Skip to contentSkip to conversation

scsi: target/tcmu: Fix queue_cmd_ring() declaration

PATCH In review Sasha Levin 1 participant 6 messages Nov 22, 2019

Subsystems (3)
lore
Share & tools
lore

Sasha Levin

author maintainer
#0 lore raw

PATCH 1/219 scsi: target/tcmu: Fix queue_cmd_ring() declaration

From: Bart Van Assche <[email protected]>

[ Upstream commit e7f411049f5164ee6db6c3434c07302846f09990 ]

This patch does not change any functionality but avoids that sparse complains about the queue_cmd_ring() function and its callers.

Fixes 6fd0ce79724d ("tcmu: prep queue_cmd_ring to be used by unmap wq")
Cc Nicholas Bellinger <[email protected]>
Cc Mike Christie <[email protected]>
Cc Christoph Hellwig <[email protected]>
Cc Hannes Reinecke <[email protected]>
---
 drivers/target/target_core_user.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/target/target_core_user.c +1 -1
index 7159e8363b83b..7ee0a75ce4526 100644
@@ -962,7 +962,7 @@ static int add_to_qfull_queue(struct tcmu_cmd *tcmu_cmd)
  *  0 success
  *  1 internally queued to wait for ring memory to free.
  */
-static sense_reason_t queue_cmd_ring(struct tcmu_cmd *tcmu_cmd, int *scsi_err)
+static int queue_cmd_ring(struct tcmu_cmd *tcmu_cmd, sense_reason_t *scsi_err)
 {
 	struct tcmu_dev *udev = tcmu_cmd->tcmu_dev;
 	struct se_cmd *se_cmd = tcmu_cmd->se_cmd;
Signature · 1 line
2.20.1
author maintainer
#1 lore raw

PATCH 2/219 scsi: lpfc: Fix kernel Oops due to null pring pointers

From: James Smart <[email protected]>

[ Upstream commit 5a9eeff57f340238c39c95d8e7e54c96fc722de7 ]

Driver is hitting null pring pointers in lpfc_do_work().

Pointer assignment occurs based on SLI-revision. If recovering after an error, its possible the sli revision for the port was cleared, making the lpfc_phba_elsring() not return a ring pointer, thus the null pointer.

Add SLI revision checking to lpfc_phba_elsring() and status checking to all callers.

---
 drivers/scsi/lpfc/lpfc.h      | 6 ++++++
 drivers/scsi/lpfc/lpfc_els.c  | 2 ++
 drivers/scsi/lpfc/lpfc_init.c | 7 ++++++-
 drivers/scsi/lpfc/lpfc_sli.c  | 2 ++
 4 files changed, 16 insertions(+), 1 deletion(-)
drivers/scsi/lpfc/lpfc.h +6 -0
index 43732e8d13473..79c3b2f53828e 100644
@@ -1235,6 +1235,12 @@ lpfc_sli_read_hs(struct lpfc_hba *phba)
 static inline struct lpfc_sli_ring *
 lpfc_phba_elsring(struct lpfc_hba *phba)
 {
+	/* Return NULL if sli_rev has become invalid due to bad fw */
+	if (phba->sli_rev != LPFC_SLI_REV4  &&
+	    phba->sli_rev != LPFC_SLI_REV3  &&
+	    phba->sli_rev != LPFC_SLI_REV2)
+		return NULL;
+
 	if (phba->sli_rev == LPFC_SLI_REV4) {
 		if (phba->sli4_hba.els_wq)
 			return phba->sli4_hba.els_wq->pring;
drivers/scsi/lpfc/lpfc_els.c +2 -0
index f3c6801c0b312..3651c0fc88197 100644
@@ -1340,6 +1340,8 @@ lpfc_els_abort_flogi(struct lpfc_hba *phba)
 			Fabric_DID);
 
 	pring = lpfc_phba_elsring(phba);
+	if (unlikely(!pring))
+		return -EIO;
 
 	/*
 	 * Check the txcmplq for an iocb that matches the nport the driver is
drivers/scsi/lpfc/lpfc_init.c +6 -1
index 9acb5b44ce4c1..dae0dd8066f04 100644
@@ -1801,7 +1801,12 @@ lpfc_sli4_port_sta_fn_reset(struct lpfc_hba *phba, int mbx_action,
 	lpfc_offline(phba);
 	/* release interrupt for possible resource change */
 	lpfc_sli4_disable_intr(phba);
-	lpfc_sli_brdrestart(phba);
+	rc = lpfc_sli_brdrestart(phba);
+	if (rc) {
+		lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
+				"6309 Failed to restart board\n");
+		return rc;
+	}
 	/* request and enable interrupt */
 	intr_mode = lpfc_sli4_enable_intr(phba, phba->intr_mode);
 	if (intr_mode == LPFC_INTR_ERROR) {
drivers/scsi/lpfc/lpfc_sli.c +2 -0
index a490e63c94b67..266e7101902ef 100644
@@ -4644,6 +4644,8 @@ lpfc_sli_brdrestart_s4(struct lpfc_hba *phba)
 	hba_aer_enabled = phba->hba_flag & HBA_AER_ENABLED;
 
 	rc = lpfc_sli4_brdreset(phba);
+	if (rc)
+		return rc;
 
 	spin_lock_irq(&phba->hbalock);
 	phba->pport->stopped = 0;
Signature · 1 line
2.20.1
author maintainer
#4 lore raw

PATCH 5/219 ARM: dts: Fix up SQ201 flash access

In reply to Sasha Levin

From: Linus Walleij <[email protected]>

[ Upstream commit d88b11ef91b15d0af9c0676cbf4f441a0dff0c56 ]

This sets the partition information on the SQ201 to be read out from the RedBoot partition table, removes the static partition table and sets our boot options to mount root from /dev/mtdblock2 where the squashfs+JFFS2 resides.

---
 arch/arm/boot/dts/gemini-sq201.dts | 37 ++++--------------------------
 1 file changed, 5 insertions(+), 32 deletions(-)
arch/arm/boot/dts/gemini-sq201.dts +5 -32
index 3787cf3763c41..e9e4a8a02600b 100644
@@ -20,7 +20,7 @@
 	};
 
 	chosen {
-		bootargs = "console=ttyS0,115200n8";
+		bootargs = "console=ttyS0,115200n8 root=/dev/mtdblock2 rw rootfstype=squashfs,jffs2 rootwait";
 		stdout-path = &uart0;
 	};
 
@@ -138,37 +138,10 @@
 			/* 16MB of flash */
 			reg = <0x30000000 0x01000000>;
 
-			partition@0 {
-				label = "RedBoot";
-				reg = <0x00000000 0x00120000>;
-				read-only;
-			};
-			partition@120000 {
-				label = "Kernel";
-				reg = <0x00120000 0x00200000>;
-			};
-			partition@320000 {
-				label = "Ramdisk";
-				reg = <0x00320000 0x00600000>;
-			};
-			partition@920000 {
-				label = "Application";
-				reg = <0x00920000 0x00600000>;
-			};
-			partition@f20000 {
-				label = "VCTL";
-				reg = <0x00f20000 0x00020000>;
-				read-only;
-			};
-			partition@f40000 {
-				label = "CurConf";
-				reg = <0x00f40000 0x000a0000>;
-				read-only;
-			};
-			partition@fe0000 {
-				label = "FIS directory";
-				reg = <0x00fe0000 0x00020000>;
-				read-only;
+			partitions {
+				compatible = "redboot-fis";
+				/* Eraseblock at 0xfe0000 */
+				fis-index-block = <0x1fc>;
 			};
 		};
 
Signature · 1 line
2.20.1
author maintainer
#3 lore raw

PATCH 4/219 arm64: dts: marvell: armada-37xx: Enable emmc on espressobin

In reply to Sasha Levin

From: Ding Tao <[email protected]>

[ Upstream commit 43ebc7c1b3ed8198b9acf3019eca16e722f7331c ]

The ESPRESSObin board has a emmc interface available on U11: declare it and let the bootloader enable it if the emmc is present.

---
 .../dts/marvell/armada-3720-espressobin.dts   | 22 +++++++++++++++++++
 1 file changed, 22 insertions(+)
arch/arm64/boot/dts/marvell/armada-3720-espressobin.dts +22 -0
index 3ab25ad402b90..846003bb480cd 100644
@@ -60,9 +60,31 @@
 	cd-gpios = <&gpionb 3 GPIO_ACTIVE_LOW>;
 	marvell,pad-type = "sd";
 	vqmmc-supply = <&vcc_sd_reg1>;
+
+	pinctrl-names = "default";
+	pinctrl-0 = <&sdio_pins>;
 	status = "okay";
 };
 
+/* U11 */
+&sdhci0 {
+	non-removable;
+	bus-width = <8>;
+	mmc-ddr-1_8v;
+	mmc-hs400-1_8v;
+	marvell,xenon-emmc;
+	marvell,xenon-tun-count = <9>;
+	marvell,pad-type = "fixed-1-8v";
+
+	pinctrl-names = "default";
+	pinctrl-0 = <&mmc_pins>;
+/*
+ * This eMMC is not populated on all boards, so disable it by
+ * default and let the bootloader enable it, if it is present
+ */
+	status = "disabled";
+};
+
 &spi0 {
 	status = "okay";
 
Signature · 7 lines
2.20.1


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
author maintainer
#5 lore raw

PATCH 6/219 tracing: Lock event_mutex before synth_event_mutex

In reply to Sasha Levin

From: Masami Hiramatsu <[email protected]>

[ Upstream commit fc800a10be26017f8f338bc8e500d48e3e6429d9 ]

synthetic event is using synth_event_mutex for protecting synth_event_list, and event_trigger_write() path acquires locks as below order.

event_trigger_write(event_mutex)
->trigger_process_regex(trigger_cmd_mutex)

    ->event_hist_trigger_func(synth_event_mutex)

On the other hand, synthetic event creation and deletion paths call trace_add_event_call() and trace_remove_event_call() which acquires event_mutex. In that case, if we keep the synth_event_mutex locked while registering/unregistering synthetic events, its dependency will be inversed.

To avoid this issue, current synthetic event is using a 2 phase process to create/delete events. For example, it searches existing events under synth_event_mutex to check for event-name conflicts, and unlocks synth_event_mutex, then registers a new event under event_mutex locked. Finally, it locks synth_event_mutex and tries to add the new event to the list. But it can introduce complexity and a chance for name conflicts.

To solve this simpler, this introduces trace_add_event_call_nolock() and trace_remove_event_call_nolock() which don't acquire event_mutex inside. synthetic event can lock event_mutex before synth_event_mutex to solve the lock dependency issue simpler.

---
 include/linux/trace_events.h     |  2 ++
 kernel/trace/trace_events.c      | 34 ++++++++++++++++++++++++++------
 kernel/trace/trace_events_hist.c | 24 ++++++++++------------
 3 files changed, 40 insertions(+), 20 deletions(-)
include/linux/trace_events.h +2 -0
index 78a010e19ed41..0643c083ed862 100644
@@ -529,6 +529,8 @@ extern int trace_event_raw_init(struct trace_event_call *call);
 extern int trace_define_field(struct trace_event_call *call, const char *type,
 			      const char *name, int offset, int size,
 			      int is_signed, int filter_type);
+extern int trace_add_event_call_nolock(struct trace_event_call *call);
+extern int trace_remove_event_call_nolock(struct trace_event_call *call);
 extern int trace_add_event_call(struct trace_event_call *call);
 extern int trace_remove_event_call(struct trace_event_call *call);
 extern int trace_event_get_offsets(struct trace_event_call *call);
kernel/trace/trace_events.c +28 -6
index 7345f5f8f3fe6..017f737237e60 100644
@@ -2302,11 +2302,11 @@ __trace_early_add_new_event(struct trace_event_call *call,
 struct ftrace_module_file_ops;
 static void __add_event_to_tracers(struct trace_event_call *call);
 
-/* Add an additional event_call dynamically */
-int trace_add_event_call(struct trace_event_call *call)
+int trace_add_event_call_nolock(struct trace_event_call *call)
 {
 	int ret;
-	mutex_lock(&event_mutex);
+	lockdep_assert_held(&event_mutex);
+
 	mutex_lock(&trace_types_lock);
 
 	ret = __register_event(call, NULL);
@@ -2314,6 +2314,16 @@ int trace_add_event_call(struct trace_event_call *call)
 		__add_event_to_tracers(call);
 
 	mutex_unlock(&trace_types_lock);
+	return ret;
+}
+
+/* Add an additional event_call dynamically */
+int trace_add_event_call(struct trace_event_call *call)
+{
+	int ret;
+
+	mutex_lock(&event_mutex);
+	ret = trace_add_event_call_nolock(call);
 	mutex_unlock(&event_mutex);
 	return ret;
 }
@@ -2363,17 +2373,29 @@ static int probe_remove_event_call(struct trace_event_call *call)
 	return 0;
 }
 
-/* Remove an event_call */
-int trace_remove_event_call(struct trace_event_call *call)
+/* no event_mutex version */
+int trace_remove_event_call_nolock(struct trace_event_call *call)
 {
 	int ret;
 
-	mutex_lock(&event_mutex);
+	lockdep_assert_held(&event_mutex);
+
 	mutex_lock(&trace_types_lock);
 	down_write(&trace_event_sem);
 	ret = probe_remove_event_call(call);
 	up_write(&trace_event_sem);
 	mutex_unlock(&trace_types_lock);
+
+	return ret;
+}
+
+/* Remove an event_call */
+int trace_remove_event_call(struct trace_event_call *call)
+{
+	int ret;
+
+	mutex_lock(&event_mutex);
+	ret = trace_remove_event_call_nolock(call);
 	mutex_unlock(&event_mutex);
 
 	return ret;
kernel/trace/trace_events_hist.c +10 -14
index dac518977e7d0..11d952650fa72 100644
@@ -914,7 +914,7 @@ static int register_synth_event(struct synth_event *event)
 	call->data = event;
 	call->tp = event->tp;
 
-	ret = trace_add_event_call(call);
+	ret = trace_add_event_call_nolock(call);
 	if (ret) {
 		pr_warn("Failed to register synthetic event: %s\n",
 			trace_event_name(call));
@@ -938,7 +938,7 @@ static int unregister_synth_event(struct synth_event *event)
 	struct trace_event_call *call = &event->call;
 	int ret;
 
-	ret = trace_remove_event_call(call);
+	ret = trace_remove_event_call_nolock(call);
 
 	return ret;
 }
@@ -1015,12 +1015,10 @@ static void add_or_delete_synth_event(struct synth_event *event, int delete)
 	if (delete)
 		free_synth_event(event);
 	else {
-		mutex_lock(&synth_event_mutex);
 		if (!find_synth_event(event->name))
 			list_add(&event->list, &synth_event_list);
 		else
 			free_synth_event(event);
-		mutex_unlock(&synth_event_mutex);
 	}
 }
 
@@ -1032,6 +1030,7 @@ static int create_synth_event(int argc, char **argv)
 	int i, consumed = 0, n_fields = 0, ret = 0;
 	char *name;
 
+	mutex_lock(&event_mutex);
 	mutex_lock(&synth_event_mutex);
 
 	/*
@@ -1104,8 +1103,6 @@ static int create_synth_event(int argc, char **argv)
 		goto err;
 	}
  out:
-	mutex_unlock(&synth_event_mutex);
-
 	if (event) {
 		if (delete_event) {
 			ret = unregister_synth_event(event);
@@ -1115,10 +1112,13 @@ static int create_synth_event(int argc, char **argv)
 			add_or_delete_synth_event(event, ret);
 		}
 	}
+	mutex_unlock(&synth_event_mutex);
+	mutex_unlock(&event_mutex);
 
 	return ret;
  err:
 	mutex_unlock(&synth_event_mutex);
+	mutex_unlock(&event_mutex);
 
 	for (i = 0; i < n_fields; i++)
 		free_synth_field(fields[i]);
@@ -1129,12 +1129,10 @@ static int create_synth_event(int argc, char **argv)
 
 static int release_all_synth_events(void)
 {
-	struct list_head release_events;
 	struct synth_event *event, *e;
 	int ret = 0;
 
