From: Arun Sharma <arun.sha...@intel.com>
Open and bind event channels; map ioreq and buffered ioreq rings.
Signed-off-by: Arun Sharma <arun.sha...@intel.com>
Signed-off-by: Anthony PERARD <anthony.per...@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabell...@eu.citrix.com>
---
hw/xen_common.h | 3 +
xen-all.c | 407 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 410 insertions(+), 0 deletions(-)
diff --git a/hw/xen_common.h b/hw/xen_common.h
index 0cc935a..a85b1c4 100644
--- a/hw/xen_common.h
+++ b/hw/xen_common.h
@@ -26,6 +26,7 @@
/* Xen unstable */
#if CONFIG_XEN_CTRL_INTERFACE_VERSION < 410
typedef int qemu_xc_interface;
+# define XC_INTERFACE_FMT "%i"
# define XC_HANDLER_INITIAL_VALUE -1
static inline int xc_fd(int xen_xc)
{
@@ -33,6 +34,7 @@ static inline int xc_fd(int xen_xc)
}
#else
typedef xc_interface *qemu_xc_interface;
+# define XC_INTERFACE_FMT "%p"
# define XC_HANDLER_INITIAL_VALUE NULL
/* FIXME The fd of xen_xc is now xen_xc->fd */
/* fd is the first field, so this works */
@@ -43,5 +45,6 @@ static inline int xc_fd(xc_interface *xen_xc)
#endif
qemu_irq *i8259_xen_init(void);
+void destroy_hvm_domain(void);
#endif /* QEMU_HW_XEN_COMMON_H */
diff --git a/xen-all.c b/xen-all.c
index 2b9e71c..36bdf30 100644
--- a/xen-all.c
+++ b/xen-all.c
@@ -8,6 +8,8 @@
#include "config.h"
+#include <sys/mman.h>
+
#include "hw/pci.h"
#include "hw/xen_common.h"
#include "hw/xen_backend.h"
@@ -15,6 +17,50 @@
#include "xen-mapcache.h"
+#include <xen/hvm/ioreq.h>
+#include <xen/hvm/params.h>
+
+//#define DEBUG_XEN
+
+#ifdef DEBUG_XEN
+#define DPRINTF(fmt, ...) \
+ do { fprintf(stderr, "xen: " fmt, ## __VA_ARGS__); } while (0)
+#else
+#define DPRINTF(fmt, ...) \
+ do { } while (0)
+#endif
+
+/* Compatibility with older version */
+#if __XEN_LATEST_INTERFACE_VERSION__ < 0x0003020a
+# define xen_vcpu_eport(shared_page, i) \
+ (shared_page->vcpu_iodata[i].vp_eport)
+# define xen_vcpu_ioreq(shared_page, vcpu) \
+ (shared_page->vcpu_iodata[vcpu].vp_ioreq)
+# define FMT_ioreq_size PRIx64
+#else
+# define xen_vcpu_eport(shared_page, i) \
+ (shared_page->vcpu_ioreq[i].vp_eport)
+# define xen_vcpu_ioreq(shared_page, vcpu) \
+ (shared_page->vcpu_ioreq[vcpu])
+# define FMT_ioreq_size "u"
+#endif
+
+#define BUFFER_IO_MAX_DELAY 100
+
+typedef struct XenIOState {
+ shared_iopage_t *shared_page;
+ buffered_iopage_t *buffered_io_page;
+ QEMUTimer *buffered_io_timer;
+ /* the evtchn port for polling the notification, */
+ evtchn_port_t *ioreq_local_port;
+ /* the evtchn fd for polling */
+ int xce_handle;
+ /* which vcpu we are serving */
+ int send_vcpu;
+
+ Notifier exit;
+} XenIOState;
+
/* Xen specific function for piix pci */
int xen_pci_slot_get_pirq(PCIDevice *pci_dev, int irq_num)
@@ -115,10 +161,308 @@ void xen_ram_alloc(ram_addr_t ram_addr, ram_addr_t size)
}
+/* VCPU Operations, MMIO, IO ring ... */
+
+/* get the ioreq packets from share mem */
+static ioreq_t *cpu_get_ioreq_from_shared_memory(XenIOState *state, int vcpu)
+{
+ ioreq_t *req = xen_vcpu_ioreq(&state->shared_page, vcpu);
+
+ if (req->state != STATE_IOREQ_READY) {
+ DPRINTF("I/O request not ready: "
+ "%x, ptr: %x, port: %"PRIx64", "
+ "data: %"PRIx64", count: %" FMT_ioreq_size ", size: %"
FMT_ioreq_size "\n",
+ req->state, req->data_is_ptr, req->addr,
+ req->data, req->count, req->size);
+ return NULL;
+ }
+
+ xen_rmb(); /* see IOREQ_READY /then/ read contents of ioreq */
+
+ req->state = STATE_IOREQ_INPROCESS;
+ return req;
+}
+
+/* use poll to get the port notification */
+/* ioreq_vec--out,the */
+/* retval--the number of ioreq packet */
+static ioreq_t *cpu_get_ioreq(XenIOState *state)
+{
+ int i;
+ evtchn_port_t port;
+
+ port = xc_evtchn_pending(state->xce_handle);
+ if (port != -1) {
+ for (i = 0; i < smp_cpus; i++) {
+ if (state->ioreq_local_port[i] == port) {
+ break;
+ }
+ }
+
+ if (i == smp_cpus) {
+ hw_error("Fatal error while trying to get io event!\n");
+ }
+
+ /* unmask the wanted port again */
+ xc_evtchn_unmask(state->xce_handle, port);
+
+ /* get the io packet from shared memory */
+ state->send_vcpu = i;
+ return cpu_get_ioreq_from_shared_memory(state, i);
+ }
+
+ /* read error or read nothing */
+ return NULL;
+}
+
+static uint32_t do_inp(pio_addr_t addr, unsigned long size)
+{
+ switch (size) {
+ case 1:
+ return cpu_inb(addr);
+ case 2:
+ return cpu_inw(addr);
+ case 4:
+ return cpu_inl(addr);
+ default:
+ hw_error("inp: bad size: %04"FMT_pioaddr" %lx", addr, size);
+ }
+}
+
+static void do_outp(pio_addr_t addr,
+ unsigned long size, uint32_t val)
+{
+ switch (size) {
+ case 1:
+ return cpu_outb(addr, val);
+ case 2:
+ return cpu_outw(addr, val);
+ case 4:
+ return cpu_outl(addr, val);
+ default:
+ hw_error("outp: bad size: %04"FMT_pioaddr" %lx", addr, size);
+ }
+}
+
+static void cpu_ioreq_pio(ioreq_t *req)
+{
+ int i, sign;
+
+ sign = req->df ? -1 : 1;
+
+ if (req->dir == IOREQ_READ) {
+ if (!req->data_is_ptr) {
+ req->data = do_inp(req->addr, req->size);
+ } else {
+ uint32_t tmp;
+
+ for (i = 0; i < req->count; i++) {
+ tmp = do_inp(req->addr, req->size);
+ cpu_physical_memory_write(req->data + (sign * i * req->size),
+ (uint8_t *) &tmp, req->size);
+ }
+ }
+ } else if (req->dir == IOREQ_WRITE) {
+ if (!req->data_is_ptr) {
+ do_outp(req->addr, req->size, req->data);
+ } else {
+ for (i = 0; i < req->count; i++) {
+ uint32_t tmp = 0;
+
+ cpu_physical_memory_read(req->data + (sign * i * req->size),
+ (uint8_t*) &tmp, req->size);
+ do_outp(req->addr, req->size, tmp);
+ }
+ }
+ }
+}
+
+static void cpu_ioreq_move(ioreq_t *req)
+{
+ int i, sign;
+
+ sign = req->df ? -1 : 1;
+
+ if (!req->data_is_ptr) {
+ if (req->dir == IOREQ_READ) {
+ for (i = 0; i < req->count; i++) {
+ cpu_physical_memory_read(req->addr + (sign * i * req->size),
+ (uint8_t *) &req->data, req->size);
+ }
+ } else if (req->dir == IOREQ_WRITE) {
+ for (i = 0; i < req->count; i++) {
+ cpu_physical_memory_write(req->addr + (sign * i * req->size),
+ (uint8_t *) &req->data, req->size);
+ }
+ }
+ } else {
+ target_ulong tmp;
+
+ if (req->dir == IOREQ_READ) {
