Now we no longer have dynamic state affecting things we can remove the
additional fields in cpu.h and simplify the TB hash calculation.
For the benchmark:
hyperfine -w 2 -m 20 \
"./arm-softmmu/qemu-system-arm -cpu cortex-a15 \
-machine type=virt,highmem=off \
-display none -m 2048 \
-serial mon:stdio \
-netdev user,id=unet,hostfwd=tcp::2222-:22 \
-device virtio-net-pci,netdev=unet \
-device virtio-scsi-pci \
-blockdev driver=raw,node-name=hd,discard=unmap,file.driver=host_device,file.filename=/dev/zen-disk/debian-bullseye-armhf \
-device scsi-hd,drive=hd -smp 4 \
-kernel /home/alex/lsrc/linux.git/builds/arm/arch/arm/boot/zImage \
-append 'console=ttyAMA0 root=/dev/sda2 systemd.unit=benchmark.service' \
-snapshot"
It has a marginal effect on runtime, before:
Time (mean ± σ): 26.279 s ± 2.438 s [User: 41.113 s, System: 1.843 s]
Range (min … max): 24.420 s … 32.565 s 20 runs
after:
Time (mean ± σ): 24.440 s ± 2.885 s [User: 34.474 s, System: 2.028 s]
Range (min … max): 21.663 s … 29.937 s 20 runs
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1358
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 20230526165401.574474-10-alex.bennee@linaro.org
Message-Id: <20230524133952.3971948-9-alex.bennee@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
PAGE_WRITE is current writability, as modified by TB protection;
PAGE_WRITE_ORG is the original page writability.
Fixes: cdfac37be0 ("accel/tcg: Honor atomicity of loads")
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This had been set since the beginning, is never undefined,
and it would seem to be harmful to debugging to do so.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Remove the locally defined load_atomic16 and store_atomic16,
along with HAVE_al16 and HAVE_al16_fast in favor of the
routines defined in atomic128.h.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
These symbols will shortly become dynamic runtime tests and
therefore not appropriate for the preprocessor. Use the
matching CONFIG_* symbols for that purpose.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Now that load/store are gone, we're always passing
PAGE_READ | PAGE_WRITE for RMW atomic operations.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Atomic load/store of 128-byte quantities is now handled
by cpu_{ld,st}16_mmu.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
With the current structure of cputlb.c, there is no difference
between the little-endian and big-endian entry points, aside
from the assert. Unify the pairs of functions.
The only use of the functions with explicit endianness was in
target/sparc64, and that was only to satisfy the assert: the
correct endianness is already built into memop.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Not only the routines in ldst_atomicity.c.inc need markup,
but also the ones in the headers.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
In fcdab382c8 we removed a tcg_gen_extu_tl_i64 from gen_empty_mem_cb,
and failed to adjust the associated copy, leading to a failed assert.
Fixes: fcdab382c8 ("accel/tcg: Widen plugin_gen_empty_mem_callback to i64")
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230518145813.2940745-1-richard.henderson@linaro.org>
arm64 has different capability from x86 to enable the dirty ring, which
is KVM_CAP_DIRTY_LOG_RING_ACQ_REL. Besides, arm64 also needs the backup
bitmap extension (KVM_CAP_DIRTY_LOG_RING_WITH_BITMAP) when 'kvm-arm-gicv3'
or 'arm-its-kvm' device is enabled. Here the extension is always enabled
and the unnecessary overhead to do the last stage of dirty log synchronization
when those two devices aren't used is introduced, but the overhead should
be very small and acceptable. The benefit is cover future cases where those
two devices are used without modifying the code.
Signed-off-by: Gavin Shan <gshan@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Tested-by: Zhenyu Zhang <zhenyzha@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Message-Id: <20230509022122.20888-5-gshan@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Due to multiple capabilities associated with the dirty ring for different
architectures: KVM_CAP_DIRTY_{LOG_RING, LOG_RING_ACQ_REL} for x86 and
arm64 separately. There will be more to be done in order to support the
dirty ring for arm64.
Lets add helper kvm_dirty_ring_init() to enable the dirty ring. With this,
the code looks a bit clean.
No functional change intended.
Signed-off-by: Gavin Shan <gshan@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Tested-by: Zhenyu Zhang <zhenyzha@redhat.com>
Message-Id: <20230509022122.20888-4-gshan@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
In the last stage of live migration or memory slot removal, the
backup bitmap needs to be synchronized when it has been enabled.
