We want to support memory devices that can automatically decide how many
memslots they will use. In the worst case, they have to use a single
memslot.
The target use cases are virtio-mem and the hyper-v balloon.
Let's calculate a reasonable limit such a memory device may use, and
instruct the device to make a decision based on that limit. Use a simple
heuristic that considers:
* A memslot soft-limit for all memory devices of 256; also, to not
consume too many memslots -- which could harm performance.
* Actually still free and unreserved memslots
* The percentage of the remaining device memory region that memory device
will occupy.
Further, while we properly check before plugging a memory device whether
there still is are free memslots, we have other memslot consumers (such as
boot memory, PCI BARs) that don't perform any checks and might dynamically
consume memslots without any prior reservation. So we might succeed in
plugging a memory device, but once we dynamically map a PCI BAR we would
be in trouble. Doing accounting / reservation / checks for all such
users is problematic (e.g., sometimes we might temporarily split boot
memory into two memslots, triggered by the BIOS).
We use the historic magic memslot number of 509 as orientation to when
supporting 256 memory devices -> memslots (leaving 253 for boot memory and
other devices) has been proven to work reliable. We'll fallback to
suggesting a single memslot if we don't have at least 509 total memslots.
Plugging vhost devices with less than 509 memslots available while we
have memory devices plugged that consume multiple memslots due to
automatic decisions can be problematic. Most configurations might just fail
due to "limit < used + reserved", however, it can also happen that these
memory devices would suddenly consume memslots that would actually be
required by other memslot consumers (boot, PCI BARs) later. Note that this
has always been sketchy with vhost devices that support only a small number
of memslots; but we don't want to make it any worse.So let's keep it simple
and simply reject plugging such vhost devices in such a configuration.
Eventually, all vhost devices that want to be fully compatible with such
memory devices should support a decent number of memslots (>= 509).
Message-ID: <20230926185738.277351-13-david@redhat.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Let's add vhost_get_max_memslots().
Message-ID: <20230926185738.277351-12-david@redhat.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
We'll need the stub soon from memory device context.
While at it, use "unsigned int" as return value and place the
declaration next to kvm_get_free_memslots().
Message-ID: <20230926185738.277351-11-david@redhat.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
We want to support memory devices that have a dynamically managed memory
region container as device memory region. This device memory region maps
multiple RAM memory subregions (e.g., aliases to the same RAM memory
region), whereby these subregions can be (un)mapped on demand.
Each RAM subregion will consume a memslot in KVM and vhost, resulting in
such a new device consuming memslots dynamically, and initially usually
0. We already track the number of used vs. required memslots for all
memslots. From that, we can derive the number of reserved memslots that
must not be used otherwise.
The target use case is virtio-mem and the hyper-v balloon, which will
dynamically map aliases to RAM memory region into their device memory
region container.
Properly document what's supported and what's not and extend the vhost
memslot check accordingly.
Message-ID: <20230926185738.277351-10-david@redhat.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Let's track how many memslots are required by plugged memory devices and
how many are currently actually getting used by plugged memory
devices.
"required - used" is the number of reserved memslots. For now, the number
of used and required memslots is always equal, and there are no
reservations. This is a preparation for memory devices that want to
dynamically consume memslots after initially specifying how many they
require -- where we'll end up with reserved memslots.
To track the number of used memslots, create a new address space for
our device memory and register a memory listener (add/remove) for that
address space.
Message-ID: <20230926185738.277351-9-david@redhat.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
We want to place non-qmp stubs in there, so let's rename it. While at
it, put it into the MAINTAINERS file under "Memory devices".
Message-ID: <20230926185738.277351-8-david@redhat.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
We want to support memory devices that have a memory region container as
device memory region that maps multiple RAM memory regions. Let's start
by supporting memory devices that statically map multiple RAM memory
regions and, thereby, consume multiple memslots.
We already have one device that uses a container as device memory region:
NVDIMMs. However, a NVDIMM always ends up consuming exactly one memslot.
Let's add support for that by asking the memory device via a new
callback how many memslots it requires.
Message-ID: <20230926185738.277351-7-david@redhat.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Let's return the number of free slots instead of only checking if there
is a free slot. Required to support memory devices that consume multiple
memslots.
This is a preparation for memory devices that consume multiple memslots.
