YAMAMOTO Takashi <yamt@mwd.biglobe.ne.jp> writes:

> not exactly same.  return value is different when VOP_BMAP fails.
> i don't know if it's important to continue processing async putpages
> even when VOP_BMAP fails, though.

Hmm.  I also wonder if it really makes sense.

> > BTW, is this change results noticable effect?
> 
> no human noticable effects, i think. :-)

I'm interested in the number of i/o issued, espcially on low memery
envronment.

> -			error = GOP_WRITE(vp, pgs, npages, flags);
> +			if (!async && iodesc == NULL) {
> +				/*
> +				 * XXX should use stack or pool? 
> +				 */
> +				MALLOC(iodesc, struct genfs_iodesc *,

As we already doing PHOLD when traversing pages by list, putting it on
stack may be an option.

enami.