xref: /openbmc/qemu/docs/system/riscv/virt.rst (revision 243705aa)
1'virt' Generic Virtual Platform (``virt``)
2==========================================
3
4The ``virt`` board is a platform which does not correspond to any real hardware;
5it is designed for use in virtual machines. It is the recommended board type
6if you simply want to run a guest such as Linux and do not care about
7reproducing the idiosyncrasies and limitations of a particular bit of
8real-world hardware.
9
10Supported devices
11-----------------
12
13The ``virt`` machine supports the following devices:
14
15* Up to 8 generic RV32GC/RV64GC cores, with optional extensions
16* Core Local Interruptor (CLINT)
17* Platform-Level Interrupt Controller (PLIC)
18* CFI parallel NOR flash memory
19* 1 NS16550 compatible UART
20* 1 Google Goldfish RTC
21* 1 SiFive Test device
22* 8 virtio-mmio transport devices
23* 1 generic PCIe host bridge
24* The fw_cfg device that allows a guest to obtain data from QEMU
25
26The hypervisor extension has been enabled for the default CPU, so virtual
27machines with hypervisor extension can simply be used without explicitly
28declaring.
29
30Hardware configuration information
31----------------------------------
32
33The ``virt`` machine automatically generates a device tree blob ("dtb")
34which it passes to the guest, if there is no ``-dtb`` option. This provides
35information about the addresses, interrupt lines and other configuration of
36the various devices in the system. Guest software should discover the devices
37that are present in the generated DTB.
38
39If users want to provide their own DTB, they can use the ``-dtb`` option.
40These DTBs should have the following requirements:
41
42* The number of subnodes of the /cpus node should match QEMU's ``-smp`` option
43* The /memory reg size should match QEMU’s selected ram_size via ``-m``
44* Should contain a node for the CLINT device with a compatible string
45  "riscv,clint0" if using with OpenSBI BIOS images
46
47Boot options
48------------
49
50The ``virt`` machine can start using the standard -kernel functionality
51for loading a Linux kernel, a VxWorks kernel, an S-mode U-Boot bootloader
52with the default OpenSBI firmware image as the -bios. It also supports
53the recommended RISC-V bootflow: U-Boot SPL (M-mode) loads OpenSBI fw_dynamic
54firmware and U-Boot proper (S-mode), using the standard -bios functionality.
55
56Using flash devices
57-------------------
58
59By default, the first flash device (pflash0) is expected to contain
60S-mode firmware code. It can be configured as read-only, with the
61second flash device (pflash1) available to store configuration data.
62
63For example, booting edk2 looks like
64
65.. code-block:: bash
66
67  $ qemu-system-riscv64 \
68     -blockdev node-name=pflash0,driver=file,read-only=on,filename=<edk2_code> \
69     -blockdev node-name=pflash1,driver=file,filename=<edk2_vars> \
70     -M virt,pflash0=pflash0,pflash1=pflash1 \
71     ... other args ....
72
73For TCG guests only, it is also possible to boot M-mode firmware from
74the first flash device (pflash0) by additionally passing ``-bios
75none``, as in
76
77.. code-block:: bash
78
79  $ qemu-system-riscv64 \
80     -bios none \
81     -blockdev node-name=pflash0,driver=file,read-only=on,filename=<m_mode_code> \
82     -M virt,pflash0=pflash0 \
83     ... other args ....
84
85Firmware images used for pflash must be exactly 32 MiB in size.
86
87Machine-specific options
88------------------------
89
90The following machine-specific options are supported:
91
92- aclint=[on|off]
93
94  When this option is "on", ACLINT devices will be emulated instead of
95  SiFive CLINT. When not specified, this option is assumed to be "off".
96
97- aia=[none|aplic|aplic-imsic]
98
99  This option allows selecting interrupt controller defined by the AIA
100  (advanced interrupt architecture) specification. The "aia=aplic" selects
101  APLIC (advanced platform level interrupt controller) to handle wired
102  interrupts whereas the "aia=aplic-imsic" selects APLIC and IMSIC (incoming
103  message signaled interrupt controller) to handle both wired interrupts and
104  MSIs. When not specified, this option is assumed to be "none" which selects
105  SiFive PLIC to handle wired interrupts.
106
107- aia-guests=nnn
108
109  The number of per-HART VS-level AIA IMSIC pages to be emulated for a guest
110  having AIA IMSIC (i.e. "aia=aplic-imsic" selected). When not specified,
111  the default number of per-HART VS-level AIA IMSIC pages is 0.
112
113Running Linux kernel
114--------------------
115
116Linux mainline v5.12 release is tested at the time of writing. To build a
117Linux mainline kernel that can be booted by the ``virt`` machine in
11864-bit mode, simply configure the kernel using the defconfig configuration:
119
120.. code-block:: bash
121
122  $ export ARCH=riscv
123  $ export CROSS_COMPILE=riscv64-linux-
124  $ make defconfig
125  $ make
126
127To boot the newly built Linux kernel in QEMU with the ``virt`` machine:
128
129.. code-block:: bash
130
131  $ qemu-system-riscv64 -M virt -smp 4 -m 2G \
132      -display none -serial stdio \
133      -kernel arch/riscv/boot/Image \
134      -initrd /path/to/rootfs.cpio \
135      -append "root=/dev/ram"
136
137To build a Linux mainline kernel that can be booted by the ``virt`` machine
138in 32-bit mode, use the rv32_defconfig configuration. A patch is required to
139fix the 32-bit boot issue for Linux kernel v5.12.
140
141.. code-block:: bash
142
143  $ export ARCH=riscv
144  $ export CROSS_COMPILE=riscv64-linux-
145  $ curl https://patchwork.kernel.org/project/linux-riscv/patch/20210627135117.28641-1-bmeng.cn@gmail.com/mbox/ > riscv.patch
146  $ git am riscv.patch
147  $ make rv32_defconfig
148  $ make
149
150Replace ``qemu-system-riscv64`` with ``qemu-system-riscv32`` in the command
151line above to boot the 32-bit Linux kernel. A rootfs image containing 32-bit
152applications shall be used in order for kernel to boot to user space.
153
154Running U-Boot
155--------------
156
157U-Boot mainline v2021.04 release is tested at the time of writing. To build an
158S-mode U-Boot bootloader that can be booted by the ``virt`` machine, use
159the qemu-riscv64_smode_defconfig with similar commands as described above for Linux:
160
161.. code-block:: bash
162
163  $ export CROSS_COMPILE=riscv64-linux-
164  $ make qemu-riscv64_smode_defconfig
165
166Boot the 64-bit U-Boot S-mode image directly:
167
168.. code-block:: bash
169
170  $ qemu-system-riscv64 -M virt -smp 4 -m 2G \
171      -display none -serial stdio \
172      -kernel /path/to/u-boot.bin
173
174To test booting U-Boot SPL which in M-mode, which in turn loads a FIT image
175that bundles OpenSBI fw_dynamic firmware and U-Boot proper (S-mode) together,
176build the U-Boot images using riscv64_spl_defconfig:
177
178.. code-block:: bash
179
180  $ export CROSS_COMPILE=riscv64-linux-
181  $ export OPENSBI=/path/to/opensbi-riscv64-generic-fw_dynamic.bin
182  $ make qemu-riscv64_spl_defconfig
183
184The minimal QEMU commands to run U-Boot SPL are:
185
186.. code-block:: bash
187
188  $ qemu-system-riscv64 -M virt -smp 4 -m 2G \
189      -display none -serial stdio \
190      -bios /path/to/u-boot-spl \
191      -device loader,file=/path/to/u-boot.itb,addr=0x80200000
192
193To test 32-bit U-Boot images, switch to use qemu-riscv32_smode_defconfig and
194riscv32_spl_defconfig builds, and replace ``qemu-system-riscv64`` with
195``qemu-system-riscv32`` in the command lines above to boot the 32-bit U-Boot.
196
197Enabling TPM
198------------
199
200A TPM device can be connected to the virt board by following the steps below.
201
202First launch the TPM emulator:
203
204.. code-block:: bash
205
206  $ swtpm socket --tpm2 -t -d --tpmstate dir=/tmp/tpm \
207        --ctrl type=unixio,path=swtpm-sock
208
209Then launch QEMU with some additional arguments to link a TPM device to the backend:
210
211.. code-block:: bash
212
213  $ qemu-system-riscv64 \
214    ... other args .... \
215    -chardev socket,id=chrtpm,path=swtpm-sock \
216    -tpmdev emulator,id=tpm0,chardev=chrtpm \
217    -device tpm-tis-device,tpmdev=tpm0
218
219The TPM device can be seen in the memory tree and the generated device
220tree and should be accessible from the guest software.
221