+ for (i = 0; i < req->count; i++) {
+ cpu_physical_memory_read(req->addr + (sign * i * req->size),
+ (uint8_t*) &tmp, req->size);
+ cpu_physical_memory_write(req->data + (sign * i * req->size),
+ (uint8_t*) &tmp, req->size);
+ }
+ } else if (req->dir == IOREQ_WRITE) {
+ for (i = 0; i < req->count; i++) {
+ cpu_physical_memory_read(req->data + (sign * i * req->size),
+ (uint8_t*) &tmp, req->size);
+ cpu_physical_memory_write(req->addr + (sign * i * req->size),
+ (uint8_t*) &tmp, req->size);
+ }
+ }
+ }
+}
+
+static void handle_ioreq(ioreq_t *req)
+{
+ if (!req->data_is_ptr && (req->dir == IOREQ_WRITE) &&
+ (req->size < sizeof (target_ulong))) {
+ req->data &= ((target_ulong) 1 << (8 * req->size)) - 1;
+ }
+
+ switch (req->type) {
+ case IOREQ_TYPE_PIO:
+ cpu_ioreq_pio(req);
+ break;
+ case IOREQ_TYPE_COPY:
+ cpu_ioreq_move(req);
+ break;
+ case IOREQ_TYPE_TIMEOFFSET:
+ break;
+ case IOREQ_TYPE_INVALIDATE:
+ qemu_invalidate_map_cache();
+ break;
+ default:
+ hw_error("Invalid ioreq type 0x%x\n", req->type);
+ }
+}
+
+static void handle_buffered_iopage(XenIOState *state)
+{
+ buf_ioreq_t *buf_req = NULL;
+ ioreq_t req;
+ int qw;
+
+ if (!state->buffered_io_page) {
+ return;
+ }
+
+ while (state->buffered_io_page->read_pointer !=
state->buffered_io_page->write_pointer) {
+ buf_req = &state->buffered_io_page->buf_ioreq[
+ state->buffered_io_page->read_pointer % IOREQ_BUFFER_SLOT_NUM];
+ req.size = 1UL << buf_req->size;
+ req.count = 1;
+ req.addr = buf_req->addr;
+ req.data = buf_req->data;
+ req.state = STATE_IOREQ_READY;
+ req.dir = buf_req->dir;
+ req.df = 1;
+ req.type = buf_req->type;
+ req.data_is_ptr = 0;
+ qw = (req.size == 8);
+ if (qw) {
+ buf_req = &state->buffered_io_page->buf_ioreq[
+ (state->buffered_io_page->read_pointer + 1) %
IOREQ_BUFFER_SLOT_NUM];
+ req.data |= ((uint64_t)buf_req->data) << 32;
+ }
+
+ handle_ioreq(&req);
+
+ xen_mb();
+ state->buffered_io_page->read_pointer += qw ? 2 : 1;
+ }
+}
+
+static void handle_buffered_io(void *opaque)
+{
+ XenIOState *state = opaque;
+
+ handle_buffered_iopage(state);
+ qemu_mod_timer(state->buffered_io_timer,
+ BUFFER_IO_MAX_DELAY + qemu_get_clock(rt_clock));
+}
+
+static void cpu_handle_ioreq(void *opaque)
+{
+ XenIOState *state = opaque;
+ ioreq_t *req = cpu_get_ioreq(state);
+
+ handle_buffered_iopage(state);
+ if (req) {
+ handle_ioreq(req);
+
+ if (req->state != STATE_IOREQ_INPROCESS) {
+ fprintf(stderr, "Badness in I/O request ... not in service?!: "
+ "%x, ptr: %x, port: %"PRIx64", "
+ "data: %"PRIx64", count: %" FMT_ioreq_size ", size: %"
FMT_ioreq_size "\n",
+ req->state, req->data_is_ptr, req->addr,
+ req->data, req->count, req->size);
+ destroy_hvm_domain();
+ return;
+ }
+
+ xen_wmb(); /* Update ioreq contents /then/ update state. */
+
+ /*
+ * We do this before we send the response so that the tools
+ * have the opportunity to pick up on the reset before the
+ * guest resumes and does a hlt with interrupts disabled which
+ * causes Xen to powerdown the domain.
