From: Bagas Sanjaya <bagasdotme@gmail.com>
To: Deepak Gupta <debug@rivosinc.com>,
paul.walmsley@sifive.com, palmer@sifive.com, conor@kernel.org,
linux-doc@vger.kernel.org, linux-riscv@lists.infradead.org,
linux-kernel@vger.kernel.org, devicetree@vger.kernel.org,
linux-fsdevel@vger.kernel.org, linux-mm@kvack.org,
linux-arch@vger.kernel.org, linux-kselftest@vger.kernel.org
Cc: corbet@lwn.net, palmer@dabbelt.com, aou@eecs.berkeley.edu,
robh@kernel.org, krzk+dt@kernel.org, oleg@redhat.com,
tglx@linutronix.de, mingo@redhat.com, bp@alien8.de,
dave.hansen@linux.intel.com, x86@kernel.org, hpa@zytor.com,
peterz@infradead.org, akpm@linux-foundation.org, arnd@arndb.de,
ebiederm@xmission.com, kees@kernel.org, Liam.Howlett@oracle.com,
vbabka@suse.cz, lorenzo.stoakes@oracle.com, shuah@kernel.org,
brauner@kernel.org, samuel.holland@sifive.com,
andy.chiu@sifive.com, jerry.shih@sifive.com,
greentime.hu@sifive.com, charlie@rivosinc.com, evan@rivosinc.com,
cleger@rivosinc.com, xiao.w.wang@intel.com,
ajones@ventanamicro.com, anup@brainfault.org,
mchitale@ventanamicro.com, atishp@rivosinc.com,
sameo@rivosinc.com, bjorn@rivosinc.com, alexghiti@rivosinc.com,
david@redhat.com, libang.li@antgroup.com, jszhang@kernel.org,
leobras@redhat.com, guoren@kernel.org, samitolvanen@google.com,
songshuaishuai@tinylab.org, costa.shul@redhat.com,
bhe@redhat.com, zong.li@sifive.com, puranjay@kernel.org,
namcaov@gmail.com, antonb@tenstorrent.com, sorear@fastmail.com,
quic_bjorande@quicinc.com, ancientmodern4@gmail.com,
ben.dooks@codethink.co.uk, quic_zhonhan@quicinc.com,
cuiyunhui@bytedance.com, yang.lee@linux.alibaba.com,
ke.zhao@shingroup.cn, sunilvl@ventanamicro.com,
tanzhasanwork@gmail.com, schwab@suse.de, dawei.li@shingroup.cn,
rppt@kernel.org, willy@infradead.org, usama.anjum@collabora.com,
osalvador@suse.de, ryan.roberts@arm.com, andrii@kernel.org,
alx@kernel.org, catalin.marinas@arm.com, broonie@kernel.org,
revest@chromium.org, bgray@linux.ibm.com, deller@gmx.de,
zev@bewilderbeest.net
Subject: Re: [PATCH v4 28/30] riscv: Documentation for landing pad / indirect branch tracking
Date: Mon, 16 Sep 2024 09:41:09 +0700 [thread overview]
Message-ID: <ZueaxRZgIf0crs4a@archie.me> (raw)
In-Reply-To: <20240912231650.3740732-29-debug@rivosinc.com>
[-- Attachment #1: Type: text/plain, Size: 11329 bytes --]
On Thu, Sep 12, 2024 at 04:16:47PM -0700, Deepak Gupta wrote:
> Adding documentation on landing pad aka indirect branch tracking on riscv
> and kernel interfaces exposed so that user tasks can enable it.
>
> Signed-off-by: Deepak Gupta <debug@rivosinc.com>
> ---
> Documentation/arch/riscv/zicfilp.rst | 104 +++++++++++++++++++++++++++
> 1 file changed, 104 insertions(+)
> create mode 100644 Documentation/arch/riscv/zicfilp.rst
Don't forget to add toctree entry:
---- >8 ----
diff --git a/Documentation/arch/riscv/index.rst b/Documentation/arch/riscv/index.rst
index eecf347ce84944..be7237b6968213 100644
--- a/Documentation/arch/riscv/index.rst
+++ b/Documentation/arch/riscv/index.rst
@@ -14,6 +14,7 @@ RISC-V architecture
uabi
vector
cmodx
+ zicfilp
features
> +Function pointers live in read-write memory and thus are susceptible to corruption
> +and allows an adversary to reach any program counter (PC) in address space. On
> +RISC-V zicfilp extension enforces a restriction on such indirect control transfers
> +
> + - indirect control transfers must land on a landing pad instruction `lpad`.
