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

/*
 * 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/module.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.
	 */
	pci_read_config_word(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)) {
		queue_work(xen_pcibk_wq, &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;

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