vhost-user-gpu: fix OOB write in 'virgl_cmd_get_capset' (CVE-2021-3546)

Git-commit: 9f22893adcb02580aee5968f32baa2cd109b3ec2
References: CVE-2021-3546 bsc#1185981

If 'virgl_cmd_get_capset' set 'max_size' to 0,
the 'virgl_renderer_fill_caps' will write the data after the 'resp'.
This patch avoid this by checking the returned 'max_size'.

virtio-gpu fix: abd7f08b23 ("display: virtio-gpu-3d: check
virgl capabilities max_size")

Fixes: CVE-2021-3546
Reported-by: Li Qiang <liq3ea@163.com>
Reviewed-by: Prasad J Pandit <pjp@fedoraproject.org>
Signed-off-by: Li Qiang <liq3ea@163.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20210516030403.107723-8-liq3ea@163.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Jose R. Ziviani <jziviani@suse.de>

vhost-user-gpu: fix leak in 'virgl_resource_attach_backing' (CVE-2021-3544)

Git-commit: 63736af5a6571d9def93769431e0d7e38c6677bf
References: CVE-2021-3544 bsc#1186010

If 'virgl_renderer_resource_attach_iov' failed, the 'res_iovs' will
be leaked.

Fixes: CVE-2021-3544
Reported-by: Li Qiang <liq3ea@163.com>
virtio-gpu fix: 33243031da ("virtio-gpu-3d: fix memory leak
in resource attach backing")

Signed-off-by: Li Qiang <liq3ea@163.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20210516030403.107723-7-liq3ea@163.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Jose R. Ziviani <jziviani@suse.de>
[jrz: tweak title to not break spec file]

vhost-user-gpu: fix leak in 'virgl_cmd_resource_unref' (CVE-2021-3544)

Git-comit: f6091d86ba9ea05f4e111b9b42ee0005c37a6779
References: CVE-2021-3544 bsc#1186010

The 'res->iov' will be leaked if the guest trigger following sequences:

virgl_cmd_create_resource_2d
virgl_resource_attach_backing
virgl_cmd_resource_unref

This patch fixes this.

Fixes: CVE-2021-3544
Reported-by: Li Qiang <liq3ea@163.com>
virtio-gpu fix: 5e8e3c4c75 ("virtio-gpu: fix resource leak
in virgl_cmd_resource_unref"

Signed-off-by: Li Qiang <liq3ea@163.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20210516030403.107723-6-liq3ea@163.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Jose R. Ziviani <jziviani@suse.de>
[jrz: tweaked title to not break spec file]

vhost-user-gpu: fix memory leak while calling 'vg_resource_unref' (CVE-2021-3544)

Git-commit: b7afebcf9e6ecf3cf9b5a9b9b731ed04bca6aa3e
References: CVE-2021-3544 bsc#1186010

If the guest trigger following sequences, the attach_backing will be leaked:

vg_resource_create_2d
vg_resource_attach_backing
vg_resource_unref

This patch fix this by freeing 'res->iov' in vg_resource_destroy.

Fixes: CVE-2021-3544
Reported-by: Li Qiang <liq3ea@163.com>
virtio-gpu fix: 5e8e3c4c75 ("virtio-gpu: fix resource leak
in virgl_cmd_resource_unref")

Reviewed-by: Prasad J Pandit <pjp@fedoraproject.org>
Signed-off-by: Li Qiang <liq3ea@163.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20210516030403.107723-5-liq3ea@163.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Jose R. Ziviani <jziviani@suse.de>

vhost-user-gpu: fix memory leak in vg_resource_attach_backing (CVE-2021-3544)

Git-commit: b9f79858a614d95f5de875d0ca31096eaab72c3b
References: CVE-2021-3544 bsc#1186010

Check whether the 'res' has already been attach_backing to avoid
memory leak.

Fixes: CVE-2021-3544
Reported-by: Li Qiang <liq3ea@163.com>
virtio-gpu fix: 204f01b309 ("virtio-gpu: fix memory leak
in resource attach backing")

Signed-off-by: Li Qiang <liq3ea@163.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20210516030403.107723-4-liq3ea@163.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Jose R. Ziviani <jziviani@suse.de>

vhost-user-gpu: fix resource leak in 'vg_resource_create_2d' (CVE-2021-3544)

Git-commit: 86dd8fac2acc366930a5dc08d3fb1b1e816f4e1e
References: CVE-2021-3544 bsc#1186010

Call 'vugbm_buffer_destroy' in error path to avoid resource leak.

Fixes: CVE-2021-3544
Reported-by: Li Qiang <liq3ea@163.com>
Reviewed-by: Prasad J Pandit <pjp@fedoraproject.org>
Signed-off-by: Li Qiang <liq3ea@163.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20210516030403.107723-3-liq3ea@163.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Jose R. Ziviani <jziviani@suse.de>

vhost-user-gpu: fix memory disclosure in virgl_cmd_get_capset_info (CVE-2021-3545)

Git-commit: 121841b25d72d13f8cad554363138c360f1250ea
References: CVE-2021-3545 bsc#1185990

Otherwise some of the 'resp' will be leaked to guest.

Fixes: CVE-2021-3545
Reported-by: Li Qiang <liq3ea@163.com>
virtio-gpu fix: 42a8dadc74 ("virtio-gpu: fix information leak
in getting capset info dispatch")

Signed-off-by: Li Qiang <liq3ea@163.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20210516030403.107723-2-liq3ea@163.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Jose R. Ziviani <jziviani@suse.de>

pc-bios/s390-ccw: don't try to read the next block if end of chunk is reached

Git-commit: a6625d38cce3901a7c1cba069f0abcf743a293f1
References: bsc#1186290

Don't read the block if a null block number is reached, because this means that
the end of chunk is reached.

Reviewed-by: Collin Walling <walling@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Message-Id: <20210416074736.17409-1-mhartmay@linux.ibm.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Cho, Yu-Chen <acho@suse.com>

qom: handle case of chardev-spice module unavailability

When qemu is built with modules, but a given module doesn't load
qemu should handle that gracefully. When chardev-spice.so isn't
able to be loaded and qemu is invoked with -display spice-app,
qemu will reach an abort call. Explicitly detect these conditions
and error out in a normal way before we reach the abort.

Signed-off-by: Bruce Rogers <brogers@suse.com>

module: for virtio-gpu, pre-load module to avoid abort on missing module

If the hw-display-virtio-gpu module is not loadable when the virtio-gpu
device is referenced either on the command line or the monitor, qemu
will call abort. We can fail gracefully by moving the attempted module
load to a context better situated to handle errors properly. (bsc#1181103)

Signed-off-by: Bruce Rogers <brogers@suse.com>

usb: Help compiler out to avoid a warning on x86 compilation

Include-If: %ifarch %arm %ix86 ppc

There is an assert present which already should give the compiler
enough information about the value of i as used in the snprintf,
but if I remember right, for x86, because memory is tighter some of
the compiler smarts are turned off, so we get the uninformed warning
there and not on other archs. So on x86 only we'll add some code to
help the compiler out, so we can again compile qemu with
--enable-werror.

Signed-off-by: Bruce Rogers <brogers@suse.com>

roms/Makefile: add --cross-file to qboot meson setup for aarch64

Include-If: %ifarch aarch64

We conditionally add a --cross-file reference so that we can do
cross compilation of qboot from an aarch64 build.

Signed-off-by: Bruce Rogers <brogers@suse.com>

Makefile: Don't check pc-bios as pre-requisite for config-host.mak

This check isn't needed when we know this is a fresh build, which of
course it is when we are building the qemu packages.

Signed-off-by: Bruce Rogers <brogers@suse.com>

Revert "roms/efirom, tests/uefi-test-tools: update edk2's own submodules first"

This reverts commit ec87b5daca761039bbcf781eedbe4987f790836f.

No need. In our build system submodules are checked out.

