2.1.114 won't compile on i586 UP

Horst von Brand (vonbrand@sleipnir.valparaiso.cl)
Mon, 03 Aug 1998 21:37:12 -0500

It bombs in arch/i386/kernel/init_task.c. The trouble is the empty
definition for INIT_LOCKS in include/linux/sched.h, which goes away and
leaves two consecutive commas... but the whole point of INIT_LOCKS was to
be able to move the offending empty structure initialization to the end,
and there just omit it: gcc-2.7.2.3 doesn't like empty structure
initializations, period. So either the spinlock has to be moved to the end
again, or gcc-2.7.2.3 dropped ;-). BTW, egcs-19980727 doesn't mind about
empty structure initializations.

--- linux/include/linux/sched.h-dist-2.1.114 Mon Aug 3 20:07:56 1998
+++ linux/include/linux/sched.h Mon Aug 3 21:34:29 1998
@@ -289,12 +289,16 @@
/* memory management info */
struct mm_struct *mm;
/* signal handlers */
- spinlock_t sigmask_lock; /* Protects signal and blocked */
struct signal_struct *sig;
sigset_t signal, blocked;
struct signal_queue *sigqueue, **sigqueue_tail;
unsigned long sas_ss_sp;
size_t sas_ss_size;
+/*
+ * This MUST be last in order for the ugly kludge around the gcc-2.7.2.3
+ * empty structure initialization bug (triggered on UP) to work - HvB
+ */
+ spinlock_t sigmask_lock; /* Protects signal and blocked */
};

/*
@@ -367,7 +371,8 @@
/* fs */ &init_fs, \
/* files */ &init_files, \
/* mm */ &init_mm, \
-/* signals */ INIT_LOCKS, &init_signals, {{0}}, {{0}}, NULL, &init_task.sigqueue, 0, 0, \
+/* signals */ &init_signals, {{0}}, {{0}}, NULL, &init_task.sigqueue, 0, 0, \
+ INIT_LOCKS \
}

union task_union {

--
Horst von Brand                             vonbrand@sleipnir.valparaiso.cl
Casilla 9G, Viņa del Mar, Chile                               +56 32 672616


-
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.altern.org/andrebalsa/doc/lkml-faq.html