Add libdw-based functions for loading and querying debuginfo. Load
debuginfo from the system and the linux-user loaders.
This is useful for the upcoming perf support, which can then put
human-readable guest symbols instead of raw guest PCs into perfmap and
jitdump files.
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-Id: <20230112152013.125680-3-iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
If we update an existing memslot (e.g., resize, split), we temporarily
remove the memslot to re-add it immediately afterwards. These updates
are not atomic, especially not for KVM VCPU threads, such that we can
get spurious faults.
Let's inhibit most KVM ioctls while performing relevant updates, such
that we can perform the update just as if it would happen atomically
without additional kernel support.
We capture the add/del changes and apply them in the notifier commit
stage instead. There, we can check for overlaps and perform the ioctl
inhibiting only if really required (-> overlap).
To keep things simple we don't perform additional checks that wouldn't
actually result in an overlap -- such as !RAM memory regions in some
cases (see kvm_set_phys_mem()).
To minimize cache-line bouncing, use a separate indicator
(in_ioctl_lock) per CPU. Also, make sure to hold the kvm_slots_lock
while performing both actions (removing+re-adding).
We have to wait until all IOCTLs were exited and block new ones from
getting executed.
This approach cannot result in a deadlock as long as the inhibitor does
not hold any locks that might hinder an IOCTL from getting finished and
exited - something fairly unusual. The inhibitor will always hold the BQL.
AFAIKs, one possible candidate would be userfaultfd. If a page cannot be
placed (e.g., during postcopy), because we're waiting for a lock, or if the
userfaultfd thread cannot process a fault, because it is waiting for a
lock, there could be a deadlock. However, the BQL is not applicable here,
because any other guest memory access while holding the BQL would already
result in a deadlock.
Nothing else in the kernel should block forever and wait for userspace
intervention.
Note: pause_all_vcpus()/resume_all_vcpus() or
start_exclusive()/end_exclusive() cannot be used, as they either drop
the BQL or require to be called without the BQL - something inhibitors
cannot handle. We need a low-level locking mechanism that is
deadlock-free even when not releasing the BQL.
Signed-off-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Tested-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20221111154758.1372674-4-eesposit@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Using the new accel-blocker API, mark where ioctls are being called
in KVM. Next, we will implement the critical section that will take
care of performing memslots modifications atomically, therefore
preventing any new ioctl from running and allowing the running ones
to finish.
Signed-off-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20221111154758.1372674-3-eesposit@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This API allows the accelerators to prevent vcpus from issuing
new ioctls while execting a critical section marked with the
accel_ioctl_inhibit_begin/end functions.
Note that all functions submitting ioctls must mark where the
ioctl is being called with accel_{cpu_}ioctl_begin/end().
This API requires the caller to always hold the BQL.
API documentation is in sysemu/accel-blocker.h
Internally, it uses a QemuLockCnt together with a per-CPU QemuLockCnt
(to minimize cache line bouncing) to keep avoid that new ioctls
run when the critical section starts, and a QemuEvent to wait
that all running ioctls finish.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221111154758.1372674-2-eesposit@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
* Always send errors to logfile when daemonized (Greg)
* Add support for IDE CompactFlash card (Lubomir)
* First round of build system cleanups (myself)
* First round of feature removals (myself)
* Reduce "qemu/accel.h" inclusion (Philippe)
-----BEGIN PGP SIGNATURE-----
iQFIBAABCAAyFiEE8TM4V0tmI4mGbHaCv/vSX3jHroMFAmO3Ym0UHHBib256aW5p
QHJlZGhhdC5jb20ACgkQv/vSX3jHroNYmwf+LHEw+4T0fk1+2NfgIzH3+8s1EqDm
Ai56EjxO/p5NUptflXAnhn4P3LawswmmNE0ZIFFFBgwG5E9L+Jj/u5efuLu4uYPg
bboEBDn8nxSNN2l08u9TyS6kSWSxbwwrs7i2+V+4uQIlVIcCHu+A0vpXns4vWwY0
zZGF8CgJKDQdPIxdXrH8+6/xtadQ8uDkYsAWDiY/nhozCsCUTAZGTXWEQbHJLARI
Z4X+Cmz/NFB9G4ka6K/y0HbQw99KA8G/EMPUSglN0ya10yjpyzrmeI7IlIves+5U
8lhCZXyBhaV9GXlIK1vIgEXlHf83C19a+v0DpW0bpxK631n2VR5y3CArBg==
=2Koq
-----END PGP SIGNATURE-----
Merge tag 'for-upstream' of https://gitlab.com/bonzini/qemu into staging
* Atomic memslot updates for KVM (Emanuele, David)
* Always send errors to logfile when daemonized (Greg)
* Add support for IDE CompactFlash card (Lubomir)
* First round of build system cleanups (myself)
* First round of feature removals (myself)
* Reduce "qemu/accel.h" inclusion (Philippe)
# gpg: Signature made Thu 05 Jan 2023 23:51:09 GMT
# gpg: using RSA key F13338574B662389866C7682BFFBD25F78C7AE83
# gpg: issuer "pbonzini@redhat.com"
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [full]
# gpg: aka "Paolo Bonzini <pbonzini@redhat.com>" [full]
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4 E2F7 7E15 100C CD36 69B1
# Subkey fingerprint: F133 3857 4B66 2389 866C 7682 BFFB D25F 78C7 AE83
* tag 'for-upstream' of https://gitlab.com/bonzini/qemu: (24 commits)
i386: SGX: remove deprecated member of SGXInfo