Signed-off-by: Bruce Rogers <brogers@suse.com>

meson: install ivshmem-client and ivshmem-server

Turn on the meson install flag for these executables

Signed-off-by: Bruce Rogers <brogers@suse.com>

qht: Revert some constification in qht.c

This change partially addresses https://bugs.launchpad.net/qemu/+bug/1886155
where a pre-release gcc 11 warns about const qualifier abuse.

Signed-off-by: Bruce Rogers <brogers@suse.com>

Revert "qht: constify qht_statistics_init"

This reverts commit 6579f10779b5b5ed2e978e7b8cae7bcbf8665254.

This change partially addresses https://bugs.launchpad.net/qemu/+bug/1886155
where a pre-release gcc 11 warns about const qualifier abuse.

Signed-off-by: Bruce Rogers <brogers@suse.com>

s390x: Fix stringop-truncation issue reported by gcc 11

Signed-off-by: Bruce Rogers <brogers@suse.com>

docs: add SUSE support statements to html docs

Include-If: %if %{legacy_qemu_kvm}

We can fairly easily produce an html version of our support statements.
Now that qemu includes fairly good html-based documentation, leverage it
to expose our SUSE specific in-package support documentation.

Signed-off-by: Bruce Rogers <brogers@suse.com>

configure: remove $pkgversion from CONFIG_STAMP input to broaden compatibility

As part of the effort to close the gap with Leap I think we are fine
removing the $pkgversion component to creating a unique CONFIG_STAMP.
This stamp is only used in creating a unique symbol used in ensuring the
dynamically loaded modules correspond correctly to the loading qemu.
The default inputs to producing this unique symbol are somewhat reasonable
as a generic mechanism, but specific packaging and maintenance practices
might require the default to be modified for best use. This is an example
of that.

Signed-off-by: Bruce Rogers <brogers@suse.com>

test: add mapping from arch of i686 to qemu_arch=i386

While we don't specifically set QEMU_PROG, the code which detects the
host architecture needs a little help mapping the output of uname -m to
what the qemu project uses to reference that architecture.

Signed-off-by: Bruce Rogers <brogers@suse.com>

roms: change cross compiler naming to be suse specific

Signed-off-by: Bruce Rogers <brogers@suse.com>

pc-bios/s390-ccw/net: avoid warning about packed structure members

This is hopefully temporary. Simply disable the warning about taking
the address of packed structure members which is new in gcc9.

Signed-off-by: Bruce Rogers <brogers@suse.com>

configure: only populate roms if softmmu

Currently roms are mistakenly getting built in a linux-user only
configuration. Add check for softmmu in all places where our list of
roms is being added to.

Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/intc/exynos4210_gic: provide more room when formatting alias names

sprintf related parameter validation complains about the size of the
buffer being written to in exynos4210_gic_realize(). Provide a bit more
space to avoid the following warning:
/home/abuild/rpmbuild/BUILD/qemu-4.0.0/hw/intc/exynos4210_gic.c: In function 'exynos4210_gic_realize':
/home/abuild/rpmbuild/BUILD/qemu-4.0.0/hw/intc/exynos4210_gic.c:316:36: error: '%x' directive writing between 1 and 7 bytes into a region of size between 4 and 28 [-Werror=format-overflow=]
  316 |         sprintf(cpu_alias_name, "%s%x", cpu_prefix, i);
      |                                    ^~
/home/abuild/rpmbuild/BUILD/qemu-4.0.0/hw/intc/exynos4210_gic.c:316:33: note: directive argument in the range [0, 29020050]
  316 |         sprintf(cpu_alias_name, "%s%x", cpu_prefix, i);
      |                                 ^~~~~~
In file included from /usr/include/stdio.h:867,
                 from /home/abuild/rpmbuild/BUILD/qemu-4.0.0/include/qemu/osdep.h:99,
                 from /home/abuild/rpmbuild/BUILD/qemu-4.0.0/hw/intc/exynos4210_gic.c:23:
/usr/include/bits/stdio2.h:36:10: note: '__builtin___sprintf_chk' output between 2 and 32 bytes into a destination of size 28
   36 |   return __builtin___sprintf_chk (__s, __USE_FORTIFY_LEVEL - 1,
      |          ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   37 |       __bos (__s), __fmt, __va_arg_pack ());
      |       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/abuild/rpmbuild/BUILD/qemu-4.0.0/hw/intc/exynos4210_gic.c:326:37: error: '%x' directive writing between 1 and 7 bytes into a region of size between 3 and 28 [-Werror=format-overflow=]
  326 |         sprintf(dist_alias_name, "%s%x", dist_prefix, i);
      |                                     ^~
/home/abuild/rpmbuild/BUILD/qemu-4.0.0/hw/intc/exynos4210_gic.c:326:34: note: directive argument in the range [0, 29020050]
  326 |         sprintf(dist_alias_name, "%s%x", dist_prefix, i);
      |                                  ^~~~~~
In file included from /usr/include/stdio.h:867,
                 from /home/abuild/rpmbuild/BUILD/qemu-4.0.0/include/qemu/osdep.h:99,
                 from /home/abuild/rpmbuild/BUILD/qemu-4.0.0/hw/intc/exynos4210_gic.c:23:
/usr/include/bits/stdio2.h:36:10: note: '__builtin___sprintf_chk' output between 2 and 33 bytes into a destination of size 28
   36 |   return __builtin___sprintf_chk (__s, __USE_FORTIFY_LEVEL - 1,
      |          ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   37 |       __bos (__s), __fmt, __va_arg_pack ());
      |       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Signed-off-by: Bruce Rogers <brogers@suse.com>

tests: change error message in test 162

Since we have a quite restricted execution environment, as far as
networking is concerned, we need to change the error message we expect
in test 162. There is actually no routing set up so the error we get is
"Network is unreachable". Change the expected output accordingly.

Signed-off-by: Bruce Rogers <brogers@suse.com>

xen: ignore live parameter from xen-save-devices-state

References: bsc#1079730, bsc#1101982, bsc#1063993

The final step of xl migrate|save for an HVM domU is saving the state of
qemu. This also involves releasing all block devices. While releasing
backends ought to be a separate step, such functionality is not
implemented.

Unfortunately, releasing the block devices depends on the optional
'live' option. This breaks offline migration with 'virsh migrate domU
dom0' because the sending side does not release the disks, as a result
the receiving side can not properly claim write access to the disks.

As a minimal fix, remove the dependency on the 'live' option. Upstream
may fix this in a different way, like removing the newly added 'live'
parameter entirely.

Fixes: 5d6c599fe1 ("migration, xen: Fix block image lock issue on live migration")

Signed-off-by: Olaf Hering <olaf@aepfle.de>
Signed-off-by: Bruce Rogers <brogers@suse.com>

tests: Fix block tests to be compatible with membarrier configuration

The use of membarriers collides with the block test's practice of
SIGKILLing test vm's. Have them quit politely. Tests: 130, 153 - and
though test 161 seems to have the same issue, it is not yet fixed, but
just marked here as possibly needing a fix.

Signed-off-by: Bruce Rogers <brogers@suse.com>

tests/qemu-iotests: Triple timeout of i/o tests due to obs environment

Executing tests in obs is very fickle, since you aren't guaranteed
reliable cpu time. Triple the timeout for each test to help ensure
we don't fail a test because the stars align against us.

Signed-off-by: Bruce Rogers <brogers@suse.com>

xen: add block resize support for xen disks

Provide monitor naming of xen disks, and plumb guest driver
notification through xenstore of resizing instigated via the
monitor.

[BR: minor edits to pass qemu's checkpatch script]
[BR: significant rework needed due to upstream xen disk qdevification]
[BR: At this point, monitor_add_blk call is all we need to add!]
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/smbios: handle both file formats regardless of machine type

References: bsc#994082, bsc#1084316, boo#1131894

It's easy enough to handle either per-spec or legacy smbios structures
in the smbios file input without regard to the machine type used, by
simply applying the basic smbios formatting rules. then depending on
what is detected. terminal numm bytes are added or removed for machine
type specific processing.

