All of lore.kernel.org
 help / color / mirror / Atom feed
From: Christophe Leroy <christophe.leroy@csgroup.eu>
To: "Christopher M. Riedl" <cmr@codefail.de>, linuxppc-dev@lists.ozlabs.org
Subject: Re: [PATCH v3 1/8] powerpc/uaccess: Add unsafe_copy_from_user
Date: Tue, 19 Jan 2021 08:29:11 +0100	[thread overview]
Message-ID: <5863d631-08a2-29ed-09b5-78d9009ca1df@csgroup.eu> (raw)
In-Reply-To: <C8LLSM3UC598.L4E2VMGJOI06@geist>



Le 17/01/2021 à 18:19, Christopher M. Riedl a écrit :
> On Mon Jan 11, 2021 at 7:22 AM CST, Christophe Leroy wrote:
>>
>>
>> Le 09/01/2021 à 04:25, Christopher M. Riedl a écrit :
>>> Implement raw_copy_from_user_allowed() which assumes that userspace read
>>> access is open. Use this new function to implement raw_copy_from_user().
>>> Finally, wrap the new function to follow the usual "unsafe_" convention
>>> of taking a label argument.
>>
>> I think there is no point implementing raw_copy_from_user_allowed(), see
>> https://github.com/linuxppc/linux/commit/4b842e4e25b1 and
>> https://patchwork.ozlabs.org/project/linuxppc-dev/patch/8c74fc9ce8131cabb10b3e95dc0e430f396ee83e.1610369143.git.christophe.leroy@csgroup.eu/
>>
>> You should simply do:
>>
>> #define unsafe_copy_from_user(d, s, l, e) \
>> unsafe_op_wrap(__copy_tofrom_user((__force void __user *)d, s, l), e)
>>
> 
> I gave this a try and the signal ops decreased by ~8K. Now, to be
> honest, I am not sure what an "acceptable" benchmark number here
> actually is - so maybe this is ok? Same loss with both radix and hash:

I don't think this is ok, but it probably means that you are using unsafe_copy_from_user() to copy 
small constant size data that should be copied with unsafe_get_user() instead.

> 
> 	|                                      | hash   | radix  |
> 	| ------------------------------------ | ------ | ------ |
> 	| linuxppc/next                        | 118693 | 133296 |
> 	| linuxppc/next w/o KUAP+KUEP          | 228911 | 228654 |
> 	| unsafe-signal64                      | 200480 | 234067 |
> 	| unsafe-signal64 (__copy_tofrom_user) | 192467 | 225119 |
> 
> To put this into perspective, prior to KUAP and uaccess flush, signal
> performance in this benchmark was ~290K on hash.
> 
>>
>> Christophe
>>
>>>
>>> The new raw_copy_from_user_allowed() calls non-inline __copy_tofrom_user()
>>> internally. This is still safe to call inside user access blocks formed
>>> with user_*_access_begin()/user_*_access_end() since asm functions are not
>>> instrumented for tracing.
>>>
>>> Signed-off-by: Christopher M. Riedl <cmr@codefail.de>
>>> ---
>>>    arch/powerpc/include/asm/uaccess.h | 28 +++++++++++++++++++---------
>>>    1 file changed, 19 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/arch/powerpc/include/asm/uaccess.h b/arch/powerpc/include/asm/uaccess.h
>>> index 501c9a79038c..698f3a6d6ae5 100644
>>> --- a/arch/powerpc/include/asm/uaccess.h
>>> +++ b/arch/powerpc/include/asm/uaccess.h
>>> @@ -403,38 +403,45 @@ raw_copy_in_user(void __user *to, const void __user *from, unsigned long n)
>>>    }
>>>    #endif /* __powerpc64__ */
>>>    
>>> -static inline unsigned long raw_copy_from_user(void *to,
>>> -		const void __user *from, unsigned long n)
>>> +static inline unsigned long
>>> +raw_copy_from_user_allowed(void *to, const void __user *from, unsigned long n)
>>>    {
>>> -	unsigned long ret;
>>>    	if (__builtin_constant_p(n) && (n <= 8)) {
>>> -		ret = 1;
>>> +		unsigned long ret = 1;
>>>    
>>>    		switch (n) {
>>>    		case 1:
>>>    			barrier_nospec();
>>> -			__get_user_size(*(u8 *)to, from, 1, ret);
>>> +			__get_user_size_allowed(*(u8 *)to, from, 1, ret);
>>>    			break;
>>>    		case 2:
>>>    			barrier_nospec();
>>> -			__get_user_size(*(u16 *)to, from, 2, ret);
>>> +			__get_user_size_allowed(*(u16 *)to, from, 2, ret);
>>>    			break;
>>>    		case 4:
>>>    			barrier_nospec();
>>> -			__get_user_size(*(u32 *)to, from, 4, ret);
>>> +			__get_user_size_allowed(*(u32 *)to, from, 4, ret);
>>>    			break;
>>>    		case 8:
>>>    			barrier_nospec();
>>> -			__get_user_size(*(u64 *)to, from, 8, ret);
>>> +			__get_user_size_allowed(*(u64 *)to, from, 8, ret);
>>>    			break;
>>>    		}
>>>    		if (ret == 0)
>>>    			return 0;
>>>    	}
>>>    
>>> +	return __copy_tofrom_user((__force void __user *)to, from, n);
>>> +}
>>> +
>>> +static inline unsigned long
>>> +raw_copy_from_user(void *to, const void __user *from, unsigned long n)
>>> +{
>>> +	unsigned long ret;
>>> +
>>>    	barrier_nospec();
>>>    	allow_read_from_user(from, n);
>>> -	ret = __copy_tofrom_user((__force void __user *)to, from, n);
>>> +	ret = raw_copy_from_user_allowed(to, from, n);
>>>    	prevent_read_from_user(from, n);
>>>    	return ret;
>>>    }
>>> @@ -542,6 +549,9 @@ user_write_access_begin(const void __user *ptr, size_t len)
>>>    #define unsafe_get_user(x, p, e) unsafe_op_wrap(__get_user_allowed(x, p), e)
>>>    #define unsafe_put_user(x, p, e) __put_user_goto(x, p, e)
>>>    
>>> +#define unsafe_copy_from_user(d, s, l, e) \
>>> +	unsafe_op_wrap(raw_copy_from_user_allowed(d, s, l), e)
>>> +
>>>    #define unsafe_copy_to_user(d, s, l, e) \
>>>    do {									\
>>>    	u8 __user *_dst = (u8 __user *)(d);				\
>>>

