[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#153022: Note that this bug is arch-specific



> On debussy (arm), only FE_ROUNDTONEAREST is defined (is this a bug?
> bits/fenv.h seems to imply that the other rounding modes are available
> on arm but accessed through a different API), and it gives the correct
> results.

This is the best we can do with the current ARM instruction set.

Although notionally the ARM FPA does support other rounding modes, the
selection is done by encoding the desired mode into the instruction,
rather than using a separate control register.  There is no way to
change rounding modes at runtime: it can only be done by recompiling the
program.

Clearly this is a fairly losing scheme, and we will be dropping it along
with the rest of the historical ARM floating-point baggage at some point
in the fairly near future.  But, right now, we're stuck with it and
there is nothing we can change in glibc that will improve matters.

p.




Reply to: