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;
> }
next prev parent 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