Re: [Gcl-devel] SGC debugging

[Camm Maguire]

Greetings!

Aurelien Chanudet <address@hidden> writes:

> The default for the signal raised when a segfault
> occurs on MacOS X is nopass.
> 
> > If you are in an infinite loop, and if our earlier
> > discussion was correct, the segfault signals ar
> > being trapped.
> 
> They are.
> 
> > You should then (or prior) be able to set 'b
> > memprotect_handler', 'c', and wind up in the
> > handler.
> 
> I don't seem to be able to wind up in the handler.
> When 'c'-ing, the program counter isn't increased. Gdb
> seems to hang at the place the segfault occurred. (By
> contrast, things work fine when the program isn't run
> in gdb).
> 

Yuck.  This definitely appears to be a gdb bug on your platform.  If
so, this leaves printf debugging.  I'd first print out all the calls
to memprotect, and fadrr in the handler relative to sgc_type_map, then
go over the page logic to determine which element of sgc_type_map
governs sgc_type_map itself, and make sure it is set appropriately
(PERM_WRITABLE).  If not, then maybe print out the element settings as
they are made.

Take care,

> 
>       
> 
>       
>               
> Créez gratuitement votre Yahoo! Mail avec 100 Mo de stockage !
> Créez votre Yahoo! Mail sur http://fr.benefits.yahoo.com/
> 
> Yahoo! Messenger: dialoguez instantanément avec vos amis. Téléchargez 
> GRATUITEMENT sur http://fr.messenger.yahoo.com
> 
> 
> _______________________________________________
> Gcl-devel mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/gcl-devel
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah