binutils-gdb/gdb/progspace.h
Pedro Alves 55b11ddf16 Redesign mock environment for gdbarch selftests
A following patch will remove this hack from within regcache's
implementation:

  struct regcache *
  get_thread_arch_regcache (ptid_t ptid, struct gdbarch *gdbarch)
  {
    struct address_space *aspace;

    /* For the benefit of "maint print registers" & co when debugging an
       executable, allow dumping the regcache even when there is no
       thread selected (target_thread_address_space internal-errors if
       no address space is found).  Note that normal user commands will
       fail higher up on the call stack due to no
       target_has_registers.  */
    aspace = (ptid_equal (null_ptid, ptid)
	      ? NULL
	      : target_thread_address_space (ptid));

i.e., it'll no longer be possible to try to build a regcache for
null_ptid.  That change alone would regress the gdbarch self tests
though, causing this:

  (gdb) maintenance selftest
  [...]
  Running selftest register_to_value.
  src/gdb/inferior.c:309: internal-error: inferior* find_inferior_pid(int): Assertion `pid != 0' failed.
  A problem internal to GDB has been detected,
  further debugging may prove unreliable.
  Quit this debugging session? (y or n) FAIL: gdb.gdb/unittest.exp: maintenance selftest (GDB internal error)

The problem is that the way the mocking environment for those unit
tests is written is a bit fragile: it creates a special purpose
regcache (and sentinel's frame), using whatever is the current
inferior_ptid (usually null_ptid), and assumes get_current_regcache
will find that in the regcache::current_regcache list.

This commit changes the way the mock environment is created.  It
eliminates the special regcache and frame and instead creates a fuller
mock environment, with a custom mock target_ops, and then a mock
inferior and thread "running" on that target.

If there's already a running target when you type "maint selftest",
then we error out, instead of pushing a new target on top of the
existing one (and thus killing the debug session).  This results in:

  (gdb) maint selftest
  (...)
  Self test failed: arch i386: target already pushed
  Self test failed: arch i386:x86-64: target already pushed
  Self test failed: arch i386:x64-32: target already pushed
  Self test failed: arch i8086: target already pushed
  Self test failed: arch i386:intel: target already pushed
  Self test failed: arch i386:x86-64:intel: target already pushed
  Self test failed: arch i386:x64-32:intel: target already pushed
  Self test failed: arch i386:nacl: target already pushed
  Self test failed: arch i386:x86-64:nacl: target already pushed
  Self test failed: arch i386:x64-32:nacl: target already pushed
  Self test failed: self-test failed at /home/pedro/gdb/mygit/src/gdb/selftest-arch.c:86
  (...)
  Ran 19 unit tests, 1 failed

I think that's OK, because self tests are really meant to be run from
a clean state right after GDB is started.  I'm adding that erroring
out just as safe measure just in case someone types "maint selftest"
on the command line while already debugging something (as I've done
it).

(In my multi-target branch, where this patch originated from, we don't
actually need to error out, because there each inferior has its own
target stack).

Also, note that the current code was doing:

 current_inferior()->gdbarch = gdbarch;

without taking care to restore the previous gdbarch.  This means that
GDB's state was being left inconsistent after running the self tests,
further supporting the point that there's probably not much
expectation that mixing "maint selftests" and regular debugging in the
same GDB invocation really works.  This patch fixes that, regardless.

gdb/ChangeLog:
2017-10-04  Pedro Alves  <palves@redhat.com>

	* frame.c (create_test_frame): Delete.
	* frame.h (create_test_frame): Delete.
	* gdbarch-selftests.c: Include gdbthread.h and target.h.
	(class regcache_test): Delete.
	(test_target_has_registers, test_target_has_stack)
	(test_target_has_memory, test_target_prepare_to_store)
	(test_target_store_registers): New functions.
	(test_target_ops): New class.
	(register_to_value_test): Error out if there's already a
	process_stratum (or higher) target pushed.  Create a fuller mock
	environment, with mock target_ops, inferior, address space, thread
	and inferior_ptid.
	* progspace.c (struct address_space): Move to ...
	* progspace.h (struct address_space): ... here.
	* regcache.h (regcache::~regcache, regcache::raw_write)
	[GDB_SELF_TEST]: No longer virtual.
2017-10-04 18:21:09 +01:00

