All of lore.kernel.org
 help / color / mirror / Atom feed
From: Arnd Bergmann <arnd@arndb.de>
To: "Russell King - ARM Linux" <linux@arm.linux.org.uk>
Cc: linux-arm-kernel@lists.infradead.org,
	Jimmy Rubin <jimmy.rubin@stericsson.com>,
	Dan Johansson <dan.johansson@stericsson.com>,
	linux-fbdev@vger.kernel.org,
	Linus Walleij <linus.walleij@stericsson.com>,
	linux-media@vger.kernel.org
Subject: Re: [PATCH 02/10] MCDE: Add configuration registers
Date: Mon, 15 Nov 2010 15:25:54 +0100	[thread overview]
Message-ID: <201011151525.54380.arnd@arndb.de> (raw)
In-Reply-To: <20101112153423.GC3619@n2100.arm.linux.org.uk>

On Friday 12 November 2010, Russell King - ARM Linux wrote:
> On Fri, Nov 12, 2010 at 04:14:51PM +0100, Arnd Bergmann wrote:
> > Some people prefer to express all this in C instead of macros:
> > 
> > struct mcde_registers {
> > 	enum {
> > 		mcde_cr_dsicmd2_en = 0x00000001,
> > 		mcde_cr_dsicmd1_en = 0x00000002,
> > 		...
> > 	} cr;
> > 	enum {
> > 		mcde_conf0_syncmux0 = 0x00000001,
> > 		...
> > 	} conf0;
> > 	...
> > };
> > 
> > This gives you better type safety, but which one you choose is your decision.
> 
> It is a bad idea to describe device registers using C structures, and
> especially enums.
> 
> The only thing C guarantees about structure layout is that the elements
> are arranged in the same order which you specify them in your definition.
> It doesn't make any guarantees about placement of those elements within
> the structure.

Right, I got carried away when seeing the macro overload. My example
would work on a given architecture since the ABI is not changing, but
we should of course not advocate nonportable code.

Normally what I do is to describe the data structure in C and define the
values in a separate enum. The main advantage of using the struct instead
of offset defines is that you have a bit more type safety, i.e. you cannot
accidentally do readw() on a __be32 member.

Using #define for the actual values makes it possible to interleave the
values with the structure definition like 

struct mcde_registers {
 	__le32 cr;
#define MCDE_CR_DSICMD2_EN 0x00000001
#define MCDE_CR_DSICMD1_EN 0x00000002
	__le32 conf0;
 	...
};

whereas the enum has the small advantage of putting the identifiers
into the C language namespace rather than the preprocessor macro
namespace.

	Arnd

WARNING: multiple messages have this Message-ID (diff)
From: Arnd Bergmann <arnd@arndb.de>
To: linux-arm-kernel@lists.infradead.org
Subject: Re: [PATCH 02/10] MCDE: Add configuration registers
Date: Mon, 15 Nov 2010 14:25:54 +0000	[thread overview]
Message-ID: <201011151525.54380.arnd@arndb.de> (raw)
In-Reply-To: <20101112153423.GC3619@n2100.arm.linux.org.uk>

On Friday 12 November 2010, Russell King - ARM Linux wrote:
> On Fri, Nov 12, 2010 at 04:14:51PM +0100, Arnd Bergmann wrote:
> > Some people prefer to express all this in C instead of macros:
> > 
> > struct mcde_registers {
> > 	enum {
> > 		mcde_cr_dsicmd2_en = 0x00000001,
> > 		mcde_cr_dsicmd1_en = 0x00000002,
> > 		...
> > 	} cr;
> > 	enum {
> > 		mcde_conf0_syncmux0 = 0x00000001,
> > 		...
> > 	} conf0;
> > 	...
> > };
> > 
> > This gives you better type safety, but which one you choose is your decision.
> 
> It is a bad idea to describe device registers using C structures, and
> especially enums.
> 
> The only thing C guarantees about structure layout is that the elements
> are arranged in the same order which you specify them in your definition.
> It doesn't make any guarantees about placement of those elements within
> the structure.

