Files
binutils-gdb/gdb/i386-obsd-tdep.c
Simon Marchi a2e3cce344 gdb/solib: C++ify solib_ops
Convert solib_ops into an abstract base class (with abstract methods,
some of them with default implementations) and convert all the existing
solib_ops instances to solib_ops derived classes / implementations.

Prior to this patch, solib_ops is a structure holding function pointers,
of which there are only a handful of global instances (in the
`solib-*.c` files).  When passing an `solib_ops *` around, it's a
pointer to one of these instances.  After this patch, there are no more
global solib_ops instances.  Instances are created as needed and stored
in struct program_space.  These instances could eventually be made to
contain the program space-specific data, which is currently kept in
per-program space registries (I have some pending patches for that).

Prior to this patch, `gdbarch_so_ops` is a gdbarch method that returns a
pointer to the appropriate solib_ops implementation for the gdbarch.
This is replaced with the `gdbarch_make_solib_ops` method, which returns
a new instance of the appropriate solib_ops implementation for this
gdbarch.  This requires introducing some factory functions for the
various solib_ops implementation, to be used as `gdbarch_make_solib_ops`
callbacks.  For instance:

    solib_ops_up
    make_linux_ilp32_svr4_solib_ops ()
    {
      return std::make_unique<linux_ilp32_svr4_solib_ops> ();
    }

The previous code is full of cases of tdep files copying some base
solib_ops implementation, and overriding one or more function pointer
(see ppc_linux_init_abi, for instance).  I tried to convert all of this
is a class hierarchy.  I like that it's now possible to get a good
static view of all the existing solib_ops variants.  The hierarchy looks
like this:

    solib_ops
    ├── aix_solib_ops
    ├── darwin_solib_ops
    ├── dsbt_solib_ops
    ├── frv_solib_ops
    ├── rocm_solib_ops
    ├── svr4_solib_ops
    │   ├── ilp32_svr4_solib_ops
    │   ├── lp64_svr4_solib_ops
    │   ├── linux_ilp32_svr4_solib_ops
    │   │   ├── mips_linux_ilp32_svr4_solib_ops
    │   │   └── ppc_linux_ilp32_svr4_solib_ops
    │   ├── linux_lp64_svr4_solib_ops
    │   │   └── mips_linux_lp64_svr4_solib_ops
    │   ├── mips_nbsd_ilp32_svr4_solib_ops
    │   ├── mips_nbsd_lp64_svr4_solib_ops
    │   ├── mips_fbsd_ilp32_svr4_solib_ops
    │   └── mips_fbsd_lp64_svr4_solib_ops
    └── target_solib_ops
        └── windows_solib_ops

The solib-svr4 code has per-arch specialization to provide a
link_map_offsets, containing the offsets of the interesting fields in
`struct link_map` on that particular architecture.  Prior to this patch,
arches would set a callback returning the appropriate link_map_offsets
by calling `set_solib_svr4_fetch_link_map_offsets`, which also happened
to set the gdbarch's so_ops to `&svr_so_ops`.  I converted this to an
abstract virtual method of `struct svr4_solib_ops`, meaning that all
classes deriving from svr4_solib_ops must provide a method returning the
appropriate link_map_offsets for the architecture.  I renamed
`set_solib_svr4_fetch_link_map_offsets` to `set_solib_svr4_ops`.  This
function is still necessary because it also calls
set_gdbarch_iterate_over_objfiles_in_search_order, but if it was not for
that, we could get rid of it.

There is an instance of CRTP in mips-linux-tdep.c, because both
mips_linux_ilp32_svr4_solib_ops and mips_linux_lp64_svr4_solib_ops need
to derive from different SVR4 base classes (linux_ilp32_svr4_solib_ops
and linux_lp64_svr4_solib_ops), but they both want to override the
in_dynsym_resolve_code method with the same implementation.

The solib_ops::supports_namespaces method is new: the support for
namespaces was previously predicated by the presence or absence of a
find_solib_ns method.  It now needs to be explicit.