Message-ID: <20230926185738.277351-6-david@redhat.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Let's return the number of free slots instead of only checking if there
is a free slot. While at it, check all address spaces, which will also
consider SMM under x86 correctly.
This is a preparation for memory devices that consume multiple memslots.
Message-ID: <20230926185738.277351-5-david@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Let's fixup the documentation (e.g., removing traces of the ram_addr
parameter that no longer exists) and move it to the header file while at
it.
Message-ID: <20230926185738.277351-4-david@redhat.com>
Suggested-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Checking whether the memory regions are equal is sufficient: if they are
equal, then most certainly the contained fd is equal.
The whole vhost-user memslot handling is suboptimal and overly
complicated. We shouldn't have to lookup a RAM memory regions we got
notified about in vhost_user_get_mr_data() using a host pointer. But that
requires a bigger rework -- especially an alternative vhost_set_mem_table()
backend call that simply consumes MemoryRegionSections.
For now, let's just drop vhost_backend_can_merge().
Message-ID: <20230926185738.277351-3-david@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Having multiple vhost devices, some filtering out fd-less memslots and
some not, can mess up the "used_memslot" accounting. Consequently our
"free memslot" checks become unreliable and we might run out of free
memslots at runtime later.
An example sequence which can trigger a potential issue that involves
different vhost backends (vhost-kernel and vhost-user) and hotplugged
memory devices can be found at [1].
Let's make the filtering mechanism less generic and distinguish between
backends that support private memslots (without a fd) and ones that only
support shared memslots (with a fd). Track the used_memslots for both
cases separately and use the corresponding value when required.
Note: Most probably we should filter out MAP_PRIVATE fd-based RAM regions
(for example, via memory-backend-memfd,...,shared=off or as default with
memory-backend-file) as well. When not using MAP_SHARED, it might not work
as expected. Add a TODO for now.
[1] https://lkml.kernel.org/r/fad9136f-08d3-3fd9-71a1-502069c000cf@redhat.com
Message-ID: <20230926185738.277351-2-david@redhat.com>
Fixes: 988a27754b ("vhost: allow backends to filter memory sections")
Cc: Tiwei Bie <tiwei.bie@intel.com>
Acked-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
We already list system/qtest.c in the qtest section, so the
corresponding header file should be listed here, too.
Message-ID: <20231012111401.871711-1-thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Associate i2c-echo with TEST_DEVICES and add a dependency on I2C.
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-ID: <20230823-i2c-echo-fixes-v1-2-ccc05a6028f0@samsung.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Add missing copyright and license notice. Also add a short description
of the device.
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Message-ID: <20230823-i2c-echo-fixes-v1-1-ccc05a6028f0@samsung.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
npcm7xx_timer-test occasionally fails due to the state of the timers
from the previous test iteration. Advancing the clock step after the
reset resolves this issue.
Fixes: https://gitlab.com/qemu-project/qemu/-/issues/1897
Signed-off-by: Chris Rauer <crauer@google.com>
Message-ID: <20230929000831.691559-1-crauer@google.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Guest driver allocates and initialize page tables to be used as a ring
of descriptors for CQ and async events.
The page table that represents the ring, along with the number of pages
in the page table is passed to the device.
Currently our device supports only one page table for a ring.
Let's make sure that the number of page table entries the driver
reports, do not exceeds the one page table size.
Reported-by: Soul Chen <soulchen8650@gmail.com>
Signed-off-by: Yuval Shaia <yuval.shaia.ml@gmail.com>
Fixes: CVE-2023-1544
Message-ID: <20230301142926.18686-1-yuval.shaia.ml@gmail.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Avoid using trivial variable names in macros, otherwise we get
the following compiler warning when compiling with -Wshadow=local:
In file included from ../../qemu/hw/display/virtio-gpu-virgl.c:19:
../../home/thuth/devel/qemu/hw/display/virtio-gpu-virgl.c:
In function ‘virgl_cmd_submit_3d’:
../../qemu/include/hw/virtio/virtio-gpu.h:228:16: error: declaration of ‘s’
shadows a previous local [-Werror=shadow=compatible-local]
228 | size_t s;
| ^
../../qemu/hw/display/virtio-gpu-virgl.c:215:5: note: in expansion of macro
‘VIRTIO_GPU_FILL_CMD’
215 | VIRTIO_GPU_FILL_CMD(cs);
| ^~~~~~~~~~~~~~~~~~~
../../qemu/hw/display/virtio-gpu-virgl.c:213:12: note: shadowed declaration
is here
213 | size_t s;
| ^
cc1: all warnings being treated as errors
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-ID: <20231009084559.41427-1-thuth@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Rename shadowing variables to make this code compilable
with -Wshadow=local.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-ID: <20231006121129.487251-1-thuth@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
No need to declare a new variable with the same name here,
we can simple re-use the one from the top of the function.
