Re: [PATCH] riscv: vector: Fix data pointer constraints in context save/restore

From: "Troy Mitchell" <troy.mitchell@linux.dev>
To: "Guo Ren" <guoren@kernel.org>, "Troy Mitchell" <troy.mitchell@linux.dev>
Cc: "Paul Walmsley" <pjw@kernel.org>,
	"Palmer Dabbelt" <palmer@dabbelt.com>,
	"Albert Ou" <aou@eecs.berkeley.edu>,
	"Andy Chiu" <andybnac@gmail.com>,
	"Vincent Chen" <vincent.chen@sifive.com>,
	"Björn Töpel" <bjorn@rivosinc.com>,
	"Alexandre Ghiti" <alex@ghiti.fr>,
	linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org,
	"Palmer Dabbelt" <palmer@rivosinc.com>,
	"Greentime Hu" <greentime.hu@sifive.com>,
	"Heiko Stuebner" <heiko@sntech.de>,
	"Conor Dooley" <conor.dooley@microchip.com>,
	"Kevin Zhang" <zhangmeng.kevin@linux.spacemit.com>
Subject: Re: [PATCH] riscv: vector: Fix data pointer constraints in context save/restore
Date: Sat, 12 Sep 2026 17:21:46 +0800	[thread overview]
Message-ID: <DLD87F4KVA70.JK2UTB6A9HZF@linux.dev> (raw)
In-Reply-To: <CAJF2gTRn4keG=peEKkncgdsu=RLP3N0Ev6cO7mzXg-J48R+bEg@mail.gmail.com>


[-- Attachment #1.1: Type: text/plain, Size: 1393 bytes --]

On Fri Sep 11, 2026 at 2:56 PM +08, Guo Ren wrote:
> On Tue, Sep 8, 2026 at 9:03 PM Troy Mitchell <troy.mitchell@linux.dev> wrote:
>>
>> The standard vector save/restore asm advances datap but declares it as
>> input-only. An inlined caller reusing the original pointer may therefore
>> use the advanced address instead.
>>
>> Declare datap as read-write so the compiler can preserve the original
>> pointer when needed.
>>
>> Fixes: 03c3fcd9941a ("riscv: Introduce struct/helpers to save/restore per-task Vector state")
>> Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>
>> ---
>>  arch/riscv/include/asm/vector.h | 4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/riscv/include/asm/vector.h b/arch/riscv/include/asm/vector.h
>> index fffe72a772080..c7fd6d50a7a47 100644
>> --- a/arch/riscv/include/asm/vector.h
>> +++ b/arch/riscv/include/asm/vector.h
>> @@ -230,7 +230,7 @@ static inline void __riscv_v_vstate_save(struct __riscv_v_ext_state *save_to,
>>                         "add            %1, %1, %0\n\t"
> Yes, %1 is modified. Good catch!
>
> Reviewed-by: GUO Ren (XuanTie) <guoren@kernel.org>
>
> PS:
> Don't forget "if (has_xtheadvector()) {" part, thx :)
The XTheadVector path copies datap into t0 and lists t0/t4 as
clobbers, so it doesn't have this issue. Thanks for the reminder!

-- 
Troy Mitchell


[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 273 bytes --]

[-- Attachment #2: Type: text/plain, Size: 161 bytes --]

_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv