| Message ID | 20200424123528.17627-2-houpu@bytedance.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Headers | show |
| Series | iscsi-target: fix login error when receiving is too fast | expand |
On 4/24/20 7:35 AM, Hou Pu wrote: > iscsi_target_sk_data_ready() could be invoked indirectly > by iscsi_target_do_login_rx() from workqueue like following: > > iscsi_target_do_login_rx() > iscsi_target_do_login() > iscsi_target_do_tx_login_io() > iscsit_put_login_tx() > iscsi_login_tx_data() > tx_data() > sock_sendmsg_nosec() > tcp_sendmsg() > release_sock() > sk_backlog_rcv() > tcp_v4_do_rcv() > tcp_data_ready() > iscsi_target_sk_data_ready() > > At that time LOGIN_FLAGS_READ_ACTIVE is not cleared. > and iscsi_target_sk_data_ready will not read data > from the socket. And some iscsi initiator(i.e. libiscsi) > will wait forever for a reply. > > LOGIN_FLAGS_READ_ACTIVE should be cleared early just after > doing the receive and before writing to the socket in > iscsi_target_do_login_rx. > > But sad news is that LOGIN_FLAGS_READ_ACTIVE is also used > by sk_state_change to do login cleanup if a socket was closed > at login time. It is supposed to be cleared after the login > pdu is successfully processed and replied. > > So introduce another flag LOGIN_FLAGS_WRITE_ACTIVE to cover > the transmit part so that sk_state_change could work well > and clear LOGIN_FLAGS_READ_ACTIVE early so that sk_data_ready > could also work. > > While at here, use login_flags more efficient. > > Signed-off-by: Hou Pu <houpu@bytedance.com> > --- > drivers/target/iscsi/iscsi_target_nego.c | 30 ++++++++++++++++++++++++++---- > include/target/iscsi/iscsi_target_core.h | 9 +++++---- > 2 files changed, 31 insertions(+), 8 deletions(-) > > diff --git a/drivers/target/iscsi/iscsi_target_nego.c b/drivers/target/iscsi/iscsi_target_nego.c > index 685d771b51d4..4cfa742e61df 100644 > --- a/drivers/target/iscsi/iscsi_target_nego.c > +++ b/drivers/target/iscsi/iscsi_target_nego.c > @@ -622,6 +622,26 @@ static void iscsi_target_do_login_rx(struct work_struct *work) > if (rc < 0) > goto err; > > + /* > + * LOGIN_FLAGS_READ_ACTIVE is cleared so that sk_data_ready > + * could be trigger again after this. > + * > + * LOGIN_FLAGS_WRITE_ACTIVE is cleared after we successfully > + * process a login pdu, so that sk_state_chage could do login > + * cleanup as needed if the socket is closed. If a delayed work is > + * ongoing (LOGIN_FLAGS_WRITE_ACTIVE or LOGIN_FLAGS_READ_ACTIVE), > + * sk_state_change will leave the cleanup to the delayed work or > + * it will schedule a delayed work to do cleanup. > + */ > + if (conn->sock) { > + struct sock *sk = conn->sock->sk; > + > + write_lock_bh(&sk->sk_callback_lock); > + clear_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags); > + set_bit(LOGIN_FLAGS_WRITE_ACTIVE, &conn->login_flags); > + write_unlock_bh(&sk->sk_callback_lock); > + } > + Hey, I had one more question. With the above code, I think we have a race where if we clear the bit above early and iscsi_target_sk_data_ready runs while iscsi_target_do_login_rx is still running then we could queue the work an extra time and get stuck. Because the bit is now not set, if iscsi_target_sk_data_ready runs it will end up calling schedule_delayed_work which will queue up the work again since the work is running and not pending. If that is correct and we hit the race what happens if this was the last login pdu, and we are supposed to go to full feature phase next? For example if iscsi_target_do_login_rx runs an extra time, will we end up stuck waiting in iscsi_target_do_login_rx's call to: rc = conn->conn_transport->iscsit_get_login_rx(conn, login); ? > pr_debug("iscsi_target_do_login_rx after rx_login_io, %p, %s:%d\n", > conn, current->comm, current->pid); > > @@ -629,7 +649,8 @@ static void iscsi_target_do_login_rx(struct work_struct *work) > if (rc < 0) { > goto err; > } else if (!rc) { > - if (iscsi_target_sk_check_and_clear(conn, LOGIN_FLAGS_READ_ACTIVE)) > + if (iscsi_target_sk_check_and_clear(conn, > + LOGIN_FLAGS_WRITE_ACTIVE)) > goto err; > } else if (rc == 1) { > iscsi_target_nego_release(conn); > @@ -670,9 +691,10 @@ static void iscsi_target_sk_state_change(struct sock *sk) > state = __iscsi_target_sk_check_close(sk); > pr_debug("__iscsi_target_sk_close_change: state: %d\n", state); > > - if (test_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags)) { > - pr_debug("Got LOGIN_FLAGS_READ_ACTIVE=1 sk_state_change" > - " conn: %p\n", conn); > + if (test_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags) || > + test_bit(LOGIN_FLAGS_WRITE_ACTIVE, &conn->login_flags)) { > + pr_debug("Got LOGIN_FLAGS_{READ|WRITE}_ACTIVE=1" > + " sk_state_change conn: %p\n", conn); > if (state) > set_bit(LOGIN_FLAGS_CLOSED, &conn->login_flags); > write_unlock_bh(&sk->sk_callback_lock); > diff --git a/include/target/iscsi/iscsi_target_core.h b/include/target/iscsi/iscsi_target_core.h > index 591cd9e4692c..844bef255e89 100644 > --- a/include/target/iscsi/iscsi_target_core.h > +++ b/include/target/iscsi/iscsi_target_core.h > @@ -566,10 +566,11 @@ struct iscsi_conn { > struct socket *sock; > void (*orig_data_ready)(struct sock *); > void (*orig_state_change)(struct sock *); > -#define LOGIN_FLAGS_READ_ACTIVE 1 > -#define LOGIN_FLAGS_CLOSED 2 > -#define LOGIN_FLAGS_READY 4 > -#define LOGIN_FLAGS_INITIAL_PDU 8 > +#define LOGIN_FLAGS_READY 0 > +#define LOGIN_FLAGS_INITIAL_PDU 1 > +#define LOGIN_FLAGS_READ_ACTIVE 2 > +#define LOGIN_FLAGS_WRITE_ACTIVE 3 > +#define LOGIN_FLAGS_CLOSED 4 > unsigned long login_flags; > struct delayed_work login_work; > struct iscsi_login *login; >
On 4/24/20 12:01 PM, Mike Christie wrote: > On 4/24/20 7:35 AM, Hou Pu wrote: >> iscsi_target_sk_data_ready() could be invoked indirectly >> by iscsi_target_do_login_rx() from workqueue like following: >> >> iscsi_target_do_login_rx() >> iscsi_target_do_login() >> iscsi_target_do_tx_login_io() >> iscsit_put_login_tx() >> iscsi_login_tx_data() >> tx_data() >> sock_sendmsg_nosec() >> tcp_sendmsg() >> release_sock() >> sk_backlog_rcv() >> tcp_v4_do_rcv() >> tcp_data_ready() >> iscsi_target_sk_data_ready() >> >> At that time LOGIN_FLAGS_READ_ACTIVE is not cleared. >> and iscsi_target_sk_data_ready will not read data >> from the socket. And some iscsi initiator(i.e. libiscsi) >> will wait forever for a reply. >> >> LOGIN_FLAGS_READ_ACTIVE should be cleared early just after >> doing the receive and before writing to the socket in >> iscsi_target_do_login_rx. >> >> But sad news is that LOGIN_FLAGS_READ_ACTIVE is also used >> by sk_state_change to do login cleanup if a socket was closed >> at login time. It is supposed to be cleared after the login >> pdu is successfully processed and replied. >> >> So introduce another flag LOGIN_FLAGS_WRITE_ACTIVE to cover >> the transmit part so that sk_state_change could work well >> and clear LOGIN_FLAGS_READ_ACTIVE early so that sk_data_ready >> could also work. >> >> While at here, use login_flags more efficient. >> >> Signed-off-by: Hou Pu <houpu@bytedance.com> >> --- >> drivers/target/iscsi/iscsi_target_nego.c | 30 ++++++++++++++++++++++++++---- >> include/target/iscsi/iscsi_target_core.h | 9 +++++---- >> 2 files changed, 31 insertions(+), 8 deletions(-) >> >> diff --git a/drivers/target/iscsi/iscsi_target_nego.c b/drivers/target/iscsi/iscsi_target_nego.c >> index 685d771b51d4..4cfa742e61df 100644 >> --- a/drivers/target/iscsi/iscsi_target_nego.c >> +++ b/drivers/target/iscsi/iscsi_target_nego.c >> @@ -622,6 +622,26 @@ static void iscsi_target_do_login_rx(struct work_struct *work) >> if (rc < 0) >> goto err; >> >> + /* >> + * LOGIN_FLAGS_READ_ACTIVE is cleared so that sk_data_ready >> + * could be trigger again after this. >> + * >> + * LOGIN_FLAGS_WRITE_ACTIVE is cleared after we successfully >> + * process a login pdu, so that sk_state_chage could do login >> + * cleanup as needed if the socket is closed. If a delayed work is >> + * ongoing (LOGIN_FLAGS_WRITE_ACTIVE or LOGIN_FLAGS_READ_ACTIVE), >> + * sk_state_change will leave the cleanup to the delayed work or >> + * it will schedule a delayed work to do cleanup. >> + */ >> + if (conn->sock) { >> + struct sock *sk = conn->sock->sk; >> + >> + write_lock_bh(&sk->sk_callback_lock); >> + clear_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags); >> + set_bit(LOGIN_FLAGS_WRITE_ACTIVE, &conn->login_flags); >> + write_unlock_bh(&sk->sk_callback_lock); >> + } >> + > Hey, > > I had one more question. > > With the above