Re: 4.15.14 crash with iscsi target and dvd

[Ming Lei]

On Mon, Apr 09, 2018 at 09:30:11PM +0000, Bart Van Assche wrote:
> On Sun, 2018-04-08 at 12:02 -0400, Wakko Warner wrote:
> > I finished with git bisect.  Here's the output:
> > 84c8590646d5b35804bac60eb58b145839b5893e is the first bad commit
> > commit 84c8590646d5b35804bac60eb58b145839b5893e
> > Author: Ming Lei <tom.leiming@xxxxxxxxx>
> > Date:   Fri Nov 11 20:05:32 2016 +0800
> > 
> >     target: avoid accessing .bi_vcnt directly
> >     
> >     When the bio is full, bio_add_pc_page() will return zero,
> >     so use this information tell when the bio is full.
> >     
> >     Also replace access to .bi_vcnt for pr_debug() with bio_segments().
> >     
> >     Reviewed-by: Christoph Hellwig <hch@xxxxxx>
> >     Signed-off-by: Ming Lei <tom.leiming@xxxxxxxxx>
> >     Reviewed-by: Sagi Grimberg <sagi@xxxxxxxxxxx>
> >     Signed-off-by: Jens Axboe <axboe@xxxxxx>
> > 
> > :040000 040000 a3ebbb71c52ee4eb8c3be4d033b81179211bf704 de39a328dbd1b18519946b3ad46d9302886e0dd0 M      drivers
> > 
> > I did a diff between HEAD^ and HEAD and manually patched the file from
> > 4.15.14.  It's not an exact revert.  I'm running it now and it's working.
> > I'll do a better test later on.  Here's the patch:
> > 
> > --- a/drivers/target/target_core_pscsi.c	2018-02-04 14:31:31.077316617 -0500
> > +++ b/drivers/target/target_core_pscsi.c	2018-04-08 11:43:49.588641374 -0400
> > @@ -915,7 +915,9 @@
> >  					bio, page, bytes, off);
> >  			pr_debug("PSCSI: bio->bi_vcnt: %d nr_vecs: %d\n",
> >  				bio_segments(bio), nr_vecs);
> > -			if (rc != bytes) {
> > +			if (rc != bytes)
> > +				goto fail;
> > +			if (bio->bi_vcnt > nr_vecs) {
> >  				pr_debug("PSCSI: Reached bio->bi_vcnt max:"
> >  					" %d i: %d bio: %p, allocating another"
> >  					" bio\n", bio->bi_vcnt, i, bio);
> 
> Hello Ming,
> 
> Can you have a look at this? The start of this e-mail thread is available at
> https://www.mail-archive.com/linux-scsi@xxxxxxxxxxxxxxx/msg72574.html.

Sure, thanks for your sharing.

Wakko, could you test the following patch and see if there is any
difference?

--
diff --git a/drivers/target/target_core_pscsi.c b/drivers/target/target_core_pscsi.c
index 0d99b242e82e..6147178f1f37 100644
--- a/drivers/target/target_core_pscsi.c
+++ b/drivers/target/target_core_pscsi.c
@@ -888,7 +888,7 @@ pscsi_map_sg(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents,
 		if (len > 0 && data_len > 0) {
 			bytes = min_t(unsigned int, len, PAGE_SIZE - off);
 			bytes = min(bytes, data_len);
-
+ new_bio:
 			if (!bio) {
 				nr_vecs = min_t(int, BIO_MAX_PAGES, nr_pages);
 				nr_pages -= nr_vecs;
@@ -931,6 +931,7 @@ pscsi_map_sg(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents,
 				 * be allocated with pscsi_get_bio() above.
 				 */
 				bio = NULL;
+				goto new_bio;
 			}
 
 			data_len -= bytes;

-- 
Ming