This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Reload fixes for EXTRA_MEMORY_CONSTRAINT problems

> Now, I'm not sure how to best fix it: either by just setting the
> operand_mode to Pmode in the 'p' case as well, or by prohibiting
> optional reloads for address operands.  (However, there is a comment
> which I don't understand that says optional reloads are necessary
> in that case or else reload inheritance might break ?!)

Where is that comment?

reloads used to be quite hard to follow, because register elimination
and address reloads could change multiple references to the same pseudo
into dissimilar code.  Most of the information about the original pseudos
that were referred to is now conveyed via the in_reg and out_reg fields
of a reload, which are not only used to drive inheritance, but also to
invalidate information used by inheritance when it becomes stale.

However, we have a mechanism to handle cases that can not be described
inside a reload: we then emit USEes and/or CLOBBERs of the original
pseudo register.

SuperH (UK) Ltd.
2410 Aztec West / Almondsbury / BRISTOL / BS32 4QX
T:+44 1454 465658

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]