On Fri, Sep 19, 2003 at 08:33:06PM +0200, Ralf Baechle wrote:
> The original reason for the kernel change was a bug report from Atushi
> Nemoto about a mismatch of the and libc msqid_ds. It seems there are
> already libcs with this patch in circulation which at least Atushi and
> me seemed to have; I don't know why this isn't in the official libc ...
>
> Adding Atushi to the cc list; maybe he can shed some light on this.
Since I haven't heared back yet, time to poke the dead horse again ;-)
So basically I like Thiemo's suggestion for the fix. But - the purpose
of the three unused 32-bit fields in struct msgid64_ds is dealing with
the year 2038 problem. So maybe we should reorder fields like:
[...]
#if defined(CONFIG_MIPS32) && !defined(CONFIG_CPU_LITTLE_ENDIAN)
unsigned long __unused1;
__kernel_time_t msg_stime;
#else if defined(CONFIG_MIPS32) && defined(CONFIG_CPU_LITTLE_ENDIAN)
__kernel_time_t msg_stime;
unsigned long __unused1;
#else
__kernel_time_t msg_stime;
#endif
[...]
?
That would eventually permit extending fields to 64-bit and take care of
endianess issues.
Comments?
Ralf
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
