[mirror_ubuntu-bionic-kernel.git] / drivers / xen / xen-pciback / pciback_ops.c

// SPDX-License-Identifier: GPL-2.0
/*
 * PCI Backend Operations - respond to PCI requests from Frontend
 *
 *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
 */

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#include <linux/moduleparam.h>
#include <linux/wait.h>
#include <linux/bitops.h>
#include <xen/events.h>
#include <linux/sched.h>
#include "pciback.h"

int verbose_request;
module_param(verbose_request, int, 0644);

static irqreturn_t xen_pcibk_guest_interrupt(int irq, void *dev_id);

/* Ensure a device is has the fake IRQ handler "turned on/off" and is
 * ready to be exported. This MUST be run after xen_pcibk_reset_device
 * which does the actual PCI device enable/disable.
 */
static void xen_pcibk_control_isr(struct pci_dev *dev, int reset)
{
	struct xen_pcibk_dev_data *dev_data;
	int rc;
	int enable = 0;

	dev_data = pci_get_drvdata(dev);
	if (!dev_data)
		return;

	/* We don't deal with bridges */
	if (dev->hdr_type != PCI_HEADER_TYPE_NORMAL)
		return;

	if (reset) {
		dev_data->enable_intx = 0;
		dev_data->ack_intr = 0;
	}
	enable =  dev_data->enable_intx;

	/* Asked to disable, but ISR isn't runnig */
	if (!enable && !dev_data->isr_on)
		return;

	/* Squirrel away the IRQs in the dev_data. We need this
	 * b/c when device transitions to MSI, the dev->irq is
	 * overwritten with the MSI vector.
	 */
	if (enable)
		dev_data->irq = dev->irq;

	/*
	 * SR-IOV devices in all use MSI-X and have no legacy
	 * interrupts, so inhibit creating a fake IRQ handler for them.
	 */
	if (dev_data->irq == 0)
		goto out;

	dev_dbg(&dev->dev, "%s: #%d %s %s%s %s-> %s\n",
		dev_data->irq_name,
		dev_data->irq,
		pci_is_enabled(dev) ? "on" : "off",
		dev->msi_enabled ? "MSI" : "",
		dev->msix_enabled ? "MSI/X" : "",
		dev_data->isr_on ? "enable" : "disable",
		enable ? "enable" : "disable");

	if (enable) {
		/*
		 * The MSI or MSI-X should not have an IRQ handler. Otherwise
		 * if the guest terminates we BUG_ON in free_msi_irqs.
		 */
		if (dev->msi_enabled || dev->msix_enabled)
			goto out;

		rc = request_irq(dev_data->irq,
				xen_pcibk_guest_interrupt, IRQF_SHARED,
				dev_data->irq_name, dev);
		if (rc) {
			dev_err(&dev->dev, "%s: failed to install fake IRQ " \
				"handler for IRQ %d! (rc:%d)\n",
				dev_data->irq_name, dev_data->irq, rc);
			goto out;
		}
	} else {
		free_irq(dev_data->irq, dev);
		dev_data->irq = 0;
	}
	dev_data->isr_on = enable;
	dev_data->ack_intr = enable;
out:
	dev_dbg(&dev->dev, "%s: #%d %s %s%s %s\n",
		dev_data->irq_name,
		dev_data->irq,
		pci_is_enabled(dev) ? "on" : "off",
		dev->msi_enabled ? "MSI" : "",
		dev->msix_enabled ? "MSI/X" : "",
		enable ? (dev_data->isr_on ? "enabled" : "failed to enable") :
			(dev_data->isr_on ? "failed to disable" : "disabled"));
}

/* Ensure a device is "turned off" and ready to be exported.
 * (Also see xen_pcibk_config_reset to ensure virtual configuration space is
 * ready to be re-exported)
 */
void xen_pcibk_reset_device(struct pci_dev *dev)
{
	u16 cmd;

	xen_pcibk_control_isr(dev, 1 /* reset device */);

