This is the mail archive of the gdb-patches@sources.redhat.com mailing list for the GDB project.


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

Re: RFA: patch to fix 'info dcache' crash


>>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:
Andrew> Want to be listed as ``dcache'' maintainer?

Sure, but after I fill you in on my plans you might not want me to...

At present, the dcache is beneath the target's xfer_memory function.
I'd like to hoist it up a level so it's between target_xfer_memory()
(and target_xfer_memory_partial()) and the target's xfer_memory
function.  Then then dcache will be available for all targets without
requiring explicit support.

* To start, I'd rename dcache_flush() to dcache_invd() (or perhaps
  dcache_invalidate(), what do you think?)  :-)

* Then I'd change the dcache memory read and write functions to be a
  single xfer_memory function with the same prototype as xfer_memory
  vector function.

* After that is done, I'd create a single 'target' dcache.  I'll keep
  the infrastructure in dcache so that we'll be able to have more than
  one dcache when/if GDB is extended to control to multiple targets
  simultaneously.  In target_open, target_resume, etc. I'd add calls
  to invalidate the 'target' dcache.  

* Finally since the memory functions were already modified in step 2,
  it would be trival to move the cache up a level.

Thoughts?

        --jtc

-- 
J.T. Conklin
RedBack Networks

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