2014-02-24 08:16:48

by Alexander Gordeev

Subject: [PATCH v2 00/23] scsi: Use pci_enable_msix_range() instead of pci_enable_msix()

Hello!

This series is against James Bottomley's SCSI tree [1], but it needs
commit f7fc32c ("PCI/MSI: Add pci_enable_msi_exact() and
pci_enable_msix_exact()") from from Bjorn Helgaas's PCI tree [2]:

1. git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi.git for-next
2. git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci.git pci/msi

Recently pci_enable_msix_exact() function has been accepted to
the mainline. That is a variation of pci_enable_msix_range() which
allows a device driver to request a particular number of MSI-Xs.

As result, most of the changes posted in version 1 of this series
are invalidated and need to use pci_enable_msix_exact() instead of
originally posted pci_enable_msix_range() usages.

I removed almost all ACKs, since unlike pci_enable_msix_range()
function which returns the number of MSI-Xs allocated or negative
errno, pci_enable_msix_exact() returns either zero success code or
a negative errno. Although this change is simple, it still entails
an updated error code analysis and would be better reviewed by
driver maintainers.

Thanks!

Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]

Alexander Gordeev (23):
be2iscsi: Use pci_enable_msix_exact() instead of pci_enable_msix()
bfa: Do not call pci_enable_msix() after it failed once
bfa: Cleanup bfad_setup_intr() function
bfa: Use pci_enable_msix_exact() instead of pci_enable_msix()
csiostor: Remove superfluous call to pci_disable_msix()
csiostor: Use pci_enable_msix_range() instead of pci_enable_msix()
fnic: Use pci_enable_msix_exact() instead of pci_enable_msix()
hpsa: Fallback to MSI rather than to INTx if MSI-X failed
hpsa: Use pci_enable_msix_exact() instead of pci_enable_msix()
isci: Use pci_enable_msix_exact() instead of pci_enable_msix()
lpfc: Remove superfluous call to pci_disable_msix()
lpfc: Use pci_enable_msix_range() instead of pci_enable_msix()
megaraid: Fail resume if MSI-X re-initialization failed
megaraid: Use pci_enable_msix_range() instead of pci_enable_msix()
mpt2sas: Use pci_enable_msix_exact() instead of pci_enable_msix()
mpt3sas: Use pci_enable_msix_exact() instead of pci_enable_msix()
pm8001: Fix invalid return when request_irq() failed
pm8001: Use pci_enable_msix_exact() instead of pci_enable_msix()
pmcraid: Get rid of a redundant assignment
pmcraid: Use pci_enable_msix_range() instead of pci_enable_msix()
qla2xxx: Use pci_enable_msix_range() instead of pci_enable_msix()
qla4xxx: Use pci_enable_msix_exact() instead of pci_enable_msix()
vmw_pvscsi: Use pci_enable_msix_exact() instead of pci_enable_msix()

drivers/scsi/be2iscsi/be_main.c | 6 +--
drivers/scsi/bfa/bfad.c | 62 ++++++++++++-----------------
drivers/scsi/csiostor/csio_hw.h | 2 +-
drivers/scsi/csiostor/csio_isr.c | 24 ++++-------
drivers/scsi/fnic/fnic_isr.c | 4 +-
drivers/scsi/hpsa.c | 12 +-----
drivers/scsi/isci/init.c | 2 +-
drivers/scsi/lpfc/lpfc_init.c | 54 ++++++++++++-------------
drivers/scsi/megaraid/megaraid_sas_base.c | 24 +++++------
drivers/scsi/mpt2sas/mpt2sas_base.c | 6 +-
drivers/scsi/mpt3sas/mpt3sas_base.c | 4 +-
drivers/scsi/pm8001/pm8001_init.c | 44 +++++++++++----------
drivers/scsi/pmcraid.c | 14 +------
drivers/scsi/qla2xxx/qla_isr.c | 27 +++++-------
drivers/scsi/qla4xxx/ql4_nx.c | 2 +-
drivers/scsi/vmw_pvscsi.c | 2 +-
16 files changed, 121 insertions(+), 168 deletions(-)

--
1.7.7.6

2014-02-24 08:00:53

by Alexander Gordeev

Subject: [PATCH v2 03/23] bfa: Cleanup bfad_setup_intr() function

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Anil Gurumurthy <[email protected]>
Cc: Vijaya Mohan Guvva <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/bfa/bfad.c | 18 ++++++++----------
1 files changed, 8 insertions(+), 10 deletions(-)

diff --git a/drivers/scsi/bfa/bfad.c b/drivers/scsi/bfa/bfad.c
index 972ff8d..e7e4774 100644
--- a/drivers/scsi/bfa/bfad.c
+++ b/drivers/scsi/bfa/bfad.c
@@ -1219,7 +1219,7 @@ bfad_install_msix_handler(struct bfad_s *bfad)
int
bfad_setup_intr(struct bfad_s *bfad)
{
- int error = 0;
+ int error;
u32 mask = 0, i, num_bit = 0, max_bit = 0;
struct msix_entry msix_entries[MAX_MSIX_ENTRY];
struct pci_dev *pdev = bfad->pcidev;
@@ -1279,20 +1279,18 @@ bfad_setup_intr(struct bfad_s *bfad)

bfad->bfad_flags |= BFAD_MSIX_ON;

- return error;
+ return 0;
}

line_based:
- error = 0;
- if (request_irq
- (bfad->pcidev->irq, (irq_handler_t) bfad_intx, BFAD_IRQ_FLAGS,
- BFAD_DRIVER_NAME, bfad) != 0) {
- /* Enable interrupt handler failed */
- return 1;
- }
+ error = request_irq(bfad->pcidev->irq, (irq_handler_t)bfad_intx,
+ BFAD_IRQ_FLAGS, BFAD_DRIVER_NAME, bfad);
+ if (error)
+ return error;
+
bfad->bfad_flags |= BFAD_INTX_ON;