With this change, the file now compiles fine with -Wshadow=local.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-ID: <20231006120819.480792-1-thuth@redhat.com>
Reviewed-by: Xie Yongji <xieyongji@bytedance.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
The operator (fwmacc16) of vfwmaccbf16.vf helper function should be
replaced by fwmaccbf16.
Fixes: adf772b0f7 ("target/riscv: Add support for Zvfbfwma extension")
Signed-off-by: Max Chou <max.chou@sifive.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20231005095734.567575-1-max.chou@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
At this moment there are eleven CPU extension properties that starts
with capital 'Z': Zifencei, Zicsr, Zihintntl, Zihintpause, Zawrs, Zfa,
Zfh, Zfhmin, Zve32f, Zve64f and Zve64d. All other extensions are named
with lower-case letters.
We want all properties to be named with lower-case letters since it's
consistent with the riscv-isa string that we create in the FDT. Having
these 11 properties to be exceptions can be confusing.
Deprecate all of them. Create their lower-case counterpart to be used as
maintained CPU properties. When trying to use any deprecated property a
warning message will be displayed, recommending users to switch to the
lower-case variant:
./build/qemu-system-riscv64 -M virt -cpu rv64,Zifencei=true --nographic
qemu-system-riscv64: warning: CPU property 'Zifencei' is deprecated. Please use 'zifencei' instead
This will give users some time to change their scripts before we remove
the capital 'Z' properties entirely.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Message-ID: <20231009112817.8896-2-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
RISCV_CPU(cs) uses a checked cast. When QOM cast debugging is enabled
this adds about 5% total overhead when emulating RV64 on x86-64 host.
Using a RISC-V guest with 16 vCPUs, 16 GB of guest RAM, virtio-blk
disk. The guest has a copy of the qemu source tree. The test
involves compiling the qemu source tree with 'make clean; time make -j16'.
Before making this change the compile step took 449 & 447 seconds over
two consecutive runs.
After making this change: 428 & 421 seconds.
The saving is over 5%.
Thanks: Paolo Bonzini
Thanks: Philippe Mathieu-Daudé
Signed-off-by: Richard W.M. Jones <rjones@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20231009124859.3373696-2-rjones@redhat.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Vendor CPUs that set RVG are displaying user warnings about other
extensions that RVG must enable, one warning per CPU. E.g.:
$ ./build/qemu-system-riscv64 -smp 8 -M virt -cpu veyron-v1 -nographic
qemu-system-riscv64: warning: Setting G will also set IMAFD_Zicsr_Zifencei
qemu-system-riscv64: warning: Setting G will also set IMAFD_Zicsr_Zifencei
qemu-system-riscv64: warning: Setting G will also set IMAFD_Zicsr_Zifencei
qemu-system-riscv64: warning: Setting G will also set IMAFD_Zicsr_Zifencei
qemu-system-riscv64: warning: Setting G will also set IMAFD_Zicsr_Zifencei
qemu-system-riscv64: warning: Setting G will also set IMAFD_Zicsr_Zifencei
qemu-system-riscv64: warning: Setting G will also set IMAFD_Zicsr_Zifencei
qemu-system-riscv64: warning: Setting G will also set IMAFD_Zicsr_Zifencei
This happens because we decided a while ago that, for simplicity, vendor
CPUs could set RVG instead of setting each G extension individually in
their cpu_init(). Our warning isn't taking that into account, and we're
bugging users with a warning that we're causing ourselves.
