From c7702b8c22712a06080e10f1d2dee1a133ec8809 Mon Sep 17 00:00:00 2001 From: "Milan P. Gandhi" Date: Sat, 24 Dec 2016 22:02:46 +0530 Subject: [PATCH 01/12] scsi: qla2xxx: Get mutex lock before checking optrom_state There is a race condition with qla2xxx optrom functions where one thread might modify optrom buffer, optrom_state while other thread is still reading from it. In couple of crashes, it was found that we had successfully passed the following 'if' check where we confirm optrom_state to be QLA_SREADING. But by the time we acquired mutex lock to proceed with memory_read_from_buffer function, some other thread/process had already modified that option rom buffer and optrom_state from QLA_SREADING to QLA_SWAITING. Then we got ha->optrom_buffer 0x0 and crashed the system: if (ha->optrom_state != QLA_SREADING) return 0; mutex_lock(&ha->optrom_mutex); rval = memory_read_from_buffer(buf, count, &off, ha->optrom_buffer, ha->optrom_region_size); mutex_unlock(&ha->optrom_mutex); With current optrom function we get following crash due to a race condition: [ 1479.466679] BUG: unable to handle kernel NULL pointer dereference at (null) [ 1479.466707] IP: [] memcpy+0x6/0x110 [...] [ 1479.473673] Call Trace: [ 1479.474296] [] ? memory_read_from_buffer+0x3c/0x60 [ 1479.474941] [] qla2x00_sysfs_read_optrom+0x9c/0xc0 [qla2xxx] [ 1479.475571] [] read+0xdb/0x1f0 [ 1479.476206] [] vfs_read+0x9e/0x170 [ 1479.476839] [] SyS_read+0x7f/0xe0 [ 1479.477466] [] system_call_fastpath+0x16/0x1b Below patch modifies qla2x00_sysfs_read_optrom, qla2x00_sysfs_write_optrom functions to get the mutex_lock before checking ha->optrom_state to avoid similar crashes. The patch was applied and tested and same crashes were no longer observed again. Tested-by: Milan P. Gandhi Signed-off-by: Milan P. Gandhi Reviewed-by: Laurence Oberman Acked-by: Himanshu Madhani Signed-off-by: Martin K. Petersen --- drivers/scsi/qla2xxx/qla_attr.c | 20 ++++++++++++++------ 1 file changed, 14 insertions(+), 6 deletions(-) diff --git a/drivers/scsi/qla2xxx/qla_attr.c b/drivers/scsi/qla2xxx/qla_attr.c index 47eb4d545d13..83f8527c0363 100644 --- a/drivers/scsi/qla2xxx/qla_attr.c +++ b/drivers/scsi/qla2xxx/qla_attr.c @@ -243,12 +243,15 @@ qla2x00_sysfs_read_optrom(struct file *filp, struct kobject *kobj, struct qla_hw_data *ha = vha->hw; ssize_t rval = 0; - if (ha->optrom_state != QLA_SREADING) - return 0; - mutex_lock(&ha->optrom_mutex); + + if (ha->optrom_state != QLA_SREADING) + goto out; + rval = memory_read_from_buffer(buf, count, &off, ha->optrom_buffer, ha->optrom_region_size); + +out: mutex_unlock(&ha->optrom_mutex); return rval; @@ -263,14 +266,19 @@ qla2x00_sysfs_write_optrom(struct file *filp, struct kobject *kobj, struct device, kobj))); struct qla_hw_data *ha = vha->hw; - if (ha->optrom_state != QLA_SWRITING) + mutex_lock(&ha->optrom_mutex); + + if (ha->optrom_state != QLA_SWRITING) { + mutex_unlock(&ha->optrom_mutex); return -EINVAL; - if (off > ha->optrom_region_size) + } + if (off > ha->optrom_region_size) { + mutex_unlock(&ha->optrom_mutex); return -ERANGE; + } if (off + count > ha->optrom_region_size) count = ha->optrom_region_size - off; - mutex_lock(&ha->optrom_mutex); memcpy(&ha->optrom_buffer[off], buf, count); mutex_unlock(&ha->optrom_mutex); From c3c4239465e11b2cc25fcf375c7909a342bcf4dc Mon Sep 17 00:00:00 2001 From: Dave Jones Date: Tue, 27 Dec 2016 13:13:21 -0500 Subject: [PATCH 02/12] scsi: qla2xxx: Fix apparent cut-n-paste error. Commit 093df73771ba ("scsi: qla2xxx: Fix Target mode handling with Multiqueue changes.") introduces two bodies of code that look similar but with s/req/rsp/ in the second instance. But in one case, it looks like this conversion was missed. Signed-off-by: Dave Jones Reviewed-by: Laurence Oberman Acked-by: Quinn Tran Signed-off-by: Martin K. Petersen --- drivers/scsi/qla2xxx/qla_os.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c index 8521cfe302e9..ad4edc13ebcf 100644 --- a/drivers/scsi/qla2xxx/qla_os.c +++ b/drivers/scsi/qla2xxx/qla_os.c @@ -466,7 +466,7 @@ static void qla2x00_free_queues(struct qla_hw_data *ha) continue; rsp = ha->rsp_q_map[cnt]; - clear_bit(cnt, ha->req_qid_map); + clear_bit(cnt, ha->rsp_qid_map); ha->rsp_q_map[cnt] = NULL; spin_unlock_irqrestore(&ha->hardware_lock, flags); qla2x00_free_rsp_que(ha, rsp); From 17e5fc58588b5e3df8220c90a9d8af55201d6b45 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 11 Jan 2017 17:55:45 +0100 Subject: [PATCH 03/12] scsi: qla2xxx: fix MSI-X vector affinity The first two or three vectors in qla2xxx adapter are global and not associated with a specific queue. They should not have IRQ affinity assigned. Signed-off-by: Christoph Hellwig Acked-by: Himanshu Madhani Signed-off-by: Martin K. Petersen --- drivers/scsi/qla2xxx/qla_def.h | 2 +- drivers/scsi/qla2xxx/qla_isr.c | 15 +++++++++++---- 2 files changed, 12 insertions(+), 5 deletions(-) diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h index f7df01b76714..69587a21cb06 100644 --- a/drivers/scsi/qla2xxx/qla_def.h +++ b/drivers/scsi/qla2xxx/qla_def.h @@ -2732,7 +2732,7 @@ struct isp_operations { #define QLA_MSIX_FW_MODE(m) (((m) & (BIT_7|BIT_8|BIT_9)) >> 7) #define QLA_MSIX_FW_MODE_1(m) (QLA_MSIX_FW_MODE(m) == 1) -#define QLA_MSIX_DEFAULT 0x00 +#define QLA_BASE_VECTORS 2 /* default + RSP */ #define QLA_MSIX_RSP_Q 0x01 #define QLA_ATIO_VECTOR 0x02 #define QLA_MSIX_QPAIR_MULTIQ_RSP_Q 0x03 diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c index af840bf587d5..e2ef303b2bcf 100644 --- a/drivers/scsi/qla2xxx/qla_isr.c +++ b/drivers/scsi/qla2xxx/qla_isr.c @@ -3017,13 +3017,20 @@ static struct qla_init_msix_entry qla82xx_msix_entries[] = { static int qla24xx_enable_msix(struct qla_hw_data *ha, struct rsp_que *rsp) { -#define MIN_MSIX_COUNT 2 int i, ret; struct qla_msix_entry *qentry; scsi_qla_host_t *vha = pci_get_drvdata(ha->pdev); + struct irq_affinity desc = { + .pre_vectors = QLA_BASE_VECTORS, + }; + + if (QLA_TGT_MODE_ENABLED() && IS_ATIO_MSIX_CAPABLE(ha)) + desc.pre_vectors++; + + ret = pci_alloc_irq_vectors_affinity(ha->pdev, QLA_BASE_VECTORS, + ha->msix_count, PCI_IRQ_MSIX | PCI_IRQ_AFFINITY, + &desc); - ret = pci_alloc_irq_vectors(ha->pdev, MIN_MSIX_COUNT, ha->msix_count, - PCI_IRQ_MSIX | PCI_IRQ_AFFINITY); if (ret < 0) { ql_log(ql_log_fatal, vha, 0x00c7, "MSI-X: Failed to enable support, " @@ -3074,7 +3081,7 @@ qla24xx_enable_msix(struct qla_hw_data *ha, struct rsp_que *rsp) } /* Enable MSI-X vectors for the base queue */ - for (i = 0; i < (QLA_MSIX_RSP_Q + 1); i++) { + for (i = 0; i < QLA_BASE_VECTORS; i++) { qentry = &ha->msix_entries[i]; qentry->handle = rsp; rsp->msix = qentry; From 98624c4fed0abd848b291fbd3da18c2251b79429 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 11 Jan 2017 17:55:46 +0100 Subject: [PATCH 04/12] scsi: qla2xxx: remove irq_affinity_notifier Now that qla2xxx uses the IRQ layer affinity assignment, affinity won't change over the life time of a device and the notifiers are useless. Signed-off-by: Christoph Hellwig Acked-by: Himanshu Madhani Signed-off-by: Martin K. Petersen --- drivers/scsi/qla2xxx/qla_def.h | 1 - drivers/scsi/qla2xxx/qla_isr.c | 73 ---------------------------------- 2 files changed, 74 deletions(-) diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h index 69587a21cb06..1549aace296a 100644 --- a/drivers/scsi/qla2xxx/qla_def.h +++ b/drivers/scsi/qla2xxx/qla_def.h @@ -2754,7 +2754,6 @@ struct qla_msix_entry { uint16_t entry; char name[30]; void *handle; - struct irq_affinity_notify irq_notify; int cpuid; }; diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c index e2ef303b2bcf..5815403d1d65 100644 --- a/drivers/scsi/qla2xxx/qla_isr.c +++ b/drivers/scsi/qla2xxx/qla_isr.c @@ -18,10 +18,6 @@ static void qla2x00_status_entry(scsi_qla_host_t *, struct rsp_que *, void *); static void qla2x00_status_cont_entry(struct rsp_que *, sts_cont_entry_t *); static void qla2x00_error_entry(scsi_qla_host_t *, struct rsp_que *, sts_entry_t *); -static void qla_irq_affinity_notify(struct irq_affinity_notify *, - const cpumask_t *); -static void qla_irq_affinity_release(struct kref *); - /** * qla2100_intr_handler() - Process interrupts for the ISP2100 and ISP2200. @@ -2571,14 +2567,6 @@ void qla24xx_process_response_queue(struct scsi_qla_host *vha, if (!vha->flags.online) return; - if (rsp->msix && rsp->msix->cpuid != smp_processor_id()) { - /* if kernel does not notify qla of IRQ's CPU change, - * then set it here. - */ - rsp->msix->cpuid = smp_processor_id(); - ha->tgt.rspq_vector_cpuid = rsp->msix->cpuid; - } - while (rsp->ring_ptr->signature != RESPONSE_PROCESSED) { pkt = (struct sts_entry_24xx *)rsp->ring_ptr; @@ -3075,9 +3063,6 @@ qla24xx_enable_msix(struct qla_hw_data *ha, struct rsp_que *rsp) qentry->have_irq = 0; qentry->in_use = 0; qentry->handle = NULL; - qentry->irq_notify.notify = qla_irq_affinity_notify; - qentry->irq_notify.release = qla_irq_affinity_release; - qentry->cpuid = -1; } /* Enable MSI-X vectors for the base queue */ @@ -3099,18 +3084,6 @@ qla24xx_enable_msix(struct qla_hw_data *ha, struct rsp_que *rsp) goto msix_register_fail; qentry->have_irq = 1; qentry->in_use = 1; - - /* Register for CPU affinity notification. */ - irq_set_affinity_notifier(qentry->vector, &qentry->irq_notify); - - /* Schedule work (ie. trigger a notification) to read cpu - * mask for this specific irq. - * kref_get is required because - * irq_affinity_notify() will do - * kref_put(). - */ - kref_get(&qentry->irq_notify.kref); - schedule_work(&qentry->irq_notify.work); } /* @@ -3307,49 +3280,3 @@ int qla25xx_request_irq(struct qla_hw_data *ha, struct qla_qpair *qpair, msix->handle = qpair; return ret; } - - -/* irq_set_affinity/irqbalance will trigger notification of cpu mask update */ -static void qla_irq_affinity_notify(struct irq_affinity_notify *notify, - const cpumask_t *mask) -{ - struct qla_msix_entry *e = - container_of(notify, struct qla_msix_entry, irq_notify); - struct qla_hw_data *ha; - struct scsi_qla_host *base_vha; - struct rsp_que *rsp = e->handle; - - /* user is recommended to set mask to just 1 cpu */ - e->cpuid = cpumask_first(mask); - - ha = rsp->hw; - base_vha = pci_get_drvdata(ha->pdev); - - ql_dbg(ql_dbg_init, base_vha, 0xffff, - "%s: host %ld : vector %d cpu %d \n", __func__, - base_vha->host_no, e->vector, e->cpuid); - - if (e->have_irq) { - if ((IS_QLA83XX(ha) || IS_QLA27XX(ha)) && - (e->entry == QLA83XX_RSPQ_MSIX_ENTRY_NUMBER)) { - ha->tgt.rspq_vector_cpuid = e->cpuid; - ql_dbg(ql_dbg_init, base_vha, 0xffff, - "%s: host%ld: rspq vector %d cpu %d runtime change\n", - __func__, base_vha->host_no, e->vector, e->cpuid); - } - } -} - -static void qla_irq_affinity_release(struct kref *ref) -{ - struct irq_affinity_notify *notify = - container_of(ref, struct irq_affinity_notify, kref); - struct qla_msix_entry *e = - container_of(notify, struct qla_msix_entry, irq_notify); - struct rsp_que *rsp = e->handle; - struct scsi_qla_host *base_vha = pci_get_drvdata(rsp->hw->pdev); - - ql_dbg(ql_dbg_init, base_vha, 0xffff, - "%s: host%ld: vector %d cpu %d\n", __func__, - base_vha->host_no, e->vector, e->cpuid); -} From cd60be4916ae689387d04b86b6fc15931e4c95ae Mon Sep 17 00:00:00 2001 From: Roberto Sassu Date: Wed, 11 Jan 2017 11:06:42 +0100 Subject: [PATCH 05/12] scsi: lpfc: avoid double free of resource identifiers Set variables initialized in lpfc_sli4_alloc_resource_identifiers() to NULL if an error occurred. Otherwise, lpfc_sli4_driver_resource_unset() attempts to free the memory again. Signed-off-by: Roberto Sassu Signed-off-by: Johannes Thumshirn Acked-by: James Smart Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_sli.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c index 4faa7672fc1d..a78a3df68f67 100644 --- a/drivers/scsi/lpfc/lpfc_sli.c +++ b/drivers/scsi/lpfc/lpfc_sli.c @@ -5954,18 +5954,25 @@ lpfc_sli4_alloc_resource_identifiers(struct lpfc_hba *phba) free_vfi_bmask: kfree(phba->sli4_hba.vfi_bmask); + phba->sli4_hba.vfi_bmask = NULL; free_xri_ids: kfree(phba->sli4_hba.xri_ids); + phba->sli4_hba.xri_ids = NULL; free_xri_bmask: kfree(phba->sli4_hba.xri_bmask); + phba->sli4_hba.xri_bmask = NULL; free_vpi_ids: kfree(phba->vpi_ids); + phba->vpi_ids = NULL; free_vpi_bmask: kfree(phba->vpi_bmask); + phba->vpi_bmask = NULL; free_rpi_ids: kfree(phba->sli4_hba.rpi_ids); + phba->sli4_hba.rpi_ids = NULL; free_rpi_bmask: kfree(phba->sli4_hba.rpi_bmask); + phba->sli4_hba.rpi_bmask = NULL; err_exit: return rc; } From 67d35e70af9cabb663c827e03bc5c1e89b43db72 Mon Sep 17 00:00:00 2001 From: Fam Zheng Date: Fri, 13 Jan 2017 11:40:01 +0800 Subject: [PATCH 06/12] scsi: libfc: Fix variable name in fc_set_wwpn The parameter name should be wwpn instead of wwnn. Signed-off-by: Fam Zheng Acked-by: Johannes Thumshirn Signed-off-by: Martin K. Petersen --- include/scsi/libfc.h | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/include/scsi/libfc.h b/include/scsi/libfc.h index 96dd0b3f70d7..da5033dd8cbc 100644 --- a/include/scsi/libfc.h +++ b/include/scsi/libfc.h @@ -809,11 +809,11 @@ static inline void fc_set_wwnn(struct fc_lport *lport, u64 wwnn) /** * fc_set_wwpn() - Set the World Wide Port Name of a local port * @lport: The local port whose WWPN is to be set - * @wwnn: The new WWPN + * @wwpn: The new WWPN */ -static inline void fc_set_wwpn(struct fc_lport *lport, u64 wwnn) +static inline void fc_set_wwpn(struct fc_lport *lport, u64 wwpn) { - lport->wwpn = wwnn; + lport->wwpn = wwpn; } /** From 9373eba6cfae48911b977d14323032cd5d161aae Mon Sep 17 00:00:00 2001 From: "Ewan D. Milne" Date: Mon, 9 Jan 2017 16:33:36 -0500 Subject: [PATCH 07/12] scsi: ses: Fix SAS device detection in enclosure The call to scsi_is_sas_rphy() needs to be made on the SAS end_device, not on the SCSI device. Fixes: 835831c57e9b ("ses: use scsi_is_sas_rphy instead of is_sas_attached") Signed-off-by: Ewan D. Milne Reviewed-by: Johannes Thumshirn Reviewed-by: James Bottomley Signed-off-by: Martin K. Petersen --- drivers/scsi/ses.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/ses.c b/drivers/scsi/ses.c index 8c9a35c91705..50adabbb5808 100644 --- a/drivers/scsi/ses.c +++ b/drivers/scsi/ses.c @@ -587,7 +587,7 @@ static void ses_match_to_enclosure(struct enclosure_device *edev, ses_enclosure_data_process(edev, to_scsi_device(edev->edev.parent), 0); - if (scsi_is_sas_rphy(&sdev->sdev_gendev)) + if (scsi_is_sas_rphy(sdev->sdev_target->dev.parent)) efd.addr = sas_get_address(sdev); if (efd.addr) { From 4633773799940b1b8b3ff98ea05e6c1ef072febd Mon Sep 17 00:00:00 2001 From: Johannes Thumshirn Date: Tue, 10 Jan 2017 12:54:36 +0100 Subject: [PATCH 08/12] scsi: bfa: fix wrongly initialized variable in bfad_im_bsg_els_ct_request() Commit 01e0e15c8b3b ("scsi: don't use fc_bsg_job::request and fc_bsg_job::reply directly") introduced a typo, which causes that the bsg_request variable in bfad_im_bsg_els_ct_request() is initialized to itself instead of pointing to the bsg job's request. Reported-by: Nicolas Iooss Signed-off-by: Johannes Thumshirn Signed-off-by: Martin K. Petersen --- drivers/scsi/bfa/bfad_bsg.