ahci: Tweaks to the driver's PCI function

Changes since v1:
- ahci_is_mrsm() open-coded inside ahci_init_interrupt()
- changelog elaborated

Alexander Gordeev (2):
ahci: Ensure "MSI Revert to Single Message" mode is not enforced
ahci: Use pci_enable_msi_exact() instead of pci_enable_msi_range()

 drivers/ata/ahci.c |   19 ++++++++++++++-----
 drivers/ata/ahci.h |    1 +
 2 files changed, 15 insertions(+), 5 deletions(-)

1.7.7.6

The AHCI specification allows hardware to choose to revert to single MSI mode when fewer messages are allocated than requested. Yet, at least ICH10 chipset reverts to single MSI mode even when enough messages are allocated in some cases (see below).

This update forces the driver do not rely on initialization of multiple MSIs mode alone and always check if "MSI Revert to Single Message" (MRSM) mode was enforced by the controller and fallback to the single MSI mode in case it did.

That prevents a situation when the driver configured multiple per-port IRQ handlers, but the controller sends all port's interrupts to a single IRQ, which could easily screw up the interrupt handling and lead to delays and possibly crashes.

The fix was tested on a 6-port controller that successfully reverted to the single MSI mode:

00:1f.2 SATA controller: Intel Corporation 82801JI (ICH10 Family) SATA AHCI Controller (prog-if 01 [AHCI 1.0])

	Subsystem: Super Micro Computer Inc Device 10a7
	Flags: bus master, 66MHz, medium devsel, latency 0, IRQ 101
	I/O ports at f110 [size=8]
	I/O ports at f100 [size=4]
	I/O ports at f0f0 [size=8]
	I/O ports at f0e0 [size=4]
	I/O ports at f020 [size=32]
	Memory at fbf00000 (32-bit, non-prefetchable) [size=2K]
	Capabilities: [80] MSI: Enable+ Count=1/16 Maskable- 64bit-
	Capabilities: [70] Power Management version 3
	Capabilities: [a8] SATA HBA v1.0
	Capabilities: [b0] PCI Advanced Features
	Kernel driver in use: ahci

With 6 ports just 8 MSI vectors should be enough, but the adapter enforces the MRSM mode when less than 16 vectors are written to the Multiple Messages Enable PCI register. I instigated MRSM mode with the patch below:

index 1e493e6..14799f2 100644
@@ -1171,7 +1171,7 @@ static int ahci_init_interrupts(struct pci_dev *pdev, unsigned int n_ports,
 	if (hpriv->flags & AHCI_HFLAG_NO_MSI)
 		goto intx;
 
-	nvec = pci_msi_vec_count(pdev);
+	nvec = 8;
 	if (nvec < 0)
 		goto intx;
 

---
 drivers/ata/ahci.c |   11 ++++++++++-
 drivers/ata/ahci.h |    1 +
 2 files changed, 11 insertions(+), 1 deletions(-)

index 5a0bf8e..649ac1c 100644
@@ -1164,7 +1164,7 @@ static inline void ahci_gtf_filter_workaround(struct ata_host *host)
 #endif
 
 static int ahci_init_interrupts(struct pci_dev *pdev, unsigned int n_ports,
-			 struct ahci_host_priv *hpriv)
+				struct ahci_host_priv *hpriv)
 {
 	int nvec;
 

@@ -1189,6 +1189,15 @@ static int ahci_init_interrupts(struct pci_dev *pdev, unsigned int n_ports,
 	else if (nvec < 0)
 		goto intx;
 
+	/*
+	 * Fallback to single MSI mode if the controller enforced MRSM mode
+	 */
+	if (readl(hpriv->mmio + HOST_CTL) & HOST_MRSM) {
+		pci_disable_msi(pdev);
+		printk(KERN_INFO "ahci: MRSM is on, fallback to single MSI\n");
+		goto single_msi;
+	}
+
 	return nvec;
 
 single_msi:

index 51af275..b5eb886 100644
@@ -94,6 +94,7 @@ enum {
 	/* HOST_CTL bits */
 	HOST_RESET		= (1 << 0),  /* reset controller; self-clear */
 	HOST_IRQ_EN		= (1 << 1),  /* global IRQ enable */
+	HOST_MRSM		= (1 << 2),  /* MSI Revert to Single Message */
 	HOST_AHCI_EN		= (1 << 31), /* AHCI enabled */
 
 	/* HOST_CAP bits */

1.7.7.6

The driver calls pci_enable_msi_range() function with the range of [nvec..nvec] which what pci_enable_msi_exact() function is for.

---
 drivers/ata/ahci.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)

index 649ac1c..0927d5f 100644
@@ -1166,7 +1166,7 @@ static inline void ahci_gtf_filter_workaround(struct ata_host *host)
 static int ahci_init_interrupts(struct pci_dev *pdev, unsigned int n_ports,
 				struct ahci_host_priv *hpriv)
 {
-	int nvec;
+	int rc, nvec;
 
 	if (hpriv->flags & AHCI_HFLAG_NO_MSI)
 		goto intx;

@@ -1183,10 +1183,10 @@ static int ahci_init_interrupts(struct pci_dev *pdev, unsigned int n_ports,
 	if (nvec < n_ports)
 		goto single_msi;
 
-	nvec = pci_enable_msi_range(pdev, nvec, nvec);
-	if (nvec == -ENOSPC)
+	rc = pci_enable_msi_exact(pdev, nvec);
+	if (rc == -ENOSPC)
 		goto single_msi;
-	else if (nvec < 0)
+	else if (rc < 0)
 		goto intx;
 
 	/*

1.7.7.6

Applied to libata/for-3.15-fixes.

Thanks.

tejun

Oh-oh.
It seems the changelog's demo patch portion applied to the source.

Regards,
Alexander Gordeev
[email protected]

Urgh... fixed it. Please don't include proper patch in the description. It'll mess up a lot of people's workflow.

Thanks.

tejun