From: hpa@zytor.com (H. Peter Anvin)
To: linux-kernel@vger.kernel.org
Subject: Re: [RFC] Splitting kernel headers and deprecating __KERNEL__
Date: Mon, 29 Nov 2004 20:01:39 +0000 (UTC) [thread overview]
Message-ID: <cofv73$us3$1@terminus.zytor.com> (raw)
In-Reply-To: 19865.1101395592@redhat.com
Followup to: <19865.1101395592@redhat.com>
By author: David Howells <dhowells@redhat.com>
In newsgroup: linux.dev.kernel
>
> What we've come up with is this:
>
> (1) Create new directories in the linux sources to shadow existing include
> directories:
>
> NEW DIRECTORY DIRECTORY SHADOWED
> ============= ==================
> include/user/ include/linux/
> include/user-*/ include/asm-*/
>
> Note that this doesn't take account of the other directories under
> include/, but I don't think they're relevant.
>
I'm not sure if user is a good choice, and user-* is even worse. Most
people seem to have suggested include/linux-abi for this; I would
personally prefer include/linux-abi/arch for the second.
> (2) Take each file from the shadowed directory. If it has any userspace
> relevant stuff, then:
>
> (a) Transfer this stuff into a file of the same name in the new
> directory. So, for example, the syscall number list from
> include/asm-i386/unistd.h will be transferred to
> include/user-i386/unistd.h.
I'm not sure you can do such a 1:1 mapping. In fact, there are cases
where you definitely don't want to, because the current structure
doesn't make much sense.
>
> (b) Make kernel file #include the user file. So:
>
> [include/asm-i386/unistd.h]
> ...
> #include <user-i386/unistd.h>
> ...
Good...
> (c) Where a user header file requires something from another header file
> (such as a type), that file should include a suitable user header file
> directly:
>
> [include/user-i386/termio.h]
> ...
> #include <user/types.h>
> ...
Good...
> (d) stdint types should be used where possible.
>
> [include/user-i386/termios.h]
> struct winsize {
> uint16_t ws_row;
> uint16_t ws_col;
> uint16_t ws_xpixel;
> uint16_t ws_ypixel;
> };
Good, except your "struct winsize" is bad; you're stepping on
namespace which belongs to userspace. Since we can't use typedefs on
struct tags, I suggest:
struct __kstruct_winsize {
/* ... */
};
.. and userspace can do:
#define __kstruct_winsize winsize
> (e) These header files should be bounded with __USER_XXXXX_H conditionals:
>
> [include/user-i386/termios.h]
> #ifndef __USER_I386_TERMIOS_H
> #define __USER_I386_TERMIOS_H
> ...
> #endif /* __USER_I386_TERMIOS_H */
Good...
> (3) Remove all #if(n)def __KERNEL__ clauses.
>
> (4) Remove the -D__KERNEL__ from the master kernel Makefile.
Bad! There is code in the kernel which can compile in userspace for
testing. This is highly valuable and should be kept.
-hpa
next prev parent reply other threads:[~2004-11-29 20:05 UTC|newest]
Thread overview: 185+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-11-25 15:13 [RFC] Splitting kernel headers and deprecating __KERNEL__ David Howells
2004-11-25 15:23 ` David Woodhouse
2004-11-25 16:54 ` Matthew Wilcox
2004-11-25 18:17 ` David Woodhouse
2004-11-27 3:24 ` Greg KH
2004-11-28 23:37 ` Paul Mackerras
2004-11-29 1:28 ` Linus Torvalds
2004-11-29 4:36 ` Jeff Garzik
2004-11-29 4:57 ` Al Viro
2004-11-29 5:27 ` Tim Hockin
2004-11-29 7:51 ` Arjan van de Ven
2004-12-05 0:49 ` Rob Landley
2004-12-05 2:26 ` Andries Brouwer
2004-12-05 13:19 ` David Greaves
2004-12-05 15:57 ` Andries Brouwer
2004-12-05 23:05 ` Rob Landley
2004-12-06 9:56 ` Andries Brouwer
2004-11-29 9:42 ` David Woodhouse
2004-11-30 8:08 ` Alex Riesen
2004-11-30 8:57 ` Alex Riesen
2004-12-01 11:46 ` Ralf Baechle
2004-12-02 1:21 ` H. Peter Anvin
2004-11-29 9:53 ` Paul Mackerras
2004-11-29 9:57 ` Arjan van de Ven
2004-11-29 10:01 ` David Woodhouse
2004-11-29 11:41 ` Mariusz Mazur
2004-11-29 17:09 ` Alexandre Oliva
2004-11-29 17:41 ` Linus Torvalds
2004-11-29 22:44 ` Al Viro
2004-12-01 11:36 ` Roman Zippel
[not found] ` <oract0thnj.fsf@livre.redhat.lsd.ic.unicamp.br>
2004-11-29 23:00 ` Linus Torvalds
2004-11-30 0:45 ` H. Peter Anvin
2004-11-30 5:31 ` Herbert Poetzl
2004-11-30 6:44 ` bert hubert
2004-11-30 21:22 ` Alexandre Oliva
2004-11-30 22:22 ` Linus Torvalds
2004-11-30 22:57 ` Alexandre Oliva
2004-11-30 23:12 ` Linus Torvalds
2004-12-01 19:41 ` Alexandre Oliva
2004-12-01 19:54 ` Matthew Wilcox
2004-12-02 3:58 ` Linus Torvalds
2004-12-02 4:28 ` Bernd Eckenfels
2004-12-02 5:11 ` Peter Williams
2004-12-02 11:30 ` Tomas Carnecky
2004-12-02 22:17 ` Peter Williams
2004-12-02 9:29 ` Pekka Enberg
2004-12-02 15:35 ` Linus Torvalds
2004-12-02 21:46 ` Pekka Enberg
2004-11-30 15:33 ` David Howells
2004-11-30 15:51 ` Linus Torvalds
2004-11-30 16:18 ` David Howells
2004-11-30 20:47 ` Alexandre Oliva
2004-11-30 20:56 ` Linus Torvalds
2004-11-30 21:30 ` Alexandre Oliva
2004-11-30 21:39 ` Alexandre Oliva
2004-11-30 23:33 ` Matthew Wilcox
2004-11-30 22:25 ` Linus Torvalds
2004-11-30 22:34 ` David Woodhouse
2004-11-30 22:51 ` Linus Torvalds
2004-11-30 23:50 ` David Woodhouse
2004-12-01 0:10 ` Linus Torvalds
2004-12-01 0:18 ` Linus Torvalds
2004-12-01 8:10 ` Arjan van de Ven
2004-12-01 0:24 ` David Woodhouse
2004-12-01 0:37 ` Linus Torvalds
2004-12-01 0:47 ` David Woodhouse
2004-12-01 0:57 ` Linus Torvalds
2004-12-01 1:06 ` David Woodhouse
2004-12-01 1:23 ` Linus Torvalds
2004-12-01 6:48 ` BAIN
2004-12-01 2:02 ` cdrom.h (was Re: [RFC] Splitting kernel headers...) Matthew Wilcox
2004-11-30 22:52 ` [RFC] Splitting kernel headers and deprecating __KERNEL__ Mariusz Mazur
2004-11-30 23:05 ` Linus Torvalds
2004-11-30 23:05 ` Alexandre Oliva
2004-11-30 15:35 ` David Woodhouse
2004-11-30 15:58 ` Linus Torvalds
2004-11-30 16:28 ` David Woodhouse
2004-11-30 16:53 ` Linus Torvalds
2004-11-30 17:52 ` David Woodhouse
2004-11-30 18:21 ` Linus Torvalds
2004-11-30 20:28 ` Mariusz Mazur
2004-11-30 20:47 ` Linus Torvalds
2004-11-30 22:33 ` Sam Ravnborg
2004-11-30 22:44 ` Mariusz Mazur
2004-11-30 23:03 ` Al Viro
2004-11-30 23:08 ` Mariusz Mazur
2004-12-01 5:23 ` Sam Ravnborg
2004-12-01 10:52 ` Mariusz Mazur
2004-12-01 18:08 ` Sam Ravnborg
2004-11-30 23:13 ` Alexandre Oliva
2004-11-30 23:36 ` Chris Friesen
2004-12-01 5:50 ` H. Peter Anvin
2004-11-30 22:48 ` Matt Mackall
2004-11-30 22:51 ` Christoph Hellwig
2004-11-30 23:09 ` Chris Friesen
2004-11-30 23:35 ` Matt Mackall
2004-12-01 0:23 ` Miquel van Smoorenburg
2004-11-30 22:55 ` Linus Torvalds
2004-11-30 23:29 ` Matt Mackall
2004-12-03 1:03 ` Krzysztof Halasa
2004-11-30 23:51 ` David Woodhouse
2004-11-30 17:55 ` Al Viro