	/* Disable devices (but not bridges) */
	if (dev->hdr_type == PCI_HEADER_TYPE_NORMAL) {
#ifdef CONFIG_PCI_MSI
		/* The guest could have been abruptly killed without
		 * disabling MSI/MSI-X interrupts.*/
		if (dev->msix_enabled)
			pci_disable_msix(dev);
		if (dev->msi_enabled)
			pci_disable_msi(dev);
#endif
		if (pci_is_enabled(dev))
			pci_disable_device(dev);

		pci_write_config_word(dev, PCI_COMMAND, 0);

		dev->is_busmaster = 0;
	} else {
		pci_read_config_word(dev, PCI_COMMAND, &cmd);
		if (cmd & (PCI_COMMAND_INVALIDATE)) {
			cmd &= ~(PCI_COMMAND_INVALIDATE);
			pci_write_config_word(dev, PCI_COMMAND, cmd);

			dev->is_busmaster = 0;
		}
	}
}

#ifdef CONFIG_PCI_MSI
static
int xen_pcibk_enable_msi(struct xen_pcibk_device *pdev,
			 struct pci_dev *dev, struct xen_pci_op *op)
{
	struct xen_pcibk_dev_data *dev_data;
	int status;

	if (unlikely(verbose_request))
		printk(KERN_DEBUG DRV_NAME ": %s: enable MSI\n", pci_name(dev));

	if (dev->msi_enabled)
		status = -EALREADY;
	else if (dev->msix_enabled)
		status = -ENXIO;
	else
		status = pci_enable_msi(dev);

	if (status) {
		pr_warn_ratelimited("%s: error enabling MSI for guest %u: err %d\n",
				    pci_name(dev), pdev->xdev->otherend_id,
				    status);
		op->value = 0;
		return XEN_PCI_ERR_op_failed;
	}

	/* The value the guest needs is actually the IDT vector, not the
	 * the local domain's IRQ number. */

	op->value = dev->irq ? xen_pirq_from_irq(dev->irq) : 0;
	if (unlikely(verbose_request))
		printk(KERN_DEBUG DRV_NAME ": %s: MSI: %d\n", pci_name(dev),
			op->value);

	dev_data = pci_get_drvdata(dev);
	if (dev_data)
		dev_data->ack_intr = 0;

	return 0;
}

static
int xen_pcibk_disable_msi(struct xen_pcibk_device *pdev,
			  struct pci_dev *dev, struct xen_pci_op *op)
{
	if (unlikely(verbose_request))
		printk(KERN_DEBUG DRV_NAME ": %s: disable MSI\n",
		       pci_name(dev));

	if (dev->msi_enabled) {
		struct xen_pcibk_dev_data *dev_data;

		pci_disable_msi(dev);

		dev_data = pci_get_drvdata(dev);
		if (dev_data)
			dev_data->ack_intr = 1;
	}
	op->value = dev->irq ? xen_pirq_from_irq(dev->irq) : 0;
	if (unlikely(verbose_request))
		printk(KERN_DEBUG DRV_NAME ": %s: MSI: %d\n", pci_name(dev),
			op->value);
	return 0;
}

static
int xen_pcibk_enable_msix(struct xen_pcibk_device *pdev,
			  struct pci_dev *dev, struct xen_pci_op *op)
{
	struct xen_pcibk_dev_data *dev_data;
	int i, result;
	struct msix_entry *entries;
	u16 cmd;

	if (unlikely(verbose_request))
		printk(KERN_DEBUG DRV_NAME ": %s: enable MSI-X\n",
		       pci_name(dev));

	if (op->value > SH_INFO_MAX_VEC)
		return -EINVAL;

	if (dev->msix_enabled)
		return -EALREADY;

	/*
	 * PCI_COMMAND_MEMORY must be enabled, otherwise we may not be able
	 * to access the BARs where the MSI-X entries reside.
	 * But VF devices are unique in which the PF needs to be checked.
	 */
	pci_read_config_word(pci_physfn(dev), PCI_COMMAND, &cmd);
	if (dev->msi_enabled || !(cmd & PCI_COMMAND_MEMORY))
		return -ENXIO;

	entries = kmalloc(op->value * sizeof(*entries), GFP_KERNEL);
	if (entries == NULL)
		return -ENOMEM;

	for (i = 0; i < op->value; i++) {
		entries[i].entry = op->msix_entries[i].entry;
		entries[i].vector = op->msix_entries[i].vector;
	}

	result = pci_enable_msix_exact(dev, entries, op->value);
	if (result == 0) {
		for (i = 0; i < op->value; i++) {
			op->msix_entries[i].entry = entries[i].entry;
			if (entries[i].vector) {
				op->msix_entries[i].vector =
					xen_pirq_from_irq(entries[i].vector);
				if (unlikely(verbose_request))
					printk(KERN_DEBUG DRV_NAME ": %s: " \
						"MSI-X[%d]: %d\n",
						pci_name(dev), i,
						op->msix_entries[i].vector);
			}
		}
	} else
		pr_warn_ratelimited("%s: error enabling MSI-X for guest %u: err %d!\n",
				    pci_name(dev), pdev->xdev->otherend_id,
				    result);
	kfree(entries);