target/i386: Add SGX aex-notify and EDECCSSA support
util: remove support -chardev tty and -chardev parport
util: remove support for hex numbers with a scaling suffix
KVM: remove support for kernel-irqchip=off
docs: do not talk about past removal as happening in the future
meson: accept relative symlinks in "meson introspect --installed" data
meson: cleanup compiler detection
meson: support meson 0.64 -Doptimization=plain
configure: test all warnings
tests/qapi-schema: remove Meson workaround
meson: cleanup dummy-cpus.c rules
meson: tweak hardening options for Windows
configure: remove backwards-compatibility and obsolete options
configure: preserve qemu-ga variables
configure: cleanup $cpu tests
configure: remove dead function
configure: remove useless write_c_skeleton
ide: Add "ide-cf" driver, a CompactFlash card
ide: Add 8-bit data mode
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Now that qtest is available on all targets including Windows, dummy-cpus.c
is included unconditionally in the build. It also does not need to be
compiled per-target.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
As in page_get_flags, we need to try again with the mmap
lock held if we fail a page lookup.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Because we allow lockless lookups, we have to be careful
when it is freed. Use rcu to delay the free until safe.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
When called from syscall(), we are not within a TB and pc == 0.
We can skip the check for invalidating the current TB.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We have been allocating a worst case number of arguments
to support calls. Instead, allow the size to vary.
By default leave space for 4 args, to maximize reuse,
but allow calls to increase the number of args to 32.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
[PMD: Split patch in two]
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221218211832.73312-3-philmd@linaro.org>
In order to have variable size allocated TCGOp, pass the number
of arguments we use (and would allocate) up to tcg_op_alloc().
This alters tcg_emit_op(), tcg_op_insert_before() and
tcg_op_insert_after() prototypes.
In tcg_op_alloc() ensure the number of arguments is in range.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
[PMD: Extracted from bigger patch]
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221218211832.73312-2-philmd@linaro.org>
Better to re-use the existing function for copying ops.
Acked-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We copied all of the arguments in copy_op_nocheck.
We only need to replace the one argument that we change.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The function pointer is immediately after the output and input
operands; no need to search.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Narrow the scope of the lock to the actual read/write,
moving the cpu_transation_failed call outside the lock.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Only the system emulation part of TB maintainance uses the
page_collection structure. Restrict its declaration (and the
functions requiring it) to tb-maint.c.
Convert the 'len' argument of tb_invalidate_phys_page_fast__locked()
from signed to unsigned.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221209093649.43738-6-philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221209093649.43738-5-philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Emphasize this function is called with pages locked.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221209093649.43738-4-philmd@linaro.org>
[rth: Use "__locked" suffix, to match other instances.]
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Commit d9bb58e510 ("tcg: move tcg related files into accel/tcg/
subdirectory") introduced accel/tcg/trace-events, so we don't
need to use the root trace-events anymore.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221209093649.43738-3-philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Missed in commit 6526919224 ("accel/tcg: Restrict cpu_io_recompile()
from other accelerators").
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221209093649.43738-2-philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The only thing that still touches PageDesc in translate-all.c
are some locking routines related to tb-maint.c which have not
yet been moved. Do so now.
Move some code up in tb-maint.c as well, to untangle the maze
of ifdefs, and allow a sensible final ordering.
Move some declarations from exec/translate-all.h to internal.h,
as they are only used within accel/tcg/.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Now that PageDesc is not used for user-only, and for system
it is only used for tb maintenance, move the implementation
into tb-main.c appropriately ifdefed.
We have not yet eliminated all references to PageDesc for
user-only, so retain a typedef to the structure without definition.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This page tracking implementation is specific to user-only,
since the system softmmu version is in cputlb.c. Move it
out of translate-all.c to user-exec.c.
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>
Make bsd-user match linux-user in not marking host pages
as reserved. This isn't especially effective anyway, as
it doesn't take into account any heap memory that qemu
may allocate after startup.