> + There are two exception to this rule
> + - rs1 = x1 or rs1 = x5, i.e. a return from a function and returns are
> + protected using shadow stack (see zicfiss.rst)
> +
> + - rs1 = x7. On RISC-V compiler usually does below to reach function
> + which is beyond the offset possible J-type instruction.
> +
> + "auipc x7, <imm>"
> + "jalr (x7)"
> +
> + Such form of indirect control transfer are still immutable and don't rely
> + on memory and thus rs1=x7 is exempted from tracking and considered software
> + guarded jumps.
Sphinx reports new htmldocs warnings:
Documentation/arch/riscv/zicfilp.rst:30: ERROR: Unexpected indentation.
Documentation/arch/riscv/zicfilp.rst:96: ERROR: Unexpected indentation.
I have to fix up the lists:
---- >8 ----
diff --git a/Documentation/arch/riscv/zicfilp.rst b/Documentation/arch/riscv/zicfilp.rst
index 23013ee711ac2c..c0fad1b5caa3d8 100644
--- a/Documentation/arch/riscv/zicfilp.rst
+++ b/Documentation/arch/riscv/zicfilp.rst
@@ -23,22 +23,24 @@ flow integrity (CFI) of the program.
Function pointers live in read-write memory and thus are susceptible to corruption
and allows an adversary to reach any program counter (PC) in address space. On
-RISC-V zicfilp extension enforces a restriction on such indirect control transfers
+RISC-V zicfilp extension enforces a restriction on such indirect control
+transfers:
- - indirect control transfers must land on a landing pad instruction `lpad`.
- There are two exception to this rule
- - rs1 = x1 or rs1 = x5, i.e. a return from a function and returns are
- protected using shadow stack (see zicfiss.rst)
+- indirect control transfers must land on a landing pad instruction `lpad`.
+ There are two exception to this rule:
- - rs1 = x7. On RISC-V compiler usually does below to reach function
- which is beyond the offset possible J-type instruction.
+ - rs1 = x1 or rs1 = x5, i.e. a return from a function and returns are
+ protected using shadow stack (see zicfiss.rst)
- "auipc x7, <imm>"
- "jalr (x7)"
+ - rs1 = x7. On RISC-V compiler usually does below to reach function
+ which is beyond the offset possible J-type instruction.
- Such form of indirect control transfer are still immutable and don't rely
- on memory and thus rs1=x7 is exempted from tracking and considered software
- guarded jumps.
+ "auipc x7, <imm>"
+ "jalr (x7)"
+
+ Such form of indirect control transfer are still immutable and don't rely
+ on memory and thus rs1=x7 is exempted from tracking and considered software
+ guarded jumps.
`lpad` instruction is pseudo of `auipc rd, <imm_20bit>` with `rd=x0`` and is a HINT
nop. `lpad` instruction must be aligned on 4 byte boundary and compares 20 bit
@@ -92,10 +94,11 @@ to lock current settings.
--------------------------------------------------
Pertaining to indirect branch tracking, CPU raises software check exception in
-following conditions
- - missing `lpad` after indirect call / jmp
- - `lpad` not on 4 byte boundary
- - `imm_20bit` embedded in `lpad` instruction doesn't match with `x7`
+following conditions:
+
+- missing `lpad` after indirect call / jmp
+- `lpad` not on 4 byte boundary
+- `imm_20bit` embedded in `lpad` instruction doesn't match with `x7`
In all 3 cases, `*tval = 2` is captured and software check exception is raised
(cause=18)
> +
> +`lpad` instruction is pseudo of `auipc rd, <imm_20bit>` with `rd=x0`` and is a HINT
> +nop. `lpad` instruction must be aligned on 4 byte boundary and compares 20 bit
> +immediate withx7. If `imm_20bit` == 0, CPU don't perform any comparision with x7. If
> +`imm_20bit` != 0, then `imm_20bit` must match x7 else CPU will raise
> +`software check exception` (cause=18)with `*tval = 2`.
> +
Also inline identifiers/keywords to be consistent with rest of riscv docs:
---- >8 ----
diff --git a/Documentation/arch/riscv/zicfilp.rst b/Documentation/arch/riscv/zicfilp.rst
index c0fad1b5caa3d8..b0a766098f2335 100644
--- a/Documentation/arch/riscv/zicfilp.rst
+++ b/Documentation/arch/riscv/zicfilp.rst
@@ -26,38 +26,38 @@ and allows an adversary to reach any program counter (PC) in address space. On
RISC-V zicfilp extension enforces a restriction on such indirect control
transfers:
-- indirect control transfers must land on a landing pad instruction `lpad`.
