While discussing a problem with how we're (not) setting vstart_eq_zero
Richard had the following to say w.r.t the conditional mark_vs_dirty()
calls on load/store functions [1]:
"I think it's required to have stores set dirty unconditionally, before
the operation.
Consider a store that traps on the 2nd element, leaving vstart = 2, and
exiting to the main loop via exception. The exception enters the kernel
page fault handler. The kernel may need to fault in the page for the
process, and in the meantime task switch.
If vs dirty is not already set, the kernel won't know to save vector
state on task switch."
Do a mark_vs_dirty() before both loads and stores.
[1] https://lore.kernel.org/qemu-riscv/72c7503b-0f43-44b8-aa82-fbafed2aac0c@linaro.org/
Suggested-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240306171932.549549-2-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
mcountinhibit, mcounteren, scounteren and hcounteren must always be 32-bit
by privileged spec
Signed-off-by: Vadim Shakirov <vadim.shakirov@syntacore.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Andrew Jones <ajones@ventanamicro.com>
Message-ID: <20240202113919.18236-1-vadim.shakirov@syntacore.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
smaia and ssaia were ratified in August 25th 2023 [1].
zvfh and zvfhmin were ratified in August 2nd 2023 [2].
zfbfmin and zvfbf(min|wma) are frozen and moved to public review since
Dec 16th 2023 [3].
zaamo and zalrsc are both marked as "Frozen" since January 24th 2024
[4].
[1] https://jira.riscv.org/browse/RVS-438
[2] https://jira.riscv.org/browse/RVS-871
[3] https://jira.riscv.org/browse/RVS-704
[4] https://jira.riscv.org/browse/RVS-1995
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240301144053.265964-1-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
The last KVM extensions added were back in 6.6. Sync them to Linux 6.8.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240304134732.386590-3-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
The Ztso extension is already ratified, this adds it as a CPU property
and adds various fences throughout the port in order to allow TSO
targets to function on weaker hosts. We need no fences for AMOs as
they're already SC, the places we need barriers are described.
These fences are placed in the RISC-V backend rather than TCG as is
planned for x86-on-arm64 because RISC-V allows heterogeneous (and
likely soon dynamic) hart memory models.
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Christoph Müllner <christoph.muellner@vrull.eu>
Message-ID: <20240207122256.902627-2-christoph.muellner@vrull.eu>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Add missing include guard in pmu.h to avoid the problem of double
inclusion.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Atish Patra <atishp@rivosinc.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240220110907.10479-1-frank.chang@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Added xATP_MODE validation for vsatp/hgatp CSRs.
The xATP register is an SXLEN-bit read/write WARL register, so
the legal value must be returned (See riscv-privileged-20211203, SATP/VSATP/HGATP CSRs).
Signed-off-by: Irina Ryapolova <irina.ryapolova@syntacore.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240109145923.37893-2-irina.ryapolova@syntacore.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
The SATP register is an SXLEN-bit read/write WARL register. It means that CSR fields are only defined
for a subset of bit encodings, but allow any value to be written while guaranteeing to return a legal
value whenever read (See riscv-privileged-20211203, SATP CSR).
For example on rv64 we are trying to write to SATP CSR val = 0x1000000000000000 (SATP_MODE = 1 - Reserved for standard use)
and after that we are trying to read SATP_CSR. We read from the SATP CSR value = 0x1000000000000000, which is not a correct
operation (return illegal value).
Signed-off-by: Irina Ryapolova <irina.ryapolova@syntacore.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240109145923.37893-1-irina.ryapolova@syntacore.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Named features are extensions which don't make sense for users to
control and are therefore not exposed on the command line. However,
svade is an extension which makes sense for users to control, so treat
it like a "normal" extension. The default is false, even for the max
cpu type, since QEMU has always implemented hardware A/D PTE bit
updating, so users must opt into svade (or get it from a CPU type
which enables it by default).
