Re: [PATCH v12 22/28] riscv: enable kernel access to shadow stack memory via FWFT sbi call

From: "Radim Krčmář" <rkrcmar@ventanamicro.com>
To: "Deepak Gupta" <debug@rivosinc.com>
Cc: "Thomas Gleixner" <tglx@linutronix.de>,
	"Ingo Molnar" <mingo@redhat.com>,
	"Borislav Petkov" <bp@alien8.de>,
	"Dave Hansen" <dave.hansen@linux.intel.com>, <x86@kernel.org>,
	"H. Peter Anvin" <hpa@zytor.com>,
	"Andrew Morton" <akpm@linux-foundation.org>,
	"Liam R. Howlett" <Liam.Howlett@oracle.com>,
	"Vlastimil Babka" <vbabka@suse.cz>,
	"Lorenzo Stoakes" <lorenzo.stoakes@oracle.com>,
	"Paul Walmsley" <paul.walmsley@sifive.com>,
	"Palmer Dabbelt" <palmer@dabbelt.com>,
	"Albert Ou" <aou@eecs.berkeley.edu>,
	"Conor Dooley" <conor@kernel.org>,
	"Rob Herring" <robh@kernel.org>,
	"Krzysztof Kozlowski" <krzk+dt@kernel.org>,
	"Arnd Bergmann" <arnd@arndb.de>,
	"Christian Brauner" <brauner@kernel.org>,
	"Peter Zijlstra" <peterz@infradead.org>,
	"Oleg Nesterov" <oleg@redhat.com>,
	"Eric Biederman" <ebiederm@xmission.com>,
	"Kees Cook" <kees@kernel.org>, "Jonathan Corbet" <corbet@lwn.net>,
	"Shuah Khan" <shuah@kernel.org>, "Jann Horn" <jannh@google.com>,
	"Conor Dooley" <conor+dt@kernel.org>,
	<linux-kernel@vger.kernel.org>, <linux-fsdevel@vger.kernel.org>,
	<linux-mm@kvack.org>, <linux-riscv@lists.infradead.org>,
	<devicetree@vger.kernel.org>, <linux-arch@vger.kernel.org>,
	<linux-doc@vger.kernel.org>, <linux-kselftest@vger.kernel.org>,
	<alistair.francis@wdc.com>, <richard.henderson@linaro.org>,
	<jim.shu@sifive.com>, <andybnac@gmail.com>,
	<kito.cheng@sifive.com>, <charlie@rivosinc.com>,
	<atishp@rivosinc.com>, <evan@rivosinc.com>, <cleger@rivosinc.com>,
	<alexghiti@rivosinc.com>, <samitolvanen@google.com>,
	<broonie@kernel.org>, <rick.p.edgecombe@intel.com>,
	"Zong Li" <zong.li@sifive.com>,
	"linux-riscv" <linux-riscv-bounces@lists.infradead.org>
Subject: Re: [PATCH v12 22/28] riscv: enable kernel access to shadow stack memory via FWFT sbi call
Date: Fri, 21 Mar 2025 08:35:49 +0100	[thread overview]
Message-ID: <D8LRS46QQ2FE.ZF1856CR80JM@ventanamicro.com> (raw)
In-Reply-To: <CAKC1njTyiaBkmHvAM8VT_MG4Cdch=H9P8r3C-m-=QQEuzyrRNA@mail.gmail.com>

2025-03-20T15:42:44-07:00, Deepak Gupta <debug@rivosinc.com>:
> On Thu, Mar 20, 2025 at 3:10 PM Radim Krčmář <rkrcmar@ventanamicro.com> wrote:
>>
>> 2025-03-14T14:39:41-07:00, Deepak Gupta <debug@rivosinc.com>:
>> > Kernel will have to perform shadow stack operations on user shadow stack.
>> > Like during signal delivery and sigreturn, shadow stack token must be
>> > created and validated respectively. Thus shadow stack access for kernel
>> > must be enabled.
>>
>> Why can't kernel access the user shadow stack through an aliased WR
>> mapping?
>
> It can, although that opens up a can of worms. If this alternating
> mapping is user mode
> then ensuring that another threat in userspace can't write to this
> address in this window
> of signal handling.

Right, it must not be user mode.

>                     A kernel alternate mapping can be created, but
> that can lead to all
> sorts of requirements of ensuring the page is pinned down. IIRC, It
> has been debated
> on x86 shadow stack merge time as well on how a flaky alias mapping approach can
> become and weaken the threat model it is supposed to protect against.

True.

> Simply using `ssamoswap` is simple and serves the purpose. Enabling shadow stack
> access for the kernel doesn't have any adverse effect on the kernel.

Makes sense.  We just depend on an extra feature, because we should
consider the case when M-mode doesn't allow S-mode shadow stack, but
S-mode can enable U-mode shadow stack:

>> > ---
>> > diff --git a/arch/riscv/kernel/head.S b/arch/riscv/kernel/head.S
>> > @@ -320,6 +326,12 @@ SYM_CODE_START(_start_kernel)
>> >       la tp, init_task
>> >       la sp, init_thread_union + THREAD_SIZE
>> >       addi sp, sp, -PT_SIZE_ON_STACK
>> > +     li a7, SBI_EXT_FWFT
>> > +     li a6, SBI_EXT_FWFT_SET
>> > +     li a0, SBI_FWFT_SHADOW_STACK
>> > +     li a1, 1 /* enable supervisor to access shadow stack access */
>> > +     li a2, SBI_FWFT_SET_FLAG_LOCK
>> > +     ecall
>>
>> I think the ecall can fail even on machines that have Zicfiss, so it
>> would be good to disable user shadow stack if that happens.