	op->value = result;
	dev_data = pci_get_drvdata(dev);
	if (dev_data)
		dev_data->ack_intr = 0;

	return result > 0 ? 0 : result;
}

static
int xen_pcibk_disable_msix(struct xen_pcibk_device *pdev,
			   struct pci_dev *dev, struct xen_pci_op *op)
{
	if (unlikely(verbose_request))
		printk(KERN_DEBUG DRV_NAME ": %s: disable MSI-X\n",
			pci_name(dev));

	if (dev->msix_enabled) {
		struct xen_pcibk_dev_data *dev_data;

		pci_disable_msix(dev);

		dev_data = pci_get_drvdata(dev);
		if (dev_data)
			dev_data->ack_intr = 1;
	}
	/*
	 * SR-IOV devices (which don't have any legacy IRQ) have
	 * an undefined IRQ value of zero.
	 */
	op->value = dev->irq ? xen_pirq_from_irq(dev->irq) : 0;
	if (unlikely(verbose_request))
		printk(KERN_DEBUG DRV_NAME ": %s: MSI-X: %d\n",
		       pci_name(dev), op->value);
	return 0;
}
#endif
/*
* Now the same evtchn is used for both pcifront conf_read_write request
* as well as pcie aer front end ack. We use a new work_queue to schedule
* xen_pcibk conf_read_write service for avoiding confict with aer_core
* do_recovery job which also use the system default work_queue
*/
void xen_pcibk_test_and_schedule_op(struct xen_pcibk_device *pdev)
{
	/* Check that frontend is requesting an operation and that we are not
	 * already processing a request */
	if (test_bit(_XEN_PCIF_active, (unsigned long *)&pdev->sh_info->flags)
	    && !test_and_set_bit(_PDEVF_op_active, &pdev->flags)) {
		schedule_work(&pdev->op_work);
	}
	/*_XEN_PCIB_active should have been cleared by pcifront. And also make
	sure xen_pcibk is waiting for ack by checking _PCIB_op_pending*/
	if (!test_bit(_XEN_PCIB_active, (unsigned long *)&pdev->sh_info->flags)
	    && test_bit(_PCIB_op_pending, &pdev->flags)) {
		wake_up(&xen_pcibk_aer_wait_queue);
	}
}

/* Performing the configuration space reads/writes must not be done in atomic
 * context because some of the pci_* functions can sleep (mostly due to ACPI
 * use of semaphores). This function is intended to be called from a work
 * queue in process context taking a struct xen_pcibk_device as a parameter */

void xen_pcibk_do_op(struct work_struct *data)
{
	struct xen_pcibk_device *pdev =
		container_of(data, struct xen_pcibk_device, op_work);
	struct pci_dev *dev;
	struct xen_pcibk_dev_data *dev_data = NULL;
	struct xen_pci_op *op = &pdev->op;
	int test_intx = 0;
#ifdef CONFIG_PCI_MSI
	unsigned int nr = 0;
#endif

	*op = pdev->sh_info->op;
	barrier();
	dev = xen_pcibk_get_pci_dev(pdev, op->domain, op->bus, op->devfn);

	if (dev == NULL)
		op->err = XEN_PCI_ERR_dev_not_found;
	else {
		dev_data = pci_get_drvdata(dev);
		if (dev_data)
			test_intx = dev_data->enable_intx;
		switch (op->cmd) {
		case XEN_PCI_OP_conf_read:
			op->err = xen_pcibk_config_read(dev,
				  op->offset, op->size, &op->value);
			break;
		case XEN_PCI_OP_conf_write:
			op->err = xen_pcibk_config_write(dev,
				  op->offset, op->size,	op->value);
			break;
#ifdef CONFIG_PCI_MSI
		case XEN_PCI_OP_enable_msi:
			op->err = xen_pcibk_enable_msi(pdev, dev, op);
			break;
		case XEN_PCI_OP_disable_msi:
			op->err = xen_pcibk_disable_msi(pdev, dev, op);
			break;
		case XEN_PCI_OP_enable_msix:
			nr = op->value;
			op->err = xen_pcibk_enable_msix(pdev, dev, op);
			break;
		case XEN_PCI_OP_disable_msix:
			op->err = xen_pcibk_disable_msix(pdev, dev, op);
			break;
#endif
		default:
			op->err = XEN_PCI_ERR_not_implemented;
			break;
		}
	}
	if (!op->err && dev && dev_data) {
		/* Transition detected */
		if ((dev_data->enable_intx != test_intx))
			xen_pcibk_control_isr(dev, 0 /* no reset */);
	}
	pdev->sh_info->op.err = op->err;
	pdev->sh_info->op.value = op->value;
#ifdef CONFIG_PCI_MSI
	if (op->cmd == XEN_PCI_OP_enable_msix && op->err == 0) {
		unsigned int i;

		for (i = 0; i < nr; i++)
			pdev->sh_info->op.msix_entries[i].vector =
				op->msix_entries[i].vector;
	}
#endif
	/* Tell the driver domain that we're done. */
	wmb();
	clear_bit(_XEN_PCIF_active, (unsigned long *)&pdev->sh_info->flags);
	notify_remote_via_irq(pdev->evtchn_irq);

