From af17092810a887178195276255b7b31f8fbe7dbe Mon Sep 17 00:00:00 2001 From: Chris Leech Date: Mon, 9 Apr 2018 15:15:28 -0700 Subject: [PATCH 01/10] scsi: iscsi: respond to netlink with unicast when appropriate Instead of always multicasting responses, send a unicast netlink message directed at the correct pid. This will be needed if we ever want to support multiple userspace processes interacting with the kernel over iSCSI netlink simultaneously. Limitations can currently be seen if you attempt to run multiple iscsistart commands in parallel. We've fixed up the userspace issues in iscsistart that prevented multiple instances from running, so now attempts to speed up booting by bringing up multiple iscsi sessions at once in the initramfs are just running into misrouted responses that this fixes. Signed-off-by: Chris Leech Reviewed-by: Lee Duncan Signed-off-by: Martin K. Petersen --- drivers/scsi/scsi_transport_iscsi.c | 29 ++++++++++++++++++----------- 1 file changed, 18 insertions(+), 11 deletions(-) diff --git a/drivers/scsi/scsi_transport_iscsi.c b/drivers/scsi/scsi_transport_iscsi.c index f4b52b44b966..65f6c94f2e9b 100644 --- a/drivers/scsi/scsi_transport_iscsi.c +++ b/drivers/scsi/scsi_transport_iscsi.c @@ -2322,6 +2322,12 @@ iscsi_multicast_skb(struct sk_buff *skb, uint32_t group, gfp_t gfp) return nlmsg_multicast(nls, skb, 0, group, gfp); } +static int +iscsi_unicast_skb(struct sk_buff *skb, u32 portid) +{ + return nlmsg_unicast(nls, skb, portid); +} + int iscsi_recv_pdu(struct iscsi_cls_conn *conn, struct iscsi_hdr *hdr, char *data, uint32_t data_size) { @@ -2524,14 +2530,11 @@ void iscsi_ping_comp_event(uint32_t host_no, struct iscsi_transport *transport, EXPORT_SYMBOL_GPL(iscsi_ping_comp_event); static int -iscsi_if_send_reply(uint32_t group, int seq, int type, int done, int multi, - void *payload, int size) +iscsi_if_send_reply(u32 portid, int type, void *payload, int size) { struct sk_buff *skb; struct nlmsghdr *nlh; int len = nlmsg_total_size(size); - int flags = multi ? NLM_F_MULTI : 0; - int t = done ? NLMSG_DONE : type; skb = alloc_skb(len, GFP_ATOMIC); if (!skb) { @@ -2539,10 +2542,9 @@ iscsi_if_send_reply(uint32_t group, int seq, int type, int done, int multi, return -ENOMEM; } - nlh = __nlmsg_put(skb, 0, 0, t, (len - sizeof(*nlh)), 0); - nlh->nlmsg_flags = flags; + nlh = __nlmsg_put(skb, 0, 0, type, (len - sizeof(*nlh)), 0); memcpy(nlmsg_data(nlh), payload, size); - return iscsi_multicast_skb(skb, group, GFP_ATOMIC); + return iscsi_unicast_skb(skb, portid); } static int @@ -3470,6 +3472,7 @@ static int iscsi_if_recv_msg(struct sk_buff *skb, struct nlmsghdr *nlh, uint32_t *group) { int err = 0; + u32 portid; struct iscsi_uevent *ev = nlmsg_data(nlh); struct iscsi_transport *transport = NULL; struct iscsi_internal *priv; @@ -3490,10 +3493,12 @@ iscsi_if_recv_msg(struct sk_buff *skb, struct nlmsghdr *nlh, uint32_t *group) if (!try_module_get(transport->owner)) return -EINVAL; + portid = NETLINK_CB(skb).portid; + switch (nlh->nlmsg_type) { case ISCSI_UEVENT_CREATE_SESSION: err = iscsi_if_create_session(priv, ep, ev, - NETLINK_CB(skb).portid, + portid, ev->u.c_session.initial_cmdsn, ev->u.c_session.cmds_max, ev->u.c_session.queue_depth); @@ -3506,7 +3511,7 @@ iscsi_if_recv_msg(struct sk_buff *skb, struct nlmsghdr *nlh, uint32_t *group) } err = iscsi_if_create_session(priv, ep, ev, - NETLINK_CB(skb).portid, + portid, ev->u.c_bound_session.initial_cmdsn, ev->u.c_bound_session.cmds_max, ev->u.c_bound_session.queue_depth); @@ -3664,6 +3669,8 @@ iscsi_if_recv_msg(struct sk_buff *skb, struct nlmsghdr *nlh, uint32_t *group) static void iscsi_if_rx(struct sk_buff *skb) { + u32 portid = NETLINK_CB(skb).portid; + mutex_lock(&rx_queue_mutex); while (skb->len >= NLMSG_HDRLEN) { int err; @@ -3699,8 +3706,8 @@ iscsi_if_rx(struct sk_buff *skb) break; if (ev->type == ISCSI_UEVENT_GET_CHAP && !err) break; - err = iscsi_if_send_reply(group, nlh->nlmsg_seq, - nlh->nlmsg_type, 0, 0, ev, sizeof(*ev)); + err = iscsi_if_send_reply(portid, nlh->nlmsg_type, + ev, sizeof(*ev)); } while (err < 0 && err != -ECONNREFUSED && err != -ESRCH); skb_pull(skb, rlen); } From 4f2c8bf6bdff306e8f918fa4bacf636c6b9ef760 Mon Sep 17 00:00:00 2001 From: Douglas Gilbert Date: Tue, 10 Apr 2018 13:00:36 -0400 Subject: [PATCH 02/10] scsi: scsi_debug: IMMED related delay adjustments A patch titled: "[PATCH v2] scsi_debug: implement IMMED bit" introduced long delays to the Start stop unit (SSU) and Synchronize cache (SC) commands when the IMMED bit is clear. This patch makes those delays more realistic. It causes SSU to only delay when the start stop state is changed; SC only delays when there's been a write since the previous SC. It also reduced the SC delay from 1 second to 50 milliseconds. Signed-off-by: Douglas Gilbert Tested-by: Ming Lei Reported-by: Ming Lei Reviewed-by: Johannes Thumshirn Signed-off-by: Martin K. Petersen --- drivers/scsi/scsi_debug.c | 33 ++++++++++++++++++++++++--------- 1 file changed, 24 insertions(+), 9 deletions(-) diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c index 9ef5e3b810f6..656c98e116a9 100644 --- a/drivers/scsi/scsi_debug.c +++ b/drivers/scsi/scsi_debug.c @@ -234,11 +234,13 @@ static const char *sdebug_version_date = "20180128"; #define F_INV_OP 0x200 #define F_FAKE_RW 0x400 #define F_M_ACCESS 0x800 /* media access */ -#define F_LONG_DELAY 0x1000 +#define F_SSU_DELAY 0x1000 +#define F_SYNC_DELAY 0x2000 #define FF_RESPOND (F_RL_WLUN_OK | F_SKIP_UA | F_DELAY_OVERR) #define FF_MEDIA_IO (F_M_ACCESS | F_FAKE_RW) #define FF_SA (F_SA_HIGH | F_SA_LOW) +#define F_LONG_DELAY (F_SSU_DELAY | F_SYNC_DELAY) #define SDEBUG_MAX_PARTS 4 @@ -510,7 +512,7 @@ static const struct opcode_info_t release_iarr[] = { }; static const struct opcode_info_t sync_cache_iarr[] = { - {0, 0x91, 0, F_LONG_DELAY | F_M_ACCESS, resp_sync_cache, NULL, + {0, 0x91, 0, F_SYNC_DELAY | F_M_ACCESS, resp_sync_cache, NULL, {16, 0x6, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x3f, 0xc7} }, /* SYNC_CACHE (16) */ }; @@ -553,7 +555,7 @@ static const struct opcode_info_t opcode_info_arr[SDEB_I_LAST_ELEMENT + 1] = { resp_write_dt0, write_iarr, /* WRITE(16) */ {16, 0xfa, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xc7} }, - {0, 0x1b, 0, F_LONG_DELAY, resp_start_stop, NULL,/* START STOP UNIT */ + {0, 0x1b, 0, F_SSU_DELAY, resp_start_stop, NULL,/* START STOP UNIT */ {6, 0x1, 0, 0xf, 0xf7, 0xc7, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0} }, {ARRAY_SIZE(sa_in_16_iarr), 0x9e, 0x10, F_SA_LOW | F_D_IN, resp_readcap16, sa_in_16_iarr, /* SA_IN(16), READ CAPACITY(16) */ @@ -606,7 +608,7 @@ static const struct opcode_info_t opcode_info_arr[SDEB_I_LAST_ELEMENT + 1] = { resp_write_same_10, write_same_iarr, /* WRITE SAME(10) */ {10, 0xff, 0xff, 0xff, 0xff, 0xff, 0x3f, 0xff, 0xff, 0xc7, 0, 0, 0, 0, 0, 0} }, - {ARRAY_SIZE(sync_cache_iarr), 0x35, 0, F_LONG_DELAY | F_M_ACCESS, + {ARRAY_SIZE(sync_cache_iarr), 0x35, 0, F_SYNC_DELAY | F_M_ACCESS, resp_sync_cache, sync_cache_iarr, {10, 0x7, 0xff, 0xff, 0xff, 0xff, 0x3f, 0xff, 0xff, 0xc7, 0, 0, 0, 0, 0, 0} }, /* SYNC_CACHE (10) */ @@ -667,6 +669,7 @@ static bool sdebug_strict = DEF_STRICT; static bool sdebug_any_injecting_opt; static bool sdebug_verbose; static bool have_dif_prot; +static bool write_since_sync; static bool sdebug_statistics = DEF_STATISTICS; static unsigned int sdebug_store_sectors; @@ -1607,6 +1610,7 @@ static int resp_start_stop(struct scsi_cmnd *scp, { unsigned char *cmd = scp->cmnd; int power_cond, stop; + bool changing; power_cond = (cmd[4] & 0xf0) >> 4; if (power_cond) { @@ -1614,8 +1618,12 @@ static int resp_start_stop(struct scsi_cmnd *scp, return check_condition_result; } stop = !