Merge branch 'xenfv-kvm-15' of git://git.infradead.org/users/dwmw2/qemu into HEAD

This adds support for emulating Xen under Linux/KVM, based on kernel
patches which have been present since Linux v5.12. As with the kernel
support, it's derived from work started by João Martins of Oracle in
2018.

This series just adds the basic platform support — CPUID, hypercalls,
event channels, a stub of XenStore.

A full single-tenant internal implementation of XenStore, and patches
to make QEMU's Xen PV drivers work with this Xen emulation, are waiting
in the wings to be submitted in a follow-on patch series.

As noted in the documentation, it's enabled by setting the xen-version
property on the KVM accelerator, e.g.:

 qemu-system-x86_64 -serial mon:stdio -M q35 -display none -m 1G -smp 2 \
    -accel kvm,xen-version=0x4000e,kernel-irqchip=split \
    -kernel vmlinuz-6.0.7-301.fc37.x86_64 \
    -append "console=ttyS0 root=/dev/sda1" \
    -drive file=/var/lib/libvirt/images/fedora28.qcow2,if=none,id=disk \
    -device ahci,id=ahci -device ide-hd,drive=disk,bus=ahci.0

Even before this was merged, we've already been using it to find and fix
bugs in the Linux kernel Xen guest support:

https://lore.kernel.org/all/4bffa69a949bfdc92c4a18e5a1c3cbb3b94a0d32.camel@infradead.org/
https://lore.kernel.org/all/871qnunycr.ffs@tglx/

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

1 files changed, 44 insertions, 36 deletions

diff --git a/hw/ide/core.c b/hw/ide/core.c
index 5d1039378f..2d034731cf 100644
--- a/hw/ide/core.c
+++ b/hw/ide/core.c
@@ -24,6 +24,7 @@
  */
 
 #include "qemu/osdep.h"
+#include "hw/irq.h"
 #include "hw/isa/isa.h"
 #include "migration/vmstate.h"
 #include "qemu/error-report.h"
@@ -653,7 +654,7 @@ void ide_set_sector(IDEState *s, int64_t sector_num)
 
 static void ide_rw_error(IDEState *s) {
     ide_abort_command(s);
-    ide_set_irq(s->bus);
+    ide_bus_set_irq(s->bus);
 }
 
 static void ide_buffered_readv_cb(void *opaque, int ret)
@@ -772,7 +773,7 @@ static void ide_sector_read_cb(void *opaque, int ret)
     s->nsector -= n;
     /* Allow the guest to read the io_buffer */
     ide_transfer_start(s, s->io_buffer, n * BDRV_SECTOR_SIZE, ide_sector_read);
-    ide_set_irq(s->bus);
+    ide_bus_set_irq(s->bus);
 }
 
 static void ide_sector_read(IDEState *s)
@@ -836,7 +837,7 @@ void ide_dma_error(IDEState *s)
     dma_buf_commit(s, 0);
     ide_abort_command(s);
     ide_set_inactive(s, false);
-    ide_set_irq(s->bus);
+    ide_bus_set_irq(s->bus);
 }
 
 int ide_handle_rw_error(IDEState *s, int error, int op)
@@ -906,7 +907,7 @@ static void ide_dma_cb(void *opaque, int ret)
     /* end of transfer ? */
     if (s->nsector == 0) {
         s->status = READY_STAT | SEEK_STAT;
-        ide_set_irq(s->bus);
+        ide_bus_set_irq(s->bus);
         goto eot;
     }
 
@@ -1006,7 +1007,7 @@ static void ide_sector_write(IDEState *s);
 static void ide_sector_write_timer_cb(void *opaque)
 {
     IDEState *s = opaque;
-    ide_set_irq(s->bus);
+    ide_bus_set_irq(s->bus);
 }
 
 static void ide_sector_write_cb(void *opaque, int ret)
@@ -1054,7 +1055,7 @@ static void ide_sector_write_cb(void *opaque, int ret)
         timer_mod(s->sector_write_timer, qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) +
                   (NANOSECONDS_PER_SECOND / 1000));
     } else {
-        ide_set_irq(s->bus);
+        ide_bus_set_irq(s->bus);
     }
 }
 
