Re: Remote protocol extension for register ranges

[jtc@redback.com (J.T. Conklin)]

>>>>> "Fernando" == Fernando Nasser <fnasser@redhat.com> writes:
Fernando> Presently, the remote protocol can only read all registers
Fernando> of a target.  This is OK for the context registers (the ones
Fernando> that are saved by the OD on a context switch) but this is a
Fernando> serious limitation for machines that have lots of additional
Fernando> registers.  It is just not feasible to read them all at
Fernando> every "g" packet.

If a target has so many registers that you don't want to read them all
with the "g" command packet, I believe you are going to need something
similar for the "G" command packet.  Otherwise, any read/modify/write
operations GDB does will end up reading and writing the entire set.

Fernando> I would like to add parameters to the "g" packet.  Something
Fernando> similar has been proposed before -- the only difference is
Fernando> that I would like the parameters to be a register number or
Fernando> a register number range range.

You could probably get away with adding parameters to the "g" command
packet --- as far as I can tell none of the sample stubs completely
parse the command.  But if GDB sent "g" command packet containing a
register range to such stubs, it would get the entire register set.
You would have to write a some sort of run time test to determine
whether it handled register ranges.

Adding parameters to the "G" command packet is more troublesome ---
the register values immediately follow the command character.

For these reasons, I recommend implementing a new set of commands
which would deprecate G/g/P/p.

Perhaps something like:

   Get Registers:
        y<start>,<end>

   Set Registers:
        Y<start>,<end>,<register-contents>

But while we are adding commands, we might want to consider whether
there are any other arguments we should add.  

One that comes to mind is a thread-id.  Currently, when we want to get
the registers from several threads on the target, GDB has to issue a 
set thread command before each store/fetch register command.  If the
thread-id was part of the request, it could significantly improve 
remote protocol performance, especially on low-bandwidth and/or high
latency connections.

For amusement purposes only, I've enclosed the store and fetch
sections from the specification I've been writing for a replacement
remote protocol.  Other than the register ranges being specified by
offset and length, it's pretty much the same as the above.

        --jtc


4.1 REGISTER FETCH

input:
	context_t	id;
	int		offset;		offset within register data
	int		length;		length of transfer

output:
	int		status;
	char		data[];

This command may yeild unpredicable results if context <id> has not
been suspended.

issues:
	This assumes a single flat address space for all registers.
	It might be convienent to have separate (but potentially
	overlapping) register files for integer registers, floating
	point registers, system registers, miscellaneous registers,
	etc.  

        For example, one register file could contain the PC, FP, SP,
        and whatever other registers are needed on a particular
	architecture for GDB after a signal/exception/breakpoint
	event.  This would solve the problem of what to return in 
	those events.


4.2 REGISTER STORE

input:
	context_t	id;
	int		offset;		offset within register data
	int		length;		length of transfer
	char		data[];

output:
	int		status;

This command may cause unpredictable behavior if context <id> has not
been suspended.

issues:
	see also register fetch.


4.3 MEMORY FETCH

input:
	addr_t		address;	address of transfer
	int		length;		length of transfer
	mode_t		mode;

output:
	int		status;
	char		data[length];

description:
	Returns the contents of the memory region at <address> through
	<address> + <length>.

This command may yeild unpredictable results if all contexts which
access this memory region have not been suspended (unless some mutual
exclusion mechanism is employed).

issues:
	should this command take a "address space" argument?  This
	might be useful for I/O space and bus spaces.  If so, should
	registers simply be considered a "space".

issues:
	should this command take a context id argument as well?  This
	might be desirable if different contexts can have different
	memory mappings.


4.4 MEMORY STORE

input:
	addr_t		address;	address of transfer
	int		length;		length of transfer
	mode_t		mode;
	char		data[length];

output:
	int		status;

description:
	Writes data[] at <address> through <address> + <length>

This command may cause unpredictable results if all contexts which
access this memory have not been suspended (unless some mutual
exclusion mechanisms is employed).

issues:
	see also memory fetch.