(cmd[4] & 1); + changing = atomic_read(&devip->stopped) == !stop; atomic_xchg(&devip->stopped, stop); - return (cmd[1] & 0x1) ? SDEG_RES_IMMED_MASK : 0; /* check IMMED bit */ + if (!changing || cmd[1] & 0x1) /* state unchanged or IMMED set */ + return SDEG_RES_IMMED_MASK; + else + return 0; } static sector_t get_sdebug_capacity(void) @@ -2473,6 +2481,7 @@ static int do_device_access(struct scsi_cmnd *scmd, u32 sg_skip, u64 lba, if (do_write) { sdb = scsi_out(scmd); dir = DMA_TO_DEVICE; + write_since_sync = true; } else { sdb = scsi_in(scmd); dir = DMA_FROM_DEVICE; @@ -3583,6 +3592,7 @@ static int resp_get_lba_status(struct scsi_cmnd *scp, static int resp_sync_cache(struct scsi_cmnd *scp, struct sdebug_dev_info *devip) { + int res = 0; u64 lba; u32 num_blocks; u8 *cmd = scp->cmnd; @@ -3598,7 +3608,11 @@ static int resp_sync_cache(struct scsi_cmnd *scp, mk_sense_buffer(scp, ILLEGAL_REQUEST, LBA_OUT_OF_RANGE, 0); return check_condition_result; } - return (cmd[1] & 0x2) ? SDEG_RES_IMMED_MASK : 0; /* check IMMED bit */ + if (!write_since_sync || cmd[1] & 0x2) + res = SDEG_RES_IMMED_MASK; + else /* delay if write_since_sync and IMMED clear */ + write_since_sync = false; + return res; } #define RL_BUCKET_ELEMS 8 @@ -5777,13 +5791,14 @@ fini: return schedule_resp(scp, devip, errsts, pfp, 0, 0); else if ((sdebug_jdelay || sdebug_ndelay) && (flags & F_LONG_DELAY)) { /* - * If any delay is active, want F_LONG_DELAY to be at least 1 + * If any delay is active, for F_SSU_DELAY want at least 1 * second and if sdebug_jdelay>0 want a long delay of that - * many seconds. + * many seconds; for F_SYNC_DELAY want 1/20 of that. */ int jdelay = (sdebug_jdelay < 2) ? 1 : sdebug_jdelay; + int denom = (flags & F_SYNC_DELAY) ? 20 : 1; - jdelay = mult_frac(USER_HZ * jdelay, HZ, USER_HZ); + jdelay = mult_frac(USER_HZ * jdelay, HZ, denom * USER_HZ); return schedule_resp(scp, devip, errsts, pfp, jdelay, 0); } else return schedule_resp(scp, devip, errsts, pfp, sdebug_jdelay, From 0cfce53a7880eb949e55fcfbe0d82c613f056fe0 Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Wed, 11 Apr 2018 08:36:49 +0100 Subject: [PATCH 03/10] scsi: fnic: fix spelling mistake in fnic stats "Abord" -> "Abort" Trivial fix to spelling mistake in fnic stats message text. Signed-off-by: Colin Ian King Signed-off-by: Martin K. Petersen --- drivers/scsi/fnic/fnic_trace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/fnic/fnic_trace.c b/drivers/scsi/fnic/fnic_trace.c index abddde11982b..98597b59c12a 100644 --- a/drivers/scsi/fnic/fnic_trace.c +++ b/drivers/scsi/fnic/fnic_trace.c @@ -296,7 +296,7 @@ int fnic_get_stats_data(struct stats_debug_info *debug, "Number of Abort FW Timeouts: %lld\n" "Number of Abort IO NOT Found: %lld\n" - "Abord issued times: \n" + "Abort issued times: \n" " < 6 sec : %lld\n" " 6 sec - 20 sec : %lld\n" " 20 sec - 30 sec : %lld\n" From 94e5395d2403c8bc2504a7cbe4c4caaacb7b8b84 Mon Sep 17 00:00:00 2001 From: "Martin K. Petersen" Date: Wed, 18 Apr 2018 22:54:59 -0400 Subject: [PATCH 04/10] scsi: mptsas: Disable WRITE SAME First generation MPT Fusion controllers can not translate WRITE SAME when the attached device is a SATA drive. Disable WRITE SAME support. Reported-by: Nikola Ciprich Cc: Signed-off-by: Martin K. Petersen --- drivers/message/fusion/mptsas.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/message/fusion/mptsas.c b/drivers/message/fusion/mptsas.c index 231f3a1e27bf..86503f60468f 100644 --- a/drivers/message/fusion/mptsas.c +++ b/drivers/message/fusion/mptsas.c @@ -1994,6 +1994,7 @@ static struct scsi_host_template mptsas_driver_template = { .cmd_per_lun = 7, .use_clustering = ENABLE_CLUSTERING, .shost_attrs = mptscsih_host_attrs, + .no_write_same = 1, }; static int mptsas_get_linkerrors(struct sas_phy *phy) From 4d79e7d0f0cfe22c05c21a1008fa000f153119aa Mon Sep 17 00:00:00 2001 From: John Pittman Date: Sat, 24 Mar 2018 09:30:48 -0400 Subject: [PATCH 05/10] scsi: core: remove reference to scsi_show_extd_sense() In commit 2104551969e8 ("scsi: use per-cpu buffer for formatting sense"), function scsi_show_extd_sense() was removed, switching use over to scsi_format_extd_sense(). Remove last reference to scsi_show_extd_sense() in include/scsi/scsi_dbg.h. Signed-off-by: John Pittman Reviewed-by: Bart Van Assche Signed-off-by: Martin K. Petersen --- include/scsi/scsi_dbg.h | 2 -- 1 file changed, 2 deletions(-) diff --git a/include/scsi/scsi_dbg.h b/include/scsi/scsi_dbg.h index 04e0679767f6..e03bd9d41fa8 100644 --- a/include/scsi/scsi_dbg.h +++ b/include/scsi/scsi_dbg.h @@ -11,8 +11,6 @@ struct scsi_sense_hdr; extern void scsi_print_command(struct scsi_cmnd *); extern size_t __scsi_format_command(char *, size_t, const unsigned char *, size_t); -extern void scsi_show_extd_sense(const struct scsi_device *, const char *, - unsigned char, unsigned char); extern void scsi_print_sense_hdr(const struct scsi_device *, const char *, const struct scsi_sense_hdr *); extern void scsi_print_sense(const struct scsi_cmnd *); From 6667e6d91c88a788d5ab1b1e999d245473e84a90 Mon Sep 17 00:00:00 2001 From: Ohad Sharabi Date: Wed, 28 Mar 2018 12:42:18 +0300 Subject: [PATCH 06/10] scsi: ufs: add trace event for ufs upiu Add UFS Protocol Information Units(upiu) trace events for ufs driver, used to trace various ufs transaction types- command, task-management and device management. The trace-point format is generic and can be easily adapted to trace other upius if needed. Currently tracing ufs transaction of type 'device management', which this patch introduce, cannot be obtained from any other trace. Device management transactions are used for communication with the device such as reading and writing descriptor or attributes etc. Signed-off-by: Ohad Sharabi Reviewed-by: Stanislav Nijnikov Reviewed-by: Bart Van Assche Signed-off-by: Martin K. Petersen --- drivers/scsi/ufs/ufshcd.c | 40 ++++++++++++++++++++++++++++++++++++++ include/trace/events/ufs.h | 27 +++++++++++++++++++++++++ 2 files changed, 67 insertions(+) diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c index c5b1bf1cadcb..00e79057f870 100644 --- a/drivers/scsi/ufs/ufshcd.c +++ b/drivers/scsi/ufs/ufshcd.c @@ -276,6 +276,35 @@ static inline void ufshcd_remove_non_printable(char *val) *val = ' '; } +static void ufshcd_add_cmd_upiu_trace(struct ufs_hba *hba, unsigned int tag, + const char *str) +{ + struct utp_upiu_req *rq = hba->lrb[tag].ucd_req_ptr; + + trace_ufshcd_upiu(dev_name(hba->dev), str, &rq->header, &rq->sc.cdb); +} + +static void ufshcd_add_query_upiu_trace(struct ufs_hba *hba, unsigned int tag, + const char *str) +{ + struct utp_upiu_req *rq = hba->lrb[tag].ucd_req_ptr; + + trace_ufshcd_upiu(dev_name(hba->dev), str, &rq->header, &rq->qr); +} + +static void ufshcd_add_tm_upiu_trace(struct ufs_hba *hba, unsigned int tag, + const char *str) +{ + struct utp_task_req_desc *descp; + struct utp_upiu_task_req *task_req; + int off = (int)tag - hba->nutrs; + + descp = &hba->utmrdl_base_addr[off]; + task_req = (struct utp_upiu_task_req *)descp->task_req_upiu; + trace_ufshcd_upiu(dev_name(hba->dev), str, &task_req->header, + &task_req->input_param1); +} + static void ufshcd_add_command_trace(struct ufs_hba *hba, unsigned int tag, const char *str) { @@ -285,6 +314,9 @@ static void ufshcd_add_command_trace(struct ufs_hba *hba, struct ufshcd_lrb *lrbp; int transfer_len = -1; + /* trace UPIU also */ + ufshcd_add_cmd_upiu_trace(hba, tag, str); + if (!trace_ufshcd_command_enabled()) return; @@ -2550,6 +2582,7 @@ static int ufshcd_exec_dev_cmd(struct ufs_hba *hba, hba->dev_cmd.complete = &wait; + ufshcd_add_query_upiu_trace(hba, tag, "query_send"); /* Make sure descriptors are ready before ringing the doorbell */ wmb(); spin_lock_irqsave(hba->host->host_lock, flags); @@ -2559,6 +2592,9 @@ static int ufshcd_exec_dev_cmd(struct ufs_hba *hba, err = ufshcd_wait_for_dev_cmd(hba, lrbp, timeout); + ufshcd_add_query_upiu_trace(hba, tag, + err ? "query_complete_err" : "query_complete"); + out_put_tag: ufshcd_put_dev_cmd_tag(hba, tag); wake_up(&hba->dev_cmd.tag_wq); @@ -5443,11 +5479,14 @@ static int ufshcd_issue_tm_cmd(struct ufs_hba *hba, int lun_id, int task_id, spin_unlock_irqrestore(host->host_lock, flags); + ufshcd_add_tm_upiu_trace(hba, task_tag, "tm_send"); + /* wait until the task management command is completed */ err = wait_event_timeout(hba->tm_wq, test_bit(free_slot, &hba->tm_condition), msecs_to_jiffies(TM_CMD_TIMEOUT)); if (!err) { + ufshcd_add_tm_upiu_trace(hba, task_tag, "tm_complete_err"); dev_err(hba->dev, "%s: task management cmd 0x%.2x timed-out\n", __func__, tm_function); if (ufshcd_clear_tm_cmd(hba, free_slot)) @@ -5456,6 +5495,7 @@ static int ufshcd_issue_tm_cmd(struct ufs_hba *hba, int lun_id, int task_id, err = -ETIMEDOUT; } else { err = ufshcd_task_req_compl(hba, free_slot, tm_response); + ufshcd_add_tm_upiu_trace(hba, task_tag, "tm_complete"); } clear_bit(free_slot, &hba->tm_condition); diff --git a/include/trace/events/ufs.h b/include/trace/events/ufs.h index bf6f82673492..f8260e5c79ad 100644 --- a/include/trace/events/ufs.h +++ b/include/trace/events/ufs.h @@ -257,6 +257,33 @@ TRACE_EVENT(ufshcd_command, ) ); +TRACE_EVENT(ufshcd_upiu, + TP_PROTO(const char *dev_name, const char *str, void *hdr, void *tsf), + + TP_ARGS(dev_name, str, hdr, tsf), + + TP_STRUCT__entry( + __string(dev_name, dev_name) + __string(str, str) + __array(unsigned char, hdr, 12) + __array(unsigned char, tsf, 16) + ), + + TP_fast_assign( + __assign_str(dev_name, dev_name); + __assign_str(str, str); + memcpy(__entry->hdr, hdr, sizeof(__entry->hdr)); + memcpy(__entry->tsf, tsf, sizeof(__entry->tsf)); + ), + + TP_printk( + "%s: %s: HDR:%s, CDB:%s", + __get_str(str), __get_str(dev_name), + __print_hex(__entry->hdr, sizeof(__entry->hdr)), + __print_hex(__entry->tsf, sizeof(__entry->tsf)) + ) +); + #endif /* if !defined(_TRACE_UFS_H) || defined(TRACE_HEADER_MULTI_READ) */ /* This part must be outside protection */ From fb1633d56b0025233ed3dc49b44544748d509d9d Mon Sep 17 00:00:00 2001 From: Vinson Lee Date: Wed, 21 Mar 2018 21:04:12 +0000 Subject: [PATCH 07/10] scsi: megaraid_sas: Do not log an error if FW successfully initializes. Fixes: 2d2c2331673c ("scsi: megaraid_sas: modified few prints in OCR and IOC INIT path") Signed-off-by: Vinson Lee Acked-by: Shivasharan S Signed-off-by: Martin K. Petersen --- drivers/scsi/megaraid/megaraid_sas_fusion.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/megaraid/megaraid_sas_fusion.c b/drivers/scsi/megaraid/megaraid_sas_fusion.c index ce97cde3b41c..f4d988dd1e9d 100644 --- a/drivers/scsi/megaraid/megaraid_sas_fusion.c +++ b/drivers/scsi/megaraid/megaraid_sas_fusion.c @@ -1124,12 +1124,12 @@ megasas_ioc_init_fusion(struct megasas_instance *instance) goto fail_fw_init; } - ret = 0; + return 0; fail_fw_init: dev_err(&instance->pdev->dev, - "Init cmd return status %s for SCSI host %d\n", - ret ? "FAILED" : "SUCCESS", instance->host->host_no); + "Init cmd return status FAILED for SCSI host %d\n", + instance->host->host_no); return ret; } From 505aa4b6a8834a2300971c5220c380c3271ebde3 Mon Sep 17 00:00:00 2001 From: Mahesh Rajashekhara Date: Tue, 17 Apr 2018 17:03:12 +0530 Subject: [PATCH 08/10] scsi: sd: Defer spinning up drive while SANITIZE is in progress A drive being sanitized will return NOT READY / ASC 0x4 / ASCQ 0x1b ("LOGICAL UNIT NOT READY. SANITIZE IN PROGRESS"). Prevent spinning up the drive until this condition clears. [mkp: tweaked commit message] Signed-off-by: Mahesh Rajashekhara Cc: Signed-off-by: Martin K. Petersen --- drivers/scsi/sd.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c index a6201e696ab9..9421d9877730 100644 --- a/drivers/scsi/sd.c +++ b/drivers/scsi/sd.c @@ -2121,6 +2121,8 @@ sd_spinup_disk(struct scsi_disk *sdkp) break; /* standby */ if (sshdr.asc == 4 && sshdr.ascq == 0xc) break; /* unavailable */ + if (sshdr.asc == 4 && sshdr.ascq == 0x1b) + break; /* sanitize in progress */ /* * Issue command to spin up drive when not ready */ From ccce20fc7968d546fb1e8e147bf5cdc8afc4278a Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Mon, 16 Apr 2018 18:04:41 -0700 Subject: [PATCH 09/10] scsi: sd_zbc: Avoid that resetting a zone fails sporadically Since SCSI scanning occurs asynchronously, since sd_revalidate_disk() is called from sd_probe_async() and since sd_revalidate_disk() calls sd_zbc_read_zones() it can happen that sd_zbc_read_zones() is called concurrently with blkdev_report_zones() and/or blkdev_reset_zones(). That can cause these functions to fail with -EIO because sd_zbc_read_zones() e.g. sets q->nr_zones to zero before restoring it to the actual value, even if no drive characteristics have changed. Avoid that this can happen by making the following changes: - Protect the code that updates zone information with blk_queue_enter() and blk_queue_exit(). - Modify sd_zbc_setup_seq_zones_bitmap() and sd_zbc_setup() such that these functions do not modify struct scsi_disk before all zone information has been obtained. Note: since commit 055f6e18e08f ("block: Make q_usage_counter also track legacy requests"; kernel v4.15) the request queue freezing mechanism also affects legacy request queues. Fixes: 89d947561077 ("sd: Implement support for ZBC devices") Signed-off-by: Bart Van Assche Cc: Jens Axboe Cc: Damien Le Moal Cc: Christoph Hellwig Cc: Hannes Reinecke Cc: stable@vger.kernel.org # v4.16 Reviewed-by: Damien Le Moal Signed-off-by: Martin K. Petersen --- drivers/scsi/sd_zbc.c | 136 ++++++++++++++++++++++++----------------- include/linux/blkdev.h | 5 ++ 2 files changed, 85 insertions(+), 56 deletions(-) diff --git a/drivers/scsi/sd_zbc.c b/drivers/scsi/sd_zbc.c index 41df75eea57b..210407cd2341 100644 --- a/drivers/scsi/sd_zbc.c +++ b/drivers/scsi/sd_zbc.c @@ -400,8 +400,10 @@ static int sd_zbc_check_capacity(struct scsi_disk *sdkp, unsigned char *buf) * * Check that all zones of the device are equal. The last zone can however * be smaller. The zone size must also be a power of two number of LBAs. + * + * Returns the zone size in bytes upon success or an error code upon failure. */ -static int sd_zbc_check_zone_size(struct scsi_disk *sdkp) +static s64 sd_zbc_check_zone_size(struct scsi_disk *sdkp) { u64 zone_blocks = 0; sector_t block = 0; @@ -412,8 +414,6 @@ static int sd_zbc_check_zone_size(struct scsi_disk *sdkp) int ret; u8 same; - sdkp->zone_blocks = 0; - /* Get a buffer */ buf = kmalloc(SD_ZBC_BUF_SIZE, GFP_KERNEL); if (!buf) @@ -445,16 +445,17 @@ static int sd_zbc_check_zone_size(struct scsi_disk *sdkp) /* Parse zone descriptors */ while (rec < buf + buf_len) { - zone_blocks = get_unaligned_be64(&rec[8]); - if (sdkp->zone_blocks == 0) { - sdkp->zone_blocks = zone_blocks; - } else if (zone_blocks != sdkp->zone_blocks && - (block + zone_blocks < sdkp->capacity - || zone_blocks > sdkp->zone_blocks)) { - zone_blocks = 0; + u64 this_zone_blocks = get_unaligned_be64(&rec[8]); + + if (zone_blocks == 0) { + zone_blocks = this_zone_blocks; + } else if (this_zone_blocks != zone_blocks && + (block + this_zone_blocks < sdkp->capacity + || this_zone_blocks > zone_blocks)) { + this_zone_blocks = 0; goto out; } - block += zone_blocks; + block += this_zone_blocks; rec += 64; } @@ -467,8 +468,6 @@ static int sd_zbc_check_zone_size(struct scsi_disk *sdkp) } while (block < sdkp->capacity); - zone_blocks = sdkp->zone_blocks; - out: if (!zone_blocks) { if (sdkp->first_scan) @@ -488,8 +487,7 @@ out: "Zone size too large\n"); ret = -ENODEV; } else { - sdkp->zone_blocks = zone_blocks; - sdkp->zone_shift = ilog2(zone_blocks); + ret = zone_blocks; } out_free: @@ -500,15 +498,14 @@ out_free: /** * sd_zbc_alloc_zone_bitmap - Allocate a zone bitmap (one bit per zone). - * @sdkp: The disk of the bitmap + * @nr_zones: Number of zones to allocate space for. + * @numa_node: NUMA node to allocate the memory from. */ -static inline unsigned long *sd_zbc_alloc_zone_bitmap(struct scsi_disk *sdkp) +static inline unsigned long * +sd_zbc_alloc_zone_bitmap(u32 nr_zones, int numa_node) { - struct request_queue *q = sdkp->disk->queue; - - return kzalloc_node(BITS_TO_LONGS(sdkp->nr_zones) - * sizeof(unsigned long), - GFP_KERNEL, q->node); + return