	/* Mark that we're done. */
	smp_mb__before_atomic(); /* /after/ clearing PCIF_active */
	clear_bit(_PDEVF_op_active, &pdev->flags);
	smp_mb__after_atomic(); /* /before/ final check for work */

	/* Check to see if the driver domain tried to start another request in
	 * between clearing _XEN_PCIF_active and clearing _PDEVF_op_active.
	*/
	xen_pcibk_test_and_schedule_op(pdev);
}

irqreturn_t xen_pcibk_handle_event(int irq, void *dev_id)
{
	struct xen_pcibk_device *pdev = dev_id;

	xen_pcibk_test_and_schedule_op(pdev);

	return IRQ_HANDLED;
}
static irqreturn_t xen_pcibk_guest_interrupt(int irq, void *dev_id)
{
	struct pci_dev *dev = (struct pci_dev *)dev_id;
	struct xen_pcibk_dev_data *dev_data = pci_get_drvdata(dev);

	if (dev_data->isr_on && dev_data->ack_intr) {
		dev_data->handled++;
		if ((dev_data->handled % 1000) == 0) {
			if (xen_test_irq_shared(irq)) {
				pr_info("%s IRQ line is not shared "
					"with other domains. Turning ISR off\n",
					 dev_data->irq_name);
				dev_data->ack_intr = 0;
			}
		}
		return IRQ_HANDLED;
	}
	return IRQ_NONE;
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
30edc14b KRW	2	/*
	3	* PCI Backend Operations - respond to PCI requests from Frontend
	4	*
	5	* Author: Ryan Wilson <hap9@epoch.ncsc.mil>
	6	*/
283c0972 JP	7
	8	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
	9
59aa56bf	10	#include <linux/moduleparam.h>
30edc14b KRW	11	#include <linux/wait.h>
	12	#include <linux/bitops.h>
	13	#include <xen/events.h>
	14	#include <linux/sched.h>
	15	#include "pciback.h"
	16
	17	int verbose_request;
	18	module_param(verbose_request, int, 0644);
	19
a92336a1 KRW	20	static irqreturn_t xen_pcibk_guest_interrupt(int irq, void *dev_id);
a92336a1 KRW	21
0513fe9e	22	/* Ensure a device is has the fake IRQ handler "turned on/off" and is
a92336a1	23	* ready to be exported. This MUST be run after xen_pcibk_reset_device
0513fe9e KRW	24	* which does the actual PCI device enable/disable.
0513fe9e KRW	25	*/
a92336a1	26	static void xen_pcibk_control_isr(struct pci_dev *dev, int reset)
0513fe9e	27	{
a92336a1	28	struct xen_pcibk_dev_data *dev_data;
0513fe9e KRW	29	int rc;
	30	int enable = 0;
	31
	32	dev_data = pci_get_drvdata(dev);
	33	if (!dev_data)
	34	return;
	35
	36	/* We don't deal with bridges */
	37	if (dev->hdr_type != PCI_HEADER_TYPE_NORMAL)
	38	return;
	39
	40	if (reset) {
	41	dev_data->enable_intx = 0;
	42	dev_data->ack_intr = 0;
	43	}
	44	enable = dev_data->enable_intx;
	45
	46	/* Asked to disable, but ISR isn't runnig */
	47	if (!enable && !dev_data->isr_on)
	48	return;
	49
	50	/* Squirrel away the IRQs in the dev_data. We need this
	51	* b/c when device transitions to MSI, the dev->irq is
	52	* overwritten with the MSI vector.
	53	*/
	54	if (enable)
	55	dev_data->irq = dev->irq;
	56
e17ab35f KRW	57	/*
	58	* SR-IOV devices in all use MSI-X and have no legacy
	59	* interrupts, so inhibit creating a fake IRQ handler for them.
	60	*/
	61	if (dev_data->irq == 0)
	62	goto out;
	63
0513fe9e KRW	64	dev_dbg(&dev->dev, "%s: #%d %s %s%s %s-> %s\n",
	65	dev_data->irq_name,
	66	dev_data->irq,
	67	pci_is_enabled(dev) ? "on" : "off",
	68	dev->msi_enabled ? "MSI" : "",
	69	dev->msix_enabled ? "MSI/X" : "",