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240215223955.969568-7-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Gate hardware A/D PTE bit updating on {m,h}envcfg.ADUE and only
enable menvcfg.ADUE on reset if svade has not been selected. Now
that we also consider svade, we have four possible configurations:
1) !svade && !svadu
use hardware updating and there's no way to disable it
(the default, which maintains past behavior. Maintaining
the default, even with !svadu is a change that fixes [1])
2) !svade && svadu
use hardware updating, but also provide {m,h}envcfg.ADUE,
allowing software to switch to exception mode
(being able to switch is a change which fixes [1])
3) svade && !svadu
use exception mode and there's no way to switch to hardware
updating
(this behavior change fixes [2])
4) svade && svadu
use exception mode, but also provide {m,h}envcfg.ADUE,
allowing software to switch to hardware updating
(this behavior change fixes [2])
Fixes: 0af3f115e6 ("target/riscv: Add *envcfg.HADE related check in address translation") [1]
Fixes: 48531f5adb ("target/riscv: implement svade") [2]
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240215223955.969568-6-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
The hypervisor should decide what it wants to enable. Zero all
configuration enable bits on reset.
Also, commit ed67d63798 ("target/riscv: Update CSR bits name for
svadu extension") missed one reference to 'hade'. Change it now.
Fixes: 0af3f115e6 ("target/riscv: Add *envcfg.HADE related check in address translation")
Fixes: ed67d63798 ("target/riscv: Update CSR bits name for svadu extension")
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Signed-off-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240215223955.969568-5-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
The RVA22U64 and RVA22S64 profiles mandates certain extensions that,
until now, we were implying that they were available.
We can't do this anymore since named features also has a riscv,isa
entry. Let's add them to riscv_cpu_named_features[].
Instead of adding one bool for each named feature that we'll always
implement, i.e. can't be turned off, add a 'ext_always_enabled' bool in
cpu->cfg. This bool will be set to 'true' in TCG accel init, and all
named features will point to it. This also means that KVM won't see
these features as always enable, which is our intention.
If any accelerator adds support to disable one of these features, we'll
have to promote them to regular extensions and allow users to disable it
via command line.
After this patch, here's the riscv,isa from a buildroot using the
'rva22s64' CPU:
# cat /proc/device-tree/cpus/cpu@0/riscv,isa
rv64imafdc_zic64b_zicbom_zicbop_zicboz_ziccamoa_ziccif_zicclsm_ziccrse_
zicntr_zicsr_zifencei_zihintpause_zihpm_za64rs_zfhmin_zca_zcd_zba_zbb_
zbs_zkt_ssccptr_sscounterenw_sstvala_sstvecd_svade_svinval_svpbmt#
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: <20240215223955.969568-4-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Further discussions after the introduction of rva22 support in QEMU
revealed that what we've been calling 'named features' are actually
regular extensions, with their respective riscv,isa DTs. This is
clarified in [1]. [2] is a bug tracker asking for the profile spec to be
less cryptic about it.
As far as QEMU goes we understand extensions as something that the user
can enable/disable in the command line. This isn't the case for named
features, so we'll have to reach a middle ground.
We'll keep our existing nomenclature 'named features' to refer to any
extension that the user can't control in the command line. We'll also do
the following:
- 'svade' and 'zic64b' flags are renamed to 'ext_svade' and
'ext_zic64b'. 'ext_svade' and 'ext_zic64b' now have riscv,isa strings and
priv_spec versions;
- skip name feature check in cpu_bump_multi_ext_priv_ver(). Now that
named features have a riscv,isa and an entry in isa_edata_arr[] we
don't need to gate the call to cpu_cfg_ext_get_min_version() anymore.
[1] https://github.com/riscv/riscv-profiles/issues/121
[2] https://github.com/riscv/riscv-profiles/issues/142
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: <20240215223955.969568-3-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Recent changes in options handling removed the 'mmu' default the bare
CPUs had, meaning that we must enable 'mmu' by hand when using the
rva22s64 profile CPU.
Given that this profile is setting a satp mode, it already implies that
we need a 'mmu'. Enable the 'mmu' in this case.
Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-ID: <20240215223955.969568-2-dbarboza@ventanamicro.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
The original implementation sets $pc to the address read from the jump
vector table first and links $ra with the address of the next instruction
after the updated $pc. After jumping to the updated $pc and executing the
next ret instruction, the program jumps to $ra, which is in the same
function currently executing, which results in an infinite loop.
This commit stores the jump address in a temporary, updates $ra with the
current $pc, and copies the temporary to $pc.
Signed-off-by: Jason Chien <jason.chien@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20240207081820.28559-1-jason.chien@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
This makes the output suitable when used for plugins.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20240305121005.3528075-29-alex.bennee@linaro.org>
If translation is enabled, and the PTE memory type is Device,
enable checking alignment via TLB_CHECK_ALIGNMENT. While the
check is done later than it should be per the ARM, it's better
than not performing the check at all.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20240301204110.656742-7-richard.henderson@linaro.org
[PMM: tweaks to comment text]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
If translation is disabled, the default memory type is Device, which
requires alignment checking. This is more optimally done early via
the MemOp given to the TCG memory operation.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reported-by: Idan Horowitz <idan.horowitz@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20240301204110.656742-6-richard.henderson@linaro.org
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1204
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Allow the target to set tlb flags to apply to all of the
comparators. Remove MemTxAttrs.byte_swap, as the bit is
not relevant to memory transactions, only the page mapping.
Adjust target/sparc to set TLB_BSWAP directly.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20240301204110.656742-4-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Now that we have removed TARGET_PAGE_BITS_MIN-6 from
TLB_FLAGS_MASK, we can test for 32-byte alignment.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20240301204110.656742-2-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
When calculating the IOR for the exception handlers, the current
unwind_breg value is needed on 64-bit hppa machines.
Restore that value by calling cpu_restore_state() earlier, which in turn
calls hppa_restore_state_to_opc() which restores the unwind_breg for the
current instruction.
Signed-off-by: Helge Deller <deller@gmx.de>
Fixes: 3824e0d643 ("target/hppa: Export function hppa_set_ior_and_isr()")
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Unaligned 64-bit accesses were found in Linux to clobber carry bits,
resulting in bad results if an arithmetic operation involving a
carry bit was executed after an unaligned 64-bit operation.
hppa 2.0 defines additional carry bits in PSW register bits 32..39.
When restoring PSW after executing an unaligned instruction trap, those
bits were not cleared and ended up to be active all the time. Since there
are no bits other than the upper carry bits needed in the upper 32 bit of
env->psw and since those are stored in env->psw_cb, just clear the entire
upper 32 bit when storing psw to solve the problem unconditionally.
Fixes: 931adff314 ("target/hppa: Update cpu_hppa_get/put_psw for hppa64")
Cc: Richard Henderson <richard.henderson@linaro.org>
Cc: Charlie Jenkins <charlie@rivosinc.com>
Cc: Helge Deller <deller@gmx.de>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Helge Deller <deller@gmx.de>
Since alpha binaries are generally built for multiple
page sizes, it is trivial to allow the page size to vary.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com>
Acked-by: Helge Deller <deller@gmx.de>
Message-Id: <20240102015808.132373-34-richard.henderson@linaro.org>
Since ppc binaries are generally built for multiple
page sizes, it is trivial to allow the page size to vary.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Ilya Leoshkevich <iii@linux.ibm.com>
Acked-by: Helge Deller <deller@gmx.de>
Message-Id: <20240102015808.132373-33-richard.henderson@linaro.org>
Since aarch64 binaries are generally built for multiple
page sizes, it is trivial to allow the page size to vary.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Acked-by: Helge Deller <deller@gmx.de>
Message-Id: <20240102015808.132373-31-richard.henderson@linaro.org>
These members will be used to help plugins to identify registers.
The added members in instances of GDBFeature dynamically generated by
CPUs will be filled in later changes.
Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-Id: <20231213-gdb-v17-10-777047380591@daynix.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20240227144335.1196131-15-alex.bennee@linaro.org>
This function is no longer used.
Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20231213-gdb-v17-9-777047380591@daynix.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20240227144335.1196131-14-alex.bennee@linaro.org>
GDBFeature has the num_regs member so use it where applicable to
remove magic numbers.
Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-Id: <20231213-gdb-v17-8-777047380591@daynix.com>
[AJB: remove core reg check from microblaze read reg]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20240227144335.1196131-13-alex.bennee@linaro.org>
Align the parameters of gdb_get_reg_cb and gdb_set_reg_cb with the
gdb_read_register and gdb_write_register members of CPUClass to allow
to unify the logic to access registers of the core and coprocessors
in the future.
Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20231213-gdb-v17-6-777047380591@daynix.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20240227144335.1196131-11-alex.bennee@linaro.org>
This is a tree-wide change to introduce GDBFeature parameter to
gdb_register_coprocessor(). The new parameter just replaces num_regs
and xml parameters for now. GDBFeature will be utilized to simplify XML
lookup in a following change.
Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Acked-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20231213-gdb-v17-4-777047380591@daynix.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20240227144335.1196131-9-alex.bennee@linaro.org>
In preparation for a change to use GDBFeature as a parameter of
gdb_register_coprocessor(), convert the internal representation of
dynamic feature from plain XML to GDBFeature.
Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-Id: <20231213-gdb-v17-3-777047380591@daynix.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20240227144335.1196131-8-alex.bennee@linaro.org>
In preparation for a change to use GDBFeature as a parameter of
gdb_register_coprocessor(), convert the internal representation of
dynamic feature from plain XML to GDBFeature.
Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20231213-gdb-v17-2-777047380591@daynix.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20240227144335.1196131-7-alex.bennee@linaro.org>
In preparation for a change to use GDBFeature as a parameter of
gdb_register_coprocessor(), convert the internal representation of
dynamic feature from plain XML to GDBFeature.
Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Acked-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20231213-gdb-v17-1-777047380591@daynix.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20240227144335.1196131-6-alex.bennee@linaro.org>
The A20 mask is only applied to the final memory access. Nested
page tables are always walked with the raw guest-physical address.
Unlike the previous patch, in this one the masking must be kept, but
it was done too early.
Cc: qemu-stable@nongnu.org
Fixes: 4a1e9d4d11 ("target/i386: Use atomic operations for pte updates", 2022-10-18)
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
If ptw_translate() does a MMU_PHYS_IDX access, the A20 mask is already
applied in get_physical_address(), which is called via probe_access_full()
and x86_cpu_tlb_fill().
If ptw_translate() on the other hand does a MMU_NESTED_IDX access,
the A20 mask must not be applied to the address that is looked up in
the nested page tables; it must be applied only to the addresses that
hold the NPT entries (which is achieved via MMU_PHYS_IDX, per the
previous paragraph).
Therefore, we can remove A20 masking from the computation of the page
table entry's address, and let get_physical_address() or mmu_translate()
apply it when they know they are returning a host-physical address.
Cc: qemu-stable@nongnu.org
Fixes: 4a1e9d4d11 ("target/i386: Use atomic operations for pte updates", 2022-10-18)
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The address translation logic in get_physical_address() will currently
truncate physical addresses to 32 bits unless long mode is enabled.
This is incorrect when using physical address extensions (PAE) outside
of long mode, with the result that a 32-bit operating system using PAE
to access memory above 4G will experience undefined behaviour.
The truncation code was originally introduced in commit 33dfdb5 ("x86:
only allow real mode to access 32bit without LMA"), where it applied
only to translations performed while paging is disabled (and so cannot
affect guests using PAE).
Commit 9828198 ("target/i386: Add MMU_PHYS_IDX and MMU_NESTED_IDX")
rearranged the code such that the truncation also applied to the use
of MMU_PHYS_IDX and MMU_NESTED_IDX. Commit 4a1e9d4 ("target/i386: Use
atomic operations for pte updates") brought this truncation into scope
for page table entry accesses, and is the first commit for which a
Windows 10 32-bit guest will reliably fail to boot if memory above 4G
is present.
The truncation code however is not completely redundant. Even though the
maximum address size for any executed instruction is 32 bits, helpers for
operations such as BOUND, FSAVE or XSAVE may ask get_physical_address()
to translate an address outside of the 32-bit range, if invoked with an
argument that is close to the 4G boundary. Likewise for processor
accesses, for example TSS or IDT accesses, when EFER.LMA==0.
