Skip to contentSkip to conversation

saa7164: use an MSI interrupt when available

PATCH Superseded Brendan McGrath 5 participants 14 messages Feb 26, 2015 – Jun 5, 2015 Media (V4L/DVB)

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 saa7164: use an MSI interrupt when available

From: Brendan McGrath <[email protected]>

[media] saa7164: use an MSI interrupt when available

Fixes a known crash which most commonly occurs when multiple saa7164 chips are in use.

---
 drivers/media/pci/saa7164/saa7164-core.c | 34 
++++++++++++++++++++++++++++++--
 drivers/media/pci/saa7164/saa7164.h      |  1 +
 2 files changed, 33 insertions(+), 2 deletions(-)

This patch falls back to the original code - a 'shared' interrupt - when MSI is not available (or encounters an error).

Many of today's cards that use the saa7164 chip operate on a PCI-E bus (thus MSI should be available). Examples: the Hauppage HVR-2200 and HVR-2250

This enhancement also fixes an issue that was causing the driver to crash:
http://permalink.gmane.org/gmane.linux.drivers.video-input-infrastructure/83948

I believe the root cause of the crash is due to a DMA/IRQ race condition. It most commonly occurs when the saa7164 driver is dealing with more than one saa7164 chip (the HVR-2200 and HVR-2250 for example have two - one for each tuner). Given MSI avoids DMA/IRQ race conditions - this would explain why the patch works as a fix.

drivers/media/pci/saa7164/saa7164-core.c +32 -2
index 4b0bec3..083bea4 100644
@@ -1230,8 +1230,33 @@ static int saa7164_initdev(struct pci_dev *pci_dev,
         goto fail_irq;
     }
 
-    err = request_irq(pci_dev->irq, saa7164_irq,
-        IRQF_SHARED, dev->name, dev);
+    /* irq bit */
+    err = pci_enable_msi(pci_dev);
+
+    if (!err) {
+        /* no error - so request an msi interrupt */
+        err = request_irq(pci_dev->irq, saa7164_irq, 0,
+                  dev->name, dev);
+
+        if (err) {
+            /* fall back to legacy interrupt */
+            printk(KERN_ERR "%s() Failed to get an MSI interrupt."
+                " Falling back to a shared IRQ\n", __func__);
+            pci_disable_msi(pci_dev);
+        } else {
+            dev->msi = true;
+        }
+    }
+
+    if (err) {
+        dev->msi = false;
+        /* if we have an error (i.e. we don't have an interrupt)
+             - fallback to legacy interrupt */
+
+        err = request_irq(pci_dev->irq, saa7164_irq,
+                    IRQF_SHARED, dev->name, dev);
+    }
+
     if (err < 0) {
         printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,
             pci_dev->irq);
@@ -1441,6 +1466,11 @@ static void saa7164_finidev(struct pci_dev *pci_dev)
     /* unregister stuff */
     free_irq(pci_dev->irq, dev);
 
+    if (dev->msi) {
+        pci_disable_msi(pci_dev);
+        dev->msi = false;
+    }
+
     mutex_lock(&devlist);
     list_del(&dev->devlist);
     mutex_unlock(&devlist);
drivers/media/pci/saa7164/saa7164.h +1 -0
index cd1a07c..6df4b252 100644
@@ -459,6 +459,7 @@ struct saa7164_dev {
     /* Interrupt status and ack registers */
     u32 int_status;
     u32 int_ack;
+    u32 msi;
 
     struct cmd            cmds[SAA_CMD_MAX_MSG_UNITS];
     struct mutex            lock;
Signature · 1 line
1.9.1
#1 lore raw
Quoted 5 lines

I believe the root cause of the crash is due to a DMA/IRQ race condition. It most commonly occurs when the saa7164 driver is dealing with more than one saa7164 chip (the HVR-2200 and HVR-2250 for example have two - one for each tuner). Given MSI avoids DMA/IRQ race conditions - this would explain why the patch works as a fix.

Brendan, thanks.

With MSI I've had some people report complete success, others still have the issues.

In my experience this does help with i2c timeout issues but not completely in every case. I've also seen it with single card instances so you descripton above is close - but not quiet accurate in all cases.

While I'm generally OK with changing the driver behaviour to enable MSI by default, please add a module option to allow the behaviour to be disabled, reverting the driver back to existing behaviour.

Once this is done, I'll be happy to Ack it.

Thanks again.

- Steve

Signature · 2 lines
Steven Toth - Kernel Labs
http://www.kernellabs.com
Tested-by
#2 lore raw

The patch allows the card to support using both internal tuners without crashing out. I've been recording significantly since the 23rd, it just works now in comparison to constant maintenance.

I am under the impression it was against the spec to have a PCI-E card without MSI support. Wouldn't the fallback code as well work in this regard?

Thanks a ton for the patch,
Kyle.

Quoted 33 lines · Steven Toth
Quoted 5 lines

I believe the root cause of the crash is due to a DMA/IRQ race condition. It most commonly occurs when the saa7164 driver is dealing with more than one saa7164 chip (the HVR-2200 and HVR-2250 for example have two - one for each tuner). Given MSI avoids DMA/IRQ race conditions - this would explain why the patch works as a fix.

Brendan, thanks.

With MSI I've had some people report complete success, others still have the issues.