Right, I got carried away when seeing the macro overload. My example
would work on a given architecture since the ABI is not changing, but
we should of course not advocate nonportable code.

Normally what I do is to describe the data structure in C and define the
values in a separate enum. The main advantage of using the struct instead
of offset defines is that you have a bit more type safety, i.e. you cannot
accidentally do readw() on a __be32 member.

Using #define for the actual values makes it possible to interleave the
values with the structure definition like 

struct mcde_registers {
 	__le32 cr;
#define MCDE_CR_DSICMD2_EN 0x00000001
#define MCDE_CR_DSICMD1_EN 0x00000002
	__le32 conf0;
 	...
};

whereas the enum has the small advantage of putting the identifiers
into the C language namespace rather than the preprocessor macro
namespace.

	Arnd

WARNING: multiple messages have this Message-ID (diff)
From: arnd@arndb.de (Arnd Bergmann)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH 02/10] MCDE: Add configuration registers
Date: Mon, 15 Nov 2010 15:25:54 +0100	[thread overview]
Message-ID: <201011151525.54380.arnd@arndb.de> (raw)
In-Reply-To: <20101112153423.GC3619@n2100.arm.linux.org.uk>

On Friday 12 November 2010, Russell King - ARM Linux wrote:
> On Fri, Nov 12, 2010 at 04:14:51PM +0100, Arnd Bergmann wrote:
> > Some people prefer to express all this in C instead of macros:
> > 
> > struct mcde_registers {
> > 	enum {
> > 		mcde_cr_dsicmd2_en = 0x00000001,
> > 		mcde_cr_dsicmd1_en = 0x00000002,
> > 		...
> > 	} cr;
> > 	enum {
> > 		mcde_conf0_syncmux0 = 0x00000001,
> > 		...
> > 	} conf0;
> > 	...
> > };
> > 
> > This gives you better type safety, but which one you choose is your decision.
> 
> It is a bad idea to describe device registers using C structures, and
> especially enums.
> 
> The only thing C guarantees about structure layout is that the elements
> are arranged in the same order which you specify them in your definition.
> It doesn't make any guarantees about placement of those elements within
> the structure.

Right, I got carried away when seeing the macro overload. My example
would work on a given architecture since the ABI is not changing, but
we should of course not advocate nonportable code.

Normally what I do is to describe the data structure in C and define the
values in a separate enum. The main advantage of using the struct instead
of offset defines is that you have a bit more type safety, i.e. you cannot
accidentally do readw() on a __be32 member.

Using #define for the actual values makes it possible to interleave the
values with the structure definition like 

struct mcde_registers {
 	__le32 cr;
#define MCDE_CR_DSICMD2_EN 0x00000001
#define MCDE_CR_DSICMD1_EN 0x00000002
	__le32 conf0;
 	...
};

whereas the enum has the small advantage of putting the identifiers
into the C language namespace rather than the preprocessor macro
namespace.

	Arnd

  reply	other threads:[~2010-11-15 14:25 UTC|newest]