Reviewed-by: Warner Losh <imp@bsdimp.com>
Tested-by: Warner Losh <imp@bsdimp.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Continue weaning user-only away from PageDesc.
Use an interval tree to record target data.
Chunk the data, to minimize allocation overhead.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Begin weaning user-only away from PageDesc.
Since, for user-only, all TB (and page) manipulation is done with
a single mutex, and there is no virtual/physical discontinuity to
split a TB across discontinuous pages, place all of the TBs into
a single IntervalTree. This makes it trivial to find all of the
TBs intersecting a range.
Retain the existing PageDesc + linked list implementation for
system mode. Move the portion of the implementation that overlaps
the new user-only code behind the common ifdef.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Rename to tb_remove_all, to remove the PageDesc "page" from the name,
and to avoid suggesting a "flush" in the icache sense.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Commit 012d4c96e2 changed the visitor functions taking Error ** to
return bool instead of void, and the commits following it used the new
return value to simplify error checking. Since then a few more uses
in need of the same treatment crept in. Do that. All pretty
mechanical except for
* balloon_stats_get_all()
This is basically the same transformation commit 012d4c96e2 applied
to the virtual walk example in include/qapi/visitor.h.
* set_max_queue_size()
Additionally replace "goto end of function" by return.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20221121085054.683122-10-armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Isolate the code protected by setjmp. Fixes:
translate-all.c: In function ‘tb_gen_code’:
translate-all.c:748:51: error: argument ‘cflags’ might be clobbered by ‘longjmp’ or ‘vfork’ [-Werror=clobbered]
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
if QEMU is configured with modules enabled, it is possible that the
load of an accelerator module will fail.
Exit in this case, relying on module_object_class_by_name to report
the specific load error if any.
Signed-off-by: Claudio Fontana <cfontana@suse.de>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
[claudio: changed abort() to exit(1)]
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20220929093035.4231-6-cfontana@suse.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Delay cpu_list_add until realize is complete, so that cross-cpu
interaction does not happen with incomplete cpu state. For this,
we must delay plugin initialization out of tcg_exec_realizefn,
because no cpu_index has been assigned.
Fixes a problem with cross-cpu jump cache flushing, when the
jump cache has not yet been allocated.
Fixes: a976a99a29 ("include/hw/core: Create struct CPUJumpCache")
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Reported-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The value passed is always true, and if the target's
synchronize_from_tb hook is non-trivial, not exiting
may be erroneous.
Reviewed-by: Claudio Fontana <cfontana@suse.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Add a way to examine the unwind data without actually
restoring the data back into env.
Reviewed-by: Claudio Fontana <cfontana@suse.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Currently signal SIGIPI [=SIGUSR1] is used to kick the dummy CPU
when qtest accelerator is used. However SIGUSR1 is unsupported on
Windows. To support Windows, we add a QemuSemaphore CPUState::sem
to kick the dummy CPU instead for Windows.
Signed-off-by: Xuzhou Cheng <xuzhou.cheng@windriver.com>
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20221028045736.679903-2-bin.meng@windriver.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
All targets have been updated. Use the tcg_ops target hook
exclusively, which allows the compat code to be removed.
Reviewed-by: Claudio Fontana <cfontana@suse.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Add a tcg_ops hook to replace the restore_state_to_opc
function call. Because these generic hooks cannot depend
on target-specific types, temporarily, copy the current
target_ulong data[] into uint64_t d64[].
Reviewed-by: Claudio Fontana <cfontana@suse.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Since the only user, Arm MTE, always requires allocation,
merge the get and alloc functions to always produce a
non-null result. Also assume that the user has already
checked page validity.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Since "target data" is always user-only, move it out of
translate-all.c to user-exec.c.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Flush translation blocks in bulk, rather than page-by-page.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Use the existing function for clearing target data.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
When PAGE_RESET is set, we are replacing pages with new
content, which means that we need to invalidate existing
cached data, such as TranslationBlocks. Perform the
reset invalidate while we're doing other invalidates,
which allows us to remove the separate invalidates from
the user-only mmap/munmap/mprotect routines.
In addition, restrict invalidation to PAGE_EXEC pages.
Since cdf7130851, we have validated PAGE_EXEC is present
before translation, which means we can assume that if the
bit is not present, there are no translations to invalidate.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We do not require detection of overlapping TBs here,
so use the more appropriate function.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We missed this function when we introduced tb_page_addr_t.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This function is is never called with a real range,
only for a single page. Drop the second parameter
and rename to tb_invalidate_phys_page.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Rename to tb_invalidate_phys_page_unwind to emphasize that
we also detect invalidating the current TB, and also to free
up that name for other usage.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This data structure will be replaced for user-only: add accessors.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
When we added the fast path, we initialized page_addr[] early.
These stores in and around tb_page_add() are redundant; remove them.
