aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/scsi/scsi_lib.c
diff options
context:
space:
mode:
authorJames Bottomley <jejb@mulgrave.(none)>2005-09-09 11:44:16 -0400
committerJames Bottomley <jejb@mulgrave.(none)>2005-09-09 11:44:16 -0400
commite91442b635be776ea205fba233bdd5bc74b62bc3 (patch)
treec04066f8d17be121244d870ab9347ca6e8c7cda3 /drivers/scsi/scsi_lib.c
parent286f3e13a1dc7f32407629fbd7aabc8ea78c62b5 (diff)
[SCSI] SCSI core: fix leakage of scsi_cmnd's
From: Alan Stern <stern@rowland.harvard.edu> This patch (as559b) adds a new routine, scsi_unprep_request, which gets called every place a request is requeued. (That includes scsi_queue_insert as well as scsi_requeue_command.) It also changes scsi_kill_requests to make it call __scsi_done with result equal to DID_NO_CONNECT << 16. (I'm not sure if it's necessary to call scsi_init_cmd_errh here; maybe you can check on that.) Finally, the patch changes the return value from scsi_end_request, to avoid returning a stale pointer in the case where the request was requeued. Fortunately the return value is used in only place, and the change actually simplified it. Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Rejections fixed up and Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Diffstat (limited to 'drivers/scsi/scsi_lib.c')
-rw-r--r--drivers/scsi/scsi_lib.c111
1 files changed, 71 insertions, 40 deletions
diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index 2ad60f1dbc63..003f8cf47cf1 100644
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -97,6 +97,30 @@ int scsi_insert_special_req(struct scsi_request *sreq, int at_head)
97} 97}
98 98
99static void scsi_run_queue(struct request_queue *q); 99static void scsi_run_queue(struct request_queue *q);
100static void scsi_release_buffers(struct scsi_cmnd *cmd);
101
102/*
103 * Function: scsi_unprep_request()
104 *
105 * Purpose: Remove all preparation done for a request, including its
106 * associated scsi_cmnd, so that it can be requeued.
107 *
108 * Arguments: req - request to unprepare
109 *
110 * Lock status: Assumed that no locks are held upon entry.
111 *
112 * Returns: Nothing.
113 */
114static void scsi_unprep_request(struct request *req)
115{
116 struct scsi_cmnd *cmd = req->special;
117
118 req->flags &= ~REQ_DONTPREP;
119 req->special = (req->flags & REQ_SPECIAL) ? cmd->sc_request : NULL;
120
121 scsi_release_buffers(cmd);
122 scsi_put_command(cmd);
123}
100 124
101/* 125/*
102 * Function: scsi_queue_insert() 126 * Function: scsi_queue_insert()
@@ -116,12 +140,14 @@ static void scsi_run_queue(struct request_queue *q);
116 * commands. 140 * commands.
117 * Notes: This could be called either from an interrupt context or a 141 * Notes: This could be called either from an interrupt context or a
118 * normal process context. 142 * normal process context.
143 * Notes: Upon return, cmd is a stale pointer.
119 */ 144 */
120int scsi_queue_insert(struct scsi_cmnd *cmd, int reason) 145int scsi_queue_insert(struct scsi_cmnd *cmd, int reason)
121{ 146{
122 struct Scsi_Host *host = cmd->device->host; 147 struct Scsi_Host *host = cmd->device->host;
123 struct scsi_device *device = cmd->device; 148 struct scsi_device *device = cmd->device;
124 struct request_queue *q = device->request_queue; 149 struct request_queue *q = device->request_queue;
150 struct request *req = cmd->request;
125 unsigned long flags; 151 unsigned long flags;
126 152
127 SCSI_LOG_MLQUEUE(1, 153 SCSI_LOG_MLQUEUE(1,
@@ -162,8 +188,9 @@ int scsi_queue_insert(struct scsi_cmnd *cmd, int reason)
162 * function. The SCSI request function detects the blocked condition 188 * function. The SCSI request function detects the blocked condition
163 * and plugs the queue appropriately. 189 * and plugs the queue appropriately.
