[LU-8811] mdc_close() may return a freed request in *request

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.10.0
• Affects Version/s: Lustre 2.9.0
• Labels: mdc
• Severity: 3
• Assignee: John Hammond
• Reporter: John Hammond

Description

In mdc_close() if ptlrpc_request_pack() fails then a freed request is returned in *request.

static int mdc_close(struct obd_export *exp, struct md_op_data *op_data,
                     struct md_open_data *mod, struct ptlrpc_request **request)
{
...
        *request = NULL;
        if (OBD_FAIL_CHECK(OBD_FAIL_MDC_CLOSE))
                req = NULL;
        else
                req = ptlrpc_request_alloc(class_exp2cliimp(exp), req_fmt);
...
        rc = ptlrpc_request_pack(req, LUSTRE_MDS_VERSION, MDS_CLOSE);
        if (rc) {
                ptlrpc_request_free(req);
                GOTO(out, rc);
        }
...
out:
        if (mod) {
                if (rc != 0)
                        mod->mod_close_req = NULL;
                /* Since now, mod is accessed through open_req only,       
                 * thus close req does not keep a reference on mod anymore. */
                obd_mod_put(mod);
        }
        *request = req;

        RETURN(rc < 0 ? rc : saved_rc);
}

This will result in a use after free in ll_close_inode_openhandle() or ll_open_cleanup().

This was introduced by

commit 62713a8530349a75b7202c9bfc6be121409a0203
Author: Alexander Boyko <[e-mail]>
Date:   Thu Dec 3 09:57:36 2015 +0300

    LU-5282 mdc: fix panic at mdc_free_open()
    
    Assertion was happened for open request when rq_replay is set
    to 1.
        ASSERTION(mod->mod_open_req->rq_replay == 0)
    But this situation is not fatal for client, and could happened
    when mdc_close() failed.
    The fix allow to free such requests. If mdc_close fail, MDS doesn`t
    receive close request from client. And in a worst case client would
    be evicted.
    
    The test recreates issue when mdc_close failed and
    client asserts:
       ASSERTION( mod->mod_open_req->rq_replay == 0 ) failed
    
    Signed-off-by: Alexander Boyko <[e-mail]>
    Seagate-bug-id: MRP-3156
    Change-Id: I5f98901f633355849fc107149eadc9cf171819af
    Reviewed-on: http://review.whamcloud.com/17495
    Reviewed-by: Alex Zhuravlev <[e-mail]>
    Tested-by: Jenkins
    Tested-by: Maloo <[e-mail]>
    Reviewed-by: Andreas Dilger <[e-mail]>
    Reviewed-by: Oleg Drokin <[e-mail]>

Issue Links

• is related to: LU-8582 Interop: master<->b2_8 - sanity test_255a: test failed to respond and timed out