There is a new progspace::release_solib_ops method, which is only needed
for rocm_solib_ops.  For the moment, rocm_solib_ops replaces and wraps
the existing svr4_solib_ops instance, in order to combine the results of
the two.  The plan is to have a subsequent patch to allow program spaces to have
multiple solib_ops, removing the need for release_solib_ops.

Speaking of rocm_solib_ops: it previously overrode only a few methods by
copying svr4_solib_ops and overwriting some function pointers.  Now, it
needs to implement all the methods that svr4_solib_ops implements, in
order to forward the call.  Otherwise, the default solib_ops method
would be called, hiding the svr4_solib_ops implementation.  Again, this
can be removed once we have support for multiple solib_ops in a
program_space.

There is also a small change in how rocm_solib_ops is activated.  Prior
to this patch, it's done at the end of rocm_update_solib_list.  Since it
overrides the function pointer in the static svr4_solib_ops, and then
overwrites the host gdbarch, so_ops field, it's something that happens
only once.  After the patch though, we need to set rocm_solib_ops in all
the program spaces that appear.  We do this in
rocm_solib_target_inferior_created and in the new
rocm_solib_target_inferior_execd.  After this, I will explore doing a
change where rocm_solib_ops is only set when we detect the ROCm runtime
is loaded.

Change-Id: I5896b5bcbf8bdb024d67980380feba1ffefaa4c9
Approved-By: Pedro Alves <pedro@palves.net>
2025-06-26 14:08:31 -04:00