- return error;
+ return 0;
}

void
--
1.7.7.6

2014-02-24 08:00:56

by Alexander Gordeev

Subject: [PATCH v2 06/23] csiostor: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Naresh Kumar Inna <[email protected]>
Cc: Arvind Bhushan <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/csiostor/csio_hw.h | 2 +-
drivers/scsi/csiostor/csio_isr.c | 22 +++++++++-------------
2 files changed, 10 insertions(+), 14 deletions(-)

diff --git a/drivers/scsi/csiostor/csio_hw.h b/drivers/scsi/csiostor/csio_hw.h
index 49b1daa..5db2d85 100644
--- a/drivers/scsi/csiostor/csio_hw.h
+++ b/drivers/scsi/csiostor/csio_hw.h
@@ -94,7 +94,7 @@ enum {
};

struct csio_msix_entries {
- unsigned short vector; /* Vector assigned by pci_enable_msix */
+ unsigned short vector; /* Assigned MSI-X vector */
void *dev_id; /* Priv object associated w/ this msix*/
char desc[24]; /* Description of this vector */
};
diff --git a/drivers/scsi/csiostor/csio_isr.c b/drivers/scsi/csiostor/csio_isr.c
index 91ba91d..a8c748a 100644
--- a/drivers/scsi/csiostor/csio_isr.c
+++ b/drivers/scsi/csiostor/csio_isr.c
@@ -499,7 +499,7 @@ csio_reduce_sqsets(struct csio_hw *hw, int cnt)
static int
csio_enable_msix(struct csio_hw *hw)
{
- int rv, i, j, k, n, min, cnt;
+ int i, j, k, n, min, cnt;
struct csio_msix_entries *entryp;
struct msix_entry *entries;
int extra = CSIO_EXTRA_VECS;
@@ -521,19 +521,15 @@ csio_enable_msix(struct csio_hw *hw)

csio_dbg(hw, "FW supp #niq:%d, trying %d msix's\n", hw->cfg_niq, cnt);

- while ((rv = pci_enable_msix(hw->pdev, entries, cnt)) >= min)
- cnt = rv;
- if (!rv) {
- if (cnt < (hw->num_sqsets + extra)) {
- csio_dbg(hw, "Reducing sqsets to %d\n", cnt - extra);
- csio_reduce_sqsets(hw, cnt - extra);
- }
- } else {
- if (rv > 0)
- csio_info(hw, "Not using MSI-X, remainder:%d\n", rv);
-
+ cnt = pci_enable_msix_range(hw->pdev, entries, min, cnt);
+ if (cnt < 0) {
kfree(entries);
- return -ENOMEM;
+ return cnt;
+ }
+
+ if (cnt < (hw->num_sqsets + extra)) {
+ csio_dbg(hw, "Reducing sqsets to %d\n", cnt - extra);
+ csio_reduce_sqsets(hw, cnt - extra);
}

/* Save off vectors */
--
1.7.7.6

2014-02-24 08:01:06

by Alexander Gordeev

Subject: [PATCH v2 07/23] fnic: Use pci_enable_msix_exact() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Hiral Patel <[email protected]>
Cc: Suma Ramars <[email protected]>
Cc: Brian Uchino <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/fnic/fnic_isr.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/fnic/fnic_isr.c b/drivers/scsi/fnic/fnic_isr.c
index 7d9b54a..a0dd1b6 100644
--- a/drivers/scsi/fnic/fnic_isr.c
+++ b/drivers/scsi/fnic/fnic_isr.c
@@ -257,8 +257,8 @@ int fnic_set_intr_mode(struct fnic *fnic)
fnic->raw_wq_count >= m &&
fnic->wq_copy_count >= o &&
fnic->cq_count >= n + m + o) {
- if (!pci_enable_msix(fnic->pdev, fnic->msix_entry,
- n + m + o + 1)) {
+ if (!pci_enable_msix_exact(fnic->pdev, fnic->msix_entry,
+ n + m + o + 1)) {
fnic->rq_count = n;
fnic->raw_wq_count = m;
fnic->wq_copy_count = o;
--
1.7.7.6

2014-02-24 08:01:29

by Alexander Gordeev

Subject: [PATCH v2 23/23] vmw_pvscsi: Use pci_enable_msix_exact() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Arvind Kumar <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/vmw_pvscsi.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/scsi/vmw_pvscsi.c b/drivers/scsi/vmw_pvscsi.c
index b9755ec..a046618 100644
--- a/drivers/scsi/vmw_pvscsi.c
+++ b/drivers/scsi/vmw_pvscsi.c
@@ -1078,7 +1078,7 @@ static int pvscsi_setup_msix(const struct pvscsi_adapter *adapter,
struct msix_entry entry = { 0, PVSCSI_VECTOR_COMPLETION };
int ret;

- ret = pci_enable_msix(adapter->dev, &entry, 1);
+ ret = pci_enable_msix_exact(adapter->dev, &entry, 1);
if (ret)
return ret;

--
1.7.7.6

2014-02-24 08:01:27

by Alexander Gordeev

Subject: [PATCH v2 18/23] pm8001: Use pci_enable_msix_exact() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/pm8001/pm8001_init.c | 39 +++++++++++++++++++------------------
1 files changed, 20 insertions(+), 19 deletions(-)

diff --git a/drivers/scsi/pm8001/pm8001_init.c b/drivers/scsi/pm8001/pm8001_init.c
index efffbb9..0aa9059 100644
--- a/drivers/scsi/pm8001/pm8001_init.c
+++ b/drivers/scsi/pm8001/pm8001_init.c
@@ -724,34 +724,35 @@ static u32 pm8001_setup_msix(struct pm8001_hba_info *pm8001_ha)
sizeof(pm8001_ha->msix_entries[0]);
for (i = 0; i < max_entry ; i++)
pm8001_ha->msix_entries[i].entry = i;
- rc = pci_enable_msix(pm8001_ha->pdev, pm8001_ha->msix_entries,
+ rc = pci_enable_msix_exact(pm8001_ha->pdev, pm8001_ha->msix_entries,
number_of_intr);
pm8001_ha->number_of_intr = number_of_intr;
- if (!rc) {
- PM8001_INIT_DBG(pm8001_ha, pm8001_printk(
- "pci_enable_msix request ret:%d no of intr %d\n",
- rc, pm8001_ha->number_of_intr));
+ if (rc)
+ return rc;

