linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
To: Deepak Gupta <debug@rivosinc.com>
Cc: linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org,
	linux-kbuild@vger.kernel.org, linux-mm@kvack.org,
	llvm@lists.linux.dev, rick.p.edgecombe@intel.com,
	broonie@kernel.org, cleger@rivosinc.com, samitolvanen@google.com,
	apatel@ventanamicro.com, ajones@ventanamicro.com,
	conor.dooley@microchip.com, charlie@rivosinc.com,
	samuel.holland@sifive.com, bjorn@rivosinc.com,
	fweimer@redhat.com, jeffreyalaw@gmail.com, andrew@sifive.com,
	ved@rivosinc.com, Paul Walmsley <paul.walmsley@sifive.com>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	Albert Ou <aou@eecs.berkeley.edu>,
	Alexandre Ghiti <alex@ghiti.fr>,
	Masahiro Yamada <masahiroy@kernel.org>,
	Nathan Chancellor <nathan@kernel.org>,
	Nicolas Schier <nicolas.schier@linux.dev>,
	Andrew Morton <akpm@linux-foundation.org>,
	David Hildenbrand <david@redhat.com>,
	Lorenzo Stoakes <lorenzo.stoakes@oracle.com>,
	"Liam R. Howlett" <Liam.Howlett@oracle.com>,
	Vlastimil Babka <vbabka@suse.cz>, Mike Rapoport <rppt@kernel.org>,
	Suren Baghdasaryan <surenb@google.com>,
	Michal Hocko <mhocko@suse.com>,
	Nick Desaulniers <nick.desaulniers+lkml@gmail.com>,
	Bill Wendling <morbo@google.com>,
	Monk Chiang <monk.chiang@sifive.com>,
	Kito Cheng <kito.cheng@sifive.com>,
	Justin Stitt <justinstitt@google.com>
Subject: Re: [PATCH 11/11] riscv: Kconfig & Makefile for riscv kernel control flow integrity
Date: Fri, 25 Jul 2025 16:39:02 +0200	[thread overview]
Message-ID: <06955781-118d-4208-af28-cfbabd7d57c2@canonical.com> (raw)
In-Reply-To: <aIOTauktSYmw2LUV@debug.ba.rivosinc.com>

On 25.07.25 16:23, Deepak Gupta wrote:
> On Fri, Jul 25, 2025 at 01:26:44PM +0200, Heinrich Schuchardt wrote:
>> On 25.07.25 01:37, Deepak Gupta wrote:
>>> Defines `CONFIG_RISCV_KERNEL_CFI` and selects SHADOW_CALL_STACK
>>> and ARCH_HAS_KERNEL_SHADOW_STACK both so that zicfiss can be wired up.
>>>
>>> Makefile checks if CONFIG_RISCV_KERNEL_CFI is enabled, then light
>>> up zicfiss and zicfilp compiler flags. CONFIG_RISCV_KERNEL_CFI is
>>> dependent on CONFIG_RISCV_USER_CFI. There is no reason for user to
>>> not select support for user cfi while enabling for kernel.
>>>
>>> compat vdso don't need fcf-protection (toolchain lacks support).
>>>
>>> Signed-off-by: Deepak Gupta <debug@rivosinc.com>
>>> ---
>>>  arch/riscv/Kconfig                     | 37 ++++++++++++++++++++++++ 
>>> +++++++++-
>>>  arch/riscv/Makefile                    |  8 ++++++++
>>>  arch/riscv/kernel/compat_vdso/Makefile |  2 +-
>>>  arch/riscv/kernel/vdso/Makefile        |  2 +-
>>>  4 files changed, 46 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
>>> index 385c3d93e378..305ba5787f74 100644
>>> --- a/arch/riscv/Kconfig
>>> +++ b/arch/riscv/Kconfig
>>> @@ -245,7 +245,7 @@ config GCC_SUPPORTS_DYNAMIC_FTRACE
>>>      depends on CC_HAS_MIN_FUNCTION_ALIGNMENT || !RISCV_ISA_C
>>>  config HAVE_SHADOW_CALL_STACK
>>> -    def_bool $(cc-option,-fsanitize=shadow-call-stack)
>>> +    def_bool $(cc-option,-fsanitize=shadow-call-stack) || $(cc- 
>>> option,-mabi=lp64 -march=rv64ima_zicfilp_zicfiss)
>>>      # https://github.com/riscv-non-isa/riscv-elf-psabi-doc/commit/ 
>>> a484e843e6eeb51f0cb7b8819e50da6d2444d769
>>>      depends on $(ld-option,--no-relax-gp)
>>> @@ -864,6 +864,16 @@ config RISCV_ISA_ZICBOP
>>>        If you don't know what to do here, say Y.
