> Hi. After yesterday's discussion about problems with my 32-bit UID patches
> (namely the fact that they broke compiles of programs using msgctl(),
> semctl(), and shmctl() under libc5), I've made the following changes:
> - struct ipc_perm, msqid_ds, semid_ds, and shmid_ds (in
> include/linux/{ipc,msg,sem,shm}.h) do not change, so that people
> can compile applications under libc5 and the modified kernel
> headers
> - include/asm-{arch}/unistd.h does not rename existing system
> calls; this would confuse builds of the C library
> There are a few things that I need to get clarified before I can fix up
> the rest of my patches, however:
> - When the C library is being built, several kernel header
> files are referenced. Does this mean that nothing can ever
> change in the kernel header files that would result in an old C
> library being built incorrectly?
No - btw. for glibc isssues, you should contact
libc-alpha@sourceware.cygnus.com.
> - in other words, can the following never change (to the extent
> that existing types or structure fields cannot be renamed):
> __kernel_uid_t in include/asm-{arch}/posix_types.h
> struct stat in include/asm-{arch}/stat.h
> union siginfo_t in include/asm-{arch}/siginfo.h
AFAIK those shouldn't be changed - or need similar changes to glibc.
But don't take this for granted.
> - if this is the case, I will then go ahead and do the following:
> define __kernel_uid32_t on selected architectures, without
> changing __kernel_uid_t
> change include/linux/types.h to typedef uid_t and gid_t
> to __kernel_uid32_t and __kernel_gid32_t on these
> architectures
> fix up all places using __kernel_uid_t and make sure they
> use the right one (uid_t or uid32_t)
> not change any existing structure fields from uid to
> uid16, but rather define new fields called uid32
> Thanks to all who have helped me with this; there are a lot of people who
> really need 32-bit UIDs in Linux, and I intend to help them out by making
> the 32-bit UID patches as correct as possible.
Glad to hear!
Andreas
-- Andreas Jaeger SuSE Labs aj@suse.de private aj@arthur.rhein-neckar.de- 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/