Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Guinevere Larsen <guinevere@redhat.com>
To: "Schimpe, Christina" <christina.schimpe@intel.com>,
	gdb-patches@sourceware.org
Subject: Re: [PATCH 08/12] gdb: Handle shadow stack pointer register unwinding for amd64 linux.
Date: Thu, 30 Jan 2025 11:29:17 -0300	[thread overview]
Message-ID: <4f873c4a-a4ba-451f-8908-bc40ab93c43a@redhat.com> (raw)
In-Reply-To: <20241220200501.324191-9-christina.schimpe@intel.com>

On 12/20/24 5:04 PM, Schimpe, Christina wrote:
> Unwind the $pl3_ssp register.
> We now have an updated value for the shadow stack pointer when
> moving up or down the frame level.  Note that $pl3_ssp can
> become unavailable when moving to a frame before the shadow
> stack enablement.  In the example below, shadow stack is enabled
> in the function 'call1'.  Thus, when moving to a frame level above
> the function, $pl3_ssp will become unavaiable.
> Following the restriction of the linux kernel, implement the unwinding
> for amd64 linux only.
>
> Before this patch:
> ~~~
> Breakpoint 1, call2 (j=3) at sample.c:44
> 44	  return 42;
> (gdb) p $pl3_ssp
> $1 = (void *) 0x7ffff79ffff8
> (gdb) up
> 55	  call2 (3);
> (gdb) p $pl3_ssp
> $2 = (void *) 0x7ffff79ffff8
> (gdb) up
> 68	  call1 (43);
> (gdb) p $pl3_ssp
> $3 = (void *) 0x7ffff79ffff8
> ~~~
>
> After this patch:
> ~~~
> Breakpoint 1, call2 (j=3) at sample.c:44
> 44	  return 42;
> (gdb) p $pl3_ssp
> $1 = (void *) 0x7ffff79ffff8
> (gdb) up
> 55	  call2 (3);
> (gdb) p $pl3_ssp
> $2 = (void *) 0x7ffff7a00000
> (gdb) up
> 68	  call1 (43i);
> (gdb) p $pl3_ssp
> $3 = <unavailable>
> ~~~
>
> As we now have an updated value for each selected frame, the
> return command is now enabled for shadow stack enabled programs, too.
>
> We therefore add a test for the return command and shadow stack support,
> and for an updated shadow stack pointer after a frame level change.
> ---
>   gdb/amd64-linux-tdep.c                        | 69 +++++++++++++++
>   gdb/linux-tdep.c                              | 47 ++++++++++
>   gdb/linux-tdep.h                              |  7 ++
>   .../gdb.arch/amd64-shadow-stack-cmds.exp      | 88 +++++++++++++++++++
>   gdb/testsuite/gdb.arch/amd64-shadow-stack.c   | 13 +++
>   5 files changed, 224 insertions(+)
>   create mode 100644 gdb/testsuite/gdb.arch/amd64-shadow-stack-cmds.exp
>
> diff --git a/gdb/amd64-linux-tdep.c b/gdb/amd64-linux-tdep.c
> index 95f643b1217..895feac85e8 100644
> --- a/gdb/amd64-linux-tdep.c
> +++ b/gdb/amd64-linux-tdep.c
> @@ -45,6 +45,8 @@
>   #include "arch/amd64-linux-tdesc.h"
>   #include "inferior.h"
>   #include "x86-tdep.h"
> +#include "dwarf2/frame.h"
> +#include "frame-unwind.h"
>   
>   /* The syscall's XML filename for i386.  */
>   #define XML_SYSCALL_FILENAME_AMD64 "syscalls/amd64-linux.xml"
> @@ -1873,6 +1875,72 @@ amd64_linux_remove_non_address_bits_watchpoint (gdbarch *gdbarch,
>     return (addr & amd64_linux_lam_untag_mask ());
>   }
>   
> +static value *
> +amd64_linux_dwarf2_prev_ssp (const frame_info_ptr &this_frame,
> +			     void **this_cache, int regnum)
> +{
> +  value *v = frame_unwind_got_register (this_frame, regnum, regnum);
> +  gdb_assert (v != nullptr);
> +
> +  gdbarch *gdbarch = get_frame_arch (this_frame);
> +
> +  if (v->entirely_available () && !v->optimized_out ())
> +    {
> +      int size = register_size (gdbarch, regnum);
> +      bfd_endian byte_order = gdbarch_byte_order (gdbarch);
> +      CORE_ADDR ssp = extract_unsigned_integer (v->contents_all ().data (),
> +						size, byte_order);
> +
> +      /* Starting with v6.6., the Linux kernel supports CET shadow stack.