452 lines
13 KiB
C
Raw Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

/* Target-dependent code for OpenBSD/i386.
Copyright (C) 1988-2025 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/>. */
#include "arch-utils.h"
#include "extract-store-integer.h"
#include "frame.h"
#include "frame-unwind.h"
#include "gdbcore.h"
#include "regcache.h"
#include "regset.h"
#include "symtab.h"
#include "objfiles.h"
#include "osabi.h"
#include "target.h"
#include "trad-frame.h"
#include "obsd-tdep.h"
#include "i386-tdep.h"
#include "i387-tdep.h"
#include "solib-svr4.h"
#include "bsd-uthread.h"
/* Support for signal handlers. */
/* Since OpenBSD 3.2, the sigtramp routine is mapped at a random page
in virtual memory. The randomness makes it somewhat tricky to
detect it, but fortunately we can rely on the fact that the start
of the sigtramp routine is page-aligned. We recognize the
trampoline by looking for the code that invokes the sigreturn
system call. The offset where we can find that code varies from
release to release.
By the way, the mapping mentioned above is read-only, so you cannot
place a breakpoint in the signal trampoline. */
/* Default page size. */
static const int i386obsd_page_size = 4096;
/* Offset for sigreturn(2). */
static const int i386obsd_sigreturn_offset[] = {
0x0a, /* OpenBSD 3.2 */
0x14, /* OpenBSD 3.6 */
0x3a, /* OpenBSD 3.8 */
-1
};
/* Return whether THIS_FRAME corresponds to an OpenBSD sigtramp
routine. */
static int
i386obsd_sigtramp_p (const frame_info_ptr &this_frame)
{
CORE_ADDR pc = get_frame_pc (this_frame);
CORE_ADDR start_pc = (pc & ~(i386obsd_page_size - 1));
/* The call sequence invoking sigreturn(2). */
const gdb_byte sigreturn[] =
{
0xb8,
0x67, 0x00, 0x00, 0x00, /* movl $SYS_sigreturn, %eax */
0xcd, 0x80 /* int $0x80 */
};
size_t buflen = sizeof sigreturn;
const int *offset;
gdb_byte *buf;
const char *name;
/* If the function has a valid symbol name, it isn't a
trampoline. */
find_pc_partial_function (pc, &name, NULL, NULL);
if (name != NULL)
return 0;
/* If the function lives in a valid section (even without a starting
point) it isn't a trampoline. */
if (find_pc_section (pc) != NULL)
return 0;
/* Allocate buffer. */
buf = (gdb_byte *) alloca (buflen);
/* Loop over all offsets. */
for (offset = i386obsd_sigreturn_offset; *offset != -1; offset++)
{
/* If we can't read the instructions, return zero. */
if (!safe_frame_unwind_memory (this_frame, start_pc + *offset,
{buf, buflen}))
return 0;
/* Check for sigreturn(2). */
if (memcmp (buf, sigreturn, buflen) == 0)
return 1;
}
return 0;
}
/* Mapping between the general-purpose registers in `struct reg'
format and GDB's register cache layout. */
/* From <machine/reg.h>. */
static int i386obsd_r_reg_offset[] =
{
0 * 4, /* %eax */
1 * 4, /* %ecx */
2 * 4, /* %edx */
3 * 4, /* %ebx */
4 * 4, /* %esp */
5 * 4, /* %ebp */
6 * 4, /* %esi */
7 * 4, /* %edi */
8 * 4, /* %eip */
9 * 4, /* %eflags */
10 * 4, /* %cs */
11 * 4, /* %ss */
12 * 4, /* %ds */
13 * 4, /* %es */
14 * 4, /* %fs */
15 * 4 /* %gs */
};
/* Sigtramp routine location for OpenBSD 3.1 and earlier releases. */
CORE_ADDR i386obsd_sigtramp_start_addr = 0xbfbfdf20;
CORE_ADDR i386obsd_sigtramp_end_addr = 0xbfbfdff0;
/* From <machine/signal.h>. */
int i386obsd_sc_reg_offset[I386_NUM_GREGS] =
{
10 * 4, /* %eax */
9 * 4, /* %ecx */
8 * 4, /* %edx */
7 * 4, /* %ebx */
14 * 4, /* %esp */
6 * 4, /* %ebp */
5 * 4, /* %esi */
4 * 4, /* %edi */
11 * 4, /* %eip */
13 * 4, /* %eflags */
12 * 4, /* %cs */
15 * 4, /* %ss */
3 * 4, /* %ds */
2 * 4, /* %es */
1 * 4, /* %fs */
0 * 4 /* %gs */
};
/* From /usr/src/lib/libpthread/arch/i386/uthread_machdep.c. */
static int i386obsd_uthread_reg_offset[] =
{
11 * 4, /* %eax */
10 * 4, /* %ecx */
9 * 4, /* %edx */
8 * 4, /* %ebx */
-1, /* %esp */
6 * 4, /* %ebp */
5 * 4, /* %esi */
4 * 4, /* %edi */
12 * 4, /* %eip */
-1, /* %eflags */
13 * 4, /* %cs */
-1, /* %ss */
3 * 4, /* %ds */
2 * 4, /* %es */
1 * 4, /* %fs */
0 * 4 /* %gs */
};
/* Offset within the thread structure where we can find the saved
stack pointer (%esp). */
#define I386OBSD_UTHREAD_ESP_OFFSET 176
static void
i386obsd_supply_uthread (struct regcache *regcache,
int regnum, CORE_ADDR addr)
{
struct gdbarch *gdbarch = regcache->arch ();
enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
CORE_ADDR sp_addr = addr + I386OBSD_UTHREAD_ESP_OFFSET;
CORE_ADDR sp = 0;
gdb_byte buf[4];
int i;
gdb_assert (regnum >= -1);
if (regnum == -1 || regnum == I386_ESP_REGNUM)
{
int offset;
/* Fetch stack pointer from thread structure. */
sp = read_memory_unsigned_integer (sp_addr, 4, byte_order);
/* Adjust the stack pointer such that it looks as if we just
returned from _thread_machdep_switch. */
offset = i386obsd_uthread_reg_offset[I386_EIP_REGNUM] + 4;
store_unsigned_integer (buf, 4, byte_order, sp + offset);
regcache->raw_supply (I386_ESP_REGNUM, buf);
}
for (i = 0; i < ARRAY_SIZE (i386obsd_uthread_reg_offset); i++)
{
if (i386obsd_uthread_reg_offset[i] != -1
&& (regnum == -1 || regnum == i))
{
/* Fetch stack pointer from thread structure (if we didn't
do so already). */
if (sp == 0)
sp = read_memory_unsigned_integer (sp_addr, 4, byte_order);
/* Read the saved register from the stack frame. */
read_memory (sp + i386obsd_uthread_reg_offset[i], buf, 4);
regcache->raw_supply (i, buf);
}
}
}
static void
i386obsd_collect_uthread (const struct regcache *regcache,
int regnum, CORE_ADDR addr)
{
struct gdbarch *gdbarch = regcache->arch ();
enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
CORE_ADDR sp_addr = addr + I386OBSD_UTHREAD_ESP_OFFSET;
CORE_ADDR sp = 0;
gdb_byte buf[4];
int i;
gdb_assert (regnum >= -1);
if (regnum == -1 || regnum == I386_ESP_REGNUM)
{
int offset;
/* Calculate the stack pointer (frame pointer) that will be
stored into the thread structure. */
offset = i386obsd_uthread_reg_offset[I386_EIP_REGNUM] + 4;
regcache->raw_collect (I386_ESP_REGNUM, buf);
sp = extract_unsigned_integer (buf, 4, byte_order) - offset;
/* Store the stack pointer. */
write_memory_unsigned_integer (sp_addr, 4, byte_order, sp);
/* The stack pointer was (potentially) modified. Make sure we
build a proper stack frame. */
regnum = -1;
}
for (i = 0; i < ARRAY_SIZE (i386obsd_uthread_reg_offset); i++)
{
if (i386obsd_uthread_reg_offset[i] != -1
&& (regnum == -1 || regnum == i))
{
/* Fetch stack pointer from thread structure (if we didn't
calculate it already). */
if (sp == 0)
sp = read_memory_unsigned_integer (sp_addr, 4, byte_order);
/* Write the register into the stack frame. */
regcache->raw_collect (i, buf);
write_memory (sp + i386obsd_uthread_reg_offset[i], buf, 4);
}
}
}
/* Kernel debugging support. */
/* From <machine/frame.h>. Note that %esp and %ess are only saved in
a trap frame when entering the kernel from user space. */
static int i386obsd_tf_reg_offset[] =
{
10 * 4, /* %eax */
9 * 4, /* %ecx */
8 * 4, /* %edx */
7 * 4, /* %ebx */
-1, /* %esp */
6 * 4, /* %ebp */
5 * 4, /* %esi */
4 * 4, /* %edi */
13 * 4, /* %eip */
15 * 4, /* %eflags */
14 * 4, /* %cs */
-1, /* %ss */
3 * 4, /* %ds */
2 * 4, /* %es */
0 * 4, /* %fs */
1 * 4 /* %gs */
};
static struct trad_frame_cache *
i386obsd_trapframe_cache (const frame_info_ptr &this_frame, void **this_cache)
{
struct gdbarch *gdbarch = get_frame_arch (this_frame);
enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
struct trad_frame_cache *cache;
CORE_ADDR func, sp, addr;
ULONGEST cs;
const char *name;
int i;
if (*this_cache)
return (struct trad_frame_cache *) *this_cache;
cache = trad_frame_cache_zalloc (this_frame);
*this_cache = cache;
func = get_frame_func (this_frame);
sp = get_frame_register_unsigned (this_frame, I386_ESP_REGNUM);
find_pc_partial_function (func, &name, NULL, NULL);
if (name && startswith (name, "Xintr"))
addr = sp + 8; /* It's an interrupt frame. */
else
addr = sp;
for (i = 0; i < ARRAY_SIZE (i386obsd_tf_reg_offset); i++)
if (i386obsd_tf_reg_offset[i] != -1)
trad_frame_set_reg_addr (cache, i, addr + i386obsd_tf_reg_offset[i]);
/* Read %cs from trap frame. */
addr += i386obsd_tf_reg_offset[I386_CS_REGNUM];
cs = read_memory_unsigned_integer (addr, 4, byte_order);
if ((cs & I386_SEL_RPL) == I386_SEL_UPL)
{
/* Trap from user space; terminate backtrace. */
trad_frame_set_id (cache, outer_frame_id);
}
else
{
/* Construct the frame ID using the function start. */
trad_frame_set_id (cache, frame_id_build (sp + 8, func));
}
return cache;
}
static void
i386obsd_trapframe_this_id (const frame_info_ptr &this_frame,
void **this_cache, struct frame_id *this_id)
{
struct trad_frame_cache *cache =
i386obsd_trapframe_cache (this_frame, this_cache);
trad_frame_get_id (cache, this_id);
}
static struct value *
i386obsd_trapframe_prev_register (const frame_info_ptr &this_frame,
void **this_cache, int regnum)
{
struct trad_frame_cache *cache =
i386obsd_trapframe_cache (this_frame, this_cache);
return trad_frame_get_register (cache, this_frame, regnum);
}
static int
i386obsd_trapframe_sniffer (const struct frame_unwind *self,
const frame_info_ptr &this_frame,
void **this_prologue_cache)
{
ULONGEST cs;
const char *name;
/* Check Current Privilege Level and bail out if we're not executing
in kernel space. */
cs = get_frame_register_unsigned (this_frame, I386_CS_REGNUM);
if ((cs & I386_SEL_RPL) == I386_SEL_UPL)
return 0;
find_pc_partial_function (get_frame_pc (this_frame), &name, NULL, NULL);
return (name && (strcmp (name, "calltrap") == 0
|| strcmp (name, "syscall1") == 0
|| startswith (name, "Xintr")
|| startswith (name, "Xsoft")));
}
static const struct frame_unwind_legacy i386obsd_trapframe_unwind (
"i386 openbsd trap",
/* FIXME: kettenis/20051219: This really is more like an interrupt
frame, but SIGTRAMP_FRAME would print <signal handler called>,
which really is not what we want here. */
NORMAL_FRAME,
FRAME_UNWIND_ARCH,
default_frame_unwind_stop_reason,
i386obsd_trapframe_this_id,
i386obsd_trapframe_prev_register,
NULL,
i386obsd_trapframe_sniffer
);
static void
i386obsd_init_abi (struct gdbarch_info info, struct gdbarch *gdbarch)
{
i386_gdbarch_tdep *tdep = gdbarch_tdep<i386_gdbarch_tdep> (gdbarch);
/* Obviously OpenBSD is BSD-based. */
i386bsd_init_abi (info, gdbarch);
obsd_init_abi (info, gdbarch);
i386_elf_init_abi (info, gdbarch);
/* OpenBSD has a different `struct reg'. */
tdep->gregset_reg_offset = i386obsd_r_reg_offset;
tdep->gregset_num_regs = ARRAY_SIZE (i386obsd_r_reg_offset);
tdep->sizeof_gregset = 16 * 4;
/* OpenBSD uses -freg-struct-return by default. */
tdep->struct_return = reg_struct_return;
/* OpenBSD uses a different memory layout. */
tdep->sigtramp_start = i386obsd_sigtramp_start_addr;
tdep->sigtramp_end = i386obsd_sigtramp_end_addr;
tdep->sigtramp_p = i386obsd_sigtramp_p;
/* OpenBSD has a `struct sigcontext' that's different from the
original 4.3 BSD. */
tdep->sc_reg_offset = i386obsd_sc_reg_offset;
tdep->sc_num_regs = ARRAY_SIZE (i386obsd_sc_reg_offset);
/* OpenBSD provides a user-level threads implementation. */
bsd_uthread_set_supply_uthread (gdbarch, i386obsd_supply_uthread);
bsd_uthread_set_collect_uthread (gdbarch, i386obsd_collect_uthread);
/* Unwind kernel trap frames correctly. */
frame_unwind_prepend_unwinder (gdbarch, &i386obsd_trapframe_unwind);
/* OpenBSD ELF uses SVR4-style shared libraries. */
set_solib_svr4_ops (gdbarch, make_svr4_ilp32_solib_ops);
}
INIT_GDB_FILE (i386obsd_tdep)
{
gdbarch_register_osabi (bfd_arch_i386, 0, GDB_OSABI_OPENBSD,
i386obsd_init_abi);
}