-- 
J.T. Conklin
RedBack Networks
From insulaner_andi@yahoo.com Sat Apr 01 00:00:00 2000
From: Andreas Kuepper <insulaner_andi@yahoo.com>
To: gdb@sourceware.cygnus.com
Subject: building GDB on Cygwin fails
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <20000311184701.12922.qmail@web3401.mail.yahoo.com>
X-SW-Source: 2000-q1/msg00675.html
Content-length: 2739

Hi there,

I am running a Win98 host system and installed Cygwin
successfully. The building of a GCC cross compiler was
also successfully for a DOS based target system using
DJGPP.

Now I want to install GDB on my host(on Cygwin) and
build it for my DJGPP target. By using the latest GDB
(gdb-20000305.tar.bz2) the configure script works
fine. But when I want to build it, I get the following
error message while 'make' is inside the gdb-folder :

bash-2.02$ make
rm -f gdb.exe
gcc -g        -o gdb.exe main.o libgdb.a   
../bfd/libbfd.a ../readline/libreadline.a
../opcodes/libopcodes.a ./../intl/libintl.a
../libiberty/libiberty.a `if test -r
../libtermcap/libtermcap.a; then echo
../libtermcap/libtermcap.a; else echo -ltermcap; fi`  
  -lm  ../libiberty/libiberty.a -luser32 
libgdb.a(stack.o): In function `frame_info':
/djgpp/cross/gdb2000/gdb/../../gdb-20000305/gdb/stack.c:952:
undefined reference to `sigtramp_saved_pc'
libgdb.a(blockframe.o): In function
`file_frame_chain_valid':
/djgpp/cross/gdb2000/gdb/../../gdb-20000305/gdb/blockframe.c:48:
undefined reference to `sigtramp_saved_pc'
libgdb.a(blockframe.o): In function `get_prev_frame':
/djgpp/cross/gdb2000/gdb/../../gdb-20000305/gdb/blockframe.c:479:
undefined reference to `sigtramp_saved_pc'
libgdb.a(blockframe.o): In function
`generic_file_frame_chain_valid':
/djgpp/cross/gdb2000/gdb/../../gdb-20000305/gdb/blockframe.c:1276:
undefined reference to `sigtramp_saved_pc'
/djgpp/cross/gdb2000/gdb/../../gdb-20000305/gdb/blockframe.c:1279:
undefined reference to `sigtramp_saved_pc'
/cygdrive/c/CYGNUS/CYGWIN~1/H-I586~1/BIN/../lib/gcc-lib/i586-cygwin32/2.95.2/../../../libtermcap.a(tputs.o):
In function `tputs':
/home/noer/src/b20/comp-tools/devo/libtermcap/tputs.c:72:
undefined reference to `_ctype_'
/home/noer/src/b20/comp-tools/devo/libtermcap/tputs.c:75:
undefined reference to `_ctype_'
/home/noer/src/b20/comp-tools/devo/libtermcap/tputs.c:80:
undefined reference to `_ctype_'
/cygdrive/c/CYGNUS/CYGWIN~1/H-I586~1/BIN/../lib/gcc-lib/i586-cygwin32/2.95.2/../../../libtermcap.a(termcap.o):
In function `tgetent':
/home/noer/src/b20/comp-tools/devo/libtermcap/termcap.c:251:
undefined reference to `_ctype_'
/cygdrive/c/CYGNUS/CYGWIN~1/H-I586~1/BIN/../lib/gcc-lib/i586-cygwin32/2.95.2/../../../libtermcap.a(getcap.o):
In function `cgetnext':
/home/noer/src/b20/comp-tools/devo/libtermcap/getcap.c:686:
undefined reference to `_ctype_'
collect2: ld returned 1 exit status
make: *** [gdb.exe] Error 1
bash-2.02$ 


I really don't know where to search for the fault
***[gdb.exe] Error

Thank you for your help !!