164 */ 190 */
191 scsi_unprep_request(req);
165 spin_lock_irqsave(q->queue_lock, flags); 192 spin_lock_irqsave(q->queue_lock, flags);
166 blk_requeue_request(q, cmd->request); 193 blk_requeue_request(q, req);
167 spin_unlock_irqrestore(q->queue_lock, flags); 194 spin_unlock_irqrestore(q->queue_lock, flags);
168 195
169 scsi_run_queue(q); 196 scsi_run_queue(q);
@@ -552,15 +579,16 @@ static void scsi_run_queue(struct request_queue *q)
552 * I/O errors in the middle of the request, in which case 579 * I/O errors in the middle of the request, in which case
553 * we need to request the blocks that come after the bad 580 * we need to request the blocks that come after the bad
554 * sector. 581 * sector.
582 * Notes: Upon return, cmd is a stale pointer.
555 */ 583 */
556static void scsi_requeue_command(struct request_queue *q, struct scsi_cmnd *cmd) 584static void scsi_requeue_command(struct request_queue *q, struct scsi_cmnd *cmd)
557{ 585{
586 struct request *req = cmd->request;
558 unsigned long flags; 587 unsigned long flags;
559 588
560 cmd->request->flags &= ~REQ_DONTPREP; 589 scsi_unprep_request(req);
561
562 spin_lock_irqsave(q->queue_lock, flags); 590 spin_lock_irqsave(q->queue_lock, flags);
563 blk_requeue_request(q, cmd->request); 591 blk_requeue_request(q, req);
564 spin_unlock_irqrestore(q->queue_lock, flags); 592 spin_unlock_irqrestore(q->queue_lock, flags);
565 593
566 scsi_run_queue(q); 594 scsi_run_queue(q);
@@ -595,13 +623,14 @@ void scsi_run_host_queues(struct Scsi_Host *shost)
595 * 623 *
596 * Lock status: Assumed that lock is not held upon entry. 624 * Lock status: Assumed that lock is not held upon entry.
597 * 625 *
598 * Returns: cmd if requeue done or required, NULL otherwise 626 * Returns: cmd if requeue required, NULL otherwise.
599 * 627 *
600 * Notes: This is called for block device requests in order to 628 * Notes: This is called for block device requests in order to
601 * mark some number of sectors as complete. 629 * mark some number of sectors as complete.
602 * 630 *
603 * We are guaranteeing that the request queue will be goosed 631 * We are guaranteeing that the request queue will be goosed
604 * at some point during this call. 632 * at some point during this call.
633 * Notes: If cmd was requeued, upon return it will be a stale pointer.
605 */ 634 */
606static struct scsi_cmnd *scsi_end_request(struct scsi_cmnd *cmd, int uptodate, 635static struct scsi_cmnd *scsi_end_request(struct scsi_cmnd *cmd, int uptodate,
607 int bytes, int requeue) 636 int bytes, int requeue)
@@ -624,14 +653,15 @@ static struct scsi_cmnd *scsi_end_request(struct scsi_cmnd *cmd, int uptodate,
624 if (!uptodate && blk_noretry_request(req)) 653 if (!uptodate && blk_noretry_request(req))
625 end_that_request_chunk(req, 0, leftover); 654 end_that_request_chunk(req, 0, leftover);
626 else { 655 else {
627 if (requeue) 656 if (requeue) {
628 /* 657 /*
629 * Bleah. Leftovers again. Stick the 658 * Bleah. Leftovers again. Stick the
630 * leftovers in the front of the 659 * leftovers in the front of the
631 * queue, and goose the queue again. 660 * queue, and goose the queue again.
632 */ 661 */
633 scsi_requeue_command(q, cmd); 662 scsi_requeue_command(q, cmd);
634 663 cmd = NULL;
664 }
635 return cmd; 665 return cmd;
636 } 666 }
637 } 667 }
@@ -857,15 +887,13 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes,
857 * requeueing right here - we will requeue down below 887 * requeueing right here - we will requeue down below
858 * when we handle the bad sectors. 888 * when we handle the bad sectors.
859 */ 889 */
860 cmd = scsi_end_request(cmd, 1, good_bytes, result == 0);
861 890
862 /* 891 /*
863 * If the command completed without error, then either finish off the 892 * If the command completed without error, then either
864 * rest of the command, or start a new one. 893 * finish off the rest of the command, or start a new one.
865 */ 894 */
866 if (result == 0 || cmd == NULL ) { 895 if (scsi_end_request(cmd, 1, good_bytes, result == 0) == NULL)
867 return; 896 return;
868 }
869 } 897 }
870 /* 898 /*
871 * Now, if we were good little boys and girls, Santa left us a request 899 * Now, if we were good little boys and girls, Santa left us a request
@@ -880,7 +908,7 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes,
880 * and quietly refuse further access. 908 * and quietly refuse further access.
881 */ 909 */
882 cmd->device->changed = 1; 910 cmd->device->changed = 1;
883 cmd = scsi_end_request(cmd, 0, 911 scsi_end_request(cmd, 0,
884 this_count, 1); 912 this_count, 1);
885 return; 913 return;
886 } else { 914 } else {
@@ -914,7 +942,7 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes,
914 scsi_requeue_command(q, cmd); 942 scsi_requeue_command(q, cmd);
915 result = 0; 943 result = 0;
916 } else { 944 } else {
917 cmd = scsi_end_request(cmd, 0, this_count, 1); 945 scsi_end_request(cmd, 0, this_count, 1);
918 return; 946 return;
919 } 947 }
920 break; 948 break;
@@ -931,7 +959,7 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes,
931 dev_printk(KERN_INFO, 959 dev_printk(KERN_INFO,
932 &cmd->device->sdev_gendev, 960 &cmd->device->sdev_gendev,
933 "Device not ready.\n"); 961 "Device not ready.\n");
934 cmd = scsi_end_request(cmd, 0, this_count, 1); 962 scsi_end_request(cmd, 0, this_count, 1);
935 return; 963 return;
936 case VOLUME_OVERFLOW: 964 case VOLUME_OVERFLOW:
937 if (!(req->flags & REQ_QUIET)) { 965 if (!(req->flags & REQ_QUIET)) {
@@ -941,7 +969,7 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes,
941 __scsi_print_command(cmd->data_cmnd); 969 __scsi_print_command(cmd->data_cmnd);
942 scsi_print_sense("", cmd); 970 scsi_print_sense("", cmd);
943 } 971 }
944 cmd = scsi_end_request(cmd, 0, block_bytes, 1); 972 scsi_end_request(cmd, 0, block_bytes, 1);
945 return; 973 return;
946 default: 974 default:
947 break; 975 break;
@@ -972,7 +1000,7 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes,
972 block_bytes = req->hard_cur_sectors << 9; 1000 block_bytes = req->hard_cur_sectors << 9;
973 if (!block_bytes) 1001 if (!block_bytes)
974 block_bytes = req->data_len; 1002 block_bytes = req->data_len;
975 cmd = scsi_end_request(cmd, 0, block_bytes, 1); 1003 scsi_end_request(cmd, 0, block_bytes, 1);
976 } 1004 }
977} 1005}
978EXPORT_SYMBOL(scsi_io_completion); 1006EXPORT_SYMBOL(scsi_io_completion);
@@ -1336,19 +1364,24 @@ static inline int scsi_host_queue_ready(struct request_queue *q,
1336} 1364}
1337 1365
1338/* 1366/*
1339 * Kill requests for a dead device 1367 * Kill a request for a dead device
1340 */ 1368 */
1341static void scsi_kill_requests(request_queue_t *q) 1369static void scsi_kill_request(struct request *req, request_queue_t *q)
1342{ 1370{
1343 struct request *req; 1371 struct scsi_cmnd *cmd = req->special;
1344 1372
1345 while ((req = elv_next_request(q)) != NULL) { 1373 spin_unlock(q->queue_lock);
