| Message ID | 1607935317-263599-1-git-send-email-wubo40@huawei.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Headers | show |
| Series | [RFC] scsi:libiscsi:Fix possible NULL dereference in iscsi_eh_cmd_timed_out | expand |
On 12/14/20 2:41 AM, Wu Bo wrote: > When testing kernel 4.18 version, NULL pointer dereference problem occurs > in iscsi_eh_cmd_timed_out function. > > I think this bug in the upstream is still exists. > > The analysis reasons are as follows: > 1) For some reason, I/O command did not complete within > the timeout period. The block layer timer works, > call scsi_times_out() to handle I/O timeout logic. > At the same time the command just completes. > > 2) scsi_times_out() call iscsi_eh_cmd_timed_out() > to processing timeout logic. although there is an NULL judgment > for the task, the task has not been released yet now. > > 3) iscsi_complete_task() call __iscsi_put_task(), > The task reference count reaches zero, the conditions for free task > is met, then iscsi_free_task () free the task, > and let sc->SCp.ptr = NULL. After iscsi_eh_cmd_timed_out passes > the task judgment check, there may be NULL dereference scenarios > later. > I have a patch for this I think. This is broken out of patchset I was trying to fixup the back lock usage for offload drivers, so I have only compile tested it. There is another issue where the for lun reset cleanup we could race. The comments mention suspending the rx side, but we only do that for session level cleaup. The basic idea is we don't want to add more frwd lock uses in the completion patch like in your patch. In these non perf paths, like the tmf/timeout case we can just take a ref to the cmd so it's not freed from under us. diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi/libiscsi.c index f9314f1..f07f8c1 100644 --- a/drivers/scsi/libiscsi.c +++ b/drivers/scsi/libiscsi.c @@ -573,18 +573,9 @@ void iscsi_complete_scsi_task(struct iscsi_task *task, static void fail_scsi_task(struct iscsi_task *task, int err) { struct iscsi_conn *conn = task->conn; - struct scsi_cmnd *sc; + struct scsi_cmnd *sc = task->sc; int state; - /* - * if a command completes and we get a successful tmf response - * we will hit this because the scsi eh abort code does not take - * a ref to the task. - */ - sc = task->sc; - if (!sc) - return; - if (task->state == ISCSI_TASK_PENDING) { /* * cmd never made it to the xmit thread, so we should not count @@ -1855,26 +1846,34 @@ static int iscsi_exec_task_mgmt_fn(struct iscsi_conn *conn, } /* - * Fail commands. session lock held and recv side suspended and xmit - * thread flushed + * Fail commands. session frwd lock held and and xmit thread flushed. */ static void fail_scsi_tasks(struct iscsi_conn *conn, u64 lun, int error) { + struct iscsi_session *session = conn->session; struct iscsi_task *task; int i; - for (i = 0; i < conn->session->cmds_max; i++) { - task = conn->session->cmds[i]; - if (!task->sc || task->state == ISCSI_TASK_FREE) + for (i = 0; i < session->cmds_max; i++) { + spin_lock_bh(&session->back_lock); + task = session->cmds[i]; + if (!task->sc || task->state == ISCSI_TASK_FREE) { + spin_unlock_bh(&session->back_lock); continue; + } - if (lun != -1 && lun != task->sc->device->lun) + if (lun != -1 && lun != task->sc->device->lun) { + spin_unlock_bh(&session->back_lock); continue; + } + __iscsi_get_task(task); + spin_unlock_bh(&session->back_lock); - ISCSI_DBG_SESSION(conn->session, + ISCSI_DBG_SESSION(session, "failing sc %p itt 0x%x state %d\n", task->sc, task->itt, task->state); fail_scsi_task(task, error); + iscsi_put_task(task); } } @@ -1953,6 +1952,7 @@ enum blk_eh_timer_return iscsi_eh_cmd_timed_out(struct scsi_cmnd *sc) ISCSI_DBG_EH(session, "scsi cmd %p timedout\n", sc); spin_lock_bh(&session->frwd_lock); + spin_lock(&session->back_lock); task = (struct iscsi_task *)sc->SCp.ptr; if (!task) { /* @@ -1960,8 +1960,11 @@ enum blk_eh_timer_return iscsi_eh_cmd_timed_out(struct scsi_cmnd *sc) * so let timeout code complete it now. */ rc = BLK_EH_DONE; + spin_unlock(&session->back_lock); goto done; } + __iscsi_get_task(task); + spin_unlock(&session->back_lock); if (session->state != ISCSI_STATE_LOGGED_IN) { /* @@ -2077,9 +2080,12 @@ enum blk_eh_timer_return iscsi_eh_cmd_timed_out(struct scsi_cmnd *sc) rc = BLK_EH_RESET_TIMER; done: - if (task) - task->last_timeout = jiffies; spin_unlock_bh(&session->frwd_lock); + + if (task) { + task->last_timeout = jiffies; + iscsi_put_task(task); + } ISCSI_DBG_EH(session, "return %s\n", rc == BLK_EH_RESET_TIMER ? "timer reset" : "shutdown or nh"); return rc; @@ -2187,15 +2193,20 @@ int iscsi_eh_abort(struct scsi_cmnd *sc) conn->eh_abort_cnt++; age = session->age; - task = (struct iscsi_task *)sc->SCp.ptr; - ISCSI_DBG_EH(session, "aborting [sc %p itt 0x%x]\n", - sc, task->itt); - - /* task completed before time out */ - if (!task->sc) { + spin_lock(&session->back_lock); + task = (struct iscsi_task *)sc->SCp.ptr; + if (!task || !task->sc) { + /* task completed before time out */ ISCSI_DBG_EH(session, "sc completed while abort in progress\n"); - goto success; + + spin_unlock(&session->back_lock); + spin_unlock_bh(&session->frwd_lock); + mutex_unlock(&session->eh_mutex); + return SUCCESS; } + ISCSI_DBG_EH(session, "aborting [sc %p itt 0x%x]\n", sc, task->itt); + __iscsi_get_task(task); + spin_unlock(&session->back_lock); if (task->state == ISCSI_TASK_PENDING) { fail_scsi_task(task, DID_ABORT); @@ -2258,6 +2269,8 @@ int iscsi_eh_abort(struct scsi_cmnd *sc) ISCSI_DBG_EH(session, "abort success [sc %p itt 0x%x]\n", sc, task->itt); mutex_unlock(&session->eh_mutex); + + iscsi_put_task(task); return SUCCESS; failed: @@ -2266,6 +2279,8 @@ int iscsi_eh_abort(struct scsi_cmnd *sc) ISCSI_DBG_EH(session, "abort failed [sc %p itt 0x%x]\n", sc, task ? task->itt : 0); mutex_unlock(&session->eh_mutex); + + iscsi_put_task(task); return FAILED; } EXPORT_SYMBOL_GPL(iscsi_eh_abort);
On 2020/12/15 1:36, Mike Christie wrote: > On 12/14/20 2:41 AM, Wu Bo wrote: >> When testing kernel 4.18 version, NULL pointer dereference problem occurs >> in iscsi_eh_cmd_timed_out function. >> >> I think this bug in the upstream is still exists. >> >> The analysis reasons are as follows: >> 1) For some reason, I/O command did not complete within >> the timeout period. The block layer timer works, >> call scsi_times_out() to handle I/O timeout logic. >> At the same time the command just completes. >> >> 2) scsi_times_out() call iscsi_eh_cmd_timed_out() >> to processing timeout logic. although there is an NULL judgment >> for the task, the task has not been released yet now. >> >> 3) iscsi_complete_task() call __iscsi_put_task(), >> The task reference count reaches zero, the conditions for free task >> is met, then iscsi_free_task () free the task, >> and let sc->SCp.ptr = NULL. After iscsi_eh_cmd_timed_out passes >> the task judgment check, there may be NULL dereference scenarios >> later. >> > > I have a patch for this I think. This is broken out of patchset I was > trying to fixup the back lock usage for offload drivers, so I have only > compile tested it. > > There is another issue where the for lun reset cleanup we could race. The > comments mention suspending the rx side, but we only do that for session level > cleaup. > > The basic idea is we don't want to add more frwd lock uses in the completion > patch like in your patch. In these non perf paths, like the tmf/timeout case > we can just take a ref to the cmd so it's not freed from under us. > You are right, add more frwd lock does affect performance in the completion. > > > diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi/libiscsi.c > index f9314f1..f07f8c1 100644 > --- a/drivers/scsi/libiscsi.c > +++ b/drivers/scsi/libiscsi.c > @@ -573,18 +573,9 @@ void iscsi_complete_scsi_task(struct iscsi_task *task, > static void fail_scsi_task(struct iscsi_task *task, int err) > { > struct iscsi_conn *conn = task->conn; > - struct scsi_cmnd *sc; > + struct scsi_cmnd *sc = task->sc; > int state; > > - /* > - * if a command completes and we get a successful tmf response > - * we will hit this because the scsi eh abort code does not take > - * a ref to the task. > - */ > - sc = task->sc; > - if (!sc) > - return; > - > if (task->state == ISCSI_TASK_PENDING) { > /* > * cmd never made it to the xmit thread, so we should not count > @@ -1855,26 +1846,34 @@ static int iscsi_exec_task_mgmt_fn(struct iscsi_conn *conn, > } > > /* > - * Fail commands. session lock held and recv side suspended and xmit > - * thread flushed > + * Fail commands. session frwd lock held and and xmit thread flushed. > */ > static void fail_scsi_tasks(struct iscsi_conn *conn, u64 lun, int error) > { > + struct iscsi_session *session = conn->session; > struct iscsi_task *task; > int i; > > - for (i = 0; i < conn->session->cmds_max; i++) { > - task = conn->session->cmds[i]; > - if (!task->sc || task->state == ISCSI_TASK_FREE) > + for (i = 0; i < session->cmds_max; i++) { > + spin_lock_bh(&session->back_lock); > + task = session->cmds[i]; > + if (!task->sc || task->state == ISCSI_TASK_FREE) { > + spin_unlock_bh(&session->back_lock); > continue; > + } > > - if (lun != -1 && lun != task->sc->device->lun) > + if (lun != -1 && lun != task->sc->device->lun) { > + spin_unlock_bh(&session->back_lock); > continue; > + } > + __iscsi_get_task(task); > + spin_unlock_bh(&session->back_lock); > > - ISCSI_DBG_SESSION(conn->session, > + ISCSI_DBG_SESSION(session, > "failing sc %p itt 0x%x state %d\n", > task->sc, task->itt, task->state); > fail_scsi_task(task, error); > + iscsi_put_task(task); > } > } > > @@ -1953,6 +1952,7 @@ enum blk_eh_timer_return iscsi_eh_cmd_timed_out(struct scsi_cmnd *sc) > ISCSI_DBG_EH(session, "scsi cmd %p timedout\n", sc); > > spin_lock_bh(&session->frwd_lock); > + spin_lock(&session->back_lock); > task = (struct iscsi_task *)sc->SCp.ptr; > if (!task) { > /* > @@ -1960,8 +1960,11 @@ enum blk_eh_timer_return iscsi_eh_cmd_timed_out(struct scsi_cmnd *sc) > * so let timeout code complete it now. > */ > rc = BLK_EH_DONE; > + spin_unlock(&session->back_lock); > goto done; > } > + __iscsi_get_task(task); > + spin_unlock(&session->back_lock); > > if (session->state != ISCSI_STATE_LOGGED_IN) { > /* > @@ -2077,9 +2080,12 @@ enum blk_eh_timer_return iscsi_eh_cmd_timed_out(struct scsi_cmnd *sc) > rc = BLK_EH_RESET_TIMER; > > done: > - if (task) > - task->last_timeout = jiffies; > spin_unlock_bh(&session->frwd_lock); > + > + if (task) { > + task->last_timeout = jiffies; > + iscsi_put_task(task); > + } > ISCSI_DBG_EH(session, "return %s\n", rc == BLK_EH_RESET_TIMER ? > "timer reset" : "shutdown or nh"); > return rc; > @@ -2187,15 +2193,20 @@ int iscsi_eh_abort(struct scsi_cmnd *sc) > conn->eh_abort_cnt++; > age = session->age; > > - task = (struct iscsi_task *)sc->SCp.ptr; > - ISCSI_DBG_EH(session, "aborting [sc %p itt 0x%x]\n", > - sc, task->itt); > - > - /* task completed before time out */ > - if (!task->sc) { > + spin_lock(&session->back_lock); > + task = (struct iscsi_task *)sc->SCp.ptr; > + if (!task || !task->sc) { > + /* task completed before time out */ > ISCSI_DBG_EH(session, "sc completed