target/i386/sev: Fix printf formats

hwaddr uses HWADDR_PRIx, sizeof yields size_t so uses %zu,
and gsize uses G_GSIZE_FORMAT.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Link: https://lore.kernel.org/r/20240626194950.1725800-4-richard.henderson@linaro.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/sev: Use size_t for object sizes

This code was using both uint32_t and uint64_t for len.
Consistently use size_t instead.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Link: https://lore.kernel.org/r/20240626194950.1725800-3-richard.henderson@linaro.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: SEV: store pointer to decoded id_auth in SevSnpGuest

Do not rely on finish->id_auth_uaddr, so that there are no casts from
pointer to uint64_t.  They break on 32-bit hosts.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: SEV: rename sev_snp_guest->id_auth

Free the "id_auth" name for the binary version of the data.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: SEV: store pointer to decoded id_block in SevSnpGuest

Do not rely on finish->id_block_uaddr, so that there are no casts from
pointer to uint64_t.  They break on 32-bit hosts.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: SEV: rename sev_snp_guest->id_block

Free the "id_block" name for the binary version of the data.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: remove unused enum

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: give CC_OP_POPCNT low bits corresponding to MO_TL

Handle it like the other arithmetic cc_ops.  This simplifies a
bit the implementation of bit test instructions.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: use cpu_cc_dst for CC_OP_POPCNT

It is the only CCOp, among those that compute ZF from one of the cc_op_*
registers, that uses cpu_cc_src.  Do not make it the odd one off,
instead use cpu_cc_dst like the others.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: fix CC_OP dump

POPCNT was missing, and the entries were all out of order after
ADCX/ADOX/ADCOX were moved close to EFLAGS.  Just use designated
initializers.

Fixes: 4885c3c4953 ("target-i386: Use ctpop helper", 2017-01-10)
Fixes: cc155f19717 ("target/i386: rewrite flags writeback for ADCX/ADOX", 2024-06-11)
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

include: move typeof_strip_qual to compiler.h, use it in QAPI_LIST_LENGTH()

The typeof_strip_qual() is most useful for the atomic fetch-and-modify
operations in atomic.h, but it can be used elsewhere as well.  For example,
QAPI_LIST_LENGTH() assumes that the argument is not const, which is not a
requirement.

Move the macro to compiler.h and, while at it, move it under #ifndef
__cplusplus to emphasize that it uses C-only constructs.  A C++ version
of typeof_strip_qual() using type traits is possible[1], but beyond the
scope of this patch because the little C++ code that is in QEMU does not
use QAPI.

The patch was tested by changing the declaration of strv_from_str_list()
in qapi/qapi-type-helpers.c to:

    char **strv_from_str_list(const strList *const list)

This is valid C code, and it fails to compile without this change.

[1] https://lore.kernel.org/qemu-devel/20240624205647.112034-1-flwu@google.com/

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Tested-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

exec: don't use void* in pointer arithmetic in headers

void* pointer arithmetic is a GCC extentension which could not be
available in other build tools (e.g. C++). This changes removes this
assumption.

Signed-off-by: Roman Kiryanov <rkir@google.com>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Link: https://lore.kernel.org/r/20240620201654.598024-1-rkir@google.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

exec: avoid using C++ keywords in function parameters

to use the QEMU headers with a C++ compiler.

Signed-off-by: Roman Kiryanov <rkir@google.com>
Link: https://lore.kernel.org/r/20240618224553.878869-1-rkir@google.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

block: rename former bdrv_file_open callbacks

Since there is no bdrv_file_open callback anymore, rename the implementations
so that they end with "_open" instead of "_file_open".  NFS is the exception
because all the functions are named nfs_file_*.

Suggested-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

block: remove separate bdrv_file_open callback

bdrv_file_open and bdrv_open are completely equivalent, they are
never checked except to see which one to invoke.  So merge them
into a single one.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

block: do not check bdrv_file_open

The set of BlockDrivers that have .bdrv_file_open coincides with those
that have .protocol_name and guess what---checking drv->bdrv_file_open
is done to see if the driver is a protocol.  So check drv->protocol_name
instead.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

block: make assertion more generic

.bdrv_needs_filename is only set for drivers that also set bdrv_file_open,
i.e. protocol drivers.

So we can make the assertion always, it will always pass for those drivers
that use bdrv_open.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

meson: remove dead optimization option

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

meson: allow configuring the x86-64 baseline

Add a Meson option to configure which x86-64 instruction
set to use.  QEMU will now default to x86-64-v1 + cmpxchg16b for
64-bit builds (that corresponds to a Pentium 4 for 32-bit builds).

The baseline can be tuned down to Pentium Pro for 32-bit builds (with
-Dx86_version=0), or up as desired.

Acked-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Revert "host/i386: assume presence of SSE2"

This reverts commit b18236897ca15c3db1506d8edb9a191dfe51429c.
The x86-64 instruction set can now be tuned down to x86-64 v1
or i386 Pentium Pro.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Revert "host/i386: assume presence of SSSE3"

This reverts commit 433cd6d94a8256af70a5200f236dc8047c3c1468.
The x86-64 instruction set can now be tuned down to x86-64 v1
or i386 Pentium Pro.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Revert "host/i386: assume presence of POPCNT"

This reverts commit 45ccdbcb24baf99667997fac5cf60318e5e7db51.
The x86-64 instruction set can now be tuned down to x86-64 v1
or i386 Pentium Pro.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: detect --cpu=mipsisa64r6

Treat it as a MIPS64 machine.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Merge tag 'pull-riscv-to-apply-20240627-1' of https://github.com/alistair23/qemu into staging