Thread overview: 130+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-11-10 12:04 [PATCH 00/10] MCDE: Add frame buffer device driver Jimmy Rubin
2010-11-10 12:04 ` Jimmy Rubin
2010-11-10 12:04 ` Jimmy Rubin
2010-11-10 12:04 ` [PATCH 01/10] MCDE: Add hardware abstraction layer Jimmy Rubin
2010-11-10 12:04   ` Jimmy Rubin
2010-11-10 12:04   ` Jimmy Rubin
2010-11-10 12:04   ` [PATCH 02/10] MCDE: Add configuration registers Jimmy Rubin
2010-11-10 12:04     ` Jimmy Rubin
2010-11-10 12:04     ` Jimmy Rubin
2010-11-10 12:04     ` [PATCH 03/10] MCDE: Add pixel processing registers Jimmy Rubin
2010-11-10 12:04       ` Jimmy Rubin
2010-11-10 12:04       ` Jimmy Rubin
2010-11-10 12:04       ` [PATCH 04/10] MCDE: Add formatter registers Jimmy Rubin
2010-11-10 12:04         ` Jimmy Rubin
2010-11-10 12:04         ` Jimmy Rubin
2010-11-10 12:04         ` [PATCH 05/10] MCDE: Add dsi link registers Jimmy Rubin
2010-11-10 12:04           ` Jimmy Rubin
2010-11-10 12:04           ` Jimmy Rubin
2010-11-10 12:04           ` [PATCH 06/10] MCDE: Add generic display Jimmy Rubin
2010-11-10 12:04             ` Jimmy Rubin
2010-11-10 12:04             ` Jimmy Rubin
2010-11-10 12:04             ` [PATCH 07/10] MCDE: Add display subsystem framework Jimmy Rubin
2010-11-10 12:04               ` Jimmy Rubin
2010-11-10 12:04               ` Jimmy Rubin
2010-11-10 12:04               ` [PATCH 08/10] MCDE: Add frame buffer device Jimmy Rubin
2010-11-10 12:04                 ` Jimmy Rubin
2010-11-10 12:04                 ` Jimmy Rubin
2010-11-10 12:04                 ` [PATCH 09/10] MCDE: Add build files and bus Jimmy Rubin
2010-11-10 12:04                   ` Jimmy Rubin
2010-11-10 12:04                   ` Jimmy Rubin
2010-11-10 12:04                   ` [PATCH 10/10] ux500: MCDE: Add platform specific data Jimmy Rubin
2010-11-10 12:04                     ` Jimmy Rubin
2010-11-10 12:04                     ` Jimmy Rubin
2010-11-12 16:03                     ` Arnd Bergmann
2010-11-12 16:03                       ` Arnd Bergmann
2010-11-12 16:03                       ` Arnd Bergmann
2010-11-25 11:20                       ` Jimmy RUBIN
2010-11-25 11:20                         ` Jimmy RUBIN
2010-11-12 16:23                   ` [PATCH 09/10] MCDE: Add build files and bus Arnd Bergmann
2010-11-12 16:23                     ` Arnd Bergmann
2010-11-12 16:23                     ` Arnd Bergmann
2010-11-12 16:29                 ` [PATCH 08/10] MCDE: Add frame buffer device Arnd Bergmann
2010-11-12 16:29                   ` Arnd Bergmann
2010-11-12 16:29                   ` Arnd Bergmann
2010-11-25 11:52                   ` Jimmy RUBIN
2010-11-25 11:52                     ` Jimmy RUBIN
2010-11-12 16:38               ` [PATCH 07/10] MCDE: Add display subsystem framework Arnd Bergmann
2010-11-12 16:38                 ` Arnd Bergmann
2010-11-12 16:38                 ` Arnd Bergmann
2010-11-25  7:16                 ` Jimmy RUBIN
2010-11-25  7:16                   ` Jimmy RUBIN
2010-11-12 15:46       ` [PATCH 03/10] MCDE: Add pixel processing registers Arnd Bergmann
2010-11-12 15:46         ` Arnd Bergmann
2010-11-12 15:46         ` Arnd Bergmann
2010-11-12 15:14     ` [PATCH 02/10] MCDE: Add configuration registers Arnd Bergmann
2010-11-12 15:14       ` Arnd Bergmann
2010-11-12 15:14       ` Arnd Bergmann
2010-11-12 15:34       ` Russell King - ARM Linux
2010-11-12 15:34         ` Russell King - ARM Linux
2010-11-12 15:34         ` Russell King - ARM Linux