In my experience this does help with i2c timeout issues but not completely in every case. I've also seen it with single card instances so you descripton above is close - but not quiet accurate in all cases.

While I'm generally OK with changing the driver behaviour to enable MSI by default, please add a module option to allow the behaviour to be disabled, reverting the driver back to existing behaviour.

Once this is done, I'll be happy to Ack it.

Thanks again.

- Steve

--
Steven Toth - Kernel Labs
http://www.kernellabs.com
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#3 lore raw
Quoted 3 lines

I am under the impression it was against the spec to have a PCI-E card without MSI support. Wouldn't the fallback code as well work in this regard?

Not if the motherboard MSI implementation is flakey, like we've seen in the past with other PCIe bridges and their interaction with root controllers. I've actually seen this with some commercial customers using the 7164. With this new patch MSI gets enabled, the 'solution' doesn't work properly and now we expect users to compile their own kernels, just to get back to a previously working solution.

I'm happy to take the patch if the option can be disabled.

Signature · 2 lines
Steven Toth - Kernel Labs
http://www.kernellabs.com
#4 lore raw

This isn't my patch, however shouldn't these bridges already exist in drivers/pci/quirks.c?

Kyle.

Quoted 16 lines · Steven Toth
Quoted 3 lines

I am under the impression it was against the spec to have a PCI-E card without MSI support. Wouldn't the fallback code as well work in this regard?

Not if the motherboard MSI implementation is flakey, like we've seen in the past with other PCIe bridges and their interaction with root controllers. I've actually seen this with some commercial customers using the 7164. With this new patch MSI gets enabled, the 'solution' doesn't work properly and now we expect users to compile their own kernels, just to get back to a previously working solution.

I'm happy to take the patch if the option can be disabled.

--
Steven Toth - Kernel Labs
http://www.kernellabs.com

#5 lore raw

PATCH v2 1/1 saa7164: use an MSI interrupt when available

In reply to Steven Toth

From: Brendan McGrath <[email protected]> [media] saa7164: use an MSI interrupt when available

Enhances driver to use an MSI interrupt when available.

Adds the module option 'enable_msi' (type bool) which by default is enabled. Can be set to 'N' to disable.

Fixes (or can reduce the occurrence of) a crash which is most commonly reported when multiple saa7164 chips are in use.

---

Thanks for the feedback Steven (and thanks for all your hard work on this driver!).

And thank-you Kyle for taking the time to test this patch. Much appreciated.

 drivers/media/pci/saa7164/saa7164-core.c | 40 
++++++++++++++++++++++++++++++--
 drivers/media/pci/saa7164/saa7164.h      |  1 +
 2 files changed, 39 insertions(+), 2 deletions(-)
drivers/media/pci/saa7164/saa7164-core.c +38 -2
index 4b0bec3..1aff2ee 100644
@@ -85,6 +85,11 @@ module_param(guard_checking, int, 0644);
 MODULE_PARM_DESC(guard_checking,
    "enable dma sanity checking for buffer overruns");
 
+static bool enable_msi = true;
+module_param(enable_msi, bool, 0444);
+MODULE_PARM_DESC(enable_msi,
+     "enable the use of an msi interrupt if available");
+
 static unsigned int saa7164_devcount;
 
 static DEFINE_MUTEX(devlist);
@@ -1230,8 +1235,34 @@ static int saa7164_initdev(struct pci_dev *pci_dev,
        goto fail_irq;
    }
 
-    err = request_irq(pci_dev->irq, saa7164_irq,
-        IRQF_SHARED, dev->name, dev);
+    /* irq bit */
+    if (enable_msi)
+        err = pci_enable_msi(pci_dev);
+
+    if (!err && enable_msi) {
+        /* no error - so request an msi interrupt */
+        err = request_irq(pci_dev->irq, saa7164_irq, 0,
+                  dev->name, dev);
+
+        if (err) {
+            /* fall back to legacy interrupt */
+            printk(KERN_ERR "%s() Failed to get an MSI interrupt."
+                " Falling back to a shared IRQ\n", __func__);
+            pci_disable_msi(pci_dev);
+        } else {
+            dev->msi = true;
+        }
+    }
+
+    if ((!enable_msi) || err) {
+        dev->msi = false;
+        /* if we have an error (i.e. we don't have an interrupt)
+             or msi is not enabled - fallback to shared interrupt */
+
+        err = request_irq(pci_dev->irq, saa7164_irq,
+                    IRQF_SHARED, dev->name, dev);
+    }
+
    if (err < 0) {
        printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,
            pci_dev->irq);
@@ -1441,6 +1472,11 @@ static void saa7164_finidev(struct pci_dev *pci_dev)
    /* unregister stuff */
    free_irq(pci_dev->irq, dev);
 
+    if (dev->msi) {
+        pci_disable_msi(pci_dev);
+        dev->msi = false;
+    }
+
    mutex_lock(&devlist);
    list_del(&dev->devlist);
    mutex_unlock(&devlist);
drivers/media/pci/saa7164/saa7164.h +1 -0
index cd1a07c..6df4b252 100644
@@ -459,6 +459,7 @@ struct saa7164_dev {
    /* Interrupt status and ack registers */
    u32 int_status;
    u32 int_ack;
+    u32 msi;
 
    struct cmd            cmds[SAA_CMD_MAX_MSG_UNITS];
    struct mutex            lock;

--
1.9.1

Quoted 25 lines · Steven Toth
Quoted 5 lines

I believe the root cause of the crash is due to a DMA/IRQ race condition. It most commonly occurs when the saa7164 driver is dealing with more than one saa7164 chip (the HVR-2200 and HVR-2250 for example have two - one for each tuner). Given MSI avoids DMA/IRQ race conditions - this would explain why the patch works as a fix.

