Use the carry-out vector as the basis to compute AF, CF and OF. The cost
is pretty much the same, because the carry-out is just four boolean
operations, and the code is much smaller because add/adc/sub/sbb now
share most of it.
A similar algorithm to what is used in target/i386/emulate can also be
used for APX, in order to build the result of CCMP/CTEST with a new CC_OP_*.
CCMP needs to place into the flags from either a subtraction or a constant
value; CTEST likewise place into the flags either an AND or a constant
value. The new CC_OP for CCMP and CTEST would store for a successful
predcate:
- in DST and SRC2, the result of the operation;
- in SRC, a carry-out vector for CCMP or zero for CTEST;
If the default flag value is used, DST/SRC/SRC2 can be filled with
constants:
- in DST the negated ZF;
- in SRC's top 2 bits, a value that results in the desired OF and CF;
- in SRC2 a suitable value (any of 0/1/~0/~1) that can be used
instead of DST to compute the desired SF and PF.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The logic is the same, but the majority(NOT a, b, c) is brought out
to a separate macro and implemented without NOT operations.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
No difference in generated code, but the XOR-based formula is
easily understood on its own. This will make more sense once
ADD/SUB stop using dst^src1^src2 to compute AF.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Remove all uses of 32-bit temporaries in emit.c.inc. Remove uses
in translate.c outside the large multiplexed generator functions.
tmp3_i32 is not used anymore and can go away.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Apply some of the simplifications used for RCL and RCR. tmp4 is not
used anywhere else, so remove it.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Commit 3f2a05b31e ("target/i386: Reset TSCs of parked vCPUs too on VM
reset") introduced a way to reset TSCs of parked vCPUs during VM reset to
prevent them getting desynchronized with the online vCPUs and therefore
causing the KVM PV clock to lose PVCLOCK_TSC_STABLE_BIT.
The way this was done was by registering a parked vCPU-specific QEMU reset
callback via qemu_register_reset().
However, it turns out that on particularly device-rich VMs QEMU reset
callbacks can take a long time to execute (which isn't surprising,
considering that they involve resetting all of VM devices).
In particular, their total runtime can exceed the 1-second TSC
synchronization window introduced in KVM commit 5d3cb0f6a8e3 ("KVM:
Improve TSC offset matching").
Since the TSCs of online vCPUs are only reset from "synchronize_post_reset"
AccelOps handler (which runs after all qemu_register_reset() handlers) this
essentially makes that fix ineffective on these VMs.
The easiest way to guarantee that these parked vCPUs are reset at the same
time as the online ones (regardless how long it takes for VM devices to
reset) is to piggyback on post-reset vCPU synchronization handler for one
of online vCPUs - as there is no generic post-reset AccelOps handler that
isn't per-vCPU.
The first online vCPU was selected for that since it is easily available
under "first_cpu" define.
This does not create an ordering issue since the order of vCPU TSC resets
does not matter.
Fixes: 3f2a05b31e ("target/i386: Reset TSCs of parked vCPUs too on VM reset")
Signed-off-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Link: https://lore.kernel.org/r/e8b85a5915f79aa177ca49eccf0e9b534470c1cd.1743099810.git.maciej.szmigiero@oracle.com
Cc: qemu-stable@nongnu.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The model number was mistakenly set to 0x0b (11) in commit ff04bc1ac4.
The correct value is 0x5b. This mistake occurred because the extended
model bits in cpuid[eax=0x1].eax were overlooked, and only the base
model was used.
Using the wrong model number can affect guest behavior. One known issue
is that vPMU (which relies on the model number) may fail to operate
correctly.
This patch corrects the model field by introducing a new vCPU version.
Fixes: ff04bc1ac4 ("target/i386: Introduce Zhaoxin Yongfeng CPU model")
Signed-off-by: Ewan Hai <ewanhai-oc@zhaoxin.com>
Link: https://lore.kernel.org/r/20250414075342.411626-1-ewanhai-oc@zhaoxin.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This is mostly for completeness; I noticed it because ENODEV is used internally
within scsi-disk.c, but when scsi_sense_from_errno(ENODEV) is called the resulting
sense is never used and instead scsi_sense_from_host_status() is called later
by scsi_req_complete_failed().