2004-11-30 16:33 ` David Howells
2004-11-30 16:57 ` Linus Torvalds
2004-12-14 5:51 ` Werner Almesberger
2004-12-14 15:49 ` Linus Torvalds
2004-12-14 16:50 ` Werner Almesberger
2004-12-14 17:58 ` Linus Torvalds
2004-12-14 21:46 ` Werner Almesberger
2004-12-14 21:58 ` Linus Torvalds
2004-12-14 23:49 ` Krzysztof Halasa
2004-12-15 0:09 ` Werner Almesberger
2004-12-16 0:58 ` Krzysztof Halasa
2004-12-14 19:23 ` Horst von Brand
2004-12-14 19:45 ` Sam Ravnborg
2004-12-14 19:58 ` Linus Torvalds
2004-12-14 20:25 ` Andreas Schwab
2004-12-14 20:33 ` Linus Torvalds
2004-12-14 16:55 ` Andreas Schwab
2004-12-14 17:59 ` Linus Torvalds
2004-11-26 11:52 ` David Howells
2004-11-27 20:39 ` Sam Ravnborg
2004-11-25 18:20 ` Alexandre Oliva
2004-11-25 21:01 ` Matthew Wilcox
2004-11-25 22:35 ` David Woodhouse
2004-11-26 11:42 ` Alexandre Oliva
2004-11-27 0:13 ` Andreas Steinmetz
2004-11-27 0:26 ` David Woodhouse
2004-11-27 0:30 ` Andreas Steinmetz
2004-11-27 0:38 ` David Woodhouse
2004-11-26 11:47 ` Alexandre Oliva
2004-11-26 11:53 ` David Woodhouse
2004-11-27 16:12 ` Krzysztof Halasa
2004-11-26 11:58 ` David Howells
2004-11-26 12:00 ` David Woodhouse
2004-11-26 14:19 ` Matthew Wilcox
2004-11-26 14:33 ` David Woodhouse
2004-11-27 12:10 ` Alexander Stohr
2004-11-27 3:46 ` Adam Heath
2004-11-30 12:18 ` Horst von Brand
2004-11-29 16:34 ` Alexandre Oliva
2004-11-25 18:45 ` Matthew Wilcox
2004-11-26 23:56 ` Grzegorz Kulewski
2004-11-27 1:27 ` Tomas Carnecky
2004-11-27 1:41 ` Grzegorz Kulewski
2004-11-27 2:03 ` Tomas Carnecky
2004-11-29 11:08 ` David Howells
2004-11-27 3:07 ` Kyle Moffett
2004-11-27 4:29 ` Tonnerre
2004-11-27 3:51 ` Matthew Wilcox
2004-11-27 16:47 ` Randy.Dunlap
2004-11-27 17:16 ` David Woodhouse
2004-11-27 22:53 ` Arnd Bergmann
2004-11-27 23:12 ` David Woodhouse
2004-11-27 23:27 ` Randy.Dunlap
2004-11-27 23:32 ` David Woodhouse
2004-11-27 23:56 ` Arnd Bergmann
2004-11-28 7:13 ` Arjan van de Ven
2004-11-28 12:03 ` Arnd Bergmann
2004-11-28 12:19 ` Arjan van de Ven
2004-11-28 16:34 ` Kevin Puetz
2004-11-28 12:28 ` Wichert Akkerman
2004-11-28 12:44 ` Arjan van de Ven
2004-11-28 16:21 ` Jakub Jelinek
2004-11-28 7:20 ` Christoph Hellwig
2004-11-27 4:05 ` Chris Friesen
2004-11-27 4:30 ` Tomas Carnecky
2004-11-27 21:03 ` Sam Ravnborg
2004-11-27 21:11 ` Andreas Steinmetz
2004-11-27 21:19 ` Sam Ravnborg
2004-11-27 21:49 ` Andreas Steinmetz
2004-11-27 23:21 ` Arnd Bergmann
2004-11-28 13:24 ` Andreas Steinmetz
2004-11-28 13:40 ` Arnd Bergmann
2004-11-29 20:01 ` H. Peter Anvin [this message]
2004-11-30 0:34 ` Kyle Moffett
2004-11-30 0:51 ` H. Peter Anvin
2004-12-14 7:07 ` Werner Almesberger
-- strict thread matches above, loose matches on Subject: below --
2004-11-27 19:43 Dan Kegel
2004-11-27 21:06 ` Sam Ravnborg
[not found] <34Xo6-2P0-19@gated-at.bofh.it>
[not found] ` <35i9f-2vZ-25@gated-at.bofh.it>
[not found] ` <35iLS-2Uo-1@gated-at.bofh.it>
[not found] ` <35kb6-46Q-25@gated-at.bofh.it>
2004-11-28 0:55 ` ak
2004-11-28 13:32 ` Arnd Bergmann
2004-11-30 1:43 Jean Tourrilhes
2004-11-30 4:22 ` Linus Torvalds
2004-11-30 6:51 ` H. Peter Anvin
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='cofv73$us3$1@terminus.zytor.com' \
--to=hpa@zytor.com \
--cc=linux-kernel@vger.kernel.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.