while abort in progress\n"); > - goto success; > + > + spin_unlock(&session->back_lock); > + spin_unlock_bh(&session->frwd_lock); > + mutex_unlock(&session->eh_mutex); > + return SUCCESS; > } > + ISCSI_DBG_EH(session, "aborting [sc %p itt 0x%x]\n", sc, task->itt); > + __iscsi_get_task(task); > + spin_unlock(&session->back_lock); > > if (task->state == ISCSI_TASK_PENDING) { > fail_scsi_task(task, DID_ABORT); > @@ -2258,6 +2269,8 @@ int iscsi_eh_abort(struct scsi_cmnd *sc) > ISCSI_DBG_EH(session, "abort success [sc %p itt 0x%x]\n", > sc, task->itt); > mutex_unlock(&session->eh_mutex); > + > + iscsi_put_task(task); > return SUCCESS; > > failed: > @@ -2266,6 +2279,8 @@ int iscsi_eh_abort(struct scsi_cmnd *sc) > ISCSI_DBG_EH(session, "abort failed [sc %p itt 0x%x]\n", sc, > task ? task->itt : 0); > mutex_unlock(&session->eh_mutex); > + > + iscsi_put_task(task); > return FAILED; > } > EXPORT_SYMBOL_GPL(iscsi_eh_abort); > I have tested this patch, covering IO timeout and IO abort error handling scenarios, it is works well. It is lgtm, Thanks
diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi/libiscsi.c index 0bb5d76..e2cacdd 100644 --- a/drivers/scsi/libiscsi.c +++ b/drivers/scsi/libiscsi.c @@ -876,7 +876,9 @@ static void iscsi_scsi_cmd_rsp(struct iscsi_conn *conn, struct iscsi_hdr *hdr, ISCSI_DBG_SESSION(session, "cmd rsp done [sc %p res %d itt 0x%x]\n", sc, sc->result, task->itt); conn->scsirsp_pdus_cnt++; + spin_lock_bh(&session->frwd_lock); iscsi_complete_task(task, ISCSI_TASK_COMPLETED); + spin_unlock_bh(&session->frwd_lock); } /** @@ -917,7 +919,9 @@ static void iscsi_scsi_cmd_rsp(struct iscsi_conn *conn, struct iscsi_hdr *hdr, "[sc %p res %d itt 0x%x]\n", sc, sc->result, task->itt); conn->scsirsp_pdus_cnt++; + spin_lock_bh(&conn->session->frwd_lock); iscsi_complete_task(task, ISCSI_TASK_COMPLETED); + spin_unlock_bh(&conn->session->frwd_lock); } static void iscsi_tmf_rsp(struct iscsi_conn *conn, struct iscsi_hdr *hdr) @@ -1001,7 +1005,10 @@ static int iscsi_nop_out_rsp(struct iscsi_task *task, rc = ISCSI_ERR_CONN_FAILED; } else mod_timer(&conn->transport_timer, jiffies + conn->recv_timeout); + spin_lock_bh(&conn->session->frwd_lock); iscsi_complete_task(task, ISCSI_TASK_COMPLETED); + spin_unlock_bh(&conn->session->frwd_lock); + return rc; } @@ -1241,7 +1248,9 @@ int __iscsi_complete_pdu(struct iscsi_conn *conn, struct iscsi_hdr *hdr, } iscsi_tmf_rsp(conn, hdr); + spin_lock_bh(&session->frwd_lock); iscsi_complete_task(task, ISCSI_TASK_COMPLETED); + spin_unlock_bh(&session->frwd_lock); break; case ISCSI_OP_NOOP_IN: iscsi_update_cmdsn(session, (struct iscsi_nopin*)hdr); @@ -1264,7 +1273,10 @@ int __iscsi_complete_pdu(struct iscsi_conn *conn, struct iscsi_hdr *hdr, recv_pdu: if (iscsi_recv_pdu(conn->cls_conn, hdr, data, datalen)) rc = ISCSI_ERR_CONN_FAILED; + spin_lock_bh(&session->frwd_lock); iscsi_complete_task(task, ISCSI_TASK_COMPLETED); + spin_unlock_bh(&session->frwd_lock); + return rc; } EXPORT_SYMBOL_GPL(__iscsi_complete_pdu);
When testing kernel 4.18 version, NULL pointer dereference problem occurs in iscsi_eh_cmd_timed_out function. I think this bug in the upstream is still exists. The analysis reasons are as follows: 1) For some reason, I/O command did not complete within the timeout period. The block layer timer works, call scsi_times_out() to handle I/O timeout logic. At the same time the command just completes. 