So, move the address truncation in get_physical_address() so that it
applies to 32-bit MMU indexes, but not to MMU_PHYS_IDX and MMU_NESTED_IDX.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2040
Fixes: 4a1e9d4d11 ("target/i386: Use atomic operations for pte updates", 2022-10-18)
Cc: qemu-stable@nongnu.org
Co-developed-by: Michael Brown <mcb30@ipxe.org>
Signed-off-by: Michael Brown <mcb30@ipxe.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Accesses from a 32-bit environment (32-bit code segment for instruction
accesses, EFER.LMA==0 for processor accesses) have to mask away the
upper 32 bits of the address. While a bit wasteful, the easiest way
to do so is to use separate MMU indexes. These days, QEMU anyway is
compiled with a fixed value for NB_MMU_MODES. Split MMU_USER_IDX,
MMU_KSMAP_IDX and MMU_KNOSMAP_IDX in two.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Remove knowledge of specific MMU indexes (other than MMU_NESTED_IDX and
MMU_PHYS_IDX) from mmu_translate(). This will make it possible to split
32-bit and 64-bit MMU indexes.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
MSR_VM_HSAVE_PA bits 0-11 are reserved, as are the bits above the
maximum physical address width of the processor. Setting them to
1 causes a #GP (see "15.30.4 VM_HSAVE_PA MSR" in the AMD manual).
The same is true of VMCB addresses passed to VMRUN/VMLOAD/VMSAVE,
even though the manual is not clear on that.
Cc: qemu-stable@nongnu.org
Fixes: 4a1e9d4d11 ("target/i386: Use atomic operations for pte updates", 2022-10-18)
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
CR3 bits 63:32 are ignored in 32-bit mode (either legacy 2-level
paging or PAE paging). Do this in mmu_translate() to remove
the last where get_physical_address() meaningfully drops the high
bits of the address.
Cc: qemu-stable@nongnu.org
Suggested-by: Richard Henderson <richard.henderson@linaro.org>
Fixes: 4a1e9d4d11 ("target/i386: Use atomic operations for pte updates", 2022-10-18)
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
I'm far from confident this handling here is correct. Hence
RFC. In particular not sure on what locks I should hold for this
to be even moderately safe.
The function already appears to be inconsistent in what it returns
as the CONFIG_ATOMIC64 block returns the endian converted 'eventual'
value of the cmpxchg whereas the TCG_OVERSIZED_GUEST case returns
the previous value.
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-id: 20240219161229.11776-1-Jonathan.Cameron@huawei.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The Cortex-A53 r0p4 revision that QEMU emulates is affected by a CatA
erratum #843419 (i.e., the most severe), which requires workarounds in
the toolchain as well as the OS.
Since the emulation is obviously not affected in the same way, we can
indicate this via REVIDR bit #8, which on r0p4 has the meaning that no
workarounds for erratum #843419 are needed.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20240215160202.2803452-1-ardb+git@google.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Filter TLB flushing by PID and mmuidx.
Zoltan reports that, together with the previous TLB flush changes,
performance of a sam460ex machine running 'lame' to convert a wav to
mp3 is improved nearly 10%:
CPU time TLB partial flushes TLB elided flushes
Before 37s 508238 7680722
After 34s 73 1143
Tested-by: BALATON Zoltan <balaton@eik.bme.hu>
Acked-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Have 440 tlbwe flush only the range corresponding to the addresses
covered by the software TLB entry being modified rather than the
entire TLB. This matches what 4xx does.
Tested-by: BALATON Zoltan <balaton@eik.bme.hu>
Acked-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Rather than tlbwe_lo always flushing all TCG TLBs, have it flush just
those corresponding to the old software TLB, and only if it was valid.
Tested-by: BALATON Zoltan <balaton@eik.bme.hu>
Acked-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
BookE software TLB is implemented by flushing old translations from the
relevant TCG TLB whenever software TLB entries change. This means a new
software TLB entry should not have any corresponding cached TCG TLB
translations, so there is nothing to flush. The exception is multiple
software TLBs that cover the same address and address space, but that is
a programming error and results in undefined behaviour, and flushing
does not give an obviously better outcome in that case either.
Remove the unnecessary flush of a newly written software TLB entry.
Tested-by: BALATON Zoltan <balaton@eik.bme.hu>
Acked-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Flushing the TCG TLB pages that cache a software TLB is a common
operation, factor it into its own function.
Tested-by: BALATON Zoltan <balaton@eik.bme.hu>
Acked-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