330 lines
13 KiB
C++

/* Program and address space management, for GDB, the GNU debugger.
Copyright (C) 2009-2017 Free Software Foundation, Inc.
This file is part of GDB.
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/>. */
#ifndef PROGSPACE_H
#define PROGSPACE_H
#include "target.h"
#include "vec.h"
#include "gdb_vecs.h"
#include "registry.h"
struct target_ops;
struct bfd;
struct objfile;
struct inferior;
struct exec;
struct address_space;
struct program_space_data;
struct address_space_data;
typedef struct so_list *so_list_ptr;
DEF_VEC_P (so_list_ptr);
/* A program space represents a symbolic view of an address space.
Roughly speaking, it holds all the data associated with a
non-running-yet program (main executable, main symbols), and when
an inferior is running and is bound to it, includes the list of its
mapped in shared libraries.
In the traditional debugging scenario, there's a 1-1 correspondence
among program spaces, inferiors and address spaces, like so:
pspace1 (prog1) <--> inf1(pid1) <--> aspace1
In the case of debugging more than one traditional unix process or
program, we still have:
|-----------------+------------+---------|
| pspace1 (prog1) | inf1(pid1) | aspace1 |
|----------------------------------------|
| pspace2 (prog1) | no inf yet | aspace2 |
|-----------------+------------+---------|
| pspace3 (prog2) | inf2(pid2) | aspace3 |
|-----------------+------------+---------|
In the former example, if inf1 forks (and GDB stays attached to
both processes), the new child will have its own program and
address spaces. Like so:
|-----------------+------------+---------|
| pspace1 (prog1) | inf1(pid1) | aspace1 |
|-----------------+------------+---------|
| pspace2 (prog1) | inf2(pid2) | aspace2 |
|-----------------+------------+---------|
However, had inf1 from the latter case vforked instead, it would
share the program and address spaces with its parent, until it
execs or exits, like so:
|-----------------+------------+---------|
| pspace1 (prog1) | inf1(pid1) | aspace1 |
| | inf2(pid2) | |
|-----------------+------------+---------|
When the vfork child execs, it is finally given new program and
address spaces.
|-----------------+------------+---------|
| pspace1 (prog1) | inf1(pid1) | aspace1 |
|-----------------+------------+---------|
| pspace2 (prog1) | inf2(pid2) | aspace2 |
|-----------------+------------+---------|
There are targets where the OS (if any) doesn't provide memory
management or VM protection, where all inferiors share the same
address space --- e.g. uClinux. GDB models this by having all
inferiors share the same address space, but, giving each its own
program space, like so:
|-----------------+------------+---------|
| pspace1 (prog1) | inf1(pid1) | |
|-----------------+------------+ |
| pspace2 (prog1) | inf2(pid2) | aspace1 |
|-----------------+------------+ |
| pspace3 (prog2) | inf3(pid3) | |
|-----------------+------------+---------|
The address space sharing matters for run control and breakpoints
management. E.g., did we just hit a known breakpoint that we need
to step over? Is this breakpoint a duplicate of this other one, or
do I need to insert a trap?
Then, there are targets where all symbols look the same for all
inferiors, although each has its own address space, as e.g.,
Ericsson DICOS. In such case, the model is:
|---------+------------+---------|
| | inf1(pid1) | aspace1 |
| +------------+---------|
| pspace | inf2(pid2) | aspace2 |
| +------------+---------|
| | inf3(pid3) | aspace3 |
|---------+------------+---------|
Note however, that the DICOS debug API takes care of making GDB
believe that breakpoints are "global". That is, although each
process does have its own private copy of data symbols (just like a
bunch of forks), to the breakpoints module, all processes share a
single address space, so all breakpoints set at the same address
are duplicates of each other, even breakpoints set in the data
space (e.g., call dummy breakpoints placed on stack). This allows
a simplification in the spaces implementation: we avoid caring for
a many-many links between address and program spaces. Either
there's a single address space bound to the program space
(traditional unix/uClinux), or, in the DICOS case, the address
space bound to the program space is mostly ignored. */
/* The program space structure. */
struct program_space
{
/* Pointer to next in linked list. */
struct program_space *next;
/* Unique ID number. */
int num;
/* The main executable loaded into this program space. This is
managed by the exec target. */
/* The BFD handle for the main executable. */
bfd *ebfd;
/* The last-modified time, from when the exec was brought in. */
long ebfd_mtime;
/* Similar to bfd_get_filename (exec_bfd) but in original form given
by user, without symbolic links and pathname resolved.
It needs to be freed by xfree. It is not NULL iff EBFD is not NULL. */
char *pspace_exec_filename;
/* The address space attached to this program space. More than one
program space may be bound to the same address space. In the
traditional unix-like debugging scenario, this will usually
match the address space bound to the inferior, and is mostly
used by the breakpoints module for address matches. If the
target shares a program space for all inferiors and breakpoints
are global, then this field is ignored (we don't currently
support inferiors sharing a program space if the target doesn't
make breakpoints global). */
struct address_space *aspace;
/* True if this program space's section offsets don't yet represent
the final offsets of the "live" address space (that is, the
section addresses still require the relocation offsets to be
