Cleanup XDR parsing for LAYOUTGET, GETDEVICEINFO

changes LAYOUTGET and GETDEVICEINFO XDR parsing to:
 - not use vmap, which doesn't work on incoherent archs
 - use xdr_stream parsing for all xdr

Signed-off-by: Weston Andros Adamson <dros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

1 files changed, 122 insertions, 56 deletions

diff --git a/fs/nfs/nfs4filelayoutdev.c b/fs/nfs/nfs4filelayoutdev.c
index 68143c162e3..de5350f2b24 100644
--- a/fs/nfs/nfs4filelayoutdev.c
+++ b/fs/nfs/nfs4filelayoutdev.c
@@ -261,7 +261,7 @@ out:
  * Currently only support ipv4, and one multi-path address.
  */
 static struct nfs4_pnfs_ds *
-decode_and_add_ds(__be32 **pp, struct inode *inode)
+decode_and_add_ds(struct xdr_stream *streamp, struct inode *inode)
 {
         struct nfs4_pnfs_ds *ds = NULL;
         char *buf;
@@ -269,25 +269,34 @@ decode_and_add_ds(__be32 **pp, struct inode *inode)
         u32 ip_addr, port;
         int nlen, rlen, i;
         int tmp[2];
-        __be32 *r_netid, *r_addr, *p = *pp;
+        __be32 *p;
 
         /* r_netid */
+        p = xdr_inline_decode(streamp, 4);
+        if (unlikely(!p))
+                goto out_err;
         nlen = be32_to_cpup(p++);
-        r_netid = p;
-        p += XDR_QUADLEN(nlen);
 
-        /* r_addr */
-        rlen = be32_to_cpup(p++);
-        r_addr = p;
-        p += XDR_QUADLEN(rlen);
-        *pp = p;
+        p = xdr_inline_decode(streamp, nlen);
+        if (unlikely(!p))
+                goto out_err;
 
         /* Check that netid is "tcp" */
-        if (nlen != 3 ||  memcmp((char *)r_netid, "tcp", 3)) {
+        if (nlen != 3 ||  memcmp((char *)p, "tcp", 3)) {
                 dprintk("%s: ERROR: non ipv4 TCP r_netid\n", __func__);
                 goto out_err;
         }
 
+        /* r_addr */
+        p = xdr_inline_decode(streamp, 4);
+        if (unlikely(!p))
+                goto out_err;
+        rlen = be32_to_cpup(p);
+
+        p = xdr_inline_decode(streamp, rlen);
+        if (unlikely(!p))
+                goto out_err;
+
         /* ipv6 length plus port is legal */
         if (rlen > INET6_ADDRSTRLEN + 8) {
                 dprintk("%s: Invalid address, length %d\n", __func__,
@@ -300,7 +309,7 @@ decode_and_add_ds(__be32 **pp, struct inode *inode)
                 goto out_err;
         }
         buf[rlen] = '\0';
-        memcpy(buf, r_addr, rlen);
+        memcpy(buf, p, rlen);
 
         /* replace the port dots with dashes for the in4_pton() delimiter*/
         for (i = 0; i < 2; i++) {
@@ -336,90 +345,154 @@ out_err:
 static struct nfs4_file_layout_dsaddr*
 decode_device(struct inode *ino, struct pnfs_device *pdev)
 {
-        int i, dummy;
+        int i;
         u32 cnt, num;
         u8 *indexp;
-        __be32 *p = (__be32 *)pdev->area, *indicesp;
-        struct nfs4_file_layout_dsaddr *dsaddr;
+        __be32 *p;
+        u8 *stripe_indices;
+        u8 max_stripe_index;
+        struct nfs4_file_layout_dsaddr *dsaddr = NULL;
+        struct xdr_stream stream;
+        struct xdr_buf buf = {
+                .pages = pdev->pages,
+                .page_len = pdev->pglen,
+                .buflen = pdev->pglen,
+                .len = pdev->pglen,
+        };
+        struct page *scratch;
+
+        /* set up xdr stream */
+        scratch = alloc_page(GFP_KERNEL);
+        if (!scratch)
+                goto out_err;
+
+        xdr_init_decode(&stream, &buf, NULL);
+        xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);
 
         /* Get the stripe count (number of stripe index) */
-        cnt = be32_to_cpup(p++);
+        p = xdr_inline_decode(&stream, 4);
+        if (unlikely(!p))
+                goto out_err_free_scratch;
+
+        cnt = be32_to_cpup(p);
         dprintk("%s stripe count  %d\n", __func__, cnt);
         if (cnt > NFS4_PNFS_MAX_STRIPE_CNT) {
                 printk(KERN_WARNING "%s: stripe count %d greater than "
                        "supported maximum %d\n", __func__,
                         cnt, NFS4_PNFS_MAX_STRIPE_CNT);
-                goto out_err;
+                goto out_err_free_scratch;
+        }
+
+        /* read stripe indices */
+        stripe_indices = kcalloc(cnt, sizeof(u8), GFP_KERNEL);
+        if (!stripe_indices)
+                goto out_err_free_scratch;
+
+        p = xdr_inline_decode(&stream, cnt << 2);
+        if (unlikely(!p))
+                goto out_err_free_stripe_indices;
+
+        indexp = &stripe_indices[0];
+        max_stripe_index = 0;
+        for (i = 0; i < cnt; i++) {
+                *indexp = be32_to_cpup(p++);
+                max_stripe_index = max(max_stripe_index, *indexp);
+                indexp++;
         }
 
