On Tue, Feb 26, 2008 at 03:56:36PM +0100, Vladimir Marek wrote:
> N_conreq:entry {
> self->x=1;
>
> calledaddr=(struct xaddrf *)arg3;
> callingaddr=(struct xaddrf *)arg4;
>
> trace(calledaddr->link_id);
> tracemem(calledaddr->DTE_MAC.lsap_add, 80);
>
> trace(callingaddr->link_id);
> tracemem(callingaddr->DTE_MAC.lsap_add, 80);
> }
>
> 0 -> N_conreq 255
> <===== first link_id
> 0 1 2 3 4 5 6 7 8 9 a b c d e f 0123456789abcdef
> <===== first tracemem
> 0: 12 30 00 00 00 00 00 00 00 00 00 00 00 00 00 00 .0..............
> 10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> 20: 00 ff 00 03 45 60 00 00 00 00 00 00 00 00 00 00 ....E`..........
> 30: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> 40: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> 255 E`
> ^ ^-- dtrace being smart here? The memory is 0x45 0x60 0x00 ...
> which is E` in ascii.
> 2nd link_id
>
> Is there a way to really trace 80 bytes of memory?
As is usually the case with DTrace, there _is_ a way, but, as is ocassionally
the case, it's not beautiful. There's a bug here in that DTrace -- as you
suspected -- tries to be a little bit too clever when printing bytes by
assuming that you must want something printed as a string if the first few
bytes are printable.
What you can do is create a buffer to hold your real data, prepend it with
a non-printable character, and trace the whole thing (remembering to mentally
discard the first byte). For example:
syscall::write:entry
/execname == "cat"/
{
this->a = copyin(arg1, 80);
this->p = (char *)alloca(81);
*this->p = '\0';
bcopy(this->a, this->p + 1, 80);
tracemem(this->p, 81);
}
Despite this glorious work-around, please file a bug; if someone is interested
in getting involved with DTrace development, this would be a good fix to try
out.
Adam
--
Adam Leventhal, Fishworks http://blogs.sun.com/ahl
_______________________________________________
dtrace-discuss mailing list
[email protected]