-	INIT_LIST_HEAD(&release_events);
-
+	mutex_lock(&event_mutex);
 	mutex_lock(&synth_event_mutex);
 
 	list_for_each_entry(event, &synth_event_list, list) {
@@ -1144,16 +1142,14 @@ static int release_all_synth_events(void)
 		}
 	}
 
-	list_splice_init(&event->list, &release_events);
-
-	mutex_unlock(&synth_event_mutex);
-
-	list_for_each_entry_safe(event, e, &release_events, list) {
+	list_for_each_entry_safe(event, e, &synth_event_list, list) {
 		list_del(&event->list);
 
 		ret = unregister_synth_event(event);
 		add_or_delete_synth_event(event, !ret);
 	}
+	mutex_unlock(&synth_event_mutex);
+	mutex_unlock(&event_mutex);
 
 	return ret;
 }
Signature · 1 line
2.20.1
author maintainer
#2 lore raw

PATCH 3/219 scsi: lpfc: Fix dif and first burst use in write commands

In reply to Sasha Levin

From: James Smart <[email protected]>

[ Upstream commit 7c4042a4d0b7532cfbc90478fd3084b2dab5849e ]

When dif and first burst is used in a write command wqe, the driver was not properly setting fields in the io command request. This resulted in no dif bytes being sent and invalid xfer_rdy's, resulting in the io being aborted by the hardware.

Correct the wqe initializaton when both dif and first burst are used.

---
 drivers/scsi/lpfc/lpfc_scsi.c | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)
drivers/scsi/lpfc/lpfc_scsi.c +18 -0
index 666495f21c246..425b83618a2e5 100644
@@ -2732,6 +2732,7 @@ lpfc_bg_scsi_prep_dma_buf_s3(struct lpfc_hba *phba,
 	int datasegcnt, protsegcnt, datadir = scsi_cmnd->sc_data_direction;
 	int prot_group_type = 0;
 	int fcpdl;
+	struct lpfc_vport *vport = phba->pport;
 
 	/*
 	 * Start the lpfc command prep by bumping the bpl beyond fcp_cmnd
@@ -2837,6 +2838,14 @@ lpfc_bg_scsi_prep_dma_buf_s3(struct lpfc_hba *phba,
 	 */
 	iocb_cmd->un.fcpi.fcpi_parm = fcpdl;
 
+	/*
+	 * For First burst, we may need to adjust the initial transfer
+	 * length for DIF
+	 */
+	if (iocb_cmd->un.fcpi.fcpi_XRdy &&
+	    (fcpdl < vport->cfg_first_burst_size))
+		iocb_cmd->un.fcpi.fcpi_XRdy = fcpdl;
+
 	return 0;
 err:
 	if (lpfc_cmd->seg_cnt)
@@ -3401,6 +3410,7 @@ lpfc_bg_scsi_prep_dma_buf_s4(struct lpfc_hba *phba,
 	int datasegcnt, protsegcnt, datadir = scsi_cmnd->sc_data_direction;
 	int prot_group_type = 0;
 	int fcpdl;
+	struct lpfc_vport *vport = phba->pport;
 
 	/*
 	 * Start the lpfc command prep by bumping the sgl beyond fcp_cmnd
@@ -3516,6 +3526,14 @@ lpfc_bg_scsi_prep_dma_buf_s4(struct lpfc_hba *phba,
 	 */
 	iocb_cmd->un.fcpi.fcpi_parm = fcpdl;
 
+	/*
+	 * For First burst, we may need to adjust the initial transfer
+	 * length for DIF
+	 */
+	if (iocb_cmd->un.fcpi.fcpi_XRdy &&
+	    (fcpdl < vport->cfg_first_burst_size))
+		iocb_cmd->un.fcpi.fcpi_XRdy = fcpdl;
+
 	/*
 	 * If the OAS driver feature is enabled and the lun is enabled for
 	 * OAS, set the oas iocb related flags.
Signature · 1 line
2.20.1