Phil Blundell wrote: > On Fri, 2002-03-01 at 09:24, jeroen dobbelaere wrote: > >>I'm wondering now where this signal _should_ be disabled. >>- in the linux kernel >>- in the startup code, provided by the compiler (crti.o) >>- in the program itself >> > > This should already be happening in the C library startup code. I don't > know why it's apparently not working for you. Can you assemble a > testcase that exhibits the bug? > > p. > >
Maybe I was too fast to start asking around for this :( My testcases don't show the problem : division by 0.0, 0.0/0.0, conversion of very large values to long : they all just continue without trapping, returning the expected values. Even when done in multiple threads. So, now I'm confused why mozilla would signal a SIGFPE, while the default settings should be ok... ...still looking at it, removing my 'sigprocmask' code from the mozilla source, recompiling... Greetings, -- Jeroen Dobbelaere Embedded Software Engineer ACUNIA Embedded Solutions http://www.acunia.com _______________________________________________ http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm http://www.arm.linux.org.uk/armlinux/mailinglists.php Please visit the above addresses for information on this list.
