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>

2 files changed, 1 insertions, 29 deletions

diff --git a/target/alpha/STATUS b/target/alpha/STATUS
deleted file mode 100644
index 6c9744569e..0000000000
--- a/target/alpha/STATUS
+++ /dev/null
@@ -1,28 +0,0 @@
-(to be completed)
-
-Alpha emulation structure:
-cpu.h           : CPU definitions globally exported
-exec.h          : CPU definitions used only for translated code execution
-helper.c        : helpers that can be called either by the translated code
-                  or the QEMU core, including the exception handler.
-op_helper.c     : helpers that can be called only from TCG
-helper.h        : TCG helpers prototypes
-translate.c     : Alpha instructions to micro-operations translator
-
-Code translator status:
-The Alpha CPU instruction emulation should be quite complete with the
-limitation that the VAX floating-point load and stores are not tested.
-The 4 MMU modes are implemented.
-
-Linux user mode emulation status:
-a few programs start to run. Most crash at a certain point, dereferencing a
-NULL pointer. It seems that the UNIQUE register is not initialized properly.
-It may appear that old executables, not relying on TLS support, run but
-this is to be proved...
-
-Full system emulation status:
-* Alpha PALCode emulation is in a very early stage and is not sufficient
-  to run any real OS. The alpha-softmmu target is not enabled for now.
-* no hardware platform description is implemented
-* there might be problems in the Alpha PALCode dedicated instructions
-  that would prevent to use a native PALCode image.
diff --git a/target/alpha/cpu.h b/target/alpha/cpu.h
index d0abc949a8..5e67304d81 100644
--- a/target/alpha/cpu.h
+++ b/target/alpha/cpu.h
@@ -276,9 +276,9 @@ extern const VMStateDescription vmstate_alpha_cpu;
 
 void alpha_cpu_do_interrupt(CPUState *cpu);
 bool alpha_cpu_exec_interrupt(CPUState *cpu, int int_req);
+hwaddr alpha_cpu_get_phys_page_debug(CPUState *cpu, vaddr addr);
 #endif /* !CONFIG_USER_ONLY */
 void alpha_cpu_dump_state(CPUState *cs, FILE *f, int flags);
-hwaddr alpha_cpu_get_phys_page_debug(CPUState *cpu, vaddr addr);
 int alpha_cpu_gdb_read_register(CPUState *cpu, GByteArray *buf, int reg);
 int alpha_cpu_gdb_write_register(CPUState *cpu, uint8_t *buf, int reg);