aboutsummaryrefslogtreecommitdiffstats
path: root/fs
diff options
context:
space:
mode:
authorTrond Myklebust <Trond.Myklebust@netapp.com>2012-08-26 14:44:43 -0400
committerTrond Myklebust <Trond.Myklebust@netapp.com>2012-09-06 11:11:53 -0400
commit1f1ea6c2d9d8c0be9ec56454b05315273b5de8ce (patch)
tree4718cdc2e494f18ab13c2c8f09930953b77c31ff /fs
parent21f498c2f73bd6150d82931f09965826dca0b5f2 (diff)
NFSv4: Fix buffer overflow checking in __nfs4_get_acl_uncached
Pass the checks made by decode_getacl back to __nfs4_get_acl_uncached so that it knows if the acl has been truncated. The current overflow checking is broken, resulting in Oopses on user-triggered nfs4_getfacl calls, and is opaque to the point where several attempts at fixing it have failed. This patch tries to clean up the code in addition to fixing the Oopses by ensuring that the overflow checks are performed in a single place (decode_getacl). If the overflow check failed, we will still be able to report the acl length, but at least we will no longer attempt to cache the acl or copy the truncated contents to user space. Reported-by: Sachin Prabhu <sprabhu@redhat.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com> Tested-by: Sachin Prabhu <sprabhu@redhat.com>
Diffstat (limited to 'fs')
-rw-r--r--fs/nfs/nfs4proc.c31
-rw-r--r--fs/nfs/nfs4xdr.c14
2 files changed, 17 insertions, 28 deletions
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 6b94f2d52532..1e50326d00dd 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -3739,7 +3739,7 @@ static void nfs4_write_cached_acl(struct inode *inode, struct page **pages, size
3739 struct nfs4_cached_acl *acl; 3739 struct nfs4_cached_acl *acl;
3740 size_t buflen = sizeof(*acl) + acl_len; 3740 size_t buflen = sizeof(*acl) + acl_len;
3741 3741
3742 if (pages && buflen <= PAGE_SIZE) { 3742 if (buflen <= PAGE_SIZE) {
3743 acl = kmalloc(buflen, GFP_KERNEL); 3743 acl = kmalloc(buflen, GFP_KERNEL);
3744 if (acl == NULL) 3744 if (acl == NULL)
3745 goto out; 3745 goto out;
@@ -3784,7 +3784,6 @@ static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t bu
3784 }; 3784 };
3785 unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE); 3785 unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE);
3786 int ret = -ENOMEM, i; 3786 int ret = -ENOMEM, i;
3787 size_t acl_len = 0;
3788 3787
3789 /* As long as we're doing a round trip to the server anyway, 3788 /* As long as we're doing a round trip to the server anyway,
3790 * let's be prepared for a page of acl data. */ 3789 * let's be prepared for a page of acl data. */
@@ -3807,11 +3806,6 @@ static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t bu
3807 args.acl_len = npages * PAGE_SIZE; 3806 args.acl_len = npages * PAGE_SIZE;
3808 args.acl_pgbase = 0; 3807 args.acl_pgbase = 0;
3809 3808
3810 /* Let decode_getfacl know not to fail if the ACL data is larger than
3811 * the page we send as a guess */
3812 if (buf == NULL)
3813 res.acl_flags |= NFS4_ACL_LEN_REQUEST;
3814
3815 dprintk("%s buf %p buflen %zu npages %d args.acl_len %zu\n", 3809 dprintk("%s buf %p buflen %zu npages %d args.acl_len %zu\n",
3816 __func__, buf, buflen, npages, args.acl_len); 3810 __func__, buf, buflen, npages, args.acl_len);
3817 ret = nfs4_call_sync(NFS_SERVER(inode)->client, NFS_SERVER(inode), 3811 ret = nfs4_call_sync(NFS_SERVER(inode)->client, NFS_SERVER(inode),
@@ -3819,20 +3813,19 @@ static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t bu
3819 if (ret) 3813 if (ret)
3820 goto out_free; 3814 goto out_free;
3821 3815
3822 acl_len = res.acl_len; 3816 /* Handle the case where the passed-in buffer is too short */
3823 if (acl_len > args.acl_len) 3817 if (res.acl_flags & NFS4_ACL_TRUNC) {
3824 nfs4_write_cached_acl(inode, NULL, 0, acl_len); 3818 /* Did the user only issue a request for the acl length? */
3825 else 3819 if (buf == NULL)
3826 nfs4_write_cached_acl(inode, pages, res.acl_data_offset, 3820 goto out_ok;
3827 acl_len);
3828 if (buf) {
3829 ret = -ERANGE; 3821 ret = -ERANGE;
3830 if (acl_len > buflen) 3822 goto out_free;
3831 goto out_free;
3832 _copy_from_pages(buf, pages, res.acl_data_offset,
3833 acl_len);
3834 } 3823 }
3835 ret = acl_len; 3824 nfs4_write_cached_acl(inode, pages, res.acl_data_offset, res.acl_len);
3825 if (buf)
3826 _copy_from_pages(buf, pages, res.acl_data_offset, res.acl_len);
3827out_ok:
3828 ret = res.acl_len;
3836out_free: 3829out_free:
3837 for (i = 0; i < npages; i++) 3830 for (i = 0; i < npages; i++)
3838 if (pages[i]) 3831 if (pages[i])
diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index 1bfbd67c556d..541e796e6db5 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -5072,18 +5072,14 @@ static int decode_getacl(struct xdr_stream *xdr, struct rpc_rqst *req,
5072 * are stored with the acl data to handle the problem of 5072 * are stored with the acl data to handle the problem of
5073 * variable length bitmaps.*/ 5073 * variable length bitmaps.*/
5074 res->acl_data_offset = xdr_stream_pos(xdr) - pg_offset; 5074 res->acl_data_offset = xdr_stream_pos(xdr) - pg_offset;
5075
5076 /* We ignore &savep and don't do consistency checks on
5077 * the attr length. Let userspace figure it out.... */
5078 res->acl_len = attrlen; 5075 res->acl_len = attrlen;
5079 if (attrlen > (xdr->nwords << 2)) { 5076
5080 if (res->acl_flags & NFS4_ACL_LEN_REQUEST) { 5077 /* Check for receive buffer overflow */
5081 /* getxattr interface called with a NULL buf */ 5078 if (res->acl_len > (xdr->nwords << 2) ||
5082 goto out; 5079 res->acl_len + res->acl_data_offset > xdr->buf->page_len) {
5083 } 5080 res->acl_flags |= NFS4_ACL_TRUNC;
5084 dprintk("NFS: acl reply: attrlen %u > page_len %u\n", 5081 dprintk("NFS: acl reply: attrlen %u > page_len %u\n",
5085 attrlen, xdr->nwords << 2); 5082 attrlen, xdr->nwords << 2);
5086 return -EINVAL;
5087 } 5083 }
5088 } else 5084 } else
5089 status = -EOPNOTSUPP; 5085 status = -EOPNOTSUPP;