Re: hexagon: signal handling bugs

[Linas Vepstas]

On 11 February 2012 17:27, Al Viro <viro@xxxxxxxxxxxxxxxxxx> wrote:
> Â Â Â Â1) unless I'm seriously misreading hexagon vm_entry.S, once we'd
> called do_notify_resume, we do *not* recheck if there's more work to
> be done.

> Â Â Â Â2) sigreturn should *NOT* be restartable at all, or you'll get
....

Is there an architecture that "does this best"?  I (re-)wrote a
significant part of the hexagon sig-handling code, and I did it via
"surgical hacking": by reading the code for 4-6 other arches, and
trying to pick out the "best practices" of each.  But its easy to
build a Frankenstein when using body parts from mis-matched places, so
perhaps many of these mistakes were mine...

Perhaps its desirable to modify most/all architectures, so as to evoke
more commonality?  Some of the other, more minor arches sure did look
incomplete to me...

-- Linas
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/