[PATCH AUTOSEL 5.2 29/76] scsi: qla2xxx: Fix gnl.l memory leak on adapter init failure

[PATCH AUTOSEL 5.2 29/76] scsi: qla2xxx: Fix gnl.l memory leak on adapter init failure

[Sasha Levin]

From: Bill Kuzeja <William.Kuzeja@stratus.com>

[ Upstream commit 26fa656e9a0cbccddf7db132ea020d2169dbe46e ]

If HBA initialization fails unexpectedly (exiting via probe_failed:), we
may fail to free vha->gnl.l. So that we don't attempt to double free, set
this pointer to NULL after a free and check for NULL at probe_failed: so we
know whether or not to call dma_free_coherent.

Signed-off-by: Bill Kuzeja <william.kuzeja@stratus.com>
Acked-by: Himanshu Madhani <hmadhani@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/scsi/qla2xxx/qla_attr.c |  2 ++
 drivers/scsi/qla2xxx/qla_os.c   | 11 ++++++++++-
 2 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/drivers/scsi/qla2xxx/qla_attr.c b/drivers/scsi/qla2xxx/qla_attr.c
index 8d560c562e9c0..6b7b390b2e522 100644
--- a/drivers/scsi/qla2xxx/qla_attr.c
+++ b/drivers/scsi/qla2xxx/qla_attr.c
@@ -2956,6 +2956,8 @@ qla24xx_vport_delete(struct fc_vport *fc_vport)
 	dma_free_coherent(&ha->pdev->dev, vha->gnl.size, vha->gnl.l,
 	    vha->gnl.ldma);
 
+	vha->gnl.l = NULL;
+
 	vfree(vha->scan.l);
 
 	if (vha->qpair && vha->qpair->vp_idx == vha->vp_idx) {
diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c
index d056f5e7cf930..794478e5f7ec8 100644
--- a/drivers/scsi/qla2xxx/qla_os.c
+++ b/drivers/scsi/qla2xxx/qla_os.c
@@ -3440,6 +3440,12 @@ qla2x00_probe_one(struct pci_dev *pdev, const struct pci_device_id *id)
 	return 0;
 
 probe_failed:
+	if (base_vha->gnl.l) {
+		dma_free_coherent(&ha->pdev->dev, base_vha->gnl.size,
+				base_vha->gnl.l, base_vha->gnl.ldma);
+		base_vha->gnl.l = NULL;
+	}
+
 	if (base_vha->timer_active)
 		qla2x00_stop_timer(base_vha);
 	base_vha->flags.online = 0;
@@ -3673,7 +3679,7 @@ qla2x00_remove_one(struct pci_dev *pdev)
 	if (!atomic_read(&pdev->enable_cnt)) {
 		dma_free_coherent(&ha->pdev->dev, base_vha->gnl.size,
 		    base_vha->gnl.l, base_vha->gnl.ldma);
-
+		base_vha->gnl.l = NULL;
 		scsi_host_put(base_vha->host);
 		kfree(ha);
 		pci_set_drvdata(pdev, NULL);
@@ -3713,6 +3719,8 @@ qla2x00_remove_one(struct pci_dev *pdev)
 	dma_free_coherent(&ha->pdev->dev,
 		base_vha->gnl.size, base_vha->gnl.l, base_vha->gnl.ldma);
 
+	base_vha->gnl.l = NULL;
+
 	vfree(base_vha->scan.l);
 
 	if (IS_QLAFX00(ha))
@@ -4817,6 +4825,7 @@ struct scsi_qla_host *qla2x00_create_host(struct scsi_host_template *sht,
 		    "Alloc failed for scan database.\n");
 		dma_free_coherent(&ha->pdev->dev, vha->gnl.size,
 		    vha->gnl.l, vha->gnl.ldma);
+		vha->gnl.l = NULL;
 		scsi_remove_host(vha->host);
 		return NULL;
 	}
-- 
2.20.1

[PATCH AUTOSEL 5.2 30/76] scsi: target: tcmu: avoid use-after-free after command timeout

[Sasha Levin]

From: Dmitry Fomichev <dmitry.fomichev@wdc.com>

[ Upstream commit a86a75865ff4d8c05f355d1750a5250aec89ab15 ]

In tcmu_handle_completion() function, the variable called read_len is
always initialized with a value taken from se_cmd structure. If this
function is called to complete an expired (timed out) out command, the
session command pointed by se_cmd is likely to be already deallocated by
the target core at that moment. As the result, this access triggers a
use-after-free warning from KASAN.

This patch fixes the code not to touch se_cmd when completing timed out
TCMU commands. It also resets the pointer to se_cmd at the time when the
TCMU_CMD_BIT_EXPIRED flag is set because it is going to become invalid
after calling target_complete_cmd() later in the same function,
tcmu_check_expired_cmd().