kzalloc_node(BITS_TO_LONGS(nr_zones) * sizeof(unsigned long), + GFP_KERNEL, numa_node); } /** @@ -516,6 +513,7 @@ static inline unsigned long *sd_zbc_alloc_zone_bitmap(struct scsi_disk *sdkp) * @sdkp: disk used * @buf: report reply buffer * @buflen: length of @buf + * @zone_shift: logarithm base 2 of the number of blocks in a zone * @seq_zones_bitmap: bitmap of sequential zones to set * * Parse reported zone descriptors in @buf to identify sequential zones and @@ -525,7 +523,7 @@ static inline unsigned long *sd_zbc_alloc_zone_bitmap(struct scsi_disk *sdkp) * Return the LBA after the last zone reported. */ static sector_t sd_zbc_get_seq_zones(struct scsi_disk *sdkp, unsigned char *buf, - unsigned int buflen, + unsigned int buflen, u32 zone_shift, unsigned long *seq_zones_bitmap) { sector_t lba, next_lba = sdkp->capacity; @@ -544,7 +542,7 @@ static sector_t sd_zbc_get_seq_zones(struct scsi_disk *sdkp, unsigned char *buf, if (type != ZBC_ZONE_TYPE_CONV && cond != ZBC_ZONE_COND_READONLY && cond != ZBC_ZONE_COND_OFFLINE) - set_bit(lba >> sdkp->zone_shift, seq_zones_bitmap); + set_bit(lba >> zone_shift, seq_zones_bitmap); next_lba = lba + get_unaligned_be64(&rec[8]); rec += 64; } @@ -553,12 +551,16 @@ static sector_t sd_zbc_get_seq_zones(struct scsi_disk *sdkp, unsigned char *buf, } /** - * sd_zbc_setup_seq_zones_bitmap - Initialize the disk seq zone bitmap. + * sd_zbc_setup_seq_zones_bitmap - Initialize a seq zone bitmap. * @sdkp: target disk + * @zone_shift: logarithm base 2 of the number of blocks in a zone + * @nr_zones: number of zones to set up a seq zone bitmap for * * Allocate a zone bitmap and initialize it by identifying sequential zones. */ -static int sd_zbc_setup_seq_zones_bitmap(struct scsi_disk *sdkp) +static unsigned long * +sd_zbc_setup_seq_zones_bitmap(struct scsi_disk *sdkp, u32 zone_shift, + u32 nr_zones) { struct request_queue *q = sdkp->disk->queue; unsigned long *seq_zones_bitmap; @@ -566,9 +568,9 @@ static int sd_zbc_setup_seq_zones_bitmap(struct scsi_disk *sdkp) unsigned char *buf; int ret = -ENOMEM; - seq_zones_bitmap = sd_zbc_alloc_zone_bitmap(sdkp); + seq_zones_bitmap = sd_zbc_alloc_zone_bitmap(nr_zones, q->node); if (!seq_zones_bitmap) - return -ENOMEM; + return ERR_PTR(-ENOMEM); buf = kmalloc(SD_ZBC_BUF_SIZE, GFP_KERNEL); if (!buf) @@ -579,7 +581,7 @@ static int sd_zbc_setup_seq_zones_bitmap(struct scsi_disk *sdkp) if (ret) goto out; lba = sd_zbc_get_seq_zones(sdkp, buf, SD_ZBC_BUF_SIZE, - seq_zones_bitmap); + zone_shift, seq_zones_bitmap); } if (lba != sdkp->capacity) { @@ -591,12 +593,9 @@ out: kfree(buf); if (ret) { kfree(seq_zones_bitmap); - return ret; + return ERR_PTR(ret); } - - q->seq_zones_bitmap = seq_zones_bitmap; - - return 0; + return seq_zones_bitmap; } static void sd_zbc_cleanup(struct scsi_disk *sdkp) @@ -612,44 +611,64 @@ static void sd_zbc_cleanup(struct scsi_disk *sdkp) q->nr_zones = 0; } -static int sd_zbc_setup(struct scsi_disk *sdkp) +static int sd_zbc_setup(struct scsi_disk *sdkp, u32 zone_blocks) { struct request_queue *q = sdkp->disk->queue; + u32 zone_shift = ilog2(zone_blocks); + u32 nr_zones; int ret; - /* READ16/WRITE16 is mandatory for ZBC disks */ - sdkp->device->use_16_for_rw = 1; - sdkp->device->use_10_for_rw = 0; - /* chunk_sectors indicates the zone size */ - blk_queue_chunk_sectors(sdkp->disk->queue, - logical_to_sectors(sdkp->device, sdkp->zone_blocks)); - sdkp->nr_zones = - round_up(sdkp->capacity, sdkp->zone_blocks) >> sdkp->zone_shift; + blk_queue_chunk_sectors(q, + logical_to_sectors(sdkp->device, zone_blocks)); + nr_zones = round_up(sdkp->capacity, zone_blocks) >> zone_shift; /* * Initialize the device request queue information if the number * of zones changed. */ - if (sdkp->nr_zones != q->nr_zones) { + if (nr_zones != sdkp->nr_zones || nr_zones != q->nr_zones) { + unsigned long *seq_zones_wlock = NULL, *seq_zones_bitmap = NULL; + size_t zone_bitmap_size; - sd_zbc_cleanup(sdkp); - - q->nr_zones = sdkp->nr_zones; - if (sdkp->nr_zones) { - q->seq_zones_wlock = sd_zbc_alloc_zone_bitmap(sdkp); - if (!q->seq_zones_wlock) { + if (nr_zones) { + seq_zones_wlock = sd_zbc_alloc_zone_bitmap(nr_zones, + q->node); + if (!seq_zones_wlock) { ret = -ENOMEM; goto err; } - ret = sd_zbc_setup_seq_zones_bitmap(sdkp); - if (ret) { - sd_zbc_cleanup(sdkp); + seq_zones_bitmap = sd_zbc_setup_seq_zones_bitmap(sdkp, + zone_shift, nr_zones); + if (IS_ERR(seq_zones_bitmap)) { + ret = PTR_ERR(seq_zones_bitmap); + kfree(seq_zones_wlock); goto err; } } + zone_bitmap_size = BITS_TO_LONGS(nr_zones) * + sizeof(unsigned long); + blk_mq_freeze_queue(q); + if (q->nr_zones != nr_zones) { + /* READ16/WRITE16 is mandatory for ZBC disks */ + sdkp->device->use_16_for_rw = 1; + sdkp->device->use_10_for_rw = 0; + sdkp->zone_blocks = zone_blocks; + sdkp->zone_shift = zone_shift; + sdkp->nr_zones = nr_zones; + q->nr_zones = nr_zones; + swap(q->seq_zones_wlock, seq_zones_wlock); + swap(q->seq_zones_bitmap, seq_zones_bitmap); + } else if (memcmp(q->seq_zones_bitmap, seq_zones_bitmap, + zone_bitmap_size) != 0) { + memcpy(q->seq_zones_bitmap, seq_zones_bitmap, + zone_bitmap_size); + } + blk_mq_unfreeze_queue(q); + kfree(seq_zones_wlock); + kfree(seq_zones_bitmap); } return 0; @@ -661,6 +680,7 @@ err: int sd_zbc_read_zones(struct scsi_disk *sdkp, unsigned char *buf) { + int64_t zone_blocks; int ret; if (!sd_is_zoned(sdkp)) @@ -697,12 +717,16 @@ int sd_zbc_read_zones(struct scsi_disk *sdkp, unsigned char *buf) * Check zone size: only devices with a constant zone size (except * an eventual last runt zone) that is a power of 2 are supported. */ - ret = sd_zbc_check_zone_size(sdkp); - if (ret) + zone_blocks = sd_zbc_check_zone_size(sdkp); + ret = -EFBIG; + if (zone_blocks != (u32)zone_blocks) + goto err; + ret = zone_blocks; + if (ret < 0) goto err; /* The drive satisfies the kernel restrictions: set it up */ - ret = sd_zbc_setup(sdkp); + ret = sd_zbc_setup(sdkp, zone_blocks); if (ret) goto err; diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 9af3e0f430bc..21e21f273a21 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -605,6 +605,11 @@ struct request_queue { * initialized by the low level device driver (e.g. scsi/sd.c). * Stacking drivers (device mappers) may or may not initialize * these fields. + * + * Reads of this information must be protected with blk_queue_enter() / + * blk_queue_exit(). Modifying this information is only allowed while + * no requests are being processed. See also blk_mq_freeze_queue() and + * blk_mq_unfreeze_queue(). */ unsigned int nr_zones; unsigned long *seq_zones_bitmap; From 8e1ceafe50ec4d1bcfae154dd70e7cb6946a6177 Mon Sep 17 00:00:00 2001 From: Ming Lei Date: Mon, 16 Apr 2018 17:48:41 +0800 Subject: [PATCH 10/10] scsi: target: fix crash with iscsi target and dvd When the current page can't be added to bio, one new bio should be created for adding this page again, instead of ignoring this page. This patch fixes kernel crash with iscsi target and dvd, as reported by Wakko. Cc: Wakko Warner Cc: Bart Van Assche Cc: target-devel@vger.kernel.org Cc: linux-scsi@vger.kernel.org Cc: "Nicholas A. Bellinger" Cc: Christoph Hellwig Fixes: 84c8590646d5b35804 ("target: avoid accessing .bi_vcnt directly") Signed-off-by: Ming Lei Reviewed-by: Christoph Hellwig Signed-off-by: Martin K. Petersen --- drivers/target/target_core_pscsi.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/target/target_core_pscsi.c b/drivers/target/target_core_pscsi.c index 0d99b242e82e..6cb933ecc084 100644 --- a/drivers/target/target_core_pscsi.c +++ b/drivers/target/target_core_pscsi.c @@ -890,6 +890,7 @@ pscsi_map_sg(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents, bytes = min(bytes, data_len); if (!bio) { +new_bio: nr_vecs = min_t(int, BIO_MAX_PAGES, nr_pages); nr_pages -= nr_vecs; /* @@ -931,6 +932,7 @@ pscsi_map_sg(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents, * be allocated with pscsi_get_bio() above. */ bio = NULL; + goto new_bio; } data_len -= bytes;