[src/trunk]: src/sys/arch/xen/xen Add back non-512-byte aligned buffers suppo...

[bouyer <bouyer%NetBSD.org@localhost>]

details:   https://anonhg.NetBSD.org/src/rev/a09c309975b6
branches:  trunk
changeset: 365654:a09c309975b6
user:      bouyer <bouyer%NetBSD.org@localhost>
date:      Sun Aug 26 11:31:56 2018 +0000

description:
Add back non-512-byte aligned buffers support (reverting 1.84).
Userland users of raw devices don't have to use sector-aligned buffers,
and e.g. disklabel doesn't.
anita tests on Xen should at last complete the install phase again.

diffstat:

 sys/arch/xen/xen/xbd_xenbus.c |  62 +++++++++++++++++++++++++++++++++++-------
 1 files changed, 51 insertions(+), 11 deletions(-)

diffs (132 lines):

diff -r 303b9f3ae654 -r a09c309975b6 sys/arch/xen/xen/xbd_xenbus.c
--- a/sys/arch/xen/xen/xbd_xenbus.c     Sun Aug 26 08:54:03 2018 +0000
+++ b/sys/arch/xen/xen/xbd_xenbus.c     Sun Aug 26 11:31:56 2018 +0000
@@ -1,4 +1,4 @@
-/*      $NetBSD: xbd_xenbus.c,v 1.86 2018/08/21 18:55:08 jdolecek Exp $      */
+/*      $NetBSD: xbd_xenbus.c,v 1.87 2018/08/26 11:31:56 bouyer Exp $      */
 
 /*
  * Copyright (c) 2006 Manuel Bouyer.
@@ -50,7 +50,7 @@
  */
 
 #include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: xbd_xenbus.c,v 1.86 2018/08/21 18:55:08 jdolecek Exp $");
+__KERNEL_RCSID(0, "$NetBSD: xbd_xenbus.c,v 1.87 2018/08/26 11:31:56 bouyer Exp $");
 
 #include "opt_xen.h"
 
@@ -105,6 +105,7 @@
                grant_ref_t req_gntref[BLKIF_MAX_SEGMENTS_PER_REQUEST];
                int req_nr_segments; /* number of segments in this request */
                struct buf *req_bp; /* buffer associated with this request */
+               void *req_data; /* pointer to the data buffer */
            } req_rw;
            struct {
                int s_error;
@@ -115,6 +116,7 @@
 #define req_gntref     u.req_rw.req_gntref
 #define req_nr_segments        u.req_rw.req_nr_segments
 #define req_bp         u.req_rw.req_bp
+#define req_data       u.req_rw.req_data
 #define req_sync       u.req_sync
 
 struct xbd_xenbus_softc {
@@ -169,6 +171,9 @@
 static void xbd_backend_changed(void *, XenbusState);
 static void xbd_connect(struct xbd_xenbus_softc *);
 
+static int  xbd_map_align(struct xbd_req *);
+static void xbd_unmap_align(struct xbd_req *);
+
 static void xbdminphys(struct buf *);
 
 CFATTACH_DECL3_NEW(xbd, sizeof(struct xbd_xenbus_softc),
@@ -694,7 +699,6 @@
 
                bp = xbdreq->req_bp;
                KASSERT(bp != NULL);
-               xbdreq->req_bp = NULL;
                DPRINTF(("%s(%p): b_bcount = %ld\n", __func__,
                    bp, (long)bp->b_bcount));
 
@@ -704,7 +708,11 @@
                        goto next;
                }
                /* b_resid was set in dk_start */
+               if (__predict_false(
+                   xbdreq->req_data != NULL && bp->b_data != xbdreq->req_data))
+                       xbd_unmap_align(xbdreq);
 next:
+               xbdreq->req_bp = NULL;
                dk_done(&sc->sc_dksc, bp);
 
                SLIST_INSERT_HEAD(&sc->sc_xbdreq_head, xbdreq, req_next);
@@ -971,14 +979,16 @@
                goto out;
        }
 
-       if ((vaddr_t)bp->b_data & (XEN_BSIZE - 1)) {
-               DPRINTF(("xbd_diskstart: no align\n"));
-               error = EINVAL;
-               goto out;
+       xbdreq->req_bp = bp;
+       xbdreq->req_data = bp->b_data;
+       if (__predict_false((vaddr_t)bp->b_data & (XEN_BSIZE - 1))) {
+               if (__predict_false(xbd_map_align(xbdreq) != 0)) {
+                       DPRINTF(("xbd_diskstart: no align\n"));
+                       error = EAGAIN;
+                       goto out;
+               }
        }
 
-       xbdreq->req_bp = bp;
-
        SLIST_REMOVE_HEAD(&sc->sc_xbdreq_head, req_next);
        req = RING_GET_REQUEST(&sc->sc_ring, sc->sc_ring.req_prod_pvt);
        req->id = xbdreq->req_id;
@@ -987,8 +997,8 @@
        req->sector_number = bp->b_rawblkno;
        req->handle = sc->sc_handle;
 
-       va = (vaddr_t)bp->b_data & ~PAGE_MASK;
-       off = (vaddr_t)bp->b_data & PAGE_MASK;
+       va = (vaddr_t)xbdreq->req_data & ~PAGE_MASK;
+       off = (vaddr_t)xbdreq->req_data & PAGE_MASK;
        bcount = bp->b_bcount;
        bp->b_resid = 0;
        for (seg = 0; bcount > 0;) {
@@ -1028,3 +1038,33 @@
 err:
        return error;
 }
+
+static int
+xbd_map_align(struct xbd_req *req)
+{
+       int s = splvm(); /* XXXSMP - bogus? */
+       int rc;
+
+       rc = uvm_km_kmem_alloc(kmem_va_arena,
+           req->req_bp->b_bcount, (VM_NOSLEEP | VM_INSTANTFIT),
+           (vmem_addr_t *)&req->req_data);
+       splx(s);
+       if (__predict_false(rc != 0))
+               return ENOMEM;
+       if ((req->req_bp->b_flags & B_READ) == 0)
+               memcpy(req->req_data, req->req_bp->b_data,
+                   req->req_bp->b_bcount);
+       return 0;
+}
+
+static void
+xbd_unmap_align(struct xbd_req *req)
+{
+       int s;
+       if (req->req_bp->b_flags & B_READ)
+               memcpy(req->req_bp->b_data, req->req_data,
+                   req->req_bp->b_bcount);
+       s = splvm(); /* XXXSMP - bogus? */
+       uvm_km_kmem_free(kmem_va_arena, (vaddr_t)req->req_data, req->req_bp->b_bcount);
+       splx(s);
+}