+- indirect control transfers must land on a landing pad instruction ``lpad``.
There are two exception to this rule:
- rs1 = x1 or rs1 = x5, i.e. a return from a function and returns are
protected using shadow stack (see zicfiss.rst)
- rs1 = x7. On RISC-V compiler usually does below to reach function
- which is beyond the offset possible J-type instruction.
+ which is beyond the offset possible J-type instruction::
- "auipc x7, <imm>"
- "jalr (x7)"
+ auipc x7, <imm>
+ jalr (x7)
Such form of indirect control transfer are still immutable and don't rely
on memory and thus rs1=x7 is exempted from tracking and considered software
guarded jumps.
-`lpad` instruction is pseudo of `auipc rd, <imm_20bit>` with `rd=x0`` and is a HINT
-nop. `lpad` instruction must be aligned on 4 byte boundary and compares 20 bit
-immediate withx7. If `imm_20bit` == 0, CPU don't perform any comparision with x7. If
-`imm_20bit` != 0, then `imm_20bit` must match x7 else CPU will raise
-`software check exception` (cause=18)with `*tval = 2`.
+``lpad`` instruction is pseudo of ``auipc rd, <imm_20bit>`` with ``rd=x0`` and
+is a HINT nop. ``lpad`` instruction must be aligned on 4 byte boundary and
+compares 20 bit immediate with x7. If ``imm_20bit`` == 0, CPU don't perform any
+comparision with x7. If ``imm_20bit`` != 0, then ``imm_20bit`` must match x7
+else CPU will raise software check exception (cause=18) with ``*tval = 2``.
Compiler can generate a hash over function signatures and setup them (truncated
-to 20bit) in x7 at callsites and function prologues can have `lpad` with same
+to 20bit) in x7 at callsites and function prologues can have ``lpad`` with same
function hash. This further reduces number of program counters a call site can
reach.
2. ELF and psABI
-----------------
-Toolchain sets up `GNU_PROPERTY_RISCV_FEATURE_1_FCFI` for property
-`GNU_PROPERTY_RISCV_FEATURE_1_AND` in notes section of the object file.
+Toolchain sets up ``GNU_PROPERTY_RISCV_FEATURE_1_FCFI`` for property
+``GNU_PROPERTY_RISCV_FEATURE_1_AND`` in notes section of the object file.
3. Linux enabling
------------------
@@ -70,25 +70,26 @@ indirect branch tracking for the program.
4. prctl() enabling
--------------------
-`PR_SET_INDIR_BR_LP_STATUS` / `PR_GET_INDIR_BR_LP_STATUS` /
-`PR_LOCK_INDIR_BR_LP_STATUS` are three prctls added to manage indirect branch
+``PR_SET_INDIR_BR_LP_STATUS`` / ``PR_GET_INDIR_BR_LP_STATUS`` /
+``PR_LOCK_INDIR_BR_LP_STATUS`` are three prctls added to manage indirect branch
tracking. prctls are arch agnostic and returns -EINVAL on other arches.
-`PR_SET_INDIR_BR_LP_STATUS`: If arg1 `PR_INDIR_BR_LP_ENABLE` and if CPU supports
-`zicfilp` then kernel will enabled indirect branch tracking for the task.
-Dynamic loader can issue this `prctl` once it has determined that all the objects
-loaded in address space support indirect branch tracking. Additionally if there is
-a `dlopen` to an object which wasn't compiled with `zicfilp`, dynamic loader can
-issue this prctl with arg1 set to 0 (i.e. `PR_INDIR_BR_LP_ENABLE` being clear)
+``PR_SET_INDIR_BR_LP_STATUS``: If arg1 ``PR_INDIR_BR_LP_ENABLE`` and if CPU
+supports ``zicfilp`` then kernel will enabled indirect branch tracking for the
+task. Dynamic loader can issue this ``prctl`` once it has determined that all
+the objects loaded in address space support indirect branch tracking.
+Additionally if there is a ``dlopen`` to an object which wasn't compiled with
+``zicfilp``, dynamic loader can issue this prctl with arg1 set to 0 (i.e.
+``PR_INDIR_BR_LP_ENABLE`` being clear)
-`PR_GET_INDIR_BR_LP_STATUS`: Returns current status of indirect branch tracking.
