Union the blkif_request request specific fields]
Subsystems (3)
PATCH 1/2 Union the blkif_request request specific fields]
Jens,
Are you OK with this patch? I was thinking to push it for 2.6.39.
I've it in my tree, but if you would rather carry it for the merge window it is ingit://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.git stable/blkfront-cleanup
commit cd50f830627fd8ec0d97a73c970ccec9811c6bae
Author: Owen Smith <[email protected]> Date: Wed Dec 22 15:05:00 2010 +0000
xen: Union the blkif_request request specific fields
Prepare for extending the block device ring to allow request specific fields, by moving the request specific fields for reads, writes and barrier requests to a union member.
drivers/block/xen-blkfront.c +4 -4
@@ -281,7 +281,7 @@ static int blkif_queue_request(struct request *req) info->shadow[id].request = req; ring_req->id = id; - ring_req->sector_number = (blkif_sector_t )blk_rq_pos (req); + ring_req->u.rw.sector_number = (blkif_sector_t )blk_rq_pos (req); ring_req->handle = info->handle; ring_req->operation = rq_data_dir(req) ?
@@ -317,7 +317,7 @@ static int blkif_queue_request(struct request *req) rq_data_dir(req) ); info->shadow[id].frame[i] =mfn_to_pfn (buffer_mfn); - ring_req->seg[i] = + ring_req->u.rw.seg[i] = (struct blkif_request_segment ) { .gref = ref, .first_sect = fsect,
@@ -615,7 +615,7 @@ static void blkif_completion(struct blk_shadow *s) {int i;for (i =0 ; i < s->req.nr_segments; i++) -gnttab_end_foreign_access (s->req.seg[i].gref,0 ,0UL ); +gnttab_end_foreign_access (s->req.u.rw.seg[i].gref,0 ,0UL ); }static irqreturn_t blkif_interrupt (int irq,void *dev_id)
@@ -932,7 +932,7 @@ static int blkif_recover(struct blkfront_info *info)/* Rewrite any grant references invalidated by susp/resume. */ for (j =0 ; j < req->nr_segments; j++)gnttab_grant_foreign_access_ref ( - req->seg[j].gref, + req->u.rw.seg[j].gref, info->xbdev->otherend_id,pfn_to_mfn (info->shadow[req->id].frame[j]),rq_data_dir (info->shadow[req->id].request));
include/xen/interface/io/blkif.h +11 -5
@@ -51,11 +51,7 @@ typedef uint64_t blkif_sector_t;*/ #define BLKIF_MAX_SEGMENTS_PER_REQUEST 11 -struct blkif_request { -uint8_t operation;/* BLKIF_OP_??? */ -uint8_t nr_segments;/* number of segments */ -blkif_vdev_t handle;/* only for read/write requests */ -uint64_t id;/* private guest value, echoed in resp */ +struct blkif_request_rw {blkif_sector_t sector_number;/* start sector idx on disk (r/w only) */ struct blkif_request_segment {grant_ref_t gref;/* reference to I/O buffer frame */
@@ -65,6 +61,16 @@ struct blkif_request { } seg[BLKIF_MAX_SEGMENTS_PER_REQUEST ]; }; +struct blkif_request { +uint8_t operation;/* BLKIF_OP_??? */ +uint8_t nr_segments;/* number of segments */ +blkif_vdev_t handle;/* only for read/write requests */ +uint64_t id;/* private guest value, echoed in resp */ +union { +struct blkif_request_rw rw; + } u; +}; +struct blkif_response {uint64_t id;/* copied from request */ uint8_t operation;/* copied from request */
Quoted 6 lines · Konrad Rzeszutek Wilk
Jens,
Are you OK with this patch? I was thinking to push it for 2.6.39.
I've it in my tree, but if you would rather carry it for the merge window it is ingit://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.gitstable/blkfront-cleanup
I'd prefer block drivers going through the block tree, but if you have it queued and since it's completely local to that driver, go ahead and keep it in your tree. No worries on my side :-)
You can add my acked-by, if you want.
Signature · 1 line
Jens Axboe
Quoted 11 lines · Jens Axboe
Quoted 6 lines · Konrad Rzeszutek Wilk
Jens,
Are you OK with this patch? I was thinking to push it for 2.6.39.
I've it in my tree, but if you would rather carry it for the merge window it is ingit://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.gitstable/blkfront-cleanupI'd prefer block drivers going through the block tree, but if you have it queued and since it's completely local to that driver, go ahead and keep it in your tree. No worries on my side :-)
OK. Thank you.
Quoted 2 lines
You can add my acked-by, if you want.
Will do.