Re: [PATCH] PCI: export MSI mode using attributes, not kobjects

[Veaceslav Falico <vfalico@redhat.com>]

On Wed, Nov 27, 2013 at 10:46:52AM -0800, Greg Kroah-Hartman wrote:
>From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>
>The PCI MSI sysfs code is a mess with kobjects for things that don't
>really need to be kobjects.  This patch creates attributes dynamically
>for the MSI interrupts instead of using kobjects.
>
>Note, this removes a directory from the current MSI interrupt sysfs
>code:
>
>old MSI kobjects:
>pci_device
>   └── msi_irqs
>       └── 40
>           └── mode
>
>new MSI attributes:
>pci_device
>   └── msi_irqs
>       └── 40
>
>As there was only one file "mode" with the kobject model, the interrupt
>number is now a file that returns the "mode" of the interrupt (msi vs.
>msix).
>
>Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Works like a charm for me.

FWIW,

Reviewed-by: Veaceslav Falico <vfalico@redhat.com>

>
>---
>
>Neil, can you test this patch to see if irqbalance works as-is or not?
>
>thanks,
>
>greg k-h
>
> drivers/pci/msi.c   |  152 +++++++++++++++++++++++++---------------------------
> include/linux/pci.h |    2
> 2 files changed, 76 insertions(+), 78 deletions(-)
>
>--- a/drivers/pci/msi.c
>+++ b/drivers/pci/msi.c
>@@ -363,6 +363,9 @@ void write_msi_msg(unsigned int irq, str
> static void free_msi_irqs(struct pci_dev *dev)
> {
> 	struct msi_desc *entry, *tmp;
>+	struct attribute **msi_attrs;
>+	struct device_attribute *dev_attr;
>+	int count = 0;
>
> 	list_for_each_entry(entry, &dev->msi_list, list) {
> 		int i, nvec;
>@@ -398,6 +401,22 @@ static void free_msi_irqs(struct pci_dev
> 		list_del(&entry->list);
> 		kfree(entry);
> 	}
>+
>+	if (dev->msi_irq_groups) {
>+		sysfs_remove_groups(&dev->dev.kobj, dev->msi_irq_groups);
>+		msi_attrs = dev->msi_irq_groups[0]->attrs;
>+		list_for_each_entry(entry, &dev->msi_list, list) {
>+			dev_attr = container_of(msi_attrs[count],
>+						struct device_attribute, attr);
>+			kfree(dev_attr->attr.name);
>+			kfree(dev_attr);
>+			++count;
>+		}
>+		kfree(msi_attrs);
>+		kfree(dev->msi_irq_groups[0]);
>+		kfree(dev->msi_irq_groups);
>+		dev->msi_irq_groups = NULL;
>+	}
> }
>
> static struct msi_desc *alloc_msi_entry(struct pci_dev *dev)
>@@ -471,96 +490,79 @@ void pci_restore_msi_state(struct pci_de
> }
> EXPORT_SYMBOL_GPL(pci_restore_msi_state);
>
>-
>-#define to_msi_attr(obj) container_of(obj, struct msi_attribute, attr)
>-#define to_msi_desc(obj) container_of(obj, struct msi_desc, kobj)
>-
>-struct msi_attribute {
>-	struct attribute        attr;
>-	ssize_t (*show)(struct msi_desc *entry, struct msi_attribute *attr,
>-			char *buf);
>-	ssize_t (*store)(struct msi_desc *entry, struct msi_attribute *attr,
>-			 const char *buf, size_t count);
>-};
>-
>-static ssize_t show_msi_mode(struct msi_desc *entry, struct msi_attribute *atr,
>+static ssize_t msi_mode_show(struct device *dev, struct device_attribute *attr,
> 			     char *buf)
> {
>-	return sprintf(buf, "%s\n", entry->msi_attrib.is_msix ? "msix" : "msi");
>-}
>-
>-static ssize_t msi_irq_attr_show(struct kobject *kobj,
>-				 struct attribute *attr, char *buf)
>-{
>-	struct msi_attribute *attribute = to_msi_attr(attr);
>-	struct msi_desc *entry = to_msi_desc(kobj);
>-
>-	if (!attribute->show)
>-		return -EIO;
>-
>-	return attribute->show(entry, attribute, buf);
>-}
>-
>-static const struct sysfs_ops msi_irq_sysfs_ops = {
>-	.show = msi_irq_attr_show,
>-};
>-
>-static struct msi_attribute mode_attribute =
>-	__ATTR(mode, S_IRUGO, show_msi_mode, NULL);
>-
>+	struct pci_dev *pdev = to_pci_dev(dev);
>+	struct msi_desc *entry;
