2014-06-25 10:39:58

by Sreekanth Reddy

[permalink] [raw]
Subject: [RESEND][PATCH 08/10][SCSI]mpt2sas: Get IOC_FACTS information using handshake protocol only after HBA card gets into READY or Operational state

The driver would send IOC facts only if HBA is in operational or ready
state. If it is in fault state, a diagnostic reset would be issued. It
would wait for 10 seconds to exit out of reset state. If the HBA continues
to be in reset state, then the HBA wouldn't be claimed by the driver.

Signed-off-by: Sreekanth Reddy <[email protected]>
---
drivers/scsi/mpt2sas/mpt2sas_base.c | 416 ++++++++++++++++++++---------------
1 files changed, 239 insertions(+), 177 deletions(-)

diff --git a/drivers/scsi/mpt2sas/mpt2sas_base.c b/drivers/scsi/mpt2sas/mpt2sas_base.c
index 81f1d58..697c841 100644
--- a/drivers/scsi/mpt2sas/mpt2sas_base.c
+++ b/drivers/scsi/mpt2sas/mpt2sas_base.c
@@ -1737,6 +1737,238 @@ _base_handshake_req_reply_wait(struct MPT2SAS_ADAPTER *ioc, int request_bytes,
}

/**
+ * mpt2sas_base_get_iocstate - Get the current state of a MPT adapter.
+ * @ioc: Pointer to MPT_ADAPTER structure
+ * @cooked: Request raw or cooked IOC state
+ *
+ * Returns all IOC Doorbell register bits if cooked==0, else just the
+ * Doorbell bits in MPI_IOC_STATE_MASK.
+ */
+u32
+mpt2sas_base_get_iocstate(struct MPT2SAS_ADAPTER *ioc, int cooked)
+{
+ u32 s, sc;
+
+ s = readl(&ioc->chip->Doorbell);
+ sc = s & MPI2_IOC_STATE_MASK;
+ return cooked ? sc : s;
+}
+
+/**
+ * _base_wait_on_iocstate - waiting on a particular ioc state
+ * @ioc_state: controller state { READY, OPERATIONAL, or RESET }
+ * @timeout: timeout in second
+ * @sleep_flag: CAN_SLEEP or NO_SLEEP
+ *
+ * Returns 0 for success, non-zero for failure.
+ */
+static int
+_base_wait_on_iocstate(struct MPT2SAS_ADAPTER *ioc, u32 ioc_state, int timeout,
+ int sleep_flag)
+{
+ u32 count, cntdn;
+ u32 current_state;
+
+ count = 0;
+ cntdn = (sleep_flag == CAN_SLEEP) ? 1000*timeout : 2000*timeout;
+ do {
+ current_state = mpt2sas_base_get_iocstate(ioc, 1);
+ if (current_state == ioc_state)
+ return 0;
+ if (count && current_state == MPI2_IOC_STATE_FAULT)
+ break;
+ if (sleep_flag == CAN_SLEEP)
+ msleep(1);
+ else
+ udelay(500);
+ count++;
+ } while (--cntdn);
+
+ return current_state;
+}
+
+/**
+ * _base_diag_reset - the "big hammer" start of day reset
+ * @ioc: per adapter object
+ * @sleep_flag: CAN_SLEEP or NO_SLEEP
+ *
+ * Returns 0 for success, non-zero for failure.
+ */
+static int
+_base_diag_reset(struct MPT2SAS_ADAPTER *ioc, int sleep_flag)
+{
+ u32 host_diagnostic;
+ u32 ioc_state;
+ u32 count;
+ u32 hcb_size;
