Skip to contentSkip to conversation

pm80xx updates.

PATCH Applied Viswas G 2 participants 11 messages Jan 9, 2021 – Jan 15, 2021

Subsystems (2)
lore
Share & tools
lore

Viswas G

author
#0 lore raw

v2 cover pm80xx updates.

This patch set include some bug fixes and enhancements for pm80xx driver.

Changes from v1:
Corrected email id.

Bhavesh Jashnani (1):
pm80xx: Simultaneous poll for all FW readiness.

Vishakha Channapattan (2):
pm80xx: Log SATA IOMB completion status on failure.
pm80xx: Add sysfs attribute for ioc health

Viswas G (1):
pm80xx: fix driver fatal dump failure.

akshatzen (4):
pm80xx: No busywait in MPI init check
pm80xx: check fatal error
pm80xx: check main config table address
pm80xx: fix missing tag_free in NVMD DATA req

 drivers/scsi/pm8001/pm8001_ctl.c  |  42 ++++++++
 drivers/scsi/pm8001/pm8001_hwi.c  |  15 ++-
 drivers/scsi/pm8001/pm8001_init.c |  11 ++-
 drivers/scsi/pm8001/pm8001_sas.c  |   9 ++
 drivers/scsi/pm8001/pm8001_sas.h  |   2 +
 drivers/scsi/pm8001/pm80xx_hwi.c  | 202 ++++++++++++++++++++++++--------------
 drivers/scsi/pm8001/pm80xx_hwi.h  |  17 +++-
 7 files changed, 216 insertions(+), 82 deletions(-)
Signature · 1 line
2.16.3
author
#1 lore raw

PATCH v2 1/8 pm80xx: No busywait in MPI init check

From: akshatzen <[email protected]>

We do not need to busy wait during mpi_init_check. I confirmed that mpi_init_check is not being invoked in an ATOMIC context. It is being called from pm8001_pci_resume, pm8001_pci_probe. Hence we are replacing the udelay which busy waits with msleep.

---
 drivers/scsi/pm8001/pm80xx_hwi.c | 6 +++---
 drivers/scsi/pm8001/pm80xx_hwi.h | 4 ++--
 2 files changed, 5 insertions(+), 5 deletions(-)
drivers/scsi/pm8001/pm80xx_hwi.c +3 -3
index 6772b0924dac..9c4b8b374ab8 100644
@@ -997,7 +997,7 @@ static int mpi_init_check(struct pm8001_hba_info *pm8001_ha)
 		max_wait_count = SPC_DOORBELL_CLEAR_TIMEOUT;
 	}
 	do {
-		udelay(1);
+		msleep(FW_READY_INTERVAL);
 		value = pm8001_cr32(pm8001_ha, 0, MSGU_IBDB_SET);
 		value &= SPCv_MSGU_CFG_TABLE_UPDATE;
 	} while ((value != 0) && (--max_wait_count));
@@ -1010,9 +1010,9 @@ static int mpi_init_check(struct pm8001_hba_info *pm8001_ha)
 		return -EBUSY;
 	}
 	/* check the MPI-State for initialization upto 100ms*/
-	max_wait_count = 100 * 1000;/* 100 msec */
+	max_wait_count = 5;/* 100 msec */
 	do {
-		udelay(1);
+		msleep(FW_READY_INTERVAL);
 		gst_len_mpistate =
 			pm8001_mr32(pm8001_ha->general_stat_tbl_addr,
 					GST_GSTLEN_MPIS_OFFSET);
drivers/scsi/pm8001/pm80xx_hwi.h +2 -2
index ec48bc276de6..2b6b52551968 100644
@@ -220,8 +220,8 @@
 #define SAS_DOPNRJT_RTRY_TMO            128
 #define SAS_COPNRJT_RTRY_TMO            128
 
-#define SPCV_DOORBELL_CLEAR_TIMEOUT	(30 * 1000 * 1000) /* 30 sec */
-#define SPC_DOORBELL_CLEAR_TIMEOUT	(15 * 1000 * 1000) /* 15 sec */
+#define SPCV_DOORBELL_CLEAR_TIMEOUT	(30 * 50) /* 30 sec */
+#define SPC_DOORBELL_CLEAR_TIMEOUT	(15 * 50) /* 15 sec */
 
 /*
   Making ORR bigger than IT NEXUS LOSS which is 2000000us = 2 second.
Signature · 1 line
2.16.3
author
#2 lore raw

PATCH v2 2/8 pm80xx: check fatal error

In reply to Viswas G

From: akshatzen <[email protected]>

When controller runs into fatal error, commands which expect response get stuck due to no response. If the controller is in fatal error state, abort request issued to the controller gets hung too. Hence we should fail it without trying.

---
 drivers/scsi/pm8001/pm8001_hwi.c |  1 +
 drivers/scsi/pm8001/pm8001_sas.c |  9 +++++++++
 drivers/scsi/pm8001/pm8001_sas.h |  2 ++
 drivers/scsi/pm8001/pm80xx_hwi.c | 36 ++++++++++++++++++++++++++++++++++++
 drivers/scsi/pm8001/pm80xx_hwi.h | 13 +++++++++++++
 5 files changed, 61 insertions(+)
drivers/scsi/pm8001/pm8001_hwi.c +1 -0
index c8d4d87c5473..f147193d67bd 100644
@@ -4998,4 +4998,5 @@ const struct pm8001_dispatch pm8001_8001_dispatch = {
 	.fw_flash_update_req	= pm8001_chip_fw_flash_update_req,
 	.set_dev_state_req	= pm8001_chip_set_dev_state_req,
 	.sas_re_init_req	= pm8001_chip_sas_re_initialization,
+	.fatal_errors		= pm80xx_fatal_errors,
 };
drivers/scsi/pm8001/pm8001_sas.c +9 -0
index d1e9dba2ef19..f8d142f9b9ad 100644
@@ -1183,12 +1183,21 @@ int pm8001_abort_task(struct sas_task *task)
 	int rc = TMF_RESP_FUNC_FAILED, ret;
 	u32 phy_id;
 	struct sas_task_slow slow_task;
+
 	if (unlikely(!task || !task->lldd_task || !task->dev))
 		return TMF_RESP_FUNC_FAILED;
+
 	dev = task->dev;
 	pm8001_dev = dev->lldd_dev;
 	pm8001_ha = pm8001_find_ha_by_dev(dev);
 	phy_id = pm8001_dev->attached_phy;
+
+	if (PM8001_CHIP_DISP->fatal_errors(pm8001_ha)) {
+		// If the controller is seeing fatal errors
+		// abort task will not get a response from the controller
+		return TMF_RESP_FUNC_FAILED;
+	}
+
 	ret = pm8001_find_tag(task, &tag);
 	if (ret == 0) {
 		pm8001_info(pm8001_ha, "no tag for task:%p\n", task);
drivers/scsi/pm8001/pm8001_sas.h +2 -0
index f2c8cbad3853..039ed91e9841 100644
@@ -215,6 +215,7 @@ struct pm8001_dispatch {
 	int (*sas_diag_execute_req)(struct pm8001_hba_info *pm8001_ha,
 		u32 state);
 	int (*sas_re_init_req)(struct pm8001_hba_info *pm8001_ha);
+	int (*fatal_errors)(struct pm8001_hba_info *pm8001_ha);
 };
 
 struct pm8001_chip_info {
@@ -725,6 +726,7 @@ ssize_t pm80xx_get_fatal_dump(struct device *cdev,
 ssize_t pm80xx_get_non_fatal_dump(struct device *cdev,
 		struct device_attribute *attr, char *buf);
 ssize_t pm8001_get_gsm_dump(struct device *cdev, u32, char *buf);
+int pm80xx_fatal_errors(struct pm8001_hba_info *pm8001_ha);
 /* ctl shared API */
 extern struct device_attribute *pm8001_host_attrs[];
 
drivers/scsi/pm8001/pm80xx_hwi.c +36 -0
index 9c4b8b374ab8..86a3d483749c 100644
@@ -1525,6 +1525,41 @@ static int mpi_uninit_check(struct pm8001_hba_info *pm8001_ha)
 	return 0;
 }
 
+/**
+ * pm80xx_fatal_errors - returns non zero *ONLY* when fatal errors
+ * @pm8001_ha: our hba card information
+ *
+ * Fatal errors are recoverable only after a host reboot.
+ */
+int
+pm80xx_fatal_errors(struct pm8001_hba_info *pm8001_ha)
+{
+	int ret = 0;
+	u32 scratch_pad_rsvd0 = pm8001_cr32(pm8001_ha, 0,
+					MSGU_HOST_SCRATCH_PAD_6);
+	u32 scratch_pad_rsvd1 = pm8001_cr32(pm8001_ha, 0,
+					MSGU_HOST_SCRATCH_PAD_7);
+	u32 scratch_pad1 = pm8001_cr32(pm8001_ha, 0, MSGU_SCRATCH_PAD_1);
+	u32 scratch_pad2 = pm8001_cr32(pm8001_ha, 0, MSGU_SCRATCH_PAD_2);
+	u32 scratch_pad3 = pm8001_cr32(pm8001_ha, 0, MSGU_SCRATCH_PAD_3);
+
+	if (pm8001_ha->chip_id != chip_8006 &&
+			pm8001_ha->chip_id != chip_8074 &&
+			pm8001_ha->chip_id != chip_8076) {
+		return 0;
+	}
+
+	if (MSGU_SCRATCHPAD1_STATE_FATAL_ERROR(scratch_pad1)) {
+		pm8001_dbg(pm8001_ha, FAIL,
+			"Fatal error SCRATCHPAD1 = 0x%x SCRATCHPAD2 = 0x%x SCRATCHPAD3 = 0x%x SCRATCHPAD_RSVD0 = 0x%x SCRATCHPAD_RSVD1 = 0x%x\n",
+				scratch_pad1, scratch_pad2, scratch_pad3,
+				scratch_pad_rsvd0, scratch_pad_rsvd1);
+		ret = 1;
+	}
+
+	return ret;
+}
+
 /**
  * pm8001_chip_soft_rst - soft reset the PM8001 chip, so that the clear all
  * the FW register status to the originated status.
@@ -4959,4 +4994,5 @@ const struct pm8001_dispatch pm8001_80xx_dispatch = {
 	.set_nvmd_req		= pm8001_chip_set_nvmd_req,
 	.fw_flash_update_req	= pm8001_chip_fw_flash_update_req,
 	.set_dev_state_req	= pm8001_chip_set_dev_state_req,
+	.fatal_errors		= pm80xx_fatal_errors,
 };
drivers/scsi/pm8001/pm80xx_hwi.h +13 -0
index 2b6b52551968..2c8e85cfdbc4 100644
@@ -1368,6 +1368,19 @@ typedef struct SASProtocolTimerConfig SASProtocolTimerConfig_t;
 #define MSGU_HOST_SCRATCH_PAD_6			0x6C
 #define MSGU_HOST_SCRATCH_PAD_7			0x70
 
+#define MSGU_SCRATCHPAD1_RAAE_STATE_ERR(x) ((x & 0x3) == 0x2)
+#define MSGU_SCRATCHPAD1_ILA_STATE_ERR(x) (((x >> 2) & 0x3) == 0x2)
+#define MSGU_SCRATCHPAD1_BOOTLDR_STATE_ERR(x) ((((x >> 4) & 0x7) == 0x7) || \
+						(((x >> 4) & 0x7) == 0x4))
+#define MSGU_SCRATCHPAD1_IOP0_STATE_ERR(x) (((x >> 10) & 0x3) == 0x2)
+#define MSGU_SCRATCHPAD1_IOP1_STATE_ERR(x) (((x >> 12) & 0x3) == 0x2)
+#define MSGU_SCRATCHPAD1_STATE_FATAL_ERROR(x)  \
+			(MSGU_SCRATCHPAD1_RAAE_STATE_ERR(x) ||      \
+			 MSGU_SCRATCHPAD1_ILA_STATE_ERR(x) ||       \
+			 MSGU_SCRATCHPAD1_BOOTLDR_STATE_ERR(x) ||   \
+			 MSGU_SCRATCHPAD1_IOP0_STATE_ERR(x) ||      \
+			 MSGU_SCRATCHPAD1_IOP1_STATE_ERR(x))
+
 /* bit definition for ODMR register */
 #define ODMR_MASK_ALL			0xFFFFFFFF/* mask all
 					interrupt vector */
Signature · 1 line
2.16.3
author
#3 lore raw

PATCH v2 3/8 pm80xx: check main config table address

In reply to Viswas G

From: akshatzen <[email protected]>

pm8001 driver initializes main configuration, general status, inbound queue and outbound queue table address based on value read from MSGU_SCRATCH_PAD_0 register.

We should validate these addresses before dereferencing them.

This change adds two validations:
1. Check if main configuration table offset lies within the pcibar mapped
2. Check if first dword of main configuration table reads "PMCS"

There are two calls to init_pci_device_addresses() done during pm8001_pci_probe() in this sequence:
1. First inside chip_soft_rst, where if init_pci_device_addresses fails we will go ahead assuming MPI state is not ready and reset the device as long as bootloader is okay. This gives chance to second call of init_pci_device_addresses to set up the addresses after reset.

2. The second call is via pm80xx_chip_init, after soft reset is done and firmware is checked to be ready. Once that is done we are safe to go ahead and initialize default table values and use them.

Tested:

1. Enabled debugging logs and observed no issues during initialization, with a controller with No issues.

pm80xx0:: pm8001_setup_msix 1034: pci_alloc_irq_vectors request ret:64 no of intr 64
pm80xx0:: init_pci_device_addresses 917: Scratchpad 0 Offset: 0x2000 value 0x40002000
pm80xx0:: init_pci_device_addresses 925: Scratchpad 0 PCI BAR: 0 pm80xx0:: init_pci_device_addresses 952: VALID main config signature 0x53434d50
pm80xx0:: init_pci_device_addresses 975: GST OFFSET 0xc4 pm80xx0:: init_pci_device_addresses 978: INBND OFFSET 0x20000128 pm80xx0:: init_pci_device_addresses 981: OBND OFFSET 0x24000928 pm80xx0:: init_pci_device_addresses 984: IVT OFFSET 0x8001408 pm80xx0:: init_pci_device_addresses 987: PSPA OFFSET 0x8001608 pm80xx0:: init_pci_device_addresses 991: addr - main cfg (ptrval) general status (ptrval)
pm80xx0:: init_pci_device_addresses 995: addr - inbnd (ptrval) obnd (ptrval)
pm80xx0:: init_pci_device_addresses 999: addr - pspa (ptrval) ivt (ptrval)
pm80xx0:: pm80xx_chip_soft_rst 1446: reset register before write : 0x0 pm80xx0:: pm80xx_chip_soft_rst 1478: reset register after write 0x40 pm80xx0:: pm80xx_chip_soft_rst 1544: SPCv soft reset Complete pm80xx0:: init_pci_device_addresses 917: Scratchpad 0 Offset: 0x2000 value 0x40002000
pm80xx0:: init_pci_device_addresses 925: Scratchpad 0 PCI BAR: 0 pm80xx0:: init_pci_device_addresses 952: VALID main config signature 0x53434d50
pm80xx0:: init_pci_device_addresses 975: GST OFFSET 0xc4 pm80xx0:: init_pci_device_addresses 978: INBND OFFSET 0x20000128 pm80xx0:: init_pci_device_addresses 981: OBND OFFSET 0x24000928 pm80xx0:: init_pci_device_addresses 984: IVT OFFSET 0x8001408 pm80xx0:: init_pci_device_addresses 987: PSPA OFFSET 0x8001608 pm80xx0:: init_pci_device_addresses 991: addr - main cfg (ptrval) general status (ptrval)
pm80xx0:: init_pci_device_addresses 995: addr - inbnd (ptrval) obnd (ptrval)
pm80xx0:: init_pci_device_addresses 999: addr - pspa (ptrval) ivt (ptrval)
pm80xx0:: pm80xx_chip_init 1329: MPI initialize successful!

2. Tested controller with firmware known to have initialization issue and observed no crashes with this fix

pm80xx 0000:01:00.0: pm80xx: driver version 0.1.38
pm80xx 0000:01:00.0: Removing from 1:1 domain
pm80xx 0000:01:00.0: Requesting non-1:1 mappings
pm80xx0:: init_pci_device_addresses 948: BAD main config signature 0x0 pm80xx0:: mpi_uninit_check 1365: Failed to init pci addresses pm80xx0:: pm80xx_chip_soft_rst 1435: MPI state is not ready scratch:0:8:62a01000:0
pm80xx0:: pm80xx_chip_soft_rst 1518: Firmware is not ready! pm80xx0:: pm80xx_chip_soft_rst 1532: iButton Feature is not Available!!! pm80xx0:: pm80xx_chip_init 1301: Firmware is not ready! pm80xx0:: pm8001_pci_probe 1215: chip_init failed [ret: -16]
pm80xx: probe of 0000:01:00.0 failed with error -16
pm80xx 0000:07:00.0: pm80xx: driver version 0.1.38
pm80xx 0000:07:00.0: Removing from 1:1 domain
pm80xx 0000:07:00.0: Requesting non-1:1 mappings
scsi host6: pm80xx
pm80xx1:: pm8001_setup_sgpio 5568: failed sgpio_req timeout pm80xx1:: mpi_phy_start_resp 3447: phy start resp status:0x0, phyid:0x0 pm80xx 0000:08:00.0: pm80xx: driver version 0.1.38
pm80xx 0000:08:00.0: Removing from 1:1 domain
pm80xx 0000:08:00.0: Requesting non-1:1 mappings

3. Without this fix we observe crash on the same controller.

pm80xx 0000:01:00.0: pm80xx: driver version 0.1.38 pm80xx 0000:01:00.0: Removing from 1:1 domain pm80xx 0000:01:00.0: Requesting non-1:1 mappings

[<ffffffffc0451b3b>] pm80xx_chip_soft_rst+0x6b/0x4c0 [pm80xx]
[<ffffffffc043a933>] pm8001_pci_probe+0xa43/0x1630 [pm80xx]
RIP: 0010:pm80xx_chip_soft_rst+0x71/0x4c0 [pm80xx]
[<ffffffffc0451b3b>] ? pm80xx_chip_soft_rst+0x6b/0x4c0 [pm80xx]
[<ffffffffc043a933>] pm8001_pci_probe+0xa43/0x1630 [pm80xx]

pm80xx0:: mpi_uninit_check 1339: TIMEOUT:IBDB value/=2 pm80xx0:: pm80xx_chip_soft_rst 1387: MPI state is not ready scratch:0:8:62a01000:0
pm80xx0:: pm80xx_chip_soft_rst 1470: Firmware is not ready! pm80xx0:: pm80xx_chip_soft_rst 1484: iButton Feature is not Available!!! pm80xx0:: pm80xx_chip_init 1266: Firmware is not ready! pm80xx0:: pm8001_pci_probe 1207: chip_init failed [ret: -16]
pm80xx: probe of 0000:01:00.0 failed with error -16

---
 drivers/scsi/pm8001/pm8001_init.c | 11 +++++---
 drivers/scsi/pm8001/pm80xx_hwi.c  | 53 ++++++++++++++++++++++++++++++++++++---
 2 files changed, 58 insertions(+), 6 deletions(-)
drivers/scsi/pm8001/pm8001_init.c +8 -3
index ee2de177d0d0..d21078ca7fb3 100644
@@ -466,9 +466,12 @@ static int pm8001_ioremap(struct pm8001_hba_info *pm8001_ha)
 			pm8001_ha->io_mem[logicalBar].memvirtaddr =
 				ioremap(pm8001_ha->io_mem[logicalBar].membase,
 				pm8001_ha->io_mem[logicalBar].memsize);
-			pm8001_dbg(pm8001_ha, INIT,
-				   "PCI: bar %d, logicalBar %d\n",
+			if (!pm8001_ha->io_mem[logicalBar].memvirtaddr) {
+				pm8001_dbg(pm8001_ha, INIT,
+					"Failed to ioremap bar %d, logicalBar %d",
 				   bar, logicalBar);
+				return -ENOMEM;
+			}
 			pm8001_dbg(pm8001_ha, INIT,
 				   "base addr %llx virt_addr=%llx len=%d\n",
 				   (u64)pm8001_ha->io_mem[logicalBar].membase,
@@ -540,9 +543,11 @@ static struct pm8001_hba_info *pm8001_pci_alloc(struct pci_dev *pdev,
 			tasklet_init(&pm8001_ha->tasklet[j], pm8001_tasklet,
 				(unsigned long)&(pm8001_ha->irq_vector[j]));
 #endif
-	pm8001_ioremap(pm8001_ha);
+	if (pm8001_ioremap(pm8001_ha))
+		goto failed_pci_alloc;
 	if (!pm8001_alloc(pm8001_ha, ent))
 		return pm8001_ha;
+failed_pci_alloc:
 	pm8001_free(pm8001_ha);
 	return NULL;
 }
drivers/scsi/pm8001/pm80xx_hwi.c +50 -3

92 lines not on this page · raw message

Signature · 1 line
2.16.3
author
#4 lore raw

PATCH v2 4/8 pm80xx: fix missing tag_free in NVMD DATA req

In reply to Viswas G

From: akshatzen <[email protected]>

Tag is not free'd in NVMD get/set data request failure scenario, which would have caused tag leak each time the request fails.

---
 drivers/scsi/pm8001/pm8001_hwi.c | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)
drivers/scsi/pm8001/pm8001_hwi.c +10 -4
index f147193d67bd..9cd6a654f8b2 100644
@@ -3038,8 +3038,8 @@ void pm8001_mpi_set_nvmd_resp(struct pm8001_hba_info *pm8001_ha, void *piomb)
 	complete(pm8001_ha->nvmd_completion);
 	pm8001_dbg(pm8001_ha, MSG, "Set nvm data complete!\n");
 	if ((dlen_status & NVMD_STAT) != 0) {
-		pm8001_dbg(pm8001_ha, FAIL, "Set nvm data error!\n");
-		return;
+		pm8001_dbg(pm8001_ha, FAIL, "Set nvm data error %x\n",
+				dlen_status);
 	}
 	ccb->task = NULL;
 	ccb->ccb_tag = 0xFFFFFFFF;
@@ -3062,11 +3062,17 @@ pm8001_mpi_get_nvmd_resp(struct pm8001_hba_info *pm8001_ha, void *piomb)
 
 	pm8001_dbg(pm8001_ha, MSG, "Get nvm data complete!\n");
 	if ((dlen_status & NVMD_STAT) != 0) {
-		pm8001_dbg(pm8001_ha, FAIL, "Get nvm data error!\n");
+		pm8001_dbg(pm8001_ha, FAIL, "Get nvm data error %x\n",
+				dlen_status);
 		complete(pm8001_ha->nvmd_completion);
+		/* We should free tag during failure also, the tag is not being
+		 * free'd by requesting path anywhere.
+		 */
+		ccb->task = NULL;
+		ccb->ccb_tag = 0xFFFFFFFF;
+		pm8001_tag_free(pm8001_ha, tag);
 		return;
 	}
-
 	if (ir_tds_bn_dps_das_nvm & IPMode) {
 		/* indirect mode - IR bit set */
 		pm8001_dbg(pm8001_ha, MSG, "Get NVMD success, IR=1\n");
Signature · 1 line
2.16.3
author
#5 lore raw

PATCH v2 5/8 pm80xx: fix driver fatal dump failure.

In reply to Viswas G

The fatal dump function pm80xx_get_fatal_dump() has two issues that result in the fatal dump not being completed successfully.

1. When trying collect fatal_logs from the application it is getting failed, because we are not shifting MEMBASE-II register properly. Once we read 64K region of data we have to shift the MEMBASE-II register and read the next chunk of data, then only we would be able to get complete data.

2. If timeout occurs our application will get stuck because we are not handling this case. In this patch it resolves all these issues.

---
 drivers/scsi/pm8001/pm80xx_hwi.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)
drivers/scsi/pm8001/pm80xx_hwi.c +9 -5
index 7d0eada11d3c..407c0cf6ab5f 100644
@@ -349,10 +349,15 @@ ssize_t pm80xx_get_fatal_dump(struct device *cdev,
 				sprintf(
 				pm8001_ha->forensic_info.data_buf.direct_data,
 				"%08x ", 0xFFFFFFFF);
-				pm8001_cw32(pm8001_ha, 0,
+				return((char *)pm8001_ha->forensic_info.data_buf.direct_data -
+						(char *)buf);
+			}
+	/* reset fatal_forensic_shift_offset back to zero and reset MEMBASE 2 register to zero */
+			pm8001_ha->fatal_forensic_shift_offset = 0; /* location in 64k region */
+			pm8001_cw32(pm8001_ha, 0,
 					MEMBASE_II_SHIFT_REGISTER,
 					pm8001_ha->fatal_forensic_shift_offset);
-			}
+		}
 			/* Read the next block of the debug data.*/
 			length_to_read = pm8001_mr32(fatal_table_address,
 			MPI_FATAL_EDUMP_TABLE_ACCUM_LEN) -
@@ -373,13 +378,12 @@ ssize_t pm80xx_get_fatal_dump(struct device *cdev,
 								= 0;
 				pm8001_ha->forensic_info.data_buf.read_len = 0;
 			}
-		}
 	}
 	offset = (int)((char *)pm8001_ha->forensic_info.data_buf.direct_data
 			- (char *)buf);
 	pm8001_dbg(pm8001_ha, IO, "get_fatal_spcv: return4 0x%x\n", offset);
-	return (char *)pm8001_ha->forensic_info.data_buf.direct_data -
-		(char *)buf;
+	return ((char *)pm8001_ha->forensic_info.data_buf.direct_data -
+		(char *)buf);
 }
 
 /* pm80xx_get_non_fatal_dump - dump the nonfatal data from the dma
Signature · 1 line
2.16.3
author
#6 lore raw

PATCH v2 6/8 pm80xx: Simultaneous poll for all FW readiness.

In reply to Viswas G

From: Bhavesh Jashnani <[email protected]>

In check_fw_ready() we first wait for ILA to come up and then we wait for RAAE to come up and IOPs and so on. This is a sequential check. Because of this ILA image seems to be not ready in the allocated time and so the driver marks it as "not ready" and then move on to other FW images. But ILA does become ready eventually, but is not checked again. In this way driver concludes that FW is not ready, when it actually is.

Fix: Instead of sequentially polling each image, we keep polling for all images to be ready. The timeout for the polling has been set to the sum of what was used for each individual image.

---
 drivers/scsi/pm8001/pm80xx_hwi.c | 80 ++++++++++++----------------------------
 1 file changed, 23 insertions(+), 57 deletions(-)
drivers/scsi/pm8001/pm80xx_hwi.c +23 -57
index 407c0cf6ab5f..df679e36954a 100644
@@ -1043,6 +1043,7 @@ static int check_fw_ready(struct pm8001_hba_info *pm8001_ha)
 	u32 value;
 	u32 max_wait_count;
 	u32 max_wait_time;
+	u32 expected_mask;
 	int ret = 0;
 
 	/* reset / PCIe ready */
@@ -1052,70 +1053,35 @@ static int check_fw_ready(struct pm8001_hba_info *pm8001_ha)
 		value = pm8001_cr32(pm8001_ha, 0, MSGU_SCRATCH_PAD_1);
 	} while ((value == 0xFFFFFFFF) && (--max_wait_count));
 
-	/* check ila status */
-	max_wait_time = max_wait_count = 50;	/* 1000 milli sec */
-	do {
-		msleep(FW_READY_INTERVAL);
-		value = pm8001_cr32(pm8001_ha, 0, MSGU_SCRATCH_PAD_1);
-	} while (((value & SCRATCH_PAD_ILA_READY) !=
-			SCRATCH_PAD_ILA_READY) && (--max_wait_count));
-	if (!max_wait_count)
-		ret = -1;
-	else {
-		pm8001_dbg(pm8001_ha, MSG,
-			   " ila ready status in %d millisec\n",
-			   (max_wait_time - max_wait_count));
-	}
-
-	/* check RAAE status */
-	max_wait_time = max_wait_count = 90;	/* 1800 milli sec */
-	do {
-		msleep(FW_READY_INTERVAL);
-		value = pm8001_cr32(pm8001_ha, 0, MSGU_SCRATCH_PAD_1);
-	} while (((value & SCRATCH_PAD_RAAE_READY) !=
-				SCRATCH_PAD_RAAE_READY) && (--max_wait_count));
-	if (!max_wait_count)
-		ret = -1;
-	else {
-		pm8001_dbg(pm8001_ha, MSG,
-			   " raae ready status in %d millisec\n",
-			   (max_wait_time - max_wait_count));
+	/* check ila, RAAE and iops status */
+	if ((pm8001_ha->chip_id != chip_8008) &&
+			(pm8001_ha->chip_id != chip_8009)) {
+		max_wait_time = max_wait_count = 180;   /* 3600 milli sec */
+		expected_mask = SCRATCH_PAD_ILA_READY |
+			SCRATCH_PAD_RAAE_READY |
+			SCRATCH_PAD_IOP0_READY |
+			SCRATCH_PAD_IOP1_READY;
+	} else {
+		max_wait_time = max_wait_count = 170;   /* 3400 milli sec */
+		expected_mask = SCRATCH_PAD_ILA_READY |
+			SCRATCH_PAD_RAAE_READY |
+			SCRATCH_PAD_IOP0_READY;
 	}
-
-	/* check iop0 status */
-	max_wait_time = max_wait_count = 30;	/* 600 milli sec */
 	do {
 		msleep(FW_READY_INTERVAL);
 		value = pm8001_cr32(pm8001_ha, 0, MSGU_SCRATCH_PAD_1);
-	} while (((value & SCRATCH_PAD_IOP0_READY) != SCRATCH_PAD_IOP0_READY) &&
-			(--max_wait_count));
-	if (!max_wait_count)
+	} while (((value & expected_mask) !=
+				 expected_mask) && (--max_wait_count));
+	if (!max_wait_count) {
+		pm8001_dbg(pm8001_ha, INIT,
+		"At least one FW component failed to load within %d millisec: Scratchpad1: 0x%x\n",
+			max_wait_time * FW_READY_INTERVAL, value);
 		ret = -1;
-	else {
+	} else {
 		pm8001_dbg(pm8001_ha, MSG,
-			   " iop0 ready status in %d millisec\n",
-			   (max_wait_time - max_wait_count));
+			"All FW components ready by %d ms\n",
+			(max_wait_time - max_wait_count) * FW_READY_INTERVAL);
 	}
-
-	/* check iop1 status only for 16 port controllers */
-	if ((pm8001_ha->chip_id != chip_8008) &&
-			(pm8001_ha->chip_id != chip_8009)) {
-		/* 200 milli sec */
-		max_wait_time = max_wait_count = 10;
-		do {
-			msleep(FW_READY_INTERVAL);
-			value = pm8001_cr32(pm8001_ha, 0, MSGU_SCRATCH_PAD_1);
-		} while (((value & SCRATCH_PAD_IOP1_READY) !=
-				SCRATCH_PAD_IOP1_READY) && (--max_wait_count));
-		if (!max_wait_count)
-			ret = -1;
-		else {
-			pm8001_dbg(pm8001_ha, MSG,
-				   "iop1 ready status in %d millisec\n",
-				   (max_wait_time - max_wait_count));
-		}
-	}
-
 	return ret;
 }
 
Signature · 1 line
2.16.3
author
#7 lore raw

PATCH v2 7/8 pm80xx: Log SATA IOMB completion status on failure.

In reply to Viswas G

From: Vishakha Channapattan <[email protected]>

Added a log message in sata completion path to log the status of failed command. If the status does not match any expected status, another message will be logged.

On IO failure with known status, log message will be

[ 1712.951735] pm80xx0:: mpi_sata_completion 2269: IO failed device_id

16385 status 0x1 tag XX

If the firmware returns unexpected status, log message of the following format will be logged -

[ 1712.951735] pm80xx0:: mpi_sata_completion XXXX: Unknown status

device_id XXXXX status 0xX tag XX

---
 drivers/scsi/pm8001/pm80xx_hwi.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)
drivers/scsi/pm8001/pm80xx_hwi.c +8 -5
index df679e36954a..e7fef42b4f6c 100644
@@ -2437,10 +2437,11 @@ mpi_sata_completion(struct pm8001_hba_info *pm8001_ha, void *piomb)
 		return;
 	}
 
-	if (unlikely(status))
-		pm8001_dbg(pm8001_ha, IOERR,
-			   "status:0x%x, tag:0x%x, task::0x%p\n",
-			   status, tag, t);
+	if (status != IO_SUCCESS) {
+		pm8001_dbg(pm8001_ha, FAIL,
+			"IO failed device_id %u status 0x%x tag %d\n",
+			pm8001_dev->device_id, status, tag);
+	}
 
 	/* Print sas address of IO failed device */
 	if ((status != IO_SUCCESS) && (status != IO_OVERFLOW) &&
@@ -2762,7 +2763,9 @@ mpi_sata_completion(struct pm8001_hba_info *pm8001_ha, void *piomb)
 			atomic_dec(&pm8001_dev->running_req);
 		break;
 	default:
-		pm8001_dbg(pm8001_ha, DEVIO, "Unknown status 0x%x\n", status);
+		pm8001_dbg(pm8001_ha, DEVIO,
+				"Unknown status device_id %u status 0x%x tag %d\n",
+			pm8001_dev->device_id, status, tag);
 		/* not allowed case. Therefore, return failed status */
 		ts->resp = SAS_TASK_COMPLETE;
 		ts->stat = SAS_DEV_NO_RESPONSE;
Signature · 1 line
2.16.3
author
#8 lore raw

PATCH v2 8/8 pm80xx: Add sysfs attribute for ioc health

In reply to Viswas G

From: Vishakha Channapattan <[email protected]>

A new sysfs variable 'health' is being introduced that tells if the controller is alive by indicating controller ticks. If on subsequent run we see the ticks changing that indicates that controller is not dead.

Tested: Using 'health' sysfs variable we can see ticks incrementing

mvae14:~# cat  /sys/class/scsi_host/host*/health
MPI-S= MPI is successfully initialized   HMI_ERR=0
MSGUTCNT = 0x00000169 IOPTCNT=0x0000016a IOP1TCNT=0x0000016a
MPI-S= MPI is successfully initialized   HMI_ERR=0
MSGUTCNT = 0x0000014d IOPTCNT=0x0000014d IOP1TCNT=0x0000014d
MPI-S= MPI is successfully initialized   HMI_ERR=0
MSGUTCNT = 0x00000149 IOPTCNT=0x00000149 IOP1TCNT=0x00000149
mvae14:~#
mvae14:~#
mvae14:~#
mvae14:~# cat  /sys/class/scsi_host/host*/health
MPI-S= MPI is successfully initialized   HMI_ERR=0
MSGUTCNT = 0x0000016c IOPTCNT=0x0000016c IOP1TCNT=0x0000016c
MPI-S= MPI is successfully initialized   HMI_ERR=0
MSGUTCNT = 0x0000014f IOPTCNT=0x0000014f IOP1TCNT=0x0000014f
MPI-S= MPI is successfully initialized   HMI_ERR=0
MSGUTCNT = 0x0000014b IOPTCNT=0x0000014b IOP1TCNT=0x0000014b
---
 drivers/scsi/pm8001/pm8001_ctl.c | 42 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)
drivers/scsi/pm8001/pm8001_ctl.c +42 -0
index 12035baf0997..f46f341132fb 100644
@@ -41,6 +41,7 @@
 #include <linux/slab.h>
 #include "pm8001_sas.h"
 #include "pm8001_ctl.h"
+#include "pm8001_chips.h"
 
 /* scsi host attributes */
 
@@ -886,6 +887,46 @@ static ssize_t pm8001_show_update_fw(struct device *cdev,
 
 static DEVICE_ATTR(update_fw, S_IRUGO|S_IWUSR|S_IWGRP,
 	pm8001_show_update_fw, pm8001_store_update_fw);
+
+/**
+ * pm8001_ctl_health_show - controller health check
+ * @cdev: pointer to embedded class device
+ * @buf: the buffer returned
+ *
+ * A sysfs 'read-only' shost attribute.
+ */
+
+char mpiStateText[][80] = {
+	"MPI is not initialized",
+	"MPI is successfully initialized",
+	"MPI termination is in progress",
+	"MPI initialization failed with error in [31:16]"
+};
+
+static ssize_t ctl_health_show(struct device *cdev,
+		struct device_attribute *attr, char *buf)
+{
+	struct Scsi_Host *shost = class_to_shost(cdev);
+	struct sas_ha_struct *sha = SHOST_TO_SAS_HA(shost);
+	struct pm8001_hba_info *pm8001_ha = sha->lldd_ha;
+	unsigned int mpiDW0 = 0;
+	unsigned int raaeCnt = 0;
+	unsigned int iop0Cnt = 0;
+	unsigned int iop1Cnt = 0;
+	int c;
+
+	pm8001_dbg(pm8001_ha, IOCTL, "%s\n", __func__);
+	mpiDW0 = pm8001_mr32(pm8001_ha->general_stat_tbl_addr, 0);
+	raaeCnt = pm8001_mr32(pm8001_ha->general_stat_tbl_addr, 12);
+	iop0Cnt = pm8001_mr32(pm8001_ha->general_stat_tbl_addr, 16);
+	iop1Cnt = pm8001_mr32(pm8001_ha->general_stat_tbl_addr, 20);
+	c = sprintf(buf, "MPI-S=%s\t HMI_ERR=%x\nMSGUTCNT=0x%08x IOPTCNT=0x%08x IOP1TCNT=0x%08x\n",
+			mpiStateText[mpiDW0 & 0x0003], ((mpiDW0 & 0xff00) >> 16),
+			raaeCnt, iop0Cnt, iop1Cnt);
+	return c;
+}
+static DEVICE_ATTR_RO(ctl_health);
+
 struct device_attribute *pm8001_host_attrs[] = {
 	&dev_attr_interface_rev,
 	&dev_attr_controller_fatal_error,
@@ -909,6 +950,7 @@ struct device_attribute *pm8001_host_attrs[] = {
 	&dev_attr_ob_log,
 	&dev_attr_ila_version,
 	&dev_attr_inc_fw_ver,
+	&dev_attr_ctl_health,
 	NULL,
 };
 
Signature · 1 line
2.16.3
#9 lore raw
In reply to Viswas G

Viswas,

Quoted 2 lines

This patch set include some bug fixes and enhancements for pm80xx driver.

Applied #1-#7 to 5.12/scsi-staging.

The health monitoring patch needs a rework. We only print one value per file in sysfs. I.e. you'll need to create separate sysfs files for each of the fields you want to monitor.

Thanks!

Signature · 1 line
Martin K. Petersen	Oracle Linux Engineering
#10 lore raw
In reply to Viswas G
Quoted 9 lines · Viswas G

This patch set include some bug fixes and enhancements for pm80xx driver.

Changes from v1:
Corrected email id.

Bhavesh Jashnani (1):
pm80xx: Simultaneous poll for all FW readiness.

[...]

Applied to 5.12/scsi-queue, thanks!

[1/8] pm80xx: No busywait in MPI init check

      https://git.kernel.org/mkp/scsi/c/d71023af4bec
[2/8] pm80xx: check fatal error
      https://git.kernel.org/mkp/scsi/c/a961ea0afd63
[3/8] pm80xx: check main config table address
      https://git.kernel.org/mkp/scsi/c/95652f98b1da
[4/8] pm80xx: fix missing tag_free in NVMD DATA req
      https://git.kernel.org/mkp/scsi/c/5d28026891c7
[5/8] pm80xx: fix driver fatal dump failure.
      https://git.kernel.org/mkp/scsi/c/ec2e7e1afff5
[6/8] pm80xx: Simultaneous poll for all FW readiness.
      https://git.kernel.org/mkp/scsi/c/6b2f2d05b581
[7/8] pm80xx: Log SATA IOMB completion status on failure.
      https://git.kernel.org/mkp/scsi/c/4f608fbce54b
Signature · 1 line
Martin K. Petersen	Oracle Linux Engineering