This change adds support for horizontal scroll to ps/2 mouse device
code. The code is implemented to match the logic of linux kernel
which is used as a reference.

Signed-off-by: Dmitry Petrov <dpetroff@gmail.com>
Message-Id: <20220108153947.171861-2-dpetroff@gmail.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

Previously, the large modes (>1080p) that were generated by Qemu in its EDID
were all 50 Hz. If we provide them to a Guest OS and the user selects
one of these modes, then the OS by default only gets 50 FPS. This is
especially true for Windows OS. With this patch, we are now exposing a
3840x2160@60 Hz which will allow the guest OS to get 60 FPS.

Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Satyeshwar Singh <satyeshwar.singh@intel.com>
Message-Id: <20211116221103.27128-1-dongwon.kim@intel.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

Currently QEMU defaults to a resolution of 1024x768 when exposing EDID
info to the guest OS. The EDID default info is important as this will
influence what resolution many guest OS will configure the screen with
on boot. It can also potentially influence what resolution the firmware
will configure the screen with, though until very recently EDK2 would
not handle EDID info.

One important thing to bear in mind is that the default graphics card
driver provided by Windows will leave the display set to whatever
resolution was enabled by the firmware on boot. Even if sufficient
VRAM is available, the resolution can't be changed without installing
new drivers. IOW, the default resolution choice is quite important
for usability of Windows.

Modern real world monitor hardware for desktop/laptop has supported
resolutions higher than 1024x768 for a long time now, perhaps as long
as 15+ years. There are quite a wide variety of native resolutions in
use today, however, and in wide screen form factors the height may not
be all that tall.

None the less, it is considered that there is scope for making the
QEMU default resolution slightly larger.

In considering what possible new default could be suitable, choices
considered were 1280x720 (720p), 1280x800 (WXGA) and 1280x1024 (SXGA).

In many ways, vertical space is the most important, and so 720p was
discarded due to loosing vertical space, despite being 25% wider.

The SXGA resolution would be good, but when taking into account
window titlebars/toolbars and window manager desktop UI, this might
be a little too tall for some users to fit the guest on their physical
montior.

This patch thus suggests a modest change to 1280x800 (WXGA). This
only consumes 1 MB per colour channel, allowing double buffered
framebuffer in 8 MB of VRAM. Width wise this is 25% larger than
QEMU's current default, but height wise this only adds 5%, so the
difference isn't massive on the QEMU side.

Overall there doesn't appear to be a compelling reason to stick
with 1024x768 resolution.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Message-Id: <20211129140508.1745130-1-berrange@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

vga_mmio_init() is used only one time and not very helpful,
inline and remove it.

Reviewed-by: BALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20211206224528.563588-5-f4bug@amsat.org>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

Introduce TYPE_VGA_MMIO, a sysbus device.

While there is no change in the vga_mmio_init()
interface, this is a migration compatibility break
of the MIPS Acer Pica 61 Jazz machine (pica61).

Suggested-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: BALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20211206224528.563588-4-f4bug@amsat.org>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

Inline vga_mm_init() in vga_mmio_init() to simplify the
next patch review. Kind of.

Reviewed-by: BALATON Zoltan <balaton@eik.bme.hu>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20211206224528.563588-3-f4bug@amsat.org>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

There is no ISA bus part in the MMIO VGA device, so rename:

 *  hw/display/vga-isa-mm.c -> hw/display/vga-mmio.c
 *  CONFIG_VGA_ISA_MM -> CONFIG_VGA_MMIO
 *  ISAVGAMMState -> VGAMmioState
 *  isa_vga_mm_init() -> vga_mmio_init()

Reviewed-by: BALATON Zoltan <balaton@eik.bme.hu>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20211206224528.563588-2-f4bug@amsat.org>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

Otherwise we run the error handling code even for successful requests.

Fixes: 13b250b1 ("uas: add stream number sanity checks.")
Reported-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211210080659.2537084-1-kraxel@redhat.com>

Quote from:
High Definition Audio Specification 1.0a, section 3.3.35

Offset 80: {IOB}SDnCTL Stream Reset (SRST): Writing a 1 causes
the corresponding stream to be reset. The Stream Descriptor
registers (except the SRST bit itself) ... are reset.

Change the code to reset the Stream Descriptor Control and Status
registers except the SRST bit.

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/757


Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20211226154017.6067-3-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

A device of USB video class usually uses larger desc structure, so
use larger buffer to avoid failure. (dev-video.c is ready)

This is an unlikely code path:
1, during guest startup, guest tries to probe device.
2, run 'lsusb' command in guest(or other similar commands).

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: zhenwei pi <pizhenwei@bytedance.com>
Message-Id: <20220112015835.900619-1-pizhenwei@bytedance.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

Linux need to fill up the HID descriptor in order to let the driver be
emulated. The descriptor was downloaded from [1]. The patch was tested
with evtest tool on top of qemu 5.2.0 with linux kernel 4.19.208.

[1] https://github.com/linuxwacom/wacom-hid-descriptors/tree/master/Wacom%20PenPartner



Signed-off-by: Michael Trimarchi <michael@amarulasolutions.com>
Co-developed-by: Michael Trimarchi <michael@amarulasolutions.com>
Signed-off-by: Dario Binacchi <dario.binacchi@amarulasolutions.com>
Message-Id: <20220112090125.381364-1-dario.binacchi@amarulasolutions.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

This reverts commit bf1d85c1.

Fixes: bf1d85c1 ("virtio: introduce macro IRTIO_CONFIG_IRQ_IDX")
Cc: "Cindy Lu" <lulu@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

This reverts commit e3480ef8.

Fixes: e3480ef8 ("virtio-pci: decouple notifier from interrupt process")
Cc: "Cindy Lu" <lulu@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

This reverts commit 316011b8.

Fixes: 316011b8 ("virtio-pci: decouple the single vector from the interrupt process")
Cc: "Cindy Lu" <lulu@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

This reverts commit 634f7c89.

Fixes: 634f7c89 ("vhost-vdpa: add support for config interrupt")
Cc: "Cindy Lu" <lulu@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

This reverts commit 081f864f.

Fixes: 081f864f ("virtio: add support for configure interrupt")
Cc: "Cindy Lu" <lulu@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

This reverts commit f7220a7c.

Fixes: f7220a7c ("vhost: add support for configure interrupt")
Cc: "Cindy Lu" <lulu@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

This reverts commit 497679d5.

Fixes: 497679d5 ("virtio-net: add support for configure interrupt")
Cc: "Cindy Lu" <lulu@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

This reverts commit d48185f1.

Fixes: d48185f1 ("virtio-mmio: add support for configure interrupt")
Cc: "Cindy Lu" <lulu@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

This reverts commit d5d24d85.

Fixes: d5d24d85 ("virtio-pci: add support for configure interrupt")
Cc: "Cindy Lu" <lulu@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

The macfb VRAM memory region was configured with coalescing rather than dirty
memory logging enabled, causing some areas of the screen not to redraw after
a full screen update.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Fixes: 8ac919a0 ("hw/m68k: add Nubus macfb video card")
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20220108164147.30813-1-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

"qemu-system-m68k -M q800 -bios /dev/null" crashes with a segfault
in q800_init().
This happens because the code doesn't check that rom_ptr() returned
a non-NULL pointer .

To avoid NULL pointer, don't allow 0 sized file and use bios_size with
rom_ptr().

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/756


Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-Id: <20220107105049.961489-1-laurent@vivier.eu>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>

Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20211218114340.1856757-1-laurent@vivier.eu>

When realising the SoC use error_fatal instead of error_abort as the
process can fail and report useful information to the user.

Currently a user can see this:

   $ ../qemu/bld/qemu-system-riscv64 -M sifive_u -S -monitor stdio -display none -drive if=pflash
    QEMU 6.1.93 monitor - type 'help' for more information
    (qemu) Unexpected error in sifive_u_otp_realize() at ../hw/misc/sifive_u_otp.c:229:
    qemu-system-riscv64: OTP drive size < 16K
    Aborted (core dumped)

Which this patch addresses

Reported-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Message-Id: <20220105213937.1113508-8-alistair.francis@opensource.wdc.com>

We can remove the original sifive_plic_irqs_pending() function and
instead just use the sifive_plic_claim() function (renamed to
sifive_plic_claimed()) to determine if any interrupts are pending.

This requires move the side effects outside of sifive_plic_claimed(),
but as they are only invoked once that isn't a problem.

We have also removed all of the old #ifdef debugging logs, so let's
cleanup the last remaining debug function while we are here.

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Message-Id: <20220105213937.1113508-5-alistair.francis@opensource.wdc.com>

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Message-Id: <20220105213937.1113508-4-alistair.francis@opensource.wdc.com>

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Message-Id: <20220105213937.1113508-3-alistair.francis@opensource.wdc.com>

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Message-Id: <20220105213937.1113508-2-alistair.francis@opensource.wdc.com>

It's obvious that PDMA supports 64-bit access of 64-bit registers, and
in previous commit, we confirm that PDMA supports 32-bit access of
both 32/64-bit registers. Thus, we configure 32/64-bit memory access
of PDMA registers as valid in general.

