From: Ranjan Kumar <ranjan.kumar@broadcom.com>
To: linux-scsi@vger.kernel.org, martin.petersen@oracle.com
Cc: sathya.prakash@broadcom.com, sreekanth.reddy@broadcom.com,
sumit.saxena@broadcom.com,
Ranjan Kumar <ranjan.kumar@broadcom.com>,
stable@vger.kernel.org
Subject: [PATCH v4 1/2] mpt3sas: Perform additional retries if Doorbell read returns 0
Date: Tue, 29 Aug 2023 14:30:19 +0530 [thread overview]
Message-ID: <20230829090020.5417-2-ranjan.kumar@broadcom.com> (raw)
In-Reply-To: <20230829090020.5417-1-ranjan.kumar@broadcom.com>
[-- Attachment #1: Type: text/plain, Size: 6776 bytes --]
The driver retries certain register reads for 3 times if the
returned value is 0, this was done based on hardware specification
where the controller could possibly return 0 for certain registers
when there is a parallel access to some of other registers from the
firmware due to BMC out of band interactions.Recently it is observed
that in certain systems with increased BMC interactions, the values
returned are 0 even for 3 retries (the proper value is returned
between 4 to 6 retries).Hence this patch changes the retry
count 3 to 30, which is a revised recommendation, to avoid
the out of band conflict.
Fixes: b899202901a8 ("mpt3sas: Add separate function for aero doorbell reads")
Cc: stable@vger.kernel.org
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpt3sas/mpt3sas_base.c | 46 +++++++++++++++++++++--------
drivers/scsi/mpt3sas/mpt3sas_base.h | 1 +
2 files changed, 34 insertions(+), 13 deletions(-)
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c
index 53f5492579cb..5284584e4cd2 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.c
@@ -138,6 +138,9 @@ _base_get_ioc_facts(struct MPT3SAS_ADAPTER *ioc);
static void
_base_clear_outstanding_commands(struct MPT3SAS_ADAPTER *ioc);
+static u32
+_base_readl_ext_retry(const volatile void __iomem *addr);
+
/**
* mpt3sas_base_check_cmd_timeout - Function
* to check timeout and command termination due
@@ -213,6 +216,20 @@ _base_readl_aero(const volatile void __iomem *addr)
return ret_val;
}
+static u32
+_base_readl_ext_retry(const volatile void __iomem *addr)
+{
+ u32 i, ret_val;
+
+ for (i = 0 ; i < 30 ; i++) {
+ ret_val = readl(addr);
+ if (ret_val == 0)
+ continue;
+ }
+
+ return ret_val;
+}
+
static inline u32
_base_readl(const volatile void __iomem *addr)
{
@@ -940,7 +957,7 @@ mpt3sas_halt_firmware(struct MPT3SAS_ADAPTER *ioc)
dump_stack();
- doorbell = ioc->base_readl(&ioc->chip->Doorbell);
+ doorbell = ioc->base_readl_ext_retry(&ioc->chip->Doorbell);
if ((doorbell & MPI2_IOC_STATE_MASK) == MPI2_IOC_STATE_FAULT) {
mpt3sas_print_fault_code(ioc, doorbell &
MPI2_DOORBELL_DATA_MASK);
@@ -6686,7 +6703,7 @@ mpt3sas_base_get_iocstate(struct MPT3SAS_ADAPTER *ioc, int cooked)
{
u32 s, sc;
- s = ioc->base_readl(&ioc->chip->Doorbell);
+ s = ioc->base_readl_ext_retry(&ioc->chip->Doorbell);
sc = s & MPI2_IOC_STATE_MASK;
return cooked ? sc : s;
}
@@ -6831,7 +6848,7 @@ _base_wait_for_doorbell_ack(struct MPT3SAS_ADAPTER *ioc, int timeout)
__func__, count, timeout));
return 0;