>>> +config TOOLCHAIN_HAS_ZICFILP
>>> +    bool
>>> +    default y
>>> +    depends on 64BIT && $(cc-option,-mabi=lp64 -march=rv64ima_zicfilp)
>>> +
>>> +config TOOLCHAIN_HAS_ZICFISS
>>> +    bool
>>> +    default y
>>> +    depends on 64BIT && $(cc-option,-mabi=lp64 -march=rv64ima_zicfiss)
>>> +
>>>  config TOOLCHAIN_NEEDS_EXPLICIT_ZICSR_ZIFENCEI
>>>      def_bool y
>>>      # https://sourceware.org/git/?p=binutils- 
>>> gdb.git;a=commit;h=aed44286efa8ae8717a77d94b51ac3614e2ca6dc
>>> @@ -1182,6 +1192,31 @@ config RISCV_USER_CFI
>>>        space does not get protection "for free".
>>>        default n.
>>> +config RISCV_KERNEL_CFI
>>> +    def_bool n
>>> +    bool "hw assisted riscv kernel control flow integrity (kcfi)"
>>> +    depends on 64BIT && $(cc-option,-mabi=lp64 - 
>>> march=rv64ima_zicfilp_zicfiss)
>>> +    depends on RISCV_USER_CFI
>>> +    select ARCH_SUPPORTS_SHADOW_CALL_STACK
>>> +    select SHADOW_CALL_STACK
>>> +    select ARCH_HAS_KERNEL_SHADOW_STACK
>>> +    help
>>> +      Provides CPU assisted control flow integrity to for riscv kernel.
>>> +      Control flow integrity is provided by implementing shadow 
>>> stack for
>>> +      backward edge and indirect branch tracking for forward edge. 
>>> Shadow
>>> +      stack protection is a hardware feature that detects function 
>>> return
>>> +      address corruption. This helps mitigate ROP attacks. 
>>> RISCV_KERNEL_CFI
>>> +      selects CONFIG_SHADOW_CALL_STACK which uses software based shadow
>>> +      stack but is unprotected against stray writes. Selecting 
>>> RISCV_KERNEL_CFI
>>> +      will select CONFIG_DYNAMIC_SCS and will enable hardware 
>>> assisted shadow
>>> +      stack protection against stray writes.
>>
>> Please, consider adding a blank line for better readability.
> 
> Noted. Will do.
> 
>>
>>> +      Indirect branch tracking enforces that all indirect branches 
>>> must land
>>> +      on a landing pad instruction else CPU will fault. This enables 
>>> forward
>>> +      control flow (call/jmp) protection in kernel and restricts all 
>>> indirect
>>> +      call or jump in kernel to a landing pad instruction which 
>>> mostly likely
>>> +      will be start of the function.
>>> +      default n
>>
>> For Linux distributions it is important that the same kernel can run 
>> both on hardware both with and without CFI support. The description 
>> provided does not help to understand if RISCV_KERNEL_CFI=y will result 
>> in such a kernel. Please, enumerate the minimum set of extensions 
>> needed for supporting a kernel built with RISCV_KERNEL_CFI=y. I guess 
>> this will at least include Zimop.
> 
> Yes, it is expected anyone selecting this config is going to take this 
> kernel to
> a RVA23 hardware. RVA23 mandates zimop and thus shouldn't be an issue on 
> such a
> hardware. Anyone selecting this config and trying to run this kernel on 
> hardware
> prior to RVA23 will run into issues. I can add a comment here to 
> highlight that.
> 
> I assume you wanted that awareness and goal is not maintain compat of same
> kernel between RVA20 and RVA23 hardware, right?

I am aware that this option is not RVA20 compatible. Could we either 
mention RVA23 or Zimop here so users will understand the implications.