Andreas Kuepper
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com
From eliz@delorie.com Sat Apr 01 00:00:00 2000
From: Eli Zaretskii <eliz@delorie.com>
To: jtc@redback.com
Cc: gdb@sourceware.cygnus.com
Subject: Re: memory region attribute CLI
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <200003211819.NAA12438@indy.delorie.com>
References: <5mr9dd5dlt.fsf@jtc.redbacknetworks.com> <200003160944.EAA01842@indy.delorie.com> <5mem9avs45.fsf@jtc.redbacknetworks.com> <5m3dplwjri.fsf@jtc.redbacknetworks.com>
X-SW-Source: 2000-q1/msg00760.html
Content-length: 585

> >>>>> "jtc" == J T Conklin <jtc@redback.com> writes:
> jtc> Yes, I guess it has been a while, hasn't it.
> jtc>
> jtc> Start at:
> jtc>         http://sourceware.cygnus.com/ml/gdb/1999-q4/msg00168.html
> 
> No comments?

Sorry, you are right.

I think this is a very useful idea.  But as far as I understand, the
information managed by the code you posted needs to be consulted by
other GDB commands, right?  I don't see any code that does that yet.

I'm not saying that this somehow invalidates your code and
descriptions, I just want to be sure I'm not missing anything
important.
From jsm@cygnus.com Sat Apr 01 00:00:00 2000
From: Jason Molenda <jsm@cygnus.com>
To: Assar Westerlund <assar@sics.se>
Cc: gdb@sourceware.cygnus.com
Subject: Re: anoncvs errors?
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <20000209091801.A1546@cygnus.com>
References: <5l7lgez651.fsf@assaris.sics.se>
X-SW-Source: 2000-q1/msg00192.html
Content-length: 465

On Wed, Feb 09, 2000 at 05:34:50PM +0100, Assar Westerlund wrote:
> Hi, I'm getting this error trying to use anonymous CVS to
> :pserver:anoncvs@anoncvs.cygnus.com:/cvs/gdb.
> 
> Sorry, you don't have read/write access to the history file /cvs/gdb/CVSROOT/history
> Permission denied

The repository has moved, use :pserver:anoncvs@anoncvs.cygnus.com:/cvs/src

(although you shouldn't have gotten an error with the old repo quite yet.
I will look in to it.)

Jason
From jimb@cygnus.com Sat Apr 01 00:00:00 2000
From: Jim Blandy <jimb@cygnus.com>
To: gdb@sourceware.cygnus.com
Subject: [cygnus.gcc] dwarfdump
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <npwvnynq1x.fsf@zwingli.cygnus.com>
X-SW-Source: 2000-q1/msg00363.html
Content-length: 675

This may be of interest to anyone working on GDB and using Dwarf 2
debugging information.

------- Start of forwarded message -------
From: jason@cygnus.com (Jason Merrill)
Newsgroups: cygnus.gcc
Subject: dwarfdump
Date: 21 Feb 2000 11:10:17 -0800
Organization: Cygnus Solutions news/mail gateway
Message-ID: <200002211910.LAA17403.cygnus.gcc@yorick.cygnus.com>
To: gcc-local@cygnus.com

FWIW, the SGI dwarf2 printing utility, dwarfdump, is available in

  http://reality.sgi.com/davea/libdwarf1999Dec14.tar.gz

I've built it for Solaris (~jason/.bin/sparc-sun-solaris2.5/dwarfdump) and
linux (/usr/unsupported/bin/dwarfdump).

Jason
------- End of forwarded message -------
From eliz@delorie.com Sat Apr 01 00:00:00 2000
From: Eli Zaretskii <eliz@delorie.com>
To: Andrew Cagney <ac131313@cygnus.com>
Cc: gdb@sourceware.cygnus.com
Subject: Re: Moving Linux-specific stuff out of i386-tdep.c
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <200003091349.IAA19958@indy.delorie.com>
References: <200003082121.e28LLRu05681@delius.kettenis.local> <1000308222742.ZM8876@ocotillo.lan> <38C7AA6D.5465C647@cygnus.com>
X-SW-Source: 2000-q1/msg00644.html
Content-length: 107

> The 8.3 uniqueness rule definitly still applies.