> +	 Using /proc/PID/smaps we can only check if the current shadow
> +	 stack pointer SSP points to shadow stack memory.  Only if this is
> +	 the case a valid previous shadow stack pointer can be
> +	 calculated.  */
> +      std::pair<CORE_ADDR, CORE_ADDR> range;
> +      if (linux_address_in_shadow_stack_mem_range (ssp, &range))
> +	{
> +	  /* The shadow stack grows downwards.  To compute the previous
> +	     shadow stack pointer, we need to increment SSP.
> +	     For x32 the shadow stack elements are still 64-bit aligned.
> +	     Thus, we cannot use gdbarch_addr_bit to compute the new stack
> +	     pointer.  */
> +	  const bfd_arch_info *binfo = gdbarch_bfd_arch_info (gdbarch);
> +	  const int bytes_per_word
> +	    = (binfo->bits_per_word / binfo->bits_per_byte);
In patch 10 of this series, you introduce 
amd64_linux_shadow_stack_element_size_aligned to simplify this 
calculation. is there any reason you didn't introduce it here?

-- 
Cheers,
Guinevere Larsen
She/Her/Hers

> +	  CORE_ADDR new_ssp = ssp + bytes_per_word;
> +
> +	  /* If NEW_SSP points to the end of or before (<=) the current
> +	     shadow stack memory range we consider NEW_SSP as valid (but
> +	     empty).  */
> +	  if (new_ssp <= range.second)
> +	    return frame_unwind_got_address (this_frame, regnum, new_ssp);
> +	}
> +    }
> +
> +  /* Return a value which is marked as unavailable in case we could not
> +     calculate a valid previous shadow stack pointer.  */
> +  value *retval
> +    = value::allocate_register (get_next_frame_sentinel_okay (this_frame),
> +				regnum, register_type (gdbarch, regnum));
> +  retval->mark_bytes_unavailable (0, retval->type ()->length ());
> +  return retval;
> +}
> +
> +static void
> +amd64_init_reg (gdbarch *gdbarch, int regnum, dwarf2_frame_state_reg *reg,
> +		const frame_info_ptr &this_frame)
> +{
> +  if (regnum == gdbarch_pc_regnum (gdbarch))
> +    reg->how = DWARF2_FRAME_REG_RA;
> +  else if (regnum == gdbarch_sp_regnum (gdbarch))
> +    reg->how = DWARF2_FRAME_REG_CFA;
> +  else if (regnum == AMD64_PL3_SSP_REGNUM)
> +    {
> +      reg->how = DWARF2_FRAME_REG_FN;
> +      reg->loc.fn = amd64_linux_dwarf2_prev_ssp;
> +    }
> +}
> +
>   static void
>   amd64_linux_init_abi_common(struct gdbarch_info info, struct gdbarch *gdbarch,
>   			    int num_disp_step_buffers)
> @@ -1927,6 +1995,7 @@ amd64_linux_init_abi_common(struct gdbarch_info info, struct gdbarch *gdbarch,
>   
>     set_gdbarch_remove_non_address_bits_watchpoint
>       (gdbarch, amd64_linux_remove_non_address_bits_watchpoint);
> +  dwarf2_frame_set_init_reg (gdbarch, amd64_init_reg);
>   }
>   
>   static void
> diff --git a/gdb/linux-tdep.c b/gdb/linux-tdep.c
> index d3452059ce2..3a20b12c3d7 100644
> --- a/gdb/linux-tdep.c
> +++ b/gdb/linux-tdep.c
> @@ -47,6 +47,7 @@
>   
>   #include <ctype.h>
>   #include <unordered_map>
> +#include <algorithm>
>   
>   /* This enum represents the values that the user can choose when
>      informing the Linux kernel about which memory mappings will be
> @@ -96,6 +97,10 @@ struct smaps_vmflags
>       /* Memory map has memory tagging enabled.  */
>   
>       unsigned int memory_tagging : 1;
> +
> +    /* Memory map used for shadow stack.  */
> +
> +    unsigned int shadow_stack_memory : 1;
>     };
>   
>   /* Data structure that holds the information contained in the
> @@ -537,6 +542,8 @@ decode_vmflags (char *p, struct smaps_vmflags *v)