-If enabled it'll return `PR_INDIR_BR_LP_ENABLE`
+``PR_GET_INDIR_BR_LP_STATUS``: Returns current status of indirect branch
+tracking. If enabled it'll return ``PR_INDIR_BR_LP_ENABLE``
-`PR_LOCK_INDIR_BR_LP_STATUS`: Locks current status of indirect branch tracking on
-the task. User space may want to run with strict security posture and wouldn't want
-loading of objects without `zicfilp` support in it and thus would want to disallow
-disabling of indirect branch tracking. In that case user space can use this prctl
-to lock current settings.
+``PR_LOCK_INDIR_BR_LP_STATUS``: Locks current status of indirect branch
+tracking on the task. User space may want to run with strict security posture
+and wouldn't want loading of objects without ``zicfilp`` support in it and thus
+would want to disallow disabling of indirect branch tracking. In that case user
+space can use this prctl to lock current settings.
5. violations related to indirect branch tracking
--------------------------------------------------
@@ -96,12 +97,12 @@ to lock current settings.
Pertaining to indirect branch tracking, CPU raises software check exception in
following conditions:
-- missing `lpad` after indirect call / jmp
-- `lpad` not on 4 byte boundary
-- `imm_20bit` embedded in `lpad` instruction doesn't match with `x7`
+- missing ``lpad`` after indirect call / jmp
+- ``lpad`` not on 4 byte boundary
+- ``imm_20bit`` embedded in ``lpad`` instruction doesn't match with x7
-In all 3 cases, `*tval = 2` is captured and software check exception is raised
+In all 3 cases, ``*tval = 2`` is captured and software check exception is raised
(cause=18)
-Linux kernel will treat this as `SIGSEV`` with code = `SEGV_CPERR` and follow
+Linux kernel will treat this as ``SIGSEV`` with code = ``SEGV_CPERR`` and follow
normal course of signal delivery.
Thanks.
--
An old man doll... just what I always wanted! - Clara
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]
next prev parent reply other threads:[~2024-09-16 2:41 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-09-12 23:16 [PATCH v4 00/30] riscv control-flow integrity for usermode Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 01/30] mm: Introduce ARCH_HAS_USER_SHADOW_STACK Deepak Gupta
2024-09-13 15:51 ` Carlos Bilbao
2024-09-12 23:16 ` [PATCH v4 02/30] mm: helper `is_shadow_stack_vma` to check shadow stack vma Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 03/30] riscv: Enable cbo.zero only when all harts support Zicboz Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 04/30] riscv: Add support for per-thread envcfg CSR values Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 05/30] riscv: Call riscv_user_isa_enable() only on the boot hart Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 06/30] riscv/Kconfig: enable HAVE_EXIT_THREAD for riscv Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 07/30] riscv: zicfilp / zicfiss in dt-bindings (extensions.yaml) Deepak Gupta
2024-09-13 0:18 ` Rob Herring (Arm)
2024-09-13 18:33 ` Conor Dooley
2024-09-12 23:16 ` [PATCH v4 08/30] riscv: zicfiss / zicfilp enumeration Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 09/30] riscv: zicfiss / zicfilp extension csr and bit definitions Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 10/30] riscv: usercfi state for task and save/restore of CSR_SSP on trap entry/exit Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 11/30] riscv/mm : ensure PROT_WRITE leads to VM_READ | VM_WRITE Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 12/30] riscv mm: manufacture shadow stack pte Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 13/30] riscv mmu: teach pte_mkwrite to manufacture shadow stack PTEs Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 14/30] riscv mmu: write protect and shadow stack Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 15/30] riscv/mm: Implement map_shadow_stack() syscall Deepak Gupta
2024-09-13 15:26 ` Mark Brown
2024-09-12 23:16 ` [PATCH v4 16/30] riscv/shstk: If needed allocate a new shadow stack on clone Deepak Gupta
2024-09-14 1:54 ` kernel test robot
2024-09-14 3:06 ` kernel test robot
2024-09-14 3:26 ` kernel test robot