>+	unsigned long irq;
>+	int retval;
>
>-static struct attribute *msi_irq_default_attrs[] = {
>-	&mode_attribute.attr,
>-	NULL
>-};
>+	retval = kstrtoul(attr->attr.name, 10, &irq);
>+	if (retval)
>+		return retval;
>
>-static void msi_kobj_release(struct kobject *kobj)
>-{
>-	struct msi_desc *entry = to_msi_desc(kobj);
>-
>-	pci_dev_put(entry->dev);
>+	list_for_each_entry(entry, &pdev->msi_list, list) {
>+		if (entry->irq == irq) {
>+			return sprintf(buf, "%s\n",
>+				       entry->msi_attrib.is_msix ? "msix" : "msi");
>+		}
>+	}
>+	return -ENODEV;
> }
>
>-static struct kobj_type msi_irq_ktype = {
>-	.release = msi_kobj_release,
>-	.sysfs_ops = &msi_irq_sysfs_ops,
>-	.default_attrs = msi_irq_default_attrs,
>-};
>-
> static int populate_msi_sysfs(struct pci_dev *pdev)
> {
>+	struct attribute **msi_attrs;
>+	struct device_attribute *msi_dev_attr;
>+	struct attribute_group *msi_irq_group;
>+	const struct attribute_group **msi_irq_groups;
> 	struct msi_desc *entry;
>-	struct kobject *kobj;
> 	int ret;
>+	int num_msi = 0;
> 	int count = 0;
>
>-	pdev->msi_kset = kset_create_and_add("msi_irqs", NULL, &pdev->dev.kobj);
>-	if (!pdev->msi_kset)
>-		return -ENOMEM;
>-
>+	/* Determine how many msi entries we have */
> 	list_for_each_entry(entry, &pdev->msi_list, list) {
>-		kobj = &entry->kobj;
>-		kobj->kset = pdev->msi_kset;
>-		pci_dev_get(pdev);
>-		ret = kobject_init_and_add(kobj, &msi_irq_ktype, NULL,
>-				     "%u", entry->irq);
>-		if (ret)
>-			goto out_unroll;
>-
>-		count++;
>+		++num_msi;
> 	}
>+	if (!num_msi)
>+		return 0;
>
>-	return 0;
>-
>-out_unroll:
>+	/* Dynamically create the MSI attributes for the PCI device */
>+	msi_attrs = kzalloc(sizeof(void *) * (num_msi + 1), GFP_KERNEL);
>+	if (!msi_attrs)
>+		return -ENOMEM;
> 	list_for_each_entry(entry, &pdev->msi_list, list) {
>-		if (!count)
>-			break;
>-		kobject_del(&entry->kobj);
>-		kobject_put(&entry->kobj);
>-		count--;
>+		char *name = kmalloc(20, GFP_KERNEL);
>+		msi_dev_attr = kzalloc(sizeof(*msi_dev_attr), GFP_KERNEL);
>+		if (!msi_dev_attr)
>+			return -ENOMEM;
>+		sprintf(name, "%d", entry->irq);
>+		msi_dev_attr->attr.name = name;
>+		msi_dev_attr->attr.mode = S_IRUGO;
>+		msi_dev_attr->show = msi_mode_show;
>+		msi_attrs[count] = &msi_dev_attr->attr;
>+		++count;
> 	}
>-	return ret;
>+
>+	msi_irq_group = kzalloc(sizeof(*msi_irq_group), GFP_KERNEL);
>+	if (!msi_irq_group)
>+		return -ENOMEM;
>+	msi_irq_group->name = "msi_irqs";
>+	msi_irq_group->attrs = msi_attrs;
>+
>+	msi_irq_groups = kzalloc(sizeof(void *) * 2, GFP_KERNEL);
>+	if (!msi_irq_groups)
>+		return -ENOMEM;
>+	msi_irq_groups[0] = msi_irq_group;
>+
>+	ret = sysfs_create_groups(&pdev->dev.kobj, msi_irq_groups);
>+	if (ret)
>+		return ret;
>+	pdev->msi_irq_groups = msi_irq_groups;
>+
>+	return 0;
> }
>
> /**
>@@ -925,8 +927,6 @@ void pci_disable_msi(struct pci_dev *dev
>
> 	pci_msi_shutdown(dev);
> 	free_msi_irqs(dev);
>-	kset_unregister(dev->msi_kset);
>-	dev->msi_kset = NULL;
> }
> EXPORT_SYMBOL(pci_disable_msi);
>
>@@ -1023,8 +1023,6 @@ void pci_disable_msix(struct pci_dev *de
>
> 	pci_msix_shutdown(dev);
> 	free_msi_irqs(dev);
>-	kset_unregister(dev->msi_kset);
>-	dev->msi_kset = NULL;
> }
> EXPORT_SYMBOL(pci_disable_msix);
>
>--- a/include/linux/pci.h
>+++ b/include/linux/pci.h
>@@ -351,7 +351,7 @@ struct pci_dev {
> 	struct bin_attribute *res_attr_wc[DEVICE_COUNT_RESOURCE]; /* sysfs file for WC mapping of resources */
> #ifdef CONFIG_PCI_MSI
> 	struct list_head msi_list;
>-	struct kset *msi_kset;
>+	const struct attribute_group **msi_irq_groups;
> #endif
> 	struct pci_vpd *vpd;
> #ifdef CONFIG_PCI_ATS