xen: privcmd: Add support for ioeventfd

Virtio guests send VIRTIO_MMIO_QUEUE_NOTIFY notification when they need
to notify the backend of an update to the status of the virtqueue. The
backend or another entity, polls the MMIO address for updates to know
when the notification is sent.

It works well if the backend does this polling by itself. But as we move
towards generic backend implementations, we end up implementing this in
a separate user-space program.

Generally, the Virtio backends are implemented to work with the Eventfd
based mechanism. In order to make such backends work with Xen, another
software layer needs to do the polling and send an event via eventfd to
the backend once the notification from guest is received. This results
in an extra context switch.

This is not a new problem in Linux though. It is present with other
hypervisors like KVM, etc. as well. The generic solution implemented in
the kernel for them is to provide an IOCTL call to pass the address to
poll and eventfd, which lets the kernel take care of polling and raise
an event on the eventfd, instead of handling this in user space (which
involves an extra context switch).

This patch adds similar support for xen.

Inspired by existing implementations for KVM, etc..

This also copies ioreq.h header file (only struct ioreq and related
macros) from Xen's source tree (Top commit 5d84f07fe6bf ("xen/pci: drop
remaining uses of bool_t")).
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/b20d83efba6453037d0c099912813c79c81f7714.1697439990.git.viresh.kumar@linaro.orgSigned-off-by: Juergen Gross <jgross@suse.com>

drivers/xen/Kconfig

@@ -269,12 +269,12 @@ config XEN_PRIVCMD
 	  disaggregated Xen setups this driver might be needed for other
 	  domains, too.
 
-config XEN_PRIVCMD_IRQFD
-	bool "Xen irqfd support"
+config XEN_PRIVCMD_EVENTFD
+	bool "Xen Ioeventfd and irqfd support"
 	depends on XEN_PRIVCMD && XEN_VIRTIO && EVENTFD
 	help
-	  Using the irqfd mechanism a virtio backend running in a daemon can
-	  speed up interrupt injection into a guest.
+	  Using the ioeventfd / irqfd mechanism a virtio backend running in a
+	  daemon can speed up interrupt delivery from / to a guest.
 
 config XEN_ACPI_PROCESSOR
 	tristate "Xen ACPI processor"

include/uapi/xen/privcmd.h

@@ -110,6 +110,22 @@ struct privcmd_irqfd {
 	__u8 pad[2];
 };
 
+/* For privcmd_ioeventfd::flags */
+#define PRIVCMD_IOEVENTFD_FLAG_DEASSIGN (1 << 0)
+
+struct privcmd_ioeventfd {
+	__u64 ioreq;
+	__u64 ports;
+	__u64 addr;
+	__u32 addr_len;
+	__u32 event_fd;
+	__u32 vcpus;
+	__u32 vq;
+	__u32 flags;
+	domid_t dom;
+	__u8 pad[2];
+};
+
 /*
  * @cmd: IOCTL_PRIVCMD_HYPERCALL
  * @arg: &privcmd_hypercall_t
@@ -139,5 +155,7 @@ struct privcmd_irqfd {
 	_IOC(_IOC_NONE, 'P', 7, sizeof(struct privcmd_mmap_resource))
 #define IOCTL_PRIVCMD_IRQFD					\
 	_IOW('P', 8, struct privcmd_irqfd)
+#define IOCTL_PRIVCMD_IOEVENTFD					\
+	_IOW('P', 9, struct privcmd_ioeventfd)
 
 #endif /* __LINUX_PUBLIC_PRIVCMD_H__ */

include/xen/interface/hvm/ioreq.h

+/* SPDX-License-Identifier: MIT */
+/*
+ * ioreq.h: I/O request definitions for device models
+ * Copyright (c) 2004, Intel Corporation.
+ */
+
+#ifndef __XEN_PUBLIC_HVM_IOREQ_H__
+#define __XEN_PUBLIC_HVM_IOREQ_H__
+
+#define IOREQ_READ      1
+#define IOREQ_WRITE     0
+
+#define STATE_IOREQ_NONE        0
+#define STATE_IOREQ_READY       1
+#define STATE_IOREQ_INPROCESS   2
+#define STATE_IORESP_READY      3
+
+#define IOREQ_TYPE_PIO          0 /* pio */
+#define IOREQ_TYPE_COPY         1 /* mmio ops */
+#define IOREQ_TYPE_PCI_CONFIG   2
+#define IOREQ_TYPE_TIMEOFFSET   7
+#define IOREQ_TYPE_INVALIDATE   8 /* mapcache */
+
+/*
+ * VMExit dispatcher should cooperate with instruction decoder to
+ * prepare this structure and notify service OS and DM by sending
+ * virq.
+ *
+ * For I/O type IOREQ_TYPE_PCI_CONFIG, the physical address is formatted
+ * as follows:
+ *
+ * 63....48|47..40|39..35|34..32|31........0
+ * SEGMENT |BUS   |DEV   |FN    |OFFSET
+ */
+struct ioreq {
+	uint64_t addr;          /* physical address */
+	uint64_t data;          /* data (or paddr of data) */
+	uint32_t count;         /* for rep prefixes */
+	uint32_t size;          /* size in bytes */
+	uint32_t vp_eport;      /* evtchn for notifications to/from device model */
+	uint16_t _pad0;
+	uint8_t state:4;
+	uint8_t data_is_ptr:1;  /* if 1, data above is the guest paddr
+				 * of the real data to use. */
+	uint8_t dir:1;          /* 1=read, 0=write */
+	uint8_t df:1;
+	uint8_t _pad1:1;
+	uint8_t type;           /* I/O type */
+};
+
+#endif /* __XEN_PUBLIC_HVM_IOREQ_H__ */