2010-11-15 14:25         ` Arnd Bergmann [this message]
2010-11-15 14:25           ` Arnd Bergmann
2010-11-15 14:25           ` Arnd Bergmann
2010-11-15 14:59           ` Russell King - ARM Linux
2010-11-15 14:59             ` Russell King - ARM Linux
2010-11-15 14:59             ` Russell King - ARM Linux
2010-11-15 18:24             ` Geert Uytterhoeven
2010-11-15 18:24               ` Geert Uytterhoeven
2010-11-15 18:24               ` Geert Uytterhoeven
2010-11-25 11:30       ` Jimmy RUBIN
2010-11-25 11:30         ` Jimmy RUBIN
2010-11-25 16:21         ` Arnd Bergmann
2010-11-25 16:21           ` Arnd Bergmann
2010-11-10 17:14   ` [PATCH 01/10] MCDE: Add hardware abstraction layer Joe Perches
2010-11-10 17:14     ` Joe Perches
2010-11-10 17:14     ` Joe Perches
2010-11-15  9:52     ` Jimmy RUBIN
2010-11-15  9:52       ` Jimmy RUBIN
2010-11-15  9:52       ` Jimmy RUBIN
2010-11-15 16:30       ` Joe Perches
2010-11-15 16:30         ` Joe Perches
2010-11-15 16:30         ` Joe Perches
2010-11-12 15:43   ` Arnd Bergmann
2010-11-12 15:43     ` Arnd Bergmann
2010-11-12 15:43     ` Arnd Bergmann
2010-11-16 15:29     ` Jimmy RUBIN
2010-11-16 15:29       ` Jimmy RUBIN
2010-11-16 15:29       ` Jimmy RUBIN
2010-11-16 16:12       ` Arnd Bergmann
2010-11-16 16:12         ` Arnd Bergmann
2010-11-16 16:12         ` Arnd Bergmann
2010-11-16 16:16         ` Arnd Bergmann
2010-11-16 16:16           ` Arnd Bergmann
2010-11-16 16:16           ` Arnd Bergmann
2010-11-16 19:46       ` Joe Perches
2010-11-16 19:46         ` Joe Perches
2010-11-16 19:46         ` Joe Perches
2010-11-17  9:55         ` Arnd Bergmann
2010-11-17  9:55           ` Arnd Bergmann
2010-11-17  9:55           ` Arnd Bergmann
2010-11-17 16:01           ` Joe Perches
2010-11-17 16:01             ` Joe Perches
2010-11-17 16:01             ` Joe Perches
2010-11-10 14:42 ` [PATCH 00/10] MCDE: Add frame buffer device driver Alex Deucher
2010-11-10 14:42   ` Alex Deucher
2010-11-10 14:42   ` Alex Deucher
2010-11-12 13:18   ` Jimmy RUBIN
2010-11-12 13:18     ` Jimmy RUBIN
2010-11-12 13:18     ` Jimmy RUBIN
2010-11-12 15:52     ` Alex Deucher
2010-11-12 15:52       ` Alex Deucher
2010-11-12 15:52       ` Alex Deucher
2010-11-12 16:46       ` Marcus LORENTZON
2010-11-12 16:46         ` Marcus LORENTZON
2010-11-12 16:46         ` Marcus LORENTZON
2010-11-12 17:22         ` Alex Deucher
2010-11-12 17:22           ` Alex Deucher
2010-11-12 17:22           ` Alex Deucher
2010-11-15 11:05           ` Michel Dänzer
2010-11-15 11:05             ` Michel Dänzer
2010-11-15 11:05             ` Michel Dänzer
2010-11-13 11:54         ` Hans Verkuil
2010-11-13 11:54           ` Hans Verkuil
2010-11-13 11:54           ` Hans Verkuil
2010-11-13 17:26           ` Marcus LORENTZON
2010-11-13 17:26             ` Marcus LORENTZON
2010-11-13 17:26             ` Marcus LORENTZON
2010-11-13 17:57             ` Hans Verkuil
2010-11-13 17:57               ` Hans Verkuil
2010-11-13 17:57               ` Hans Verkuil

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=201011151525.54380.arnd@arndb.de \
    --to=arnd@arndb.de \
    --cc=dan.johansson@stericsson.com \
    --cc=jimmy.rubin@stericsson.com \
    --cc=linus.walleij@stericsson.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-fbdev@vger.kernel.org \
    --cc=linux-media@vger.kernel.org \
    --cc=linux@arm.linux.org.uk \
    /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.