+ */
+ if (vm_running) {
+ if (qemu_shutdown_requested_get()) {
+ destroy_hvm_domain();
+ }
+ if (qemu_reset_requested_get()) {
+ qemu_system_reset();
+ }
+ }
+
+ req->state = STATE_IORESP_READY;
+ xc_evtchn_notify(state->xce_handle,
state->ioreq_local_port[state->send_vcpu]);
+ }
+}
+
+static void xen_main_loop_prepare(XenIOState *state)
+{
+ int evtchn_fd = state->xce_handle == -1 ? -1 :
xc_evtchn_fd(state->xce_handle);
+
+ state->buffered_io_timer = qemu_new_timer(rt_clock, handle_buffered_io,
+ state);
+ qemu_mod_timer(state->buffered_io_timer, qemu_get_clock(rt_clock));
+
+ if (evtchn_fd != -1) {
+ qemu_set_fd_handler(evtchn_fd, cpu_handle_ioreq, NULL, state);
+ }
+}
+
+
/* Initialise Xen */
+static void xen_vm_change_state_handler(void *opaque, int running, int reason)
+{
+ XenIOState *state = opaque;
+ if (running) {
+ xen_main_loop_prepare(state);
+ }
+}
+
+static void xen_exit_notifier(Notifier *n)
+{
+ XenIOState *state = container_of(n, XenIOState, exit);
+
+ xc_evtchn_close(state->xce_handle);
+}
+
int xen_init(int smp_cpus)
{
+ int i, rc;
+ unsigned long ioreq_pfn;
+ XenIOState *state;
+
if (xen_mode == XEN_EMULATE)
xen_mode = XEN_ATTACH;
xen_interfaces_init();
@@ -128,9 +472,72 @@ int xen_init(int smp_cpus)
return -1;
}
+ state = qemu_mallocz(sizeof (XenIOState));
+
+ state->xce_handle = xc_evtchn_open();
+ if (state->xce_handle == -1) {
+ perror("xen: event channel open");
+ return -errno;
+ }
+
+ state->exit.notify = xen_exit_notifier;
+ qemu_add_exit_notifier(&state->exit);
+
+ xc_get_hvm_param(xen_xc, xen_domid, HVM_PARAM_IOREQ_PFN, &ioreq_pfn);
+ DPRINTF("shared page at pfn %lx\n", ioreq_pfn);
+ state->shared_page = xc_map_foreign_range(xen_xc, xen_domid, XC_PAGE_SIZE,
+ PROT_READ|PROT_WRITE,
ioreq_pfn);
+ if (state->shared_page == NULL) {
+ hw_error("map shared IO page returned error %d handle="
XC_INTERFACE_FMT,
+ errno, xen_xc);
+ }
+
+ xc_get_hvm_param(xen_xc, xen_domid, HVM_PARAM_BUFIOREQ_PFN, &ioreq_pfn);
+ DPRINTF("buffered io page at pfn %lx\n", ioreq_pfn);
+ state->buffered_io_page = xc_map_foreign_range(xen_xc, xen_domid,
XC_PAGE_SIZE,
+ PROT_READ|PROT_WRITE, ioreq_pfn);
+ if (state->buffered_io_page == NULL) {
+ hw_error("map buffered IO page returned error %d", errno);
+ }
+
+ state->ioreq_local_port = qemu_mallocz(smp_cpus * sizeof (evtchn_port_t));
+
+ /* FIXME: how about if we overflow the page here? */
+ for (i = 0; i < smp_cpus; i++) {
+ rc = xc_evtchn_bind_interdomain(state->xce_handle, xen_domid,
+ xen_vcpu_eport(state->shared_page, i));
+ if (rc == -1) {
+ fprintf(stderr, "bind interdomain ioctl error %d\n", errno);
+ return -1;
+ }
+ state->ioreq_local_port[i] = rc;
+ }
+
/* Init RAM management */
qemu_map_cache_init();
xen_ram_init(ram_size);
+ qemu_add_vm_change_state_handler(xen_vm_change_state_handler, state);
+
return 0;
}
+
+void destroy_hvm_domain(void)
+{
+ qemu_xc_interface xc_handle;
+ int sts;
+
+ xc_handle = xc_interface_open(0, 0, 0);
+ if (xc_handle == XC_HANDLER_INITIAL_VALUE) {
+ fprintf(stderr, "Cannot acquire xenctrl handle\n");
+ } else {
+ sts = xc_domain_shutdown(xc_handle, xen_domid, SHUTDOWN_poweroff);
+ if (sts != 0) {
+ fprintf(stderr, "? xc_domain_shutdown failed to issue poweroff, "
+ "sts %d, %s\n", sts, strerror(errno));
+ } else {
+ fprintf(stderr, "Issued domain %d poweroff\n", xen_domid);
+ }
+ xc_interface_close(xc_handle);
+ }
+}
--
1.7.1