Signed-off-by: Gavin Shan <gshan@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Tested-by: Zhenyu Zhang <zhenyzha@redhat.com>
Message-Id: <20230509022122.20888-3-gshan@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The global dirty log synchronization is used when KVM and dirty ring
are enabled. There is a particularity for ARM64 where the backup
bitmap is used to track dirty pages in non-running-vcpu situations.
It means the dirty ring works with the combination of ring buffer
and backup bitmap. The dirty bits in the backup bitmap needs to
collected in the last stage of live migration.
In order to identify the last stage of live migration and pass it
down, an extra parameter is added to the relevant functions and
callbacks. This last stage indicator isn't used until the dirty
ring is enabled in the subsequent patches.
No functional change intended.
Signed-off-by: Gavin Shan <gshan@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Tested-by: Zhenyu Zhang <zhenyzha@redhat.com>
Message-Id: <20230509022122.20888-2-gshan@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Disconnect guest tlb parameters from TCG compilation.
Reviewed-by: Anton Johansson <anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Disconnect guest page size from TCG compilation.
While this could be done via exec/target_page.h, we want to cache
the value across multiple memory access operations, so we might
as well initialize this early.
The changes within tcg/ are entirely mechanical:
sed -i s/TARGET_PAGE_BITS/s->page_bits/g
sed -i s/TARGET_PAGE_MASK/s->page_mask/g
Reviewed-by: Anton Johansson <anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This will enable replacement of TARGET_LONG_BITS within tcg/.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Since we do this inside gen_empty_mem_cb anyway, let's
do this earlier inside tcg expansion.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
As do_gen_mem_cb is called once, merge it into gen_empty_mem_cb.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
As gen_mem_wrapped is only used in plugin_gen_empty_mem_callback,
we can avoid the curiosity of union mem_gen_fn by inlining it.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Always pass the target address as uint64_t.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Always pass the target address as uint64_t.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Always pass the target address as uint64_t.
Adjust tcg_out_{ld,st}_helper_args to match.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We already pass uint64_t to restore_state_to_opc; this changes all
of the other uses from insn_start through the encoding to decoding.
Reviewed-by: Anton Johansson <anjo@rev.ng>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
There is an edge condition prior to gcc13 for which optimization
is required to generate 16-byte atomic sequences. Detect this.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
TCG backends may need to defer to a helper to implement
the atomicity required by a given operation. Mirror the
interface used in system mode.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
With the current structure of cputlb.c, there is no difference
between the little-endian and big-endian entry points, aside
from the assert. Unify the pairs of functions.
Hoist the qemu_{ld,st}_helpers arrays to tcg.c.
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>
Create ldst_atomicity.c.inc.
Not required for user-only code loads, because we've ensured that
the page is read-only before beginning to translate code.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Instead of trying to unify all operations on uint64_t, use
mmu_lookup() to perform the basic tlb hit and resolution.
Create individual functions to handle access by size.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Instead of trying to unify all operations on uint64_t, pull out
mmu_lookup() to perform the basic tlb hit and resolution.
Create individual functions to handle access by size.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Instead of playing with offsetof in various places, use
MMUAccessType to index an array. This is easily defined
instead of the previous dummy padding array in the union.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
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>
Like cpu_in_exclusive_context, but also true if
there is no other cpu against which we could race.
Use it in tb_flush as a direct replacement.
Use it in cpu_loop_exit_atomic to ensure that there
is no loop against cpu_exec_step_atomic.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
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>
The round-robin scheduler will iterate over the CPU list with an
assigned budget until the next timer expiry and may exit early because
of a TB exit. This is fine under normal operation but with icount
enabled and SMP it is possible for a CPU to be starved of run time and
the system live-locks.
For example, booting a riscv64 platform with '-icount
shift=0,align=off,sleep=on -smp 2' we observe a livelock once the kernel
has timers enabled and starts performing TLB shootdowns. In this case
we have CPU 0 in M-mode with interrupts disabled sending an IPI to CPU
1. As we enter the TCG loop, we assign the icount budget to next timer
interrupt to CPU 0 and begin executing where the guest is sat in a busy
loop exhausting all of the budget before we try to execute CPU 1 which
is the target of the IPI but CPU 1 is left with no budget with which to
execute and the process repeats.