2) scsi_times_out() call iscsi_eh_cmd_timed_out() to processing timeout logic. although there is an NULL judgment for the task, the task has not been released yet now. 3) iscsi_complete_task() call __iscsi_put_task(), The task reference count reaches zero, the conditions for free task is met, then iscsi_free_task () free the task, and let sc->SCp.ptr = NULL. After iscsi_eh_cmd_timed_out passes the task judgment check, there may be NULL dereference scenarios later. CPU0 CPU3 |- scsi_times_out() |- iscsi_complete_task() | | |- iscsi_eh_cmd_timed_out() |- __iscsi_put_task() | | |- task=sc->SCp.ptr, task is not NUL, check passed |- iscsi_free_task(task) | | | |-> sc->SCp.ptr = NULL | | |- task is NULL now, NULL pointer dereference | | | \|/ \|/ Calltrace: [380751.840862] BUG: unable to handle kernel NULL pointer dereference at 0000000000000138 [380751.843709] PGD 0 P4D 0 [380751.844770] Oops: 0000 [#1] SMP PTI [380751.846283] CPU: 0 PID: 403 Comm: kworker/0:1H Kdump: loaded Tainted: G [380751.851467] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996) [380751.856521] Workqueue: kblockd blk_mq_timeout_work [380751.858527] RIP: 0010:iscsi_eh_cmd_timed_out+0x15e/0x2e0 [libiscsi] [380751.861129] Code: 83 ea 01 48 8d 74 d0 08 48 8b 10 48 8b 4a 50 48 85 c9 74 2c 48 39 d5 74 [380751.868811] RSP: 0018:ffffc1e280a5fd58 EFLAGS: 00010246 [380751.870978] RAX: ffff9fd1e84e15e0 RBX: ffff9fd1e84e6dd0 RCX: 0000000116acc580 [380751.873791] RDX: ffff9fd1f97a9400 RSI: ffff9fd1e84e1800 RDI: ffff9fd1e4d6d420 [380751.876059] RBP: ffff9fd1e4d49000 R08: 0000000116acc580 R09: 0000000116acc580 [380751.878284] R10: 0000000000000000 R11: 0000000000000000 R12: ffff9fd1e6e931e8 [380751.880500] R13: ffff9fd1e84e6ee0 R14: 0000000000000010 R15: 0000000000000003 [380751.882687] FS: 0000000000000000(0000) GS:ffff9fd1fac00000(0000) knlGS:0000000000000000 [380751.885236] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [380751.887059] CR2: 0000000000000138 CR3: 000000011860a001 CR4: 00000000003606f0 [380751.889308] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [380751.891523] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 [380751.893738] Call Trace: [380751.894639] scsi_times_out+0x60/0x1c0 [380751.895861] blk_mq_check_expired+0x144/0x200 [380751.897302] ? __switch_to_asm+0x35/0x70 [380751.898551] blk_mq_queue_tag_busy_iter+0x195/0x2e0 [380751.900091] ? __blk_mq_requeue_request+0x100/0x100 [380751.901611] ? __switch_to_asm+0x41/0x70 [380751.902853] ? __blk_mq_requeue_request+0x100/0x100 [380751.904398] blk_mq_timeout_work+0x54/0x130 [380751.905740] process_one_work+0x195/0x390 [380751.907228] worker_thread+0x30/0x390 [380751.908713] ? process_one_work+0x390/0x390 [380751.910350] kthread+0x10d/0x130 [380751.911470] ? kthread_flush_work_fn+0x10/0x10 [380751.913007] ret_from_fork+0x35/0x40 crash> dis -l iscsi_eh_cmd_timed_out+0x15e xxxxx/drivers/scsi/libiscsi.c: 2062 1970 enum blk_eh_timer_return iscsi_eh_cmd_timed_out(struct scsi_cmnd *sc) { ... 1984 spin_lock_bh(&session->frwd_lock); 1985 task = (struct iscsi_task *)sc->SCp.ptr; 1986 if (!task) { 1987 /* 1988 * Raced with completion. Blk layer has taken ownership 1989 * so let timeout code complete it now. 1990 */ 1991 rc = BLK_EH_DONE; 1992 goto done; 1993 } ... 2052 for (i = 0; i < conn->session->cmds_max; i++) { 2053 running_task = conn->session->cmds[i]; 2054 if (!running_task->sc || running_task == task || 2055 running_task->state != ISCSI_TASK_RUNNING) 2056 continue; 2057 2058 /* 2059 * Only check if cmds started before this one have made 2060 * progress, or this could never fail 2061 */ 2062 if (time_after(running_task->sc->jiffies_at_alloc, 2063 task->sc->jiffies_at_alloc)) <--- 2064 continue; 2065 ... } carsh> struct scsi_cmnd ffff9fd1e6e931e8 struct scsi_cmnd { ... SCp = { ptr = 0x0, <--- iscsi_task this_residual = 0, ... }, } Fixes: 3e5c28ad03 ("libiscsi: merge iscsi_mgmt_task and iscsi_cmd_task") Signed-off-by: Wu Bo <wubo40@huawei.com> --- drivers/scsi/libiscsi.c | 12 ++++++++++++ 1 file changed, 12 insertions(+)