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/bfa/bfad_bsg.c b/drivers/scsi/bfa/bfad_bsg.c index a9a00169ad91..b2e8c0dfc79c 100644 --- a/drivers/scsi/bfa/bfad_bsg.c +++ b/drivers/scsi/bfa/bfad_bsg.c @@ -3363,7 +3363,7 @@ bfad_im_bsg_els_ct_request(struct bsg_job *job) struct bfad_fcxp *drv_fcxp; struct bfa_fcs_lport_s *fcs_port; struct bfa_fcs_rport_s *fcs_rport; - struct fc_bsg_request *bsg_request = bsg_request; + struct fc_bsg_request *bsg_request = job->request; struct fc_bsg_reply *bsg_reply = job->reply; uint32_t command_type = bsg_request->msgcode; unsigned long flags; From 26f2819772af891dee2843e1f8662c58e5129d5f Mon Sep 17 00:00:00 2001 From: Damien Le Moal Date: Thu, 12 Jan 2017 15:25:10 +0900 Subject: [PATCH 09/12] scsi: sd: Fix wrong DPOFUA disable in sd_read_cache_type Zoned block devices force the use of READ/WRITE(16) commands by setting sdkp->use_16_for_rw and clearing sdkp->use_10_for_rw. This result in DPOFUA always being disabled for these drives as the assumed use of the deprecated READ/WRITE(6) commands only looks at sdkp->use_10_for_rw. Strenghten the test by also checking that sdkp->use_16_for_rw is false. Signed-off-by: Damien Le Moal Reviewed-by: Hannes Reinecke Signed-off-by: Martin K. Petersen --- drivers/scsi/sd.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c index 1622e23138e0..28713d36050e 100644 --- a/drivers/scsi/sd.c +++ b/drivers/scsi/sd.c @@ -2600,7 +2600,8 @@ sd_read_cache_type(struct scsi_disk *sdkp, unsigned char *buffer) if (sdp->broken_fua) { sd_first_printk(KERN_NOTICE, sdkp, "Disabling FUA\n"); sdkp->DPOFUA = 0; - } else if (sdkp->DPOFUA && !sdkp->device->use_10_for_rw) { + } else if (sdkp->DPOFUA && !sdkp->device->use_10_for_rw && + !sdkp->device->use_16_for_rw) { sd_first_printk(KERN_NOTICE, sdkp, "Uses READ/WRITE(6), disabling FUA\n"); sdkp->DPOFUA = 0; From 68af412c7713b55c01ffc4312320abd10ca70e77 Mon Sep 17 00:00:00 2001 From: Damien Le Moal Date: Thu, 12 Jan 2017 15:25:11 +0900 Subject: [PATCH 10/12] scsi: sd: Ignore zoned field for host-managed devices There is no good match of the zoned field of the block device characteristics page for host-managed devices. For these devices, the zoning model is derived directly from the device type. So ignore the zoned field for these drives. [mkp: typo] Signed-off-by: Damien Le Moal Reviewed-by: Hannes Reinecke Signed-off-by: Martin K. Petersen --- drivers/scsi/sd.c | 20 ++++++++++++++------ 1 file changed, 14 insertions(+), 6 deletions(-) diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c index 28713d36050e..ebb1ccef7943 100644 --- a/drivers/scsi/sd.c +++ b/drivers/scsi/sd.c @@ -2784,13 +2784,21 @@ static void sd_read_block_characteristics(struct scsi_disk *sdkp) queue_flag_clear_unlocked(QUEUE_FLAG_ADD_RANDOM, q); } - sdkp->zoned = (buffer[8] >> 4) & 3; - if (sdkp->zoned == 1) - q->limits.zoned = BLK_ZONED_HA; - else if (sdkp->device->type == TYPE_ZBC) + if (sdkp->device->type == TYPE_ZBC) { + /* Host-managed */ q->limits.zoned = BLK_ZONED_HM; - else - q->limits.zoned = BLK_ZONED_NONE; + } else { + sdkp->zoned = (buffer[8] >> 4) & 3; + if (sdkp->zoned == 1) + /* Host-aware */ + q->limits.zoned = BLK_ZONED_HA; + else + /* + * Treat drive-managed devices as + * regular block devices. + */ + q->limits.zoned = BLK_ZONED_NONE; + } if (blk_queue_is_zoned(q) && sdkp->first_scan) sd_printk(KERN_NOTICE, sdkp, "Host-%s