docs: Add documentation for iommu bypass

Signed-off-by: Xingang Wang <wangxingang5@huawei.com>
Message-Id: <1625748919-52456-10-git-send-email-wangxingang5@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/i386/acpi-build: Add IVRS support to bypass iommu

Check bypass_iommu to exclude the devices which will bypass iommu.

Signed-off-by: Xingang Wang <wangxingang5@huawei.com>
Message-Id: <1625748919-52456-9-git-send-email-wangxingang5@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/i386/acpi-build: Add DMAR support to bypass iommu

In DMAR table, the drhd is set to cover all PCI devices when intel_iommu
is on. To support bypass iommu feature, we need to walk the PCI bus with
bypass_iommu disabled and add explicit scope data in DMAR drhd structure.

/mnt/sdb/wxg/qemu-next/qemu/build/x86_64-softmmu/qemu-system-x86_64 \
 -machine q35,accel=kvm,default_bus_bypass_iommu=true \
 -cpu host \
 -m 16G \
 -smp 36,sockets=2,cores=18,threads=1 \
 -device pxb-pcie,bus_nr=0x10,id=pci.10,bus=pcie.0,addr=0x3 \
 -device pxb-pcie,bus_nr=0x20,id=pci.20,bus=pcie.0,addr=0x4,bypass_iommu=true \
 -device pcie-root-port,port=0x1,chassis=1,id=pci.11,bus=pci.10,addr=0x0 \
 -device pcie-root-port,port=0x2,chassis=2,id=pci.21,bus=pci.20,addr=0x0 \
 -device virtio-scsi-pci,id=scsi0,bus=pci.11,addr=0x0 \
 -device virtio-scsi-pci,id=scsi1,bus=pci.21,addr=0x0 \
 -drive file=/mnt/sdb/wxg/fedora-48g.qcow2,format=qcow2,if=none,id=drive-scsi0-0-0-0,cache=none,aio=native \
 -device scsi-hd,bus=scsi1.0,channel=0,scsi-id=0,lun=0,drive=drive-scsi0-0-0-0,id=scsi0-0-0-0,bootindex=1 \
 -device intel-iommu \
 -nographic \

And we get the guest configuration:

~ lspci -vt
-+-[0000:20]---00.0-[21]----00.0  Red Hat, Inc. Virtio SCSI
 +-[0000:10]---00.0-[11]----00.0  Red Hat, Inc. Virtio SCSI
 \-[0000:00]-+-00.0  Intel Corporation 82G33/G31/P35/P31 Express DRAM Controller
             +-01.0  Device 1234:1111
             +-02.0  Intel Corporation 82574L Gigabit Network Connection
             +-03.0  Red Hat, Inc. QEMU PCIe Expander bridge
             +-04.0  Red Hat, Inc. QEMU PCIe Expander bridge
             +-1f.0  Intel Corporation 82801IB (ICH9) LPC Interface Controller
             +-1f.2  Intel Corporation 82801IR/IO/IH (ICH9R/DO/DH) 6 port SATA Controller [AHCI mode]
             \-1f.3  Intel Corporation 82801I (ICH9 Family) SMBus Controller

With bypass_iommu enabled on root bus, the attached devices will bypass iommu:

/sys/class/iommu/dmar0
├── devices
│   ├── 0000:10:00.0 -> ../../../../pci0000:10/0000:10:00.0
│   └── 0000:11:00.0 -> ../../../../pci0000:10/0000:10:00.0/0000:11:00.0

Signed-off-by: Xingang Wang <wangxingang5@huawei.com>
Message-Id: <1625748919-52456-8-git-send-email-wangxingang5@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/arm/virt-acpi-build: Add IORT support to bypass SMMUv3

When we build IORT table with SMMUv3 and bypass iommu feature enabled,
we can no longer setup one map from RC to SMMUv3 covering the whole RIDs.
We need to walk the PCI bus and check whether the root bus will bypass
iommu, setup RC -> SMMUv3 -> ITS map for RC which will not bypass iommu.

When a SMMUv3 node exist, we setup the idmap from SMMUv3 to ITS
covering the whole RIDs, and only modify the map from RC to SMMUv3.
We build RC -> SMMUv3 -> ITS map for root bus with bypass_iommu
disabled, and build idmap from RC to ITS directly for the rest of
the whole RID space.

For example we run qemu with command line:

qemu/build/aarch64-softmmu/qemu-system-aarch64 \
 -kernel arch/arm64/boot/Image \
 -enable-kvm \
 -cpu host \
 -m 8G \
 -smp 8,sockets=2,cores=4,threads=1 \
 -machine virt,kernel_irqchip=on,gic-version=3,iommu=smmuv3,default_bus_bypass_iommu=true \
 -drive file=./QEMU_EFI-pflash.raw,if=pflash,format=raw,unit=0,readonly=on \
 -device pxb-pcie,bus_nr=0x10,id=pci.10,bus=pcie.0,addr=0x3.0x1 \
 -device pxb-pcie,bus_nr=0x20,id=pci.20,bus=pcie.0,addr=0x3.0x2,bypass_iommu=true \
 -device pcie-root-port,port=0x20,chassis=1,id=pci.1,bus=pcie.0,addr=0x2 \
 -device pcie-root-port,port=0x20,chassis=11,id=pci.11,bus=pci.10,addr=0x1 \
 -device pcie-root-port,port=0x20,chassis=21,id=pci.21,bus=pci.20,addr=0x1 \
 -device virtio-scsi-pci,id=scsi0,bus=pci.1,addr=0x1 \
 -device virtio-scsi-pci,id=scsi1,bus=pci.11,addr=0x1 \
 -device virtio-scsi-pci,id=scsi2,bus=pci.21,addr=0x1 \
 -initrd /mnt/davinci/wxg/kill-linux/rootfs/mfs.cpio.gz \
 -nographic \
 -append "rdinit=init console=ttyAMA0 earlycon=pl011,0x9000000 nokaslr" \

And we get guest configuration:

-+-[0000:20]---01.0-[21]--
 +-[0000:10]---01.0-[11]--
 \-[0000:00]-+-00.0  Device 1b36:0008
             +-01.0  Device 1af4:1000
             \-02.0-[01]--

With bypass_iommu enabled, the attached devices will bypass iommu.

/sys/class/iommu/smmu3.0x0000000009050000/
|-- device -> ../../../arm-smmu-v3.0.auto
|-- devices
|   `-- 0000:10:01.0 -> ../../../../../pci0000:10/0000:10:01.0

Signed-off-by: Xingang Wang <wangxingang5@huawei.com>
Message-Id: <1625748919-52456-7-git-send-email-wangxingang5@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/pci: Add pci_bus_range() to get PCI bus number range

This helps to get the min and max bus number of a PCI bus hierarchy.