Signed-off-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Acked-by: Mike Christie <mchristi@redhat.com>
Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/target/target_core_user.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/drivers/target/target_core_user.c b/drivers/target/target_core_user.c
index b43d6385a1a09..95b2371fb67b6 100644
--- a/drivers/target/target_core_user.c
+++ b/drivers/target/target_core_user.c
@@ -1132,14 +1132,16 @@ static void tcmu_handle_completion(struct tcmu_cmd *cmd, struct tcmu_cmd_entry *
 	struct se_cmd *se_cmd = cmd->se_cmd;
 	struct tcmu_dev *udev = cmd->tcmu_dev;
 	bool read_len_valid = false;
-	uint32_t read_len = se_cmd->data_length;
+	uint32_t read_len;
 
 	/*
 	 * cmd has been completed already from timeout, just reclaim
 	 * data area space and free cmd
 	 */
-	if (test_bit(TCMU_CMD_BIT_EXPIRED, &cmd->flags))
+	if (test_bit(TCMU_CMD_BIT_EXPIRED, &cmd->flags)) {
+		WARN_ON_ONCE(se_cmd);
 		goto out;
+	}
 
 	list_del_init(&cmd->queue_entry);
 
@@ -1152,6 +1154,7 @@ static void tcmu_handle_completion(struct tcmu_cmd *cmd, struct tcmu_cmd_entry *
 		goto done;
 	}
 
+	read_len = se_cmd->data_length;
 	if (se_cmd->data_direction == DMA_FROM_DEVICE &&
 	    (entry->hdr.uflags & TCMU_UFLAG_READ_LEN) && entry->rsp.read_len) {
 		read_len_valid = true;
@@ -1307,6 +1310,7 @@ static int tcmu_check_expired_cmd(int id, void *p, void *data)
 		 */
 		scsi_status = SAM_STAT_CHECK_CONDITION;
 		list_del_init(&cmd->queue_entry);
+		cmd->se_cmd = NULL;
 	} else {
 		list_del_init(&cmd->queue_entry);
 		idr_remove(&udev->commands, id);
@@ -2024,6 +2028,7 @@ static void tcmu_reset_ring(struct tcmu_dev *udev, u8 err_level)
 
 		idr_remove(&udev->commands, i);
 		if (!test_bit(TCMU_CMD_BIT_EXPIRED, &cmd->flags)) {
+			WARN_ON(!cmd->se_cmd);
 			list_del_init(&cmd->queue_entry);
 			if (err_level == 1) {
 				/*
-- 
2.20.1

[PATCH AUTOSEL 5.2 52/76] scsi: lpfc: Mitigate high memory pre-allocation by SCSI-MQ

[Sasha Levin]

From: James Smart <jsmart2021@gmail.com>

[ Upstream commit 77ffd3465ba837e9dc714e17b014e77b2eae765a ]

When SCSI-MQ is enabled, the SCSI-MQ layers will do pre-allocation of MQ
resources based on shost values set by the driver. In newer cases of the
driver, which attempts to set nr_hw_queues to the cpu count, the
multipliers become excessive, with a single shost having SCSI-MQ
pre-allocation reaching into the multiple GBytes range.  NPIV, which
creates additional shosts, only multiply this overhead. On lower-memory
systems, this can exhaust system memory very quickly, resulting in a system
crash or failures in the driver or elsewhere due to low memory conditions.

After testing several scenarios, the situation can be mitigated by limiting
the value set in shost->nr_hw_queues to 4. Although the shost values were
changed, the driver still had per-cpu hardware queues of its own that
allowed parallelization per-cpu.  Testing revealed that even with the
smallish number for nr_hw_queues for SCSI-MQ, performance levels remained
near maximum with the within-driver affiinitization.

A module parameter was created to allow the value set for the nr_hw_queues
to be tunable.

Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Ewan D. Milne <emilne@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/scsi/lpfc/lpfc.h      |  1 +
 drivers/scsi/lpfc/lpfc_attr.c | 15 +++++++++++++++
 drivers/scsi/lpfc/lpfc_init.c | 10 ++++++----
 drivers/scsi/lpfc/lpfc_sli4.h |  5 +++++
 4 files changed, 27 insertions(+), 4 deletions(-)

diff --git a/drivers/scsi/lpfc/lpfc.h b/drivers/scsi/lpfc/lpfc.h
index aafcffaa25f71..4604e1bc334c0 100644
--- a/drivers/scsi/lpfc/lpfc.h
+++ b/drivers/scsi/lpfc/lpfc.h
@@ -822,6 +822,7 @@ struct lpfc_hba {
 	uint32_t cfg_cq_poll_threshold;
 	uint32_t cfg_cq_max_proc_limit;
 	uint32_t cfg_fcp_cpu_map;
+	uint32_t cfg_fcp_mq_threshold;
 	uint32_t cfg_hdw_queue;
 	uint32_t cfg_irq_chann;
 	uint32_t cfg_suppress_rsp;
diff --git a/drivers/scsi/lpfc/lpfc_attr.c b/drivers/scsi/lpfc/lpfc_attr.c
index d4c65e2109e2f..353da12d797ba 100644
--- a/drivers/scsi/lpfc/lpfc_attr.c
+++ b/drivers/scsi/lpfc/lpfc_attr.c
@@ -5640,6 +5640,19 @@ LPFC_ATTR_RW(nvme_oas, 0, 0, 1,
 LPFC_ATTR_RW(nvme_embed_cmd, 1, 0, 2,
 	     "Embed NVME Command in WQE");
 
