[Monetdb-developers] [Monetdb-checkins] MonetDB5/src/modules/kernel batcast.mx, Feb2010, 1.31.2.7, 1.31.2.8

Stefan Manegold Stefan.Manegold at cwi.nl
Sat Mar 20 14:37:25 CET 2010


On Sat, Mar 20, 2010 at 12:53:12PM +0000, Martin Kersten wrote:
> Update of /cvsroot/monetdb/MonetDB5/src/modules/kernel
> In directory sfp-cvsdas-1.v30.ch3.sourceforge.com:/tmp/cvs-serv10537
> 
> Modified Files:
>       Tag: Feb2010
> 	batcast.mx 
> Log Message:
> Backport of possible dangerous double frees. No roll forward needed.

While your revised code looks indeed more efficient to me (as it allows
ATOMfromstr() --- well, at least strFromStr(); I cannot tell about other
"extern" (varsized) atoms --- to reuse a allocated temporary buffer),
I wonder whether/why the original experienced double frees?

Called with v==NULL,
ATOMfromstr() --- again, I refer only to strFromStr(); FromStr functions for
other "extern" (varsized) atoms might behave differently; IMHO, we don't
have any "contract" or "policies" for FromStr() implementations --- would
allocate a buffer for the result,
ATOMput() copies the value from the buffer to the BAT heap,
and then the buffer is (was) freed.

Looks very sound to me.

What am I missing?

I would be thankful for any enlightment.

Stefan

> Index: batcast.mx
> ===================================================================
> RCS file: /cvsroot/monetdb/MonetDB5/src/modules/kernel/batcast.mx,v
> retrieving revision 1.31.2.7
> retrieving revision 1.31.2.8
> diff -u -d -r1.31.2.7 -r1.31.2.8
> --- batcast.mx	18 Feb 2010 01:04:05 -0000	1.31.2.7
> +++ batcast.mx	20 Mar 2010 12:53:10 -0000	1.31.2.8
> @@ -248,6 +248,8 @@
>  	BAT *b,*bn;
>  	BATiter bi;
>  	BUN p, q, r = 0;
> +	ptr v = NULL; 
> +	int len = 0;
>  
>  	@:getBATdescriptor(bid, b, "batcalc. at 1")@
>  	@:voidresultBAT(TYPE_ at 1,"batcalc. at 1")@
> @@ -255,15 +257,12 @@
>  
>  	BATaccessBegin(b, USE_TAIL, MMAP_SEQUENTIAL);
>  	BATloop(b, p, q) {
> -		ptr v = NULL; 
> -		int len = 0;
> -
>  		ATOMfromstr(TYPE_ at 1, &v, &len, (char *) BUNtvar(bi, p));
>  		ATOMput(TYPE_ at 1, bn->T->vheap, Tloc(bn, r), v);
> -		if (ATOMextern(TYPE_ at 1))
> -			GDKfree(v);
>  		r++;
>  	}
> +	if (ATOMextern(TYPE_ at 1))
> +		GDKfree(v);
>  	BATsetcount(bn, BATcount(b));
>  bunins_failed:
>  	BATaccessEnd(b, USE_TAIL, MMAP_SEQUENTIAL);
> 
> 
> ------------------------------------------------------------------------------
> Download Intel® Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Monetdb-checkins mailing list
> Monetdb-checkins at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/monetdb-checkins
> 

-- 
| Dr. Stefan Manegold | mailto:Stefan.Manegold at cwi.nl |
| CWI,  P.O.Box 94079 | http://www.cwi.nl/~manegold/  |
| 1090 GB Amsterdam   | Tel.: +31 (20) 592-4212       |
| The Netherlands     | Fax : +31 (20) 592-4199       |




More information about the developers-list mailing list