@@ -1105,7 +1106,7 @@ static void ide_flush_cb(void *opaque, int ret)
     }
     s->status = READY_STAT | SEEK_STAT;
     ide_cmd_done(s);
-    ide_set_irq(s->bus);
+    ide_bus_set_irq(s->bus);
 }
 
 static void ide_flush_cache(IDEState *s)
@@ -1194,7 +1195,7 @@ static void ide_cd_change_cb(void *opaque, bool load, Error **errp)
     s->cdrom_changed = 1;
     s->events.new_media = true;
     s->events.eject_request = false;
-    ide_set_irq(s->bus);
+    ide_bus_set_irq(s->bus);
 }
 
 static void ide_cd_eject_request_cb(void *opaque, bool force)
@@ -1205,7 +1206,7 @@ static void ide_cd_eject_request_cb(void *opaque, bool force)
     if (force) {
         s->tray_locked = false;
     }
-    ide_set_irq(s->bus);
+    ide_bus_set_irq(s->bus);
 }
 
 static void ide_cmd_lba48_transform(IDEState *s, int lba48)
@@ -1264,7 +1265,7 @@ const char *ATA_IOPORT_WR_lookup[ATA_IOPORT_WR_NUM_REGISTERS] = {
 void ide_ioport_write(void *opaque, uint32_t addr, uint32_t val)
 {
     IDEBus *bus = opaque;
-    IDEState *s = idebus_active_if(bus);
+    IDEState *s = ide_bus_active_if(bus);
     int reg_num = addr & 7;
 
     trace_ide_ioport_write(addr, ATA_IOPORT_WR_lookup[reg_num], val, bus, s);
@@ -1326,7 +1327,7 @@ void ide_ioport_write(void *opaque, uint32_t addr, uint32_t val)
     case ATA_IOPORT_WR_COMMAND:
         ide_clear_hob(bus);
         qemu_irq_lower(bus->irq);
-        ide_exec_cmd(bus, val);
+        ide_bus_exec_cmd(bus, val);
         break;
     }
 }
@@ -1439,7 +1440,7 @@ static bool cmd_identify(IDEState *s, uint8_t cmd)
         }
         s->status = READY_STAT | SEEK_STAT;
         ide_transfer_start(s, s->io_buffer, 512, ide_transfer_stop);
-        ide_set_irq(s->bus);
+        ide_bus_set_irq(s->bus);
         return false;
     } else {
         if (s->drive_kind == IDE_CD) {
@@ -1629,7 +1630,7 @@ static bool cmd_specify(IDEState *s, uint8_t cmd)
     if (s->blk && s->drive_kind != IDE_CD) {
         s->heads = (s->select & (ATA_DEV_HS)) + 1;
         s->sectors = s->nsector;
-        ide_set_irq(s->bus);
+        ide_bus_set_irq(s->bus);
     } else {
         ide_abort_command(s);
     }
@@ -1730,7 +1731,7 @@ static bool cmd_identify_packet(IDEState *s, uint8_t cmd)
     ide_atapi_identify(s);
     s->status = READY_STAT | SEEK_STAT;
     ide_transfer_start(s, s->io_buffer, 512, ide_transfer_stop);
-    ide_set_irq(s->bus);
+    ide_bus_set_irq(s->bus);
     return false;
 }
 
