linux-kernel - Re: [PATCH] bitfield.h: add FIELD_MAX() and field

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]

Date:   Sat, 29 Feb 2020 07:10:45 -0600
From:   Alex Elder <elder@...aro.org>
To:     Jakub Kicinski <kuba@...nel.org>
Cc:     Al Viro <viro@...iv.linux.org.uk>,
        Johannes Berg <johannes@...solutions.net>,
        Arnd Bergmann <arnd@...db.de>,
        Masahiro Yamada <masahiroy@...nel.org>,
        Bjorn Andersson <bjorn.andersson@...aro.org>,
        lkml <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] bitfield.h: add FIELD_MAX() and field_max()

On 2/28/20 12:04 PM, Alex Elder wrote:
> On 2/28/20 11:56 AM, Jakub Kicinski wrote:
>> On Fri, 28 Feb 2020 10:53:43 -0600 Alex Elder wrote:
>>> Define FIELD_MAX(), which supplies the maximum value that can be
>>> represented by a field value.  Define field_max() as well, to go
>>> along with the lower-case forms of the field mask functions.
>>>
>>> Signed-off-by: Alex Elder <elder@...aro.org>
>>> ---
>>>
>>>  NOTE:	I'm not entirely sure who owns/maintains this file so
>>> 	I'm sending it to those who have committed things to it.
>>> 	I hope someone will just take it in after review; I use
>>> 	field_max() in some code I'm about to send out.
>>
>> Could you give us an example use?
>>
>> Is it that you find the current macros misnamed or there's something
>> you can't do? Or are you trying to set fields to max?
> 
> I'm trying to validate variable values are in range before attempting
> to use them in a bitfield.

I should have actually checked my code before I sent this.  Yes
I am using the macro as I described, to see if something fits.
But I'm also using it this way:

	foo = u32_get_bits(register, FOO_COUNT_FMASK);
	if (foo == field_max(FOO_COUNT_MASK))
		;	/* This has special meaning */

And another way:

	size_limit = field_max(FOO_COUNT_MASK) * sizeof(struct foo);

So field_max() is really what I need here.  It does imply a
signed interpretation of the field value, but that's true
for all of the lower-case bitfield functions.

I understand the value of FIELD_FIT() but I think field_max()
(and FIELD_MAX()) serves a purpose.  In fact one could argue it
makes FIELD_FIT() unnecessary (compare against field_max(mask)
instead) but I won't propose removing it.

So after further consideration I believe the original patch
is fine.  What are your thoughts?

					-Alex

> I find field_max() to be a good name for what I'm looking for.
> 
> 					-Alex
> 
> #define FOO_FMASK	0x000ff000
> 
> static u32 register = 0x12345678;
> 
> int foo(u32 value)
> {
> 	if (value > field_max(FOO_FMASK))
> 		return -EINVAL;
> 
> 	u32_replace_bits(&register, value, FOO_FMASK);
> 	
> 	return 0;
> }
>