Mail Archives: djgpp/2000/04/10/06:48:21
Eli Zaretskii <eliz AT is DOT elta DOT co DOT il> wrote:
> Hans-Bernhard Broeker wrote:
>> *If* the current instruction is using a segment override prefix, which
>> it doesn't understand, FSDB will look in the wrong place
> Are you saying that FSDB doesn't grok segment overrides?
Not really. That sentence was misunderstandable. I wanted to propose
this as a possible reason, not as a fact. I.e.: in case FSDB
misunderstands the seg override, this might well result in the type of
crash observed. It'd also explain why this doesn't happen to
everyone, all of the time: most functions won't be using segment
overrides, so the bug wouldn't happen, then.
> In any case, I don't think FSDB would crash even if it does try to access a
> bad address.
IIRC, getting proper handling of unmapped memory regions right in our
GDB port took quite a while, and several tries, didn't it? The
supposed error in handling segment overrides would lead to similar
problems, which haven't been tested as thoroughly as the GDB ones, to
my knowledge, so a crash would be no surprise at all, to me.
> I'd need to see a specific test case.
That's for the original poster to offer.
--
Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de)
Even if all the snow were burnt, ashes would remain.
- Raw text -