code, I think we have a race where if we clear the bit > above early and iscsi_target_sk_data_ready runs while > iscsi_target_do_login_rx is still running then we could queue the work > an extra time and get stuck. Because the bit is now not set, if > iscsi_target_sk_data_ready runs it will end up calling > schedule_delayed_work which will queue up the work again since the work > is running and not pending. > > If that is correct and we hit the race what happens if this was the last > login pdu, and we are supposed to go to full feature phase next? For > example if iscsi_target_do_login_rx runs an extra time, will we end up > stuck waiting in iscsi_target_do_login_rx's call to: > > rc = conn->conn_transport->iscsit_get_login_rx(conn, login); > > ? Just answering my own question. It looks like we do not get stuck. But we either get nothing on the session so the login timeout fires and we drop the session. Or, we get a PDU and the login thread reads it in before the normal rx thread does, but it assumes it is a login related and we most likely drop the session due to invalid fields. I think in iscsi_target_restore_sock_callbacks we want to do a: cancel_delayed_work(&conn->login_work) after we reset the callbacks and drop the sk_callback_lock lock.
On 2020/4/26 1:45 上午, Mike Christie wrote: > On 4/24/20 12:01 PM, Mike Christie wrote: >> On 4/24/20 7:35 AM, Hou Pu wrote: >>> iscsi_target_sk_data_ready() could be invoked indirectly >>> by iscsi_target_do_login_rx() from workqueue like following: >>> >>> iscsi_target_do_login_rx() >>> iscsi_target_do_login() >>> iscsi_target_do_tx_login_io() >>> iscsit_put_login_tx() >>> iscsi_login_tx_data() >>> tx_data() >>> sock_sendmsg_nosec() >>> tcp_sendmsg() >>> release_sock() >>> sk_backlog_rcv() >>> tcp_v4_do_rcv() >>> tcp_data_ready() >>> iscsi_target_sk_data_ready() >>> >>> At that time LOGIN_FLAGS_READ_ACTIVE is not cleared. >>> and iscsi_target_sk_data_ready will not read data >>> from the socket. And some iscsi initiator(i.e. libiscsi) >>> will wait forever for a reply. >>> >>> LOGIN_FLAGS_READ_ACTIVE should be cleared early just after >>> doing the receive and before writing to the socket in >>> iscsi_target_do_login_rx. >>> >>> But sad news is that LOGIN_FLAGS_READ_ACTIVE is also used >>> by sk_state_change to do login cleanup if a socket was closed >>> at login time. It is supposed to be cleared after the login >>> pdu is successfully processed and replied. >>> >>> So introduce another flag LOGIN_FLAGS_WRITE_ACTIVE to cover >>> the transmit part so that sk_state_change could work well >>> and clear LOGIN_FLAGS_READ_ACTIVE early so that sk_data_ready >>> could also work. >>> >>> While at here, use login_flags more efficient. >>> >>> Signed-off-by: Hou Pu <houpu@bytedance.com> >>> --- >>> drivers/target/iscsi/iscsi_target_nego.c | 30 ++++++++++++++++++++++++++---- >>> include/target/iscsi/iscsi_target_core.h | 9 +++++---- >>> 2 files changed, 31 insertions(+), 8 deletions(-) >>> >>> diff --git a/drivers/target/iscsi/iscsi_target_nego.c b/drivers/target/iscsi/iscsi_target_nego.c >>> index 685d771b51d4..4cfa742e61df 100644 >>> --- a/drivers/target/iscsi/iscsi_target_nego.c >>> +++ b/drivers/target/iscsi/iscsi_target_nego.c >>> @@ -622,6 +622,26 @@ static void iscsi_target_do_login_rx(struct work_struct *work) >>> if (rc < 0) >>> goto err; >>> >>> + /* >>> + * LOGIN_FLAGS_READ_ACTIVE is cleared so that sk_data_ready >>> + * could be trigger again after this. >>> + * >>> + * LOGIN_FLAGS_WRITE_ACTIVE is cleared after we successfully >>> + * process a login pdu, so that sk_state_chage could do login >>> + * cleanup as needed if the socket is closed. If a delayed work is >>> + * ongoing (LOGIN_FLAGS_WRITE_ACTIVE or LOGIN_FLAGS_READ_ACTIVE), >>> + * sk_state_change will leave the cleanup to the delayed work or >>> + * it will schedule a delayed work to do cleanup. >>> + */ >>> + if (conn->sock) { >>> + struct sock *sk = conn->sock->sk; >>> + >>> + write_lock_bh(&sk->sk_callback_lock); >>> + clear_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags); >>> + set_bit(LOGIN_FLAGS_WRITE_ACTIVE, &conn->login_flags); >>> + write_unlock_bh(&sk->sk_callback_lock); >>> + } >>> + >> Hey, >> >> I had one more question. >> >> With