Signed-off-by: Bruce Rogers <brogers@suse.com>

Make installed scripts explicitly python3

References: bsc#1077564

We want to explicitly reference python3 in the scripts we install.

Signed-off-by: Bruce Rogers <brogers@suse.com>

acpi_piix4: Fix migration from SLE11 SP2

References: bnc#812836

qemu-kvm 0.15 uses the same GPE format as qemu 1.4, but as version 2
rather than 3.

Signed-off-by: Andreas Färber <afaerber@suse.de>

i8254: Fix migration from SLE11 SP2

References: bnc#812836

qemu-kvm 0.15 had a VMSTATE_UINT32(flags, PITState) field that
qemu 1.4 does not have.

Signed-off-by: Andreas Färber <afaerber@suse.de>

increase x86_64 physical bits to 42

Allow for guests with higher amounts of ram. The current thought
is that 2TB specified on qemu commandline would be an appropriate
limit. Note that this requires the next higher bit value since
the highest address is actually more than 2TB due to the pci
memory hole.

Signed-off-by: Bruce Rogers <brogers@suse.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>

Raise soft address space limit to hard limit

For SLES we want users to be able to use large memory configurations
with KVM without fiddling with ulimit -Sv.

Signed-off-by: Andreas Färber <afaerber@suse.de>
[BR: add include for sys/resource.h]
Signed-off-by: Bruce Rogers <brogers@suse.com>

roms/Makefile: pass a packaging timestamp to subpackages with date info

References: bsc#1011213

Certain rom subpackages build from qemu git-submodules call the date
program to include date information in the packaged binaries. This
causes repeated builds of the package to be different, wkere the only
real difference is due to the fact that time build timestamp has
changed. To promote reproducible builds and avoid customers being
prompted to update packages needlessly, we'll use the timestamp of the
VERSION file as the packaging timestamp for all packages that build in a
timestamp for whatever reason.

Signed-off-by: Bruce Rogers <brogers@suse.com>

qemu-binfmt-conf: use qemu-ARCH-binfmt

Signed-off-by: Andreas Schwab <schwab@suse.de>
Signed-off-by: Andreas Färber <afaerber@suse.de>

qemu-bridge-helper: reduce security profile

References: boo#988279

Change from using glib alloc and free routines to those
from libc. Also perform safety measure of dropping privs
to user if configured no-caps.

Signed-off-by: Bruce Rogers <brogers@suse.com>
[AF: Rebased for v2.7.0-rc2]
Signed-off-by: Andreas Färber <afaerber@suse.de>

xen_disk: Add suse specific flush disable handling and map to QEMU equiv

Add code to read the suse specific suse-diskcache-disable-flush flag out
of xenstore, and set the equivalent flag within QEMU.

Patch taken from Xen's patch queue, Olaf Hering being the original author.
[bsc#879425]

[BR: minor edits to pass qemu's checkpatch script]
[BR: With qdevification of xen-block, code has changed significantly]
Signed-off-by: Bruce Rogers <brogers@suse.com>
Signed-off-by: Olaf Hering <olaf@aepfle.de>

AIO: Reduce number of threads for 32bit hosts

On hosts with limited virtual address space (32bit pointers), we can very
easily run out of virtual memory with big thread pools.

Instead, we should limit ourselves to small pools to keep memory footprint
low on those systems.

This patch fixes random VM stalls like

  (process:25114): GLib-ERROR **: gmem.c:103: failed to allocate 1048576 bytes

on 32bit ARM systems for me.

Signed-off-by: Alexander Graf <agraf@suse.de>

linux-user: lseek: explicitly cast non-set offsets to signed

When doing lseek, SEEK_SET indicates that the offset is an unsigned variable.
Other seek types have parameters that can be negative.

When converting from 32bit to 64bit parameters, we need to take this into
account and enable SEEK_END and SEEK_CUR to be negative, while SEEK_SET stays
absolute positioned which we need to maintain as unsigned.

Signed-off-by: Alexander Graf <agraf@suse.de>

Make char muxer more robust wrt small FIFOs

Virtio-Console can only process one character at a time. Using it on S390
gave me strange "lags" where I got the character I pressed before when
pressing one. So I typed in "abc" and only received "a", then pressed "d"
but the guest received "b" and so on.

While the stdio driver calls a poll function that just processes on its
queue in case virtio-console can't take multiple characters at once, the
muxer does not have such callbacks, so it can't empty its queue.

To work around that limitation, I introduced a new timer that only gets
active when the guest can not receive any more characters. In that case
it polls again after a while to check if the guest is now receiving input.

This patch fixes input when using -nographic on s390 for me.

[AF: Rebased for v2.7.0-rc2]
[BR: minor edits to pass qemu's checkpatch script]
Signed-off-by: Bruce Rogers <brogers@suse.com>

linux-user: use target_ulong

Linux syscalls pass pointers or data length or other information of that sort
to the kernel. This is all stuff you don't want to have sign extended.
Otherwise a host 64bit variable parameter with a size parameter will extend
it to a negative number, breaking lseek for example.

Pass syscall arguments as ulong always.

Signed-off-by: Alexander Graf <agraf@suse.de>

linux-user: Fake /proc/cpuinfo

Fedora 17 for ARM reads /proc/cpuinfo and fails if it doesn't contain
ARM related contents. This patch implements a quick hack to expose real
/proc/cpuinfo data taken from a real world machine.

The real fix would be to generate at least the flags automatically based
on the selected CPU. Please do not submit this patch upstream until this
has happened.

Signed-off-by: Alexander Graf <agraf@suse.de>
[AF: Rebased for v1.6 and v1.7]
Signed-off-by: Andreas Färber <afaerber@suse.de>

linux-user: binfmt: support host binaries

When we have a working host binary equivalent for the guest binary we're
trying to run, let's just use that instead as it will be a lot faster.

Signed-off-by: Alexander Graf <agraf@suse.de>

PPC: KVM: Disable mmu notifier check

When using hugetlbfs (which is required for HV mode KVM on 970), we
check for MMU notifiers that on 970 can not be implemented properly.

So disable the check for mmu notifiers on PowerPC guests, making
KVM guests work there, even if possibly racy in some odd circumstances.

Signed-off-by: Bruce Rogers <brogers@suse.com>

linux-user: add binfmt wrapper for argv[0] handling

When using qemu's linux-user binaries through binfmt, argv[0] gets lost
along the execution because qemu only gets passed in the full file name
to the executable while argv[0] can be something completely different.

This breaks in some subtile situations, such as the grep and make test
suites.

This patch adds a wrapper binary called qemu-$TARGET-binfmt that can be
used with binfmt's P flag which passes the full path _and_ argv[0] to
the binfmt handler.

The binary would be smart enough to be versatile and only exist in the
system once, creating the qemu binary path names from its own argv[0].
However, this seemed like it didn't fit the make system too well, so
we're currently creating a new binary for each target archictecture.

CC: Reinhard Max <max@suse.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
[AF: Rebased onto new Makefile infrastructure, twice]
[AF: Updated for aarch64 for v2.0.0-rc1]
[AF: Rebased onto Makefile changes for v2.1.0-rc0]
[AF: Rebased onto script rewrite for v2.7.0-rc2 - to be fixed]
Signed-off-by: Andreas Färber <afaerber@suse.de>

qemu-cvs-ioctl_nodirection

the direction given in the ioctl should be correct so we can assume the
communication is uni-directional. The alsa developers did not like this
concept though and declared ioctls IOC_R and IOC_W even though they were
IOC_RW.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Ulrich Hecht <uli@suse.de>
[BR: minor edits to pass qemu's checkpatch script]
Signed-off-by: Bruce Rogers <brogers@suse.com>

qemu-cvs-ioctl_debug

Extends unsupported ioctl debug output.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Ulrich Hecht <uli@suse.de>
[BR: minor edits to pass qemu's checkpatch script]
Signed-off-by: Bruce Rogers <brogers@suse.com>

qemu-cvs-gettimeofday

No clue what this is for.

