https://sourceware.org/bugzilla/show_bug.cgi?id=32613
Jan Beulich <jbeulich at suse dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Assignee|unassigned at sourceware dot org |jbeulich at suse dot com
Status|NEW |ASSIGNED
--- Comment #2 from Jan Beulich <jbeulich at suse dot com> ---
(In reply to Indu Bhagat from comment #1)
> Can you elaborate on "Which in turn requires them to either be constants, or
> to involve perhaps pretty complex expressions." ? It seemed to me that the
> current interface only allows a byte sequence after .cfi_escape.
Correct. Which for programmers means that they need to reduce anything they
want to pass to expressions yielding byte values. In turn meaning that fields
wider than a byte and requiring relocation can't be expressed at all.
> Also does this mean that GAS will need to validate the contents specified in
> .cfi_escape too?
No, people are still on their own to produce correct data. We'd just make it
easier for them to have data emitted which isn't naturally "bytes only". (By
now I've already coded up the LEB128 part of this. Before posting I'd like to
at least draft the multi-byte-entity one, though.)
--
You are receiving this mail because:
You are on the CC list for the bug.
