scsi: qla2xxx: Fix qla2x00_request_irqs() for MSI

Commit 4fa183455988 ("scsi: qla2xxx: Utilize pci_alloc_irq_vectors/
pci_free_irq_vectors calls.") use pci_alloc_irq_vectors() to replace
pci_enable_msi() but it didn't handle the return value correctly. This bug
make qla2x00 always fail to setup MSI if MSI-X fail, so fix it.

BTW, improve the log message of return value in qla2x00_request_irqs() to
avoid confusion.

Fixes: 4fa183455988 ("scsi: qla2xxx: Utilize pci_alloc_irq_vectors/pci_free_irq_vectors calls.")
Cc: Michael Hernandez <michael.hernandez@cavium.com>
Link: https://lore.kernel.org/r/1574314847-14280-1-git-send-email-chenhc@lemote.com
Signed-off-by: Huacai Chen <chenhc@lemote.com>
Acked-by: Himanshu Madhani <hmadhani@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c
index 1b8f297..2601d76 100644 (file)
--- a/drivers/scsi/qla2xxx/qla_isr.c
+++ b/drivers/scsi/qla2xxx/qla_isr.c
@@ -3650,7 +3650,7 @@ qla2x00_request_irqs(struct qla_hw_data *ha, struct rsp_que *rsp)
 skip_msix:
 
        ql_log(ql_log_info, vha, 0x0037,
-           "Falling back-to MSI mode -%d.\n", ret);
+           "Falling back-to MSI mode -- ret=%d.\n", ret);
 
        if (!IS_QLA24XX(ha) && !IS_QLA2532(ha) && !IS_QLA8432(ha) &&
            !IS_QLA8001(ha) && !IS_P3P_TYPE(ha) && !IS_QLAFX00(ha) &&
@@ -3658,13 +3658,13 @@ skip_msix:
                goto skip_msi;
 
        ret = pci_alloc_irq_vectors(ha->pdev, 1, 1, PCI_IRQ_MSI);
-       if (!ret) {
+       if (ret > 0) {
                ql_dbg(ql_dbg_init, vha, 0x0038,
                    "MSI: Enabled.\n");
                ha->flags.msi_enabled = 1;
        } else
                ql_log(ql_log_warn, vha, 0x0039,
-                   "Falling back-to INTa mode -- %d.\n", ret);
+                   "Falling back-to INTa mode -- ret=%d.\n", ret);
 skip_msi:
 
        /* Skip INTx on ISP82xx. */