Re: DPKG not compiling due to differences between Alpha and Intel...

["David S. Miller" <davem@caip.rutgers.edu>]

   Date: Tue, 23 Jan 96 23:51 GMT
   From: Ian Jackson <ian@chiark.chu.cam.ac.uk>

   > This is why there exists no sa_restorer member.  I hope this helps
   > people out.

   So, what am I supposed to do when I call sigaction ?

I do not know if Linux libc on the i386 dorks with sa_restorer, but I
can tell you what is up with the kernel end of things.  In the olden
days the signal state was stored wherever this sa_restorer thing
pointed to.  Now, and for the purposes of being ibcs2 compliant, the
signal save state is thrown on the processes stack.

For compatability reasons the sa_restorer is still checked, if it is
non-null the stack save state is thrown wherever that points to.  So
it seems not a bad idea to set it to NULL in your application.
*However* I would imagine that H.J. sets it to null on purpose in libc
to get the new behavior.  I would check the libc sources before doing
this.

As a point of reference, I just looked at bash, to check things out.
All it cares about is whether signal handlers return void or int, it
makes no reference to sa_restorer or similar and works.  Make whatever
conclusions you want...  I suggest checking out libc directly or
contacting H.J. Liu for the real scoop.

Later,
David S. Miller
davem@caip.rutgers.edu