Signed-off-by: Jim Shu <jim.shu@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Message-id: 20220104063408.658169-3-jim.shu@sifive.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

Real PDMA supports high 32-bit read/write memory access of 64-bit
register.

The following result is PDMA tested in U-Boot on Unmatched board:

1. Real PDMA allows high 32-bit read/write to 64-bit register.
=> mw.l 0x3000000 0x0                      <= Disclaim channel 0
=> mw.l 0x3000000 0x1                      <= Claim channel 0
=> mw.l 0x3000010 0x80000000               <= Write low 32-bit NextDest (NextDest = 0x280000000)
=> mw.l 0x3000014 0x2                      <= Write high 32-bit NextDest
=> md.l 0x3000010 1                        <= Dump low 32-bit NextDest
03000010: 80000000
=> md.l 0x3000014 1                        <= Dump high 32-bit NextDest
03000014: 00000002
=> mw.l 0x3000018 0x80001000               <= Write low 32-bit NextSrc (NextSrc = 0x280001000)
=> mw.l 0x300001c 0x2                      <= Write high 32-bit NextSrc
=> md.l 0x3000018 1                        <= Dump low 32-bit NextSrc
03000010: 80001000
=> md.l 0x300001c 1                        <= Dump high 32-bit NextSrc
03000014: 00000002

2. PDMA transfer from 0x280001000 to 0x280000000 is OK.
=> mw.q 0x3000008 0x4                      <= NextBytes = 4
=> mw.l 0x3000004 0x22000000               <= wsize = rsize = 2 (2^2 = 4 bytes)
=> mw.l 0x280000000 0x87654321             <= Fill test data to dst
=> mw.l 0x280001000 0x12345678             <= Fill test data to src
=> md.l 0x280000000 1; md.l 0x280001000 1  <= Dump src/dst memory contents
280000000: 87654321                              !Ce.
280001000: 12345678                              xV4.
=> md.l 0x3000000 8                        <= Dump PDMA status
03000000: 00000001 22000000 00000004 00000000    ......."........
03000010: 80000000 00000002 80001000 00000002    ................
=> mw.l 0x3000000 0x3                      <= Set channel 0 run and claim bits
=> md.l 0x3000000 8                        <= Dump PDMA status
03000000: 40000001 22000000 00000004 00000000    ...@..."........
03000010: 80000000 00000002 80001000 00000002    ................
=> md.l 0x280000000 1; md.l 0x280001000 1  <= Dump src/dst memory contents
280000000: 12345678                               xV4.
280001000: 12345678                               xV4.

Signed-off-by: Jim Shu <jim.shu@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Message-id: 20220104063408.658169-2-jim.shu@sifive.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

Add missing TPM device identification objects _STR and _UID. They will
appear as files 'description' and 'uid' under Linux sysfs.

Following inspection of sysfs entries for hardware TPMs we chose
uid '1'.

Cc: Shannon Zhao <shannon.zhaosl@gmail.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Ani Sinha <ani@anisinha.ca>
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/708


Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Ani Sinha <ani@anisinha.ca>
Reviewed-by: Shannon Zhao <shannon.zhaosl@gmail.com>
Message-id: 20211223022310.575496-3-stefanb@linux.ibm.com
Message-Id: <20220104175806.872996-3-stefanb@linux.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>

In case of an error during initialization in vhost_dev_init, vhostfd is
closed in vhost_dev_cleanup. Remove close from err_virtio as it's both
redundant and causes a double close on vhostfd.

Signed-off-by: Daniil Tatianin <d-tatianin@yandex-team.ru>
Message-Id: <20211129125204.1108088-1-d-tatianin@yandex-team.ru>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost_dev_init calls vhost_dev_cleanup on error, which closes vhostfd,
don't double close it.

Signed-off-by: Daniil Tatianin <d-tatianin@yandex-team.ru>
Message-Id: <20211129132358.1110372-2-d-tatianin@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost_dev_init calls vhost_dev_cleanup in case of an error during
initialization, which zeroes out the entire vsc->dev as well as the
vsc->dev.vqs pointer. This prevents us from properly freeing it in free_vqs.
Keep a local copy of the pointer so we can free it later.

Signed-off-by: Daniil Tatianin <d-tatianin@yandex-team.ru>
Message-Id: <20211129132358.1110372-1-d-tatianin@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

When running "qemu-system-x86_64 -M pc,help" I noticed that some
properties were still missing their description. Add them now so
that users get at least a slightly better idea what they are all
about.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20211206134255.94784-1-thuth@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Get rid of the static variable that keeps track of whether hotplug has been
disabled on the root pci bus. Simply use qbus_is_hotpluggable() api to
perform the same check. This eliminates additional if conditional and
simplifies the function.

