linux-kernel - Re: [RFC -tip 0/4] x86: reduce fixup of uaccess

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <49655A3C.9020908@ct.jp.nec.com>
Date:	Wed, 07 Jan 2009 17:43:24 -0800
From:	Hiroshi Shimamoto <h-shimamoto@...jp.nec.com>
To:	"H. Peter Anvin" <hpa@...or.com>
CC:	Ingo Molnar <mingo@...e.hu>, Thomas Gleixner <tglx@...utronix.de>,
	linux-kernel@...r.kernel.org
Subject: Re: [RFC -tip 0/4] x86: reduce fixup of uaccess

H. Peter Anvin wrote:
> Hiroshi Shimamoto wrote:
>> This is my second try to reduce fixup code size for exceptions of uaccess.
>>
>> This patch series reduces fixup code for exceptions of uaccess in signal.
>>
>> I gave up to make direct jump to end of function when an exception occurs.
>> However, I thought fixup code could be reduced. The concept is that to add
>> uaccess_err in thread_info and set it to -EFAULT on exception, finally check
>> this value on the last of function.
>>
>> Is this good to reduce code size?
>>
> 
> Hello Hiroshi,

Hello Peter,

> 
> The patches look technically really nice.  I have a couple of stylistic
> comments, though, which I'd like yours and others' comments on.

Thanks for comments.

> 
> This introduces a new blocking construct, and it's not immediately
> obvious in the source code.  I think introducing a technically redundant

Yeah, I think it's not friendly about readability now.

> set of braces and dropping the parens from the try construct and the
> redundant pointer might look better:
> 
> 	get_user_try {
> 		/* do stuff */
> 	} get_user_catch(err);

OK, I'll update like that.

> 
> This makes it, in my opinion, much clearer that it is a new bracing
> construct, and it also eliminates the need to form a pointer to "err"

I'm not sure which is better plain "err" or pointer to "err".

> (even though the compiler doesn't actually do so, it looks like it does
> to the programmer.)
> 
> Also, I don't think we need double underscores for the wrapping
> construct, since the get_user/__get_user (check/nocheck) etc.
> distinction doesn't directly apply there.

OK. Will drop underscores.

I'll repost update patches.

Thanks,
Hiroshi

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/