[BR: minor edits to pass qemu's checkpatch script]
Signed-off-by: Bruce Rogers <brogers@suse.com>

qemu-binfmt-conf: Modify default path

Change QEMU_PATH from /usr/local/bin to /usr/bin prefix.

Signed-off-by: Andreas Färber <afaerber@suse.de>

XXX dont dump core on sigabort

Signed-off-by: Bruce Rogers <brogers@suse.com>

net: vmxnet3: validate configuration values during activate (CVE-2021-20203)

Git-commit: 0000000000000000000000000000000000000000
References: bsc#1181639

While activating device in vmxnet3_acticate_device(), it does not
validate guest supplied configuration values against predefined
minimum - maximum limits. This may lead to integer overflow or
OOB access issues. Add checks to avoid it.

Fixes: CVE-2021-20203
Buglink: https://bugs.launchpad.net/qemu/+bug/1913873
Reported-by: Gaoning Pan <pgn@zju.edu.cn>
Signed-off-by: Prasad J Pandit <pjp@fedoraproject.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

mptsas: Remove unused MPTSASState 'pending' field (CVE-2021-3392)

Git-commit: 3791642c8d60029adf9b00bcb4e34d7d8a1aea4d

While processing SCSI i/o requests in mptsas_process_scsi_io_request(),
the Megaraid emulator appends new MPTSASRequest object 'req' to
the 's->pending' queue. In case of an error, this same object gets
dequeued in mptsas_free_request() only if SCSIRequest object
'req->sreq' is initialised. This may lead to a use-after-free issue.

Since s->pending is actually not used, simply remove it from
MPTSASState.

Cc: qemu-stable@nongnu.org
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reported-by: Cheolwoo Myung <cwmyung@snu.ac.kr>
Message-id: 20210419134247.1467982-1-f4bug@amsat.org
Message-Id: <20210416102243.1293871-1-mjt@msgid.tls.msk.ru>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Reported-by: Cheolwoo Myung <cwmyung@snu.ac.kr>
BugLink: https://bugs.launchpad.net/qemu/+bug/1914236
Fixes: e351b826112 ("hw: Add support for LSI SAS1068 (mptsas) device")
[PMD: Reworded description, added more tags]
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/isa/piix4: Migrate Reset Control Register

Git-commit: 62271205bcfaee440d06c06060ee79dac657caff

When adding the Reset register in commit 5790b757cfb we
forgot to migrate it.

While it is possible a VM using the PIIX4 is migrated just
after requesting a system shutdown, it is very unlikely.
However when restoring a migrated VM, we might have the
RCR bit #4 set on the stack and when the VM resume it
directly shutdowns.

Add a post_load() migration handler and set the default
RCR value to 0 for earlier versions, assuming the VM was
not going to shutdown before migration.

Fixes: 5790b757cfb ("piix4: Add the Reset Control Register")
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210324200334.729899-1-f4bug@amsat.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/isa/Kconfig: Add missing dependency VIA VT82C686 -> APM

Git-commit: 50fab4cc672233fee22fff2cf51543af57602c7d

TYPE_VIA_PM calls apm_init() in via_pm_realize(), so
requires APM to be selected.

Reported-by: BALATON Zoltan <balaton@eik.bme.hu>
Fixes: dd0ff8191ab ("isa: express SuperIO dependencies with Kconfig")
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210302080531.913802-1-f4bug@amsat.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/arm/virt-acpi-build: Fix GSIV values of the {GERR, Sync} interrupts

Git-commit: 0c38f607836af40921ea2b58676b7c4a9fe33bef

The GSIV values in SMMUv3 IORT node are not correct as they don't match
the SMMUIrq enumeration, which describes the IRQ<->PIN mapping used by
our emulated vSMMU.

Fixes: a703b4f6c1ee ("hw/arm/virt-acpi-build: Add smmuv3 node in IORT table")
Signed-off-by: Zenghui Yu <yuzenghui@huawei.com>
Acked-by: Eric Auger <eric.auger@redhat.com>
Message-id: 20210402084731.93-1-yuzenghui@huawei.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

cpu/core: Fix "help" of CPU core device types

Git-commit: 0b47ec4b95ad1952e55e639711d442f8ec6e1345

Calling qdev_get_machine() from a QOM instance_init function is
fragile because we can't be sure the machine object actually
exists. And this happens to break when passing ",help" on the
command line to get the list of properties for a CPU core
device types :

$ ./qemu-system-ppc64 -device power8_v2.0-spapr-cpu-core,help
qemu-system-ppc64: ../../hw/core/machine.c:1290:
 qdev_get_machine: Assertion `machine != NULL' failed.
Aborted (core dumped)

This used to work before QEMU 5.0, but commit 3df261b6676b
unwillingly introduced a subtle regression : the above command
line needs to create an instance but the instance_init function
of the base class calls qdev_get_machine() before
qemu_create_machine() has been called, which is a programming bug.

Use current_machine instead. It is okay to skip the setting of
nr_thread in this case since only its type is displayed.

Fixes: 3df261b6676b ("softmmu/vl.c: Handle '-cpu help' and '-device help' before 'no default machine'")
Reported-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Greg Kurz <groug@kaod.org>
Cc: peter.maydell@linaro.org
Message-Id: <20210409160339.500167-3-groug@kaod.org>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/block/fdc: Fix 'fallback' property on sysbus floppy disk controllers

Git-commit: da64789d3a16b2c5b5f1be9c75b00c2b8ae393a0

Setting the 'fallback' property corrupts the QOM instance state
(FDCtrlSysBus) because it accesses an incorrect offset (it uses
the offset of the FDCtrlISABus state).

Cc: qemu-stable@nongnu.org
Fixes: a73275dd6fc ("fdc: Add fallback option")
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210407133742.1680424-1-f4bug@amsat.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

block/rbd: Fix memory leak in qemu_rbd_co_create_opts()

Git-commit: b084b420d9d6347dede328fbcf18c8e4c695f7e8

When we allocate 'q_namespace', we forgot to set 'has_q_namespace'
to true. This can cause several issues, including a memory leak,
since qapi_free_BlockdevCreateOptions() does not deallocate that
memory, as reported by valgrind:

  13 bytes in 1 blocks are definitely lost in loss record 7 of 96
     at 0x4839809: malloc (vg_replace_malloc.c:307)
     by 0x48CEBB8: g_malloc (in /usr/lib64/libglib-2.0.so.0.6600.8)
     by 0x48E3FE3: g_strdup (in /usr/lib64/libglib-2.0.so.0.6600.8)
     by 0x180010: qemu_rbd_co_create_opts (rbd.c:446)
     by 0x1AE72C: bdrv_create_co_entry (block.c:492)
     by 0x241902: coroutine_trampoline (coroutine-ucontext.c:173)
     by 0x57530AF: ??? (in /usr/lib64/libc-2.32.so)
     by 0x1FFEFFFA6F: ???

Fix setting 'has_q_namespace' to true when we allocate 'q_namespace'.