applied, and hence we can't trust the section addresses for
anything that pokes at live memory). E.g., for qOffsets
targets, or for PIE executables, until we connect and ask the
target for the final relocation offsets, the symbols we've used
to set breakpoints point at the wrong addresses. */
int executing_startup;
/* True if no breakpoints should be inserted in this program
space. */
int breakpoints_not_allowed;
/* The object file that the main symbol table was loaded from
(e.g. the argument to the "symbol-file" or "file" command). */
struct objfile *symfile_object_file;
/* All known objfiles are kept in a linked list. This points to
the head of this list. */
struct objfile *objfiles;
/* The set of target sections matching the sections mapped into
this program space. Managed by both exec_ops and solib.c. */
struct target_section_table target_sections;
/* List of shared objects mapped into this space. Managed by
solib.c. */
struct so_list *so_list;
/* Number of calls to solib_add. */
unsigned solib_add_generation;
/* When an solib is added, it is also added to this vector. This
is so we can properly report solib changes to the user. */
VEC (so_list_ptr) *added_solibs;
/* When an solib is removed, its name is added to this vector.
This is so we can properly report solib changes to the user. */
VEC (char_ptr) *deleted_solibs;
/* Per pspace data-pointers required by other GDB modules. */
REGISTRY_FIELDS;
};
/* An address space. It is used for comparing if
pspaces/inferior/threads see the same address space and for
associating caches to each address space. */
struct address_space
{
int num;
/* Per aspace data-pointers required by other GDB modules. */
REGISTRY_FIELDS;
};
/* The object file that the main symbol table was loaded from (e.g. the
argument to the "symbol-file" or "file" command). */
#define symfile_objfile current_program_space->symfile_object_file
/* All known objfiles are kept in a linked list. This points to the
root of this list. */
#define object_files current_program_space->objfiles
/* The set of target sections matching the sections mapped into the
current program space. */
#define current_target_sections (&current_program_space->target_sections)
/* The list of all program spaces. There's always at least one. */
extern struct program_space *program_spaces;
/* The current program space. This is always non-null. */
extern struct program_space *current_program_space;
#define ALL_PSPACES(pspace) \
for ((pspace) = program_spaces; (pspace) != NULL; (pspace) = (pspace)->next)
/* Add a new empty program space, and assign ASPACE to it. Returns the
pointer to the new object. */
extern struct program_space *add_program_space (struct address_space *aspace);
/* Remove a program space from the program spaces list and release it. It is
an error to call this function while PSPACE is the current program space. */
extern void delete_program_space (struct program_space *pspace);
/* Returns the number of program spaces listed. */
extern int number_of_program_spaces (void);
/* Returns true iff there's no inferior bound to PSPACE. */
extern int program_space_empty_p (struct program_space *pspace);
/* Copies program space SRC to DEST. Copies the main executable file,
and the main symbol file. Returns DEST. */
extern struct program_space *clone_program_space (struct program_space *dest,
struct program_space *src);
/* Sets PSPACE as the current program space. This is usually used
instead of set_current_space_and_thread when the current
thread/inferior is not important for the operations that follow.
E.g., when accessing the raw symbol tables. If memory access is
required, then you should use switch_to_program_space_and_thread.
Otherwise, it is the caller's responsibility to make sure that the
currently selected inferior/thread matches the selected program
space. */
extern void set_current_program_space (struct program_space *pspace);
/* Save/restore the current program space. */
class scoped_restore_current_program_space
{
public:
scoped_restore_current_program_space ()
: m_saved_pspace (current_program_space)
{}
~scoped_restore_current_program_space ()
{ set_current_program_space (m_saved_pspace); }
DISABLE_COPY_AND_ASSIGN (scoped_restore_current_program_space);
private:
program_space *m_saved_pspace;
};
/* Create a new address space object, and add it to the list. */
extern struct address_space *new_address_space (void);
/* Maybe create a new address space object, and add it to the list, or
return a pointer to an existing address space, in case inferiors
share an address space. */
extern struct address_space *maybe_new_address_space (void);
/* Returns the integer address space id of ASPACE. */
extern int address_space_num (struct address_space *aspace);
/* Update all program spaces matching to address spaces. The user may
have created several program spaces, and loaded executables into
them before connecting to the target interface that will create the
inferiors. All that happens before GDB has a chance to know if the
inferiors will share an address space or not. Call this after
having connected to the target interface and having fetched the
target description, to fixup the program/address spaces
mappings. */
extern void update_address_spaces (void);
/* Reset saved solib data at the start of an solib event. This lets
us properly collect the data when calling solib_add, so it can then
later be printed. */
extern void clear_program_space_solib_cache (struct program_space *);
/* Keep a registry of per-pspace data-pointers required by other GDB
modules. */
DECLARE_REGISTRY (program_space);
/* Keep a registry of per-aspace data-pointers required by other GDB
modules. */
DECLARE_REGISTRY (address_space);
#endif