RISC-V PR for 9.1

* Extend virtual irq csrs masks to be 64 bit wide
* Move Guest irqs out of the core local irqs range
* zvbb implies zvkb
* virt: add address-cells in create_fdt_one_aplic()
* virt: add aplic nodename helper
* virt: rename aplic nodename to 'interrupt-controller'
* virt: aplic DT: add 'qemu, aplic' to 'compatible'
* virt: aplic DT: rename prop to 'riscv, delegation'
* virt: change imsic nodename to 'interrupt-controller'
* virt: imsics DT: add 'qemu, imsics' to 'compatible'
* virt: imsics DT: add '#msi-cells'
* QEMU support for KVM Guest Debug on RISC-V
* Support RISC-V privilege 1.13 spec
* Add support for RISC-V ACPI tests
* Modularize common match conditions for trigger

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEEaukCtqfKh31tZZKWr3yVEwxTgBMFAmZ9OCEACgkQr3yVEwxT
# gBM6aRAApE7Cjo6U1MrcXywh897DnMV9TT9a0jWxhw659eOSk8Uo8mXNNmGSudcl
# +//jIQr/JB5YVDcnFAGWFDaMMev2hkbXG68IaLsfqA32CWZzrOgFpFP+sgicDKmP
# A+P/z0CL332hDRxlhglbIYukCN1bEjWbUDgZaXPVc7ieMM0mwDyFZt7jH65dNmNg
# HpmVcUSooUx8weHOKp0vCKCBcQ6neahjvweTYszsWjXgu51+VrpoSAAds98HHY9H
# vilbTGtJzGEmnfTCmm1GmRn5/g5iQzDKDge+Kg+OzINEfoOBByJerK27sGNRkzxt
# ZNUd0RM2q1wkfZ3XfLU3qEnxHson8SQp2+VYTqnEv3TdR0AjrnsxL7IUuFm+iTv4
# wG3IqBwt+efdAzc9k+K2smCyp3036HNrJHQSyhtxu+yU36K+jBKVq++pRzgOFTvl
# 87VvNI/dYL4hSJ4yS00ti2axl6GNEAEePHwQ0fRXvEdSGc4vOw3ayDqyqF0Gcy9+
# MJq+nV7BX9pEyTRZEKqTi2nB7xh3kq0e+mev2ByQlracBqHIayXC3DsDvPcUbiwJ
# 8bzX28kl5C28vYUVJjzYQgZy2XXGbbsnv3ifah4hi2NJESmYHN0uxs9raW9i+lhR
# 2FAm+eO6mxuJTIH/NTou+msxZJw2jTPLkYdG2e5UkTjUN7JKGnk=
# =bag2
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 27 Jun 2024 03:00:01 AM PDT
# gpg:                using RSA key 6AE902B6A7CA877D6D659296AF7C95130C538013
# gpg: Good signature from "Alistair Francis <alistair@alistair23.me>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 6AE9 02B6 A7CA 877D 6D65  9296 AF7C 9513 0C53 8013

* tag 'pull-riscv-to-apply-20240627-1' of https://github.com/alistair23/qemu: (32 commits)
  target/riscv: Apply modularized matching conditions for icount trigger
  target/riscv: Apply modularized matching conditions for watchpoint
  target/riscv: Add functions for common matching conditions of trigger
  target/riscv: Remove extension auto-update check statements
  target/riscv: Add Zc extension implied rule
  target/riscv: Add multi extension implied rules
  target/riscv: Add MISA extension implied rules
  target/riscv: Introduce extension implied rule helpers
  target/riscv: Introduce extension implied rules definition
  target/riscv: fix instructions count handling in icount mode
  target/riscv: Fix froundnx.h nanbox check
  hw/riscv/virt.c: Make block devices default to virtio
  target/riscv: Support the version for ss1p13
  target/riscv: Reserve exception codes for sw-check and hw-err
  target/riscv: Add MEDELEGH, HEDELEGH csrs for RV32
  target/riscv: Add 'P1P13' bit in SMSTATEEN0
  target/riscv: Define macros and variables for ss1p13
  target/riscv: Reuse the conversion function of priv_spec
  target/riscv/kvm: define TARGET_KVM_HAVE_GUEST_DEBUG
  target/riscv/kvm: handle the exit with debug reason
  ...

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/riscv: Apply modularized matching conditions for icount trigger

We have implemented trigger_common_match(), which checks if the enabled
privilege levels of the trigger match CPU's current privilege level. We
can invoke trigger_common_match() to check the privilege levels of the
type 3 triggers.

Signed-off-by: Alvin Chang <alvinga@andestech.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240626132247.2761286-4-alvinga@andestech.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Apply modularized matching conditions for watchpoint

We have implemented trigger_common_match(), which checks if the enabled
privilege levels of the trigger match CPU's current privilege level.
Remove the related code in riscv_cpu_debug_check_watchpoint() and invoke
trigger_common_match() to check the privilege levels of the type 2 and
type 6 triggers for the watchpoints.

This commit also changes the behavior of looping the triggers. In
previous implementation, if we have a type 2 trigger and
env->virt_enabled is true, we directly return false to stop the loop.
Now we keep looping all the triggers until we find a matched trigger.

Only load/store bits and loaded/stored address should be further checked
in riscv_cpu_debug_check_watchpoint().

Signed-off-by: Alvin Chang <alvinga@andestech.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240626132247.2761286-3-alvinga@andestech.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Add functions for common matching conditions of trigger

According to RISC-V Debug specification version 0.13 [1] (also applied
to version 1.0 [2] but it has not been ratified yet), there are several
common matching conditions before firing a trigger, including the
enabled privilege levels of the trigger.