zoned block device\n", q->limits.zoned == BLK_ZONED_HM ? "managed" : "aware"); From 8667f515952feefebb3c0f8d9a9266c91b101a46 Mon Sep 17 00:00:00 2001 From: Johannes Thumshirn Date: Tue, 10 Jan 2017 12:05:54 +0100 Subject: [PATCH 11/12] scsi: lpfc: Set elsiocb contexts to NULL after freeing it Set the elsiocb contexts to NULL after freeing as others depend on it. Signed-off-by: Johannes Thumshirn Acked-by: Dick Kennedy Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_els.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c index 236e4e51d161..7b6bd8ed0d0b 100644 --- a/drivers/scsi/lpfc/lpfc_els.c +++ b/drivers/scsi/lpfc/lpfc_els.c @@ -3590,12 +3590,14 @@ lpfc_els_free_iocb(struct lpfc_hba *phba, struct lpfc_iocbq *elsiocb) } else { buf_ptr1 = (struct lpfc_dmabuf *) elsiocb->context2; lpfc_els_free_data(phba, buf_ptr1); + elsiocb->context2 = NULL; } } if (elsiocb->context3) { buf_ptr = (struct lpfc_dmabuf *) elsiocb->context3; lpfc_els_free_bpl(phba, buf_ptr); + elsiocb->context3 = NULL; } lpfc_sli_release_iocbq(phba, elsiocb); return 0; From ffb58456589443ca572221fabbdef3db8483a779 Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Sun, 1 Jan 2017 09:39:24 -0800 Subject: [PATCH 12/12] scsi: mpt3sas: fix hang on ata passthrough commands mpt3sas has a firmware failure where it can only handle one pass through ATA command at a time. If another comes in, contrary to the SAT standard, it will hang until the first one completes (causing long commands like secure erase to timeout). The original fix was to block the device when an ATA command came in, but this caused a regression with commit 669f044170d8933c3d66d231b69ea97cb8447338 Author: Bart Van Assche Date: Tue Nov 22 16:17:13 2016 -0800 scsi: srp_transport: Move queuecommand() wait code to SCSI core So fix the original fix of the secure erase timeout by properly returning SAM_STAT_BUSY like the SAT recommends. The original patch also had a concurrency problem since scsih_qcmd is lockless at that point (this is fixed by using atomic bitops to set and test the flag). [mkp: addressed feedback wrt. test_bit and fixed whitespace] Fixes: 18f6084a989ba1b (mpt3sas: Fix secure erase premature termination) Signed-off-by: James Bottomley Acked-by: Sreekanth Reddy Reviewed-by: Christoph Hellwig Reported-by: Ingo Molnar Tested-by: Ingo Molnar Signed-off-by: Martin K. Petersen --- drivers/scsi/mpt3sas/mpt3sas_base.h | 12 +++++++++ drivers/scsi/mpt3sas/mpt3sas_scsih.c | 40 ++++++++++++++++++---------- 2 files changed, 38 insertions(+), 14 deletions(-) diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.h b/drivers/scsi/mpt3sas/mpt3sas_base.h index 394fe1338d09..dcb33f4fa687 100644 --- a/drivers/scsi/mpt3sas/mpt3sas_base.h +++ b/drivers/scsi/mpt3sas/mpt3sas_base.h @@ -393,6 +393,7 @@ struct MPT3SAS_TARGET { * @eedp_enable: eedp support enable bit * @eedp_type: 0(type_1), 1(type_2), 2(type_3) * @eedp_block_length: block size + * @ata_command_pending: SATL passthrough outstanding for device */ struct MPT3SAS_DEVICE { struct