Fixes: 19ae9ae014 ("block/rbd: Add support for ceph namespaces")
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Message-Id: <20210329150129.121182-3-sgarzare@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>
[BR: Modified subject to acheive unique patchname]

block/rbd: fix memory leak in qemu_rbd_connect()

Git-commit: c1c1f6cf511496b985cb9a1c536d59c9be7b9317

In qemu_rbd_connect(), 'mon_host' is allocated by qemu_rbd_mon_host()
using g_strjoinv(), but it's only freed in the error path, leaking
memory in the success path as reported by valgrind:

  80 bytes in 4 blocks are definitely lost in loss record 5,028 of 6,516
     at 0x4839809: malloc (vg_replace_malloc.c:307)
     by 0x5315BB8: g_malloc (in /usr/lib64/libglib-2.0.so.0.6600.8)
     by 0x532B6FF: g_strjoinv (in /usr/lib64/libglib-2.0.so.0.6600.8)
     by 0x87D07E: qemu_rbd_mon_host (rbd.c:538)
     by 0x87D07E: qemu_rbd_connect (rbd.c:562)
     by 0x87E1CE: qemu_rbd_open (rbd.c:740)
     by 0x840EB1: bdrv_open_driver (block.c:1528)
     by 0x8453A9: bdrv_open_common (block.c:1802)
     by 0x8453A9: bdrv_open_inherit (block.c:3444)
     by 0x8464C2: bdrv_open (block.c:3537)
     by 0x8108CD: qmp_blockdev_add (blockdev.c:3569)
     by 0x8EA61B: qmp_marshal_blockdev_add (qapi-commands-block-core.c:1086)
     by 0x90B528: do_qmp_dispatch_bh (qmp-dispatch.c:131)
     by 0x907EA4: aio_bh_poll (async.c:164)

Fix freeing 'mon_host' also when qemu_rbd_connect() ends correctly.

Fixes: 0a55679b4a5061f4d74bdb1a0e81611ba3390b00
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Message-Id: <20210329150129.121182-2-sgarzare@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

s390x: css: report errors from ccw_dstream_read/write

Git-commit: d895d25ae2bb8519aa715dd2a97f09d4a66b189d

ccw_dstream_read/write functions returned values are sometime
not taking into account and reported back to the upper level
of interpretation of CCW instructions.

It follows that accessing an invalid address does not trigger
a subchannel status program check to the guest as it should.

Let's test the return values of ccw_dstream_write[_buf] and
ccw_dstream_read[_buf] and report it to the caller.

Cc: qemu-stable@nongnu.org
Signed-off-by: Pierre Morel <pmorel@linux.ibm.com>
Acked-by: Halil Pasic <pasic@linux.ibm.com>
Message-Id: <1617899529-9329-2-git-send-email-pmorel@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

virtio-pci: compat page aligned ATS

Git-commit: d83f46d189a26fa32434139954d264326f199a45

Commit 4c70875372b8 ("pci: advertise a page aligned ATS") advertises
the page aligned via ATS capability (RO) to unbrek recent Linux IOMMU
drivers since 5.2. But it forgot the compat the capability which
breaks the migration from old machine type:

(qemu) qemu-kvm: get_pci_config_device: Bad config data: i=0x104 read:
0 device: 20 cmask: ff wmask: 0 w1cmask:0

This patch introduces a new parameter "x-ats-page-aligned" for
virtio-pci device and turns it on for machine type which is newer than
5.1.

Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: qemu-stable@nongnu.org
Fixes: 4c70875372b8 ("pci: advertise a page aligned ATS")
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20210406040330.11306-1-jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/virtio-pci Added AER capability.

Git-commit: fdfa3b1d6f9edd97c807df496a0d8e9ea49240da

Added AER capability for virtio-pci devices.
Also added property for devices, by default AER is disabled.

Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Message-Id: <20201203110713.204938-3-andrew@daynix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>
[BR: needed for stable commit d83f46d189a26fa32434139954d264326f199a45]

hw/virtio-pci Added counter for pcie capabilities offsets.

Git-commit: 06e97442420b03a1e0ff05e8eb554fac684ca736

Removed hardcoded offset for ats. Added cap offset counter
for future capabilities like AER.

Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Message-Id: <20201203110713.204938-2-andrew@daynix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>
[BR: needed for stable commit d83f46d189a26fa32434139954d264326f199a45]

target/xtensa: fix meson.build rule for xtensa cores

Git-commit: 84317d57e8c61ff68eeaa1f2de93472fa930a6a4

import_core.sh tries to change Makefile.objs when importing new xtensa
core, but that file no longer exists. Rewrite meson.build rule to pick
up all source files that match core-*.c pattern and drop commands that
change Makefile.objs.

Cc: qemu-stable@nongnu.org # v5.2.0
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

util: fix use-after-free in module_load_one

Git-commit: 64e16fbbf49ce81b37841480d14b0caf5753c98e

g_hash_table_add always retains ownership of the pointer passed in as
the key. Its return status merely indicates whether the added entry was
new, or replaced an existing entry. Thus key must never be freed after
this method returns.

Spotted by ASAN:

==2407186==ERROR: AddressSanitizer: heap-use-after-free on address 0x6020003ac4f0 at pc 0x7ffff766659c bp 0x7fffffffd1d0 sp 0x7fffffffc980
READ of size 1 at 0x6020003ac4f0 thread T0
    #0 0x7ffff766659b  (/lib64/libasan.so.6+0x8a59b)
    #1 0x7ffff6bfa843 in g_str_equal ../glib/ghash.c:2303
    #2 0x7ffff6bf8167 in g_hash_table_lookup_node ../glib/ghash.c:493
    #3 0x7ffff6bf9b78 in g_hash_table_insert_internal ../glib/ghash.c:1598
    #4 0x7ffff6bf9c32 in g_hash_table_add ../glib/ghash.c:1689
    #5 0x5555596caad4 in module_load_one ../util/module.c:233
    #6 0x5555596ca949 in module_load_one ../util/module.c:225
    #7 0x5555596ca949 in module_load_one ../util/module.c:225
    #8 0x5555596cbdf4 in module_load_qom_all ../util/module.c:349

Typical C bug...

Fixes: 90629122d2e ("module: use g_hash_table_add()")
Cc: qemu-stable@nongnu.org
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20210316134456.3243102-1-marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

s390x: modularize virtio-gpu-ccw

Git-commit: adcf33a504de29feb720736051dc32889314c9e6
References: bsc#1181103

Since the virtio-gpu-ccw device depends on the hw-display-virtio-gpu
module, which provides the type virtio-gpu-device, packaging the
hw-display-virtio-gpu module as a separate package that may or may not
be installed along with the qemu package leads to problems. Namely if
the hw-display-virtio-gpu is absent, qemu continues to advertise
virtio-gpu-ccw, but it aborts not only when one attempts using
virtio-gpu-ccw, but also when libvirtd's capability probing tries
to instantiate the type to introspect it.

Let us thus introduce a module named hw-s390x-virtio-gpu-ccw that
is going to provide the virtio-gpu-ccw device. The hw-s390x prefix
was chosen because it is not a portable device.

With virtio-gpu-ccw built as a module, the correct way to package a
modularized qemu is to require that hw-display-virtio-gpu must be
installed whenever the module hw-s390x-virtio-gpu-ccw.

Signed-off-by: Halil Pasic <pasic@linux.ibm.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Tested-by: Halil Pasic <pasic@linux.ibm.com>
Message-Id: <20210317095622.2839895-4-kraxel@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

s390x: add have_virtio_ccw

Git-commit: 2dd9d8cfb4f3bd30d9cdfc2edba5cb7ee5917f4b
References: bsc#1181103

Introduce a symbol which can be used to prevent ccw modules
being loaded into system emulators without ccw support.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Tested-by: Halil Pasic <pasic@linux.ibm.com>
Message-Id: <20210317095622.2839895-3-kraxel@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

s390x: move S390_ADAPTER_SUPPRESSIBLE

Git-commit: d4c603d7be2e4173252c5b55e62d30ddd26edaca
References: bsc#1181103

The definition S390_ADAPTER_SUPPRESSIBLE was moved to "cpu.h", per
suggestion of Thomas Huth. From interface design perspective, IMHO, not
a good thing as it belongs to the public interface of
css_register_io_adapters(). We did this because CONFIG_KVM requeires
NEED_CPU_H and Thomas, and other commenters did not like the
consequences of that.

Moving the interrupt related declarations to s390_flic.h was suggested
by Cornelia Huck.

