linux-kernel - Re: [PATCH] pinctrl: tegra: use signed bitfields for optional fields

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: <fd76a3659b5096dfa6a5daeecf36ea7c@agner.ch>
Date:	Mon, 16 Mar 2015 21:17:09 +0100
From:	Stefan Agner <stefan@...er.ch>
To:	Stephen Warren <swarren@...dotorg.org>
Cc:	linus.walleij@...aro.org, thierry.reding@...il.com,
	gnurou@...il.com, linux-gpio@...r.kernel.org,
	linux-tegra@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] pinctrl: tegra: use signed bitfields for optional fields

On 2015-03-16 17:59, Stephen Warren wrote:
> On 03/14/2015 06:05 PM, Stefan Agner wrote:
>> Optional fields are set to -1 by various preprocessor macros. Make
>> sure the struct fields can actually store them.
> 
>> diff --git a/drivers/pinctrl/pinctrl-tegra.h b/drivers/pinctrl/pinctrl-tegra.h
> 
>> -	u32 mux_bit:6;
>> -	u32 pupd_bit:6;
>> -	u32 tri_bit:6;
> ...
>> +	s8 mux_bit:6;
>> +	s8 pupd_bit:6;
>> +	s8 tri_bit:6;
> 
> Could we make these s32s instead? According to the C standard, the
> type should be a signed or unsigned int, and s32 matches that better
> than s8 for existing Tegra 32-bit platforms. Equally, for bitfields
> that don't fit into the remaining space within a container (s8 above),
> implementations are allowed to either span bitfields across multiple
> containers, or pad the current container and start the bitfield in the
> next container. Using the larger s32 as the "container" yields less
> opportunity for potential padding and thus wasting space.

The reason I took the signed byte variant was that it gets filled into
that type later on. So I thought it would be kind of nice to see that
fact already in the definition of the struct...

But your argument has actual real significance. Will change that.

> Do you observe any increase in the sizes reported by
> "${CROSS_COMPILE}size pinctrl-tegra*.o" with this patch?

Baseline:
   5406	    180	      1	   5587	  
15d3	drivers/pinctrl/pinctrl-tegra-xusb.o
   5244	     64	      0	   5308	   14bc	drivers/pinctrl/pinctrl-tegra.o
  18072	   1032	      0	  19104	  
4aa0	drivers/pinctrl/pinctrl-tegra114.o
  19214	   1128	      0	  20342	  
4f76	drivers/pinctrl/pinctrl-tegra124.o
  18352	    876	      0	  19228	  
4b1c	drivers/pinctrl/pinctrl-tegra20.o
  24621	   1068	      0	  25689	  
6459	drivers/pinctrl/pinctrl-tegra30.o


With the patch above...
   text	   data	    bss	    dec	    hex	filename
   5406	    180	      1	   5587	  
15d3	drivers/pinctrl/pinctrl-tegra-xusb.o
   5252	     64	      0	   5316	   14c4	drivers/pinctrl/pinctrl-tegra.o
  18932	   1032	      0	  19964	  
4dfc	drivers/pinctrl/pinctrl-tegra114.o
  20142	   1128	      0	  21270	  
5316	drivers/pinctrl/pinctrl-tegra124.o
  18988	    876	      0	  19864	  
4d98	drivers/pinctrl/pinctrl-tegra20.o
  25781	   1068	      0	  26849	  
68e1	drivers/pinctrl/pinctrl-tegra30.o

--
Stefan

--
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/