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
On MacOS, UI event loop has to be ran in the main thread of a process.
Because of that restriction, on this platform, qemu main event loop is
ran on another thread [1].
This breaks record/replay feature, which expects thread running qemu_init
to initialize hold this lock, breaking associated functional tests on
MacOS.
Thus, as a generalization, and similar to how BQL is handled, we release
it after init, and reacquire the lock before entering main event loop,
avoiding a special case if a separate thread is used.
Tested on MacOS with:
$ meson test -C build --setup thorough --print-errorlogs \
func-x86_64-x86_64_replay func-arm-arm_replay func-aarch64-aarch64_replay
$ ./build/qemu-system-x86_64 -nographic -icount shift=auto,rr=record,rrfile=replay.log
$ ./build/qemu-system-x86_64 -nographic -icount shift=auto,rr=replay,rrfile=replay.log
[1] https://gitlab.com/qemu-project/qemu/-/commit/f5ab12caba4f1656479c1feb5248beac1c833243
Fixes: https://gitlab.com/qemu-project/qemu/-/issues/2907
Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Reviewed-by: Nicholas Piggin <npiggin@gmail.com>
Tested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20250410225550.46807-2-pierrick.bouvier@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Final fixes for 10.0
- properly exit aarch64-softmmu tcg tests on unexpected exceptions
- fix deadlock when uninstalling a plugin
# -----BEGIN PGP SIGNATURE-----
#
# iQEzBAABCgAdFiEEZoWumedRZ7yvyN81+9DbCVqeKkQFAmf2P+0ACgkQ+9DbCVqe
# KkTHdAf/b4CGLNJBS7UV/eYxhgxJ2TM4KgfXm4FvrcS4V5rpNheMuLoQzgYDY+EC
# 5ioANRcY+rOK2Yd+upO4SbaUyAvf7C5kAtMgs6AYiqbm5F8b47FS/q8f9mBBJ+qB
# tl2FU6hjVjymMRoaPqqpPkkkHy9YNHe48TP4ohGfD6336JA5MW6v8nhdK+d3fsan
# AUCSWAe8gxjmM6qnrhku6vvJ5EOiqpsb8zhsWVXdZesSG+1ATm6KCo6hDd0OkMrk
# si6ML5txU1rxx5VOkGlDUo7ZV2TE/zdgMyaP/kXGtKuj+ipLMzqNjlD50xPJu05b
# o4hTbt8erzxduVHrBKT5qhKQl8BEpw==
# =LN/t
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 09 Apr 2025 05:37:49 EDT
# gpg: using RSA key 6685AE99E75167BCAFC8DF35FBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>" [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: 6685 AE99 E751 67BC AFC8 DF35 FBD0 DB09 5A9E 2A44
* tag 'pull-10.0-final-fixes-090425-1' of https://gitlab.com/stsquad/qemu:
tests/tcg/plugins: add plugin to test reset and uninstall
plugins/loader: fix deadlock when resetting/uninstalling a plugin
tests/tcg: fix semihosting SYS_EXIT for aarch64 in boot.S
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
In msys2 distribution objdump from gcc is using single tab character
prefix, but objdump from clang is using 4 white space characters instead.
The script will not identify any dll dependencies for a QEMU build
generated with clang. This in turn will fail the build, because there
will be no files inside dlldir and no setup file will be created.
Instead of checking for whitespace in prefix use lstrip to accommodate
for differences in outputs.
Signed-off-by: Arthur Sengileyev <arthur.sengileyev@gmail.com>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Mask out unsupported bits and return failure if attempting to set
any. This is not required by the IPMI spec, but it does require that
system software not change bits it isn't aware of.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-ID: <20250401140153.685523-6-npiggin@gmail.com>
Signed-off-by: Corey Minyard <corey@minyard.net>
If the dont-log flag is set in the 'timer use' field for the
'set watchdog' command, a watchdog timeout will not get logged as
a timer use expiration.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-ID: <20250401140153.685523-5-npiggin@gmail.com>
Signed-off-by: Corey Minyard <corey@minyard.net>
This requires some adjustments to callers to avoid possible behaviour
changes for PCI devices.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-ID: <20250401140153.685523-3-npiggin@gmail.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
[Rename pci_ipmi_bt_get_fwinfo to pci_ipmi_kcs_get_fwinfo in the
pci_ipmi_kcs.c file.]