@@ -1755,7 +1756,7 @@ static bool cmd_exec_dev_diagnostic(IDEState *s, uint8_t cmd)
          * They are part of the regular output (this is why ERR_STAT isn't set)
          * Device 0 passed, Device 1 passed or not present. */
         s->error = 0x01;
-        ide_set_irq(s->bus);
+        ide_bus_set_irq(s->bus);
     }
 
     return false;
@@ -1787,7 +1788,7 @@ static bool cmd_cfa_req_ext_error_code(IDEState *s, uint8_t cmd)
 {
     s->error = 0x09;    /* miscellaneous error */
     s->status = READY_STAT | SEEK_STAT;
-    ide_set_irq(s->bus);
+    ide_bus_set_irq(s->bus);
 
     return false;
 }
@@ -1826,7 +1827,7 @@ static bool cmd_cfa_translate_sector(IDEState *s, uint8_t cmd)
     s->io_buffer[0x1a] = 0x01;                      /* Hot count */
 
     ide_transfer_start(s, s->io_buffer, 0x200, ide_transfer_stop);
-    ide_set_irq(s->bus);
+    ide_bus_set_irq(s->bus);
 
     return false;
 }
@@ -1850,7 +1851,7 @@ static bool cmd_cfa_access_metadata_storage(IDEState *s, uint8_t cmd)
 
     ide_transfer_start(s, s->io_buffer, 0x200, ide_transfer_stop);
     s->status = 0x00; /* NOTE: READY is _not_ set */
-    ide_set_irq(s->bus);
+    ide_bus_set_irq(s->bus);
 
     return false;
 }
@@ -1933,7 +1934,7 @@ static bool cmd_smart(IDEState *s, uint8_t cmd)
 
         s->status = READY_STAT | SEEK_STAT;
         ide_transfer_start(s, s->io_buffer, 0x200, ide_transfer_stop);
-        ide_set_irq(s->bus);
+        ide_bus_set_irq(s->bus);
         return false;
 
     case SMART_READ_DATA:
@@ -1974,7 +1975,7 @@ static bool cmd_smart(IDEState *s, uint8_t cmd)
 
         s->status = READY_STAT | SEEK_STAT;
         ide_transfer_start(s, s->io_buffer, 0x200, ide_transfer_stop);
-        ide_set_irq(s->bus);
+        ide_bus_set_irq(s->bus);
         return false;
 
     case SMART_READ_LOG:
@@ -2013,7 +2014,7 @@ static bool cmd_smart(IDEState *s, uint8_t cmd)
         }
         s->status = READY_STAT | SEEK_STAT;
         ide_transfer_start(s, s->io_buffer, 0x200, ide_transfer_stop);
-        ide_set_irq(s->bus);
+        ide_bus_set_irq(s->bus);
         return false;
 
     case SMART_EXECUTE_OFFLINE:
@@ -2122,13 +2123,13 @@ static bool ide_cmd_permitted(IDEState *s, uint32_t cmd)
         && (ide_cmd_table[cmd].flags & (1u << s->drive_kind));
 }
 
-void ide_exec_cmd(IDEBus *bus, uint32_t val)
+void ide_bus_exec_cmd(IDEBus *bus, uint32_t val)
 {
     IDEState *s;
     bool complete;
 
-    s = idebus_active_if(bus);
-    trace_ide_exec_cmd(bus, s, val);
+    s = ide_bus_active_if(bus);
+    trace_ide_bus_exec_cmd(bus, s, val);
 
     /* ignore commands to non existent slave */
     if (s != bus->ifs && !s->blk) {
@@ -2145,7 +2146,7 @@ void ide_exec_cmd(IDEBus *bus, uint32_t val)
 
     if (!ide_cmd_permitted(s, val)) {
         ide_abort_command(s);
-        ide_set_irq(s->bus);
+        ide_bus_set_irq(s->bus);
         return;
     }
 
@@ -2163,7 +2164,7 @@ void ide_exec_cmd(IDEBus *bus, uint32_t val)
         }
 
         ide_cmd_done(s);
-        ide_set_irq(s->bus);
+        ide_bus_set_irq(s->bus);
     }
 }
 
@@ -2194,7 +2195,7 @@ const char *ATA_IOPORT_RR_lookup[ATA_IOPORT_RR_NUM_REGISTERS] = {
 uint32_t ide_ioport_read(void *opaque, uint32_t addr)
 {
     IDEBus *bus = opaque;
-    IDEState *s = idebus_active_if(bus);
+    IDEState *s = ide_bus_active_if(bus);
     uint32_t reg_num;
     int ret, hob;
 