Brendan, thanks.

With MSI I've had some people report complete success, others still have the issues.

In my experience this does help with i2c timeout issues but not completely in every case. I've also seen it with single card instances so you descripton above is close - but not quiet accurate in all cases.

While I'm generally OK with changing the driver behaviour to enable MSI by default, please add a module option to allow the behaviour to be disabled, reverting the driver back to existing behaviour.

Once this is done, I'll be happy to Ack it.

Thanks again.

- Steve

Reviewed-by on v2 1/1
#6 lore raw
Quoted 5 lines · Brendan McGrath

From: Brendan McGrath <[email protected]> [media] saa7164: use an MSI interrupt when available

Enhances driver to use an MSI interrupt when available.

Quoted 7 lines

Adds the module option 'enable_msi' (type bool) which by default is enabled. Can be set to 'N' to disable.

Fixes (or can reduce the occurrence of) a crash which is most commonly reported when multiple saa7164 chips are in use.

Thank you Brendan.

Signature · 2 lines
Steven Toth - Kernel Labs
http://www.kernellabs.com
#7 lore raw

PATCH v3 1/1 saa7164: use an MSI interrupt when available

In reply to Brendan McGrath

Enhances driver to use an MSI interrupt when available.

Adds the module option 'enable_msi' (type bool) which by default is enabled. Can be set to 'N' to disable.

Fixes (or can reduce the occurrence of) a crash which is most commonly reported when multiple saa7164 chips are in use. A reported example can be found here:
http://permalink.gmane.org/gmane.linux.drivers.video-input-infrastructure/83948

---

Changes since v3:
* Added link to reported incident in comments
* Added Reviewed-by tag from Steven Toth
* Used git send-mail (as my patch got mangled using Thunderbird)
* Added the linux-kernel mailing list to the recipient list

Note I have not included the Tested-by tag from Kyle Sanderson as his test was on v1 (which did not include the module option 'enable_msi') - although much of the code in v3 was there is v1.

 drivers/media/pci/saa7164/saa7164-core.c | 40 ++++++++++++++++++++++++++++++--
 drivers/media/pci/saa7164/saa7164.h      |  1 +
 2 files changed, 39 insertions(+), 2 deletions(-)
drivers/media/pci/saa7164/saa7164-core.c +38 -2
index 4b0bec3..c9a6447 100644
@@ -85,6 +85,11 @@ module_param(guard_checking, int, 0644);
 MODULE_PARM_DESC(guard_checking,
 	"enable dma sanity checking for buffer overruns");
 
+static bool enable_msi = true;
+module_param(enable_msi, bool, 0444);
+MODULE_PARM_DESC(enable_msi,
+		"enable the use of an msi interrupt if available");
+
 static unsigned int saa7164_devcount;
 
 static DEFINE_MUTEX(devlist);
@@ -1230,8 +1235,34 @@ static int saa7164_initdev(struct pci_dev *pci_dev,
 		goto fail_irq;
 	}
 
-	err = request_irq(pci_dev->irq, saa7164_irq,
-		IRQF_SHARED, dev->name, dev);
+	/* irq bit */
+	if (enable_msi)
+		err = pci_enable_msi(pci_dev);
+
+	if (!err && enable_msi) {
+		/* no error - so request an msi interrupt */
+		err = request_irq(pci_dev->irq, saa7164_irq, 0,
+				  dev->name, dev);
+
+		if (err) {
+			/* fall back to legacy interrupt */
+			printk(KERN_ERR "%s() Failed to get an MSI interrupt."
+			       " Falling back to a shared IRQ\n", __func__);
+			pci_disable_msi(pci_dev);
+		} else {
+			dev->msi = true;
+		}
+	}
+
+	if ((!enable_msi) || err) {
+		dev->msi = false;
+		/* if we have an error (i.e. we don't have an interrupt)
+			 or msi is not enabled - fallback to shared interrupt */
+
+		err = request_irq(pci_dev->irq, saa7164_irq,
+				  IRQF_SHARED, dev->name, dev);
+	}
+
 	if (err < 0) {
 		printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,
 			pci_dev->irq);
@@ -1441,6 +1472,11 @@ static void saa7164_finidev(struct pci_dev *pci_dev)
 	/* unregister stuff */
 	free_irq(pci_dev->irq, dev);
 
+	if (dev->msi) {
+		pci_disable_msi(pci_dev);
+		dev->msi = false;
+	}
+
 	mutex_lock(&devlist);
 	list_del(&dev->devlist);
 	mutex_unlock(&devlist);
drivers/media/pci/saa7164/saa7164.h +1 -0
index cd1a07c..6df4b252 100644
@@ -459,6 +459,7 @@ struct saa7164_dev {
 	/* Interrupt status and ack registers */
 	u32 int_status;
 	u32 int_ack;
+	u32 msi;
 
 	struct cmd			cmds[SAA_CMD_MAX_MSG_UNITS];
 	struct mutex			lock;
Signature · 1 line
1.9.1
#8 lore raw
Quoted 5 lines · Brendan McGrath

Enhances driver to use an MSI interrupt when available.

Adds the module option 'enable_msi' (type bool) which by default is enabled. Can be set to 'N' to disable.

Quoted 8 lines

Fixes (or can reduce the occurrence of) a crash which is most commonly reported when multiple saa7164 chips are in use. A reported example can be found here:
http://permalink.gmane.org/gmane.linux.drivers.video-input-infrastructure/83948

I wanted to report that I have been running this patch for about a week now and I have had no instances of the zero free sequences issue.

Thank you very much!

David

#9 lore raw

PATCH v4 1/1 saa7164: use an MSI interrupt when available

In reply to Brendan McGrath

Enhances driver to use an MSI interrupt when available.

Adds the module option 'enable_msi' (type bool) which by default is enabled. Can be set to 'N' to disable.

Fixes (or can reduce the occurrence of) a crash which is most commonly reported when both digital tuners of the saa7164 chip is in use. A reported example can be found here:
http://permalink.gmane.org/gmane.linux.drivers.video-input-infrastructure/83948

---
Changes since v3:
- fixes a conflict with a commit (3f845f3c4cf4) made to the media_tree after v3 was created (only the unified context has been changed)
- corrected comments to reflect that the reported incident occured more commonly when multiple tuners were in use (not multiple saa7164 chips as previously stated)

 drivers/media/pci/saa7164/saa7164-core.c | 40 ++++++++++++++++++++++++++++++--
 drivers/media/pci/saa7164/saa7164.h      |  1 +
 2 files changed, 39 insertions(+), 2 deletions(-)
drivers/media/pci/saa7164/saa7164-core.c +38 -2
index 9cf3c6c..7635598 100644
@@ -85,6 +85,11 @@ module_param(guard_checking, int, 0644);
 MODULE_PARM_DESC(guard_checking,
 	"enable dma sanity checking for buffer overruns");
 
+static bool enable_msi = true;
+module_param(enable_msi, bool, 0444);
+MODULE_PARM_DESC(enable_msi,
+		"enable the use of an msi interrupt if available");
+
 static unsigned int saa7164_devcount;
 
 static DEFINE_MUTEX(devlist);
@@ -1230,8 +1235,34 @@ static int saa7164_initdev(struct pci_dev *pci_dev,
 		goto fail_irq;
 	}
 
-	err = request_irq(pci_dev->irq, saa7164_irq,
-		IRQF_SHARED, dev->name, dev);
+	/* irq bit */
+	if (enable_msi)
+		err = pci_enable_msi(pci_dev);
+
+	if (!err && enable_msi) {
+		/* no error - so request an msi interrupt */
+		err = request_irq(pci_dev->irq, saa7164_irq, 0,
+				  dev->name, dev);
+
+		if (err) {
+			/* fall back to legacy interrupt */
+			printk(KERN_ERR "%s() Failed to get an MSI interrupt."
+			       " Falling back to a shared IRQ\n", __func__);
+			pci_disable_msi(pci_dev);
+		} else {
+			dev->msi = true;
+		}
+	}
+
+	if ((!enable_msi) || err) {
+		dev->msi = false;
+		/* if we have an error (i.e. we don't have an interrupt)
+			 or msi is not enabled - fallback to shared interrupt */
+
+		err = request_irq(pci_dev->irq, saa7164_irq,
+				  IRQF_SHARED, dev->name, dev);
+	}
+
 	if (err < 0) {
 		printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,
 			pci_dev->irq);
@@ -1439,6 +1470,11 @@ static void saa7164_finidev(struct pci_dev *pci_dev)
 	/* unregister stuff */
 	free_irq(pci_dev->irq, dev);
 
+	if (dev->msi) {
+		pci_disable_msi(pci_dev);
+		dev->msi = false;
+	}
+
 	pci_disable_device(pci_dev);
 
 	mutex_lock(&devlist);
drivers/media/pci/saa7164/saa7164.h +1 -0
index cd1a07c..6df4b252 100644
@@ -459,6 +459,7 @@ struct saa7164_dev {
 	/* Interrupt status and ack registers */
 	u32 int_status;
 	u32 int_ack;
+	u32 msi;
 
 	struct cmd			cmds[SAA_CMD_MAX_MSG_UNITS];
 	struct mutex			lock;
Signature · 1 line
1.9.1
#10 lore raw

Hi Brendan,

The idea is good,
Some comments bellow.

Em Sat, 14 Mar 2015 14:27:39 +1100
Brendan McGrath <[email protected]> escreveu:

Quoted 39 lines

Enhances driver to use an MSI interrupt when available.

Adds the module option 'enable_msi' (type bool) which by default is enabled. Can be set to 'N' to disable.

