RE: [PATCH v2] ARM: unwind: improve unwinders for noreturn case

[David Laight <David.Laight@ACULAB.COM>]

...
> Whereas we already provide an abort() function because iirc the
> compiler used to emit branches to that due to noreturn functions. If
> correct, there's previous convention for doing this - and abort() is
> still exists in the kernel and in userspace since it's part of ANSI
> C. This would be a more reliable and portable solution, but probably
> not for embedded platforms - and that's probably why it got removed.

Wouldn't you want it to do a 'bl abort' so that you could do a backtrace
to find out which 'noreturn' function had returned?

But that leaves you with another 'noreturn' function that you have
difficulty generating a backtrace from.

So you'd need a compiler option to specify what to put there.
I'd suspect linux would like something that generates an 'undefined
instruction' trap - since that would be expected to fault with the
saved PC pointing to the instruction itself (but architectures may vary).

'One size' definitely doesn't 'fit all' :-)

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)