>   	v->shared_mapping = 1;
>         else if (strcmp (s, "mt") == 0)
>   	v->memory_tagging = 1;
> +      else if (strcmp (s, "ss") == 0)
> +	v->shadow_stack_memory = 1;
>       }
>   }
>   
> @@ -2744,6 +2751,46 @@ show_dump_excluded_mappings (struct ui_file *file, int from_tty,
>   		      " flag is %s.\n"), value);
>   }
>   
> +/* See linux-tdep.h.  */
> +
> +bool
> +linux_address_in_shadow_stack_mem_range
> +  (CORE_ADDR addr, std::pair<CORE_ADDR, CORE_ADDR> *range)
> +{
> +  if (!target_has_execution () || current_inferior ()->fake_pid_p)
> +    return false;
> +
> +  const int pid = current_inferior ()->pid;
> +
> +  std::string smaps_file = string_printf ("/proc/%d/smaps", pid);
> +
> +  gdb::unique_xmalloc_ptr<char> data
> +    = target_fileio_read_stralloc (nullptr, smaps_file.c_str ());
> +
> +  if (data == nullptr)
> +    return false;
> +
> +  const std::vector<smaps_data> smaps
> +    = parse_smaps_data (data.get (), std::move (smaps_file));
> +
> +  auto find_addr_mem_range = [&addr] (const smaps_data &map)
> +    {
> +      bool addr_in_mem_range
> +	= (addr >= map.start_address && addr < map.end_address);
> +      return (addr_in_mem_range && map.vmflags.shadow_stack_memory);
> +    };
> +  auto it = std::find_if (smaps.begin (), smaps.end (), find_addr_mem_range);
> +
> +  if (it != smaps.end ())
> +    {
> +      range->first = it->start_address;
> +      range->second = it->end_address;
> +      return true;
> +    }
> +
> +  return false;
> +}
> +
>   /* To be called from the various GDB_OSABI_LINUX handlers for the
>      various GNU/Linux architectures and machine types.
>   
> diff --git a/gdb/linux-tdep.h b/gdb/linux-tdep.h
> index bf4220bf75b..97dcc75a79c 100644
> --- a/gdb/linux-tdep.h
> +++ b/gdb/linux-tdep.h
> @@ -117,4 +117,11 @@ extern CORE_ADDR linux_get_hwcap2 ();
>   extern struct link_map_offsets *linux_ilp32_fetch_link_map_offsets ();
>   extern struct link_map_offsets *linux_lp64_fetch_link_map_offsets ();
>   
> +/* Returns true if ADDR belongs to a shadow stack memory range.  If this
> +   is the case, assign the shadow stack memory range to RANGE
> +   [start_address, end_address).  */
> +
> +extern bool linux_address_in_shadow_stack_mem_range
> +  (CORE_ADDR addr, std::pair<CORE_ADDR, CORE_ADDR> *range);
> +
>   #endif /* GDB_LINUX_TDEP_H */
> diff --git a/gdb/testsuite/gdb.arch/amd64-shadow-stack-cmds.exp b/gdb/testsuite/gdb.arch/amd64-shadow-stack-cmds.exp
> new file mode 100644
> index 00000000000..17f32ce3964
> --- /dev/null
> +++ b/gdb/testsuite/gdb.arch/amd64-shadow-stack-cmds.exp
> @@ -0,0 +1,88 @@
> +# Copyright 2018-2024 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/>.
> +
> +# Test shadow stack enabling for frame level update and the return command.
> +
> +require allow_ssp_tests
> +
> +standard_testfile amd64-shadow-stack.c
> +
> +save_vars { ::env(GLIBC_TUNABLES) } {
> +
> +    append_environment GLIBC_TUNABLES "glibc.cpu.hwcaps" "SHSTK"
> +
> +    if { [prepare_for_testing "failed to prepare" ${testfile} ${srcfile} \
> +	  {debug additional_flags="-fcf-protection=return"}] } {
> +	return -1
> +    }
> +
> +    clean_restart ${binfile}
> +    if { ![runto_main] } {
> +	return -1
> +    }
> +
> +    set call1_line [ gdb_get_line_number "break call1" ]
> +    set call2_line [ gdb_get_line_number "break call2" ]
> +
> +    # Extract shadow stack pointer inside main, call1 and call2 function.