} else if (int_status & MPI2_HIS_IOC2SYS_DB_STATUS) {
- doorbell = ioc->base_readl(&ioc->chip->Doorbell);
+ doorbell = ioc->base_readl_ext_retry(&ioc->chip->Doorbell);
if ((doorbell & MPI2_IOC_STATE_MASK) ==
MPI2_IOC_STATE_FAULT) {
mpt3sas_print_fault_code(ioc, doorbell);
@@ -6871,7 +6888,7 @@ _base_wait_for_doorbell_not_used(struct MPT3SAS_ADAPTER *ioc, int timeout)
count = 0;
cntdn = 1000 * timeout;
do {
- doorbell_reg = ioc->base_readl(&ioc->chip->Doorbell);
+ doorbell_reg = ioc->base_readl_ext_retry(&ioc->chip->Doorbell);
if (!(doorbell_reg & MPI2_DOORBELL_USED)) {
dhsprintk(ioc,
ioc_info(ioc, "%s: successful count(%d), timeout(%d)\n",
@@ -7019,7 +7036,7 @@ _base_handshake_req_reply_wait(struct MPT3SAS_ADAPTER *ioc, int request_bytes,
__le32 *mfp;
/* make sure doorbell is not in use */
- if ((ioc->base_readl(&ioc->chip->Doorbell) & MPI2_DOORBELL_USED)) {
+ if ((ioc->base_readl_ext_retry(&ioc->chip->Doorbell) & MPI2_DOORBELL_USED)) {
ioc_err(ioc, "doorbell is in use (line=%d)\n", __LINE__);
return -EFAULT;
}
@@ -7068,7 +7085,7 @@ _base_handshake_req_reply_wait(struct MPT3SAS_ADAPTER *ioc, int request_bytes,
}
/* read the first two 16-bits, it gives the total length of the reply */
- reply[0] = le16_to_cpu(ioc->base_readl(&ioc->chip->Doorbell)
+ reply[0] = le16_to_cpu(ioc->base_readl_ext_retry(&ioc->chip->Doorbell)
& MPI2_DOORBELL_DATA_MASK);
writel(0, &ioc->chip->HostInterruptStatus);
if ((_base_wait_for_doorbell_int(ioc, 5))) {
@@ -7076,7 +7093,7 @@ _base_handshake_req_reply_wait(struct MPT3SAS_ADAPTER *ioc, int request_bytes,
__LINE__);
return -EFAULT;
}
- reply[1] = le16_to_cpu(ioc->base_readl(&ioc->chip->Doorbell)
+ reply[1] = le16_to_cpu(ioc->base_readl_ext_retry(&ioc->chip->Doorbell)
& MPI2_DOORBELL_DATA_MASK);
writel(0, &ioc->chip->HostInterruptStatus);
@@ -7087,10 +7104,10 @@ _base_handshake_req_reply_wait(struct MPT3SAS_ADAPTER *ioc, int request_bytes,
return -EFAULT;
}
if (i >= reply_bytes/2) /* overflow case */
- ioc->base_readl(&ioc->chip->Doorbell);
+ ioc->base_readl_ext_retry(&ioc->chip->Doorbell);
else
reply[i] = le16_to_cpu(
- ioc->base_readl(&ioc->chip->Doorbell)
+ ioc->base_readl_ext_retry(&ioc->chip->Doorbell)
& MPI2_DOORBELL_DATA_MASK);
writel(0, &ioc->chip->HostInterruptStatus);
}
@@ -7949,7 +7966,7 @@ _base_diag_reset(struct MPT3SAS_ADAPTER *ioc)
goto out;
}
- host_diagnostic = ioc->base_readl(&ioc->chip->HostDiagnostic);
+ host_diagnostic = ioc->base_readl_ext_retry(&ioc->chip->HostDiagnostic);
drsprintk(ioc,
ioc_info(ioc, "wrote magic sequence: count(%d), host_diagnostic(0x%08x)\n",
count, host_diagnostic));
@@ -7969,7 +7986,7 @@ _base_diag_reset(struct MPT3SAS_ADAPTER *ioc)
for (count = 0; count < (300000000 /
MPI2_HARD_RESET_PCIE_SECOND_READ_DELAY_MICRO_SEC); count++) {
- host_diagnostic = ioc->base_readl(&ioc->chip->HostDiagnostic);
+ host_diagnostic = ioc->base_readl_ext_retry(&ioc->chip->HostDiagnostic);
if (host_diagnostic == 0xFFFFFFFF) {
ioc_info(ioc,
@@ -8359,10 +8376,13 @@ mpt3sas_base_attach(struct MPT3SAS_ADAPTER *ioc)
ioc->rdpq_array_enable_assigned = 0;
ioc->use_32bit_dma = false;
ioc->dma_mask = 64;
- if (ioc->is_aero_ioc)
+ if (ioc->is_aero_ioc) {
ioc->base_readl = &_base_readl_aero;
- else
+ ioc->base_readl_ext_retry = &_base_readl_ext_retry;
+ } else {
ioc->base_readl = &_base_readl;
+ ioc->base_readl_ext_retry = &_base_readl;
+ }
r = mpt3sas_base_map_resources(ioc);
if (r)
goto out_free_resources;
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.h b/drivers/scsi/mpt3sas/mpt3sas_base.h
index 05364aa15ecd..10055c7e4a9f 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.h
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.h
@@ -1618,6 +1618,7 @@ struct MPT3SAS_ADAPTER {
u8 diag_trigger_active;
u8 atomic_desc_capable;
BASE_READ_REG base_readl;
+ BASE_READ_REG base_readl_ext_retry;
struct SL_WH_MASTER_TRIGGER_T diag_trigger_master;
struct SL_WH_EVENT_TRIGGERS_T diag_trigger_event;
struct SL_WH_SCSI_TRIGGERS_T diag_trigger_scsi;
--
2.31.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
WARNING: multiple messages have this Message-ID (diff)
From: Ranjan Kumar <ranjan.kumar@broadcom.com>
To: linux-scsi@vger.kernel.org, martin.petersen@oracle.com
Cc: sathya.prakash@broadcom.com, sreekanth.reddy@broadcom.com,
sumit.saxena@broadcom.com,
Ranjan Kumar <ranjan.kumar@broadcom.com>,
stable@vger.kernel.org
Subject: [PATCH v4 1/2] mpt3sas: Perform additional retries if Doorbell read returns 0
Date: Tue, 29 Aug 2023 14:30:19 +0530 [thread overview]
Message-ID: <20230829090020.5417-2-ranjan.kumar@broadcom.com> (raw)
In-Reply-To: <20230829090020.5417-1-ranjan.kumar@broadcom.com>
[-- Attachment #1: Type: text/plain, Size: 6776 bytes --]
The driver retries certain register reads for 3 times if the
returned value is 0, this was done based on hardware specification
where the controller could possibly return 0 for certain registers
when there is a parallel access to some of other registers from the
firmware due to BMC out of band interactions.Recently it is observed
that in certain systems with increased BMC interactions, the values
returned are 0 even for 3 retries (the proper value is returned
between 4 to 6 retries).Hence this patch changes the retry
count 3 to 30, which is a revised recommendation, to avoid
the out of band conflict.
Fixes: b899202901a8 ("mpt3sas: Add separate function for aero doorbell reads")
Cc: stable@vger.kernel.org
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpt3sas/mpt3sas_base.c | 46 +++++++++++++++++++++--------
drivers/scsi/mpt3sas/mpt3sas_base.h | 1 +
2 files changed, 34 insertions(+), 13 deletions(-)
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c
index 53f5492579cb..5284584e4cd2 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.c
@@ -138,6 +138,9 @@ _base_get_ioc_facts(struct MPT3SAS_ADAPTER *ioc);
static void
_base_clear_outstanding_commands(struct MPT3SAS_ADAPTER *ioc);
+static u32
+_base_readl_ext_retry(const volatile void __iomem *addr);
+
/**
* mpt3sas_base_check_cmd_timeout - Function
* to check timeout and command termination due
@@ -213,6 +216,20 @@ _base_readl_aero(const volatile void __iomem *addr)
