a396f3a210
Otherwise an guest can subvert the generic MSI code to trigger
an BUG_ON condition during MSI interrupt freeing:
for (i = 0; i < entry->nvec_used; i++)
BUG_ON(irq_has_action(entry->irq + i));
Xen PCI backed installs an IRQ handler (request_irq) for
the dev->irq whenever the guest writes PCI_COMMAND_MEMORY
(or PCI_COMMAND_IO) to the PCI_COMMAND register. This is
done in case the device has legacy interrupts the GSI line
is shared by the backend devices.
To subvert the backend the guest needs to make the backend
to change the dev->irq from the GSI to the MSI interrupt line,
make the backend allocate an interrupt handler, and then command
the backend to free the MSI interrupt and hit the BUG_ON.
Since the backend only calls 'request_irq' when the guest
writes to the PCI_COMMAND register the guest needs to call
XEN_PCI_OP_enable_msi before any other operation. This will
cause the generic MSI code to setup an MSI entry and
populate dev->irq with the new PIRQ value.
Then the guest can write to PCI_COMMAND PCI_COMMAND_MEMORY
and cause the backend to setup an IRQ handler for dev->irq
(which instead of the GSI value has the MSI pirq). See
'xen_pcibk_control_isr'.
Then the guest disables the MSI: XEN_PCI_OP_disable_msi
which ends up triggering the BUG_ON condition in 'free_msi_irqs'
as there is an IRQ handler for the entry->irq (dev->irq).
Note that this cannot be done using MSI-X as the generic
code does not over-write dev->irq with the MSI-X PIRQ values.
The patch inhibits setting up the IRQ handler if MSI or
MSI-X (for symmetry reasons) code had been called successfully.
P.S.
Xen PCIBack when it sets up the device for the guest consumption
ends up writting 0 to the PCI_COMMAND (see xen_pcibk_reset_device).
XSA-120 addendum patch removed that - however when upstreaming said
addendum we found that it caused issues with qemu upstream. That
has now been fixed in qemu upstream.
This is part of XSA-157
CC: stable@vger.kernel.org
Reviewed-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
420 lines
11 KiB
C
420 lines
11 KiB
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)
|
|
{
|
|
struct xen_pcibk_dev_data *dev_data;
|
|
|
|
if (unlikely(verbose_request))
|
|
printk(KERN_DEBUG DRV_NAME ": %s: disable MSI\n",
|
|
pci_name(dev));
|
|
pci_disable_msi(dev);
|
|
|
|
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 = 1;
|
|
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;
|
|
|
|
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;
|
|
|
|
if (dev->msi_enabled)
|
|
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)
|
|
{
|
|
struct xen_pcibk_dev_data *dev_data;
|
|
if (unlikely(verbose_request))
|
|
printk(KERN_DEBUG DRV_NAME ": %s: disable MSI-X\n",
|
|
pci_name(dev));
|
|
pci_disable_msix(dev);
|
|
|
|
/*
|
|
* 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);
|
|
dev_data = pci_get_drvdata(dev);
|
|
if (dev_data)
|
|
dev_data->ack_intr = 1;
|
|
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;
|
|
}
|