This commit adds trigger_common_match() to prepare the common matching
conditions for the type 2/3/6 triggers. For now, we just implement
trigger_priv_match() to check if the enabled privilege levels of the
trigger match CPU's current privilege level.

Remove the related code in riscv_cpu_debug_check_breakpoint() and invoke
trigger_common_match() to check the privilege levels of the type 2 and
type 6 triggers for the breakpoints.

This commit also changes the behavior of looping the triggers. In
previous implementation, if we have a type 2 trigger and
env->virt_enabled is true, we directly return false to stop the loop.
Now we keep looping all the triggers until we find a matched trigger.

Only the execution bit and the executed PC should be futher checked in
riscv_cpu_debug_check_breakpoint().

[1]: https://github.com/riscv/riscv-debug-spec/releases/tag/task_group_vote
[2]: https://github.com/riscv/riscv-debug-spec/releases/tag/1.0.0-rc1-asciidoc

Signed-off-by: Alvin Chang <alvinga@andestech.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240626132247.2761286-2-alvinga@andestech.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Remove extension auto-update check statements

Remove the old-fashioned extension auto-update check statements as
they are replaced by the extension implied rules.

Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Jerry Zhang Jian <jerry.zhangjian@sifive.com>
Tested-by: Max Chou <max.chou@sifive.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Message-ID: <20240625114629.27793-7-frank.chang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Add Zc extension implied rule

Zc extension has special implied rules that need to be handled separately.

Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Jerry Zhang Jian <jerry.zhangjian@sifive.com>
Tested-by: Max Chou <max.chou@sifive.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Message-ID: <20240625114629.27793-6-frank.chang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Add multi extension implied rules

Add multi extension implied rules to enable the implied extensions of
the multi extension recursively.

Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Jerry Zhang Jian <jerry.zhangjian@sifive.com>
Tested-by: Max Chou <max.chou@sifive.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Message-ID: <20240625114629.27793-5-frank.chang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Add MISA extension implied rules

Add MISA extension implied rules to enable the implied extensions
of MISA recursively.

Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Jerry Zhang Jian <jerry.zhangjian@sifive.com>
Tested-by: Max Chou <max.chou@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Message-ID: <20240625114629.27793-4-frank.chang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Introduce extension implied rule helpers

Introduce helpers to enable the extensions based on the implied rules.
The implied extensions are enabled recursively, so we don't have to
expand all of them manually. This also eliminates the old-fashioned
ordering requirement. For example, Zvksg implies Zvks, Zvks implies
Zvksed, etc., removing the need to check the implied rules of Zvksg
before Zvks.

Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Jerry Zhang Jian <jerry.zhangjian@sifive.com>
Tested-by: Max Chou <max.chou@sifive.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Message-ID: <20240625114629.27793-3-frank.chang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Introduce extension implied rules definition

RISCVCPUImpliedExtsRule is created to store the implied rules.
'is_misa' flag is used to distinguish whether the rule is derived
from the MISA or other extensions.
'ext' stores the MISA bit if 'is_misa' is true. Otherwise, it stores
the offset of the extension defined in RISCVCPUConfig. 'ext' will also
serve as the key of the hash tables to look up the rule in the following
commit.

Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Jerry Zhang Jian <jerry.zhangjian@sifive.com>
Tested-by: Max Chou <max.chou@sifive.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240625114629.27793-2-frank.chang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: fix instructions count handling in icount mode

When icount is enabled, rather than returning the virtual CPU time, we
should return the instruction count itself. Add an instructions bool
parameter to get_ticks() to correctly return icount_get_raw() when
icount_enabled() == 1 and instruction count is queried. This will modify
the existing behavior which was returning an instructions count close to
the number of cycles (CPI ~= 1).

Signed-off-by: Clément Léger <cleger@rivosinc.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
Message-ID: <20240618112649.76683-1-cleger@rivosinc.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Fix froundnx.h nanbox check

helper_froundnx_h function mistakenly uses single percision nanbox
check instead of the half percision one. This patch fixes the issue.

Signed-off-by: Branislav Brzak <brzakbranislav@gmail.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20240608214546.226963-1-brzakbranislav@gmail.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

hw/riscv/virt.c: Make block devices default to virtio

RISC-V virt is currently missing default type for block devices. Without
this being set, proper backend is not created when option like -cdrom
is used. So, make the virt board's default block device type be
IF_VIRTIO similar to other architectures.

We also need to set no_cdrom to avoid getting a default cdrom device.

Signed-off-by: Sunil V L <sunilvl@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Message-ID: <20240620064718.275427-1-sunilvl@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Support the version for ss1p13

Add RISC-V privilege 1.13 support.

Signed-off-by: Fea.Wang <fea.wang@sifive.com>
Signed-off-by: Fea.Wang <fea.wang@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liwei1518@gmail.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Message-ID: <20240606135454.119186-7-fea.wang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Reserve exception codes for sw-check and hw-err

Based on the priv-1.13.0, add the exception codes for Software-check and
Hardware-error.

Signed-off-by: Fea.Wang <fea.wang@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240606135454.119186-6-fea.wang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Add MEDELEGH, HEDELEGH csrs for RV32

Based on privileged spec 1.13, the RV32 needs to implement MEDELEGH
and HEDELEGH for exception codes 32-47 for reserving and exception codes
48-63 for custom use. Add the CSR number though the implementation is
just reading zero and writing ignore. Besides, for accessing HEDELEGH, it
should be controlled by mstateen0 'P1P13' bit.

Signed-off-by: Fea.Wang <fea.wang@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240606135454.119186-5-fea.wang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Add 'P1P13' bit in SMSTATEEN0

Based on privilege 1.13 spec, there should be a bit56 for 'P1P13' in
mstateen0 that controls access to the hedeleg.

Signed-off-by: Fea.Wang <fea.wang@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liwei1518@gmail.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240606135454.119186-4-fea.wang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Define macros and variables for ss1p13

Add macros and variables for RISC-V privilege 1.13 support.

Signed-off-by: Fea.Wang <fea.wang@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liwei1518@gmail.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240606135454.119186-3-fea.wang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Reuse the conversion function of priv_spec

Public the conversion function of priv_spec in cpu.h, so that tcg-cpu.c
could also use it.

Signed-off-by: Jim Shu <jim.shu@sifive.com>
Signed-off-by: Fea.Wang <fea.wang@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240606135454.119186-2-fea.wang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv/kvm: define TARGET_KVM_HAVE_GUEST_DEBUG

To enable the KVM GUEST DEBUG for RISC-V at QEMU side.

Signed-off-by: Chao Du <duchao@eswincomputing.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240606014501.20763-4-duchao@eswincomputing.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv/kvm: handle the exit with debug reason

If the breakpoint belongs to the userspace then set the ret value.

Signed-off-by: Chao Du <duchao@eswincomputing.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240606014501.20763-3-duchao@eswincomputing.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv/kvm: add software breakpoints support

This patch implements insert/remove software breakpoint process.

For RISC-V, GDB treats single-step similarly to breakpoint: add a
breakpoint at the next step address, then continue. So this also
works for single-step debugging.

Implement kvm_arch_update_guest_debug(): Set the control flag
when there are active breakpoints. This will help KVM to know
the status in the userspace.

Add some stubs which are necessary for building, and will be
implemented later.

Signed-off-by: Chao Du <duchao@eswincomputing.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240606014501.20763-2-duchao@eswincomputing.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

hw/riscv/virt.c: imsics DT: add '#msi-cells'

The DT docs for riscv,imsics [1] requires a 'msi-cell' property. Add one
and set it zero.

[1] Documentation/devicetree/bindings/interrupt-controller/riscv,imsics.yaml

Reported-by: Conor Dooley <conor@kernel.org>
Fixes: 28d8c281200f ("hw/riscv: virt: Add optional AIA IMSIC support to virt machine")
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Message-ID: <20240531202759.911601-9-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

hw/riscv/virt.c: imsics DT: add 'qemu, imsics' to 'compatible'

The DT docs for riscv,imsics [1] predicts a 'qemu,imsics' enum in the
'compatible' property.

[1] Documentation/devicetree/bindings/interrupt-controller/riscv,imsics.yaml

Reported-by: Conor Dooley <conor@kernel.org>
Fixes: 28d8c281200f ("hw/riscv: virt: Add optional AIA IMSIC support to virt machine")
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240531202759.911601-8-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

hw/riscv/virt.c: change imsic nodename to 'interrupt-controller'

The Linux DT docs for imsic [1] predicts an 'interrupt-controller@addr'
node, not 'imsic@addr', given this node inherits the
'interrupt-controller' node.

[1] Documentation/devicetree/bindings/interrupt-controller/riscv,imsics.yaml

Reported-by: Conor Dooley <conor@kernel.org>
Fixes: 28d8c281200f ("hw/riscv: virt: Add optional AIA IMSIC support to virt machine")
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240531202759.911601-7-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

hw/riscv/virt.c: aplic DT: rename prop to 'riscv, delegation'

The DT docs for riscv,aplic [1] predicts a 'riscv,delegation' property.
Not 'riscv,delegate'.

[1] Documentation/devicetree/bindings/interrupt-controller/riscv,aplic.yaml

Reported-by: Conor Dooley <conor@kernel.org>
Fixes: e6faee65855b ("hw/riscv: virt: Add optional AIA APLIC support to virt machine")
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240531202759.911601-6-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

hw/riscv/virt.c: aplic DT: add 'qemu, aplic' to 'compatible'

The DT docs for riscv,aplic [1] predicts a 'qemu,aplic' enum in the
'compatible' property.

[1] Documentation/devicetree/bindings/interrupt-controller/riscv,aplic.yaml

Reported-by: Conor Dooley <conor@kernel.org>
Fixes: e6faee65855b ("hw/riscv: virt: Add optional AIA APLIC support to virt machine")
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240531202759.911601-5-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

hw/riscv/virt.c: rename aplic nodename to 'interrupt-controller'

The correct name of the aplic controller node, as per Linux kernel DT
docs [1], is 'interrupt-controller@addr'.

[1] Documentation/devicetree/bindings/interrupt-controller/riscv,aplic.yaml

Reported-by: Conor Dooley <conor@kernel.org>
Fixes: e6faee65855b ("hw/riscv: virt: Add optional AIA APLIC support to virt machine")
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240531202759.911601-4-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

hw/riscv/virt.c: add aplic nodename helper

We'll change the aplic DT nodename in the next patch and the name is
hardcoded in 2 different functions. Create a helper to change a single
place later.

While we're at it, in create_fdt_socket_aplic(), move 'aplic_name'
inside the conditional to avoid allocating a string that won't be used
when socket == NULL.

Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240531202759.911601-3-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

hw/riscv/virt.c: add address-cells in create_fdt_one_aplic()

We need #address-cells properties in all interrupt controllers that are
referred by an interrupt-map [1]. For the RISC-V machine, both PLIC and
APLIC controllers must have this property.

PLIC already sets it in create_fdt_socket_plic(). Set the property for
APLIC in create_fdt_one_aplic().

[1] https://lore.kernel.org/linux-arm-kernel/CAL_JsqJE15D-xXxmELsmuD+JQHZzxGzdXvikChn6KFWqk6NzPw@mail.gmail.com/

Suggested-by: Anup Patel <apatel@ventanamicro.com>
Fixes: e6faee65855b ("hw/riscv: virt: Add optional AIA APLIC support to virt machine")
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240531202759.911601-2-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: zvbb implies zvkb

According to RISC-V crypto spec, Zvkb extension is a
subset of the Zvbb extension [1].

1: https://github.com/riscv/riscv-crypto/blob/1769c2609bf4535632e0c0fd715778f212bb272e/doc/vector/riscv-crypto-vector-zvkb.adoc?plain=1#L10

Signed-off-by: Jerry Zhang Jian <jerry.zhangjian@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240528130349.20193-1-jerry.zhangjian@sifive.com>
[ Changes by AF:
 - Tidy up commit message
 - Rebase
]
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Move Guest irqs out of the core local irqs range.

Qemu maps IRQs 0:15 for core interrupts and 16 onward for
guest interrupts which are later translated to hgiep in
`riscv_cpu_set_irq()` function.

With virtual IRQ support added, software now can fully
use the whole local interrupt range without any actual
hardware attached.

This change moves the guest interrupt range after the
core local interrupt range to avoid clash.

Fixes: 1697837ed9 ("target/riscv: Add M-mode virtual interrupt and IRQ filtering support.")
Fixes: 40336d5b1d ("target/riscv: Add HS-mode virtual interrupt and IRQ filtering support.")
Signed-off-by: Rajnesh Kanwal <rkanwal@rivosinc.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Message-ID: <20240520125157.311503-3-rkanwal@rivosinc.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv: Extend virtual irq csrs masks to be 64 bit wide.

AIA extends the width of all IRQ CSRs to 64bit even
in 32bit systems by adding missing half CSRs.

This seems to be missed while adding support for
virtual IRQs. The whole logic seems to be correct
except the width of the masks.

Fixes: 1697837ed9 ("target/riscv: Add M-mode virtual interrupt and IRQ filtering support.")
Fixes: 40336d5b1d ("target/riscv: Add HS-mode virtual interrupt and IRQ filtering support.")
Signed-off-by: Rajnesh Kanwal <rkanwal@rivosinc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240520125157.311503-2-rkanwal@rivosinc.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

Merge tag 'pull-vfio-20240624' of https://github.com/legoater/qemu into staging

vfio queue:

* Add a host IOMMU device abstraction
* VIRTIO-IOMMU/VFIO: Fix host iommu geometry handling
* QOMify VFIOContainer

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEEoPZlSPBIlev+awtgUaNDx8/77KEFAmZ541QACgkQUaNDx8/7
# 7KFdnQ/8Dih3HI2qtY93bTxg0lmJ+ZMibojTkEkTu3kSvwoI12wkiSMFKzzTWpZE
# UtGyIqQQij8IfQtIz87uQskv7oFiZKG6JWMTAX4uJ8ZIgZiih29/e/38VGEbogBh
# yO+1Pqr3ETlyLnQcu9ruBTJ293LXovmD4d9feoaVdURBNZ1EqIh7sv/y7YdUsR+i
# tXa6kW1ZIlKBI54o/uuODHWQYyOHs39VtZ6JZvgxVVEQsNikcJsosK9ts9A1EByi
# 0roQVXm2QAK/nPXlmMGLvJWzQcdeXQ6W6hzYkO2HqGnCLURnpW+y/ZVbNcxGOOiU
# 2G6L0TASlqA3yqCJeLuZZqjM6S2VbnvrA8omyg4QnygIHppYjp2CdcCmUpg6wfze
# rkgbVLNasX+le4ss2emuHPh55dLDP20yW83DeGeqSgE//foaJWhtOK/cnvs04zV2
# D6oSAVsOsZ6ozYlQckYnaxIBANDKLRnzCXVZLUCmHxCUhxHuiNJUsHfZYIv/Zxen
# C5ZjD/JPgx3onkoKbNfTRTgwOCdXhVPjWnnp7Su49jymsekqdk1ntln4ixDT3Vol
# ghQPQLjICBc8qXiOJAcFDwqLf/telPlzUUzvlDeC4BYMnpBAP6rQ3JJ8i0vCCiWv
# zKCtmbcDqDRMDpWyJWM3XA/kVKP9i2tNa1R/ej2SleCFLgRapBw=
# =3koe
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 24 Jun 2024 02:21:24 PM PDT
# gpg:                using RSA key A0F66548F04895EBFE6B0B6051A343C7CFFBECA1
# gpg: Good signature from "Cédric Le Goater <clg@kaod.org>" [undefined]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: A0F6 6548 F048 95EB FE6B  0B60 51A3 43C7 CFFB ECA1

* tag 'pull-vfio-20240624' of https://github.com/legoater/qemu: (42 commits)
  vfio/container: Move vfio_container_destroy() to an instance_finalize() handler
  vfio/container: Introduce vfio_iommu_legacy_instance_init()
  vfio/container: Remove vfio_container_init()
  vfio/container: Remove VFIOContainerBase::ops
  vfio/container: Introduce an instance_init() handler
  vfio/container: Switch to QOM
  vfio/container: Change VFIOContainerBase to use QOM
  vfio/container: Discover IOMMU type before creating the container
  vfio/container: Introduce vfio_create_container()
  vfio/container: Introduce vfio_get_iommu_class_name()
  vfio/container: Modify vfio_get_iommu_type() to use a container fd
  vfio/container: Simplify vfio_container_init()
  vfio/container: Introduce vfio_address_space_insert()
  vfio/common: Extract vIOMMU code from vfio_sync_dirty_bitmap()
  vfio/common: Move dirty tracking ranges update to helper
  vfio: Remove unused declarations from vfio-common.h
  vfio: Make vfio_devices_dma_logging_start() return bool
  memory: Remove IOMMU MR iommu_set_iova_range API
  hw/vfio: Remove memory_region_iommu_set_iova_ranges() call
  virtio-iommu: Remove the implementation of iommu_set_iova_range
  ...