Signed-off-by: Xingang Wang <wangxingang5@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Message-Id: <1625748919-52456-6-git-send-email-wangxingang5@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/i386: Add a default_bus_bypass_iommu pc machine option

Add a default_bus_bypass_iommu pc machine option to enable/disable
bypass_iommu for default root bus. The option is disabled by default
and can be enabled with:
$QEMU -machine q35,default_bus_bypass_iommu=true

Signed-off-by: Xingang Wang <wangxingang5@huawei.com>
Message-Id: <1625748919-52456-5-git-send-email-wangxingang5@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/arm/virt: Add default_bus_bypass_iommu machine option

Add a default_bus_bypass_iommu machine option to enable/disable
bypass_iommu for default root bus. The option is disabled by
default and can be enabled with:
$QEMU -machine virt,iommu=smmuv3,default_bus_bypass_iommu=true

Signed-off-by: Xingang Wang <wangxingang5@huawei.com>
Message-Id: <1625748919-52456-4-git-send-email-wangxingang5@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/pxb: Add a bypass iommu property

Add a bypass_iommu property for pci_expander_bridge, the property
is used to indicate whether pxb root bus will bypass iommu. By
default the bypass_iommu is disabled, and it can be enabled with:
qemu -device pxb-pcie,bus_nr=0x10,addr=0x1,bypass_iommu=true

Signed-off-by: Xingang Wang <wangxingang5@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Message-Id: <1625748919-52456-3-git-send-email-wangxingang5@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/pci/pci_host: Allow PCI host to bypass iommu

Add a new bypass_iommu property for PCI host and use it to check
whether devices attached to the PCI root bus will bypass iommu.
In pci_device_iommu_address_space(), check the property and
avoid getting iommu address space for devices bypass iommu.

Signed-off-by: Xingang Wang <wangxingang5@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Message-Id: <1625748919-52456-2-git-send-email-wangxingang5@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

docs: Add '-device intel-iommu' entry

The parameters of intel-iommu device are non-trivial to understand.  Add an
entry for it so that people can reference to it when using.

There're actually a few more options there, but I hide them explicitly because
they shouldn't be used by normal QEMU users.

Cc: Chao Yang <chayang@redhat.com>
Cc: Lei Yang <leiyang@redhat.com>
Cc: Jing Zhao <jinzhao@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20210707154114.197580-1-peterx@redhat.com>
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/virtio: add vhost-user-i2c-pci boilerplate

This allows is to instantiate a vhost-user-i2c device as part of a PCI
bus. It is mostly boilerplate which looks pretty similar to the
vhost-user-fs-pci device.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Message-Id: <8a083eaa57d93feaab12acd1f94b225879212f20.1625806763.git.viresh.kumar@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/virtio: add boilerplate for vhost-user-i2c device

This creates the QEMU side of the vhost-user-i2c device which connects
to the remote daemon. It is based of vhost-user-fs code.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Message-Id: <e80591b52fea4b51631818bb92a798a3daf90399.1625806763.git.viresh.kumar@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

bios-tables-test: Update golden binaries

Add ACPI hot-plug registers to DSDT Q35 tables.
Changes in the tables:

+    Scope (_SB.PCI0)
+    {
+        OperationRegion (PCST, SystemIO, 0x0CC4, 0x08)
+        Field (PCST, DWordAcc, NoLock, WriteAsZeros)
+        {
+            PCIU,   32,
+            PCID,   32
+        }
+
+        OperationRegion (SEJ, SystemIO, 0x0CCC, 0x04)
+        Field (SEJ, DWordAcc, NoLock, WriteAsZeros)
+        {
+            B0EJ,   32
+        }
+
+        OperationRegion (BNMR, SystemIO, 0x0CD4, 0x08)
+        Field (BNMR, DWordAcc, NoLock, WriteAsZeros)
+        {
+            BNUM,   32,
+            PIDX,   32
+        }
+
+        Mutex (BLCK, 0x00)
+        Method (PCEJ, 2, NotSerialized)
+        {
+            Acquire (BLCK, 0xFFFF)
+            BNUM = Arg0
+            B0EJ = (One << Arg1)
+            Release (BLCK)
+            Return (Zero)
+        }
+
+        Method (AIDX, 2, NotSerialized)
+        {
+            Acquire (BLCK, 0xFFFF)
+            BNUM = Arg0
+            PIDX = (One << Arg1)
+            Local0 = PIDX /* \_SB_.PCI0.PIDX */
+            Release (BLCK)
+            Return (Local0)
+        }
+
+        Method (PDSM, 6, Serialized)
+        {
+            If ((Arg0 == ToUUID ("e5c937d0-3553-4d7a-9117-ea4d19c3434d") /* Device Labeling Interface */))
+            {
+                Local0 = AIDX (Arg4, Arg5)
+                If ((Arg2 == Zero))
+                {
+                    If ((Arg1 == 0x02))
+                    {
+                        If (!((Local0 == Zero) | (Local0 == 0xFFFFFFFF)))
+                        {
+                            Return (Buffer (One)
+                            {
+                                 0x81                                             // .
+                            })
+                        }
+                    }
+
+                    Return (Buffer (One)
+                    {
+                         0x00                                             // .
+                    })
+                }
+                ElseIf ((Arg2 == 0x07))
+                {
+                    Local1 = Package (0x02)
+                        {
+                            Zero,
+                            ""
+                        }
+                    Local1 [Zero] = Local0
+                    Return (Local1)
+                }
+            }
+        }
+    }
+
...

     Scope (_GPE)
     {
         Name (_HID, "ACPI0006" /* GPE Block Device */)  // _HID: Hardware ID
+        Method (_E01, 0, NotSerialized)  // _Exx: Edge-Triggered GPE, xx=0x00-0xFF
+        {
+            Acquire (\_SB.PCI0.BLCK, 0xFFFF)
+            \_SB.PCI0.PCNT ()
+            Release (\_SB.PCI0.BLCK)
+        }
...

+
+        Device (PHPR)
+        {
+            Name (_HID, "PNP0A06" /* Generic Container Device */)  // _HID: Hardware ID
+            Name (_UID, "PCI Hotplug resources")  // _UID: Unique ID
+            Name (_STA, 0x0B)  // _STA: Status
+            Name (_CRS, ResourceTemplate ()  // _CRS: Current Resource Settings
+            {
+                IO (Decode16,
+                    0x0CC4,             // Range Minimum
+                    0x0CC4,             // Range Maximum
+                    0x01,               // Alignment
+                    0x18,               // Length
+                    )
+            })
+        }
     }
...

And if there is a port in configuration:

             Device (S10)
             {
                 Name (_ADR, 0x00020000)  // _ADR: Address
+                Name (BSEL, Zero)
+                Device (S00)
+                {
+                    Name (_SUN, Zero)  // _SUN: Slot User Number
+                    Name (_ADR, Zero)  // _ADR: Address
+                    Method (_EJ0, 1, NotSerialized)  // _EJx: Eject Device, x=0-9
+                    {
+                        PCEJ (BSEL, _SUN)
+                    }
+
+                    Method (_DSM, 4, Serialized)  // _DSM: Device-Specific Method
+                    {
+                        Return (PDSM (Arg0, Arg1, Arg2, Arg3, BSEL, _SUN))
+                    }
+                }
+
...

+                Method (DVNT, 2, NotSerialized)
+                {
+                    If ((Arg0 & One))
+                    {
+                        Notify (S00, Arg1)
+                    }
...

Signed-off-by: Julia Suvorova <jusual@redhat.com>
Message-Id: <20210713004205.775386-7-jusual@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/acpi/ich9: Set ACPI PCI hot-plug as default on Q35

Q35 has three different types of PCI devices hot-plug: PCIe Native,
SHPC Native and ACPI hot-plug. This patch changes the default choice
for cold-plugged bridges from PCIe Native to ACPI Hot-plug with
ability to use SHPC and PCIe Native for hot-plugged bridges.

This is a list of the PCIe Native hot-plug issues that led to this
change:
    * no racy behavior during boot (see 110c477c2ed)
    * no delay during deleting - after the actual power off software
      must wait at least 1 second before indicating about it. This case
      is quite important for users, it even has its own bug:
          https://bugzilla.redhat.com/show_bug.cgi?id=1594168
    * no timer-based behavior - in addition to the previous example,
      the attention button has a 5-second waiting period, during which
      the operation can be canceled with a second press. While this
      looks fine for manual button control, automation will result in
      the need to queue or drop events, and the software receiving
      events in all sort of unspecified combinations of attention/power
      indicator states, which is racy and uppredictable.
    * fixes:
        * https://bugzilla.redhat.com/show_bug.cgi?id=1752465
        * https://bugzilla.redhat.com/show_bug.cgi?id=1690256

To return to PCIe Native hot-plug:
    -global ICH9-LPC.acpi-pci-hotplug-with-bridge-support=off

Known issue: older linux guests need the following flag
to allow hotplugged pci express devices to use io:
        -device pcie-root-port,io-reserve=4096.
io is unusual for pci express so this seems minor.
We'll fix this by a follow up patch.

Signed-off-by: Julia Suvorova <jusual@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20210713004205.775386-6-jusual@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>

bios-tables-test: Allow changes in DSDT ACPI tables

All DSDT Q35 tables will be modified because ACPI hot-plug is enabled
by default.

Signed-off-by: Julia Suvorova <jusual@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Message-Id: <20210713004205.775386-5-jusual@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/pci/pcie: Do not set HPC flag if acpihp is used

Instead of changing the hot-plug type in _OSC register, do not
set the 'Hot-Plug Capable' flag. This way guest will choose ACPI
hot-plug if it is preferred and leave the option to use SHPC with
pcie-pci-bridge.

The ability to control hot-plug for each downstream port is retained,
while 'hotplug=off' on the port means all hot-plug types are disabled.

Signed-off-by: Julia Suvorova <jusual@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20210713004205.775386-4-jusual@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/acpi/ich9: Enable ACPI PCI hot-plug

Add acpi_pcihp to ich9_pm as part of
'acpi-pci-hotplug-with-bridge-support' option. Set default to false.

Signed-off-by: Julia Suvorova <jusual@redhat.com>
Signed-off-by: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20210713004205.775386-3-jusual@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>

hw/i386/acpi-build: Add ACPI PCI hot-plug methods to Q35

Implement notifications and gpe to support q35 ACPI PCI hot-plug.
Use 0xcc4 - 0xcd7 range for 'acpi-pci-hotplug' io ports.

Signed-off-by: Julia Suvorova <jusual@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Message-Id: <20210713004205.775386-2-jusual@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>

Merge remote-tracking branch 'remotes/awilliam/tags/vfio-update-20210714.0' into staging

VFIO update 2021-07-14

 * Coverity fix to discard listener (David Hildenbrand)

 * MSI-X PBA quirk for BAIDU VFs, additional helper use cases (Cai Huoqing)

# gpg: Signature made Wed 14 Jul 2021 22:31:35 BST
# gpg:                using RSA key 42F6C04E540BD1A99E7B8A90239B9B6E3BB08B22
# gpg:                issuer "alex.williamson@redhat.com"
# gpg: Good signature from "Alex Williamson <alex.williamson@redhat.com>" [full]
# gpg:                 aka "Alex Williamson <alex@shazbot.org>" [full]
# gpg:                 aka "Alex Williamson <alwillia@redhat.com>" [full]
# gpg:                 aka "Alex Williamson <alex.l.williamson@gmail.com>" [full]
# Primary key fingerprint: 42F6 C04E 540B D1A9 9E7B  8A90 239B 9B6E 3BB0 8B22

* remotes/awilliam/tags/vfio-update-20210714.0:
  vfio/pci: Add pba_offset PCI quirk for BAIDU KUNLUN AI processor
  vfio/pci: Change to use vfio_pci_is()
  vfio: Fix CID 1458134 in vfio_register_ram_discard_listener()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/berrange-gitlab/tags/crypto-and-more-pull-request' into staging

Merge crypto updates and misc fixes

 * Introduce a GNUTLS backend for crypto algorithms
 * Change crypto library preference gnutls > gcrypt > nettle > built-in
 * Remove built-in DES impl
 * Remove XTS mode from built-in AES impl
 * Fix seccomp rules to allow resource info getters
 * Fix migration performance test
 * Use GDateTime in io/ and net/rocker/ code
 * Improve docs for -smp

# gpg: Signature made Wed 14 Jul 2021 15:08:00 BST
# gpg:                using RSA key DAF3A6FDB26B62912D0E8E3FBE86EBB415104FDF
# gpg: Good signature from "Daniel P. Berrange <dan@berrange.com>" [full]
# gpg:                 aka "Daniel P. Berrange <berrange@redhat.com>" [full]
# Primary key fingerprint: DAF3 A6FD B26B 6291 2D0E  8E3F BE86 EBB4 1510 4FDF

* remotes/berrange-gitlab/tags/crypto-and-more-pull-request: (26 commits)
  qemu-options: rewrite help for -smp options
  qemu-options: tweak to show that CPU count is optional
  qemu-options: re-arrange CPU topology options
  docs: fix typo s/Intel/AMD/ in CPU model notes
  tests/migration: fix unix socket migration
  seccomp: don't block getters for resource control syscalls
  io: use GDateTime for formatting timestamp for websock headers
  net/rocker: use GDateTime for formatting timestamp in debug messages
  crypto: prefer gnutls as the crypto backend if new enough
  crypto: add gnutls pbkdf provider
  crypto: add gnutls hmac provider
  crypto: add gnutls hash provider
  crypto: add gnutls cipher provider
  crypto: introduce build system for gnutls crypto backend
  crypto: flip priority of backends to prefer gcrypt
  crypto: replace 'des-rfb' cipher with 'des'
  crypto: delete built-in XTS cipher mode support
  crypto: delete built-in DES implementation
  crypto: add crypto tests for single block DES-ECB and DES-CBC
  crypto: drop custom XTS support in gcrypt driver
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/stsquad/tags/pull-testing-and-plugins-140721-5' into staging

Testing and plugin updates:

  - custom runner playbooks for configuring GitLab runners
  - integrate Cirrus jobs into GitLab via cirrus-run
  - clean-up docker package lists
  - bump NetBSD to 9.2
  - bump OpenBSD to 6.9
  - make test-mmap more hexagon friendly
  - fixup handling of hostaddr for plugins
  - disallow some incompatible plugin configurations
  - fix handling of -ldl for BSDs
  - remove some old unused symbols from the plugin symbol map
  - enable plugins by default for most TCG builds
  - honour main build -Wall settings for plugins
  - new execlog plugin
  - new cache modelling plugin
  - fix io_uring build regression
  - disable modular TCG on Darwin

# gpg: Signature made Wed 14 Jul 2021 15:56:27 BST
# gpg:                using RSA key 6685AE99E75167BCAFC8DF35FBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>" [full]
# Primary key fingerprint: 6685 AE99 E751 67BC AFC8  DF35 FBD0 DB09 5A9E 2A44

* remotes/stsquad/tags/pull-testing-and-plugins-140721-5: (44 commits)
  MAINTAINERS: Added myself as a reviewer for TCG Plugins
  docs/devel: Added cache plugin to the plugins docs
  plugins/cache: Added FIFO and LRU eviction policies
  plugins/cache: Enable cache parameterization
  plugins: Added a new cache modelling plugin
  docs/devel: tcg-plugins: add execlog plugin description
  contrib/plugins: add execlog to log instruction execution and memory access
  contrib/plugins: enable -Wall for building plugins
  tcg/plugins: enable by default for most TCG builds
  configure: stop user enabling plugins on Windows for now
  configure: add an explicit static and plugins check
  configure: don't allow plugins to be enabled for a non-TCG build
  tcg/plugins: remove some stale entries from the symbol list
  meson.build: relax the libdl test to one for the function dlopen
  meson.build: move TCG plugin summary output
  plugins: fix-up handling of internal hostaddr for 32 bit
  tests/tcg: make test-mmap a little less aggressive
  tests/vm: update openbsd to release 6.9
  tests/vm: update NetBSD to 9.2
  tests/docker: expand opensuse-leap package list
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/philmd/tags/fw-edk2-20210714' into staging

Patches related to EDK2/OVMF

- MAINTAINERS: remove Laszlo Ersek's entries
- Introduce X86_FW_OVMF Kconfig symbol
- pc_system_ovmf_table_find: Assert that flash was parsed, document
- gitlab-ci: Fix the jobs building EDK2 blobs

# gpg: Signature made Wed 14 Jul 2021 21:36:33 BST
# gpg:                using RSA key FAABE75E12917221DCFD6BB2E3E32C2CDEADC0DE
# gpg: Good signature from "Philippe Mathieu-Daudé (F4BUG) <f4bug@amsat.org>" [full]
# Primary key fingerprint: FAAB E75E 1291 7221 DCFD  6BB2 E3E3 2C2C DEAD C0DE

* remotes/philmd/tags/fw-edk2-20210714:
  MAINTAINERS: remove Laszlo Ersek's entries
  hw/i386: Introduce X86_FW_OVMF Kconfig symbol
  hw/i386/pc: Document pc_system_ovmf_table_find
  hw/i386/pc: pc_system_ovmf_table_find: Assert that flash was parsed
  gitlab-ci: Extract EDK2 job rules to reusable section

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

MAINTAINERS: remove Laszlo Ersek's entries

I've relinquished my edk2 roles with the following commit message [1] [2]
[3]:

> Maintainers.txt: remove Laszlo Ersek's entries
>
> I'm relinquishing all my roles listed in "Maintainers.txt", for personal
> reasons.
>
> My email address <lersek@redhat.com> remains functional.
>
> To my understanding, my employer is working to assign others engineers
> to the edk2 project (at their discretion).

[1] https://edk2.groups.io/g/devel/message/77585
[2] https://listman.redhat.com/archives/edk2-devel-archive/2021-July/msg00202.html
[3] http://mid.mail-archive.com/20210708070916.8937-1-lersek@redhat.com

Accordingly, remove my entries from QEMU's MAINTAINERS file as well, which
all relate to guest firmware.

Cc: Daniel P. Berrange <berrange@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Kashyap Chamarthy <kchamart@redhat.com>
Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210708071409.9671-1-lersek@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

hw/i386: Introduce X86_FW_OVMF Kconfig symbol

Introduce the X86_FW_OVMF Kconfig symbol for OVMF-specific code.
Move the OVMF-specific code from pc_sysfw.c to pc_sysfw_ovmf.c,
adding a pair of stubs.
Update MAINTAINERS to reach OVMF maintainers when these new
files are modified.

This fixes when building the microvm machine standalone:

  /usr/bin/ld: libqemu-i386-softmmu.fa.p/target_i386_monitor.c.o: in
  function `qmp_sev_inject_launch_secret':
  target/i386/monitor.c:749: undefined reference to `pc_system_ovmf_table_find'

Fixes: f522cef9b35 ("sev: update sev-inject-launch-secret to make gpa optional")
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20210616204328.2611406-22-philmd@redhat.com>

hw/i386/pc: Document pc_system_ovmf_table_find

Suggested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Dov Murik <dovmurik@linux.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210701052749.934744-3-dovmurik@linux.ibm.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

hw/i386/pc: pc_system_ovmf_table_find: Assert that flash was parsed

Add assertion in pc_system_ovmf_table_find that verifies that the flash
was indeed previously parsed (looking for the OVMF table) by
pc_system_parse_ovmf_flash.

Now pc_system_ovmf_table_find distinguishes between "no one called
pc_system_parse_ovmf_flash" (which will abort due to assertion failure)
and "the flash was parsed but no OVMF table was found, or it is invalid"
(which will return false).

Suggested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Dov Murik <dovmurik@linux.ibm.com>
Reviewed-by: Tom Lendacky <thomas.lendacky@amd.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210701052749.934744-2-dovmurik@linux.ibm.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>

gitlab-ci: Extract EDK2 job rules to reusable section

All jobs depending on 'docker-edk2' job must use at most all
the rules that triggers it. The simplest way to ensure that
is to always use the same rules. Extract all the rules to a
reusable section, and include this section (with the 'extends'
keyword) in both 'docker-edk2' and 'build-edk2' jobs.

The problem was introduced in commit 71920809cea ("gitlab-ci.yml:
Add jobs to build EDK2 firmware binaries"), but was revealed in
commit 1925468ddbf ("docker: EDK2 build job depends on EDK2
container") and eventually failed on CI:
https://gitlab.com/qemu-project/qemu/-/pipelines/335995843

Reported-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Message-Id: <20210714101003.3113726-1-philmd@redhat.com>

vfio/pci: Add pba_offset PCI quirk for BAIDU KUNLUN AI processor

Fix pba_offset initialization value for BAIDU KUNLUN Virtual
Function device. The KUNLUN hardware returns an incorrect
value for the VF PBA offset, and add a quirk to instead
return a hardcoded value of 0xb400.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210713093743.942-1-caihuoqing@baidu.com
[aw: comment & whitespace tuning]
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>

vfio/pci: Change to use vfio_pci_is()

Make use of vfio_pci_is() helper function.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210713014831.742-1-caihuoqing@baidu.com
[aw: commit log wording]
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>

vfio: Fix CID 1458134 in vfio_register_ram_discard_listener()

  CID 1458134:  Integer handling issues  (BAD_SHIFT)
    In expression "1 << ctz64(container->pgsizes)", left shifting by more
    than 31 bits has undefined behavior.  The shift amount,
    "ctz64(container->pgsizes)", is 64.

Commit 5e3b981c330c ("vfio: Support for RamDiscardManager in the !vIOMMU
case") added an assertion that our granularity is at least as big as the
page size.

Although unlikely, we could have a page size that does not fit into
32 bit. In that case, we'd try shifting by more than 31 bit.

Let's use 1ULL instead and make sure we're not shifting by more than 63
bit by asserting that any bit in container->pgsizes is set.

Fixes: CID 1458134
Cc: Alex Williamson <alex.williamson@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Auger Eric <eric.auger@redhat.com>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: teawater <teawaterz@linux.alibaba.com>
Cc: Marek Kedzierski <mkedzier@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta@ionos.com>
Link: https://lore.kernel.org/r/20210712083135.15755-1-david@redhat.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>

Merge remote-tracking branch 'remotes/cleber-gitlab/tags/python-next-pull-request' into staging

Python and Acceptance Tests

- New SMMUv3 and Intel IOMMU tests
- Respect "cpu" tags and reduce boiler plate code
- Improved logging of qemu execution output
- Other misc improvements

# gpg: Signature made Tue 13 Jul 2021 22:11:36 BST
# gpg:                using RSA key 7ABB96EB8B46B94D5E0FE9BB657E8D33A5F209F3
# gpg: Good signature from "Cleber Rosa <crosa@redhat.com>" [marginal]
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg:          It is not certain that the signature belongs to the owner.
# Primary key fingerprint: 7ABB 96EB 8B46 B94D 5E0F  E9BB 657E 8D33 A5F2 09F3

* remotes/cleber-gitlab/tags/python-next-pull-request: (23 commits)
  tests/acceptance/cpu_queries.py: use the proper logging channels
  tests/acceptance/linux_ssh_mips_malta.py: drop identical setUp
  Acceptance tests: do not try to reuse packages from the system
  python: Configure tox to skip missing interpreters
  tests/acceptance: Handle cpu tag on x86_cpu_model_versions tests
  tests/acceptance: Add set_vm_arg() to the Test class
  python/qemu: Add args property to the QEMUMachine class
  tests/acceptance: Tagging tests with "cpu:VALUE"
  tests/acceptance: Let the framework handle "cpu:VALUE" tagged tests
  tests/acceptance: Fix mismatch on cpu tagged tests
  tests/acceptance: Automatic set -cpu to the test vm
  tests/acceptance: Tag NetBSD tests as 'os:netbsd'
  avocado_qemu: Add Intel iommu tests
  avocado_qemu: Add SMMUv3 tests
  Acceptance Tests: Add default kernel params and pxeboot url to the KNOWN_DISTROS collection
  avocado_qemu: Fix KNOWN_DISTROS map into the LinuxDistro class
  tests/acceptance: Ignore binary data sent on serial console
  Acceptance Tests: support choosing specific distro and version
  Acceptance Tests: move definition of distro checksums to the framework
  Acceptance Tests: rename attribute holding the distro image checksum
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/mdroth/tags/qga-pull-2021-07-13-tag' into staging

qemu-ga patch queue for soft-freeze

* add support for Windows Server 2022 in get-osinfo command

# gpg: Signature made Tue 13 Jul 2021 19:10:05 BST
# gpg:                using RSA key CEACC9E15534EBABB82D3FA03353C9CEF108B584
# gpg: Good signature from "Michael Roth <flukshun@gmail.com>" [full]
# gpg:                 aka "Michael Roth <mdroth@utexas.edu>" [full]
# gpg:                 aka "Michael Roth <mdroth@linux.vnet.ibm.com>" [full]
# Primary key fingerprint: CEAC C9E1 5534 EBAB B82D  3FA0 3353 C9CE F108 B584

* remotes/mdroth/tags/qga-pull-2021-07-13-tag:
  qga-win: Add support of Windows Server 2022 in get-osinfo command

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

MAINTAINERS: Added myself as a reviewer for TCG Plugins

Signed-off-by: Mahmoud Mandour <ma.mandourr@gmail.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623125458.450462-6-ma.mandourr@gmail.com>
Message-Id: <20210709143005.1554-41-alex.bennee@linaro.org>

docs/devel: Added cache plugin to the plugins docs

Signed-off-by: Mahmoud Mandour <ma.mandourr@gmail.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210628053808.17422-1-ma.mandourr@gmail.com>
Message-Id: <20210709143005.1554-40-alex.bennee@linaro.org>

plugins/cache: Added FIFO and LRU eviction policies

Implemented FIFO and LRU eviction policies. Now one of the three
eviction policies can be chosen as an argument. On not specifying an
argument, LRU is used by default.

Signed-off-by: Mahmoud Mandour <ma.mandourr@gmail.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623125458.450462-4-ma.mandourr@gmail.com>
Message-Id: <20210709143005.1554-39-alex.bennee@linaro.org>

plugins/cache: Enable cache parameterization

Enabled configuring both icache and dcache parameters using plugin
parameters.

Signed-off-by: Mahmoud Mandour <ma.mandourr@gmail.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623125458.450462-3-ma.mandourr@gmail.com>
Message-Id: <20210709143005.1554-38-alex.bennee@linaro.org>

plugins: Added a new cache modelling plugin

Added a cache modelling plugin that uses a static configuration used in
many of the commercial microprocessors and uses random eviction policy.

The purpose of the plugin is to identify the most cache-thrashing
instructions for both instruction cache and data cache.

Signed-off-by: Mahmoud Mandour <ma.mandourr@gmail.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623125458.450462-2-ma.mandourr@gmail.com>
Message-Id: <20210709143005.1554-37-alex.bennee@linaro.org>

docs/devel: tcg-plugins: add execlog plugin description

This adds description of the execlog TCG plugin with an example.

Signed-off-by: Alexandre Iooss <erdnaxe@crans.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210702081307.1653644-3-erdnaxe@crans.org>
Message-Id: <20210709143005.1554-36-alex.bennee@linaro.org>

contrib/plugins: add execlog to log instruction execution and memory access

Log instruction execution and memory access to a file.
This plugin can be used for reverse engineering or for side-channel analysis
using QEMU.

Signed-off-by: Alexandre Iooss <erdnaxe@crans.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210702081307.1653644-2-erdnaxe@crans.org>
Message-Id: <20210709143005.1554-35-alex.bennee@linaro.org>

contrib/plugins: enable -Wall for building plugins

Lets spot the obvious errors.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20210709143005.1554-34-alex.bennee@linaro.org>

tcg/plugins: enable by default for most TCG builds

Aside from a minor bloat to file size the ability to have TCG plugins
has no real impact on performance unless a plugin is actively loaded.
Even then the libempty.so plugin shows only a minor degradation in
performance caused by the extra book keeping the TCG has to do to keep
track of instructions. As it's a useful feature lets just enable it by
default and reduce our testing matrix a little.

We need to move our linker testing earlier so we can be sure we can
enable the loader module required. As we have ruled out static &
plugins in an earlier patch we can also reduce the indent a little.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20210709143005.1554-33-alex.bennee@linaro.org>

configure: stop user enabling plugins on Windows for now

There are some patches on the list that enable plugins on Windows but
they still need some changes to be ready:

  https://patchew.org/QEMU/20201013002806.1447-1-luoyonggang@gmail.com/

In the meantime lets stop the user from being able to configure the
support so they don't get confused by the weird linker error messages
later.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Cc: Yonggang Luo <luoyonggang@gmail.com>
Message-Id: <20210709143005.1554-32-alex.bennee@linaro.org>

configure: add an explicit static and plugins check

Moving this check earlier will make the later re-factor for enabling
by default a bit neater.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20210709143005.1554-31-alex.bennee@linaro.org>

configure: don't allow plugins to be enabled for a non-TCG build

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20210709143005.1554-30-alex.bennee@linaro.org>

tcg/plugins: remove some stale entries from the symbol list

The OSX linker is a lot fussier about these missing symbols.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210714101623.17046-1-alex.bennee@linaro.org>

meson.build: relax the libdl test to one for the function dlopen

For the *BSD family dlopen is already part of libc so it's not a hard
dependency to have a libdl.so library.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20210714101536.16016-1-alex.bennee@linaro.org>

meson.build: move TCG plugin summary output

Let's put it with the rest of the TCG related output with the
accelerator.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20210709143005.1554-29-alex.bennee@linaro.org>

plugins: fix-up handling of internal hostaddr for 32 bit

The compiler rightly complains when we build on 32 bit that casting
uint64_t into a void is a bad idea. We are really dealing with a host
pointer at this point so treat it as such. This does involve
a uintptr_t cast of the result of the TLB addend as we know that has
to point to the host memory.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20210709143005.1554-28-alex.bennee@linaro.org>

tests/tcg: make test-mmap a little less aggressive

The check_aligned_anonymous_unfixed_mmaps and
check_aligned_anonymous_unfixed_colliding_mmaps do a lot of mmap's and
copying of data. This is especially unfriendly to targets like hexagon
which have quite large pages and need to do sanity checks on each
memory access.

While we are at it clean-up the white space and style issues from the
legacy code. As we no longer do quite so much needless memory access
we can also remove the hexagon timeout hack.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Message-Id: <20210709143005.1554-27-alex.bennee@linaro.org>

tests/vm: update openbsd to release 6.9

tests/vm: update openbsd to release 6.9

Signed-off-by: Brad Smith <brad@comstyle.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <YLRDeJV8qBrt9++c@humpty.home.comstyle.com>
Message-Id: <20210709143005.1554-26-alex.bennee@linaro.org>

tests/vm: update NetBSD to 9.2

tests/vm: update NetBSD to 9.2

Signed-off-by: Brad Smith <brad@comstyle.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Acked-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <YLe+LZ/j0pxVGeBi@humpty.home.comstyle.com>
Message-Id: <20210709143005.1554-25-alex.bennee@linaro.org>

tests/docker: expand opensuse-leap package list

This is the fully expanded list of build pre-requisites QEMU can
conceivably use in any scenario.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-16-berrange@redhat.com>
Message-Id: <20210709143005.1554-24-alex.bennee@linaro.org>

tests/docker: expand ubuntu2004 package list

This is the fully expanded list of build pre-requisites QEMU can
conceivably use in any scenario.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-15-berrange@redhat.com>
Message-Id: <20210709143005.1554-23-alex.bennee@linaro.org>

tests/docker: expand ubuntu1804 package list

This is the fully expanded list of build pre-requisites QEMU can
conceivably use in any scenario.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-14-berrange@redhat.com>
Message-Id: <20210709143005.1554-22-alex.bennee@linaro.org>

tests/docker: expand fedora package list

This is the fully expanded list of build pre-requisites QEMU can
conceivably use in any scenario.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-13-berrange@redhat.com>
Message-Id: <20210709143005.1554-21-alex.bennee@linaro.org>

tests/docker: expand centos8 package list

This is the fully expanded list of build pre-requisites QEMU can
conceivably use in any scenario.

[AJB: added centos-release-advanced-virtualization/epel-release]

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-12-berrange@redhat.com>
Message-Id: <20210709143005.1554-20-alex.bennee@linaro.org>

tests/docker: remove mingw packages from Fedora

There are dedicated containers providing mingw packages for Fedora.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-11-berrange@redhat.com>
Message-Id: <20210709143005.1554-19-alex.bennee@linaro.org>

tests/docker: fix mistakes in ubuntu package lists

librados-dev is not required by QEMU directly, only librbd-dev.

glusterfs-common is not directly needed by QEMU.

QEMU uses ncursesw only on non-Windows hosts.

The clang package is clang 10.

flex and bison are not required by QEMU.

Standardize on nmap ncat implementation to match Fedora/CentOS.

Remove vim since it is not a build pre-requisite and no other containers
include it.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-10-berrange@redhat.com>
Message-Id: <20210709143005.1554-18-alex.bennee@linaro.org>

tests/docker: fix mistakes in fedora package list

libblockdev-mpath-devel is not used by QEMU, rather it wants
device-mapper-multipath-devel.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-9-berrange@redhat.com>
Message-Id: <20210709143005.1554-17-alex.bennee@linaro.org>

tests/docker: fix mistakes in centos package lists

mesa-libEGL-devel is not used in QEMU at all, but mesa-libgbm-devel is.

spice-glib-devel is not use in QEMU at all, but spice-protocol is.
We also need the -devel package for spice-server, not the runtime.

There is no need to specifically refer to python36, we can just
use python3 as in other distros.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-8-berrange@redhat.com>
Message-Id: <20210709143005.1554-16-alex.bennee@linaro.org>

tests/docker: fix sorting in package lists

This will make diffs in later patches clearer.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-7-berrange@redhat.com>
Message-Id: <20210709143005.1554-15-alex.bennee@linaro.org>

tests/docker: remove FEATURES env var from templates

In preparation for switching to auto-generated dockerfiles, remove the
FEATURES env variable. The equivalent functionality can be achieved in
most cases by just looking for existance of a binary.

The cases which don't correspond to binaries are simply dropped because
configure/meson will probe for any requested feature anyway.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-6-berrange@redhat.com>
Message-Id: <20210709143005.1554-14-alex.bennee@linaro.org>

tests/docker: use explicit docker.io registry

It is good practice to use an explicit registry for referencing the base
image. This is because some distros will inject their own registries
into the search path. For example registry.fedoraproject.org comes ahead
of docker.io. Using an explicit registry avoids wasting time querying
multiple registries for images that they won't have.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-5-berrange@redhat.com>
Message-Id: <20210709143005.1554-13-alex.bennee@linaro.org>

tests/docker: use project specific container registries

Since Docker Hub has started to enforce pull rate limits on clients, it
is preferrable to use project specific container registries where they
are available. Fedora and OpenSUSE projects provide registries.

The images in these registries are also refreshed on a more regular
basis than the ones in docker hub, so the package update should
generally be faster.

While CentOS also has a registry it is considerably outdated compared
to docker.io, and also only provides x86 images, while docker.io images
are multi-arch.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-4-berrange@redhat.com>
Message-Id: <20210709143005.1554-12-alex.bennee@linaro.org>

tests/docker: don't use BUILDKIT in GitLab either

Using BUILDKIT breaks with certain container registries such as CentOS,
with docker build reporting an error such as

  failed to solve with frontend dockerfile.v0:
  failed to build LLB: failed to load cache key:
  unexpected status code
  https://registry.centos.org/v2/centos/manifests/7:
  403 Forbidden

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Message-Id: <20210623142245.307776-3-berrange@redhat.com>
Message-Id: <20210709143005.1554-11-alex.bennee@linaro.org>

hw/usb/ccid: remove references to NSS

The NSS package was previously pre-requisite for building CCID related
features, however, this became obsolete when the libcacard library was
spun off to a separate project:

    commit 7b02f5447c64d1854468f758398c9f6fe9e5721f
    Author: Marc-André Lureau <marcandre.lureau@redhat.com>
    Date:   Sun Aug 30 11:48:40 2015 +0200

        libcacard: use the standalone project

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210623142245.307776-2-berrange@redhat.com>
Message-Id: <20210709143005.1554-10-alex.bennee@linaro.org>

cirrus: delete FreeBSD and macOS jobs

The builds for these two platforms can now be performed from GitLab CI
using cirrus-run.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Message-Id: <20210625172211.451010-4-berrange@redhat.com>
Message-Id: <20210709143005.1554-9-alex.bennee@linaro.org>

gitlab: support for FreeBSD 12, 13 and macOS 11 via cirrus-run

This adds support for running 4 jobs via Cirrus CI runners:

 * FreeBSD 12
 * FreeBSD 13
 * macOS 11 with default XCode
 * macOS 11 with latest XCode

The gitlab job uses a container published by the libvirt-ci
project (https://gitlab.com/libvirt/libvirt-ci) that contains
the 'cirrus-run' command. This accepts a short yaml file that
describes a single Cirrus CI job, runs it using the Cirrus CI
REST API, and reports any output to the console.

In this way Cirrus CI is effectively working as an indirect
custom runner for GitLab CI pipelines. The key benefit is that
Cirrus CI job results affect the GitLab CI pipeline result and
so the user only has look at one CI dashboard.

[AJB: remove $TEMPORARILY_DISABLED condition, s/py37/py38/]

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210625172211.451010-3-berrange@redhat.com>
Message-Id: <20210709143005.1554-8-alex.bennee@linaro.org>

build: validate that system capstone works before using it

Some versions of capstone have shipped a broken pkg-config file which
puts the -I path without the trailing '/capstone' suffix. This breaks
the ability to "#include <capstone.h>". Upstream and most distros have
fixed this, but a few stragglers remain, notably FreeBSD.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20210625172211.451010-2-berrange@redhat.com>
Message-Id: <20210709143005.1554-7-alex.bennee@linaro.org>

disable modular TCG on Darwin

Accelerator modularity does not work on Darwin:

ld: illegal thread local variable reference to regular symbol _current_cpu for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)

Fix by avoiding modular TCG builds.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20210712122208.456264-1-pbonzini@redhat.com>
[AJB: manually merged typo fix]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>

meson: fix condition for io_uring stubs

CONFIG_LINUX_IO_URING is not included in config-host.mak and therefore is
not usable in "when" clauses.  Check the availability of the library,
which matches the condition for the non-stubbed version block/io_uring.c.

At this point, the difference between libraries that have config-host.mak
entries and those that do not is quite confusing.  The remaining ~dozen
should be converted in 6.2.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210712151810.508249-1-pbonzini@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>

tests/tcg: also disable the signals test for plugins

This will be more important when plugins is enabled by default.

Fixes: eba61056e4 ("tests/tcg: generalise the disabling of the signals test")
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20210709143005.1554-6-alex.bennee@linaro.org>

Jobs based on custom runners: add job definitions for QEMU's machines

The QEMU project has two machines (aarch64 and s390x) that can be used
for jobs that do build and run tests.  This introduces those jobs,
which are a mapping of custom scripts used for the same purpose.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Message-Id: <20210630012619.115262-5-crosa@redhat.com>
Message-Id: <20210709143005.1554-5-alex.bennee@linaro.org>

Jobs based on custom runners: docs and gitlab-runner setup playbook

To have the jobs dispatched to custom runners, gitlab-runner must
be installed, active as a service and properly configured.  The
variables file and playbook introduced here should help with those
steps.

The playbook introduced here covers the Linux distributions and
has been primarily tested on OS/machines that the QEMU project
has available to act as runners, namely:

 * Ubuntu 20.04 on aarch64
 * Ubuntu 18.04 on s390x

But, it should work on all other Linux distributions.  Earlier
versions were tested on FreeBSD too, so chances of success are
high.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Willian Rampazzo <willianr@redhat.com>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210630012619.115262-4-crosa@redhat.com>
Message-Id: <20210709143005.1554-4-alex.bennee@linaro.org>

Jobs based on custom runners: build environment docs and playbook

To run basic jobs on custom runners, the environment needs to be
properly set up.  The most common requirement is having the right
packages installed.

The playbook introduced here covers the QEMU's project s390x and
aarch64 machines.  At the time this is being proposed, those machines
have already had this playbook applied to them.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20210630012619.115262-3-crosa@redhat.com>
Message-Id: <20210709143005.1554-3-alex.bennee@linaro.org>

Jobs based on custom runners: documentation and configuration placeholder

As described in the included documentation, the "custom runner" jobs
extend the GitLab CI jobs already in place.  One of their primary
goals of catching and preventing regressions on a wider number of host
systems than the ones provided by GitLab's shared runners.

This sets the stage in which other community members can add their own
machine configuration documentation/scripts, and accompanying job
definitions.  As a general rule, those newly added contributed jobs
should run as "non-gating", until their reliability is verified (AKA
"allow_failure: true").

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Willian Rampazzo <willianr@redhat.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Message-Id: <20210630012619.115262-2-crosa@redhat.com>
Message-Id: <20210709143005.1554-2-alex.bennee@linaro.org>

qemu-options: rewrite help for -smp options

The -smp option help is peculiarly specific about mentioning the CPU
upper limits, but these are wrong. The "PC" target has varying max
CPU counts depending on the machine type picked. Notes about guest
OS limits are inappropriate for QEMU docs. There are way too many
machine types for it to be practical to mention actual limits, and
some limits are even modified by downstream distribtions. Thus it
is better to remove the specific limits entirely.

The CPU topology reporting is also not neccessarily specific to the
PC platform and descriptions around the rules of usage are somewhat
terse. Expand this information with some examples to show effects
of defaulting.

Reviewed-by: Pankaj Gupta <pankaj.gupta@ionos.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

qemu-options: tweak to show that CPU count is optional

The initial CPU count number is not required, if any of the topology
options are given, since it can be computed.

Reviewed-by: Pankaj Gupta <pankaj.gupta@ionos.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

qemu-options: re-arrange CPU topology options

The list of CPU topology options are presented in a fairly arbitrary
order currently. Re-arrange them so that they're ordered from largest to
smallest unit

Reviewed-by: Pankaj Gupta <pankaj.gupta@ionos.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

docs: fix typo s/Intel/AMD/ in CPU model notes

Reviewed-by: Pankaj Gupta <pankaj.gupta@ionos.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

tests/migration: fix unix socket migration

The test aborts and error message as the following be throwed:
"No such file or directory: '/var/tmp/qemu-migrate-{pid}.migrate",
when the unix socket migration test nearly done. The reason is
qemu removes the unix socket file after migration before
guestperf.py script do it. So pre-check if the socket file exists
when removing it to prevent the guestperf program from aborting.