> +    gdb_breakpoint $call1_line
> +    gdb_breakpoint $call2_line
> +    set ssp_main [get_valueof /x "\$pl3_ssp" 0 "get value of ssp in main"]
> +    gdb_continue_to_breakpoint "break call1" ".*break call1.*"
> +    set ssp_call1 [get_valueof /x "\$pl3_ssp" 0 "get value of ssp in call1"]
> +    gdb_continue_to_breakpoint "break call2" ".*break call2.*"
> +    set ssp_call2 [get_valueof /x "\$pl3_ssp" 0 "get value of ssp in call2"]
> +
> +    with_test_prefix "test frame level update" {
> +	gdb_test "up" "call1.*" "move to frame 1"
> +	gdb_test "print /x \$pl3_ssp" "= $ssp_call1" "check pl3_ssp of frame 1"
> +	gdb_test "up" "main.*" "move to frame 2"
> +	gdb_test "print /x \$pl3_ssp" "= $ssp_main" "check pl3_ssp of frame 2"
> +	gdb_test "frame 0" "call2.*" "move to frame 0"
> +	gdb_test "print /x \$pl3_ssp" "= $ssp_call2" "check pl3_ssp of frame 0"
> +    }
> +
> +    with_test_prefix "test return from current frame" {
> +	gdb_test "return (int) 1" "#0.*call1.*" \
> +	"Test shadow stack return from current frame" \
> +	"Make.*return now\\? \\(y or n\\) " "y"
> +
> +	# Potential CET violations often only occur after resuming normal execution.
> +	# Therefore, it is important to test normal program continuation after
> +	# testing the return command.
> +	gdb_continue_to_end
> +    }
> +
> +    clean_restart ${binfile}
> +    if { ![runto_main] } {
> +	return -1
> +    }
> +
> +    with_test_prefix "test return from past frame" {
> +	gdb_breakpoint $call2_line
> +	gdb_continue_to_breakpoint "break call2" ".*break call2.*"
> +
> +	gdb_test "frame 1" ".*in call1.*"
> +
> +	gdb_test "return (int) 1" "#0.*main.*" \
> +	    "Test shadow stack return from past frame" \
> +	    "Make.*return now\\? \\(y or n\\) " "y"
> +
> +	# Potential CET violations often only occur after resuming normal execution.
> +	# Therefore, it is important to test normal program continuation after
> +	# testing the return command.
> +	gdb_continue_to_end
> +    }
> +}
> diff --git a/gdb/testsuite/gdb.arch/amd64-shadow-stack.c b/gdb/testsuite/gdb.arch/amd64-shadow-stack.c
> index 643ef2d5f56..80389730bc2 100644
> --- a/gdb/testsuite/gdb.arch/amd64-shadow-stack.c
> +++ b/gdb/testsuite/gdb.arch/amd64-shadow-stack.c
> @@ -15,8 +15,21 @@
>    You should have received a copy of the GNU General Public License
>    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
>   
> +static int
> +call2 ()
> +{
> +  return 42; /* break call2.  */
> +}
> +
> +static int
> +call1 ()
> +{
> +  return call2 (); /* break call1.  */
> +}
> +
>   int
>   main ()
>   {
> +  call1 (); /* break main.  */
>     return 0;
>   }


  reply	other threads:[~2025-01-30 14:29 UTC|newest]

Thread overview: 72+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-12-20 20:04 [PATCH 00/12] Add CET shadow stack support Schimpe, Christina
2024-12-20 20:04 ` [PATCH 01/12] gdb, testsuite: Rename set_sanitizer_default to append_environment Schimpe, Christina
2025-01-28 13:45   ` Guinevere Larsen
2025-01-30 13:07     ` Schimpe, Christina
2025-01-30 14:27       ` Tom de Vries
2025-01-30 16:39         ` Schimpe, Christina
2024-12-20 20:04 ` [PATCH 02/12] gdbserver: Add optional runtime register set type Schimpe, Christina
2025-01-28 13:35   ` Guinevere Larsen
2025-01-30 10:28     ` Schimpe, Christina
2025-01-30 13:53       ` Guinevere Larsen
2025-01-30 17:43         ` Schimpe, Christina
2025-02-06  2:59       ` Thiago Jung Bauermann
2025-02-06 12:15         ` Schimpe, Christina
2024-12-20 20:04 ` [PATCH 03/12] gdbserver: Add assert in x86_linux_read_description Schimpe, Christina