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge tag 'sdmmc-20240624' of https://github.com/philmd/qemu into staging

SD/MMC patches queue

One fix and various cleanups for the SD card model.

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEE+qvnXhKRciHc/Wuy4+MsLN6twN4FAmZ5cRUACgkQ4+MsLN6t
# wN59Qw//cUdjD287pB5Ml5aQqr9sOTyVnHUceZtz7AOZ5w8RM2tlPDgOImeLOvU6
# OV7qfWvNaUxtQxhfh5jpe8Pj4eHBtRQzA6a1AWToEvnN4189QWHZpqf5TUa4AlFS
# uAk7k2TkoNv9zbNKca0bP3L1x6sT9l0VPZBLaLbgdXDIX2ycD0r3NVQxXb/bJRgM
# 6pFRcLCF/isKzLQDwqnTa11hB/JDTvOU7xnY0kazGRvyWjbSvE2sOJzLNJXHkW0I
# /FNfRbOKJo2t+47Z5qSXUFFLeIEBTy7VqNBsOQ6sMIgrWzbOSrtBcuxKp0p9NCGH
# fdZHlDVRnNGXewUya4RjbmXiCNuGL4zJ82b2BaQZVd5ZwU2opIr8xO96WCojQ4dZ
# +Dq3uv7su3PUVOh95i38Eo93OG9jXFx642XD4q2uKu5j70IoGXAkIoLUcFkZZdGS
# 9rCsaNUHyHJrN6nXf3Cekvkqxz36p6QXaUF9I1vB0JF6CrexMD35sBUK+RE9k4uW
# LnqL7ZwQDGDGVl3kPS/VCXv1mMim4aRLSEIveq7Ui6dKzaaJMIIodZ8CFMuyTTsD
# cGE+Cd053nf6SzX3+kEZftNdjtJ906O8xIAw+RNdARYx003l4kUxgsPDk7ELyzIP
# Tb+VlZl2P+ROJmeWvRMTW7ZQ49M9IEMrg8zlGF4hLCxB1JndeOA=
# =O5er
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 24 Jun 2024 06:13:57 AM PDT
# gpg:                using RSA key FAABE75E12917221DCFD6BB2E3E32C2CDEADC0DE
# gpg: Good signature from "Philippe Mathieu-Daudé (F4BUG) <f4bug@amsat.org>" [full]

* tag 'sdmmc-20240624' of https://github.com/philmd/qemu:
  hw/sd/sdcard: Add comments around registers and commands
  hw/sd/sdcard: Inline BLK_READ_BLOCK / BLK_WRITE_BLOCK macros
  hw/sd/sdcard: Add sd_invalid_mode_for_cmd to report invalid mode switch
  hw/sd/sdcard: Only call sd_req_get_address() where address is used
  hw/sd/sdcard: Factor sd_req_get_address() method out
  hw/sd/sdcard: Only call sd_req_get_rca() where RCA is used
  hw/sd/sdcard: Factor sd_req_get_rca() method out
  hw/sd/sdcard: Have cmd_valid_while_locked() return a boolean value
  hw/sd/sdcard: Trace update of block count (CMD23)
  hw/sd/sdcard: Remove explicit entries for illegal commands
  hw/sd/sdcard: Remove ACMD6 handler for SPI mode
  hw/sd/sdcard: Use Load/Store API to fill some CID/CSD registers
  hw/sd/sdcard: Use registerfield CSR::CURRENT_STATE definition
  hw/sd/sdcard: Use HWBLOCK_SHIFT definition instead of magic values
  hw/sd/sdcard: Fix typo in SEND_OP_COND command name
  hw/sd/sdcard: Rewrite sd_cmd_ALL_SEND_CID using switch case (CMD2)
  hw/sd/sdcard: Correct code indentation
  hw/sd/sdcard: Avoid OOB in sd_read_byte() during unexpected CMD switch
  bswap: Add st24_be_p() to store 24 bits in big-endian order

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

vfio/container: Move vfio_container_destroy() to an instance_finalize() handler

vfio_container_destroy() clears the resources allocated
VFIOContainerBase object. Now that VFIOContainerBase is a QOM object,
add an instance_finalize() handler to do the cleanup. It will be
called through object_unref().

Suggested-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Introduce vfio_iommu_legacy_instance_init()

Just as we did for the VFIOContainerBase object, introduce an
instance_init() handler for the legacy VFIOContainer object and do the
specific initialization there.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Remove vfio_container_init()

It's now empty.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Remove VFIOContainerBase::ops

Instead, use VFIO_IOMMU_GET_CLASS() to get the class pointer.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Introduce an instance_init() handler

This allows us to move the initialization code from vfio_container_init(),
which we will soon remove.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Switch to QOM

Instead of allocating the container struct, create a QOM object of the
appropriate type.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Change VFIOContainerBase to use QOM

VFIOContainerBase was made a QOM interface because we believed that a
QOM object would expose all the IOMMU backends to the QEMU machine and
human interface. This only applies to user creatable devices or objects.

Change the VFIOContainerBase nature from interface to object and make
the necessary adjustments in the VFIO_IOMMU hierarchy.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Discover IOMMU type before creating the container

Since the QEMU struct type representing the VFIO container is deduced
from the IOMMU type exposed by the host, this type should be well
defined *before* creating the container struct. This will be necessary
to instantiate a QOM object of the correct type in future changes.

