 32 * AARCH64_SIGCONTEXT_REG_SIZE);

The issue was fixed by this change, and I did the regression test.  It
also fixed a lot of other XFAIL and FAIL.

2014-05-20  Hui Zhu  <hui@codesourcery.com>
	    Yao Qi  <yao@codesourcery.com>

	PR backtrace/16558
	* aarch64-linux-tdep.c (aarch64_linux_sigframe_init): Update comments
	and change address of sp and pc.
2014-05-20 13:19:06 +08:00
Sergio Durigan Junior 410a0ff2df Remove `expout*' globals from parser-defs.h
This commit removes the "expout*" globals from our parser code, turning
them into a structure that is passed when an expression needs to be
evaluated.  This is the initial step to make our parser less
"globalized".

This is mostly a mechanical patch, which creates a structure containing
the "expout*" globals and then modify all the functions that handle them
in order to take the structure as argument.  It is big, and has been
reviewed at least 4 times, so I think everything is covered.

Below you can see the message links from the discussions:

- First attempt:

<https://sourceware.org/ml/gdb-patches/2012-01/msg00522.html>
Message-ID: <m3k44s7qej.fsf@gmail.com>

- Second attempt:

<https://sourceware.org/ml/gdb-patches/2012-06/msg00054.html>
Message-Id: <1338665528-5932-1-git-send-email-sergiodj@redhat.com>

- Third attempt:

<https://sourceware.org/ml/gdb-patches/2014-01/msg00949.html>
Message-Id: <1390629467-27139-1-git-send-email-sergiodj@redhat.com>

- Fourth (last) attempt:

<https://sourceware.org/ml/gdb-patches/2014-03/msg00546.html>
Message-Id: <1395463432-29750-1-git-send-email-sergiodj@redhat.com>