Fixes (or can reduce the occurrence of) a crash which is most commonly reported when both digital tuners of the saa7164 chip is in use. A reported example can be found here:
http://permalink.gmane.org/gmane.linux.drivers.video-input-infrastructure/83948

---
Changes since v3:
- fixes a conflict with a commit (3f845f3c4cf4) made to the media_tree after v3 was created (only the unified context has been changed)
- corrected comments to reflect that the reported incident occured more commonly when multiple tuners were in use (not multiple saa7164 chips as previously stated)

 drivers/media/pci/saa7164/saa7164-core.c | 40 ++++++++++++++++++++++++++++++--
 drivers/media/pci/saa7164/saa7164.h      |  1 +
 2 files changed, 39 insertions(+), 2 deletions(-)
drivers/media/pci/saa7164/saa7164-core.c +5 -0
index 9cf3c6c..7635598 100644
@@ -85,6 +85,11 @@ module_param(guard_checking, int, 0644);
 MODULE_PARM_DESC(guard_checking,
 	"enable dma sanity checking for buffer overruns");
 
+static bool enable_msi = true;
+module_param(enable_msi, bool, 0444);
+MODULE_PARM_DESC(enable_msi,
+		"enable the use of an msi interrupt if available");
+
 static unsigned int saa7164_devcount;
 
 static DEFINE_MUTEX(devlist);

drivers/media/pci/saa7164/saa7164-core.c static int saa7164_initdev(struct pci_dev *pci_dev, · line 1240

Quoted 8 lines
drivers/media/pci/saa7164/saa7164-core.c +3 -2
@@ -1230,8 +1235,34 @@ static int saa7164_initdev(struct pci_dev *pci_dev,
 		goto fail_irq;
 	}
 
-	err = request_irq(pci_dev->irq, saa7164_irq,
-		IRQF_SHARED, dev->name, dev);
+	/* irq bit */
+	if (enable_msi)
+		err = pci_enable_msi(pci_dev);

It is worth printing a warning about that.

Quoted 5 lines
drivers/media/pci/saa7164/saa7164-core.c +5 -0
+
+	if (!err && enable_msi) {
+		/* no error - so request an msi interrupt */
+		err = request_irq(pci_dev->irq, saa7164_irq, 0,
+				  dev->name, dev);

drivers/media/pci/saa7164/saa7164-core.c static int saa7164_initdev(struct pci_dev *pci_dev, · line 1255

Quoted 10 lines
drivers/media/pci/saa7164/saa7164-core.c +10 -0
+
+		if (err) {
+			/* fall back to legacy interrupt */
+			printk(KERN_ERR "%s() Failed to get an MSI interrupt."
+			       " Falling back to a shared IRQ\n", __func__);
+			pci_disable_msi(pci_dev);
+		} else {
+			dev->msi = true;
+		}
+	}

It would be better to join this if with the next one, in order to make clear that both belong to the enable_msi logic. Something like:

static bool saa7164_enable_msi(struct device *pci_dev)

{
	if (!enable_msi)
		return false;

	err = pci_enable_msi(pci_dev);
	if (err) {
		printf(KERN_ERR "%s() Failed to enable MSI"
		       " Falling back to a shared IRQ\n", __func__);
		return false;
	}
	err = request_irq(pci_dev->irq, saa7164_irq, 0,
			  dev->name, dev);
	if (err) {
		printk(KERN_ERR "%s() Failed to get an MSI interrupt."
		       " Falling back to a shared IRQ\n", __func__);
		pci_disable_msi(pci_dev);
		return false;
	}
	return true;
}

Then, at the probe function, you could simply do:

	if (saa7164_enable_msi(pci_dev)) {
		dev->msi = true;
	} else {
		/* SOME_FALLBACK_CODE */
	}

The probe function is already complex enough. Breaking it into small inlined functions makes easier to review. The removal of the if's is an extra bonus, as the code size will likely be a little bit smaller.

drivers/media/pci/saa7164/saa7164-core.c static int saa7164_initdev(struct pci_dev *pci_dev, · line 1258

Quoted 3 lines
drivers/media/pci/saa7164/saa7164-core.c +3 -0
+
+	if ((!enable_msi) || err) {
+		dev->msi = false;

No need, as dev was initialized with kzalloc(), with zeroes all fields.

Also, you can simplify the "if" clause to:

	if (!dev->msi) {

That makes clearer that the code below is to be used when MSI is not enabled or not initialized properly.

Quoted 27 lines
drivers/media/pci/saa7164/saa7164-core.c +12 -0
+		/* if we have an error (i.e. we don't have an interrupt)
+			 or msi is not enabled - fallback to shared interrupt */
+
+		err = request_irq(pci_dev->irq, saa7164_irq,
+				  IRQF_SHARED, dev->name, dev);
+	}
+
 	if (err < 0) {
 		printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,
 			pci_dev->irq);
@@ -1439,6 +1470,11 @@ static void saa7164_finidev(struct pci_dev *pci_dev)
 	/* unregister stuff */
 	free_irq(pci_dev->irq, dev);
 
+	if (dev->msi) {
+		pci_disable_msi(pci_dev);
+		dev->msi = false;
+	}
+
 	pci_disable_device(pci_dev);
 
 	mutex_lock(&devlist);
drivers/media/pci/saa7164/saa7164.h +0 -0
index cd1a07c..6df4b252 100644

drivers/media/pci/saa7164/saa7164.h struct saa7164_dev { · line 462

Quoted 4 lines
drivers/media/pci/saa7164/saa7164.h +1 -0
@@ -459,6 +459,7 @@ struct saa7164_dev {
 	/* Interrupt status and ack registers */
 	u32 int_status;
 	u32 int_ack;
+	u32 msi;

Should be bool instead of u32.

Quoted 3 lines
drivers/media/pci/saa7164/saa7164.h +0 -0
 	struct cmd			cmds[SAA_CMD_MAX_MSG_UNITS];
 	struct mutex			lock;
#11 lore raw

PATCH v5 1/1 saa7164: use an MSI interrupt when available

Enhances driver to use an MSI interrupt when available.

Adds the module option 'enable_msi' (type bool) which by default is enabled. Can be set to 'N' to disable.

Fixes (or can reduce the occurrence of) a crash which is most commonly reported when both digital tuners of the saa7164 chip is in use. A reported example can be found here:
http://permalink.gmane.org/gmane.linux.drivers.video-input-infrastructure/83948

---
Changes since v4:
- improved readability by taking on suggestions made by Mauro
- the msi variable in the saa7164_dev structure is now a bool

Thanks Mauro - good suggestions and I think I've taken on board all of them.

 drivers/media/pci/saa7164/saa7164-core.c | 66 ++++++++++++++++++++++++++++----
 drivers/media/pci/saa7164/saa7164.h      |  1 +
 2 files changed, 60 insertions(+), 7 deletions(-)
drivers/media/pci/saa7164/saa7164-core.c +59 -7
index 9cf3c6c..5e4a9f0 100644
@@ -85,6 +85,11 @@ module_param(guard_checking, int, 0644);
 MODULE_PARM_DESC(guard_checking,
 	"enable dma sanity checking for buffer overruns");
 
+static bool enable_msi = true;
+module_param(enable_msi, bool, 0444);
+MODULE_PARM_DESC(enable_msi,
+		"enable the use of an msi interrupt if available");
+
 static unsigned int saa7164_devcount;
 
 static DEFINE_MUTEX(devlist);
@@ -1184,6 +1189,39 @@ static int saa7164_thread_function(void *data)
 	return 0;
 }
 
+static bool saa7164_enable_msi(struct pci_dev *pci_dev, struct saa7164_dev *dev)
+{
+	int err;
+
+	if (!enable_msi) {
+		printk(KERN_WARNING "%s() MSI disabled by module parameter 'enable_msi'"
+		       , __func__);
+		return false;
+	}
+
+	err = pci_enable_msi(pci_dev);
+
+	if (err) {
+		printk(KERN_ERR "%s() Failed to enable MSI interrupt."
+			" Falling back to a shared IRQ\n", __func__);
+		return false;
+	}
+
+	/* no error - so request an msi interrupt */
+	err = request_irq(pci_dev->irq, saa7164_irq, 0,
+						dev->name, dev);
+
+	if (err) {
+		/* fall back to legacy interrupt */
+		printk(KERN_ERR "%s() Failed to get an MSI interrupt."
+		       " Falling back to a shared IRQ\n", __func__);
+		pci_disable_msi(pci_dev);
+		return false;
+	}
+
+	return true;
+}
+
 static int saa7164_initdev(struct pci_dev *pci_dev,
 			   const struct pci_device_id *pci_id)
 {
@@ -1230,13 +1268,22 @@ static int saa7164_initdev(struct pci_dev *pci_dev,
 		goto fail_irq;
 	}
 
-	err = request_irq(pci_dev->irq, saa7164_irq,
-		IRQF_SHARED, dev->name, dev);
-	if (err < 0) {
-		printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,
-			pci_dev->irq);
-		err = -EIO;
-		goto fail_irq;
+	/* irq bit */
+	if (saa7164_enable_msi(pci_dev, dev)) {
+		dev->msi = true;
+	} else {
+		/* if we have an error (i.e. we don't have an interrupt)
+			 or msi is not enabled - fallback to shared interrupt */
+
+		err = request_irq(pci_dev->irq, saa7164_irq,
+				IRQF_SHARED, dev->name, dev);
+
+		if (err < 0) {
+			printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,
+			       pci_dev->irq);
+			err = -EIO;
+			goto fail_irq;
+		}
 	}
 
 	pci_set_drvdata(pci_dev, dev);
@@ -1439,6 +1486,11 @@ static void saa7164_finidev(struct pci_dev *pci_dev)
 	/* unregister stuff */
 	free_irq(pci_dev->irq, dev);
 
+	if (dev->msi) {
+		pci_disable_msi(pci_dev);
+		dev->msi = false;
+	}
+
 	pci_disable_device(pci_dev);
 
 	mutex_lock(&devlist);
drivers/media/pci/saa7164/saa7164.h +1 -0
index cd1a07c..75a3f51 100644
@@ -459,6 +459,7 @@ struct saa7164_dev {
 	/* Interrupt status and ack registers */
 	u32 int_status;
 	u32 int_ack;
+	bool msi;
 
 	struct cmd			cmds[SAA_CMD_MAX_MSG_UNITS];
 	struct mutex			lock;
Signature · 1 line
1.9.1
#12 lore raw

This has been plaguing users for years (there's a number of threads on the Ubuntu board). I've been using revision 1 of the patch without issue since early February. This is from having to constantly reboot the system to flawless recording. If something has been outstanding from Brendan, please let me know and I'll happily make the requested changes.