Rework vfio_set_iommu() to extract the part doing the container
initialization and move it under vfio_create_container().

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Introduce vfio_create_container()

This routine allocates the QEMU struct type representing the VFIO
container. It is minimal currently and future changes will do more
initialization.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Introduce vfio_get_iommu_class_name()

Rework vfio_get_iommu_class() to return a literal class name instead
of a class object. We will need this name to instantiate the object
later on. Since the default case asserts, remove the error report as
QEMU will simply abort before.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Modify vfio_get_iommu_type() to use a container fd

The 'container' pointer has no other use than its 'fd' attribute.
Simplify the prototype to ease future changes.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Simplify vfio_container_init()

Assign the base container VFIOAddressSpace 'space' pointer in
vfio_address_space_insert(). The ultimate goal is to remove
vfio_container_init() and instead rely on an .instance_init() handler
to perfom the initialization of VFIOContainerBase.

To be noted that vfio_connect_container() will assign the 'space'
pointer later in the execution flow. This should not have any
consequence.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/container: Introduce vfio_address_space_insert()

It prepares ground for a future change initializing the 'space' pointer
of VFIOContainerBase. The goal is to replace vfio_container_init() by
an .instance_init() handler when VFIOContainerBase is QOMified.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/common: Extract vIOMMU code from vfio_sync_dirty_bitmap()

Extract vIOMMU code from vfio_sync_dirty_bitmap() to a new function and
restructure the code.

This is done in preparation for optimizing vIOMMU device dirty page
tracking. No functional changes intended.

Signed-off-by: Avihai Horon <avihaih@nvidia.com>
Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
[ clg: - Rebased on upstream
       - Fixed typo in commit log ]
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio/common: Move dirty tracking ranges update to helper

Separate the changes that update the ranges from the listener, to
make it reusable in preparation to expand its use to vIOMMU support.

Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
[ clg: - Rebased on upstream
       - Introduced vfio_dirty_tracking_update_range()
       - Fixed typ in commit log ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio: Remove unused declarations from vfio-common.h

These were forgotten in the recent cleanups.

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

vfio: Make vfio_devices_dma_logging_start() return bool

Since vfio_devices_dma_logging_start() takes an 'Error **' argument,
best practices suggest to return a bool. See the api/error.h Rules
section. It will simplify potential changes coming after.

vfio_container_set_dirty_page_tracking() could be modified in the same
way but the errno value can be saved in the migration stream when
called from vfio_listener_log_global_stop().

Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>

memory: Remove IOMMU MR iommu_set_iova_range API

Since the host IOVA ranges are now passed through the
PCIIOMMUOps set_host_resv_regions and we have removed
the only implementation of iommu_set_iova_range() in
the virtio-iommu and the only call site in vfio/common,
let's retire the IOMMU MR API and its memory wrapper.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

hw/vfio: Remove memory_region_iommu_set_iova_ranges() call

As we have just removed the only implementation of
iommu_set_iova_ranges IOMMU MR callback in the virtio-iommu,
let's remove the call to the memory wrapper. Usable IOVA ranges
are now conveyed through the PCIIOMMUOps in VFIO-PCI.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

virtio-iommu: Remove the implementation of iommu_set_iova_range

Now that we use PCIIOMMUOps to convey information about usable IOVA
ranges we do not to implement the iommu_set_iova_ranges IOMMU MR
callback.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

virtio-iommu: Compute host reserved regions

Compute the host reserved regions in virtio_iommu_set_iommu_device().
The usable IOVA regions are retrieved from the HostIOMMUDevice.
The virtio_iommu_set_host_iova_ranges() helper turns usable regions
into complementary reserved regions while testing the inclusion
into existing ones. virtio_iommu_set_host_iova_ranges() reuse the
implementation of virtio_iommu_set_iova_ranges() which will be
removed in subsequent patches. rebuild_resv_regions() is just moved.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

HostIOMMUDevice: Store the aliased bus and devfn

Store the aliased bus and devfn in the HostIOMMUDevice.
This will be useful to handle info that are iommu group
specific and not device specific (such as reserved
iova ranges).

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

HostIOMMUDevice: Introduce get_iova_ranges callback

Introduce a new HostIOMMUDevice callback that allows to
retrieve the usable IOVA ranges.

Implement this callback in the legacy VFIO and IOMMUFD VFIO
host iommu devices. This relies on the VFIODevice agent's
base container iova_ranges resource.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

virtio-iommu: Implement set|unset]_iommu_device() callbacks

Implement PCIIOMMUOPs [set|unset]_iommu_device() callbacks.
In set(), the HostIOMMUDevice handle is stored in a hash
table indexed by PCI BDF. The object will allow to retrieve
information related to the physical IOMMU.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

HostIOMMUDevice: Store the VFIO/VDPA agent

Store the agent device (VFIO or VDPA) in the host IOMMU device.
This will allow easy access to some of its resources.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

intel_iommu: Check compatibility with host IOMMU capabilities

If check fails, host device (either VFIO or VDPA device) is not
compatible with current vIOMMU config and should not be passed to
guest.

Only aw_bits is checked for now, we don't care about other caps
before scalable modern mode is introduced.

Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

intel_iommu: Implement [set|unset]_iommu_device() callbacks

Implement [set|unset]_iommu_device() callbacks in Intel vIOMMU.
In set call, we take a reference of HostIOMMUDevice and store it
in hash table indexed by PCI BDF.

