Sign in

username:

password:



Not a member?

Search fpga-cpu



Search tips

Subscribe to fpga-cpu



fpga-cpu by Keywords

Altera | CISCifying | IDE | ISA | Java | JHDL | JTAG | LBU | MicroBlaze | PAR | PCI | RISC | SoC | Spartan | Transputers | Verilog | VHDL | Virtex | VLIW | WebPack | Xilinx | Xsoc | YARD-1A

Ads

Discussion Groups

Discussion Groups | FPGA-CPU | Hardware for exception handling

This list is for discussion of the design and implementation of field-programmable gate array based processors and integrated systems. It is also for discussion and community support of the XSOC Project (see http://www.fpgacpu.org/xsoc).

Hardware for exception handling - Rob Finch - Mar 10 19:32:00 2003

Can anyone give me a good reference (URL) to information on hardware
support for exception handling as used in a language like Java or C++.

Thanks,
Rob





(You need to be a member of fpga-cpu -- send a blank email to fpga-cpu-subscribe@yahoogroups.com )


Re: Hardware for exception handling - Veronica Merryfield - Mar 11 11:22:00 2003

What an interesting concept. I have not seen anything on hardware support
for expection handling in C++ but I am sure someone on this list will know.
I have never even though about doing it in hardware but I would imagine it
would need a register per level (if you want to support nesting) that holds
a stack address to mark the TRY. Quite what you would do for CATCH I am not
sure.

Looking forward to hearing about this one.
--
Veronica Merryfield, somewhere in Cambridgeshire, UK
"The best things in life aren't things"
www.sheyagreyhoundrescue.com
www.dunrunnin.co.uk

----- Original Message -----
From: "Rob Finch" <>
To: <>
Sent: Tuesday, March 11, 2003 12:32 AM
Subject: [fpga-cpu] Hardware for exception handling > Can anyone give me a good reference (URL) to information on hardware
> support for exception handling as used in a language like Java or C++.
>
> Thanks,
> Rob >
> To post a message, send it to:
> To unsubscribe, send a blank message to:






(You need to be a member of fpga-cpu -- send a blank email to fpga-cpu-subscribe@yahoogroups.com )

Re: Hardware for exception handling - Jonathan Kirwan - Mar 11 11:38:00 2003

On Tue, 11 Mar 2003 16:22:55 -0000, you wrote:

>What an interesting concept. I have not seen anything on hardware support
>for expection handling in C++ but I am sure someone on this list will know.
>I have never even though about doing it in hardware but I would imagine it
>would need a register per level (if you want to support nesting) that holds
>a stack address to mark the TRY. Quite what you would do for CATCH I am not
>sure.
>
>Looking forward to hearing about this one.

I'm not sure how helpful these two references might be (they are
both circa 1995/1996, and may not cover exception handling as
well as anyone might like.) But they *do* cover some of the
details needed for understanding exception handling in C++:

"Inside the Object Model: The Sensible Use of C+"
by David M. Papurt

"Inside the C++ Object Model"
by Stanley B. Lippman

I've read the second one, thoroughly, and have had email
conversations with the author. I don't even have the first one
of those two, so I cannot be certain about it -- yet, the
description appears right.

Anyway, that's my suggestion for Rob.

Jon




(You need to be a member of fpga-cpu -- send a blank email to fpga-cpu-subscribe@yahoogroups.com )

Re: Hardware for exception handling - Rob Finch - Mar 12 22:28:00 2003

> >Looking forward to hearing about this one. > "Inside the Object Model: The Sensible Use of C+"
> by David M. Papurt
>
> "Inside the C++ Object Model"
> by Stanley B. Lippman Do these books give details of how exception handling is implemented
from the compiler writer's point of veiw ? Rob





(You need to be a member of fpga-cpu -- send a blank email to fpga-cpu-subscribe@yahoogroups.com )

Re: Re: Hardware for exception handling - Jonathan Kirwan - Mar 12 23:24:00 2003

On Thu, 13 Mar 2003 03:28:23 -0000, you wrote:

>> >Looking forward to hearing about this one.
>>
>> "Inside the Object Model: The Sensible Use of C+"
>> by David M. Papurt
>>
>> "Inside the C++ Object Model"
>> by Stanley B. Lippman
>
>Do these books give details of how exception handling is implemented
>from the compiler writer's point of veiw ?

I've got the second one and it does, but not well. That was the
reason, in fact, that I started writing to Lippman. He told me
then that he was working on a newer version which will include
that information in more detail, but it doesn't appear to be out
there, as yet. But maybe I screwed up looking, so double check
me on that point.

Anyway, it does provide a great deal of information overall and
much of that information informs you about limitations
circumscribing how exceptions *must* be handled in C++. Also,
he does talk about it, as well. I'm just not willing to call it
satisfactory.

I haven't read Papurt's book, nor do I have it. It was written
before Lippman's, but the book description specifically says
that it addresses exception handling, so perhaps it will be good
in that respect. If you find out that it is, I'll be buying the
book. ;)

Jon




(You need to be a member of fpga-cpu -- send a blank email to fpga-cpu-subscribe@yahoogroups.com )

RE: Re: Hardware for exception handling - Jan Gray - Mar 13 0:12:00 2003

You might find this article interesting -- C++: Under the Hood:
http://msdn.microsoft.com/archive/default.asp?url=/archive/en-us/dnarvc/
html/jangrayhood.asp.

This article is also near the mark:
http://groups.google.com/groups?selm=hM2R2.2498%24b73.139198%40paloalto-
snr1

and this one on Win32 Structured Exception Handling by Matt Pietrek is
instructive. (You've got to admire an article with a section heading
"Into the Inferno"):
http://www.microsoft.com/msj/defaultframe.asp?page=/msj/0197/exception/e
xception.htm For C++, efficient implementation of exception handling is fiendishly
complex. I doubt a hardware approach is helpful, beyond management of
stack frame linkages (which might include exception frame linkages).

An implementation must do sufficient bookkeeping (lots) such that
whenever and wherever a software exception is thrown it can walk up n
stack frames to:

1. determine which try block will handle the exception, a matter of
scanning the stack frame's nested try blocks in order, attempting to
match (under C++'s baroque type conversion rules) the thrown object type
with each successive catch clause type; and

2. during stack unwind (across the intervening function's stack frames),
determine which extant objects with destructors must be destroyed. This
is typically accomplished by consulting read-only tables that map
instruction pointer ranges to a table of objects to destroy, or else by
having a per-function local state variable that is updated as each
object is constructed/destructed.

Jan Gray, Gray Research LLC
(Not speaking for anyone else)




(You need to be a member of fpga-cpu -- send a blank email to fpga-cpu-subscribe@yahoogroups.com )