From c87817495b3c5c36dcca9d157e9313b7d3195eed Mon Sep 17 00:00:00 2001
From: Susant Palai <spalai@redhat.com>
Date: Fri, 27 Dec 2019 12:06:19 +0530
Subject: dht: Fix stale-layout and create issue

Problem: With lookup-optimize set to on by default, a client with
stale-layout can create a new file on a wrong subvol. This will lead to
possible duplicate files if two different clients attempt to create the
same file with two different layouts.

Solution: Send in-memory layout to be cross checked at posix before
commiting a "create". In case of a mismatch, sync the client layout with
that of the server and attempt the create fop one more time.

test: Manual, testcase(attached)

fixes: bz#1786679
Change-Id: Ife0941f105113f1c572f4363cbcee65e0dd9bd6a
Signed-off-by: Susant Palai <spalai@redhat.com>
---
 xlators/storage/posix/src/posix-entry-ops.c | 29 ++++++++++-
 xlators/storage/posix/src/posix-helpers.c   | 76 +++++++++++++++++++++++++++++
 xlators/storage/posix/src/posix.h           |  4 ++
 3 files changed, 108 insertions(+), 1 deletion(-)

(limited to 'xlators/storage')

diff --git a/xlators/storage/posix/src/posix-entry-ops.c b/xlators/storage/posix/src/posix-entry-ops.c
index 1f1e05f1dc9..667fec7d62f 100644
--- a/xlators/storage/posix/src/posix-entry-ops.c
+++ b/xlators/storage/posix/src/posix-entry-ops.c
@@ -2145,6 +2145,8 @@ posix_create(call_frame_t *frame, xlator_t *this, loc_t *loc, int32_t flags,
         0,
     };
 
+    dict_t *xdata_rsp = dict_ref(xdata);
+
     DECLARE_OLD_FS_ID_VAR;
 
     VALIDATE_OR_GOTO(frame, out);
@@ -2194,6 +2196,28 @@ posix_create(call_frame_t *frame, xlator_t *this, loc_t *loc, int32_t flags,
         was_present = 0;
     }
 
+    if (!was_present) {
+        if (posix_is_layout_stale(xdata, par_path, this)) {
+            op_ret = -1;
+            op_errno = EIO;
+            if (!xdata_rsp) {
+                xdata_rsp = dict_new();
+                if (!xdata_rsp) {
+                    op_errno = ENOMEM;
+                    goto out;
+                }
+            }
+
+            if (dict_set_int32_sizen(xdata_rsp, GF_PREOP_CHECK_FAILED, 1) ==
+                -1) {
+                gf_msg(this->name, GF_LOG_ERROR, errno, P_MSG_DICT_SET_FAILED,
+                       "setting key %s in dict failed", GF_PREOP_CHECK_FAILED);
+            }
+
+            goto out;
+        }
+    }
+
     if (priv->o_direct)
         _flags |= O_DIRECT;
 
@@ -2313,7 +2337,10 @@ out:
 
     STACK_UNWIND_STRICT(create, frame, op_ret, op_errno, fd,
                         (loc) ? loc->inode : NULL, &stbuf, &preparent,
-                        &postparent, xdata);
+                        &postparent, xdata_rsp);
+
+    if (xdata_rsp)
+        dict_unref(xdata_rsp);
 
     return 0;
 }
diff --git a/xlators/storage/posix/src/posix-helpers.c b/xlators/storage/posix/src/posix-helpers.c
index cbc271481a6..4919cbdf2e9 100644
--- a/xlators/storage/posix/src/posix-helpers.c
+++ b/xlators/storage/posix/src/posix-helpers.c
@@ -3586,3 +3586,79 @@ posix_update_iatt_buf(struct iatt *buf, int fd, char *loc, dict_t *xattr_req)
         }
     }
 }
+
+gf_boolean_t
+posix_is_layout_stale(dict_t *xdata, char *par_path, xlator_t *this)
+{
+    int op_ret = 0;
+    ssize_t size = 0;
+    char value_buf[4096] = {
+        0,
+    };
+    gf_boolean_t have_val = _gf_false;
+    data_t *arg_data = NULL;
+    char *xattr_name = NULL;
+    gf_boolean_t is_stale = _gf_false;
+
+    op_ret = dict_get_str_sizen(xdata, GF_PREOP_PARENT_KEY, &xattr_name);
+    if (xattr_name == NULL) {
+        op_ret = 0;
+        goto out;
+    }
+
+    arg_data = dict_get(xdata, xattr_name);
+    if (!arg_data) {
+        op_ret = 0;
+        goto out;
+    }
+
+    size = sys_lgetxattr(par_path, xattr_name, value_buf,
+                         sizeof(value_buf) - 1);
+
+    if (size >= 0) {
+        have_val = _gf_true;
+    } else {
+        if (errno == ERANGE) {
+            gf_msg(this->name, GF_LOG_INFO, errno, P_MSG_PREOP_CHECK_FAILED,
+                   "getxattr on key (%s) path (%s) failed due to"
+                   " buffer overflow",
+                   xattr_name, par_path);
+            size = sys_lgetxattr(par_path, xattr_name, NULL, 0);
+        }
+        if (size < 0) {
+            op_ret = -1;
+            gf_msg(this->name, GF_LOG_ERROR, errno, P_MSG_PREOP_CHECK_FAILED,
+                   "getxattr on key (%s)  failed, path : %s", xattr_name,
+                   par_path);
+            goto out;
+        }
+    }
+
+    if (!have_val) {
+        size = sys_lgetxattr(par_path, xattr_name, value_buf, size);
+        if (size < 0) {
+            gf_msg(this->name, GF_LOG_ERROR, errno, P_MSG_PREOP_CHECK_FAILED,
+                   "getxattr on key (%s) failed (%s)", xattr_name,
+                   strerror(errno));
+            goto out;
+        }
+    }
+
+    if ((arg_data->len != size) || (memcmp(arg_data->data, value_buf, size))) {
+        gf_msg(this->name, GF_LOG_INFO, EIO, P_MSG_PREOP_CHECK_FAILED,
+               "failing preop as on-disk xattr value differs from argument "
+               "value for key %s",
+               xattr_name);
+        op_ret = -1;
+    }
+
+out:
+    dict_del_sizen(xdata, xattr_name);
+    dict_del_sizen(xdata, GF_PREOP_PARENT_KEY);
+
+    if (op_ret == -1) {
+        is_stale = _gf_true;
+    }
+
+    return is_stale;
+}
diff --git a/xlators/storage/posix/src/posix.h b/xlators/storage/posix/src/posix.h
index ce4e1193639..7893a9b3e35 100644
--- a/xlators/storage/posix/src/posix.h
+++ b/xlators/storage/posix/src/posix.h
@@ -658,4 +658,8 @@ posix_spawn_ctx_janitor_thread(xlator_t *this);
 
 void
 posix_update_iatt_buf(struct iatt *buf, int fd, char *loc, dict_t *xdata);
+
+gf_boolean_t
+posix_is_layout_stale(dict_t *xdata, char *par_path, xlator_t *this);
+
 #endif /* _POSIX_H */
-- 
cgit