+
+ printk(MPT2SAS_INFO_FMT "sending diag reset !!\n", ioc->name);
+
+ drsprintk(ioc, printk(MPT2SAS_INFO_FMT "clear interrupts\n",
+ ioc->name));
+
+ count = 0;
+ do {
+ /* Write magic sequence to WriteSequence register
+ * Loop until in diagnostic mode
+ */
+ drsprintk(ioc, printk(MPT2SAS_INFO_FMT "write magic sequence\n"
+ , ioc->name));
+ writel(MPI2_WRSEQ_FLUSH_KEY_VALUE, &ioc->chip->WriteSequence);
+ writel(MPI2_WRSEQ_1ST_KEY_VALUE, &ioc->chip->WriteSequence);
+ writel(MPI2_WRSEQ_2ND_KEY_VALUE, &ioc->chip->WriteSequence);
+ writel(MPI2_WRSEQ_3RD_KEY_VALUE, &ioc->chip->WriteSequence);
+ writel(MPI2_WRSEQ_4TH_KEY_VALUE, &ioc->chip->WriteSequence);
+ writel(MPI2_WRSEQ_5TH_KEY_VALUE, &ioc->chip->WriteSequence);
+ writel(MPI2_WRSEQ_6TH_KEY_VALUE, &ioc->chip->WriteSequence);
+
+ /* wait 100 msec */
+ if (sleep_flag == CAN_SLEEP)
+ msleep(100);
+ else
+ mdelay(100);
+
+ if (count++ > 20)
+ goto out;
+
+ host_diagnostic = readl(&ioc->chip->HostDiagnostic);
+ drsprintk(ioc, printk(MPT2SAS_INFO_FMT
+ "wrote magic sequence: count(%d), host_diagnostic(0x%08x)\n",
+ ioc->name, count, host_diagnostic));
+
+ } while ((host_diagnostic & MPI2_DIAG_DIAG_WRITE_ENABLE) == 0);
+
+ hcb_size = readl(&ioc->chip->HCBSize);
+
+ drsprintk(ioc, printk(MPT2SAS_INFO_FMT "diag reset: issued\n",
+ ioc->name));
+ writel(host_diagnostic | MPI2_DIAG_RESET_ADAPTER,
+ &ioc->chip->HostDiagnostic);
+
+ /*This delay allows the chip PCIe hardware time to finish reset tasks*/
+ if (sleep_flag == CAN_SLEEP)
+ msleep(MPI2_HARD_RESET_PCIE_FIRST_READ_DELAY_MICRO_SEC/1000);
+ else
+ mdelay(MPI2_HARD_RESET_PCIE_FIRST_READ_DELAY_MICRO_SEC/1000);
+
+ /* Approximately 300 second max wait */
+ for (count = 0; count < (300000000 /
+ MPI2_HARD_RESET_PCIE_SECOND_READ_DELAY_MICRO_SEC); count++) {
+
+ host_diagnostic = readl(&ioc->chip->HostDiagnostic);
+
+ if (host_diagnostic == 0xFFFFFFFF)
+ goto out;
+ if (!(host_diagnostic & MPI2_DIAG_RESET_ADAPTER))
+ break;
+
+ /* Wait to pass the second read delay window */
+ if (sleep_flag == CAN_SLEEP)
+ msleep(MPI2_HARD_RESET_PCIE_SECOND_READ_DELAY_MICRO_SEC
+ /1000);
+ else
+ mdelay(MPI2_HARD_RESET_PCIE_SECOND_READ_DELAY_MICRO_SEC
+ /1000);
+ }
+
+ if (host_diagnostic & MPI2_DIAG_HCB_MODE) {
+
+ drsprintk(ioc, printk(MPT2SAS_INFO_FMT
+ "restart adapter assuming HCB Address points to good FW\n",
+ ioc->name));
+ host_diagnostic &= ~MPI2_DIAG_BOOT_DEVICE_SELECT_MASK;
+ host_diagnostic |= MPI2_DIAG_BOOT_DEVICE_SELECT_HCDW;
+ writel(host_diagnostic, &ioc->chip->HostDiagnostic);
+
+ drsprintk(ioc, printk(MPT2SAS_INFO_FMT
+ "re-enable the HCDW\n", ioc->name));
+ writel(hcb_size | MPI2_HCB_SIZE_HCB_ENABLE,
+ &ioc->chip->HCBSize);
+ }
+
+ drsprintk(ioc, printk(MPT2SAS_INFO_FMT "restart the adapter\n",
+ ioc->name));
+ writel(host_diagnostic & ~MPI2_DIAG_HOLD_IOC_RESET,
+ &ioc->chip->HostDiagnostic);
+
+ drsprintk(ioc, printk(MPT2SAS_INFO_FMT
+ "disable writes to the diagnostic register\n", ioc->name));
+ writel(MPI2_WRSEQ_FLUSH_KEY_VALUE, &ioc->chip->WriteSequence);
+
+ drsprintk(ioc, printk(MPT2SAS_INFO_FMT
+ "Wait for FW to go to the READY state\n", ioc->name));
+ ioc_state = _base_wait_on_iocstate(ioc, MPI2_IOC_STATE_READY, 20,
+ sleep_flag);
+ if (ioc_state) {
+ printk(MPT2SAS_ERR_FMT
+ "%s: failed going to ready state(ioc_state=0x%x)\n",
+ ioc->name, __func__, ioc_state);
+ goto out;
+ }
+
+ printk(MPT2SAS_INFO_FMT "diag reset: SUCCESS\n", ioc->name);
+ return 0;
+
+ out:
+ printk(MPT2SAS_ERR_FMT "diag reset: FAILED\n", ioc->name);
+ return -EFAULT;
+}
+
+/**
+ * _base_wait_for_iocstate - Wait until the card is in READY or OPERATIONAL
+ * @ioc: per adapter object
+ * @timeout:
+ * @sleep_flag: CAN_SLEEP or NO_SLEEP
+ *
+ * Returns 0 for success, non-zero for failure.
+ */
+static int
+_base_wait_for_iocstate(struct MPT2SAS_ADAPTER *ioc, int timeout,
+ int sleep_flag)
+{
+ u32 ioc_state;
+ int rc;
+
+ dinitprintk(ioc, printk(MPT2SAS_INFO_FMT "%s\n", ioc->name,
+ __func__));
+
+ if (ioc->pci_error_recovery)
+ return 0;
+
+ ioc_state = mpt2sas_base_get_iocstate(ioc, 0);
+ dhsprintk(ioc, printk(MPT2SAS_INFO_FMT "%s: ioc_state(0x%08x)\n",
+ ioc->name, __func__, ioc_state));
+
+ if (((ioc_state & MPI2_IOC_STATE_MASK) == MPI2_IOC_STATE_READY) ||
+ (ioc_state & MPI2_IOC_STATE_MASK) == MPI2_IOC_STATE_OPERATIONAL)
+ return 0;
+
+ if (ioc_state & MPI2_DOORBELL_USED) {
+ dhsprintk(ioc, printk(MPT2SAS_INFO_FMT
+ "unexpected doorbell activ!e\n", ioc->name));
+ goto issue_diag_reset;
+ }
+
+ if ((ioc_state & MPI2_IOC_STATE_MASK) == MPI2_IOC_STATE_FAULT) {
+ mpt2sas_base_fault_info(ioc, ioc_state &
+ MPI2_DOORBELL_DATA_MASK);
+ goto issue_diag_reset;
+ }
+
+ ioc_state = _base_wait_on_iocstate(ioc, MPI2_IOC_STATE_READY,
+ timeout, sleep_flag);
+ if (ioc_state) {
+ printk(MPT2SAS_ERR_FMT
+ "%s: failed going to ready state (ioc_state=0x%x)\n",
+ ioc->name, __func__, ioc_state);
+ return -EFAULT;
+ }
+
+ issue_diag_reset:
+ rc = _base_diag_reset(ioc, sleep_flag);
+ return rc;
+}
+/**
* _base_get_ioc_facts - obtain ioc facts reply and save in ioc
* @ioc: per adapter object
* @sleep_flag: CAN_SLEEP or NO_SLEEP
@@ -1752,8 +1984,14 @@ _base_get_ioc_facts(struct MPT2SAS_ADAPTER *ioc, int sleep_flag)
int mpi_reply_sz, mpi_request_sz, r;

dinitprintk(ioc, printk(MPT2SAS_INFO_FMT "%s\n", ioc->name,
- __func__));
+ __func__));

+ r = _base_wait_for_iocstate(ioc, 10, sleep_flag);
+ if (r) {
+ printk(MPT2SAS_ERR_FMT "%s: failed getting to correct state\n",
+ ioc->name, __func__);
+ return r;
+ }
mpi_reply_sz = sizeof(Mpi2IOCFactsReply_t);
mpi_request_sz = sizeof(Mpi2IOCFactsRequest_t);
memset(&mpi_request, 0, mpi_request_sz);
@@ -3228,57 +3466,6 @@ chain_done:
}

/**
- * mpt2sas_base_get_iocstate - Get the current state of a MPT adapter.
- * @ioc: Pointer to MPT_ADAPTER structure
- * @cooked: Request raw or cooked IOC state
- *
- * Returns all IOC Doorbell register bits if cooked==0, else just the
- * Doorbell bits in MPI_IOC_STATE_MASK.
- */
-u32
-mpt2sas_base_get_iocstate(struct MPT2SAS_ADAPTER *ioc, int cooked)
-{
- u32 s, sc;
-
- s = readl(&ioc->chip->Doorbell);
- sc = s & MPI2_IOC_STATE_MASK;
- return cooked ? sc : s;
-}
-
-/**
- * _base_wait_on_iocstate - waiting on a particular ioc state
- * @ioc_state: controller state { READY, OPERATIONAL, or RESET }
- * @timeout: timeout in second
- * @sleep_flag: CAN_SLEEP or NO_SLEEP
- *
- * Returns 0 for success, non-zero for failure.
- */
-static int
-_base_wait_on_iocstate(struct MPT2SAS_ADAPTER *ioc, u32 ioc_state, int timeout,
- int sleep_flag)
-{
- u32 count, cntdn;
- u32 current_state;
-
- count = 0;
- cntdn = (sleep_flag == CAN_SLEEP) ? 1000*timeout : 2000*timeout;
- do {
- current_state = mpt2sas_base_get_iocstate(ioc, 1);
- if (current_state == ioc_state)
- return 0;
- if (count && current_state == MPI2_IOC_STATE_FAULT)
- break;
- if (sleep_flag == CAN_SLEEP)
- msleep(1);
- else
- udelay(500);
- count++;
- } while (--cntdn);
-
- return current_state;
-}
-
-/**
* _base_send_ioc_reset - send doorbell reset
* @ioc: per adapter object
* @reset_type: currently only supports: MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET
@@ -4039,131 +4226,6 @@ mpt2sas_base_validate_event_type(struct MPT2SAS_ADAPTER *ioc, u32 *event_type)
}

/**
- * _base_diag_reset - the "big hammer" start of day reset
- * @ioc: per adapter object
- * @sleep_flag: CAN_SLEEP or NO_SLEEP
- *
- * Returns 0 for success, non-zero for failure.
- */
-static int
-_base_diag_reset(struct MPT2SAS_ADAPTER *ioc, int sleep_flag)
-{
- u32 host_diagnostic;
- u32 ioc_state;
- u32 count;
- u32 hcb_size;
-
- printk(MPT2SAS_INFO_FMT "sending diag reset !!\n", ioc->name);
- drsprintk(ioc, printk(MPT2SAS_INFO_FMT "clear interrupts\n",
- ioc->name));
-
- count = 0;
- do {
- /* Write magic sequence to WriteSequence register
- * Loop until in diagnostic mode
- */
- drsprintk(ioc, printk(MPT2SAS_INFO_FMT "write magic "
- "sequence\n", ioc->name));
- writel(MPI2_WRSEQ_FLUSH_KEY_VALUE, &ioc->chip->WriteSequence);
- writel(MPI2_WRSEQ_1ST_KEY_VALUE, &ioc->chip->WriteSequence);
- writel(MPI2_WRSEQ_2ND_KEY_VALUE, &ioc->chip->WriteSequence);
- writel(MPI2_WRSEQ_3RD_KEY_VALUE, &ioc->chip->WriteSequence);
- writel(MPI2_WRSEQ_4TH_KEY_VALUE, &ioc->chip->WriteSequence);
- writel(MPI2_WRSEQ_5TH_KEY_VALUE, &ioc->chip->WriteSequence);
- writel(MPI2_WRSEQ_6TH_KEY_VALUE, &ioc->chip->WriteSequence);
-
- /* wait 100 msec */
- if (sleep_flag == CAN_SLEEP)
- msleep(100);
- else
- mdelay(100);
-
- if (count++ > 20)
- goto out;
-
- host_diagnostic = readl(&ioc->chip->HostDiagnostic);
- drsprintk(ioc, printk(MPT2SAS_INFO_FMT "wrote magic "
- "sequence: count(%d), host_diagnostic(0x%08x)\n",
- ioc->name, count, host_diagnostic));
-
- } while ((host_diagnostic & MPI2_DIAG_DIAG_WRITE_ENABLE) == 0);
-
- hcb_size = readl(&ioc->chip->HCBSize);
-
- drsprintk(ioc, printk(MPT2SAS_INFO_FMT "diag reset: issued\n",
- ioc->name));
- writel(host_diagnostic | MPI2_DIAG_RESET_ADAPTER,
- &ioc->chip->HostDiagnostic);
-
- /* This delay allows the chip PCIe hardware time to finish reset tasks*/
- if (sleep_flag == CAN_SLEEP)
- msleep(MPI2_HARD_RESET_PCIE_FIRST_READ_DELAY_MICRO_SEC/1000);
- else
- mdelay(MPI2_HARD_RESET_PCIE_FIRST_READ_DELAY_MICRO_SEC/1000);
-
- /* Approximately 300 second max wait */
- for (count = 0; count < (300000000 /
- MPI2_HARD_RESET_PCIE_SECOND_READ_DELAY_MICRO_SEC); count++) {
-
- host_diagnostic = readl(&ioc->chip->HostDiagnostic);
-
- if (host_diagnostic == 0xFFFFFFFF)
- goto out;
- if (!(host_diagnostic & MPI2_DIAG_RESET_ADAPTER))
- break;
-
- /* Wait to pass the second read delay window */
- if (sleep_flag == CAN_SLEEP)
- msleep(MPI2_HARD_RESET_PCIE_SECOND_READ_DELAY_MICRO_SEC
- /1000);
- else
- mdelay(MPI2_HARD_RESET_PCIE_SECOND_READ_DELAY_MICRO_SEC
- /1000);
- }
-
- if (host_diagnostic & MPI2_DIAG_HCB_MODE) {
-
- drsprintk(ioc, printk(MPT2SAS_INFO_FMT "restart the adapter "
- "assuming the HCB Address points to good F/W\n",
- ioc->name));
- host_diagnostic &= ~MPI2_DIAG_BOOT_DEVICE_SELECT_MASK;
- host_diagnostic |= MPI2_DIAG_BOOT_DEVICE_SELECT_HCDW;
- writel(host_diagnostic, &ioc->chip->HostDiagnostic);
-
- drsprintk(ioc, printk(MPT2SAS_INFO_FMT
- "re-enable the HCDW\n", ioc->name));
- writel(hcb_size | MPI2_HCB_SIZE_HCB_ENABLE,
- &ioc->chip->HCBSize);
- }
-
- drsprintk(ioc, printk(MPT2SAS_INFO_FMT "restart the adapter\n",
- ioc->name));
- writel(host_diagnostic & ~MPI2_DIAG_HOLD_IOC_RESET,
- &ioc->chip->HostDiagnostic);
-
- drsprintk(ioc, printk(MPT2SAS_INFO_FMT "disable writes to the "
- "diagnostic register\n", ioc->name));
- writel(MPI2_WRSEQ_FLUSH_KEY_VALUE, &ioc->chip->WriteSequence);
-
- drsprintk(ioc, printk(MPT2SAS_INFO_FMT "Wait for FW to go to the "
- "READY state\n", ioc->name));
- ioc_state = _base_wait_on_iocstate(ioc, MPI2_IOC_STATE_READY, 20,
- sleep_flag);
- if (ioc_state) {
- printk(MPT2SAS_ERR_FMT "%s: failed going to ready state "
- " (ioc_state=0x%x)\n", ioc->name, __func__, ioc_state);
- goto out;
- }
-
- printk(MPT2SAS_INFO_FMT "diag reset: SUCCESS\n", ioc->name);
- return 0;
-
- out:
- printk(MPT2SAS_ERR_FMT "diag reset: FAILED\n", ioc->name);
- return -EFAULT;
-}
-
-/**
* _base_make_ioc_ready - put controller in READY state
* @ioc: per adapter object
* @sleep_flag: CAN_SLEEP or NO_SLEEP
--
1.7.1