Note this BDF index is device's real BDF not the aliased one which
is different from the index of VTDAddressSpace. There can be multiple
assigned devices under same virtual iommu group and share same
VTDAddressSpace, but each has its own HostIOMMUDevice.

Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Yi Sun <yi.y.sun@linux.intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

intel_iommu: Extract out vtd_cap_init() to initialize cap/ecap

Extract cap/ecap initialization in vtd_cap_init() to make code
cleaner.

No functional change intended.

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

vfio/pci: Pass HostIOMMUDevice to vIOMMU

With HostIOMMUDevice passed, vIOMMU can check compatibility with host
IOMMU, call into IOMMUFD specific methods, etc.

Originally-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
Signed-off-by: Yi Sun <yi.y.sun@linux.intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

hw/pci: Introduce pci_device_[set|unset]_iommu_device()

pci_device_[set|unset]_iommu_device() call pci_device_get_iommu_bus_devfn()
to get iommu_bus->iommu_ops and call [set|unset]_iommu_device callback to
set/unset HostIOMMUDevice for a given PCI device.

Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Yi Sun <yi.y.sun@linux.intel.com>
Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

hw/pci: Introduce helper function pci_device_get_iommu_bus_devfn()

Extract out pci_device_get_iommu_bus_devfn() from
pci_device_iommu_address_space() to facilitate
implementation of pci_device_[set|unset]_iommu_device()
in following patch.

No functional change intended.

Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Yi Sun <yi.y.sun@linux.intel.com>
Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

vfio: Create host IOMMU device instance

Create host IOMMU device instance in vfio_attach_device() and call
.realize() to initialize it further.

Introuduce attribute VFIOIOMMUClass::hiod_typename and initialize
it based on VFIO backend type. It will facilitate HostIOMMUDevice
creation in vfio_attach_device().

Suggested-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

backends/iommufd: Implement HostIOMMUDeviceClass::get_cap() handler

Suggested-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

vfio/container: Implement HostIOMMUDeviceClass::get_cap() handler

Suggested-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

vfio/iommufd: Implement HostIOMMUDeviceClass::realize() handler

It calls iommufd_backend_get_device_info() to get host IOMMU
related information and translate it into HostIOMMUDeviceCaps
for query with .get_cap().

For aw_bits, use the same way as legacy backend by calling
vfio_device_get_aw_bits() which is common for different vendor
IOMMU.

Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

backends/iommufd: Introduce helper function iommufd_backend_get_device_info()

Introduce a helper function iommufd_backend_get_device_info() to get
host IOMMU related information through iommufd uAPI.

Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Yi Sun <yi.y.sun@linux.intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

vfio/container: Implement HostIOMMUDeviceClass::realize() handler

The realize function populates the capabilities. For now only the
aw_bits caps is computed for legacy backend.

Introduce a helper function vfio_device_get_aw_bits() which calls
range_get_last_bit() to get host aw_bits and package it in
HostIOMMUDeviceCaps for query with .get_cap(). This helper will
also be used by iommufd backend.

Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

range: Introduce range_get_last_bit()

This helper get the highest 1 bit position of the upper bound.

If the range is empty or upper bound is zero, -1 is returned.

Suggested-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

backends/iommufd: Introduce TYPE_HOST_IOMMU_DEVICE_IOMMUFD[_VFIO] devices

TYPE_HOST_IOMMU_DEVICE_IOMMUFD represents a host IOMMU device under
iommufd backend. It is abstract, because it is going to be derived
into VFIO or VDPA type'd device.

It will have its own .get_cap() implementation.

TYPE_HOST_IOMMU_DEVICE_IOMMUFD_VFIO is a sub-class of
TYPE_HOST_IOMMU_DEVICE_IOMMUFD, represents a VFIO type'd host IOMMU
device under iommufd backend. It will be created during VFIO device
attaching and passed to vIOMMU.

It will have its own .realize() implementation.

Opportunistically, add missed header to include/sysemu/iommufd.h.

Suggested-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

vfio/container: Introduce TYPE_HOST_IOMMU_DEVICE_LEGACY_VFIO device

TYPE_HOST_IOMMU_DEVICE_LEGACY_VFIO represents a host IOMMU device under
VFIO legacy container backend.

It will have its own realize implementation.

Suggested-by: Eric Auger <eric.auger@redhat.com>
Suggested-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

backends/host_iommu_device: Introduce HostIOMMUDeviceCaps

HostIOMMUDeviceCaps's elements map to the host IOMMU's capabilities.
Different platform IOMMU can support different elements.

Currently only two elements, type and aw_bits, type hints the host
platform IOMMU type, i.e., INTEL vtd, ARM smmu, etc; aw_bits hints
host IOMMU address width.

Introduce .get_cap() handler to check if HOST_IOMMU_DEVICE_CAP_XXX
is supported.

Suggested-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>

backends: Introduce HostIOMMUDevice abstract

A HostIOMMUDevice is an abstraction for an assigned device that is protected
by a physical IOMMU (aka host IOMMU). The userspace interaction with this
physical IOMMU can be done either through the VFIO IOMMU type 1 legacy
backend or the new iommufd backend. The assigned device can be a VFIO device
or a VDPA device. The HostIOMMUDevice is needed to interact with the host
IOMMU that protects the assigned device. It is especially useful when the
device is also protected by a virtual IOMMU as this latter use the translation
services of the physical IOMMU and is constrained by it. In that context the
HostIOMMUDevice can be passed to the virtual IOMMU to collect physical IOMMU
capabilities such as the supported address width. In the future, the virtual
IOMMU will use the HostIOMMUDevice to program the guest page tables in the
first translation stage of the physical IOMMU.

Introduce .realize() to initialize HostIOMMUDevice further after instance init.

Suggested-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>