	70	dev_data->isr_on ? "enable" : "disable",
	71	enable ? "enable" : "disable");
	72
	73	if (enable) {
a396f3a2 KRW	74	/*
	75	* The MSI or MSI-X should not have an IRQ handler. Otherwise
	76	* if the guest terminates we BUG_ON in free_msi_irqs.
	77	*/
	78	if (dev->msi_enabled \|\| dev->msix_enabled)
	79	goto out;
	80
0513fe9e	81	rc = request_irq(dev_data->irq,
a92336a1	82	xen_pcibk_guest_interrupt, IRQF_SHARED,
0513fe9e KRW	83	dev_data->irq_name, dev);
	84	if (rc) {
	85	dev_err(&dev->dev, "%s: failed to install fake IRQ " \
	86	"handler for IRQ %d! (rc:%d)\n",
	87	dev_data->irq_name, dev_data->irq, rc);
	88	goto out;
	89	}
	90	} else {
	91	free_irq(dev_data->irq, dev);
	92	dev_data->irq = 0;
	93	}
	94	dev_data->isr_on = enable;
	95	dev_data->ack_intr = enable;
	96	out:
	97	dev_dbg(&dev->dev, "%s: #%d %s %s%s %s\n",
	98	dev_data->irq_name,
	99	dev_data->irq,
	100	pci_is_enabled(dev) ? "on" : "off",
	101	dev->msi_enabled ? "MSI" : "",
	102	dev->msix_enabled ? "MSI/X" : "",
	103	enable ? (dev_data->isr_on ? "enabled" : "failed to enable") :
	104	(dev_data->isr_on ? "failed to disable" : "disabled"));
	105	}
	106
30edc14b	107	/* Ensure a device is "turned off" and ready to be exported.
a92336a1	108	* (Also see xen_pcibk_config_reset to ensure virtual configuration space is
30edc14b KRW	109	* ready to be re-exported)
30edc14b KRW	110	*/
a92336a1	111	void xen_pcibk_reset_device(struct pci_dev *dev)
30edc14b KRW	112	{
	113	u16 cmd;
	114
a92336a1	115	xen_pcibk_control_isr(dev, 1 /* reset device */);
0513fe9e	116
30edc14b KRW	117	/* Disable devices (but not bridges) */
30edc14b KRW	118	if (dev->hdr_type == PCI_HEADER_TYPE_NORMAL) {
a2be65fd KRW	119	#ifdef CONFIG_PCI_MSI
	120	/* The guest could have been abruptly killed without
	121	* disabling MSI/MSI-X interrupts.*/
	122	if (dev->msix_enabled)
	123	pci_disable_msix(dev);
	124	if (dev->msi_enabled)
	125	pci_disable_msi(dev);
	126	#endif
bdc5c181 KRW	127	if (pci_is_enabled(dev))
bdc5c181 KRW	128	pci_disable_device(dev);
30edc14b KRW	129
	130	pci_write_config_word(dev, PCI_COMMAND, 0);
	131
	132	dev->is_busmaster = 0;
	133	} else {
	134	pci_read_config_word(dev, PCI_COMMAND, &cmd);
	135	if (cmd & (PCI_COMMAND_INVALIDATE)) {
	136	cmd &= ~(PCI_COMMAND_INVALIDATE);
	137	pci_write_config_word(dev, PCI_COMMAND, cmd);
	138
	139	dev->is_busmaster = 0;
	140	}
	141	}
	142	}
a92336a1 KRW	143
	144	#ifdef CONFIG_PCI_MSI
	145	static
	146	int xen_pcibk_enable_msi(struct xen_pcibk_device *pdev,
	147	struct pci_dev dev, struct xen_pci_op op)
	148	{
	149	struct xen_pcibk_dev_data *dev_data;
a92336a1 KRW	150	int status;
	151
	152	if (unlikely(verbose_request))
	153	printk(KERN_DEBUG DRV_NAME ": %s: enable MSI\n", pci_name(dev));
	154
56441f3c KRW	155	if (dev->msi_enabled)
	156	status = -EALREADY;
	157	else if (dev->msix_enabled)
	158	status = -ENXIO;
	159	else
	160	status = pci_enable_msi(dev);
a92336a1 KRW	161
a92336a1 KRW	162	if (status) {
283c0972	163	pr_warn_ratelimited("%s: error enabling MSI for guest %u: err %d\n",
51ac8893 JB	164	pci_name(dev), pdev->xdev->otherend_id,
51ac8893 JB	165	status);
a92336a1 KRW	166	op->value = 0;
	167	return XEN_PCI_ERR_op_failed;
	168	}
	169
	170	/* The value the guest needs is actually the IDT vector, not the