Signed-off-by: Corey Minyard <corey@minyard.net>
IPMI drivers use p/k suffix in variable names depending on bt or kcs.
The pci bt driver must have come from the kcs driver because it's
still using k suffixes in some cases. Rename.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-ID: <20250401140153.685523-2-npiggin@gmail.com>
Signed-off-by: Corey Minyard <corey@minyard.net>
It's the right way to do it now, and it handles multiple instances
properly. I need multiple instances for some testing I'm doing so this
is the right thing to do.
Tested by doing:
(qemu) migrate -d exec:cat>filen.mig
before and after the fix, then:
scripts/analyze-migration.py -d desc -f file1.mig >file1.json
scripts/analyze-migration.py -d desc -f file2.mig >file2.json
diff file1.json file2.json
with no differences.
Signed-off-by: Corey Minyard <corey@minyard.net>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
We don't expect to hit exceptions in our testing so currently all the
vectors report an un-expected exception and then attempt to exit.
However for aarch64 we should always use the extended information
block as we do in _exit. Rather than duplicate the code on the error
handler just branch to the _exit handler with a failing status code.
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20250404115641.258048-1-alex.bennee@linaro.org>
The cpu-idle-states property causes a hard boot hang. Rather than documenting
the workaround, perform the removal from the devicetree automatically.
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Bernhard Beschow <shentey@gmail.com>
[Bernhard: split patch, update documentation, adapt commit message]
Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Message-ID: <20250405214900.7114-3-shentey@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
The comments for segment_size and allocation_unit_size incorrectly
described them as 4KB. According to the UFS specification,
segment_size is expressed in units of 512 bytes.
Given segment_size = 0x2000 (8192), the actual size is 4MB.
Similarly, allocation_unit_size = 1 means 1 segment = 4MB.
This patch updates the comments to reflect the correct size.
Signed-off-by: Keoseong Park <keosung.park@samsung.com>
Reviewed-by: Jeuk Kim <jeuk20.kim@samsung.com>
Message-ID: <20250403092140epcms2p355a7f039871b3e5b409754ef450b9158@epcms2p3>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
In commit 8fd2518ef2 ("hw: Centralize handling of -machine dumpdtb
option") the call to dump was moved with respect to the init of the
machine. This resulted in the device tree missing parts of the machine
description, depending on how they construct their device tree.
The arm virt machine is missing some PSCI nodes, while the riscv one
is missing most of its content.
Move the dump to after the notifiers have been run, allowing
virt_machine_done to be called and the device tree to be fully
populated.
Fixes: 8fd2518ef2 ("hw: Centralize handling of -machine dumpdtb option")
Signed-off-by: Joel Stanley <joel@jms.id.au>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20250401041509.719153-1-joel@jms.id.au>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
loongarch bug fix for 10.0
# -----BEGIN PGP SIGNATURE-----
#
# iLMEAAEKAB0WIQS4/x2g0v3LLaCcbCxAov/yOSY+3wUCZ/THrQAKCRBAov/yOSY+
# 37asA/0YLOX9BvqG5Qk2cDYJD7ZCTuVsd7bJ0VHGbbf4vuuvmD8ubGeJf1F8cse0
# mufip5/DHlzV3gB+G/vavkG/i45mMDlkyFgvuYaEHZiYH/+3S1Np0/iyl3RFpGNz
# Srua0Flo4PppYRbQ2NG0puvpuhNziTQTMneTm2YiMMBxl4p2OA==
# =BwNB
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 08 Apr 2025 02:52:29 EDT
# gpg: using RSA key B8FF1DA0D2FDCB2DA09C6C2C40A2FFF239263EDF
# gpg: Good signature from "Song Gao <m17746591750@163.com>" [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: B8FF 1DA0 D2FD CB2D A09C 6C2C 40A2 FFF2 3926 3EDF
* tag 'pull-loongarch-20250408' of https://github.com/gaosong715/qemu:
hw/loongarch/virt: Replace destination error with error_abort
hw/loongarch/virt: Fix cpuslot::cpu set at last in virt_cpu_plug()
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Wei Liu