1346 blkdev_dequeue_request(req); 1374 if (unlikely(cmd == NULL)) {
1347 req->flags |= REQ_QUIET; 1375 printk(KERN_CRIT "impossible request in %s.\n",
1348 while (end_that_request_first(req, 0, req->nr_sectors)) 1376 __FUNCTION__);
1349 ; 1377 BUG();
1350 end_that_request_last(req);
1351 } 1378 }
1379
1380 scsi_init_cmd_errh(cmd);
1381 cmd->result = DID_NO_CONNECT << 16;
1382 atomic_inc(&cmd->device->iorequest_cnt);
1383 __scsi_done(cmd);
1384 spin_lock(q->queue_lock);
1352} 1385}
1353 1386
1354/* 1387/*
@@ -1371,7 +1404,8 @@ static void scsi_request_fn(struct request_queue *q)
1371 1404
1372 if (!sdev) { 1405 if (!sdev) {
1373 printk("scsi: killing requests for dead queue\n"); 1406 printk("scsi: killing requests for dead queue\n");
1374 scsi_kill_requests(q); 1407 while ((req = elv_next_request(q)) != NULL)
1408 scsi_kill_request(req, q);
1375 return; 1409 return;
1376 } 1410 }
1377 1411
@@ -1399,10 +1433,7 @@ static void scsi_request_fn(struct request_queue *q)
1399 printk(KERN_ERR "scsi%d (%d:%d): rejecting I/O to offline device\n", 1433 printk(KERN_ERR "scsi%d (%d:%d): rejecting I/O to offline device\n",
1400 sdev->host->host_no, sdev->id, sdev->lun); 1434 sdev->host->host_no, sdev->id, sdev->lun);
1401 blkdev_dequeue_request(req); 1435 blkdev_dequeue_request(req);
1402 req->flags |= REQ_QUIET; 1436 scsi_kill_request(req, q);
1403 while (end_that_request_first(req, 0, req->nr_sectors))
1404 ;
1405 end_that_request_last(req);
1406 continue; 1437 continue;
1407 } 1438 }
1408 1439
@@ -1415,6 +1446,14 @@ static void scsi_request_fn(struct request_queue *q)
1415 sdev->device_busy++; 1446 sdev->device_busy++;
1416 1447
1417 spin_unlock(q->queue_lock); 1448 spin_unlock(q->queue_lock);
1449 cmd = req->special;
1450 if (unlikely(cmd == NULL)) {
1451 printk(KERN_CRIT "impossible request in %s.\n"
1452 "please mail a stack trace to "
1453 "linux-scsi@vger.kernel.org",
1454 __FUNCTION__);
1455 BUG();
1456 }
1418 spin_lock(shost->host_lock); 1457 spin_lock(shost->host_lock);
1419 1458
1420 if (!scsi_host_queue_ready(q, shost, sdev)) 1459 if (!scsi_host_queue_ready(q, shost, sdev))
@@ -1433,15 +1472,6 @@ static void scsi_request_fn(struct request_queue *q)
1433 */ 1472 */
1434 spin_unlock_irq(shost->host_lock); 1473 spin_unlock_irq(shost->host_lock);
1435 1474
1436 cmd = req->special;
1437 if (unlikely(cmd == NULL)) {
1438 printk(KERN_CRIT "impossible request in %s.\n"
1439 "please mail a stack trace to "
1440 "linux-scsi@vger.kernel.org",
1441 __FUNCTION__);
1442 BUG();
1443 }
1444
1445 /* 1475 /*
1446 * Finally, initialize any error handling parameters, and set up 1476 * Finally, initialize any error handling parameters, and set up
1447 * the timers for timeouts. 1477 * the timers for timeouts.
@@ -1477,6 +1507,7 @@ static void scsi_request_fn(struct request_queue *q)
1477 * cases (host limits or settings) should run the queue at some 1507 * cases (host limits or settings) should run the queue at some
1478 * later time. 1508 * later time.
1479 */ 1509 */
1510 scsi_unprep_request(req);
1480 spin_lock_irq(q->queue_lock); 1511 spin_lock_irq(q->queue_lock);
1481 blk_requeue_request(q, req); 1512 blk_requeue_request(q, req);
1482 sdev->device_busy--; 1513 sdev->device_busy--;