Messages in this thread Patch in this message | | | Date | Fri, 21 Mar 2003 17:52:06 +0000 | From | Richard Curnow <> | Subject | struct nfs_fattr alignment problem in nfs3proc.c |
| |
In the nfs3_proc_unlink_setup function, there appears to be a bug with the way kmalloc is used to allocate storage for two structs grouped together. The second struct ends up with a non 8-byte aligned pointer, which can cause trouble later (in xdr_decode_fattr) when stores occur to the u64 fields inside it. The following patch (on 2.4.19) fixes this problem, though I'm not sure if it's the cleanest fix. (I hit this when working on the port to SH-5, which is currently baselined on 2.4.19).
--- ../linux-2.4/fs/nfs/nfs3proc.c Fri Jan 3 15:01:17 2003 +++ fs/nfs/nfs3proc.c Fri Mar 21 17:42:38 2003 @@ -294,25 +294,28 @@ nfs_refresh_inode(dir, &dir_attr); dprintk("NFS reply remove: %d\n", status); return status; } static int nfs3_proc_unlink_setup(struct rpc_message *msg, struct dentry *dir, struct qstr *name) { struct nfs3_diropargs *arg; struct nfs_fattr *res; + int arg_size_8, res_size_8; - arg = (struct nfs3_diropargs *)kmalloc(sizeof(*arg)+sizeof(*res), GFP_KERNEL); + arg_size_8 = (sizeof(*arg) + 7) & ~7; + res_size_8 = (sizeof(*res) + 7) & ~7; + arg = (struct nfs3_diropargs *)kmalloc(arg_size_8 + res_size_8, GFP_KERNEL); if (!arg) return -ENOMEM; - res = (struct nfs_fattr*)(arg + 1); + res = (struct nfs_fattr*)((char *)arg + arg_size_8); arg->fh = NFS_FH(dir->d_inode); arg->name = name->name; arg->len = name->len; res->valid = 0; msg->rpc_proc = NFS3PROC_REMOVE; msg->rpc_argp = arg; msg->rpc_resp = res; return 0; } -- Richard \\\ SuperH Core+Debug Architect /// .. At home .. P. /// richard.curnow@superh.com /// rc@rc0.org.uk Curnow \\\ http://www.superh.com/ /// www.rc0.org.uk - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |