Re: Debugging data on Alpha

[Bernd Schmidt <bernds at redhat dot com>]

On Fri, 29 Dec 2000, Richard Henderson wrote:

> So find a large test case, breaking on a particular line,
> looking at a particular variable.  I.e. record as many of
> the specifics as you can the next time it happens.
>
> This isn't the first time you've mentioned this but failed
> to provide anything at all to go on.

My guess is that the scheduler messes up the BLOCK_BEG/BLOCK_END notes.
I've seen that happen a few times; if a BLOCK_END note is found inside
a basic block, it'll move upward to the start of the block.

I had an idea yesterday.  I was working on a few patches that make us
keep the original pseudo register number in the second field of a REG,
so that we can look it up after reload has completed.  The idea is that
maybe this could be used as a starting point for better debugging
support with register renaming and LRS optimizations.  We'd stop bothering
with the BLOCK notes and re-create the necessary information during the
final pass by establishing a mapping between hard regs and user variables
at all points in the function.
I'm not really familiar with the debugging output code, so I can't tell
whether this is feasible.  Comments?


Bernd