2014-07-15 17:59:31

by Martin K. Petersen

[permalink] [raw]
Subject: Re: [RESEND][PATCH 08/10][SCSI]mpt2sas: Get IOC_FACTS information using handshake protocol only after HBA card gets into READY or Operational state

>>>>> "Sreekanth" == Reddy, Sreekanth <[email protected]> writes:

Sreekanth> The driver would send IOC facts only if HBA is in operational
Sreekanth> or ready state. If it is in fault state, a diagnostic reset
Sreekanth> would be issued. It would wait for 10 seconds to exit out of
Sreekanth> reset state. If the HBA continues to be in reset state, then
Sreekanth> the HBA wouldn't be claimed by the driver.

Your patch header describes what used to happen. What does the patched
code do?

Also, several of your patches move large functions around within the
file. That makes it really hard to figure out what the actual code
changes are. Please split changes like that in two: One patch that moves
stuff around in bulk without any functional changes and a second that
addresses the issue at hand. That makes reviewing much easier.

--
Martin K. Petersen Oracle Linux Engineering

2014-07-16 07:07:01

by Christoph Hellwig

[permalink] [raw]
Subject: Re: [RESEND][PATCH 08/10][SCSI]mpt2sas: Get IOC_FACTS information using handshake protocol only after HBA card gets into READY or Operational state

On Wed, Jul 16, 2014 at 12:00:22PM +0530, Sreekanth Reddy wrote:
> Next time onwards I will follow as per your suggestions while posting the
> patches to upstream.

Btw, it would be good to resend the series with the various updates from
Martin once he's done with the review. That makes it a lot easier for
me to pick it up.

If you want to make my life even easier send the series using
git-send-patch so that it's properly threaded.

2014-07-16 17:22:51

by Martin K. Petersen

[permalink] [raw]
Subject: Re: [RESEND][PATCH 08/10][SCSI]mpt2sas: Get IOC_FACTS information using handshake protocol only after HBA card gets into READY or Operational state

>>>>> "Christoph" == Christoph Hellwig <[email protected]> writes:

Sreekanth,

I just sent out a patch for the direct I/O type cast issue. Please add
your SoB to that and use it to replace patch 6 when you send out the
final series.

Christoph> Btw, it would be good to resend the series with the various
Christoph> updates from Martin once he's done with the review. That
Christoph> makes it a lot easier for me to pick it up.

I'm at T10 this week so I don't have a ton of time. But I'll try to get
the remaining couple of patches reviewed today or tomorrow. And then you
should send out the entire series again with my Reviewed-by: added to
the remaining patches to make life easier for Christoph.

--
Martin K. Petersen Oracle Linux Engineering