+ PM8001_INIT_DBG(pm8001_ha, pm8001_printk(
+ "pci_enable_msix_exact request ret:%d no of intr %d\n",
+ rc, pm8001_ha->number_of_intr));

- for (i = 0; i < number_of_intr; i++) {
- snprintf(intr_drvname[i], sizeof(intr_drvname[0]),
- DRV_NAME"%d", i);
- pm8001_ha->irq_vector[i].irq_id = i;
- pm8001_ha->irq_vector[i].drv_inst = pm8001_ha;
+ for (i = 0; i < number_of_intr; i++) {
+ snprintf(intr_drvname[i], sizeof(intr_drvname[0]),
+ DRV_NAME"%d", i);
+ pm8001_ha->irq_vector[i].irq_id = i;
+ pm8001_ha->irq_vector[i].drv_inst = pm8001_ha;

- rc = request_irq(pm8001_ha->msix_entries[i].vector,
- pm8001_interrupt_handler_msix, flag,
- intr_drvname[i], &(pm8001_ha->irq_vector[i]));
- if (rc) {
- for (j = 0; j < i; j++)
- free_irq(
- pm8001_ha->msix_entries[j].vector,
+ rc = request_irq(pm8001_ha->msix_entries[i].vector,
+ pm8001_interrupt_handler_msix, flag,
+ intr_drvname[i], &(pm8001_ha->irq_vector[i]));
+ if (rc) {
+ for (j = 0; j < i; j++) {
+ free_irq(pm8001_ha->msix_entries[j].vector,
&(pm8001_ha->irq_vector[i]));
- pci_disable_msix(pm8001_ha->pdev);
- break;
}
+ pci_disable_msix(pm8001_ha->pdev);
+ break;
}
}
+
return rc;
}
#endif
--
1.7.7.6

2014-02-24 08:01:25

by Alexander Gordeev

Subject: [PATCH v2 17/23] pm8001: Fix invalid return when request_irq() failed

When a call to request_irq() failed pm8001_setup_msix()
still returns the success. This udate fixes the described
misbehaviour.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Acked-by: Jack Wang <[email protected]>
---
drivers/scsi/pm8001/pm8001_init.c | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/pm8001/pm8001_init.c b/drivers/scsi/pm8001/pm8001_init.c
index 73a120d..efffbb9 100644
--- a/drivers/scsi/pm8001/pm8001_init.c
+++ b/drivers/scsi/pm8001/pm8001_init.c
@@ -739,9 +739,10 @@ static u32 pm8001_setup_msix(struct pm8001_hba_info *pm8001_ha)
pm8001_ha->irq_vector[i].irq_id = i;
pm8001_ha->irq_vector[i].drv_inst = pm8001_ha;

- if (request_irq(pm8001_ha->msix_entries[i].vector,
+ rc = request_irq(pm8001_ha->msix_entries[i].vector,
pm8001_interrupt_handler_msix, flag,
- intr_drvname[i], &(pm8001_ha->irq_vector[i]))) {
+ intr_drvname[i], &(pm8001_ha->irq_vector[i]));
+ if (rc) {
for (j = 0; j < i; j++)
free_irq(
pm8001_ha->msix_entries[j].vector,
--
1.7.7.6

2014-02-24 08:02:13

by Alexander Gordeev

Subject: [PATCH v2 08/23] hpsa: Fallback to MSI rather than to INTx if MSI-X failed

Currently the driver falls back to INTx mode when MSI-X
initialization failed. This is a suboptimal behaviour
for chips that also support MSI. This update changes that
behaviour and falls back to MSI mode in case MSI-X mode
initialization failed.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: "Stephen M. Cameron" <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/hpsa.c | 2 --
1 files changed, 0 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/hpsa.c b/drivers/scsi/hpsa.c
index 868318a..52cfd19 100644
--- a/drivers/scsi/hpsa.c
+++ b/drivers/scsi/hpsa.c
@@ -4180,11 +4180,9 @@ static void hpsa_interrupt_mode(struct ctlr_info *h)
if (err > 0) {
dev_warn(&h->pdev->dev, "only %d MSI-X vectors "
"available\n", err);
- goto default_int_mode;
} else {
dev_warn(&h->pdev->dev, "MSI-X init failed %d\n",
err);
- goto default_int_mode;
}
}
if (pci_find_capability(h->pdev, PCI_CAP_ID_MSI)) {
--
1.7.7.6

2014-02-24 08:02:16

by Alexander Gordeev

Subject: [PATCH v2 09/23] hpsa: Use pci_enable_msix_exact() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: "Stephen M. Cameron" <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/hpsa.c | 10 ++--------
1 files changed, 2 insertions(+), 8 deletions(-)

diff --git a/drivers/scsi/hpsa.c b/drivers/scsi/hpsa.c
index 52cfd19..523a60e 100644
--- a/drivers/scsi/hpsa.c
+++ b/drivers/scsi/hpsa.c
@@ -4169,7 +4169,7 @@ static void hpsa_interrupt_mode(struct ctlr_info *h)
goto default_int_mode;
if (pci_find_capability(h->pdev, PCI_CAP_ID_MSIX)) {
dev_info(&h->pdev->dev, "MSIX\n");
- err = pci_enable_msix(h->pdev, hpsa_msix_entries,
+ err = pci_enable_msix_exact(h->pdev, hpsa_msix_entries,
MAX_REPLY_QUEUES);
if (!err) {
for (i = 0; i < MAX_REPLY_QUEUES; i++)
@@ -4177,13 +4177,7 @@ static void hpsa_interrupt_mode(struct ctlr_info *h)
h->msix_vector = 1;
return;
}
- if (err > 0) {
- dev_warn(&h->pdev->dev, "only %d MSI-X vectors "
- "available\n", err);
- } else {
- dev_warn(&h->pdev->dev, "MSI-X init failed %d\n",
- err);
- }
+ dev_warn(&h->pdev->dev, "MSI-X init failed %d\n", err);
}
if (pci_find_capability(h->pdev, PCI_CAP_ID_MSI)) {
dev_info(&h->pdev->dev, "MSI\n");
--
1.7.7.6

