[PATCH 1/2] cciss: patch to add MSI/MSI-X support

From: mike . miller
Date: Fri Dec 09 2005 - 10:56:10 EST


This patch creates a new function cciss_interrupt_mode called from
cciss_pci_init. This function determines what type of interrupt mode
to use, i.e., MSI-X, MSI, or IO-APIC.
One noticeable difference is changing the interrupt field of the
controller struct to an array of 4 unsigned ints. The Smart Array HW
is capable of generating 4 distinct interrupts depending on the transport
method in use during operation. These are:

#define DOORBELL_INT 0
Used to notify the contoller of configuration updates. We only use
this feature when in polling mode.

#define PERF_MODE_INT 1
Used when the controller is in Performant Mode.

#define SIMPLE_MODE_INT 2
Used when the controller is in Simple Mode (current Linux implementation).

#define MEMQ_MODE_INT 3
Not used.

When using IO-APIC interrupts these 4 lines are OR'ed together in hardware
so when any one fires an interrupt an is generated. In MSI or MSI-X mode
this hardware OR'ing is ignored. We must register for our interrupt
depending on what mode the controller is running. For Linux we use
SIMPLE_MODE_INT exclusively at this time. That's the reason we request
4 vectors and register the third.

Please consider this for inclusion.

Signed-off-by: Mike Miller <mike.miller@xxxxxx>
--------------------------------------------------------------------------------
cciss.c | 79 +++++++++++++++++++++++++++++++++++++++++++++++++++++------
cciss.h | 8 +++++
cciss_scsi.c | 2 -
3 files changed, 80 insertions(+), 9 deletions(-)

diff --git a/drivers/block/cciss.c b/drivers/block/cciss.c
index e34104d..ede990f 100644
--- a/drivers/block/cciss.c
+++ b/drivers/block/cciss.c
@@ -166,7 +166,7 @@ static void cciss_geometry_inquiry(int c
unsigned int block_size, InquiryData_struct *inq_buff,
drive_info_struct *drv);
static void cciss_getgeometry(int cntl_num);
-
+static void __devinit cciss_interrupt_mode(ctlr_info_t *, struct pci_dev *, __u32);
static void start_io( ctlr_info_t *h);
static int sendcmd( __u8 cmd, int ctlr, void *buff, size_t size,
unsigned int use_unit_num, unsigned int log_unit, __u8 page_code,
@@ -282,7 +282,7 @@ static int cciss_proc_get_info(char *buf
h->product_name,
(unsigned long)h->board_id,
h->firm_ver[0], h->firm_ver[1], h->firm_ver[2], h->firm_ver[3],
- (unsigned int)h->intr,
+ (unsigned int)h->intr[SIMPLE_MODE_INT],
h->num_luns,
h->Qdepth, h->commands_outstanding,
h->maxQsinceinit, h->max_outstanding, h->maxSG);
@@ -2659,6 +2659,60 @@ static int find_PCI_BAR_index(struct pci
return -1;
}

+/* If MSI/MSI-X is supported by the kernel we will try to enable it on
+ * controllers that are capable. If not, we use IO-APIC mode.
+ */
+
+static void __devinit cciss_interrupt_mode(ctlr_info_t *c, struct pci_dev *pdev, __u32 board_id)
+{
+#ifdef CONFIG_PCI_MSI
+ int err;
+ struct msix_entry cciss_msix_entries[4] = {{0,0}, {0,1},
+ {0,2}, {0,3}};
+
+ /* Some boards advertise MSI but don't really support it */
+ if ((board_id == 0x40700E11) ||
+ (board_id == 0x40800E11) ||
+ (board_id == 0x40820E11) ||
+ (board_id == 0x40830E11))
+ goto default_int_mode;
+
+ if (pci_find_capability(pdev, PCI_CAP_ID_MSIX)) {
+ err = pci_enable_msix(pdev, cciss_msix_entries, 4);
+ if (!err) {
+ c->intr[0] = cciss_msix_entries[0].vector;
+ c->intr[1] = cciss_msix_entries[1].vector;
+ c->intr[2] = cciss_msix_entries[2].vector;
+ c->intr[3] = cciss_msix_entries[3].vector;
+ c->msix_vector = 1;
+ return;
+ }
+ if (err > 0) {
+ printk(KERN_WARNING "cciss: only %d MSI-X vectors "
+ "available\n", err);
+ } else {
+ printk(KERN_WARNING "cciss: MSI-X init failed %d\n",
+ err);
+ }
+ }
+ if (pci_find_capability(pdev, PCI_CAP_ID_MSI)) {
+ if (!pci_enable_msi(pdev)) {
+ c->intr[SIMPLE_MODE_INT] = pdev->irq;
+ c->msi_vector = 1;
+ return;
+ } else {
+ printk(KERN_WARNING "cciss: MSI init failed\n");
+ c->intr[SIMPLE_MODE_INT] = pdev->irq;
+ return;
+ }
+ }
+#endif /* CONFIG_PCI_MSI */
+ /* if we get here we're going to use the default interrupt mode */
+default_int_mode:
+ c->intr[SIMPLE_MODE_INT] = pdev->irq;
+ return;
+}
+
static int cciss_pci_init(ctlr_info_t *c, struct pci_dev *pdev)
{
ushort subsystem_vendor_id, subsystem_device_id, command;
@@ -2719,7 +2773,10 @@ static int cciss_pci_init(ctlr_info_t *c
printk("board_id = %x\n", board_id);
#endif /* CCISS_DEBUG */

- c->intr = pdev->irq;
+/* If the kernel supports MSI/MSI-X we will try to enable that functionality,
+ * else we use the IO-APIC interrupt assigned to us by system ROM.
+ */
+ cciss_interrupt_mode(c, pdev, board_id);

/*
* Memory base addr is first addr , the second points to the config
@@ -3073,11 +3130,11 @@ static int __devinit cciss_init_one(stru

/* make sure the board interrupts are off */
hba[i]->access.set_intr_mask(hba[i], CCISS_INTR_OFF);
- if( request_irq(hba[i]->intr, do_cciss_intr,
+ if( request_irq(hba[i]->intr[SIMPLE_MODE_INT], do_cciss_intr,
SA_INTERRUPT | SA_SHIRQ | SA_SAMPLE_RANDOM,
hba[i]->devname, hba[i])) {
printk(KERN_ERR "cciss: Unable to get irq %d for %s\n",
- hba[i]->intr, hba[i]->devname);
+ hba[i]->intr[SIMPLE_MODE_INT], hba[i]->devname);
goto clean2;
}
hba[i]->cmd_pool_bits = kmalloc(((NR_CMDS+BITS_PER_LONG-1)/BITS_PER_LONG)*sizeof(unsigned long), GFP_KERNEL);
@@ -3183,7 +3240,7 @@ clean4:
NR_CMDS * sizeof( ErrorInfo_struct),
hba[i]->errinfo_pool,
hba[i]->errinfo_pool_dhandle);
- free_irq(hba[i]->intr, hba[i]);
+ free_irq(hba[i]->intr[SIMPLE_MODE_INT], hba[i]);
clean2:
unregister_blkdev(hba[i]->major, hba[i]->devname);
clean1:
@@ -3224,7 +3281,15 @@ static void __devexit cciss_remove_one (
printk(KERN_WARNING "Error Flushing cache on controller %d\n",
i);
}
- free_irq(hba[i]->intr, hba[i]);
+ free_irq(hba[i]->intr[2], hba[i]);
+
+#ifdef CONFIG_PCI_MSI
+ if (hba[i]->msix_vector)
+ pci_disable_msix(hba[i]->pdev);
+ else if (hba[i]->msi_vector)
+ pci_disable_msi(hba[i]->pdev);
+#endif /* CONFIG_PCI_MSI */
+
pci_set_drvdata(pdev, NULL);
iounmap(hba[i]->vaddr);
cciss_unregister_scsi(i); /* unhook from SCSI subsystem */
diff --git a/drivers/block/cciss.h b/drivers/block/cciss.h
index 3b0858c..ad45e58 100644
--- a/drivers/block/cciss.h
+++ b/drivers/block/cciss.h
@@ -65,7 +65,6 @@ struct ctlr_info
unsigned long io_mem_addr;
unsigned long io_mem_length;
CfgTable_struct __iomem *cfgtable;
- unsigned int intr;
int interrupts_enabled;
int major;
int max_commands;
@@ -74,6 +73,13 @@ struct ctlr_info
int num_luns;
int highest_lun;
int usage_count; /* number of opens all all minor devices */
+# define DOORBELL_INT 0
+# define PERF_MODE_INT 1
+# define SIMPLE_MODE_INT 2
+# define MEMQ_MODE_INT 3
+ unsigned int intr[4];
+ unsigned int msix_vector;
+ unsigned int msi_vector;

// information about each logical volume
drive_info_struct drv[CISS_MAX_LUN];
diff --git a/drivers/block/cciss_scsi.c b/drivers/block/cciss_scsi.c
index 2942d32..9e35de0 100644
--- a/drivers/block/cciss_scsi.c
+++ b/drivers/block/cciss_scsi.c
@@ -714,7 +714,7 @@ cciss_scsi_detect(int ctlr)
((struct cciss_scsi_adapter_data_t *)
hba[ctlr]->scsi_ctlr)->scsi_host = (void *) sh;
sh->hostdata[0] = (unsigned long) hba[ctlr];
- sh->irq = hba[ctlr]->intr;
+ sh->irq = hba[ctlr]->intr[SIMPLE_MODE_INT];
sh->unique_id = sh->irq;
error = scsi_add_host(sh, &hba[ctlr]->pdev->dev);
if (error)

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/