Signed-off-by: Halil Pasic <pasic@linux.ibm.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Tested-by: Halil Pasic <pasic@linux.ibm.com>
Message-Id: <20210317095622.2839895-2-kraxel@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/arm/virt: Disable pl011 clock migration if needed

Git-commit: e6fa978d8343ec7cf20b9c8b2dcb390646242457

A clock is added by commit aac63e0e6ea3 ("hw/char/pl011: add a clock
input") since v5.2.0 which corresponds to virt-5.2 machine type. It
causes backwards migration failure from upstream to downstream (v5.1.0)
when the machine type is specified with virt-5.1.

This fixes the issue by following instructions from section "Connecting
subsections to properties" in docs/devel/migration.rst. With this applied,
the PL011 clock is migrated based on the machine type.

   virt-5.2 or newer:  migration
   virt-5.1 or older:  non-migration

Cc: qemu-stable@nongnu.org # v5.2.0+
Fixes: aac63e0e6ea3 ("hw/char/pl011: add a clock input")
Suggested-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Gavin Shan <gshan@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Message-id: 20210318023801.18287-1-gshan@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

xen-block: Fix removal of backend instance via xenstore

Git-commit: b807ca3fa0ca29ec015adcf4045e716337cd3635

Whenever a Xen block device is detach via xenstore, the image
associated with it remained open by the backend QEMU and an error is
logged:
    qemu-system-i386: failed to destroy drive: Node xvdz-qcow2 is in use

This happened since object_unparent() doesn't immediately frees the
object and thus keep a reference to the node we are trying to free.
The reference is hold by the "drive" property and the call
xen_block_drive_destroy() fails.

In order to fix that, we call drain_call_rcu() to run the callback
setup by bus_remove_child() via object_unparent().

Fixes: 2d24a6466154 ("device-core: use RCU for list of children of a bus")

Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Reviewed-by: Paul Durrant <paul@xen.org>
Message-Id: <20210308143232.83388-1-anthony.perard@citrix.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/sd: sdhci: Reset the data pointer of s->fifo_buffer[] when a different block size is programmed

Git-commit: cffb446e8fd19a14e1634c7a3a8b07be3f01d5c9
References: bsc#1175144, CVE-2020-17380, bsc#1176681, CVE-2020-25085
References: bsc#1182282, CVE-2021-3409

If the block size is programmed to a different value from the
previous one, reset the data pointer of s->fifo_buffer[] so that
s->fifo_buffer[] can be filled in using the new block size in
the next transfer.

With this fix, the following reproducer:

outl 0xcf8 0x80001010
outl 0xcfc 0xe0000000
outl 0xcf8 0x80001001
outl 0xcfc 0x06000000
write 0xe000002c 0x1 0x05
write 0xe0000005 0x1 0x02
write 0xe0000007 0x1 0x01
write 0xe0000028 0x1 0x10
write 0x0 0x1 0x23
write 0x2 0x1 0x08
write 0xe000000c 0x1 0x01
write 0xe000000e 0x1 0x20
write 0xe000000f 0x1 0x00
write 0xe000000c 0x1 0x32
write 0xe0000004 0x2 0x0200
write 0xe0000028 0x1 0x00
write 0xe0000003 0x1 0x40

cannot be reproduced with the following QEMU command line:

$ qemu-system-x86_64 -nographic -machine accel=qtest -m 512M \
      -nodefaults -device sdhci-pci,sd-spec-version=3 \
      -drive if=sd,index=0,file=null-co://,format=raw,id=mydrive \
      -device sd-card,drive=mydrive -qtest stdio

Cc: qemu-stable@nongnu.org
Fixes: CVE-2020-17380
Fixes: CVE-2020-25085
Fixes: CVE-2021-3409
Fixes: d7dfca0807a0 ("hw/sdhci: introduce standard SD host controller")
Reported-by: Alexander Bulekov <alxndr@bu.edu>
Reported-by: Cornelius Aschermann (Ruhr-Universität Bochum)
Reported-by: Sergej Schumilo (Ruhr-Universität Bochum)
Reported-by: Simon Wörner (Ruhr-Universität Bochum)
Buglink: https://bugs.launchpad.net/qemu/+bug/1892960
Buglink: https://bugs.launchpad.net/qemu/+bug/1909418
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1928146
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Message-Id: <20210303122639.20004-6-bmeng.cn@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/sd: sdhci: Limit block size only when SDHC_BLKSIZE register is writable

Git-commit: 5cd7aa3451b76bb19c0f6adc2b931f091e5d7fcd
References: bsc#1175144, CVE-2020-17380, bsc#1176681, CVE-2020-25085
References: bsc#1182282, CVE-2021-3409

The codes to limit the maximum block size is only necessary when
SDHC_BLKSIZE register is writable.

Tested-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Message-Id: <20210303122639.20004-5-bmeng.cn@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/sd: sdhci: Correctly set the controller status for ADMA

Git-commit: bc6f28995ff88f5d82c38afcfd65406f0ae375aa
References: bsc#1175144, CVE-2020-17380, bsc#1176681, CVE-2020-25085
References: bsc#1182282, CVE-2021-3409

When an ADMA transfer is started, the codes forget to set the
controller status to indicate a transfer is in progress.

With this fix, the following 2 reproducers:

https://paste.debian.net/plain/1185136
https://paste.debian.net/plain/1185141

cannot be reproduced with the following QEMU command line:

$ qemu-system-x86_64 -nographic -machine accel=qtest -m 512M \
      -nodefaults -device sdhci-pci,sd-spec-version=3 \
      -drive if=sd,index=0,file=null-co://,format=raw,id=mydrive \
      -device sd-card,drive=mydrive -qtest stdio

Cc: qemu-stable@nongnu.org
Fixes: CVE-2020-17380
Fixes: CVE-2020-25085
Fixes: CVE-2021-3409
Fixes: d7dfca0807a0 ("hw/sdhci: introduce standard SD host controller")
Reported-by: Alexander Bulekov <alxndr@bu.edu>
Reported-by: Cornelius Aschermann (Ruhr-Universität Bochum)
Reported-by: Sergej Schumilo (Ruhr-Universität Bochum)
Reported-by: Simon Wörner (Ruhr-Universität Bochum)
Buglink: https://bugs.launchpad.net/qemu/+bug/1892960
Buglink: https://bugs.launchpad.net/qemu/+bug/1909418
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1928146
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Message-Id: <20210303122639.20004-4-bmeng.cn@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/sd: sdhci: Don't write to SDHC_SYSAD register when transfer is in progress

Git-commit: 8be45cc947832b3c02144c9d52921f499f2d77fe
References: bsc#1175144, CVE-2020-17380, bsc#1176681, CVE-2020-25085
References: bsc#1182282, CVE-2021-3409

Per "SD Host Controller Standard Specification Version 7.00"
chapter 2.2.1 SDMA System Address Register:

This register can be accessed only if no transaction is executing
(i.e., after a transaction has stopped).