the above code, I think we have a race where if we clear the bit >> above early and iscsi_target_sk_data_ready runs while >> iscsi_target_do_login_rx is still running then we could queue the work >> an extra time and get stuck. Because the bit is now not set, if >> iscsi_target_sk_data_ready runs it will end up calling >> schedule_delayed_work which will queue up the work again since the work >> is running and not pending. Yes. I was trying to allow queuing the delayed work early. >> >> If that is correct and we hit the race what happens if this was the last >> login pdu, and we are supposed to go to full feature phase next? For >> example if iscsi_target_do_login_rx runs an extra time, will we end up >> stuck waiting in iscsi_target_do_login_rx's call to: >> >> rc = conn->conn_transport->iscsit_get_login_rx(conn, login); >> >> ? For the last login pdu, we may have race as you said. thanks for pointing it out. I was trying to image a case where we can hit the race, normally it is case a). a). initiator send last login pdu -> target received -> target replied b). initiator send last login pdu -> target received -> initiator send something -> target replied in case b). we will queue another delayed work which we should not. After the target replied the last login pdu, conn->conn_login is freed. we might visited it in the delayed work. > Just answering my own question. It looks like we do not get stuck. But > we either get nothing on the session so the login timeout fires and we > drop the session. Or, we get a PDU and the login thread reads it in > before the normal rx thread does, but it assumes it is a login related > and we most likely drop the session due to invalid fields. > > I think in iscsi_target_restore_sock_callbacks we want to do a: > > cancel_delayed_work(&conn->login_work) > > after we reset the callbacks and drop the sk_callback_lock lock. I am not very sure if we could or if it is good to cancel_delayed_work from the work itself. If it is ok then i am ok with it. Or in another way, I think we could just clear LOGIN_FLAGS_READ_ACTIVE and set LOGIN_FLAGS_WRITE_ACTIVE after iscsi_target_restore_sock_callbacks when finish process last login pdu. What do you think? Thanks, Hou >
>>>> + */ >>>> + if (conn->sock) { >>>> + struct sock *sk = conn->sock->sk; >>>> + >>>> + write_lock_bh(&sk->sk_callback_lock); >>>> + clear_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags); >>>> + set_bit(LOGIN_FLAGS_WRITE_ACTIVE, &conn->login_flags); >>>> + write_unlock_bh(&sk->sk_callback_lock); >>>> + } >>>> + >>> Hey, >>> >>> I had one more question. >>> >>> With the above code, I think we have a race where if we clear the bit >>> above early and iscsi_target_sk_data_ready runs while >>> iscsi_target_do_login_rx is still running then we could queue the work >>> an extra time and get stuck. Because the bit is now not set, if >>> iscsi_target_sk_data_ready runs it will end up calling >>> schedule_delayed_work which will queue up the work again since the work >>> is running and not pending. > > Yes. I was trying to allow queuing the delayed work early. > >>> >>> If that is correct and we hit the race what happens if this was the >>> last >>> login pdu, and we are supposed to go to full feature phase next? For >>> example if iscsi_target_do_login_rx runs an extra time, will we end up >>> stuck waiting in iscsi_target_do_login_rx's call to: >>> >>> rc = conn->conn_transport->iscsit_get_login_rx(conn, login); >>> >>> ? > > For the last login pdu, we may have race as you said. thanks for > pointing it out. > > I was trying to image a case where we can hit the race, normally it is > case a). > > a). initiator send last login pdu -> target received -> target replied > > b). initiator send last login pdu -> target received -> initiator send > something -> target replied > > in case b). we will queue another delayed work which we should not. > After the target replied > > the last login pdu, conn->conn_login is freed. we might visited it in > the delayed work. > > >> Just answering my own question. It looks like we do not get stuck. But >> we either get nothing on the session so the login timeout fires and we >> drop the session. Or, we get a PDU and the login thread