2024-09-12 23:16 ` [PATCH v4 17/30] prctl: arch-agnostic prctl for shadow stack Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 18/30] prctl: arch-agnostic prctl for indirect branch tracking Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 19/30] riscv: Implements arch agnostic shadow stack prctls Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 20/30] riscv: Implements arch agnostic indirect branch tracking prctls Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 21/30] riscv/traps: Introduce software check exception Deepak Gupta
2024-09-13 19:35 ` Andy Chiu
2024-09-17 0:00 ` Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 22/30] riscv sigcontext: cfi state struct definition for sigcontext Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 23/30] riscv signal: save and restore of shadow stack for signal Deepak Gupta
2024-09-13 19:25 ` Andy Chiu
2024-09-16 22:03 ` Deepak Gupta
2024-09-17 22:03 ` Andy Chiu
2024-09-17 22:52 ` Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 24/30] riscv/kernel: update __show_regs to print shadow stack register Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 25/30] riscv/ptrace: riscv cfi status and state via ptrace and in core files Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 26/30] riscv/hwprobe: zicfilp / zicfiss enumeration in hwprobe Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 27/30] riscv: create a config for shadow stack and landing pad instr support Deepak Gupta
2024-09-12 23:16 ` [PATCH v4 28/30] riscv: Documentation for landing pad / indirect branch tracking Deepak Gupta
2024-09-16 2:41 ` Bagas Sanjaya [this message]
2024-09-12 23:16 ` [PATCH v4 29/30] riscv: Documentation for shadow stack on riscv Deepak Gupta
2024-09-16 3:20 ` Bagas Sanjaya
2024-09-12 23:16 ` [PATCH v4 30/30] kselftest/riscv: kselftest for user mode cfi Deepak Gupta
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=ZueaxRZgIf0crs4a@archie.me \
--to=bagasdotme@gmail.com \
--cc=Liam.Howlett@oracle.com \
--cc=ajones@ventanamicro.com \
--cc=akpm@linux-foundation.org \
--cc=alexghiti@rivosinc.com \
--cc=alx@kernel.org \
--cc=ancientmodern4@gmail.com \
--cc=andrii@kernel.org \
--cc=andy.chiu@sifive.com \
--cc=antonb@tenstorrent.com \
--cc=anup@brainfault.org \
--cc=aou@eecs.berkeley.edu \
--cc=arnd@arndb.de \
--cc=atishp@rivosinc.com \
--cc=ben.dooks@codethink.co.uk \
--cc=bgray@linux.ibm.com \
--cc=bhe@redhat.com \
--cc=bjorn@rivosinc.com \
--cc=bp@alien8.de \
--cc=brauner@kernel.org \
--cc=broonie@kernel.org \
--cc=catalin.marinas@arm.com \
--cc=charlie@rivosinc.com \
--cc=cleger@rivosinc.com \
--cc=conor@kernel.org \
--cc=corbet@lwn.net \
--cc=costa.shul@redhat.com \
--cc=cuiyunhui@bytedance.com \
--cc=dave.hansen@linux.intel.com \
--cc=david@redhat.com \
--cc=dawei.li@shingroup.cn \
--cc=debug@rivosinc.com \
--cc=deller@gmx.de \
--cc=devicetree@vger.kernel.org \
--cc=ebiederm@xmission.com \
--cc=evan@rivosinc.com \
--cc=greentime.hu@sifive.com \
--cc=guoren@kernel.org \
--cc=hpa@zytor.com \
--cc=jerry.shih@sifive.com \
--cc=jszhang@kernel.org \
--cc=ke.zhao@shingroup.cn \
--cc=kees@kernel.org \
--cc=krzk+dt@kernel.org \
--cc=leobras@redhat.com \
--cc=libang.li@antgroup.com \
--cc=linux-arch@vger.kernel.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-riscv@lists.infradead.org \
--cc=lorenzo.stoakes@oracle.com \
--cc=mchitale@ventanamicro.com \
--cc=mingo@redhat.com \
--cc=namcaov@gmail.com \
--cc=oleg@redhat.com \
--cc=osalvador@suse.de \
--cc=palmer@dabbelt.com \
--cc=palmer@sifive.com \
--cc=paul.walmsley@sifive.com \
--cc=peterz@infradead.org \
--cc=puranjay@kernel.org \
--cc=quic_bjorande@quicinc.com \
--cc=quic_zhonhan@quicinc.com \
--cc=revest@chromium.org \
--cc=robh@kernel.org \
--cc=rppt@kernel.org \
--cc=ryan.roberts@arm.com \
--cc=sameo@rivosinc.com \
--cc=samitolvanen@google.com \
--cc=samuel.holland@sifive.com \
--cc=schwab@suse.de \
--cc=shuah@kernel.org \
--cc=songshuaishuai@tinylab.org \
--cc=sorear@fastmail.com \
--cc=sunilvl@ventanamicro.com \
--cc=tanzhasanwork@gmail.com \
--cc=tglx@linutronix.de \
--cc=usama.anjum@collabora.com \
--cc=vbabka@suse.cz \
--cc=willy@infradead.org \
--cc=x86@kernel.org \
--cc=xiao.w.wang@intel.com \
--cc=yang.lee@linux.alibaba.com \
--cc=zev@bewilderbeest.net \
--cc=zong.li@sifive.com \
/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;
as well as URLs for NNTP newsgroup(s).