         /* Check the multipath list count */
-        indicesp = p;
-        p += XDR_QUADLEN(cnt << 2);
-        num = be32_to_cpup(p++);
+        p = xdr_inline_decode(&stream, 4);
+        if (unlikely(!p))
+                goto out_err_free_stripe_indices;
+
+        num = be32_to_cpup(p);
         dprintk("%s ds_num %u\n", __func__, num);
         if (num > NFS4_PNFS_MAX_MULTI_CNT) {
                 printk(KERN_WARNING "%s: multipath count %d greater than "
                         "supported maximum %d\n", __func__,
                         num, NFS4_PNFS_MAX_MULTI_CNT);
-                goto out_err;
+                goto out_err_free_stripe_indices;
         }
+
+        /* validate stripe indices are all < num */
+        if (max_stripe_index >= num) {
+                printk(KERN_WARNING "%s: stripe index %u >= num ds %u\n",
+                        __func__, max_stripe_index, num);
+                goto out_err_free_stripe_indices;
+        }
+
         dsaddr = kzalloc(sizeof(*dsaddr) +
                         (sizeof(struct nfs4_pnfs_ds *) * (num - 1)),
                         GFP_KERNEL);
         if (!dsaddr)
-                goto out_err;
-
-        dsaddr->stripe_indices = kzalloc(sizeof(u8) * cnt, GFP_KERNEL);
-        if (!dsaddr->stripe_indices)
-                goto out_err_free;
+                goto out_err_free_stripe_indices;
 
         dsaddr->stripe_count = cnt;
+        dsaddr->stripe_indices = stripe_indices;
+        stripe_indices = NULL;
         dsaddr->ds_num = num;
 
         memcpy(&dsaddr->deviceid, &pdev->dev_id, sizeof(pdev->dev_id));
 
-        /* Go back an read stripe indices */
-        p = indicesp;
-        indexp = &dsaddr->stripe_indices[0];
-        for (i = 0; i < dsaddr->stripe_count; i++) {
-                *indexp = be32_to_cpup(p++);
-                if (*indexp >= num)
-                        goto out_err_free;
-                indexp++;
-        }
-        /* Skip already read multipath list count */
-        p++;
-
         for (i = 0; i < dsaddr->ds_num; i++) {
                 int j;
+                u32 mp_count;
+
+                p = xdr_inline_decode(&stream, 4);
+                if (unlikely(!p))
+                        goto out_err_free_deviceid;
 
-                dummy = be32_to_cpup(p++); /* multipath count */
-                if (dummy > 1) {
+                mp_count = be32_to_cpup(p); /* multipath count */
+                if (mp_count > 1) {
                         printk(KERN_WARNING
                                "%s: Multipath count %d not supported, "
                                "skipping all greater than 1\n", __func__,
-                                dummy);
+                                mp_count);
                 }
-                for (j = 0; j < dummy; j++) {
+                for (j = 0; j < mp_count; j++) {
                         if (j == 0) {
-                                dsaddr->ds_list[i] = decode_and_add_ds(&p, ino);
+                                dsaddr->ds_list[i] = decode_and_add_ds(&stream,
+                                        ino);
                                 if (dsaddr->ds_list[i] == NULL)
-                                        goto out_err_free;
+                                        goto out_err_free_deviceid;
                         } else {
                                 u32 len;
                                 /* skip extra multipath */
-                                len = be32_to_cpup(p++);
-                                p += XDR_QUADLEN(len);
-                                len = be32_to_cpup(p++);
-                                p += XDR_QUADLEN(len);
-                                continue;
+
+                                /* read len, skip */
+                                p = xdr_inline_decode(&stream, 4);
+                                if (unlikely(!p))
+                                        goto out_err_free_deviceid;
+                                len = be32_to_cpup(p);
+
+                                p = xdr_inline_decode(&stream, len);
+                                if (unlikely(!p))
+                                        goto out_err_free_deviceid;
+
+                                /* read len, skip */
+                                p = xdr_inline_decode(&stream, 4);
+                                if (unlikely(!p))
+                                        goto out_err_free_deviceid;
+                                len = be32_to_cpup(p);
+
+                                p = xdr_inline_decode(&stream, len);
+                                if (unlikely(!p))
+                                        goto out_err_free_deviceid;
                         }
                 }
         }
+
+        __free_page(scratch);
         return dsaddr;
 
-out_err_free:
+out_err_free_deviceid:
         nfs4_fl_free_deviceid(dsaddr);
+        /* stripe_indicies was part of dsaddr */
+        goto out_err_free_scratch;
+out_err_free_stripe_indices:
+        kfree(stripe_indices);
+out_err_free_scratch:
+        __free_page(scratch);
 out_err:
         dprintk("%s ERROR: returning NULL\n", __func__);
         return NULL;
@@ -498,11 +571,6 @@ get_device_info(struct inode *inode, struct nfs4_deviceid *dev_id)
                         goto out_free;
         }
 
-        /* set pdev->area */
-        pdev->area = vmap(pages, max_pages, VM_MAP, PAGE_KERNEL);
-        if (!pdev->area)
-                goto out_free;
-
         memcpy(&pdev->dev_id, dev_id, sizeof(*dev_id));
         pdev->layout_type = LAYOUT_NFSV4_1_FILES;
         pdev->pages = pages;
@@ -521,8 +589,6 @@ get_device_info(struct inode *inode, struct nfs4_deviceid *dev_id)
          */
         dsaddr = decode_and_add_device(inode, pdev);
 out_free:
-        if (pdev->area != NULL)
-                vunmap(pdev->area);
         for (i = 0; i < max_pages; i++)
                 __free_page(pages[i]);
         kfree(pages);