Jan Beulich
2009-Mar-04 11:56 UTC
[Xen-devel] [PATCH] linux/blkfront: use blk_rq_map_sg to generate ring entries
From: Jens Axboe <jens.axboe@oracle.com> Subject: xen/blkfront: use blk_rq_map_sg to generate ring entries On occasion, the request will apparently have more segments than we fit into the ring. Jens says:> The second problem is that the block layer then appears to create one > too many segments, but from the dump it has rq->nr_phys_segments => BLKIF_MAX_SEGMENTS_PER_REQUEST. I suspect the latter is due to > xen-blkfront not handling the merging on its own. It should check that > the new page doesn''t form part of the previous page. The > rq_for_each_segment() iterates all single bits in the request, not dma > segments. The "easiest" way to do this is to call blk_rq_map_sg() and > then iterate the mapped sg list. That will give you what you are > looking for.> Here''s a test patch, compiles but otherwise untested. I spent more > time figuring out how to enable XEN than to code it up, so YMMV! > Probably the sg list wants to be put inside the ring and only > initialized on allocation, then you can get rid of the sg on stack and > sg_init_table() loop call in the function. I''ll leave that, and the > testing, to you.[Moved sg array into info structure, and initialize once. -J] Signed-off-by: Jens Axboe <jens.axboe@oracle.com> Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> After testing on 2.6.27.19, made apply to 2.6.18 tree. Signed-off-by: Jan Beulich <jbeulich@novell.com> --- sle11-2009-03-04.orig/drivers/xen/blkfront/blkfront.c 2009-03-04 11:38:21.000000000 +0100 +++ sle11-2009-03-04/drivers/xen/blkfront/blkfront.c 2009-03-04 11:38:33.000000000 +0100 @@ -40,6 +40,7 @@ #include <linux/cdrom.h> #include <linux/sched.h> #include <linux/interrupt.h> +#include <linux/scatterlist.h> #include <scsi/scsi.h> #include <xen/evtchn.h> #include <xen/xenbus.h> @@ -232,6 +233,8 @@ static int setup_blkring(struct xenbus_d SHARED_RING_INIT(sring); FRONT_RING_INIT(&info->ring, sring, PAGE_SIZE); + sg_init_table(info->sg, BLKIF_MAX_SEGMENTS_PER_REQUEST); + err = xenbus_grant_ring(dev, virt_to_mfn(info->ring.sring)); if (err < 0) { free_page((unsigned long)sring); @@ -593,13 +596,11 @@ static int blkif_queue_request(struct re struct blkfront_info *info = req->rq_disk->private_data; unsigned long buffer_mfn; blkif_request_t *ring_req; - struct bio *bio; - struct bio_vec *bvec; - int idx; unsigned long id; unsigned int fsect, lsect; - int ref; + int i, ref; grant_ref_t gref_head; + struct scatterlist *sg; if (unlikely(info->connected != BLKIF_STATE_CONNECTED)) return 1; @@ -629,14 +631,13 @@ static int blkif_queue_request(struct re if (blk_barrier_rq(req)) ring_req->operation = BLKIF_OP_WRITE_BARRIER; - ring_req->nr_segments = 0; - rq_for_each_bio (bio, req) { - bio_for_each_segment (bvec, bio, idx) { - BUG_ON(ring_req->nr_segments - == BLKIF_MAX_SEGMENTS_PER_REQUEST); - buffer_mfn = page_to_phys(bvec->bv_page) >> PAGE_SHIFT; - fsect = bvec->bv_offset >> 9; - lsect = fsect + (bvec->bv_len >> 9) - 1; + ring_req->nr_segments = blk_rq_map_sg(req->q, req, info->sg); + BUG_ON(ring_req->nr_segments > BLKIF_MAX_SEGMENTS_PER_REQUEST); + for (i = 0; i < ring_req->nr_segments; ++i) { + sg = info->sg + i; + buffer_mfn = page_to_phys(sg->page) >> PAGE_SHIFT; + fsect = sg->offset >> 9; + lsect = fsect + (sg->length >> 9) - 1; /* install a grant reference. */ ref = gnttab_claim_grant_reference(&gref_head); BUG_ON(ref == -ENOSPC); @@ -646,17 +648,12 @@ static int blkif_queue_request(struct re buffer_mfn, rq_data_dir(req) ? GTF_readonly : 0 ); - info->shadow[id].frame[ring_req->nr_segments] - mfn_to_pfn(buffer_mfn); - - ring_req->seg[ring_req->nr_segments] + info->shadow[id].frame[i] = mfn_to_pfn(buffer_mfn); + ring_req->seg[i] (struct blkif_request_segment) { .gref = ref, .first_sect = fsect, .last_sect = lsect }; - - ring_req->nr_segments++; - } } info->ring.req_prod_pvt++; --- sle11-2009-03-04.orig/drivers/xen/blkfront/block.h 2009-03-04 11:38:21.000000000 +0100 +++ sle11-2009-03-04/drivers/xen/blkfront/block.h 2009-03-04 11:38:33.000000000 +0100 @@ -102,6 +102,7 @@ struct blkfront_info int connected; int ring_ref; blkif_front_ring_t ring; + struct scatterlist sg[BLKIF_MAX_SEGMENTS_PER_REQUEST]; unsigned int irq; struct xlbd_major_info *mi; request_queue_t *rq; _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel