Re: IA32: printing FP register variables

[Richard Henderson <rth@cygnus.com>]

On Tue, Jul 13, 1999 at 03:25:02PM -0500, Jim Blandy wrote:
> But numbering registers relative to the FP stack base will still reduce
> the number of ranges substantially, won't it?

Yes.


r~
From jimb@cygnus.com Wed Jul 14 17:06:00 1999
From: Jim Blandy <jimb@cygnus.com>
To: Eli Zaretskii <eliz@delorie.com>
Cc: gdb@sourceware.cygnus.com, DJ Delorie <dj@delorie.com>
Subject: Re: Single-stepping through INT nn instructions
Date: Wed, 14 Jul 1999 17:06:00 -0000
Message-id: <np7lo2hkam.fsf@zwingli.cygnus.com>
References: <199907140747.DAA03375@indy.delorie.com>
X-SW-Source: 1999-q3/msg00067.html
Content-length: 1023

> I'm not sure if this is the right forum to discuss the above; if not,
> please tell me where to post.

I think this is best for gdb@sourceware.cygnus.com.  I've changed the
CC'd address.

> The INT nn and INTO instructions reset the trace bit on x86.  So if
> you are single-stepping through a function that issues these
> instructions, the single-step mode is effectively turned off when you
> step over one of them.

So you're trying to step *into* an int, not over it, right?  It seems
to me that the TF flag ought to work for stepping over an int.

The pentium manual says:

    The INT instructions, however, do clear the TF flag.  Therefore,
    software debuggers which single-step code must recognize and
    emulate INT n or INTO instructions rather than executing them
    directly.

You might look into defining SOFTWARE_SINGLE_STEP for i386; it could
check the next instruction, emulate INT n and INTO, and use the TF
flag for all others.

Are you sure the other platforms can do this?  I bet they just punt.
From eliz@delorie.com Thu Jul 15 02:14:00 1999
From: Eli Zaretskii <eliz@delorie.com>
To: Jim Blandy <jimb@cygnus.com>
Cc: gdb@sourceware.cygnus.com, DJ Delorie <dj@delorie.com>
Subject: Re: Single-stepping through INT nn instructions
Date: Thu, 15 Jul 1999 02:14:00 -0000
Message-id: <199907150913.FAA01508@indy.delorie.com>
X-SW-Source: 1999-q3/msg00068.html
Content-length: 482

> So you're trying to step *into* an int, not over it, right?

I meant this: suppose the debuggee is stopped right in front of the
INT nn instruction.  Now I want to do a "stepi" in GDB.

> You might look into defining SOFTWARE_SINGLE_STEP for i386

Thanks, I will look it up.

> Are you sure the other platforms can do this?  I bet they just punt.

Well, I usually don't dare to assume that DJGPP solves problems that
other platforms punt ;-).  At least not without sound reasons.
From rogerc@ignitus.com Thu Jul 15 07:51:00 1999
From: Roger Cruz <rogerc@ignitus.com>
To: gdb@sourceware.cygnus.com
Subject: Can't build GDB 4.18 under Cygnus 2.0.1
Date: Thu, 15 Jul 1999 07:51:00 -0000
Message-id: <378DF57E.3D26CA78@ignitus.com>
X-SW-Source: 1999-q3/msg00069.html
Content-length: 779

I have CygWin installed in my WinNT 40 PC.  I downloaded GDB 4.18, unzipped
and untar'ed it and ./configure it without specifying a target or host.  It
correctly configured it for i686-pc-cygwin.  I then proceeded to run "make"
at the top level and it complained that there were no targets to be made.
Looking at the makefile created, I see that this is also true.  Why does
the make file failed to be created with any targets?  The only problem I
saw during the configuration was a warning about an unterminated sed "s"
command.  I couldn't track which sed cmd this was.

Can any one please help?

Thanks
Roger

PS: Please email me directly as I do not subscribe to this list.  Also,
note that you must first remove "_nojunk" from the email address:
rogerc_nojunk@ignitus.com