reads it in >> before the normal rx thread does, but it assumes it is a login related >> and we most likely drop the session due to invalid fields. >> >> I think in iscsi_target_restore_sock_callbacks we want to do a: >> >> cancel_delayed_work(&conn->login_work) >> >> after we reset the callbacks and drop the sk_callback_lock lock. > > I am not very sure if we could or if it is good to cancel_delayed_work > from the work itself. > > If it is ok then i am ok with it. Or in another way, I think we could > just clear > > LOGIN_FLAGS_READ_ACTIVE and set LOGIN_FLAGS_WRITE_ACTIVE > > after iscsi_target_restore_sock_callbacks when finish process last > login pdu. That would look like (in iscsi_target_do_tx_login_io): diff --git a/drivers/target/iscsi/iscsi_target_nego.c b/drivers/target/iscsi/iscsi_target_nego.c index 685d771b51d4..4d0658731382 100644 --- a/drivers/target/iscsi/iscsi_target_nego.c +++ b/drivers/target/iscsi/iscsi_target_nego.c @@ -328,6 +328,28 @@ static int iscsi_target_do_tx_login_io(struct iscsi_conn *conn, struct iscsi_log u32 padding = 0; struct iscsi_login_rsp *login_rsp; + /* + * LOGIN_FLAGS_READ_ACTIVE is cleared so that sk_data_ready + * could be trigger again after this. + * + * LOGIN_FLAGS_WRITE_ACTIVE is cleared after we successfully + * process a login pdu, so that sk_state_chage could do login + * cleanup as needed if the socket is closed. If a delayed work is + * ongoing (LOGIN_FLAGS_WRITE_ACTIVE or LOGIN_FLAGS_READ_ACTIVE), + * sk_state_change will leave the cleanup to the delayed work or + * it will schedule a delayed work to do cleanup. + */ + if (conn->sock) { + struct sock *sk = conn->sock->sk; + + write_lock_bh(&sk->sk_callback_lock); + if (!test_bit(LOGIN_FLAGS_INITIAL_PDU, &conn->login_flags)) { + clear_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags); + set_bit(LOGIN_FLAGS_WRITE_ACTIVE, &conn->login_flags); + } + write_unlock_bh(&sk->sk_callback_lock); + } > > What do you think? > > > > Thanks, > > Hou > > >>
On 4/26/20 1:09 AM, Hou Pu wrote: > >>>>> + */ >>>>> + if (conn->sock) { >>>>> + struct sock *sk = conn->sock->sk; >>>>> + >>>>> + write_lock_bh(&sk->sk_callback_lock); >>>>> + clear_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags); >>>>> + set_bit(LOGIN_FLAGS_WRITE_ACTIVE, &conn->login_flags); >>>>> + write_unlock_bh(&sk->sk_callback_lock); >>>>> + } >>>>> + >>>> Hey, >>>> >>>> I had one more question. >>>> >>>> With the above code, I think we have a race where if we clear the bit >>>> above early and iscsi_target_sk_data_ready runs while >>>> iscsi_target_do_login_rx is still running then we could queue the work >>>> an extra time and get stuck. Because the bit is now not set, if >>>> iscsi_target_sk_data_ready runs it will end up calling >>>> schedule_delayed_work which will queue up the work again since the work >>>> is running and not pending. >> >> Yes. I was trying to allow queuing the delayed work early. >> >>>> >>>> If that is correct and we hit the race what happens if this was the >>>> last >>>> login pdu, and we are supposed to go to full feature phase next? For >>>> example if iscsi_target_do_login_rx runs an extra time, will we end up >>>> stuck waiting in iscsi_target_do_login_rx's call to: >>>> >>>> rc = conn->conn_transport->iscsit_get_login_rx(conn, login); >>>> >>>> ? >> >> For the last login pdu, we may have race as you said. thanks for >> pointing it out. >> >> I was trying to image a case where we can hit the race, normally it is >> case a). >> >> a). initiator send last login pdu -> target received -> target replied >> >> b). initiator send last login pdu -> target received -> initiator send >> something -> target replied >> >> in case b). we will queue another delayed work which we should not. >> After the target replied >> >> the last login pdu, conn->conn_login is freed. we might visited it in >> the delayed wo>> >> >>> Just answering my own question. It looks like we do not get stuck. But >>> we either get nothing on the session so the login timeout fires and we >>> drop the session. Or, we get a PDU and the login thread reads it in >>> before the normal rx thread does, but it assumes it is a login related >>> and we most likely drop the session due to invalid fields. >>> >>> I