  parent reply	other threads:[~2021-01-19  8:55 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-01-09  3:25 [PATCH v3 0/8] Improve signal performance on PPC64 with KUAP Christopher M. Riedl
2021-01-09  3:25 ` [PATCH v3 1/8] powerpc/uaccess: Add unsafe_copy_from_user Christopher M. Riedl
2021-01-11 13:22   ` Christophe Leroy
2021-01-17 17:19     ` Christopher M. Riedl
2021-01-19  2:11       ` Michael Ellerman
2021-01-19 12:33         ` Christophe Leroy
2021-01-19 17:02           ` Christopher M. Riedl
2021-01-19 17:27             ` Christophe Leroy
2021-01-20  5:08               ` Christopher M. Riedl
2021-02-09 14:09         ` Christophe Leroy
2021-01-19  7:29       ` Christophe Leroy [this message]
2021-01-09  3:25 ` [PATCH v3 2/8] powerpc/signal: Add unsafe_copy_{vsx,fpr}_from_user() Christopher M. Riedl
2021-01-09  3:25 ` [PATCH v3 3/8] powerpc/signal64: Move non-inline functions out of setup_sigcontext() Christopher M. Riedl
2021-01-09  3:25 ` [PATCH v3 4/8] powerpc/signal64: Remove TM ifdefery in middle of if/else block Christopher M. Riedl
2021-01-11 13:29   ` Christophe Leroy
2021-01-17 17:16     ` Christopher M. Riedl
2021-01-09  3:25 ` [PATCH v3 5/8] powerpc/signal64: Replace setup_sigcontext() w/ unsafe_setup_sigcontext() Christopher M. Riedl
2021-01-09  3:25 ` [PATCH v3 6/8] powerpc/signal64: Replace restore_sigcontext() w/ unsafe_restore_sigcontext() Christopher M. Riedl
2021-01-09  3:25 ` [PATCH v3 7/8] powerpc/signal64: Rewrite handle_rt_signal64() to minimise uaccess switches Christopher M. Riedl
2021-01-09  3:25 ` [PATCH v3 8/8] powerpc/signal64: Rewrite rt_sigreturn() " Christopher M. Riedl

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=5863d631-08a2-29ed-09b5-78d9009ca1df@csgroup.eu \
    --to=christophe.leroy@csgroup.eu \
    --cc=cmr@codefail.de \
    --cc=linuxppc-dev@lists.ozlabs.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.