MPT3SAS_TARGET *sas_target; @@ -404,6 +405,17 @@ struct MPT3SAS_DEVICE { u8 ignore_delay_remove; /* Iopriority Command Handling */ u8 ncq_prio_enable; + /* + * Bug workaround for SATL handling: the mpt2/3sas firmware + * doesn't return BUSY or TASK_SET_FULL for subsequent + * commands while a SATL pass through is in operation as the + * spec requires, it simply does nothing with them until the + * pass through completes, causing them possibly to timeout if + * the passthrough is a long executing command (like format or + * secure erase). This variable allows us to do the right + * thing while a SATL command is pending. + */ + unsigned long ata_command_pending; }; diff --git a/drivers/scsi/mpt3sas/mpt3sas_scsih.c b/drivers/scsi/mpt3sas/mpt3sas_scsih.c index b5c966e319d3..75f3fce1c867 100644 --- a/drivers/scsi/mpt3sas/mpt3sas_scsih.c +++ b/drivers/scsi/mpt3sas/mpt3sas_scsih.c @@ -3899,9 +3899,18 @@ _scsih_temp_threshold_events(struct MPT3SAS_ADAPTER *ioc, } } -static inline bool ata_12_16_cmd(struct scsi_cmnd *scmd) +static int _scsih_set_satl_pending(struct scsi_cmnd *scmd, bool pending) { - return (scmd->cmnd[0] == ATA_12 || scmd->cmnd[0] == ATA_16); + struct MPT3SAS_DEVICE *priv = scmd->device->hostdata; + + if (scmd->cmnd[0] != ATA_12 && scmd->cmnd[0] != ATA_16) + return 0; + + if (pending) + return test_and_set_bit(0, &priv->ata_command_pending); + + clear_bit(0, &priv->ata_command_pending); + return 0; } /** @@ -3925,9 +3934,7 @@ _scsih_flush_running_cmds(struct MPT3SAS_ADAPTER *ioc) if (!scmd) continue; count++; - if (ata_12_16_cmd(scmd)) - scsi_internal_device_unblock(scmd->device, - SDEV_RUNNING); + _scsih_set_satl_pending(scmd, false); mpt3sas_base_free_smid(ioc, smid); scsi_dma_unmap(scmd); if (ioc->pci_error_recovery) @@ -4063,13 +4070,6 @@ scsih_qcmd(struct Scsi_Host *shost, struct scsi_cmnd *scmd) if (ioc->logging_level & MPT_DEBUG_SCSI) scsi_print_command(scmd); - /* - * Lock the device for any subsequent command until command is - * done. - */ - if (ata_12_16_cmd(scmd)) - scsi_internal_device_block(scmd->device); - sas_device_priv_data = scmd->device->hostdata; if (!sas_device_priv_data || !sas_device_priv_data->sas_target) { scmd->result = DID_NO_CONNECT << 16; @@ -4083,6 +4083,19 @@ scsih_qcmd(struct Scsi_Host *shost, struct scsi_cmnd *scmd) return 0; } + /* + * Bug work around for firmware SATL handling. The loop + * is based on atomic operations and ensures consistency + * since we're lockless at this point + */ + do { + if (test_bit(0, &sas_device_priv_data->ata_command_pending)) { + scmd->result = SAM_STAT_BUSY; + scmd->scsi_done(scmd); + return 0; + } + } while (_scsih_set_satl_pending(scmd, true)); + sas_target_priv_data = sas_device_priv_data->sas_target; /* invalid device handle */ @@ -4650,8 +4663,7 @@ _scsih_io_done(struct MPT3SAS_ADAPTER *ioc, u16 smid, u8 msix_index, u32 reply) if (scmd == NULL) return 1; - if (ata_12_16_cmd(scmd)) - scsi_internal_device_unblock(scmd->device, SDEV_RUNNING); + _scsih_set_satl_pending(scmd, false); mpi_request = mpt3sas_base_get_msg_frame(ioc, smid);