think in iscsi_target_restore_sock_callbacks we want to do a: >>> >>> cancel_delayed_work(&conn->login_work) >>> >>> after we reset the callbacks and drop the sk_callback_lock lock. >> >> I am not very sure if we could or if it is good to cancel_delayed_work >> from the work itself. >> >> If it is ok then i am ok with it. Or in another way, I think we could >> just clear >> >> LOGIN_FLAGS_READ_ACTIVE and set LOGIN_FLAGS_WRITE_ACTIVE >> >> after iscsi_target_restore_sock_callbacks when finish process last >> login pdu. > > That would look like (in iscsi_target_do_tx_login_io): > > diff --git a/drivers/target/iscsi/iscsi_target_nego.c > b/drivers/target/iscsi/iscsi_target_nego.c > index 685d771b51d4..4d0658731382 100644 > --- a/drivers/target/iscsi/iscsi_target_nego.c > +++ b/drivers/target/iscsi/iscsi_target_nego.c > @@ -328,6 +328,28 @@ static int iscsi_target_do_tx_login_io(struct > iscsi_conn *conn, struct iscsi_log > u32 padding = 0; > struct iscsi_login_rsp *login_rsp; > > + /* > + * LOGIN_FLAGS_READ_ACTIVE is cleared so that sk_data_ready > + * could be trigger again after this. > + * > + * LOGIN_FLAGS_WRITE_ACTIVE is cleared after we successfully > + * process a login pdu, so that sk_state_chage could do login > + * cleanup as needed if the socket is closed. If a delayed work is > + * ongoing (LOGIN_FLAGS_WRITE_ACTIVE or LOGIN_FLAGS_READ_ACTIVE), > + * sk_state_change will leave the cleanup to the delayed work or > + * it will schedule a delayed work to do cleanup. > + */ > + if (conn->sock) { > + struct sock *sk = conn->sock->sk; > + > + write_lock_bh(&sk->sk_callback_lock); > + if (!test_bit(LOGIN_FLAGS_INITIAL_PDU, > &conn->login_flags)) { > + clear_bit(LOGIN_FLAGS_READ_ACTIVE, > &conn->login_flags); > + set_bit(LOGIN_FLAGS_WRITE_ACTIVE, > &conn->login_flags); > + } > + write_unlock_bh(&sk->sk_callback_lock); > + } You lost me. I didn't understand this part. Would you still be doing the above bit manipulation in iscsi_target_do_login_rx too? Is the above code then to handle when iscsi_target_start_negotiation->iscsi_target_do_login->iscsi_target_do_tx_login_io runs? I was thinking when you mentioned the final login PDU you were going to do something when you detect login->login_complete is true. This code is not my expertise area, so I might just not be understanding something simple about how it all works.
On 2020/4/29 1:50 上午, Mike Christie wrote: > On 4/26/20 1:09 AM, Hou Pu wrote: >> >>>>>> + */ >>>>>> + if (conn->sock) { >>>>>> + struct sock *sk = conn->sock->sk; >>>>>> + >>>>>> + write_lock_bh(&sk->sk_callback_lock); >>>>>> + clear_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags); >>>>>> + set_bit(LOGIN_FLAGS_WRITE_ACTIVE, &conn->login_flags); >>>>>> + write_unlock_bh(&sk->sk_callback_lock); >>>>>> + } >>>>>> + >>>>> Hey, >>>>> >>>>> I had one more question. >>>>> >>>>> With the above code, I think we have a race where if we clear the bit >>>>> above early and iscsi_target_sk_data_ready runs while >>>>> iscsi_target_do_login_rx is still running then we could queue the work >>>>> an extra time and get stuck. Because the bit is now not set, if >>>>> iscsi_target_sk_data_ready runs it will end up calling >>>>> schedule_delayed_work which will queue up the work again since the work >>>>> is running and not pending. >>> >>> Yes. I was trying to allow queuing the delayed work early. >>> >>>>> >>>>> If that is correct and we hit the race what happens if this was the >>>>> last >>>>> login pdu, and we are supposed to go to full feature phase next? For >>>>> example if iscsi_target_do_login_rx runs an extra time, will we end up >>>>> stuck waiting in iscsi_target_do_login_rx's call to: >>>>> >>>>> rc = conn->conn_transport->iscsit_get_login_rx(conn, login); >>>>> >>>>> ? >>> >>> For the last login pdu, we may have race as you said. thanks for >>> pointing it out. >>> >>> I was trying to image a case where we can hit the race, normally it is >>> case a). >>> >>> a). initiator send last login pdu -> target received -> target replied >>> >>> b). initiator send last login pdu -> target received -> initiator send >>> something -> target replied >>> >>> in case b). we will queue another delayed work which we should not. >>> After the target replied >>> >>> the last login pdu, conn->conn_login is freed. we might visited it in >>> the delayed wo>> >>> >>>> Just answering my own question. It looks like we do not get stuck. But >>>> we either get nothing on the session so the login timeout fires and we >>>> drop the session. Or, we get a PDU and the login thread reads it in >>>> before the normal rx thread does, but it assumes it is a login related >>>> and we most likely drop the session due to invalid fields. >>>> >>>> I think in iscsi_target_restore_sock_callbacks we want to do a: >>>> >>>> cancel_delayed_work(&conn->login_work) >>>> >>>> after we reset the callbacks and drop the sk_callback_lock lock. >>> >>> I am not very sure if we could or if it is good to cancel_delayed_work >>> from the work itself. >>> >>> If it is ok then i am ok with it. Or in another way, I think we could >>> just clear >>> >>> LOGIN_FLAGS_READ_ACTIVE and set LOGIN_FLAGS_WRITE_ACTIVE >>> >>> after iscsi_target_restore_sock_callbacks when finish process last >>> login pdu. >> >> That would look like (in iscsi_target_do_tx_login_io): >> >> diff --git a/drivers/target/iscsi/iscsi_target_nego.c >> b/drivers/target/iscsi/iscsi_target_nego.c >> index 685d771b51d4..4d0658731382 100644 >> --- a/drivers/target/iscsi/iscsi_target_nego.c >> +++ b/drivers/target/iscsi/iscsi_target_nego.c >> @@ -328,6 +328,28 @@ static int iscsi_target_do_tx_login_io(struct >> iscsi_conn *conn, struct iscsi_log >> u32 padding = 0; >> struct iscsi_login_rsp *login_rsp; >> >> + /* >> + * LOGIN_FLAGS_READ_ACTIVE is cleared so that sk_data_ready >> + * could be trigger again after this. >> + * >> + * LOGIN_FLAGS_WRITE_ACTIVE is cleared after we successfully >> + * process a login pdu, so that sk_state_chage could do login >> + * cleanup as needed if the socket is closed. If a delayed work is >> + * ongoing (LOGIN_FLAGS_WRITE_ACTIVE or LOGIN_FLAGS_READ_ACTIVE), >> + * sk_state_change will leave the cleanup to the delayed work or >> + * it will schedule a delayed work to do cleanup. >> + */ >> + if (conn->sock) { >> + struct sock *sk = conn->sock->sk; >> + >> + write_lock_bh(&sk->sk_callback_lock); >> + if (!test_bit(LOGIN_FLAGS_INITIAL_PDU, >> &conn->login_flags)) { >> + clear_bit(LOGIN_FLAGS_READ_ACTIVE, >> &conn->login_flags); >> + set_bit(LOGIN_FLAGS_WRITE_ACTIVE, >> &conn->login_flags); >> + } >> + write_unlock_bh(&sk->sk_callback_lock); >> + } > > You lost me. I didn't understand this part. Would you still be doing the > above bit manipulation in iscsi_target_do_login_rx too? By moving this chunk here(from iscsi_target_do_login_rx to iscsi_target_do_tx_login_io), we have already restore sock callbacks(like below in iscsi_target_do_login), So clearing READ_ACTIVE would not cause another delayed work being scheduled. So I think we would not have the race you pointed out for last login pdu. [code snippet iscsi_target_do_login() ... login->login_complete = 1; iscsi_target_restore_sock_callbacks(conn); if (iscsi_target_do_tx_login_io(conn, login) < 0) ... ] > > Is the above code then to handle when > iscsi_target_start_negotiation->iscsi_target_do_login->iscsi_target_do_tx_login_io > runs? > > I was thinking when you mentioned the final login PDU you were going to > do something when you detect login->login_complete is true. > Even if we move above chunk here, I realize that we could still theoretically have problem. For login pdus which is not initial or last login pdus, after enable scheduling delayed work by clearing READ_ACTIVE. It is possible to queue a delayed work. (i.e. a out of order initiator could send pdus at any time). In case we received such out of order pdu, we need to cancel them when we have an error and later doing the clean up. I will try to solve this and resend these patches. Maybe it's better to cancel_delayed_work() in iscsi_target_do_login_rx. I might not express myself very clear. Please see next version of these patch. > This code is not my expertise area, so I might just not be understanding > something simple about how it all works. Your comments make sense and are helpful for me. Thanks. > Thanks, Hou