With this fix, the following reproducer:

outl 0xcf8 0x80001010
outl 0xcfc 0xfbefff00
outl 0xcf8 0x80001001
outl 0xcfc 0x06000000
write 0xfbefff2c 0x1 0x05
write 0xfbefff0f 0x1 0x37
write 0xfbefff0a 0x1 0x01
write 0xfbefff0f 0x1 0x29
write 0xfbefff0f 0x1 0x02
write 0xfbefff0f 0x1 0x03
write 0xfbefff04 0x1 0x01
write 0xfbefff05 0x1 0x01
write 0xfbefff07 0x1 0x02
write 0xfbefff0c 0x1 0x33
write 0xfbefff0e 0x1 0x20
write 0xfbefff0f 0x1 0x00
write 0xfbefff2a 0x1 0x01
write 0xfbefff0c 0x1 0x00
write 0xfbefff03 0x1 0x00
write 0xfbefff05 0x1 0x00
write 0xfbefff2a 0x1 0x02
write 0xfbefff0c 0x1 0x32
write 0xfbefff01 0x1 0x01
write 0xfbefff02 0x1 0x01
write 0xfbefff03 0x1 0x01

cannot be reproduced with the following QEMU command line:

$ qemu-system-x86_64 -nographic -machine accel=qtest -m 512M \
       -nodefaults -device sdhci-pci,sd-spec-version=3 \
       -drive if=sd,index=0,file=null-co://,format=raw,id=mydrive \
       -device sd-card,drive=mydrive -qtest stdio

Cc: qemu-stable@nongnu.org
Fixes: CVE-2020-17380
Fixes: CVE-2020-25085
Fixes: CVE-2021-3409
Fixes: d7dfca0807a0 ("hw/sdhci: introduce standard SD host controller")
Reported-by: Alexander Bulekov <alxndr@bu.edu>
Reported-by: Cornelius Aschermann (Ruhr-Universität Bochum)
Reported-by: Sergej Schumilo (Ruhr-Universität Bochum)
Reported-by: Simon Wörner (Ruhr-Universität Bochum)
Buglink: https://bugs.launchpad.net/qemu/+bug/1892960
Buglink: https://bugs.launchpad.net/qemu/+bug/1909418
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1928146
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Message-Id: <20210303122639.20004-3-bmeng.cn@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/sd: sdhci: Don't transfer any data when command time out

Git-commit: b263d8f928001b5cfa2a993ea43b7a5b3a1811e8
References: bsc#1175144, CVE-2020-17380, bsc#1176681, CVE-2020-25085
References: bsc#1182282, CVE-2021-3409

At the end of sdhci_send_command(), it starts a data transfer if the
command register indicates data is associated. But the data transfer
should only be initiated when the command execution has succeeded.

With this fix, the following reproducer:

outl 0xcf8 0x80001810
outl 0xcfc 0xe1068000
outl 0xcf8 0x80001804
outw 0xcfc 0x7
write 0xe106802c 0x1 0x0f
write 0xe1068004 0xc 0x2801d10101fffffbff28a384
write 0xe106800c 0x1f 0x9dacbbcad9e8f7061524334251606f7e8d9cabbac9d8e7f60514233241505f
write 0xe1068003 0x28 0x80d000251480d000252280d000253080d000253e80d000254c80d000255a80d000256880d0002576
write 0xe1068003 0x1 0xfe

cannot be reproduced with the following QEMU command line:

$ qemu-system-x86_64 -nographic -M pc-q35-5.0 \
      -device sdhci-pci,sd-spec-version=3 \
      -drive if=sd,index=0,file=null-co://,format=raw,id=mydrive \
      -device sd-card,drive=mydrive \
      -monitor none -serial none -qtest stdio

Cc: qemu-stable@nongnu.org
Fixes: CVE-2020-17380
Fixes: CVE-2020-25085
Fixes: CVE-2021-3409
Fixes: d7dfca0807a0 ("hw/sdhci: introduce standard SD host controller")
Reported-by: Alexander Bulekov <alxndr@bu.edu>
Reported-by: Cornelius Aschermann (Ruhr-Universität Bochum)
Reported-by: Sergej Schumilo (Ruhr-Universität Bochum)
Reported-by: Simon Wörner (Ruhr-Universität Bochum)
Buglink: https://bugs.launchpad.net/qemu/+bug/1892960
Buglink: https://bugs.launchpad.net/qemu/+bug/1909418
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1928146
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Tested-by: Alexander Bulekov <alxndr@bu.edu>
Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Message-Id: <20210303122639.20004-2-bmeng.cn@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/sd: sd: Actually perform the erase operation

Git-commit: 818a5cdcfcf0a55d60b59b2cb74482ef4ba6b205
References: bsc#1175144, CVE-2020-17380, bsc#1176681, CVE-2020-25085
References: bsc#1182282, CVE-2021-3409

At present the sd_erase() does not erase the requested range of card
data to 0xFFs. Let's make the erase operation actually happen.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Message-Id: <1613811493-58815-1-git-send-email-bmeng.cn@gmail.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/sd: sd: Fix build error when DEBUG_SD is on

Git-commit: a78d9f27b73de3c42f376540bd1d1d0570eb1fa3
References: bsc#1175144, CVE-2020-17380, bsc#1176681, CVE-2020-25085
References: bsc#1182282, CVE-2021-3409

"qemu-common.h" should be included to provide the forward declaration
of qemu_hexdump() when DEBUG_SD is on.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210228050609.24779-1-bmeng.cn@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/sd: sd: Skip write protect groups check in sd_erase() for high capacity cards

Git-commit: 2473dc4022458dcc05ec367ce97edbef29d7e50c
References: bsc#1175144, CVE-2020-17380, bsc#1176681, CVE-2020-25085
References: bsc#1182282, CVE-2021-3409

High capacity cards don't support write protection hence we should
not perform the write protect groups check in sd_erase() for them.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210216150225.27996-6-bmeng.cn@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/sd: sd: Move the sd_block_{read, write} and macros ahead

Git-commit: ce6ea2efc5fb308ebf11339276f60215fe0ec44c
References: bsc#1175144, CVE-2020-17380, bsc#1176681, CVE-2020-25085
References: bsc#1182282, CVE-2021-3409

These APIs and macros may be referenced by functions that are
currently before them. Move them ahead a little bit.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210216150225.27996-5-bmeng.cn@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Bruce Rogers <brogers@suse.com>

lan9118: switch to use qemu_receive_packet() for loopback

Git-commit: 37cee01784ff0df13e5209517e1b3594a5e792d1

This patch switches to use qemu_receive_packet() which can detect
reentrancy and return early.

This is intended to address CVE-2021-3416.

Cc: Prasad J Pandit <ppandit@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com
Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

cadence_gem: switch to use qemu_receive_packet() for loopback

Git-commit: e73adfbeec9d4e008630c814759052ed945c3fed

This patch switches to use qemu_receive_packet() which can detect
reentrancy and return early.

This is intended to address CVE-2021-3416.

Cc: Prasad J Pandit <ppandit@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

pcnet: switch to use qemu_receive_packet() for loopback

Git-commit: 99ccfaa1edafd79f7a3a0ff7b58ae4da7c514928

This patch switches to use qemu_receive_packet() which can detect
reentrancy and return early.

This is intended to address CVE-2021-3416.

Cc: Prasad J Pandit <ppandit@redhat.com>
Cc: qemu-stable@nongnu.org
Buglink: https://bugs.launchpad.net/qemu/+bug/1917085
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com
Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

rtl8139: switch to use qemu_receive_packet() for loopback

Git-commit: 5311fb805a4403bba024e83886fa0e7572265de4
References: bsc#1182968, CVE-2021-3416

This patch switches to use qemu_receive_packet() which can detect
reentrancy and return early.

This is intended to address CVE-2021-3416.

Cc: Prasad J Pandit <ppandit@redhat.com>
Cc: qemu-stable@nongnu.org
Buglink: https://bugs.launchpad.net/qemu/+bug/1910826
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com
Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

tx_pkt: switch to use qemu_receive_packet_iov() for loopback

Git-commit: 8c552542b81e56ff532dd27ec6e5328954bdda73

This patch switches to use qemu_receive_receive_iov() which can detect
reentrancy and return early.

This is intended to address CVE-2021-3416.

Cc: Prasad J Pandit <ppandit@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

sungem: switch to use qemu_receive_packet() for loopback

Git-commit: 8c92060d3c0248bd4d515719a35922cd2391b9b4

This patch switches to use qemu_receive_packet() which can detect
reentrancy and return early.

This is intended to address CVE-2021-3416.

Cc: Prasad J Pandit <ppandit@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

msf2-mac: switch to use qemu_receive_packet() for loopback

Git-commit: 26194a58f4eb83c5bdf4061a1628508084450ba1

This patch switches to use qemu_receive_packet() which can detect
reentrancy and return early.

This is intended to address CVE-2021-3416.

