提交 4e34feb5 编写于 作者: R Ronnie Sahlberg 提交者: Steve French

cifs: only wake the thread for the very last PDU in a compound

For compounded PDUs we whould only wake the waiting thread for the
very last PDU of the compound.
We do this so that we are guaranteed that the demultiplex_thread will
not process or access any of those MIDs any more once the send/recv
thread starts processing.

Else there is a race where at the end of the send/recv processing we
will try to delete all the mids of the compound. If the multiplex
thread still has other mids to process at this point for this compound
this can lead to an oops.

Needed to fix recent commit:
commit 730928c8
("cifs: update smb2_queryfs() to use compounding")
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>
上级 ddf83afb
......@@ -773,6 +773,11 @@ cifs_setup_request(struct cifs_ses *ses, struct smb_rqst *rqst)
return mid;
}
static void
cifs_noop_callback(struct mid_q_entry *mid)
{
}
int
compound_send_recv(const unsigned int xid, struct cifs_ses *ses,
const int flags, const int num_rqst, struct smb_rqst *rqst,
......@@ -827,8 +832,13 @@ compound_send_recv(const unsigned int xid, struct cifs_ses *ses,
}
midQ[i]->mid_state = MID_REQUEST_SUBMITTED;
/*
* We don't invoke the callback compounds unless it is the last
* request.
*/
if (i < num_rqst - 1)
midQ[i]->callback = cifs_noop_callback;
}
cifs_in_send_inc(ses->server);
rc = smb_send_rqst(ses->server, num_rqst, rqst, flags);
cifs_in_send_dec(ses->server);
......@@ -909,6 +919,12 @@ compound_send_recv(const unsigned int xid, struct cifs_ses *ses,
midQ[i]->resp_buf = NULL;
}
out:
/*
* This will dequeue all mids. After this it is important that the
* demultiplex_thread will not process any of these mids any futher.
* This is prevented above by using a noop callback that will not
* wake this thread except for the very last PDU.
*/
for (i = 0; i < num_rqst; i++)
cifs_delete_mid(midQ[i]);
add_credits(ses->server, credits, optype);
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册