2025-02-06  3:00   ` Thiago Jung Bauermann
2024-12-20 20:04 ` [PATCH 04/12] gdb: Sync up x86-gcc-cpuid.h with cpuid.h from gcc 14 branch Schimpe, Christina
2025-02-06  3:03   ` Thiago Jung Bauermann
2025-02-06 12:23     ` Schimpe, Christina
2024-12-20 20:04 ` [PATCH 05/12] gdb, gdbserver: Use xstate_bv for target description creation on x86 Schimpe, Christina
2025-01-30 14:51   ` Guinevere Larsen
2025-01-30 16:45     ` Schimpe, Christina
2025-02-06  3:09   ` Thiago Jung Bauermann
2025-02-06 12:33     ` Schimpe, Christina
2024-12-20 20:04 ` [PATCH 06/12] gdb, gdbserver: Add support of Intel shadow stack pointer register Schimpe, Christina
2025-02-06  3:13   ` Thiago Jung Bauermann
2025-02-06 14:33     ` Schimpe, Christina
2025-02-08  3:44       ` Thiago Jung Bauermann
2024-12-20 20:04 ` [PATCH 07/12] gdb, bfd: amd64 linux coredump support with shadow stack Schimpe, Christina
2025-02-06  3:15   ` Thiago Jung Bauermann
2025-02-07 11:54     ` Schimpe, Christina
2024-12-20 20:04 ` [PATCH 08/12] gdb: Handle shadow stack pointer register unwinding for amd64 linux Schimpe, Christina
2025-01-30 14:29   ` Guinevere Larsen [this message]
2025-01-30 16:11     ` Schimpe, Christina
2025-01-30 16:13       ` Guinevere Larsen
2025-01-30 16:40         ` Schimpe, Christina
2025-02-06  3:30   ` Thiago Jung Bauermann
2025-02-06 14:40     ` Schimpe, Christina
2024-12-20 20:04 ` [PATCH 09/12] gdb, gdbarch: Enable inferior calls for shadow stack support Schimpe, Christina
2025-02-06  3:31   ` Thiago Jung Bauermann
2025-02-06 15:07     ` Schimpe, Christina
2025-02-08  3:57       ` Thiago Jung Bauermann
2025-02-10  8:37         ` Schimpe, Christina
2024-12-20 20:04 ` [PATCH 10/12] gdb: Implement amd64 linux shadow stack support for inferior calls Schimpe, Christina
2025-02-06  3:34   ` Thiago Jung Bauermann
2025-02-07 11:55     ` Schimpe, Christina
2024-12-20 20:05 ` [PATCH 11/12] gdb, gdbarch: Introduce gdbarch method to get the shadow stack pointer Schimpe, Christina
2025-01-28 20:27   ` Guinevere Larsen
2025-01-30 10:33     ` Luis Machado
2025-01-30 12:34       ` Schimpe, Christina
2025-01-30 13:42         ` Guinevere Larsen
2025-02-06  3:35   ` Thiago Jung Bauermann
2025-02-07 12:01     ` Schimpe, Christina
2025-02-08  4:03       ` Thiago Jung Bauermann
2025-02-10  8:58         ` Schimpe, Christina
2025-02-11  1:53           ` Thiago Jung Bauermann
2025-02-15  3:45             ` Thiago Jung Bauermann
2025-02-16 10:45               ` Schimpe, Christina
2025-02-20  8:48                 ` Schimpe, Christina
2025-02-21  5:10                   ` Thiago Jung Bauermann
2025-02-21  9:41                     ` Schimpe, Christina
2024-12-20 20:05 ` [PATCH 12/12] gdb: Enable displaced stepping with shadow stack on amd64 linux Schimpe, Christina
2024-12-20 20:14   ` Eli Zaretskii
2025-01-02  9:04     ` Schimpe, Christina
2025-01-02  9:15       ` Eli Zaretskii
2025-02-06  3:37   ` Thiago Jung Bauermann
2025-01-16 14:01 ` [PING][PATCH 00/12] Add CET shadow stack support Schimpe, Christina
2025-01-27  9:44   ` [PING*2][PATCH " Schimpe, Christina
2025-01-30 15:01 ` [PATCH " Guinevere Larsen
2025-01-30 17:46   ` Schimpe, Christina
2025-02-04  3:57     ` Thiago Jung Bauermann
2025-02-04  9:40       ` Schimpe, Christina
2025-02-06  3:44         ` Thiago Jung Bauermann

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4f873c4a-a4ba-451f-8908-bc40ab93c43a@redhat.com \
    --to=guinevere@redhat.com \
    --cc=christina.schimpe@intel.com \
    --cc=gdb-patches@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox