Re: [PATCH] Use CORE_ADDR as return type from x86_dr_low_get_addr

[Simon Marchi via Gdb-patches <gdb-patches@sourceware.org>]

On 2021-09-03 1:00 p.m., Tom Tromey wrote:
> Tom> It just seemed to me that, most of the time, the value really is a
> Tom> CORE_ADDR, and that anyway a CORE_ADDR would definitely be large enough.
> 
> Thinking about this a bit more -- it seems wrong to me to use uintptr_t
> (a host type) to represent a register value (a target type).  So from
> this perspective, CORE_ADDR is more correct.  Now, it's not completely
> correct in the sense that not every DR register holds an address.  But,
> it's at least harmless in those cases.

This is in windows-nat.c, where host == target, so I think uintptr_t is
somewhat correct given the uses cases we want to support (GDB 32
debugging inferior 32, GDB 64 debugging inferior 32 and GDB 64 debugging
inferior 64).  If it were in windows-tdep.c, it would definitely be
wrong.

But uintptr_t isn't any more correct than CORE_ADDR, so I'm not against
the change either.

Simon