	171	* the local domain's IRQ number. */
	172
	173	op->value = dev->irq ? xen_pirq_from_irq(dev->irq) : 0;
	174	if (unlikely(verbose_request))
	175	printk(KERN_DEBUG DRV_NAME ": %s: MSI: %d\n", pci_name(dev),
	176	op->value);
	177
	178	dev_data = pci_get_drvdata(dev);
	179	if (dev_data)
	180	dev_data->ack_intr = 0;
	181
	182	return 0;
	183	}
	184
	185	static
	186	int xen_pcibk_disable_msi(struct xen_pcibk_device *pdev,
	187	struct pci_dev dev, struct xen_pci_op op)
	188	{
a92336a1 KRW	189	if (unlikely(verbose_request))
	190	printk(KERN_DEBUG DRV_NAME ": %s: disable MSI\n",
	191	pci_name(dev));
a92336a1	192
7cfb905b KRW	193	if (dev->msi_enabled) {
	194	struct xen_pcibk_dev_data *dev_data;
	195
	196	pci_disable_msi(dev);
	197
	198	dev_data = pci_get_drvdata(dev);
	199	if (dev_data)
	200	dev_data->ack_intr = 1;
	201	}
a92336a1 KRW	202	op->value = dev->irq ? xen_pirq_from_irq(dev->irq) : 0;
	203	if (unlikely(verbose_request))
	204	printk(KERN_DEBUG DRV_NAME ": %s: MSI: %d\n", pci_name(dev),
	205	op->value);
a92336a1 KRW	206	return 0;
	207	}
	208
	209	static
	210	int xen_pcibk_enable_msix(struct xen_pcibk_device *pdev,
	211	struct pci_dev dev, struct xen_pci_op op)
	212	{
	213	struct xen_pcibk_dev_data *dev_data;
	214	int i, result;
	215	struct msix_entry *entries;
408fb0e5	216	u16 cmd;
a92336a1 KRW	217
	218	if (unlikely(verbose_request))
	219	printk(KERN_DEBUG DRV_NAME ": %s: enable MSI-X\n",
	220	pci_name(dev));
5e0ce145	221
a92336a1 KRW	222	if (op->value > SH_INFO_MAX_VEC)
	223	return -EINVAL;
	224
5e0ce145 KRW	225	if (dev->msix_enabled)
	226	return -EALREADY;
	227
408fb0e5 KRW	228	/*
	229	* PCI_COMMAND_MEMORY must be enabled, otherwise we may not be able
	230	* to access the BARs where the MSI-X entries reside.
8d47065f	231	* But VF devices are unique in which the PF needs to be checked.
408fb0e5	232	*/
8d47065f	233	pci_read_config_word(pci_physfn(dev), PCI_COMMAND, &cmd);
408fb0e5	234	if (dev->msi_enabled \|\| !(cmd & PCI_COMMAND_MEMORY))
5e0ce145 KRW	235	return -ENXIO;
5e0ce145 KRW	236
a92336a1 KRW	237	entries = kmalloc(op->value * sizeof(*entries), GFP_KERNEL);
	238	if (entries == NULL)
	239	return -ENOMEM;
	240
	241	for (i = 0; i < op->value; i++) {
	242	entries[i].entry = op->msix_entries[i].entry;
	243	entries[i].vector = op->msix_entries[i].vector;
	244	}
	245
efdfa3ed	246	result = pci_enable_msix_exact(dev, entries, op->value);
a92336a1 KRW	247	if (result == 0) {
	248	for (i = 0; i < op->value; i++) {
	249	op->msix_entries[i].entry = entries[i].entry;
c0914e61	250	if (entries[i].vector) {
a92336a1 KRW	251	op->msix_entries[i].vector =
	252	xen_pirq_from_irq(entries[i].vector);
	253	if (unlikely(verbose_request))
	254	printk(KERN_DEBUG DRV_NAME ": %s: " \
	255	"MSI-X[%d]: %d\n",
	256	pci_name(dev), i,
	257	op->msix_entries[i].vector);
c0914e61	258	}
a92336a1	259	}
51ac8893	260	} else
283c0972	261	pr_warn_ratelimited("%s: error enabling MSI-X for guest %u: err %d!\n",
51ac8893 JB	262	pci_name(dev), pdev->xdev->otherend_id,
51ac8893 JB	263	result);
a92336a1 KRW	264	kfree(entries);
	265
	266	op->value = result;
	267	dev_data = pci_get_drvdata(dev);
	268	if (dev_data)
	269	dev_data->ack_intr = 0;
	270