gdb/
2014-03-27  Sergio Durigan Junior  <sergiodj@redhat.com>

	Remove some globals from our parser.
	* language.c (unk_lang_parser): Add "struct parser_state"
	argument.
	* language.h (struct language_defn) <la_parser>: Likewise.
	* parse.c (expout, expout_size, expout_ptr): Remove variables.
	(initialize_expout): Add "struct parser_state" argument.
	Rewrite function to use the parser state.
	(reallocate_expout, write_exp_elt, write_exp_elt_opcode,
	write_exp_elt_sym, write_exp_elt_block, write_exp_elt_objfile,
	write_exp_elt_longcst, write_exp_elt_dblcst,
	write_exp_elt_decfloatcst, write_exp_elt_type,
	write_exp_elt_intern, write_exp_string, write_exp_string_vector,
	write_exp_bitstring, write_exp_msymbol, mark_struct_expression,
	write_dollar_variable): Likewise.
	(parse_exp_in_context_1): Use parser state.
	(insert_type_address_space): Add "struct parser_state" argument.
	Use parser state.
	(increase_expout_size): New function.
	* parser-defs.h: Forward declare "struct language_defn" and
	"struct parser_state".
	(expout, expout_size, expout_ptr): Remove extern declarations.
	(parse_gdbarch, parse_language): Rewrite macro declarations to
	accept the parser state.
	(struct parser_state): New struct.
	(initialize_expout, reallocate_expout, write_exp_elt_opcode,
	write_exp_elt_sym, write_exp_elt_longcst, write_exp_elt_dblcst,
	write_exp_elt_decfloatcst, write_exp_elt_type,
	write_exp_elt_intern, write_exp_string, write_exp_string_vector,
	write_exp_bitstring, write_exp_elt_block, write_exp_elt_objfile,
	write_exp_msymbol, write_dollar_variable,
	mark_struct_expression, insert_type_address_space): Add "struct
	parser_state" argument.
	(increase_expout_size): New function.
	* utils.c (do_clear_parser_state): New function.
	(make_cleanup_clear_parser_state): Likewise.
	* utils.h (make_cleanup_clear_parser_state): New function
	prototype.
	* aarch64-linux-tdep.c (aarch64_stap_parse_special_token):
	Update calls to write_exp* in order to pass the parser state.
	* arm-linux-tdep.c (arm_stap_parse_special_token): Likewise.
	* i386-tdep.c (i386_stap_parse_special_token_triplet): Likewise.
	(i386_stap_parse_special_token_three_arg_disp): Likewise.
	* ppc-linux-tdep.c (ppc_stap_parse_special_token): Likewise.
	* stap-probe.c (stap_parse_register_operand): Likewise.
	(stap_parse_single_operand): Likewise.
	(stap_parse_argument_1): Likewise.
	(stap_parse_argument): Use parser state.
	* stap-probe.h: Include "parser-defs.h".
	(struct stap_parse_info) <pstate>: New field.
	* c-exp.y (parse_type): Rewrite to use parser state.
	(yyparse): Redefine to c_parse_internal.
	(pstate): New global variable.
	(parse_number): Add "struct parser_state" argument.
	(write_destructor_name): Likewise.
	(type_exp): Update calls to write_exp* and similars in order to
	use parser state.
	(exp1, exp, variable, qualified_name, space_identifier,
	typename, typebase): Likewise.
	(write_destructor_name, parse_number, lex_one_token,
	classify_name, classify_inner_name, c_parse): Add "struct
	parser_state" argument.  Update function to use parser state.
	* c-lang.h: Forward declare "struct parser_state".
	(c_parse): Add "struct parser_state" argument.
	* ada-exp.y (parse_type): Rewrite macro to use parser state.
	(yyparse): Redefine macro to ada_parse_internal.
	(pstate): New variable.
	(write_int, write_object_renaming, write_var_or_type,
	write_name_assoc, write_exp_op_with_string, write_ambiguous_var,
	type_int, type_long, type_long_long, type_float, type_double,
	type_long_double, type_char, type_boolean, type_system_address):
	Add "struct parser_state" argument.
	(exp1, primary, simple_exp, relation, and_exp, and_then_exp,
	or_exp, or_else_exp, xor_exp, type_prefix, opt_type_prefix,
	var_or_type, aggregate, aggregate_component_list,
	positional_list, others, component_group,
	component_associations): Update calls to write_exp* and similar
	functions in order to use parser state.
	(ada_parse, write_var_from_sym, write_int,
	write_exp_op_with_string, write_object_renaming,
	find_primitive_type, write_selectors, write_ambiguous_var,
	write_var_or_type, write_name_assoc, type_int, type_long,
	type_long_long, type_float, type_double, type_long_double,
	type_char, type_boolean, type_system_address): Add "struct
	parser_state" argument.  Adjust function to use parser state.
	* ada-lang.c (parse): Likewise.
	* ada-lang.h: Forward declare "struct parser_state".
	(ada_parse): Add "struct parser_state" argument.
	* ada-lex.l (processInt, processReal): Likewise.  Adjust all
	calls to both functions.
	* f-exp.y (parse_type, parse_f_type): Rewrite macros to use
	parser state.
	(yyparse): Redefine macro to f_parse_internal.
	(pstate): New variable.
	(parse_number): Add "struct parser_state" argument.
	(type_exp, exp, subrange, typebase): Update calls to write_exp*
	and similars in order to use parser state.
	(parse_number): Adjust code to use parser state.
	(yylex): Likewise.
	(f_parse): New function.
	* f-lang.h: Forward declare "struct parser_state".
	(f_parse): Add "struct parser_state" argument.
	* jv-exp.y (parse_type, parse_java_type): Rewrite macros to use
	parser state.
	(yyparse): Redefine macro for java_parse_internal.
	(pstate): New variable.
	(push_expression_name, push_expression_name, insert_exp): Add
	"struct parser_state" argument.
	(type_exp, StringLiteral, Literal, PrimitiveType, IntegralType,
	FloatingPointType, exp1, PrimaryNoNewArray, FieldAccess,
	FuncStart, MethodInvocation, ArrayAccess, PostfixExpression,
	PostIncrementExpression, PostDecrementExpression,
	UnaryExpression, PreIncrementExpression, PreDecrementExpression,
	UnaryExpressionNotPlusMinus, CastExpression,
	MultiplicativeExpression, AdditiveExpression, ShiftExpression,
	RelationalExpression, EqualityExpression, AndExpression,
	ExclusiveOrExpression, InclusiveOrExpression,
	ConditionalAndExpression, ConditionalOrExpression,
	ConditionalExpression, Assignment, LeftHandSide): Update
	calls to write_exp* and similars in order to use parser state.
	(parse_number): Ajust code to use parser state.
	(yylex): Likewise.
	(java_parse): New function.
	(push_variable): Add "struct parser_state" argument.  Adjust
	code to user parser state.
	(push_fieldnames, push_qualified_expression_name,
	push_expression_name, insert_exp): Likewise.
	* jv-lang.h: Forward declare "struct parser_state".
	(java_parse): Add "struct parser_state" argument.
	* m2-exp.y (parse_type, parse_m2_type): Rewrite macros to use
	parser state.
	(yyparse): Redefine macro to m2_parse_internal.
	(pstate): New variable.
	(type_exp, exp, fblock, variable, type): Update calls to
	write_exp* and similars to use parser state.
	(yylex): Likewise.
	(m2_parse): New function.
	* m2-lang.h: Forward declare "struct parser_state".
	(m2_parse): Add "struct parser_state" argument.
	* objc-lang.c (end_msglist): Add "struct parser_state" argument.
	* objc-lang.h: Forward declare "struct parser_state".
	(end_msglist): Add "struct parser_state" argument.
	* p-exp.y (parse_type): Rewrite macro to use parser state.
	(yyparse): Redefine macro to pascal_parse_internal.
	(pstate): New variable.
	(parse_number): Add "struct parser_state" argument.
	(type_exp, exp1, exp, qualified_name, variable): Update calls to
	write_exp* and similars in order to use parser state.
	(parse_number, yylex): Adjust code to use parser state.
	(pascal_parse): New function.
	* p-lang.h: Forward declare "struct parser_state".
	(pascal_parse): Add "struct parser_state" argument.
	* go-exp.y (parse_type): Rewrite macro to use parser state.
	(yyparse): Redefine macro to go_parse_internal.
	(pstate): New variable.
	(parse_number): Add "struct parser_state" argument.
	(type_exp, exp1, exp, variable, type): Update calls to
	write_exp* and similars in order to use parser state.
	(parse_number, lex_one_token, classify_name, yylex): Adjust code
	to use parser state.
	(go_parse): Likewise.
	* go-lang.h: Forward declare "struct parser_state".
	(go_parse): Add "struct parser_state" argument.
2014-03-27 19:10:40 -03:00
Joel Brobecker ecd75fc8ee Update Copyright year range in all files maintained by GDB. 2014-01-01 07:54:24 +04:00
Sergio Durigan Junior 08248ca9fe Implement SystemTap SDT probe support for AArch64
This commit implements the needed bits for SystemTap SDT probe support
on AArch64 architectures.