See also commit f9cc00346d3 ("tests/migration: fix unix socket batch
migration").

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Signed-off-by: Hyman <huangy81@chinatelecom.cn>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

seccomp: don't block getters for resource control syscalls

Recent GLibC calls sched_getaffinity in code paths related to malloc and
when QEMU blocks access, it sends it off into a bad codepath resulting
in stack exhaustion[1]. The GLibC bug is being fixed[2], but none the
less, GLibC has valid reasons to want to use sched_getaffinity.

It is not unreasonable for code to want to run many resource syscalls
for information gathering, so it is a bit too harsh for QEMU to block
them.

[1] https://bugzilla.redhat.com/show_bug.cgi?id=1975693
[2] https://sourceware.org/pipermail/libc-alpha/2021-June/128271.html
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Acked-by: Eduardo Otubo <otubo@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

io: use GDateTime for formatting timestamp for websock headers

The GDateTime APIs provided by GLib avoid portability pitfalls, such
as some platforms where 'struct timeval.tv_sec' field is still 'long'
instead of 'time_t'. When combined with automatic cleanup, GDateTime
often results in simpler code too.

Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

net/rocker: use GDateTime for formatting timestamp in debug messages

The GDateTime APIs provided by GLib avoid portability pitfalls, such
as some platforms where 'struct timeval.tv_sec' field is still 'long'
instead of 'time_t'. When combined with automatic cleanup, GDateTime
often results in simpler code too.

Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: prefer gnutls as the crypto backend if new enough

If we have gnutls >= 3.6.13, then it has enough functionality
and performance that we can use it as the preferred crypto
backend.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: add gnutls pbkdf provider

This adds support for using gnutls as a provider of the crypto
pbkdf APIs.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: add gnutls hmac provider

This adds support for using gnutls as a provider of the crypto
hmac APIs.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: add gnutls hash provider

This adds support for using gnutls as a provider of the crypto
hash APIs.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: add gnutls cipher provider

Add an implementation of the QEMU cipher APIs to the gnutls
crypto backend. XTS support is only available for gnutls
version >= 3.6.8. Since ECB mode is not exposed by gnutls
APIs, we can't use the private XTS code for compatibility.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: introduce build system for gnutls crypto backend

This introduces the build logic needed to decide whether we can
use gnutls as a crypto driver backend. The actual implementations
will be introduced in following patches. We only wish to use
gnutls if it has version 3.6.14 or newer, because that is what
finally brings HW accelerated AES-XTS mode for x86_64.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: flip priority of backends to prefer gcrypt

Originally we preferred to use nettle over gcrypt because
gnutls already links to nettle and thus it minimizes the
dependencies. In retrospect this was the wrong criteria to
optimize for.

Currently shipping versions of gcrypt have cipher impls that
are massively faster than those in nettle and this is way
more important.  The nettle library is also not capable of
enforcing FIPS compliance, since it considers that out of
scope. It merely aims to provide general purpose impls of
algorithms, and usage policy is left upto the layer above,
such as GNUTLS.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: replace 'des-rfb' cipher with 'des'

Currently the crypto layer exposes support for a 'des-rfb'
algorithm which is just normal single-DES, with the bits
in each key byte reversed. This special key munging is
required by the RFB protocol password authentication
mechanism.

Since the crypto layer is generic shared code, it makes
more sense to do the key byte munging in the VNC server
code, and expose normal single-DES support.

Replacing cipher 'des-rfb' by 'des' looks like an incompatible
interface change, but it doesn't matter.  While the QMP schema
allows any QCryptoCipherAlgorithm for the 'cipher-alg' field
in QCryptoBlockCreateOptionsLUKS, the code restricts what can
be used at runtime. Thus the only effect is a change in error
message.

Original behaviour:

 $ qemu-img create -f luks --object secret,id=sec0,data=123 -o cipher-alg=des-rfb,key-secret=sec0 demo.luks 1G
 Formatting 'demo.luks', fmt=luks size=1073741824 key-secret=sec0 cipher-alg=des-rfb
 qemu-img: demo.luks: Algorithm 'des-rfb' not supported

New behaviour:

 $ qemu-img create -f luks --object secret,id=sec0,data=123 -o cipher-alg=des-rfb,key-secret=sec0 demo.luks 1G
 Formatting 'demo.luks', fmt=luks size=1073741824 key-secret=sec0 cipher-alg=des-fish
 qemu-img: demo.luks: Invalid parameter 'des-rfb'

Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: delete built-in XTS cipher mode support

The built-in AES+XTS implementation is used for the LUKS encryption
When building system emulators it is reasonable to expect that an
external crypto library is being used instead. The performance of the
builtin XTS implementation is terrible as it has no CPU acceleration
support. It is thus not worth keeping a home grown XTS implementation
for the built-in cipher backend.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: delete built-in DES implementation

The built-in DES implementation is used for the VNC server password
authentication scheme. When building system emulators it is reasonable
to expect that an external crypto library is being used. It is thus
not worth keeping a home grown DES implementation in tree.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: add crypto tests for single block DES-ECB and DES-CBC

The GNUTLS crypto provider doesn't support DES-ECB, only DES-CBC.
We can use the latter to simulate the former, if we encrypt only
1 block (8 bytes) of data at a time, using an all-zeros IV. This
is a very inefficient way to use the QCryptoCipher APIs, but
since the VNC authentication challenge is only 16 bytes, this
is acceptable. No other part of QEMU should be using DES. This
test case demonstrates the equivalence of ECB and CBC for the
single-block case.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: drop custom XTS support in gcrypt driver

The XTS cipher mode was introduced in gcrypt 1.8.0, which
matches QEMU's current minimum version.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: drop gcrypt thread initialization code

This is only required on gcrypt < 1.6.0, and is thus obsolete
since

  commit b33a84632a3759c00320fd80923aa963c11207fc
  Author: Daniel P. Berrangé <berrange@redhat.com>
  Date:   Fri May 14 13:04:08 2021 +0100

    crypto: bump min gcrypt to 1.8.0, dropping RHEL-7 support

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: fix gcrypt min version 1.8 regression

The min gcrypt was bumped:

  commit b33a84632a3759c00320fd80923aa963c11207fc
  Author: Daniel P. Berrangé <berrange@redhat.com>
  Date:   Fri May 14 13:04:08 2021 +0100

    crypto: bump min gcrypt to 1.8.0, dropping RHEL-7 support

but this was accidentally lost in conflict resolution for

  commit 5761251138cb69c310e9df7dfc82c4c6fd2444e4
  Author: Paolo Bonzini <pbonzini@redhat.com>
  Date:   Thu Jun 3 11:15:26 2021 +0200

    configure, meson: convert crypto detection to meson

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: use &error_fatal in crypto tests

Using error_fatal provides better diagnostics when tests
failed, than using asserts, because we see the text of
the error message.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: skip essiv ivgen tests if AES+ECB isn't available

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>