0ee46eca	271	return result > 0 ? 0 : result;
a92336a1 KRW	272	}
	273
	274	static
	275	int xen_pcibk_disable_msix(struct xen_pcibk_device *pdev,
	276	struct pci_dev dev, struct xen_pci_op op)
	277	{
a92336a1 KRW	278	if (unlikely(verbose_request))
	279	printk(KERN_DEBUG DRV_NAME ": %s: disable MSI-X\n",
	280	pci_name(dev));
a92336a1	281
7cfb905b KRW	282	if (dev->msix_enabled) {
	283	struct xen_pcibk_dev_data *dev_data;
	284
	285	pci_disable_msix(dev);
	286
	287	dev_data = pci_get_drvdata(dev);
	288	if (dev_data)
	289	dev_data->ack_intr = 1;
	290	}
a92336a1 KRW	291	/*
	292	* SR-IOV devices (which don't have any legacy IRQ) have
	293	* an undefined IRQ value of zero.
	294	*/
	295	op->value = dev->irq ? xen_pirq_from_irq(dev->irq) : 0;
	296	if (unlikely(verbose_request))
7cfb905b KRW	297	printk(KERN_DEBUG DRV_NAME ": %s: MSI-X: %d\n",
7cfb905b KRW	298	pci_name(dev), op->value);
a92336a1 KRW	299	return 0;
	300	}
	301	#endif
30edc14b KRW	302	/*
	303	* Now the same evtchn is used for both pcifront conf_read_write request
	304	* as well as pcie aer front end ack. We use a new work_queue to schedule
a92336a1	305	* xen_pcibk conf_read_write service for avoiding confict with aer_core
30edc14b KRW	306	* do_recovery job which also use the system default work_queue
30edc14b KRW	307	*/
a92336a1	308	void xen_pcibk_test_and_schedule_op(struct xen_pcibk_device *pdev)
30edc14b KRW	309	{
	310	/* Check that frontend is requesting an operation and that we are not
	311	* already processing a request */
	312	if (test_bit(_XEN_PCIF_active, (unsigned long *)&pdev->sh_info->flags)
	313	&& !test_and_set_bit(_PDEVF_op_active, &pdev->flags)) {
429eafe6	314	schedule_work(&pdev->op_work);
30edc14b KRW	315	}
30edc14b KRW	316	/*_XEN_PCIB_active should have been cleared by pcifront. And also make
a92336a1	317	sure xen_pcibk is waiting for ack by checking _PCIB_op_pending*/
30edc14b KRW	318	if (!test_bit(_XEN_PCIB_active, (unsigned long *)&pdev->sh_info->flags)
30edc14b KRW	319	&& test_bit(_PCIB_op_pending, &pdev->flags)) {
a92336a1	320	wake_up(&xen_pcibk_aer_wait_queue);
30edc14b KRW	321	}
	322	}
	323
	324	/* Performing the configuration space reads/writes must not be done in atomic
	325	* context because some of the pci_* functions can sleep (mostly due to ACPI
	326	* use of semaphores). This function is intended to be called from a work
a92336a1	327	* queue in process context taking a struct xen_pcibk_device as a parameter */
30edc14b	328
a92336a1	329	void xen_pcibk_do_op(struct work_struct *data)
30edc14b	330	{
a92336a1 KRW	331	struct xen_pcibk_device *pdev =
a92336a1 KRW	332	container_of(data, struct xen_pcibk_device, op_work);
30edc14b	333	struct pci_dev *dev;
a92336a1	334	struct xen_pcibk_dev_data *dev_data = NULL;
8135cf8b	335	struct xen_pci_op *op = &pdev->op;
0513fe9e	336	int test_intx = 0;
d159457b KRW	337	#ifdef CONFIG_PCI_MSI
	338	unsigned int nr = 0;
	339	#endif
30edc14b	340
8135cf8b KRW	341	*op = pdev->sh_info->op;
8135cf8b KRW	342	barrier();
a92336a1	343	dev = xen_pcibk_get_pci_dev(pdev, op->domain, op->bus, op->devfn);
30edc14b KRW	344
	345	if (dev == NULL)
	346	op->err = XEN_PCI_ERR_dev_not_found;
	347	else {
0513fe9e KRW	348	dev_data = pci_get_drvdata(dev);
	349	if (dev_data)
