Re: [PATCH] Futex Generalization Patch

["Peter Wächtler" <pwaechtler@loewe-komp.de>]

Hubertus Franke wrote:
> 
> On Monday 15 April 2002 10:49 am, Bill Abt wrote:
> > Dealing with the realtime signal is not a problem.  Also, saving the extra
> > system call is *BIG* bonus.
> >
> 
> Cool
> 
> As of Peter's initial message. I took a look at the siginfo_t and Peter's
> statement needs to be corrected "a bit".
> All the members he listed are NOT necessarily available.
> 

Well, then we do not comply to susv2. But that's not so bad.

But what we need is there:

typedef union sigval {
	int sival_int;
	void *sival_ptr;
} sigval_t;
[...]
>                 /* POSIX.1b signals */
>                 struct {
>                         pid_t _pid;             /* sender's pid */
>                         uid_t _uid;             /* sender's uid */
>                         sigval_t _sigval;
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>                 } _rt;
> 
>         } _sifields;
> } siginfo_t;
> 
> I'd suggest we tag along the _sigfault semantics.
> We don't need to know who woke us up, just which <addr> got signalled.
> 

Please not, this would be confusing.

we can add our si_code and our own struct to the union as suggested
by Mark and me.



--- /usr/local/src/linux-2.5.7/include/asm-i386/siginfo.h ---
/*
 * si_code values
 * Digital reserves positive values for kernel-generated signals.
 */
#define SI_USER		0		/* sent by kill, sigsend, raise */
#define SI_KERNEL	0x80		/* sent by the kernel from somewhere */
#define SI_QUEUE	-1		/* sent by sigqueue */
#define SI_TIMER __SI_CODE(__SI_TIMER,-2) /* sent by timer expiration */
#define SI_MESGQ	-3		/* sent by real time mesq state change */
#define SI_ASYNCIO	-4		/* sent by AIO completion */
#define SI_SIGIO	-5		/* sent by queued SIGIO */
#define SI_TKILL	-6		/* sent by tkill system call */
#define SI_DETHREAD	-7		/* sent by execve() killing subsidiary threads */
--- snip ---