In a closer look we conclude that this warning is not warranted in any
other circumstance since we're just following the ISA [1], which states
in chapter 24:
"One goal of the RISC-V project is that it be used as a stable software
development target. For this purpose, we define a combination of a base
ISA (RV32I or RV64I) plus selected standard extensions (IMAFD, Zicsr,
Zifencei) as a 'general-purpose' ISA, and we use the abbreviation G for
the IMAFDZicsr Zifencei combination of instruction-set extensions."
With this in mind, enabling IMAFD_Zicsr_Zifencei if the user explicitly
enables 'G' is an expected behavior and the warning is unneeded. Any
user caught by surprise should refer to the ISA.
Remove the warning when handling RVG.
[1] https://github.com/riscv/riscv-isa-manual/releases/download/Ratified-IMAFDQC/riscv-spec-20191213.pdf
Reported-by: Paul A. Clarke <pclarke@ventanamicro.com>
Suggested-by: Andrew Jones <ajones@ventanamicro.com>
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Message-ID: <20231003122539.775932-1-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
KVM for RISC-V started supporting KVM_GET_REG_LIST in Linux 6.6. It
consists of a KVM ioctl() that retrieves a list of all available regs
for get_one_reg/set_one_reg. Regs that aren't present in the list aren't
supported in the host.
This simplifies our lives when initing the KVM regs since we don't have
to always attempt a KVM_GET_ONE_REG for all regs QEMU knows. We'll only
attempt a get_one_reg() if we're sure the reg is supported, i.e. it was
retrieved by KVM_GET_REG_LIST. Any error in get_one_reg() will then
always considered fatal, instead of having to handle special error codes
that might indicate a non-fatal failure.
Start by moving the current kvm_riscv_init_multiext_cfg() logic into a
new kvm_riscv_read_multiext_legacy() helper. We'll prioritize using
KVM_GET_REG_LIST, so check if we have it available and, in case we
don't, use the legacy() logic.
Otherwise, retrieve the available reg list and use it to check if the
host supports our known KVM regs, doing the usual get_one_reg() for
the supported regs and setting cpu->cfg accordingly.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Message-ID: <20231003132148.797921-3-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Our error message is returning the value of 'ret', which will be always
-1 in case of error, and will not be that useful:
qemu-system-riscv64: Unable to read ISA_EXT KVM register ssaia, error -1
Improve the error message by outputting 'errno' instead of 'ret'. Use
strerrorname_np() to output the error name instead of the error code.
This will give us what we need to know right away:
qemu-system-riscv64: Unable to read ISA_EXT KVM register ssaia, error code: ENOENT
Given that we're going to exit(1) in this condition instead of
attempting to recover, remove the 'kvm_riscv_destroy_scratch_vcpu()'
call.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20231003132148.797921-2-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
This replaces the exit calls by shutdown requests, ensuring a proper
cleanup of Qemu. Features like net/vhost-vdpa.c are expecting
qemu_cleanup to be called to remove their last residuals.
Signed-off-by: Clément Chigot <chigot@adacore.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20231003071427.188697-6-chigot@adacore.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
This replaces the exit calls by shutdown requests, ensuring a proper
cleanup of Qemu. Otherwise, some connections like gdb could be broken
before its final packet ("Wxx") is being sent. This part, being done
inside qemu_cleanup function, can be reached only when the main loop
exits after a shutdown request.
Signed-off-by: Clément Chigot <chigot@adacore.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20231003071427.188697-5-chigot@adacore.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
This replaces the exit calls by shutdown requests, ensuring a proper
cleanup of Qemu. Otherwise, some connections like gdb could be broken
before its final packet ("Wxx") is being sent. This part, being done
inside qemu_cleanup function, can be reached only when the main loop
exits after a shutdown request.
Signed-off-by: Clément Chigot <chigot@adacore.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20231003071427.188697-4-chigot@adacore.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
gdb_exit function aims to close gdb sessions and sends the exit code of
the current execution. It's being called by qemu_cleanup once the main
loop is over.
Until now, the exit code sent was always 0. Now that hardware can
shutdown this main loop with custom exit codes, these codes must be
transfered to gdb as well.
Signed-off-by: Clément Chigot <chigot@adacore.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20231003071427.188697-3-chigot@adacore.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
As of now, the exit code was either EXIT_FAILURE when a panic shutdown
was requested or EXIT_SUCCESS otherwise.
However, some hardware could want to pass more complex exit codes. Thus,
introduce a new shutdown request function allowing that.
Signed-off-by: Clément Chigot <chigot@adacore.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20231003071427.188697-2-chigot@adacore.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
At this moment we do not expose extension properties for vendor CPUs
because that would allow users to change them via command line. The
drawback is that if we were to add an API that shows all CPU properties,
e.g. qmp-query-cpu-model-expansion, we won't be able to show extensions
state of vendor CPUs.
We have the required machinery to create extension properties for vendor
CPUs while not allowing users to enable extensions. Disabling existing
extensions is allowed since it can be useful for debugging.
Change the set() callback cpu_set_multi_ext_cfg() to allow enabling
extensions only for generic CPUs. In cpu_add_multi_ext_prop() let's not
set the default values for the properties if we're not dealing with
generic CPUs, otherwise the values set in cpu_init() of vendor CPUs will
be overwritten. And finally, in tcg_cpu_instance_init(), add cpu user
properties for all CPUs.
For the veyron-v1 CPU, we're now able to disable existing extensions
like smstateen:
$ ./build/qemu-system-riscv64 --nographic -M virt \
-cpu veyron-v1,smstateen=false
But setting extensions that the CPU didn't set during cpu_init(), like
V, is not allowed:
$ ./build/qemu-system-riscv64 --nographic -M virt \
-cpu veyron-v1,v=true
qemu-system-riscv64: can't apply global veyron-v1-riscv-cpu.v=true:
'veyron-v1' CPU does not allow enabling extensions
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230926183109.165878-3-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
We'll introduce generic errors that will output a CPU type name via its
RISCVCPU pointer. Create a helper for that.
Use the helper in tcg_cpu_realizefn() instead of hardcoding the 'host'
CPU name.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230926183109.165878-2-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Priv spec validation is TCG specific. Move it to the TCG accel class.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20230925175709.35696-20-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
This array will be read by the TCG accel class, allowing it to handle
priv spec verifications on its own. The array will remain here in cpu.c
because it's also used by the riscv,isa string function.
To export it we'll finish it with an empty element since ARRAY_SIZE()
won't work outside of cpu.c. Get rid of its ARRAY_SIZE() usage now to
alleviate the changes for the next patch.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-19-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
All code related to MISA TCG properties is also moved.
At this point, all TCG properties handling is done in tcg-cpu.c, all KVM
properties handling is done in kvm-cpu.c.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-18-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
The array isn't marked as 'const' because we're initializing their
elements in riscv_cpu_add_misa_properties(), 'name' and 'description'
fields.
In a closer look we can see that we're not using these 2 fields after
creating the MISA properties. And we can create the properties by using
riscv_get_misa_ext_name() and riscv_get_misa_ext_description()
directly.
Remove the 'name' and 'description' fields from RISCVCPUMisaExtConfig
and make misa_ext_cfgs[] a const array.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-17-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
tcg_cpu_instance_init() will be the 'cpu_instance_init' impl for the TCG
accelerator. It'll be called from within riscv_cpu_post_init(), via
accel_cpu_instance_init(), similar to what happens with KVM. In fact, to
preserve behavior, the implementation will be similar to what
riscv_cpu_post_init() already does.
In this patch we'll move riscv_cpu_add_user_properties() and
riscv_init_max_cpu_extensions() and all their dependencies to tcg-cpu.c.
All multi-extension properties code was moved. The 'multi_ext_user_opts'
hash table was also moved to tcg-cpu.c since it's a TCG only structure,
meaning that we won't have to worry about initializing a TCG hash table
when running a KVM CPU anymore.
riscv_cpu_add_user_properties() will remain in cpu.c for now due to how
much code it requires to be moved at the same time. We'll do that in the
next patch.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-16-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
We'll move riscv_init_max_cpu_extensions() to tcg-cpu.c in the next
patch and set_misa() needs to be usable from there.
Rename it to riscv_cpu_set_misa() and make it public.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-15-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
riscv_cpu_add_misa_properties() is being used to fill the missing KVM
MISA properties but it is a TCG helper that was adapted to do so. We'll
move it to tcg-cpu.c in the next patches, meaning that KVM needs to fill
the remaining MISA properties on its own.