First, I started by looking at AArch64 assembly specification and
filling the necessary options on gdbarch's stap machinery in order to
make the generic asm parser (implemented in stap-probe.c) recognize
AArch64's asm.

After my last patch for the SystemTap SDT API, which extends it in order
to accept multiple prefixes and suffixes, this patch became simpler.  I
also followed Marcus suggestion and did not shared code between 32- and
64-bit ARM.

Tom asked me in a previous message how I did my tests.  I believe I
replied that, but just in case: I ran the tests on
gdb.base/stap-probe.exp by hand.  I also managed to run the tests on
real hardware, and they pass without regressions.

2013-12-28  Sergio Durigan Junior  <sergiodj@redhat.com>

	PR tdep/15653
	* NEWS: Mention SystemTap SDT probe support for AArch64 GNU/Linux.
	* aarch64-linux-tdep.c: Include necessary headers for parsing of
	SystemTap SDT probes.
	(aarch64_stap_is_single_operand): New function.
	(aarch64_stap_parse_special_token): Likewise.
	(aarch64_linux_init_abi): Declare SystemTap SDT probe argument
	prefixes and suffixes.  Initialize gdbarch with them.
2013-12-28 14:14:11 -02:00
Will Newton 05feb1933f gdb/aarch64-linux-tdep.c: Call linux_init_abi.
If we are running on a Linux platform we should call linux_init_abi
in order to get all the useful hooks it enables.

gdb/ChangeLog:

2013-10-10  Will Newton  <will.newton@linaro.org>

	* aarch64-linux-tdep.c (aarch64_linux_init_abi): Call
	linux_init_abi.
2013-10-10 10:40:42 +00:00
Yufeng Zhang b5dbc8d429 gdb/
2013-02-08  Yufeng Zhang  <yufeng.zhang@arm.com>

	* aarch64-linux-tdep.c (AARCH64_LINUX_SIZEOF_GREGSET): Change the
	number of the registers from 36 to 34.
2013-02-08 12:46:28 +00:00
Marcus Shawcroft 45e25a3641 Enable TLS support in aarch64-linux-tdep.c
2013-02-07  Marcus Shawcroft  <marcus.shawcroft@arm.com>

        * aarch64-linux-tdep.c (aarch64_linux_init_abi): Call
        set_gdbarch_fetch_tls_load_module_address.
2013-02-07 10:49:29 +00:00
Marcus Shawcroft 1ae3db1961 Adding aarch64-linux-tdep support.
2013-02-04  Jim MacArthur  <jim.macarthur@arm.com>
            Marcus Shawcroft  <marcus.shawcroft@arm.com>
            Nigel Stephens  <nigel.stephens@arm.com>
            Yufeng Zhang  <yufeng.zhang@arm.com>

        * Makefile.in (ALL_64_TARGET_OBS): Add aarch64-linux-tdep.o.
        (ALLDEPFILES): Add aarch64-linux-tdep.c.
        * aarch64-linux-tdep.c: New file.
        * aarch64-linux-tdep.h: New file.
        * aarch64-tdep.h (gdbarch_tdep): Define gregset and fpregset.
        * configure.tgt: Add aarch64-none-linux-gnu.
2013-02-04 12:53:59 +00:00