Should I post a list of files which violate that rule?
From eliz@delorie.com Sat Apr 01 00:00:00 2000
From: Eli Zaretskii <eliz@delorie.com>
To: jimb@cygnus.com
Cc: hjl@lucon.org, gdb@sourceware.cygnus.com, gdb-patches@sourceware.cygnus.com
Subject: Re: Problems with hardware watchpoint on ia32.
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <200003312323.SAA07185@indy.delorie.com>
References: <20000307132401.A20282@valinux.com> <200003081008.FAA16481@indy.delorie.com> <20000308084304.A3150@lucon.org> <200003091210.HAA19857@indy.delorie.com> <npya7c6zn7.fsf@zwingli.cygnus.com> <200003221806.NAA14225@indy.delorie.com> <npbt4040t3.fsf@zwingli.cygnus.com>
X-SW-Source: 2000-q1/msg00849.html
Content-length: 905

> > I seem to be unable to reproduce the problem, at least in a C program:
> > whenever I say "watch foo == bar" (where foo and bar are structs), GDB
> > curses thusly:
> > 
> > 	Structure has no component named operator==.
> > 
> > Am I missing something?
> 
> As Michael points out, structure comparison is broken.  But that
> doesn't change my original point --- that it's perfectly legitimate to
> have `struct value' objects representing structures, which should be
> watched in their entirety.

The current sources do support watching an entire struct, the patches
I sent (and Michael commited) have special provisions for this case.

I asked for an example because I want to look into the problem you
were describing and fix it.  Do you have any other case where this
problem could happen?

If every instance of such a situation is equally broken, then the
users won't miss this functionality ;-).
From eliz@delorie.com Sat Apr 01 00:00:00 2000
From: Eli Zaretskii <eliz@delorie.com>
To: Andrew Cagney <ac131313@cygnus.com>
Cc: gdb@sourceware.cygnus.com
Subject: Re: annotate.texi
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <200003091254.HAA19889@indy.delorie.com>
References: <38C74298.B562FDD1@cygnus.com> <200003070832.DAA14451@indy.delorie.com>
X-SW-Source: 2000-q1/msg00640.html
Content-length: 1098

> > Is there any reason why annotate.texi shouldn't be @include'd by
> > gdb.texinfo and be part of the manual?  Right now, "set annotate" is
> > not documented at all, and annotate.texi seems to be just what the
> > doctor ordered...
>
> I just wonder if we want to encourage its use :-)

Any user who types "gdb --help" will see the --annotate switch
advertised, and will want to know what that does.  Without
annotate.texi in the manual, this switch cannot be documented
properly.  The same goes for the "set annotate" command.  I think
undocumented features should be generally avoided.

Here's a case in point.  I'm going through the test suite trying to
make sure the DJGPP version passes all the tests.  When I bumped into
the test which tests annotations, I wanted to know what does that
command do.  Naturally, I looked it up in gdb.info, and when that
failed in gdbint.info.  You can guess what I found ;-).

In other words, undocumented features in the long run haunt
maintainers as well.

If we want to discourage use of --annotate, we can say that explicitly
in the manual.

Comments?
From fnasser@redhat.com Sat Apr 01 00:00:00 2000
From: Fernando Nasser <fnasser@redhat.com>
To: Andrew Cagney <ac131313@cygnus.com>
Cc: Grant Edwards <grante@visi.com>, Fernando Nasser <fnasser@cygnus.com>, gdb@sourceware.cygnus.com
Subject: Re: RDI target broken in 000215 snapshot
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <38B69A2A.ED2DC6F3@redhat.com>
References: <20000221104541.A28578@visi.com> <38B2AD14.7B0B4A4E@redhat.com> <20000224124726.A663@visi.com> <38B58292.3B11D622@cygnus.com> <20000224134607.A6354@visi.com> <38B59CE1.4CFA7F1E@cygnus.com> <20000224152638.A2092@visi.com> <38B61CF6.B4F80802@cygnus.com>
X-SW-Source: 2000-q1/msg00418.html
Content-length: 962

Andrew Cagney wrote:
> 
> It's more a question of should the endianess have changed between
> releases or should it have been little-endian anyway.  Not my problem
> :-)
> 

I inherited this one :-)

But you are right, this will at least require an entry in the NEWS, as
it alters some previous behavior.  *If* the arm users in the list decide
that the default should be little-endian.  It does seem reasonable and
if the file command straights it out, only the load users would have to
care (but they are more aware of these issues, and can always have the
set endian on their .gdbinit).