We try here to add some fairness by splitting the budget across all of
the CPUs on the thread fairly before entering each one. The CPU count
is cached on CPU list generation ID to avoid iterating the list on each
loop iteration. With this change it is possible to boot an SMP rv64
guest with icount enabled and no hangs.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Tested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Jamie Iles <quic_jiles@quicinc.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230427020925.51003-3-quic_jiles@quicinc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
A copy-paste bug had us looking at the victim cache for writes.
Cc: qemu-stable@nongnu.org
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Fixes: 08dff435e2 ("tcg: Probe the proper permissions for atomic ops")
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20230505204049.352469-1-richard.henderson@linaro.org>
The load side can use a relaxed load, which will surely happen before
the work item is run by async_safe_run_on_cpu() or before double-checking
under mmap_lock. The store side can use an atomic RMW operation.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
While the old type was correct in the ideal sense, some ABIs require
the argument to be zero-extended. Using uint32_t for all such values
is a decent compromise.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
At least RISC-V has the need to be able to perform a read
using execute permissions, outside of translation.
Add helpers to facilitate this.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn>
Tested-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Message-Id: <20230325105429.1142530-9-richard.henderson@linaro.org>
Message-Id: <20230412114333.118895-9-richard.henderson@linaro.org>
Add tcg expander and helper functions for and-compliment
vector with scalar operand.
Signed-off-by: Nazar Kazakov <nazar.kazakov@codethink.co.uk>
Message-Id: <20230428144757.57530-10-lawrence.hunter@codethink.co.uk>
[rth: Split out of larger patch.]
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
When PMP entry overlap part of the page, we'll set the tlb_size to 1, which
will make the address in tlb entry set with TLB_INVALID_MASK, and the next
access will again go through tlb_fill.However, this way will not work in
tb_gen_code() => get_page_addr_code_hostp(): the TLB host address will be
cached, and the following instructions can use this host address directly
which may lead to the bypass of PMP related check.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1542.
Signed-off-by: Weiwei Li <liweiwei@iscas.ac.cn>
Signed-off-by: Junqiang Wang <wangjunqiang@iscas.ac.cn>
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230422130329.23555-6-liweiwei@iscas.ac.cn>
Currently we report whether the TCG accelerator is in
'one-insn-per-tb' mode in the 'info status' output. This is a pretty
minor piece of TCG specific information, and we want to deprecate the
'singlestep' field of the associated QMP command. Move the
'one-insn-per-tb' reporting to 'info jit'.
We don't need a deprecate-and-drop period for this because the
HMP interface has no stability guarantees.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-id: 20230417164041.684562-8-peter.maydell@linaro.org
The only place left that looks at the old 'singlestep' global
variable is the TCG curr_cflags() function. Replace the old global
with a new 'one_insn_per_tb' which is defined in tcg-all.c and
declared in accel/tcg/internal.h. This keeps it restricted to the
TCG code, unlike 'singlestep' which was available to every file in
the system and defined in multiple different places for softmmu vs
linux-user vs bsd-user.
While we're making this change, use qatomic_read() and qatomic_set()
on the accesses to the new global, because TCG will read it without
holding a lock.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-id: 20230417164041.684562-4-peter.maydell@linaro.org
This commit adds 'one-insn-per-tb' as a property on the TCG
accelerator object, so you can enable it with
-accel tcg,one-insn-per-tb=on
It has the same behaviour as the existing '-singlestep' command line
option. We use a different name because 'singlestep' has always been
a confusing choice, because it doesn't have anything to do with
single-stepping the CPU. What it does do is force TCG emulation to
put one guest instruction in each TB, which can be useful in some
situations (such as analysing debug logs).
The existing '-singlestep' commandline options are decoupled from the
global 'singlestep' variable and instead now are syntactic sugar for
setting the accel property. (These can then go away after a
deprecation period.)
The global variable remains for the moment as:
* what the TCG code looks at to change its behaviour
* what HMP and QMP use to query and set the behaviour
In the following commits we'll clean those up to not directly
look at the global variable.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20230417164041.684562-2-peter.maydell@linaro.org
Alex Bennée
Richard Henderson
Paolo Bonzini
Gavin Shan
Jamie Iles
Nazar Kazakov
Weiwei Li
Peter Maydell