Messages in this thread |  | | | Date | Sat, 14 Nov 1998 03:07:33 -0500 | | From | Ion Badulescu <> | | Subject | Re: include file conflict |
| |
On 11 Nov 1998 02:24:54 GMT, H. Peter Anvin wrote: > > I think there are legitimate uses for the <linux/*.h> style headers, > even within glibc. For example <linux/auto_fs.h> contains the > communications protocol which the kernel uses to talk to the autofs > daemon. > > However, kernel headers should use #ifdef __KERNEL__ for most things > and only export things that are truly needed in user space.
Finally a reasonable attitude. :-) Obviously, anything bypassing the libc interface needs to use kernel protocols, data type, data structures, etc. And -- Linus knows this fact well, and uses it at times -- a compile error due to a header file that changed is much preferrable to a silent data corruption from an application that compiled correctly with its own headers.
The am-utils has the same problem as autofs, only it's larger by about one order of magnitude. Just imagine trying to compile the beast under libc5 _and_ glibc, kernel 2.0 _and_ 2.1, not to count some 10 other OS's, and you get the idea. To call the resulting configure script "complex" is a gross understatement.
I am willing to clean up the kernel headers with proper #ifdef __KERNEL__ and #ifdef __KERNEL_STRICT_NAMES, but I'd like to know that my work will have some finality. Would such patches be accepted still in 2.1 or am I wasting my time?
Thanks, Ion
-- It is better to keep your mouth shut and be thought a fool, than to open it and remove all doubt.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
|  |