Cc: Prasad J Pandit <ppandit@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

dp8393x: switch to use qemu_receive_packet() for loopback packet

Git-commit: 331d2ac9ea307c990dc86e6493e8f0c48d14bb33

This patch switches to use qemu_receive_packet() which can detect
reentrancy and return early.

This is intended to address CVE-2021-3416.

Cc: Prasad J Pandit <ppandit@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

e1000: switch to use qemu_receive_packet() for loopback

Git-commit: 1caff0340f49c93d535c6558a5138d20d475315c

This patch switches to use qemu_receive_packet() which can detect
reentrancy and return early.

This is intended to address CVE-2021-3416.

Cc: Prasad J Pandit <ppandit@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

net: introduce qemu_receive_packet()

Git-commit: 705df5466c98f3efdd2b68d3b31dad86858acad7
References: bsc#1182968, CVE-2021-3416
Some NIC supports loopback mode and this is done by calling
nc->info->receive() directly which in fact suppresses the effort of
reentrancy check that is done in qemu_net_queue_send().

Unfortunately we can't use qemu_net_queue_send() here since for
loopback there's no sender as peer, so this patch introduce a
qemu_receive_packet() which is used for implementing loopback mode
for a NIC with this check.

NIC that supports loopback mode will be converted to this helper.

This is intended to address CVE-2021-3416.

Cc: Prasad J Pandit <ppandit@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Cc: qemu-stable@nongnu.org
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

e1000: fail early for evil descriptor

Git-commit: 3de46e6fc489c52c9431a8a832ad8170a7569bd8
References: bsc#1182577, CVE-2021-20257

During procss_tx_desc(), driver can try to chain data descriptor with
legacy descriptor, when will lead underflow for the following
calculation in process_tx_desc() for bytes:

            if (tp->size + bytes > msh)
                bytes = msh - tp->size;

This will lead a infinite loop. So check and fail early if tp->size if
greater or equal to msh.

Reported-by: Alexander Bulekov <alxndr@bu.edu>
Reported-by: Cheolwoo Myung <cwmyung@snu.ac.kr>
Reported-by: Ruhr-University Bochum <bugs-syssec@rub.de>
Cc: Prasad J Pandit <ppandit@redhat.com>
Cc: qemu-stable@nongnu.org
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

vhost-user-blk: fix blkcfg->num_queues endianness

Git-commit: 535255b43898d2e96744057eb86f8497d4d7a461

Treat the num_queues field as virtio-endian. On big-endian hosts the
vhost-user-blk num_queues field was in the wrong endianness.

Move the blkcfg.num_queues store operation from realize to
vhost_user_blk_update_config() so feature negotiation has finished and
we know the endianness of the device. VIRTIO 1.0 devices are
little-endian, but in case someone wants to use legacy VIRTIO we support
all endianness cases.

Cc: qemu-stable@nongnu.org
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20210223144653.811468-2-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

lsilogic: Use PCIDevice::exit instead of DeviceState::unrealize

Git-commit: faabca42cc4ff51110116dfe44d420c668b4d8d8

PCI_DEVICE has overwritten DeviceState::unrealize (pci_qdev_unrealize).
However, LSI53C895A, which is a subclass of PCI_DEVICE, overwrites it
again and doesn't save the parent's implementation so the PCI_DEVICE's
implementation of DeviceState::unrealize will never be called when
unrealize a LSI53C895A device.  And it will lead to memory leak and
unplug failure.

For a PCI device, it's better to implement PCIDevice::exit instead of
DeviceState::unrealize.  So let's change to use PCIDevice::exit.

Fixes: a8632434c7e9 ("lsi: implement I/O memory space for Memory Move instructions")
Cc: qemu-stable@nongnu.org
Signed-off-by: Peng Liang <liangpeng10@huawei.com>
Message-Id: <20210302133016.1221081-1-liangpeng10@huawei.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

s390x/pci: restore missing Query PCI Function CLP data

Git-commit: 403af209db8c030ed1e000640cd3cd80c6882883
References: bsc#1183372

Some CLP response data was accidentally dropped when fixing endianness
issues with the Query PCI Function CLP response.  All of these values are
sent as 0s to the guest for emulated devices, so the impact is only
observed on passthrough devices.

Fixes: a4e2fff1b104 ("s390x/pci: fix endianness issues")
Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
Message-Id: <1613681609-9349-1-git-send-email-mjrosato@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

hw/s390x: fix build for virtio-9p-ccw

Git-commit: 24056cbfd577fd219d55c03f69df66e6351456e7
References: bsc#1182496

Commit  2c44220d05 ("meson: convert hw/arch*"), which migrated the old
Makefile.objs to meson.build  accidentally excluded virtio-ccw-9p.c and
thus the virtio-9p-ccw device from the build (and potentially also
included the file virtio-ccw-blk.c twice in the source set). And since
CONFIG_VIRTFS can't be used the way it was used here (see commit
2c9dce0196 ("meson: do not use CONFIG_VIRTFS")), the preconditions have
to be written differently.

Let's fix this!

Signed-off-by: Halil Pasic <pasic@linux.ibm.com>
Fixes: 2c44220d05 ("meson: convert hw/arch*")
Reported-by: Jakob Naucke <jakob.naucke@ibm.com>
Cc: qemu-stable@nongnu.org
[BR: back out the part which depends on the have_virtfs change]
Signed-off-by: Bruce Rogers <brogers@suse.com>

virtiofs: drop remapped security.capability xattr as needed

Git-commit: e586edcb410543768ef009eaa22a2d9dd4a53846
References: bsc#1183373, CVE-2021-20263

On Linux, the 'security.capability' xattr holds a set of
capabilities that can change when an executable is run, giving
a limited form of privilege escalation to those programs that
the writer of the file deemed worthy.

Any write causes the 'security.capability' xattr to be dropped,
stopping anyone from gaining privilege by modifying a blessed
file.

Fuse relies on the daemon to do this dropping, and in turn the
daemon relies on the host kernel to drop the xattr for it.  However,
with the addition of -o xattrmap, the xattr that the guest
stores its capabilities in is now not the same as the one that
the host kernel automatically clears.

Where the mapping changes 'security.capability', explicitly clear
the remapped name to preserve the same behaviour.

This bug is assigned CVE-2021-20263.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Bruce Rogers <brogers@suse.com>

i386/acpi: restore device paths for pre-5.1 vms

Git-commit: 0a343a5add75f9f90c65e932863d57ddbcb28f5c

After fixing the _UID value for the primary PCI root bridge in
af1b80ae it was discovered that this change updates Windows
configuration in an incompatible way causing network configuration
failure unless DHCP is used. More details provided on the list:

https://lists.gnu.org/archive/html/qemu-devel/2021-02/msg08484.html

This change reverts the _UID update from 1 to 0 for q35 and i440fx
VMs before version 5.2 to maintain the original behaviour when
upgrading.

Cc: qemu-stable@nongnu.org
Cc: qemu-devel@nongnu.org
Reported-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Suggested-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Vitaly Cheptsov <cheptsov@ispras.ru>
Message-Id: <20210301195919.9333-1-cheptsov@ispras.ru>
Tested-by: Thomas Lamprecht <t.lamprecht@proxmox.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>
Fixes: af1b80ae56c9 ("i386/acpi: fix inconsistent QEMU/OVMF device paths")
Signed-off-by: Bruce Rogers <brogers@suse.com>

spice-app: avoid crash when core spice module doesn't loaded

git-commit: 122e4ef6dea14a078a860ca253852e18ddebb8e2

When qemu is built with modules, but a given module doesn't load
qemu should handle that gracefully. When ui-spice-core.so isn't
able to be loaded and qemu is invoked with -display spice-app or
-spice, qemu will dereference a null pointer. With this change we
check the pointer before dereferencing and error out in a normal
way.

Signed-off-by: Bruce Rogers <brogers@suse.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210213032318.346093-1-brogers@suse.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>