Signed-off-by: Ani Sinha <ani@anisinha.ca>
Message-Id: <1640764674-7784-1-git-send-email-ani@anirban.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

if QEMU is started with used provided SLIC table blob,

  -acpitable sig=SLIC,oem_id='CRASH ',oem_table_id="ME",oem_rev=00002210,asl_compiler_id="",asl_compiler_rev=00000000,data=/dev/null
it will assert with:

  hw/acpi/aml-build.c:61:build_append_padded_str: assertion failed: (len <= maxlen)

and following backtrace:

  ...
  build_append_padded_str (array=0x555556afe320, str=0x555556afdb2e "CRASH ME", maxlen=0x6, pad=0x20) at hw/acpi/aml-build.c:61
  acpi_table_begin (desc=0x7fffffffd1b0, array=0x555556afe320) at hw/acpi/aml-build.c:1727
  build_fadt (tbl=0x555556afe320, linker=0x555557ca3830, f=0x7fffffffd318, oem_id=0x555556afdb2e "CRASH ME", oem_table_id=0x555556afdb34 "ME") at hw/acpi/aml-build.c:2064
  ...

which happens due to acpi_table_begin() expecting NULL terminated
oem_id and oem_table_id strings, which is normally the case, but
in case of user provided SLIC table, oem_id points to table's blob
directly and as result oem_id became longer than expected.

Fix issue by handling oem_id consistently and make acpi_get_slic_oem()
return NULL terminated strings.

PS:
After [1] refactoring, oem_id semantics became inconsistent, where
NULL terminated string was coming from machine and old way pointer
into byte array coming from -acpitable option. That used to work
since build_header() wasn't expecting NULL terminated string and
blindly copied the 1st 6 bytes only.

However commit [2] broke that by replacing build_header() with
acpi_table_begin(), which was expecting NULL terminated string
and was checking oem_id size.

1) 602b4582 ("acpi: Permit OEM ID and OEM table ID fields to be changed")
2)
Fixes: 4b56e1e4 ("acpi: build_fadt: use acpi_table_begin()/acpi_table_end() instead of build_header()")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/786


Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20211227193120.1084176-2-imammedo@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Denis Lisov <dennis.lissov@gmail.com>
Tested-by: Alexander Tsoy <alexander@tsoy.me>
Cc: qemu-stable@nongnu.org
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

When scalable mode is enabled, the passthrough more is not determined
by the context entry but PASID entry, so switch to use the logic of
vtd_dev_pt_enabled() to determine the passthrough mode in
vtd_do_iommu_translate().

Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20220105041945.13459-2-jasowang@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Set the new default to "auto", keeping it set to "off" for compat
machines. This property is only available for x86 targets.

Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta@ionos.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20211217134039.29670-4-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

With VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE, we signal the VM that reading
unplugged memory is not supported. We have to fail feature negotiation
in case the guest does not support VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE.

First, VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE is required to properly handle
memory backends (or architectures) without support for the shared zeropage
in the hypervisor cleanly. Without the shared zeropage, even reading an
unpopulated virtual memory location can populate real memory and
consequently consume memory in the hypervisor. We have a guaranteed shared
zeropage only on MAP_PRIVATE anonymous memory.

Second, we want VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE to be the default
long-term as even populating the shared zeropage can be problematic: for
example, without THP support (possible) or without support for the shared
huge zeropage with THP (unlikely), the PTE page tables to hold the shared
zeropage entries can consume quite some memory that cannot be reclaimed
easily.

Third, there are other optimizations+features (e.g., protection of
unplugged memory, reducing the total memory slot size and bitmap sizes)
that will require VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE.

We really only support x86 targets with virtio-mem for now (and
Linux similarly only support x86), but that might change soon, so prepare
for different targets already.

Add a new "unplugged-inaccessible" tristate property for x86 targets:
- "off" will keep VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE unset and legacy
  guests working.
- "on" will set VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE and stop legacy guests
  from using the device.
- "auto" selects the default based on support for the shared zeropage.

Warn in case the property is set to "off" and we don't have support for the
shared zeropage.

For existing compat machines, the property will default to "off", to
not change the behavior but eventually warn about a problematic setup.
Short-term, we'll set the property default to "auto" for new QEMU machines.
Mid-term, we'll set the property default to "on" for new QEMU machines.
Long-term, we'll deprecate the parameter and disallow legacy
guests completely.

The property has to match on the migration source and destination. "auto"
will result in the same VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE setting as long
as the qemu command line (esp. memdev) match -- so "auto" is good enough
for migration purposes and the parameter doesn't have to be migrated
explicitly.

Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20211217134039.29670-3-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>