2014-02-24 08:01:24

by Alexander Gordeev

Subject: [PATCH v2 22/23] qla4xxx: Use pci_enable_msix_exact() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Vikas Chaudhary <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/qla4xxx/ql4_nx.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/scsi/qla4xxx/ql4_nx.c b/drivers/scsi/qla4xxx/ql4_nx.c
index d001202..2cfcfbf 100644
--- a/drivers/scsi/qla4xxx/ql4_nx.c
+++ b/drivers/scsi/qla4xxx/ql4_nx.c
@@ -3805,7 +3805,7 @@ qla4_8xxx_enable_msix(struct scsi_qla_host *ha)
for (i = 0; i < QLA_MSIX_ENTRIES; i++)
entries[i].entry = qla4_8xxx_msix_entries[i].entry;

- ret = pci_enable_msix(ha->pdev, entries, ARRAY_SIZE(entries));
+ ret = pci_enable_msix_exact(ha->pdev, entries, ARRAY_SIZE(entries));
if (ret) {
ql4_printk(KERN_WARNING, ha,
"MSI-X: Failed to enable support -- %d/%d\n",
--
1.7.7.6

2014-02-24 08:01:20

by Alexander Gordeev

Subject: [PATCH v2 20/23] pmcraid: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Anil Ravindranath <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/pmcraid.c | 13 ++-----------
1 files changed, 2 insertions(+), 11 deletions(-)

diff --git a/drivers/scsi/pmcraid.c b/drivers/scsi/pmcraid.c
index c06af7f..6d0f208 100644
--- a/drivers/scsi/pmcraid.c
+++ b/drivers/scsi/pmcraid.c
@@ -4698,19 +4698,10 @@ pmcraid_register_interrupt_handler(struct pmcraid_instance *pinstance)
for (i = 0; i < PMCRAID_NUM_MSIX_VECTORS; i++)
entries[i].entry = i;

- rc = pci_enable_msix(pdev, entries, num_hrrq);
- if (rc < 0)
+ num_hrrq = pci_enable_msix_range(pdev, entries, 1, num_hrrq);
+ if (num_hrrq < 0)
goto pmcraid_isr_legacy;

- /* Check how many MSIX vectors are allocated and register
- * msi-x handlers for each of them giving appropriate buffer
- */
- if (rc > 0) {
- num_hrrq = rc;
- if (pci_enable_msix(pdev, entries, num_hrrq))
- goto pmcraid_isr_legacy;
- }
-
for (i = 0; i < num_hrrq; i++) {
pinstance->hrrq_vector[i].hrrq_id = i;
pinstance->hrrq_vector[i].drv_inst = pinstance;
--
1.7.7.6

2014-02-24 08:06:32

by Alexander Gordeev

Subject: [PATCH v2 21/23] qla2xxx: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Log message code 0x00c6 preserved, although it is reported
after successful call to pci_enable_msix_range(), not before
possibly unsuccessful call to pci_enable_msix(). Consumers
of the error code should not notice the difference.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/qla2xxx/qla_isr.c | 27 +++++++++++----------------
1 files changed, 11 insertions(+), 16 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c
index 9bc86b9..b2dcfa4 100644
--- a/drivers/scsi/qla2xxx/qla_isr.c
+++ b/drivers/scsi/qla2xxx/qla_isr.c
@@ -2896,27 +2896,22 @@ qla24xx_enable_msix(struct qla_hw_data *ha, struct rsp_que *rsp)
for (i = 0; i < ha->msix_count; i++)
entries[i].entry = i;

- ret = pci_enable_msix(ha->pdev, entries, ha->msix_count);
- if (ret) {
- if (ret < MIN_MSIX_COUNT)
- goto msix_failed;
-
+ ret = pci_enable_msix_range(ha->pdev,
+ entries, MIN_MSIX_COUNT, ha->msix_count);
+ if (ret < 0) {
+ ql_log(ql_log_fatal, vha, 0x00c7,
+ "MSI-X: Failed to enable support, "
+ "giving up -- %d/%d.\n",
+ ha->msix_count, ret);
+ goto msix_out;
+ } else if (ret < ha->msix_count) {
ql_log(ql_log_warn, vha, 0x00c6,
"MSI-X: Failed to enable support "
"-- %d/%d\n Retry with %d vectors.\n",
ha->msix_count, ret, ret);
- ha->msix_count = ret;
- ret = pci_enable_msix(ha->pdev, entries, ha->msix_count);
- if (ret) {
-msix_failed:
- ql_log(ql_log_fatal, vha, 0x00c7,
- "MSI-X: Failed to enable support, "
- "giving up -- %d/%d.\n",
- ha->msix_count, ret);
- goto msix_out;
- }
- ha->max_rsp_queues = ha->msix_count - 1;
}
+ ha->msix_count = ret;
+ ha->max_rsp_queues = ha->msix_count - 1;
ha->msix_entries = kzalloc(sizeof(struct qla_msix_entry) *
ha->msix_count, GFP_KERNEL);
if (!ha->msix_entries) {
--
1.7.7.6

2014-02-24 08:01:16

by Alexander Gordeev

Subject: [PATCH v2 11/23] lpfc: Remove superfluous call to pci_disable_msix()

There is no need to call pci_disable_msix() in case
the previous call to pci_enable_msix() failed

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: James Smart <[email protected]>
Cc: [email protected]
Cc: [email protected]
Acked-by: James Smart <[email protected]>
---
drivers/scsi/lpfc/lpfc_init.c | 9 ++++++---
1 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index 68c94cc..a9df97a 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -8078,7 +8078,7 @@ lpfc_sli_enable_msix(struct lpfc_hba *phba)
if (rc) {
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
"0420 PCI enable MSI-X failed (%d)\n", rc);
- goto msi_fail_out;
+ goto vec_fail_out;
}
for (i = 0; i < LPFC_MSIX_VECTORS; i++)
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
@@ -8156,6 +8156,8 @@ irq_fail_out:
msi_fail_out:
/* Unconfigure MSI-X capability structure */
pci_disable_msix(phba->pcidev);
+
+vec_fail_out:
return rc;
}

@@ -8644,7 +8646,7 @@ enable_msix_vectors:
} else if (rc) {
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
"0484 PCI enable MSI-X failed (%d)\n", rc);
- goto msi_fail_out;
+ goto vec_fail_out;
}

/* Log MSI-X vector assignment */
@@ -8696,9 +8698,10 @@ cfg_fail_out:
&phba->sli4_hba.fcp_eq_hdl[index]);
}

-msi_fail_out:
/* Unconfigure MSI-X capability structure */
pci_disable_msix(phba->pcidev);
+
+vec_fail_out:
return rc;
}

--
1.7.7.6

2014-02-24 08:07:00

by Alexander Gordeev

Subject: [PATCH v2 19/23] pmcraid: Get rid of a redundant assignment

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Anil Ravindranath <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/pmcraid.c | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)

diff --git a/drivers/scsi/pmcraid.c b/drivers/scsi/pmcraid.c
index be8ce54..c06af7f 100644
--- a/drivers/scsi/pmcraid.c
+++ b/drivers/scsi/pmcraid.c
@@ -4746,7 +4746,6 @@ pmcraid_isr_legacy:
pinstance->hrrq_vector[0].drv_inst = pinstance;
pinstance->hrrq_vector[0].vector = pdev->irq;
pinstance->num_hrrq = 1;
- rc = 0;

rc = request_irq(pdev->irq, pmcraid_isr, IRQF_SHARED,
PMCRAID_DRIVER_NAME, &pinstance->hrrq_vector[0]);
--
1.7.7.6

2014-02-24 08:01:13

by Alexander Gordeev

Subject: [PATCH v2 14/23] megaraid: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Neela Syam Kolli <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/megaraid/megaraid_sas_base.c | 20 +++++++-------------
1 files changed, 7 insertions(+), 13 deletions(-)

diff --git a/drivers/scsi/megaraid/megaraid_sas_base.c b/drivers/scsi/megaraid/megaraid_sas_base.c
index 01afe38..b276aa3 100644
--- a/drivers/scsi/megaraid/megaraid_sas_base.c
+++ b/drivers/scsi/megaraid/megaraid_sas_base.c
@@ -3731,17 +3731,11 @@ static int megasas_init_fw(struct megasas_instance *instance)
(unsigned int)num_online_cpus());
for (i = 0; i < instance->msix_vectors; i++)
instance->msixentry[i].entry = i;
- i = pci_enable_msix(instance->pdev, instance->msixentry,
- instance->msix_vectors);
- if (i >= 0) {
- if (i) {
- if (!pci_enable_msix(instance->pdev,
- instance->msixentry, i))
- instance->msix_vectors = i;
- else
- instance->msix_vectors = 0;
- }
- } else
+ i = pci_enable_msix_range(instance->pdev, instance->msixentry,
+ 1, instance->msix_vectors);
+ if (i)
+ instance->msix_vectors = i;
+ else
instance->msix_vectors = 0;

dev_info(&instance->pdev->dev, "[scsi%d]: FW supports"
@@ -4668,8 +4662,8 @@ megasas_resume(struct pci_dev *pdev)

/* Now re-enable MSI-X */
if (instance->msix_vectors &&
- pci_enable_msix(instance->pdev, instance->msixentry,
- instance->msix_vectors))
+ pci_enable_msix_exact(instance->pdev, instance->msixentry,
+ instance->msix_vectors))
goto fail_reenable_msix;

switch (instance->pdev->device) {
--
1.7.7.6

2014-02-24 08:08:30

by Alexander Gordeev

Subject: [PATCH v2 16/23] mpt3sas: Use pci_enable_msix_exact() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Nagalakshmi Nandigama <[email protected]>
Cc: Sreekanth Reddy <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/mpt3sas/mpt3sas_base.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c
index 0cf4f70..7b65ec2 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.c
@@ -1747,10 +1747,10 @@ _base_enable_msix(struct MPT3SAS_ADAPTER *ioc)
for (i = 0, a = entries; i < ioc->reply_queue_count; i++, a++)
a->entry = i;

- r = pci_enable_msix(ioc->pdev, entries, ioc->reply_queue_count);
+ r = pci_enable_msix_exact(ioc->pdev, entries, ioc->reply_queue_count);
if (r) {
dfailprintk(ioc, pr_info(MPT3SAS_FMT
- "pci_enable_msix failed (r=%d) !!!\n",
+ "pci_enable_msix_exact failed (r=%d) !!!\n",
ioc->name, r));
kfree(entries);
goto try_ioapic;
--
1.7.7.6

2014-02-24 08:08:52

by Alexander Gordeev

Subject: [PATCH v2 15/23] mpt2sas: Use pci_enable_msix_exact() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Nagalakshmi Nandigama <[email protected]>
Cc: Sreekanth Reddy <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/mpt2sas/mpt2sas_base.c | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/scsi/mpt2sas/mpt2sas_base.c b/drivers/scsi/mpt2sas/mpt2sas_base.c
index bde63f7..984b605 100644
--- a/drivers/scsi/mpt2sas/mpt2sas_base.c
+++ b/drivers/scsi/mpt2sas/mpt2sas_base.c
@@ -1432,10 +1432,10 @@ _base_enable_msix(struct MPT2SAS_ADAPTER *ioc)
for (i = 0, a = entries; i < ioc->reply_queue_count; i++, a++)
a->entry = i;

- r = pci_enable_msix(ioc->pdev, entries, ioc->reply_queue_count);
+ r = pci_enable_msix_exact(ioc->pdev, entries, ioc->reply_queue_count);
if (r) {
- dfailprintk(ioc, printk(MPT2SAS_INFO_FMT "pci_enable_msix "
- "failed (r=%d) !!!\n", ioc->name, r));
+ dfailprintk(ioc, printk(MPT2SAS_INFO_FMT
+ "pci_enable_msix_exact failed (r=%d) !!!\n", ioc->name, r));
kfree(entries);
goto try_ioapic;
}
--
1.7.7.6

2014-02-24 08:11:11

by Alexander Gordeev

Subject: [PATCH v2 01/23] be2iscsi: Use pci_enable_msix_exact() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Jayamohan Kallickal <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/be2iscsi/be_main.c | 6 ++----
1 files changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c
index 1f37505..580192f 100644
--- a/drivers/scsi/be2iscsi/be_main.c
+++ b/drivers/scsi/be2iscsi/be_main.c
@@ -5284,12 +5284,10 @@ static void beiscsi_msix_enable(struct beiscsi_hba *phba)
for (i = 0; i <= phba->num_cpus; i++)
phba->msix_entries[i].entry = i;

- status = pci_enable_msix(phba->pcidev, phba->msix_entries,
- (phba->num_cpus + 1));
+ status = pci_enable_msix_exact(phba->pcidev, phba->msix_entries,
+ phba->num_cpus + 1);
if (!status)
phba->msix_enabled = true;
-
- return;
}

/*
--
1.7.7.6

2014-02-24 08:14:56

by Alexander Gordeev

Subject: [PATCH v2 10/23] isci: Use pci_enable_msix_exact() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Lukasz Dorau <[email protected]>
Cc: Maciej Patelczyk <[email protected]>
Cc: Dave Jiang <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/isci/init.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/scsi/isci/init.c b/drivers/scsi/isci/init.c
index d25d0d8..ce40538 100644
--- a/drivers/scsi/isci/init.c
+++ b/drivers/scsi/isci/init.c
@@ -356,7 +356,7 @@ static int isci_setup_interrupts(struct pci_dev *pdev)
for (i = 0; i < num_msix; i++)
pci_info->msix_entries[i].entry = i;

- err = pci_enable_msix(pdev, pci_info->msix_entries, num_msix);
+ err = pci_enable_msix_exact(pdev, pci_info->msix_entries, num_msix);
if (err)
goto intx;

--
1.7.7.6

2014-02-24 08:14:55

by Alexander Gordeev

Subject: [PATCH v2 12/23] lpfc: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: James Smart <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/lpfc/lpfc_init.c | 45 ++++++++++++++++++----------------------
1 files changed, 20 insertions(+), 25 deletions(-)

diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index a9df97a..5368f75 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -8047,9 +8047,9 @@ lpfc_sli4_pci_mem_unset(struct lpfc_hba *phba)
* @phba: pointer to lpfc hba data structure.
*
* This routine is invoked to enable the MSI-X interrupt vectors to device
- * with SLI-3 interface specs. The kernel function pci_enable_msix() is
- * called to enable the MSI-X vectors. Note that pci_enable_msix(), once
- * invoked, enables either all or nothing, depending on the current
+ * with SLI-3 interface specs. The kernel function pci_enable_msix_exact()
+ * is called to enable the MSI-X vectors. Note that pci_enable_msix_exact(),
+ * once invoked, enables either all or nothing, depending on the current
* availability of PCI vector resources. The device driver is responsible
* for calling the individual request_irq() to register each MSI-X vector
* with a interrupt handler, which is done in this function. Note that
@@ -8073,8 +8073,8 @@ lpfc_sli_enable_msix(struct lpfc_hba *phba)
phba->msix_entries[i].entry = i;

/* Configure MSI-X capability structure */
- rc = pci_enable_msix(phba->pcidev, phba->msix_entries,
- ARRAY_SIZE(phba->msix_entries));
+ rc = pci_enable_msix_exact(phba->pcidev, phba->msix_entries,
+ ARRAY_SIZE(phba->msix_entries));
if (rc) {
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
"0420 PCI enable MSI-X failed (%d)\n", rc);
@@ -8187,8 +8187,8 @@ lpfc_sli_disable_msix(struct lpfc_hba *phba)
* @phba: pointer to lpfc hba data structure.
*
* This routine is invoked to enable the MSI interrupt mode to device with
- * SLI-3 interface spec. The kernel function pci_enable_msi() is called to
- * enable the MSI vector. The device driver is responsible for calling the
+ * SLI-3 interface spec. The kernel function pci_enable_msi() is called
+ * to enable the MSI vector. The device driver is responsible for calling the
* request_irq() to register MSI vector with a interrupt the handler, which
* is done in this function.
*
@@ -8611,16 +8611,14 @@ out:
* @phba: pointer to lpfc hba data structure.
*
* This routine is invoked to enable the MSI-X interrupt vectors to device
- * with SLI-4 interface spec. The kernel function pci_enable_msix() is called
- * to enable the MSI-X vectors. Note that pci_enable_msix(), once invoked,
- * enables either all or nothing, depending on the current availability of
- * PCI vector resources. The device driver is responsible for calling the
- * individual request_irq() to register each MSI-X vector with a interrupt
- * handler, which is done in this function. Note that later when device is
- * unloading, the driver should always call free_irq() on all MSI-X vectors
- * it has done request_irq() on before calling pci_disable_msix(). Failure
- * to do so results in a BUG_ON() and a device will be left with MSI-X
- * enabled and leaks its vectors.
+ * with SLI-4 interface spec. The kernel function pci_enable_msix_range()
+ * is called to enable the MSI-X vectors. The device driver is responsible
+ * for calling the individual request_irq() to register each MSI-X vector
+ * with a interrupt handler, which is done in this function. Note that
+ * later when device is unloading, the driver should always call free_irq()
+ * on all MSI-X vectors it has done request_irq() on before calling
+ * pci_disable_msix(). Failure to do so results in a BUG_ON() and a device
+ * will be left with MSI-X enabled and leaks its vectors.
*
* Return codes
* 0 - successful
@@ -8636,17 +8634,14 @@ lpfc_sli4_enable_msix(struct lpfc_hba *phba)
phba->sli4_hba.msix_entries[index].entry = index;

/* Configure MSI-X capability structure */
- vectors = phba->cfg_fcp_io_channel;
-enable_msix_vectors:
- rc = pci_enable_msix(phba->pcidev, phba->sli4_hba.msix_entries,
- vectors);
- if (rc > 1) {
- vectors = rc;
- goto enable_msix_vectors;
- } else if (rc) {
+ rc = pci_enable_msix_range(phba->pcidev, phba->sli4_hba.msix_entries,
+ 2, phba->cfg_fcp_io_channel);
+ if (rc < 0) {
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
"0484 PCI enable MSI-X failed (%d)\n", rc);
goto vec_fail_out;
+ } else {
+ vectors = rc;
}

/* Log MSI-X vector assignment */
--
1.7.7.6

2014-02-24 08:01:04

by Alexander Gordeev

Subject: [PATCH v2 05/23] csiostor: Remove superfluous call to pci_disable_msix()

There is no need to call pci_disable_msix() in case
the previous call to pci_enable_msix() failed

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Naresh Kumar Inna <[email protected]>
Cc: Arvind Bhushan <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/csiostor/csio_isr.c | 4 +---
1 files changed, 1 insertions(+), 3 deletions(-)

diff --git a/drivers/scsi/csiostor/csio_isr.c b/drivers/scsi/csiostor/csio_isr.c
index 7ee9777..91ba91d 100644
--- a/drivers/scsi/csiostor/csio_isr.c
+++ b/drivers/scsi/csiostor/csio_isr.c
@@ -529,10 +529,8 @@ csio_enable_msix(struct csio_hw *hw)
csio_reduce_sqsets(hw, cnt - extra);
}
} else {
- if (rv > 0) {
- pci_disable_msix(hw->pdev);
+ if (rv > 0)
csio_info(hw, "Not using MSI-X, remainder:%d\n", rv);
- }

kfree(entries);
return -ENOMEM;
--
1.7.7.6

2014-02-24 08:15:45

by Alexander Gordeev

Subject: [PATCH v2 13/23] megaraid: Fail resume if MSI-X re-initialization failed

Currently the driver fails to analize MSI-X re-enablement
status on resuming and always assumes the success. This
update checks the MSI-X initialization result and fails
to resume if MSI-Xs re-enablement failed.

Signed-off-by: Alexander Gordeev <[email protected]>
---
drivers/scsi/megaraid/megaraid_sas_base.c | 8 +++++---
1 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/scsi/megaraid/megaraid_sas_base.c b/drivers/scsi/megaraid/megaraid_sas_base.c
index 3b7ad10..01afe38 100644
--- a/drivers/scsi/megaraid/megaraid_sas_base.c
+++ b/drivers/scsi/megaraid/megaraid_sas_base.c
@@ -4667,9 +4667,10 @@ megasas_resume(struct pci_dev *pdev)
goto fail_ready_state;

/* Now re-enable MSI-X */
- if (instance->msix_vectors)
- pci_enable_msix(instance->pdev, instance->msixentry,
- instance->msix_vectors);
+ if (instance->msix_vectors &&
+ pci_enable_msix(instance->pdev, instance->msixentry,
+ instance->msix_vectors))
+ goto fail_reenable_msix;

switch (instance->pdev->device) {
case PCI_DEVICE_ID_LSI_FUSION:
@@ -4756,6 +4757,7 @@ fail_init_mfi:

fail_set_dma_mask:
fail_ready_state:
+fail_reenable_msix:

pci_disable_device(pdev);

--
1.7.7.6

2014-02-24 08:00:51

by Alexander Gordeev

Subject: [PATCH v2 04/23] bfa: Use pci_enable_msix_exact() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Anil Gurumurthy <[email protected]>
Cc: Vijaya Mohan Guvva <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/bfa/bfad.c | 18 +++++-------------
1 files changed, 5 insertions(+), 13 deletions(-)

diff --git a/drivers/scsi/bfa/bfad.c b/drivers/scsi/bfa/bfad.c
index e7e4774..b7b0733 100644
--- a/drivers/scsi/bfa/bfad.c
+++ b/drivers/scsi/bfa/bfad.c
@@ -1234,26 +1234,18 @@ bfad_setup_intr(struct bfad_s *bfad)
if ((bfa_asic_id_ctc(pdev->device) && !msix_disable_ct) ||
(bfa_asic_id_cb(pdev->device) && !msix_disable_cb)) {

- error = pci_enable_msix(bfad->pcidev, msix_entries, bfad->nvec);
+ error = pci_enable_msix_exact(bfad->pcidev,
+ msix_entries, bfad->nvec);
/* In CT1 & CT2, try to allocate just one vector */
- if (error > 0 && bfa_asic_id_ctc(pdev->device)) {
+ if (error == -ENOSPC && bfa_asic_id_ctc(pdev->device)) {
printk(KERN_WARNING "bfa %s: trying one msix "
"vector failed to allocate %d[%d]\n",
bfad->pci_name, bfad->nvec, error);
bfad->nvec = 1;
- error = pci_enable_msix(bfad->pcidev,
- msix_entries, bfad->nvec);
+ error = pci_enable_msix_exact(bfad->pcidev,
+ msix_entries, 1);
}

- /*
- * Only error number of vector is available.
- * We don't have a mechanism to map multiple
- * interrupts into one vector, so even if we
- * can try to request less vectors, we don't
- * know how to associate interrupt events to
- * vectors. Linux doesn't duplicate vectors
- * in the MSIX table for this case.
- */
if (error) {
printk(KERN_WARNING "bfad%d: "
"pci_enable_msix failed (%d), "
--
1.7.7.6

2014-02-24 08:00:49

by Alexander Gordeev

Subject: [PATCH v2 02/23] bfa: Do not call pci_enable_msix() after it failed once

Function pci_enable_msix() should not be called in case
it threw a negative errno from a previous call.

Signed-off-by: Alexander Gordeev <[email protected]>
Cc: Anil Gurumurthy <[email protected]>
Cc: Vijaya Mohan Guvva <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/scsi/bfa/bfad.c | 48 ++++++++++++++++++++++------------------------
1 files changed, 23 insertions(+), 25 deletions(-)

diff --git a/drivers/scsi/bfa/bfad.c b/drivers/scsi/bfa/bfad.c
index cc0fbcd..972ff8d 100644
--- a/drivers/scsi/bfa/bfad.c
+++ b/drivers/scsi/bfa/bfad.c
@@ -1235,33 +1235,31 @@ bfad_setup_intr(struct bfad_s *bfad)
(bfa_asic_id_cb(pdev->device) && !msix_disable_cb)) {

error = pci_enable_msix(bfad->pcidev, msix_entries, bfad->nvec);
- if (error) {
- /* In CT1 & CT2, try to allocate just one vector */
- if (bfa_asic_id_ctc(pdev->device)) {
- printk(KERN_WARNING "bfa %s: trying one msix "
- "vector failed to allocate %d[%d]\n",
- bfad->pci_name, bfad->nvec, error);
- bfad->nvec = 1;
- error = pci_enable_msix(bfad->pcidev,
+ /* In CT1 & CT2, try to allocate just one vector */
+ if (error > 0 && bfa_asic_id_ctc(pdev->device)) {
+ printk(KERN_WARNING "bfa %s: trying one msix "
+ "vector failed to allocate %d[%d]\n",
+ bfad->pci_name, bfad->nvec, error);
+ bfad->nvec = 1;
+ error = pci_enable_msix(bfad->pcidev,
msix_entries, bfad->nvec);
- }
+ }

- /*
- * Only error number of vector is available.
- * We don't have a mechanism to map multiple
- * interrupts into one vector, so even if we
- * can try to request less vectors, we don't
- * know how to associate interrupt events to
- * vectors. Linux doesn't duplicate vectors
- * in the MSIX table for this case.
- */
- if (error) {
- printk(KERN_WARNING "bfad%d: "
- "pci_enable_msix failed (%d), "
- "use line based.\n",
- bfad->inst_no, error);
- goto line_based;
- }
+ /*
+ * Only error number of vector is available.
+ * We don't have a mechanism to map multiple
+ * interrupts into one vector, so even if we
+ * can try to request less vectors, we don't
+ * know how to associate interrupt events to
+ * vectors. Linux doesn't duplicate vectors
+ * in the MSIX table for this case.
+ */
+ if (error) {
+ printk(KERN_WARNING "bfad%d: "
+ "pci_enable_msix failed (%d), "
+ "use line based.\n",
+ bfad->inst_no, error);
+ goto line_based;
}

/* Disable INTX in MSI-X mode */
--
1.7.7.6

2014-02-25 08:35:27

by Alexander Gordeev

Subject: Re: [PATCH v2 00/23] scsi: Use pci_enable_msix_range() instead of pci_enable_msix()

On Mon, Feb 24, 2014 at 09:02:00AM +0100, Alexander Gordeev wrote:
> This series is against James Bottomley's SCSI tree [1], but it needs
> commit f7fc32c ("PCI/MSI: Add pci_enable_msi_exact() and
> pci_enable_msix_exact()") from from Bjorn Helgaas's PCI tree [2]:

The prerequisite commit 3ce4e86 is in 3.14-rc4 now.

--
Regards,
Alexander Gordeev
[email protected]

2014-04-02 11:47:48

by Alexander Gordeev

Subject: Re: [PATCH v2 00/23] scsi: Use pci_enable_msix_range() instead of pci_enable_msix()

On Wed, Mar 12, 2014 at 10:04:40PM -0600, Bjorn Helgaas wrote:
> Hi James,
>
> I think Alexander sent these to linux-scsi hoping that you would handle
> them, but I know it's a hassle because they depend on f7fc32c, which went
> in after the merge window.
>
> I'd be glad to review these and apply them through my tree, unless you want
> to do it. I'd like to get these merged in the v3.15 merge window so
> Alexander can move on to something else. I haven't checked for merge
> conflicts with scsi.git yet, but I assume they'd be pretty trivial if there
> are any.

Hi Bjorn,

It has shifted to v3.16, right?

> Bjorn

--
Regards,
Alexander Gordeev
[email protected]

2014-04-02 17:00:55

by Bjorn Helgaas

Subject: Re: [PATCH v2 00/23] scsi: Use pci_enable_msix_range() instead of pci_enable_msix()

On Wed, Apr 2, 2014 at 5:49 AM, Alexander Gordeev <[email protected]> wrote:
> On Wed, Mar 12, 2014 at 10:04:40PM -0600, Bjorn Helgaas wrote:
>> Hi James,
>>
>> I think Alexander sent these to linux-scsi hoping that you would handle
>> them, but I know it's a hassle because they depend on f7fc32c, which went
>> in after the merge window.
>>
>> I'd be glad to review these and apply them through my tree, unless you want
>> to do it. I'd like to get these merged in the v3.15 merge window so
>> Alexander can move on to something else. I haven't checked for merge
>> conflicts with scsi.git yet, but I assume they'd be pretty trivial if there
>> are any.
>
> Hi Bjorn,
>
> It has shifted to v3.16, right?

Yep, sorry, I meant to get to these, but was too busy, so they'll
probably go in v3.16.

Bjorn