return ret_val;
}
+static u32
+_base_readl_ext_retry(const volatile void __iomem *addr)
+{
+ u32 i, ret_val;
+
+ for (i = 0 ; i < 30 ; i++) {
+ ret_val = readl(addr);
+ if (ret_val == 0)
+ continue;
+ }
+
+ return ret_val;
+}
+
static inline u32
_base_readl(const volatile void __iomem *addr)
{
@@ -940,7 +957,7 @@ mpt3sas_halt_firmware(struct MPT3SAS_ADAPTER *ioc)
dump_stack();
- doorbell = ioc->base_readl(&ioc->chip->Doorbell);
+ doorbell = ioc->base_readl_ext_retry(&ioc->chip->Doorbell);
if ((doorbell & MPI2_IOC_STATE_MASK) == MPI2_IOC_STATE_FAULT) {
mpt3sas_print_fault_code(ioc, doorbell &
MPI2_DOORBELL_DATA_MASK);
@@ -6686,7 +6703,7 @@ mpt3sas_base_get_iocstate(struct MPT3SAS_ADAPTER *ioc, int cooked)
{
u32 s, sc;
- s = ioc->base_readl(&ioc->chip->Doorbell);
+ s = ioc->base_readl_ext_retry(&ioc->chip->Doorbell);
sc = s & MPI2_IOC_STATE_MASK;
return cooked ? sc : s;
}
@@ -6831,7 +6848,7 @@ _base_wait_for_doorbell_ack(struct MPT3SAS_ADAPTER *ioc, int timeout)
__func__, count, timeout));
return 0;
} else if (int_status & MPI2_HIS_IOC2SYS_DB_STATUS) {
- doorbell = ioc->base_readl(&ioc->chip->Doorbell);
+ doorbell = ioc->base_readl_ext_retry(&ioc->chip->Doorbell);
if ((doorbell & MPI2_IOC_STATE_MASK) ==
MPI2_IOC_STATE_FAULT) {
mpt3sas_print_fault_code(ioc, doorbell);
@@ -6871,7 +6888,7 @@ _base_wait_for_doorbell_not_used(struct MPT3SAS_ADAPTER *ioc, int timeout)
count = 0;
cntdn = 1000 * timeout;
do {
- doorbell_reg = ioc->base_readl(&ioc->chip->Doorbell);
+ doorbell_reg = ioc->base_readl_ext_retry(&ioc->chip->Doorbell);
if (!(doorbell_reg & MPI2_DOORBELL_USED)) {
dhsprintk(ioc,
ioc_info(ioc, "%s: successful count(%d), timeout(%d)\n",
@@ -7019,7 +7036,7 @@ _base_handshake_req_reply_wait(struct MPT3SAS_ADAPTER *ioc, int request_bytes,
__le32 *mfp;
/* make sure doorbell is not in use */
- if ((ioc->base_readl(&ioc->chip->Doorbell) & MPI2_DOORBELL_USED)) {
+ if ((ioc->base_readl_ext_retry(&ioc->chip->Doorbell) & MPI2_DOORBELL_USED)) {
ioc_err(ioc, "doorbell is in use (line=%d)\n", __LINE__);
return -EFAULT;
}
@@ -7068,7 +7085,7 @@ _base_handshake_req_reply_wait(struct MPT3SAS_ADAPTER *ioc, int request_bytes,
}
/* read the first two 16-bits, it gives the total length of the reply */
- reply[0] = le16_to_cpu(ioc->base_readl(&ioc->chip->Doorbell)
+ reply[0] = le16_to_cpu(ioc->base_readl_ext_retry(&ioc->chip->Doorbell)
& MPI2_DOORBELL_DATA_MASK);
writel(0, &ioc->chip->HostInterruptStatus);
if ((_base_wait_for_doorbell_int(ioc, 5))) {
@@ -7076,7 +7093,7 @@ _base_handshake_req_reply_wait(struct MPT3SAS_ADAPTER *ioc, int request_bytes,
__LINE__);
return -EFAULT;
}
- reply[1] = le16_to_cpu(ioc->base_readl(&ioc->chip->Doorbell)
+ reply[1] = le16_to_cpu(ioc->base_readl_ext_retry(&ioc->chip->Doorbell)
& MPI2_DOORBELL_DATA_MASK);
writel(0, &ioc->chip->HostInterruptStatus);
@@ -7087,10 +7104,10 @@ _base_handshake_req_reply_wait(struct MPT3SAS_ADAPTER *ioc, int request_bytes,
return -EFAULT;
}
if (i >= reply_bytes/2) /* overflow case */
- ioc->base_readl(&ioc->chip->Doorbell);
+ ioc->base_readl_ext_retry(&ioc->chip->Doorbell);
else
reply[i] = le16_to_cpu(
- ioc->base_readl(&ioc->chip->Doorbell)
+ ioc->base_readl_ext_retry(&ioc->chip->Doorbell)
& MPI2_DOORBELL_DATA_MASK);
writel(0, &ioc->chip->HostInterruptStatus);
}
@@ -7949,7 +7966,7 @@ _base_diag_reset(struct MPT3SAS_ADAPTER *ioc)
goto out;
}
- host_diagnostic = ioc->base_readl(&ioc->chip->HostDiagnostic);
+ host_diagnostic = ioc->base_readl_ext_retry(&ioc->chip->HostDiagnostic);
drsprintk(ioc,
ioc_info(ioc, "wrote magic sequence: count(%d), host_diagnostic(0x%08x)\n",
count, host_diagnostic));
@@ -7969,7 +7986,7 @@ _base_diag_reset(struct MPT3SAS_ADAPTER *ioc)
for (count = 0; count < (300000000 /
MPI2_HARD_RESET_PCIE_SECOND_READ_DELAY_MICRO_SEC); count++) {
- host_diagnostic = ioc->base_readl(&ioc->chip->HostDiagnostic);
+ host_diagnostic = ioc->base_readl_ext_retry(&ioc->chip->HostDiagnostic);
if (host_diagnostic == 0xFFFFFFFF) {
ioc_info(ioc,
@@ -8359,10 +8376,13 @@ mpt3sas_base_attach(struct MPT3SAS_ADAPTER *ioc)
ioc->rdpq_array_enable_assigned = 0;
ioc->use_32bit_dma = false;
ioc->dma_mask = 64;
- if (ioc->is_aero_ioc)
+ if (ioc->is_aero_ioc) {
ioc->base_readl = &_base_readl_aero;
- else
+ ioc->base_readl_ext_retry = &_base_readl_ext_retry;
+ } else {
ioc->base_readl = &_base_readl;
+ ioc->base_readl_ext_retry = &_base_readl;
+ }
r = mpt3sas_base_map_resources(ioc);
if (r)
goto out_free_resources;
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.h b/drivers/scsi/mpt3sas/mpt3sas_base.h
index 05364aa15ecd..10055c7e4a9f 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.h
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.h
@@ -1618,6 +1618,7 @@ struct MPT3SAS_ADAPTER {
u8 diag_trigger_active;
u8 atomic_desc_capable;
BASE_READ_REG base_readl;
+ BASE_READ_REG base_readl_ext_retry;
struct SL_WH_MASTER_TRIGGER_T diag_trigger_master;
struct SL_WH_EVENT_TRIGGERS_T diag_trigger_event;
struct SL_WH_SCSI_TRIGGERS_T diag_trigger_scsi;
--
2.31.1
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]
next prev parent reply other threads:[~2023-08-29 9:02 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-29 9:00 [PATCH v4 0/2] mpt3sas: Additional retries when reading specific registers Ranjan Kumar
2023-08-29 9:00 ` Ranjan Kumar [this message]
2023-08-29 9:00 ` [PATCH v4 1/2] mpt3sas: Perform additional retries if Doorbell read returns 0 Ranjan Kumar
2023-08-29 9:00 ` [PATCH v4 2/2] mpt3sas: Removing volatile qualifier Ranjan Kumar
2023-08-31 1:39 ` [PATCH v4 0/2] mpt3sas: Additional retries when reading specific registers Martin K. Petersen
2023-12-09 20:21 ` patrick.strateman
2023-12-11 9:42 ` Ranjan Kumar
2023-09-05 10:18 ` Martin K. Petersen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230829090020.5417-2-ranjan.kumar@broadcom.com \
--to=ranjan.kumar@broadcom.com \
--cc=linux-scsi@vger.kernel.org \
--cc=martin.petersen@oracle.com \
--cc=sathya.prakash@broadcom.com \
--cc=sreekanth.reddy@broadcom.com \
--cc=stable@vger.kernel.org \
--cc=sumit.saxena@broadcom.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.