Do not use riscv_cpu_add_misa_properties(). Let's create a new array
with all available MISA bits we support that can be read by KVM. The
array is zero terminate to allow us to iterate through it without
knowing its size.
Then, inside kvm_riscv_add_cpu_user_properties(), we'll create all KVM
MISA properties as usual and then use this array to add any missing MISA
properties with the riscv_cpu_add_kvm_unavail_prop() helper.
Note that we're creating misa_bits[], and not using the existing
'riscv_single_letter_exts[]', because the latter is tuned for riscv,isa
related functions and it doesn't have all MISA bits we support. Commit
0e2c377023 ("target/riscv: misa to ISA string conversion fix") has the
full context.
While we're at it, move both satp and the multi-letter extension
properties to kvm_riscv_add_cpu_user_properties() as well.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-14-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Move the files to a 'kvm' dir to promote more code separation between
accelerators and making our lives easier supporting build options such
as --disable-tcg.
Rename kvm.c to kvm-cpu.c to keep it in line with its TCG counterpart.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-13-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Add a KVM accelerator class like we did with TCG. The difference is
that, at least for now, we won't be using a realize() implementation for
this accelerator.
We'll start by assiging kvm_riscv_cpu_add_kvm_properties(), renamed to
kvm_cpu_instance_init(), as a 'cpu_instance_init' implementation. Change
riscv_cpu_post_init() to invoke accel_cpu_instance_init(), which will go
through the 'cpu_instance_init' impl of the current acceleration (if
available) and execute it. The end result is that the KVM initial setup,
i.e. starting registers and adding its specific properties, will be done
via this hook.
Add a 'tcg_enabled()' condition in riscv_cpu_post_init() to avoid
calling riscv_cpu_add_user_properties() when running KVM. We'll remove
this condition when the TCG accel class get its own 'cpu_instance_init'
implementation.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-12-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
This file is not needed for some time now. Both kvm_riscv_reset_vcpu()
and kvm_riscv_set_irq() have public declarations in kvm_riscv.h and are
wrapped in 'if kvm_enabled()' blocks that the compiler will rip it out
in non-KVM builds.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-11-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
This function is used for both accelerators. Make it public, and call it
from kvm_riscv_cpu_add_kvm_properties(). This will make it easier to
split KVM specific code for the KVM accelerator class in the next patch.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-10-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
We'll introduce the KVM accelerator class with a 'cpu_instance_init'
implementation that is going to be invoked during the common
riscv_cpu_post_init() (via accel_cpu_instance_init()). This
instance_init will execute KVM exclusive code that TCG doesn't care
about, such as adding KVM specific properties, initing registers using a
KVM scratch CPU and so on.
The core of the forementioned cpu_instance_init impl is the current
riscv_cpu_add_kvm_properties() that is being used by the common code via
riscv_cpu_add_user_properties() in cpu.c. Move it to kvm.c, together
will all the relevant artifacts, exporting and renaming it to
kvm_riscv_cpu_add_kvm_properties() so cpu.c can keep using it for now.
To make this work we'll need to export riscv_cpu_extensions,
riscv_cpu_vendor_exts and riscv_cpu_experimental_exts from cpu.c as
well. The TCG accelerator will also need to access those in the near
future so this export will benefit us in the long run.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-9-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
We'll need to export these arrays to the accelerator classes in the next
patches. Mark them as 'const' now because they should not be modified at
runtime.
Note that 'riscv_cpu_options' will also be exported, but can't be marked
as 'const', because the properties are changed via
qdev_property_add_static().
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-8-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
This CPU only exists if we're compiling with KVM so move it to the kvm
specific file.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-7-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
All generic CPUs call riscv_cpu_add_user_properties(). The 'max' CPU
calls riscv_init_max_cpu_extensions(). Both can be moved to a common
instance_post_init() callback, implemented in riscv_cpu_post_init(),
called by all CPUs. The call order then becomes:
riscv_cpu_init() -> cpu_init() of each CPU -> .instance_post_init()
In the near future riscv_cpu_post_init() will call the init() function
of the current accelerator, providing a hook for KVM and TCG accel
classes to change the init() process of the CPU.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-6-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Move the remaining of riscv_tcg_ops now that we have a working realize()
implementation.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20230925175709.35696-5-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