	350	test_intx = dev_data->enable_intx;
30edc14b KRW	351	switch (op->cmd) {
30edc14b KRW	352	case XEN_PCI_OP_conf_read:
a92336a1	353	op->err = xen_pcibk_config_read(dev,
30edc14b KRW	354	op->offset, op->size, &op->value);
	355	break;
	356	case XEN_PCI_OP_conf_write:
a92336a1	357	op->err = xen_pcibk_config_write(dev,
30edc14b KRW	358	op->offset, op->size, op->value);
	359	break;
	360	#ifdef CONFIG_PCI_MSI
	361	case XEN_PCI_OP_enable_msi:
a92336a1	362	op->err = xen_pcibk_enable_msi(pdev, dev, op);
30edc14b KRW	363	break;
30edc14b KRW	364	case XEN_PCI_OP_disable_msi:
a92336a1	365	op->err = xen_pcibk_disable_msi(pdev, dev, op);
30edc14b KRW	366	break;
30edc14b KRW	367	case XEN_PCI_OP_enable_msix:
d159457b	368	nr = op->value;
a92336a1	369	op->err = xen_pcibk_enable_msix(pdev, dev, op);
30edc14b KRW	370	break;
30edc14b KRW	371	case XEN_PCI_OP_disable_msix:
a92336a1	372	op->err = xen_pcibk_disable_msix(pdev, dev, op);
30edc14b KRW	373	break;
	374	#endif
	375	default:
	376	op->err = XEN_PCI_ERR_not_implemented;
	377	break;
	378	}
	379	}
0513fe9e KRW	380	if (!op->err && dev && dev_data) {
	381	/* Transition detected */
	382	if ((dev_data->enable_intx != test_intx))
a92336a1	383	xen_pcibk_control_isr(dev, 0 /* no reset */);
0513fe9e	384	}
8135cf8b KRW	385	pdev->sh_info->op.err = op->err;
	386	pdev->sh_info->op.value = op->value;
	387	#ifdef CONFIG_PCI_MSI
	388	if (op->cmd == XEN_PCI_OP_enable_msix && op->err == 0) {
	389	unsigned int i;
	390
d159457b	391	for (i = 0; i < nr; i++)
8135cf8b KRW	392	pdev->sh_info->op.msix_entries[i].vector =
	393	op->msix_entries[i].vector;
	394	}
	395	#endif
30edc14b KRW	396	/* Tell the driver domain that we're done. */
	397	wmb();
	398	clear_bit(_XEN_PCIF_active, (unsigned long *)&pdev->sh_info->flags);
	399	notify_remote_via_irq(pdev->evtchn_irq);
	400
	401	/* Mark that we're done. */
4e857c58	402	smp_mb__before_atomic(); /* /after/ clearing PCIF_active */
30edc14b	403	clear_bit(_PDEVF_op_active, &pdev->flags);
4e857c58	404	smp_mb__after_atomic(); /* /before/ final check for work */
30edc14b KRW	405
	406	/* Check to see if the driver domain tried to start another request in
	407	* between clearing _XEN_PCIF_active and clearing _PDEVF_op_active.
	408	*/
a92336a1	409	xen_pcibk_test_and_schedule_op(pdev);
30edc14b KRW	410	}
30edc14b KRW	411
a92336a1	412	irqreturn_t xen_pcibk_handle_event(int irq, void *dev_id)
30edc14b	413	{
a92336a1	414	struct xen_pcibk_device *pdev = dev_id;
30edc14b	415
a92336a1	416	xen_pcibk_test_and_schedule_op(pdev);
30edc14b KRW	417
	418	return IRQ_HANDLED;
	419	}
a92336a1	420	static irqreturn_t xen_pcibk_guest_interrupt(int irq, void *dev_id)
0513fe9e KRW	421	{
0513fe9e KRW	422	struct pci_dev dev = (struct pci_dev )dev_id;
a92336a1	423	struct xen_pcibk_dev_data *dev_data = pci_get_drvdata(dev);
0513fe9e KRW	424
	425	if (dev_data->isr_on && dev_data->ack_intr) {
	426	dev_data->handled++;
	427	if ((dev_data->handled % 1000) == 0) {
	428	if (xen_test_irq_shared(irq)) {
283c0972	429	pr_info("%s IRQ line is not shared "
0513fe9e KRW	430	"with other domains. Turning ISR off\n",
	431	dev_data->irq_name);
	432	dev_data->ack_intr = 0;
	433	}
	434	}
	435	return IRQ_HANDLED;
	436	}
	437	return IRQ_NONE;
	438	}