+/*
+ * lpfc_fcp_mq_threshold: Set the maximum number of Hardware Queues
+ * the driver will advertise it supports to the SCSI layer.
+ *
+ *      0    = Set nr_hw_queues by the number of CPUs or HW queues.
+ *      1,128 = Manually specify the maximum nr_hw_queue value to be set,
+ *
+ * Value range is [0,128]. Default value is 8.
+ */
+LPFC_ATTR_R(fcp_mq_threshold, LPFC_FCP_MQ_THRESHOLD_DEF,
+	    LPFC_FCP_MQ_THRESHOLD_MIN, LPFC_FCP_MQ_THRESHOLD_MAX,
+	    "Set the number of SCSI Queues advertised");
+
 /*
  * lpfc_hdw_queue: Set the number of Hardware Queues the driver
  * will advertise it supports to the NVME and  SCSI layers. This also
@@ -5961,6 +5974,7 @@ struct device_attribute *lpfc_hba_attrs[] = {
 	&dev_attr_lpfc_cq_poll_threshold,
 	&dev_attr_lpfc_cq_max_proc_limit,
 	&dev_attr_lpfc_fcp_cpu_map,
+	&dev_attr_lpfc_fcp_mq_threshold,
 	&dev_attr_lpfc_hdw_queue,
 	&dev_attr_lpfc_irq_chann,
 	&dev_attr_lpfc_suppress_rsp,
@@ -7042,6 +7056,7 @@ lpfc_get_cfgparam(struct lpfc_hba *phba)
 	/* Initialize first burst. Target vs Initiator are different. */
 	lpfc_nvme_enable_fb_init(phba, lpfc_nvme_enable_fb);
 	lpfc_nvmet_fb_size_init(phba, lpfc_nvmet_fb_size);
+	lpfc_fcp_mq_threshold_init(phba, lpfc_fcp_mq_threshold);
 	lpfc_hdw_queue_init(phba, lpfc_hdw_queue);
 	lpfc_irq_chann_init(phba, lpfc_irq_chann);
 	lpfc_enable_bbcr_init(phba, lpfc_enable_bbcr);
diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index eaaef682de251..2fd8f15f99975 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -4308,10 +4308,12 @@ lpfc_create_port(struct lpfc_hba *phba, int instance, struct device *dev)
 	shost->max_cmd_len = 16;
 
 	if (phba->sli_rev == LPFC_SLI_REV4) {
-		if (phba->cfg_fcp_io_sched == LPFC_FCP_SCHED_BY_HDWQ)
-			shost->nr_hw_queues = phba->cfg_hdw_queue;
-		else
-			shost->nr_hw_queues = phba->sli4_hba.num_present_cpu;
+		if (!phba->cfg_fcp_mq_threshold ||
+		    phba->cfg_fcp_mq_threshold > phba->cfg_hdw_queue)
+			phba->cfg_fcp_mq_threshold = phba->cfg_hdw_queue;
+
+		shost->nr_hw_queues = min_t(int, 2 * num_possible_nodes(),
+					    phba->cfg_fcp_mq_threshold);
 
 		shost->dma_boundary =
 			phba->sli4_hba.pc_sli4_params.sge_supp_len-1;
diff --git a/drivers/scsi/lpfc/lpfc_sli4.h b/drivers/scsi/lpfc/lpfc_sli4.h
index 8e4fd1a98023c..986594ec40e2a 100644
--- a/drivers/scsi/lpfc/lpfc_sli4.h
+++ b/drivers/scsi/lpfc/lpfc_sli4.h
@@ -44,6 +44,11 @@
 #define LPFC_HBA_HDWQ_MAX	128
 #define LPFC_HBA_HDWQ_DEF	0
 
+/* FCP MQ queue count limiting */
+#define LPFC_FCP_MQ_THRESHOLD_MIN	0
+#define LPFC_FCP_MQ_THRESHOLD_MAX	128
+#define LPFC_FCP_MQ_THRESHOLD_DEF	8
+
 /* Common buffer size to accomidate SCSI and NVME IO buffers */
 #define LPFC_COMMON_IO_BUF_SZ	768
 
-- 
2.20.1