Re: No support for hardware single step in gdb?

[Peter Reilley]

It is not clear if it is a pipe or cache problem.   In looking
further at the gdb code for istep, I think that removing
breakpoints before doing an istep is a good solution.
The breakpoints are inserted and then removed for before
an istep.   This is unnecessary but not too bad for performance
and it avoids the work about the instruction pipe/cache.

Thanks,
Pete.


-----Original Message-----
From: Art Berggreen <address@hidden>
Newsgroups: gnu.gdb.bug
To: address@hidden <address@hidden>
Date: Wednesday, October 11, 2000 10:59 AM
Subject: Re: No support for hardware single step in gdb?


>Peter Reilley wrote:
>> This problem is seen in the 8260 chip only.   I have the IBM 403 chip
>> working OK with the current way of handling istep.   On the
>> 8260 it seems that the instruction queue has already read the
>> breakpoint instruction from memory and restoring the original
>> instruction to memory does not cause the queue to be flushed
>> and the correct data to be read.
>
>Is it an instruction pipeline issue or an instruction cache problem?
>If your machine has an I cache which is enabled, you have to be sure
>that you flush and invalidate your caches when resuming execution.
>
>Art
>
>_______________________________________________
>Bug-gdb mailing list
>address@hidden
>http://mail.gnu.org/mailman/listinfo/bug-gdb
>