Re: [PATCH] loop.c patches, take two

[Mika Penttilä]

Ben Slusky wrote:

> On Mon, 22 Dec 2003 01:00:39 +0200, Mika Penttil? wrote:
>  
>
> > AFAICS, this code path is never taken. You don't queue block device writes 
> > for the loop thread.
> >    
>
> Yes I do, in loop_end_io_transfer. If we allocated fewer pages for the copy
> bio than contained in the original bio, then those pages are recycled for
> the next write.
>
> @@ -413,7 +411,7 @@ static int loop_end_io_transfer(struct b
> 	if (bio->bi_size)
> 		return 1;
>
> -	if (err || bio_rw(bio) == WRITE) {
> +	if (err || (bio_rw(bio) == WRITE && bio->bi_vcnt == rbh->bi_vcnt)) {
> 		bio_endio(rbh, rbh->bi_size, err);
> 		if (atomic_dec_and_test(&lo->lo_pending))
> 			up(&lo->lo_bh_mutex);
>  
I see, subtle...

--Mika


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/