mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "Clément Léger" <cleger@rivosinc.com>
To: David Laight <david.laight.linux@gmail.com>
Cc: linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org,
	Paul Walmsley <paul.walmsley@sifive.com>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	Albert Ou <aou@eecs.berkeley.edu>,
	Alexandre Ghiti <alex@ghiti.fr>
Subject: Re: [PATCH 2/2] riscv: uaccess: do not do misaligned accesses in get/put_user()
Date: Mon, 2 Jun 2025 09:34:17 +0200	[thread overview]
Message-ID: <836dffaf-c1c4-4e1a-957a-355deaa2d963@rivosinc.com> (raw)
In-Reply-To: <20250531192801.5d6329c3@pumpkin>



On 31/05/2025 20:28, David Laight wrote:
> On Fri, 30 May 2025 22:56:58 +0200
> Clément Léger <cleger@rivosinc.com> wrote:
> 
>> Doing misaligned access to userspace memory would make a trap on
>> platform where it is emulated. Latest fixes removed the kernel
>> capability to do unaligned accesses to userspace memory safely since
>> interrupts are kept disabled at all time during that. Thus doing so
>> would crash the kernel.
>>
>> Such behavior was detected with GET_UNALIGN_CTL() that was doing
>> a put_user() with an unsigned long* address that should have been an
>> unsigned int*. Reenabling kernel misaligned access emulation is a bit
>> risky and it would also degrade performances. Rather than doing that,
>> we will try to avoid any misaligned accessed by using copy_from/to_user()
>> which does not do any misaligned accesses. This can be done only for
>> !CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS and thus allows to only generate
>> a bit more code for this config.
> 
> For get_user() you are much better off reading the two words that contain
> the value and then doing 'shift' and 'or' to get the correct value.
> 
> Even for put_user() doing the explicit byte accesses will be faster than
> going though the generic copy_to/from_user() function.

Hi David,

Alexandre tried that approach as well but that added a bit more code and
it was more complex than just calling copy_from/to_user(). That can
still be done in another commit if we need more performance later. As a
side note, prior to that patch, these misaligned accesses were using
trap-and-emulate so that is still a performance improvement.

Thanks,

Clément

> 
> 	David
> 
>>
>> Signed-off-by: Clément Léger <cleger@rivosinc.com>
>> ---
>>  arch/riscv/include/asm/uaccess.h | 28 ++++++++++++++++++++++------
>>  1 file changed, 22 insertions(+), 6 deletions(-)
>>
>> diff --git a/arch/riscv/include/asm/uaccess.h b/arch/riscv/include/asm/uaccess.h
>> index 046de7ced09c..b542c05f394f 100644
>> --- a/arch/riscv/include/asm/uaccess.h
>> +++ b/arch/riscv/include/asm/uaccess.h
>> @@ -169,8 +169,21 @@ do {								\
>>  
>>  #endif /* CONFIG_64BIT */
>>  
>> +unsigned long __must_check __asm_copy_to_user(void __user *to,
>> +	const void *from, unsigned long n);
>> +unsigned long __must_check __asm_copy_from_user(void *to,
>> +	const void __user *from, unsigned long n);
>> +
>>  #define __get_user_nocheck(x, __gu_ptr, label)			\
>>  do {								\
>> +	if (!IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS)) {			\
>> +		if (!IS_ALIGNED((uintptr_t)__gu_ptr, sizeof(*__gu_ptr))) {		\
>> +			if (__asm_copy_from_user(&(x), __gu_ptr, sizeof(*__gu_ptr)))	\
>> +				goto label;			\
>> +			else					\
>> +				break;				\
>> +		}						\
>> +	}							\
>>  	switch (sizeof(*__gu_ptr)) {				\
>>  	case 1:							\
>>  		__get_user_asm("lb", (x), __gu_ptr, label);	\
>> @@ -297,6 +310,15 @@ do {								\
>>  
>>  #define __put_user_nocheck(x, __gu_ptr, label)			\
>>  do {								\
>> +	if (!IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS)) {			\
>> +		if (!IS_ALIGNED((uintptr_t)__gu_ptr, sizeof(*__gu_ptr))) {		\
>> +			unsigned long val = (unsigned long)(x);				\
>> +			if (__asm_copy_to_user(__gu_ptr, &(val), sizeof(*__gu_ptr)))	\
>> +				goto label;			\
>> +			else					\
>> +				break;				\
>> +		}						\
>> +	}							\
>>  	switch (sizeof(*__gu_ptr)) {				\
>>  	case 1:							\
>>  		__put_user_asm("sb", (x), __gu_ptr, label);	\
>> @@ -385,12 +407,6 @@ err_label:							\
>>  		-EFAULT;					\
>>  })
>>  
>> -
>> -unsigned long __must_check __asm_copy_to_user(void __user *to,
>> -	const void *from, unsigned long n);
>> -unsigned long __must_check __asm_copy_from_user(void *to,
>> -	const void __user *from, unsigned long n);
>> -
>>  static inline unsigned long
>>  raw_copy_from_user(void *to, const void __user *from, unsigned long n)
>>  {
> 


  parent reply	other threads:[~2025-06-02  7:34 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-05-30 20:56 [PATCH 0/2] riscv: misaligned: fix misaligned accesses handling in put/get_user() Clément Léger
2025-05-30 20:56 ` [PATCH 1/2] riscv: process: use unsigned int instead of unsigned long for put_user() Clément Léger
2025-05-31 12:29   ` Alexandre Ghiti
2025-05-30 20:56 ` [PATCH 2/2] riscv: uaccess: do not do misaligned accesses in get/put_user() Clément Léger
2025-05-31 12:35   ` Alexandre Ghiti
2025-06-02  7:37     ` Clément Léger
2025-06-02 15:22       ` Alexandre Ghiti
2025-05-31 18:28   ` David Laight
2025-06-01 17:35     ` Maciej W. Rozycki
2025-06-02  7:35       ` Clément Léger
2025-06-02  7:34     ` Clément Léger [this message]
2025-05-31 13:32 ` [PATCH 0/2] riscv: misaligned: fix misaligned accesses handling in put/get_user() Alexandre Ghiti
2025-06-02  7:19   ` Clément Léger

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=836dffaf-c1c4-4e1a-957a-355deaa2d963@rivosinc.com \
    --to=cleger@rivosinc.com \
    --cc=alex@ghiti.fr \
    --cc=aou@eecs.berkeley.edu \
    --cc=david.laight.linux@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=palmer@dabbelt.com \
    --cc=paul.walmsley@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

all inboxes | Powered by JetHome®