Can we please merge this? There are at-least three consumers in this thread alone that have confirmed this fixes the saa7164 driver for the HVR-2250 device.
Kyle.

PS: I can't seem to source out who owns this in the MAINTAINERS file?

Quoted 141 lines · Brendan McGrath

Enhances driver to use an MSI interrupt when available.

Adds the module option 'enable_msi' (type bool) which by default is enabled. Can be set to 'N' to disable.

Fixes (or can reduce the occurrence of) a crash which is most commonly reported when both digital tuners of the saa7164 chip is in use. A reported example can be found here:
http://permalink.gmane.org/gmane.linux.drivers.video-input-infrastructure/83948

---
Changes since v4:
- improved readability by taking on suggestions made by Mauro
- the msi variable in the saa7164_dev structure is now a bool

Thanks Mauro - good suggestions and I think I've taken on board all of them.

 drivers/media/pci/saa7164/saa7164-core.c | 66 ++++++++++++++++++++++++++++----
 drivers/media/pci/saa7164/saa7164.h      |  1 +
 2 files changed, 60 insertions(+), 7 deletions(-)
drivers/media/pci/saa7164/saa7164-core.c +59 -7
index 9cf3c6c..5e4a9f0 100644
@@ -85,6 +85,11 @@ module_param(guard_checking, int, 0644);
 MODULE_PARM_DESC(guard_checking,
        "enable dma sanity checking for buffer overruns");
 
+static bool enable_msi = true;
+module_param(enable_msi, bool, 0444);
+MODULE_PARM_DESC(enable_msi,
+               "enable the use of an msi interrupt if available");
+
 static unsigned int saa7164_devcount;
 
 static DEFINE_MUTEX(devlist);
@@ -1184,6 +1189,39 @@ static int saa7164_thread_function(void *data)
        return 0;
 }
 
+static bool saa7164_enable_msi(struct pci_dev *pci_dev, struct saa7164_dev *dev)
+{
+       int err;
+
+       if (!enable_msi) {
+               printk(KERN_WARNING "%s() MSI disabled by module parameter 'enable_msi'"
+                      , __func__);
+               return false;
+       }
+
+       err = pci_enable_msi(pci_dev);
+
+       if (err) {
+               printk(KERN_ERR "%s() Failed to enable MSI interrupt."
+                       " Falling back to a shared IRQ\n", __func__);
+               return false;
+       }
+
+       /* no error - so request an msi interrupt */
+       err = request_irq(pci_dev->irq, saa7164_irq, 0,
+                                               dev->name, dev);
+
+       if (err) {
+               /* fall back to legacy interrupt */
+               printk(KERN_ERR "%s() Failed to get an MSI interrupt."
+                      " Falling back to a shared IRQ\n", __func__);
+               pci_disable_msi(pci_dev);
+               return false;
+       }
+
+       return true;
+}
+
 static int saa7164_initdev(struct pci_dev *pci_dev,
                           const struct pci_device_id *pci_id)
 {
@@ -1230,13 +1268,22 @@ static int saa7164_initdev(struct pci_dev *pci_dev,
                goto fail_irq;
        }
 
-       err = request_irq(pci_dev->irq, saa7164_irq,
-               IRQF_SHARED, dev->name, dev);
-       if (err < 0) {
-               printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,
-                       pci_dev->irq);
-               err = -EIO;
-               goto fail_irq;
+       /* irq bit */
+       if (saa7164_enable_msi(pci_dev, dev)) {
+               dev->msi = true;
+       } else {
+               /* if we have an error (i.e. we don't have an interrupt)
+                        or msi is not enabled - fallback to shared interrupt */
+
+               err = request_irq(pci_dev->irq, saa7164_irq,
+                               IRQF_SHARED, dev->name, dev);
+
+               if (err < 0) {
+                       printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,
+                              pci_dev->irq);
+                       err = -EIO;
+                       goto fail_irq;
+               }
        }
 
        pci_set_drvdata(pci_dev, dev);
@@ -1439,6 +1486,11 @@ static void saa7164_finidev(struct pci_dev *pci_dev)
        /* unregister stuff */
        free_irq(pci_dev->irq, dev);
 
+       if (dev->msi) {
+               pci_disable_msi(pci_dev);
+               dev->msi = false;
+       }
+
        pci_disable_device(pci_dev);
 
        mutex_lock(&devlist);
drivers/media/pci/saa7164/saa7164.h +1 -1
index cd1a07c..75a3f51 100644
@@ -459,6 +459,7 @@ struct saa7164_dev {
        /* Interrupt status and ack registers */
        u32 int_status;
        u32 int_ack;
+       bool msi;
 
        struct cmd                      cmds[SAA_CMD_MAX_MSG_UNITS];
        struct mutex                    lock;
--

1.9.1

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

#13 lore raw

Hi Kyle,

Great to hear you haven't had any problems since applying this patch! I'm looking forward to seeing it in the Linux master branch too.

Version 5 of the patch has been accepted and committed to the media tree by Mauro:
http://git.linuxtv.org/cgit.cgi/media_tree.git/commit/?id=77978089ddc90347644cc057e6b6cd169ac9abd4

I'm guessing it will therefore go in to the main Linux tree with the release of kernel version v4.2? (or it can be submitted as a fix for v4.1 - but I have no idea how a patch is selected on that criteria).

Hopefully someone can confirm or elaborate.

Regards,
Brendan McGrath

Quoted 157 lines · Kyle Sanderson

This has been plaguing users for years (there's a number of threads on the Ubuntu board). I've been using revision 1 of the patch without issue since early February. This is from having to constantly reboot the system to flawless recording. If something has been outstanding from Brendan, please let me know and I'll happily make the requested changes.

Can we please merge this? There are at-least three consumers in this thread alone that have confirmed this fixes the saa7164 driver for the HVR-2250 device.
Kyle.

PS: I can't seem to source out who owns this in the MAINTAINERS file?

Quoted 141 lines · Brendan McGrath

Enhances driver to use an MSI interrupt when available.

Adds the module option 'enable_msi' (type bool) which by default is enabled. Can be set to 'N' to disable.

Fixes (or can reduce the occurrence of) a crash which is most commonly reported when both digital tuners of the saa7164 chip is in use. A reported example can be found here:
http://permalink.gmane.org/gmane.linux.drivers.video-input-infrastructure/83948

---
Changes since v4:
- improved readability by taking on suggestions made by Mauro
- the msi variable in the saa7164_dev structure is now a bool

Thanks Mauro - good suggestions and I think I've taken on board all of them.

  drivers/media/pci/saa7164/saa7164-core.c | 66 ++++++++++++++++++++++++++++----
  drivers/media/pci/saa7164/saa7164.h      |  1 +
  2 files changed, 60 insertions(+), 7 deletions(-)
drivers/media/pci/saa7164/saa7164-core.c +59 -7
index 9cf3c6c..5e4a9f0 100644
@@ -85,6 +85,11 @@ module_param(guard_checking, int, 0644);
  MODULE_PARM_DESC(guard_checking,
         "enable dma sanity checking for buffer overruns");
 
+static bool enable_msi = true;
+module_param(enable_msi, bool, 0444);
+MODULE_PARM_DESC(enable_msi,
+               "enable the use of an msi interrupt if available");
+
  static unsigned int saa7164_devcount;
 
  static DEFINE_MUTEX(devlist);
@@ -1184,6 +1189,39 @@ static int saa7164_thread_function(void *data)
         return 0;
  }
 
+static bool saa7164_enable_msi(struct pci_dev *pci_dev, struct saa7164_dev *dev)
+{
+       int err;
+
+       if (!enable_msi) {
+               printk(KERN_WARNING "%s() MSI disabled by module parameter 'enable_msi'"
+                      , __func__);
+               return false;
+       }
+
+       err = pci_enable_msi(pci_dev);
+
+       if (err) {
+               printk(KERN_ERR "%s() Failed to enable MSI interrupt."
+                       " Falling back to a shared IRQ\n", __func__);
+               return false;
+       }
+
+       /* no error - so request an msi interrupt */
+       err = request_irq(pci_dev->irq, saa7164_irq, 0,
+                                               dev->name, dev);
+
+       if (err) {
+               /* fall back to legacy interrupt */
+               printk(KERN_ERR "%s() Failed to get an MSI interrupt."
+                      " Falling back to a shared IRQ\n", __func__);
+               pci_disable_msi(pci_dev);
+               return false;
+       }
+
+       return true;
+}
+
  static int saa7164_initdev(struct pci_dev *pci_dev,
                            const struct pci_device_id *pci_id)
  {
@@ -1230,13 +1268,22 @@ static int saa7164_initdev(struct pci_dev *pci_dev,
                 goto fail_irq;
         }
 
-       err = request_irq(pci_dev->irq, saa7164_irq,
-               IRQF_SHARED, dev->name, dev);
-       if (err < 0) {
-               printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,
-                       pci_dev->irq);
-               err = -EIO;
-               goto fail_irq;
+       /* irq bit */
+       if (saa7164_enable_msi(pci_dev, dev)) {
+               dev->msi = true;
+       } else {
+               /* if we have an error (i.e. we don't have an interrupt)
+                        or msi is not enabled - fallback to shared interrupt */
+
+               err = request_irq(pci_dev->irq, saa7164_irq,
+                               IRQF_SHARED, dev->name, dev);
+
+               if (err < 0) {
+                       printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,
+                              pci_dev->irq);
+                       err = -EIO;
+                       goto fail_irq;
+               }
         }
 
         pci_set_drvdata(pci_dev, dev);
@@ -1439,6 +1486,11 @@ static void saa7164_finidev(struct pci_dev *pci_dev)
         /* unregister stuff */
         free_irq(pci_dev->irq, dev);
 
+       if (dev->msi) {
+               pci_disable_msi(pci_dev);
+               dev->msi = false;
+       }
+
         pci_disable_device(pci_dev);
 
         mutex_lock(&devlist);
drivers/media/pci/saa7164/saa7164.h +1 -1
index cd1a07c..75a3f51 100644
@@ -459,6 +459,7 @@ struct saa7164_dev {
         /* Interrupt status and ack registers */
         u32 int_status;
         u32 int_ack;
+       bool msi;
 
         struct cmd                      cmds[SAA_CMD_MAX_MSG_UNITS];
         struct mutex                    lock;
--

1.9.1

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