ARM: Imply frame pointer for arm-linux profiling
Richard Earnshaw
rearnsha@arm.com
Wed May 11 15:09:00 GMT 2005
On Wed, 2005-05-11 at 15:47, Daniel Jacobowitz wrote:
> > 2) The address of the count word is always stored PIC, even in non-pic
> > code. That means we can profile both normal and PIC code in the same
> > manner.
>
> This is only adequately PIC on systems where the text and data segments
> are loaded at fixed offsets from each other. I've got at least one
> (VxWorks) where this isn't the case. I think some other platforms have
> similar constraints.
>
> I suppose those platforms can use a similar approach to Linux's though;
> no counters at all.
>
If the platform environment relocates the code segment dynamically (eg,
symbianos), then there's no need to make the sequence PIC at all. The
sequence I cited is unashamedly SVr4 centric.
> I'd probably give the with and without counters versions different
> names, for robustness.
>
Yes, that would be sensible. Indeed, with different names the need to
have the same sequence in each point can be reduced, so you could factor
out the thumb/arm and the PIC/absolute/no-counter cases as appropriate.
R.
More information about the Gcc-patches
mailing list