[Dwarf-discuss] dwarf - debug line number information program state machine registers
Ramanaiah Nalamothu
ramana.venkat83@gmail.com
Mon May 14 04:13:27 GMT 2007
So, basically an end_sequence tells the end of sequential instructions. am I
wrong??
In that case, the basic_block, tells start of new sequence, and 'memory',
tells the address, pair is enough to represent what end_sequence tells.
Again i am repeating, why we need or in what exact scenario we need to use
this end_sequence???
Sorry, i am not convinced with the above answers/explanation. Please help
me, i need this urgent.
Thanks for your replies.
Regards,
Ramana
On 5/11/07, Michael Eager <eager at eagercon.com> wrote:
>
> Daniel Jacobowitz wrote:
> > On Fri, May 11, 2007 at 11:28:13AM +0530, Ramanaiah Nalamothu wrote:
> >> The immediate instruction following .LM2 and .LM3 sets start of the
> basic
> >> block i.e. "basic_block". And that instructions' address flags end of
> >> sequence i.e. "end_sequence".
> >
> > That's not what end_sequence means. It should be used for each
> > segment of the line number table that is not contiguous. Suppose you
> > have a file containing one instruction; it should have a line table
> > entry for that instruction, advance to after the instruction and have
> > an end of sequence. This ensures that random memory after that object
> > file in memory will not be associated with the same line.
>
> For another example, switch statements can be compiled into a jump table:
>
> ld r1,value ; switch (value) {
> shl r1,2
> ld r1,tbl(r1)
> br r1
> tbl .word target0
> .word target1
> .word target2
> ...
>
> There would be a end_sequence after the br instruction,
> indicating that this is the end of the line with switch.
>
>
> --
> Michael Eager eager at eagercon.com
> 1960 Park Blvd., Palo Alto, CA 94306 650-325-8077
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.dwarfstd.org/private.cgi/dwarf-discuss-dwarfstd.org/attachments/20070514/a08159b8/attachment.htm>
More information about the Dwarf-discuss
mailing list