diff --git a/drivers/target/iscsi/iscsi_target_nego.c b/drivers/target/iscsi/iscsi_target_nego.c index 685d771b51d4..4cfa742e61df 100644 --- a/drivers/target/iscsi/iscsi_target_nego.c +++ b/drivers/target/iscsi/iscsi_target_nego.c @@ -622,6 +622,26 @@ static void iscsi_target_do_login_rx(struct work_struct *work) if (rc < 0) goto err; + /* + * LOGIN_FLAGS_READ_ACTIVE is cleared so that sk_data_ready + * could be trigger again after this. + * + * LOGIN_FLAGS_WRITE_ACTIVE is cleared after we successfully + * process a login pdu, so that sk_state_chage could do login + * cleanup as needed if the socket is closed. If a delayed work is + * ongoing (LOGIN_FLAGS_WRITE_ACTIVE or LOGIN_FLAGS_READ_ACTIVE), + * sk_state_change will leave the cleanup to the delayed work or + * it will schedule a delayed work to do cleanup. + */ + if (conn->sock) { + struct sock *sk = conn->sock->sk; + + write_lock_bh(&sk->sk_callback_lock); + clear_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags); + set_bit(LOGIN_FLAGS_WRITE_ACTIVE, &conn->login_flags); + write_unlock_bh(&sk->sk_callback_lock); + } + pr_debug("iscsi_target_do_login_rx after rx_login_io, %p, %s:%d\n", conn, current->comm, current->pid); @@ -629,7 +649,8 @@ static void iscsi_target_do_login_rx(struct work_struct *work) if (rc < 0) { goto err; } else if (!rc) { - if (iscsi_target_sk_check_and_clear(conn, LOGIN_FLAGS_READ_ACTIVE)) + if (iscsi_target_sk_check_and_clear(conn, + LOGIN_FLAGS_WRITE_ACTIVE)) goto err; } else if (rc == 1) { iscsi_target_nego_release(conn); @@ -670,9 +691,10 @@ static void iscsi_target_sk_state_change(struct sock *sk) state = __iscsi_target_sk_check_close(sk); pr_debug("__iscsi_target_sk_close_change: state: %d\n", state); - if (test_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags)) { - pr_debug("Got LOGIN_FLAGS_READ_ACTIVE=1 sk_state_change" - " conn: %p\n", conn); + if (test_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags) || + test_bit(LOGIN_FLAGS_WRITE_ACTIVE, &conn->login_flags)) { + pr_debug("Got LOGIN_FLAGS_{READ|WRITE}_ACTIVE=1" + " sk_state_change conn: %p\n", conn); if (state) set_bit(LOGIN_FLAGS_CLOSED, &conn->login_flags); write_unlock_bh(&sk->sk_callback_lock); diff --git a/include/target/iscsi/iscsi_target_core.h b/include/target/iscsi/iscsi_target_core.h index 591cd9e4692c..844bef255e89 100644 --- a/include/target/iscsi/iscsi_target_core.h +++ b/include/target/iscsi/iscsi_target_core.h @@ -566,10 +566,11 @@ struct iscsi_conn { struct socket *sock; void (*orig_data_ready)(struct sock *); void (*orig_state_change)(struct sock *); -#define LOGIN_FLAGS_READ_ACTIVE 1 -#define LOGIN_FLAGS_CLOSED 2 -#define LOGIN_FLAGS_READY 4 -#define LOGIN_FLAGS_INITIAL_PDU 8 +#define LOGIN_FLAGS_READY 0 +#define LOGIN_FLAGS_INITIAL_PDU 1 +#define LOGIN_FLAGS_READ_ACTIVE 2 +#define LOGIN_FLAGS_WRITE_ACTIVE 3 +#define LOGIN_FLAGS_CLOSED 4 unsigned long login_flags; struct delayed_work login_work; struct iscsi_login *login;
iscsi_target_sk_data_ready() could be invoked indirectly by iscsi_target_do_login_rx() from workqueue like following: iscsi_target_do_login_rx() iscsi_target_do_login() iscsi_target_do_tx_login_io() iscsit_put_login_tx() iscsi_login_tx_data() tx_data() sock_sendmsg_nosec() tcp_sendmsg() release_sock() sk_backlog_rcv() tcp_v4_do_rcv() tcp_data_ready() iscsi_target_sk_data_ready() At that time LOGIN_FLAGS_READ_ACTIVE is not cleared. and iscsi_target_sk_data_ready will not read data from the socket. And some iscsi initiator(i.e. libiscsi) will wait forever for a reply. LOGIN_FLAGS_READ_ACTIVE should be cleared early just after doing the receive and before writing to the socket in iscsi_target_do_login_rx. But sad news is that LOGIN_FLAGS_READ_ACTIVE is also used by sk_state_change to do login cleanup if a socket was closed at login time. It is supposed to be cleared after the login pdu is successfully processed and replied. So introduce another flag LOGIN_FLAGS_WRITE_ACTIVE to cover the transmit part so that sk_state_change could work well and clear LOGIN_FLAGS_READ_ACTIVE early so that sk_data_ready could also work. While at here, use login_flags more efficient. Signed-off-by: Hou Pu <houpu@bytedance.com> --- drivers/target/iscsi/iscsi_target_nego.c | 30 ++++++++++++++++++++++++++---- include/target/iscsi/iscsi_target_core.h | 9 +++++---- 2 files changed, 31 insertions(+), 8 deletions(-)