It is a shame that Angel is so buggy.  The return code should
appropriately indicate the endianess of the target  --  it would be a
nice feature to use :-(

-- 
Fernando Nasser
Red Hat, Inc. - Toronto                 E-Mail:  fnasser@redhat.com
2323 Yonge Street, Suite #300           Tel:  416-482-2661 ext. 311
Toronto, Ontario   M4P 2C9              Fax:  416-482-6299
From leisner@rochester.rr.com Sat Apr 01 00:00:00 2000
From: "Marty Leisner" <leisner@rochester.rr.com>
To: shebs@apple.com
Cc: gcc@gcc.gnu.org, gdb@sourceware.cygnus.com, leisner@rochester.rr.com
Subject: Re: Should GCC tell GDB about its optimizations? 
Date: Sat, 01 Apr 2000 00:00:00 -0000
Message-id: <200003040446.XAA04330@rochester.rr.com>
References: <38C051C3.260D666B@apple.com>
X-SW-Source: 2000-q1/msg00533.html
Content-length: 3255

Yes...!!!

I proposed something like this years ago...I almost hacked up
gcc to put this in (I think I started...)

It is incredibly useful in the real world to encapsulate what
compiler flags were used in the binary (essentially the gcc command
line with preprocessor flags).

I just compiled a trivial C program with -O3 to see..

The .s file has
  .ident  "GCC: (GNU) 2.95.2 19991024 (release)"

Is there a way for nm to decode this?  (strings can show it, 
but its pretty meaningless in the binary).

For example:
...
GCC: (GNU) egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)
GCC: (GNU) egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)
GCC: (GNU) 2.95.2 19991024 (release)
GCC: (GNU) 2.95.2 19991024 (release)
GCC: (GNU) 2.95.2 19991024 (release)
GCC: (GNU) egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)
....

I don't know how to take apart elf binaries off the top
of my head, but I assume there is information someplace encoded
to match this up with specific files.

gdb doesn't have to warn, but you should be able to use a tool
to figure this out.

Marty Leisner


Stan Shebs <shebs@apple.com> writes  on Fri, 03 Mar 2000 15:58:59 PST
     > I'm sure this subject has come up before, but I don't remember any
     > consensus...
     > 
     > In the process of finishing Mac OS X, Apple has hundreds of engineers
     > beating on GCC and GDB every day.  The system has literally hundreds
     > of subcomponents; everything from the kernel to speech output to Hebrew
     > font support (sort of like a Linux system preloaded with everything
     > you can find on freshmeat :-) ).  By default, everything is built with
     > optimization turned on.  While an engineer can turn off optimization
     > for a specific bit of code, it's not practical to build an entire system
     > with optimization turned off.  So in practice any single program 
consists
     > of a combination of optimized and unoptimized code.
     > 
     > Ideally of course, GCC would issue lots of amazingly detailed debug 
info,
     > and GDB would use it to reconstruct and report program state just as the
     > programmer expects to see it.  But today, the result is just lame; 
hackers
     > trying to debug get lots of squirrelly behavior from GDB.  The problem 
is
     > that they don't know whether the randomness is due to bugs in the 
program,
     > or to the effect of the optimizer.  So the suggestion came up to have 
GCC
     > issue debug info stating what optimizations have been applied to a file,
     > and to have GDB report that information per-function, so that users 
could
     > lower their expectations appropriately.
     > 
     > Although my first reaction was to say "bleah", upon reflection I 
wondered
     > if I was too easily dismissing the concerns of real users of the tools.
     > This sort of thing routinely confuses users; even with years of GNU
     > experience, I still find myself being caught by surprise because I've
     > forgotten that part of the code was optimized.  A simple warning from 
the
     > debugger would have saved me some time and trouble.
     
     > If there's consensus that this would be a worthwhile addition, I'll
     > work up a specific design and publish it.
     > 
     > Stan