@@ -2280,7 +2281,7 @@ uint32_t ide_ioport_read(void *opaque, uint32_t addr)
 uint32_t ide_status_read(void *opaque, uint32_t addr)
 {
     IDEBus *bus = opaque;
-    IDEState *s = idebus_active_if(bus);
+    IDEState *s = ide_bus_active_if(bus);
     int ret;
 
     if ((!bus->ifs[0].blk && !bus->ifs[1].blk) ||
@@ -2369,7 +2370,7 @@ static bool ide_is_pio_out(IDEState *s)
 void ide_data_writew(void *opaque, uint32_t addr, uint32_t val)
 {
     IDEBus *bus = opaque;
-    IDEState *s = idebus_active_if(bus);
+    IDEState *s = ide_bus_active_if(bus);
     uint8_t *p;
 
     trace_ide_data_writew(addr, val, bus, s);
@@ -2405,7 +2406,7 @@ void ide_data_writew(void *opaque, uint32_t addr, uint32_t val)
 uint32_t ide_data_readw(void *opaque, uint32_t addr)
 {
     IDEBus *bus = opaque;
-    IDEState *s = idebus_active_if(bus);
+    IDEState *s = ide_bus_active_if(bus);
     uint8_t *p;
     int ret;
 
@@ -2443,7 +2444,7 @@ uint32_t ide_data_readw(void *opaque, uint32_t addr)
 void ide_data_writel(void *opaque, uint32_t addr, uint32_t val)
 {
     IDEBus *bus = opaque;
-    IDEState *s = idebus_active_if(bus);
+    IDEState *s = ide_bus_active_if(bus);
     uint8_t *p;
 
     trace_ide_data_writel(addr, val, bus, s);
@@ -2471,7 +2472,7 @@ void ide_data_writel(void *opaque, uint32_t addr, uint32_t val)
 uint32_t ide_data_readl(void *opaque, uint32_t addr)
 {
     IDEBus *bus = opaque;
-    IDEState *s = idebus_active_if(bus);
+    IDEState *s = ide_bus_active_if(bus);
     uint8_t *p;
     int ret;
 
@@ -2710,7 +2711,7 @@ static void ide_restart_bh(void *opaque)
         return;
     }
 
-    s = idebus_active_if(bus);
+    s = ide_bus_active_if(bus);
     is_read = (bus->error_status & IDE_RETRY_READ) != 0;
 
     /* The error status must be cleared before resubmitting the request: The
@@ -2758,7 +2759,7 @@ static void ide_restart_cb(void *opaque, bool running, RunState state)
     }
 }
 
-void ide_register_restart_cb(IDEBus *bus)
+void ide_bus_register_restart_cb(IDEBus *bus)
 {
     if (bus->dma->ops->restart_dma) {
         bus->vmstate = qemu_add_vm_change_state_handler(ide_restart_cb, bus);
@@ -2770,7 +2771,7 @@ static IDEDMA ide_dma_nop = {
     .aiocb = NULL,
 };
 
-void ide_init2(IDEBus *bus, qemu_irq irq)
+void ide_bus_init_output_irq(IDEBus *bus, qemu_irq irq_out)
 {
     int i;
 
@@ -2778,10 +2779,17 @@ void ide_init2(IDEBus *bus, qemu_irq irq)
         ide_init1(bus, i);
         ide_reset(&bus->ifs[i]);
     }
-    bus->irq = irq;
+    bus->irq = irq_out;
     bus->dma = &ide_dma_nop;
 }
 
+void ide_bus_set_irq(IDEBus *bus)
+{
+    if (!(bus->cmd & IDE_CTRL_DISABLE_IRQ)) {
+        qemu_irq_raise(bus->irq);
+    }
+}
+
 void ide_exit(IDEState *s)
 {
     timer_free(s->sector_write_timer);