Best regards

Heinrich

> 
>>
>> Best regards
>>
>> Heinrich
>>
>>> +
>>>  endmenu # "Kernel features"
>>>  menu "Boot options"
>>> diff --git a/arch/riscv/Makefile b/arch/riscv/Makefile
>>> index 7128df832b28..6ef30a3d2bc4 100644
>>> --- a/arch/riscv/Makefile
>>> +++ b/arch/riscv/Makefile
>>> @@ -61,8 +61,10 @@ else ifeq ($(CONFIG_LTO_CLANG),y)
>>>  endif
>>>  ifeq ($(CONFIG_SHADOW_CALL_STACK),y)
>>> +ifndef CONFIG_ARCH_HAS_KERNEL_SHADOW_STACK
>>>      KBUILD_LDFLAGS += --no-relax-gp
>>>  endif
>>> +endif
>>>  # ISA string setting
>>>  riscv-march-$(CONFIG_ARCH_RV32I)    := rv32ima
>>> @@ -91,6 +93,12 @@ riscv-march-$(CONFIG_TOOLCHAIN_HAS_ZABHA) := 
>>> $(riscv-march-y)_zabha
>>>  KBUILD_BASE_ISA = -march=$(shell echo $(riscv-march-y) | sed -E 's/ 
>>> (rv32ima|rv64ima)fd([^v_]*)v?/\1\2/')
>>>  export KBUILD_BASE_ISA
>>> +ifeq ($(CONFIG_RISCV_KERNEL_CFI),y)
>>> +riscv-march-$(CONFIG_TOOLCHAIN_HAS_ZICFILP) := $(riscv-march-y)_zicfilp
>>> +riscv-march-$(CONFIG_TOOLCHAIN_HAS_ZICFISS) := $(riscv-march-y)_zicfiss
>>> +KBUILD_CFLAGS += -fcf-protection=full
>>> +KBUILD_AFLAGS += -fcf-protection=full
>>> +endif
>>>  # Remove F,D,V from isa string for all. Keep extensions between "fd" 
>>> and "v" by
>>>  # matching non-v and non-multi-letter extensions out with the filter 
>>> ([^v_]*)
>>>  KBUILD_CFLAGS += $(KBUILD_BASE_ISA)
>>> diff --git a/arch/riscv/kernel/compat_vdso/Makefile b/arch/riscv/ 
>>> kernel/compat_vdso/Makefile
>>> index 24e37d1ef7ec..552131bc34d7 100644
>>> --- a/arch/riscv/kernel/compat_vdso/Makefile
>>> +++ b/arch/riscv/kernel/compat_vdso/Makefile
>>> @@ -69,4 +69,4 @@ quiet_cmd_compat_vdsold = VDSOLD  $@
>>>  # actual build commands
>>>  quiet_cmd_compat_vdsoas = VDSOAS  $@
>>> -      cmd_compat_vdsoas = $(COMPAT_CC) $(a_flags) $(COMPAT_CC_FLAGS) 
>>> -c -o $@ $<
>>> +      cmd_compat_vdsoas = $(COMPAT_CC) $(filter-out -fcf- 
>>> protection=full, $(a_flags)) $(COMPAT_CC_FLAGS) -c -o $@ $<
>>> diff --git a/arch/riscv/kernel/vdso/Makefile b/arch/riscv/kernel/ 
>>> vdso/Makefile
>>> index 2b528d82fa7d..7b1446b63ebc 100644
>>> --- a/arch/riscv/kernel/vdso/Makefile
>>> +++ b/arch/riscv/kernel/vdso/Makefile
>>> @@ -17,7 +17,7 @@ ifdef CONFIG_VDSO_GETRANDOM
>>>  vdso-syms += getrandom
>>>  endif
>>> -ifdef CONFIG_RISCV_USER_CFI
>>> +ifneq ($(CONFIG_RISCV_USER_CFI), $(CONFIG_RISCV_KERNEL_CFI))
>>>  CFI_MARCH = _zicfilp_zicfiss
>>>  CFI_FULL = -fcf-protection=full
>>>  endif
>>>
>>



  reply	other threads:[~2025-07-25 14:39 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-07-24 23:36 [PATCH 00/11] riscv: fine grained hardware assisted kernel control-flow integrity Deepak Gupta
2025-07-24 23:36 ` [PATCH 01/11] riscv: add landing pad for asm routines Deepak Gupta
2025-07-25  6:13   ` Heinrich Schuchardt
2025-07-25 14:10     ` Deepak Gupta
2025-07-25 15:27   ` Sami Tolvanen
2025-07-25 17:01     ` Deepak Gupta
2025-07-24 23:36 ` [PATCH 02/11] riscv: update asm call site in `call_on_irq_stack` to setup correct label Deepak Gupta
2025-07-25  6:23   ` Heinrich Schuchardt
2025-07-25 14:16     ` Deepak Gupta
2025-07-25 15:33   ` Sami Tolvanen
2025-07-25 16:56     ` Deepak Gupta
2025-07-24 23:36 ` [PATCH 03/11] riscv: indirect jmp in asm that's static in nature to use sw guarded jump Deepak Gupta
2025-07-25  6:26   ` Heinrich Schuchardt
2025-07-24 23:36 ` [PATCH 04/11] riscv: exception handlers can be software guarded transfers Deepak Gupta
2025-07-24 23:36 ` [PATCH 05/11] riscv: enable landing pad enforcement Deepak Gupta
2025-07-25  6:33   ` Heinrich Schuchardt
2025-07-25 14:20     ` Deepak Gupta
2025-07-25 14:43       ` Heinrich Schuchardt
2025-07-24 23:36 ` [PATCH 06/11] mm: Introduce ARCH_HAS_KERNEL_SHADOW_STACK Deepak Gupta
2025-07-26  7:42   ` Mike Rapoport
2025-07-29  0:36     ` Deepak Gupta
2025-07-24 23:37 ` [PATCH 07/11] scs: place init shadow stack in .shadowstack section Deepak Gupta
2025-07-24 23:37 ` [PATCH 08/11] riscv/mm: prepare shadow stack for init task Deepak Gupta
2025-07-24 23:37 ` [PATCH 09/11] riscv: scs: add hardware shadow stack support to scs Deepak Gupta
2025-07-24 23:37 ` [PATCH 10/11] scs: generic scs code updated to leverage hw assisted shadow stack Deepak Gupta
2025-07-25 16:13   ` Sami Tolvanen
2025-07-25 16:42     ` Deepak Gupta
2025-07-25 16:47       ` Deepak Gupta
2025-07-25 16:46     ` Mark Brown
2025-07-28 12:47     ` Will Deacon
2025-07-28 16:37       ` Deepak Gupta
2025-07-25 17:06   ` Edgecombe, Rick P
2025-07-25 17:19     ` Deepak Gupta
2025-07-25 18:05       ` Edgecombe, Rick P
2025-07-28 19:23         ` Deepak Gupta
2025-07-28 21:19           ` Deepak Gupta
2025-07-24 23:37 ` [PATCH 11/11] riscv: Kconfig & Makefile for riscv kernel control flow integrity Deepak Gupta
2025-07-25 11:26   ` Heinrich Schuchardt
2025-07-25 14:23     ` Deepak Gupta
2025-07-25 14:39       ` Heinrich Schuchardt [this message]
2025-07-24 23:38 ` [PATCH 00/11] riscv: fine grained hardware assisted kernel control-flow integrity 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=06955781-118d-4208-af28-cfbabd7d57c2@canonical.com \
    --to=heinrich.schuchardt@canonical.com \
    --cc=Liam.Howlett@oracle.com \
    --cc=ajones@ventanamicro.com \
    --cc=akpm@linux-foundation.org \
    --cc=alex@ghiti.fr \
    --cc=andrew@sifive.com \
    --cc=aou@eecs.berkeley.edu \
    --cc=apatel@ventanamicro.com \
    --cc=bjorn@rivosinc.com \
    --cc=broonie@kernel.org \
    --cc=charlie@rivosinc.com \
    --cc=cleger@rivosinc.com \
    --cc=conor.dooley@microchip.com \
    --cc=david@redhat.com \
    --cc=debug@rivosinc.com \
    --cc=fweimer@redhat.com \
    --cc=jeffreyalaw@gmail.com \
    --cc=justinstitt@google.com \
    --cc=kito.cheng@sifive.com \
    --cc=linux-kbuild@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=llvm@lists.linux.dev \
    --cc=lorenzo.stoakes@oracle.com \
    --cc=masahiroy@kernel.org \
    --cc=mhocko@suse.com \
    --cc=monk.chiang@sifive.com \
    --cc=morbo@google.com \
    --cc=nathan@kernel.org \
    --cc=nick.desaulniers+lkml@gmail.com \
    --cc=nicolas.schier@linux.dev \
    --cc=palmer@dabbelt.com \
    --cc=paul.walmsley@sifive.com \
    --cc=rick.p.edgecombe@intel.com \
    --cc=rppt@kernel.org \
    --cc=samitolvanen@google.com \
    --cc=samuel.holland@sifive.com \
    --cc=surenb@google.com \
    --cc=vbabka@suse.cz \
    --cc=ved@rivosinc.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