Fixes: 50627f1b7b ("accel/tcg: Add fast path for translator_ld*")
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The results of the calls to cpu_get_tb_cpu_state,
current_{pc,cs_base,flags}, are not used.
In tb_invalidate_phys_page, use bool for current_tb_modified.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
There are no users outside of accel/tcg; this function
does not need to be defined in exec-all.h.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Move all of the TranslationBlock flushing and page linking
code from translate-all.c to tb-maint.c.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
These items printf, and could be replaced with proper
tracepoints if we really cared.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Use a constant target data allocation size for all pages.
This will be necessary to reduce overhead of page tracking.
Since TARGET_PAGE_DATA_SIZE is now required, we can use this
to omit data tracking for targets that don't require it.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Profiling QEMU during Fedora 35 for PPC64 boot revealed that a
considerable amount of time was being spent in
check_for_breakpoints() (0.61% of total time on PPC64 and 2.19% on
amd64), even though it was just checking that its queue was empty
and returning, when no breakpoints were set. It turns out this
function is not inlined by the compiler and it's always called by
helper_lookup_tb_ptr(), one of the most called functions.
By leaving only the check for empty queue in
check_for_breakpoints() and moving the remaining code to
check_for_breakpoints_slow(), called only when the queue is not
empty, it's possible to avoid the call overhead. An improvement of
about 3% in total time was measured on POWER9.
Signed-off-by: Leandro Lupori <leandro.lupori@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20221025202424.195984-2-leandro.lupori@eldorado.org.br>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Commit a82fd5a4ec was intended to be a code cleanup, but
unfortunately it has a bug. It moves the initialization of the
TCG cflags from the "start a new vcpu" function to the
thread handler; this is fine when each vcpu has its own thread,
but when we are doing round-robin of vcpus on a single thread
we end up only initializing the cflags for CPU 0, not for any
of the others.
The most obvious effect of this bug is that running in icount
mode with more than one CPU is broken; typically the guest
hangs shortly after it brings up the secondary CPUs.
This reverts commit a82fd5a4ec.
Cc: qemu-stable@nongnu.org
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20221021163409.3674911-1-peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
There are cases that malicious virtual machine can cause CPU stuck (due
to event windows don't open up), e.g., infinite loop in microcode when
nested #AC (CVE-2015-5307). No event window means no event (NMI, SMI and
IRQ) can be delivered. It leads the CPU to be unavailable to host or
other VMs. Notify VM exit is introduced to mitigate such kind of
attacks, which will generate a VM exit if no event window occurs in VM
non-root mode for a specified amount of time (notify window).
A new KVM capability KVM_CAP_X86_NOTIFY_VMEXIT is exposed to user space
so that the user can query the capability and set the expected notify
window when creating VMs. The format of the argument when enabling this
capability is as follows:
Bit 63:32 - notify window specified in qemu command
Bit 31:0 - some flags (e.g. KVM_X86_NOTIFY_VMEXIT_ENABLED is set to
enable the feature.)
Users can configure the feature by a new (x86 only) accel property:
qemu -accel kvm,notify-vmexit=run|internal-error|disable,notify-window=n
The default option of notify-vmexit is run, which will enable the
capability and do nothing if the exit happens. The internal-error option
raises a KVM internal error if it happens. The disable option does not
enable the capability. The default value of notify-window is 0. It is valid
only when notify-vmexit is not disabled. The valid range of notify-window
is non-negative. It is even safe to set it to zero since there's an
internal hardware threshold to be added to ensure no false positive.
Because a notify VM exit may happen with VM_CONTEXT_INVALID set in exit
qualification (no cases are anticipated that would set this bit), which
means VM context is corrupted. It would be reflected in the flags of
KVM_EXIT_NOTIFY exit. If KVM_NOTIFY_CONTEXT_INVALID bit is set, raise a KVM
internal error unconditionally.
Acked-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
Message-Id: <20220929072014.20705-5-chenyi.qiang@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Expose struct KVMState out of kvm-all.c so that the field of struct
KVMState can be accessed when defining target-specific accelerator
properties.
Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
Message-Id: <20220929072014.20705-4-chenyi.qiang@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Several hypervisor capabilities in KVM are target-specific. When exposed
to QEMU users as accelerator properties (i.e. -accel kvm,prop=value), they
should not be available for all targets.
Add a hook for targets to add their own properties to -accel kvm, for
now no such property is defined.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220929072014.20705-3-chenyi.qiang@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This removes the final hard coding of kvm_enabled() in gdbstub and
moves the check to an AccelOps.
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Mads Ynddal <mads@ynddal.dk>
Message-Id: <20220929114231.583801-46-alex.bennee@linaro.org>
As HW virtualization requires specific support to handle breakpoints
lets push out special casing out of the core gdbstub code and into
AccelOpsClass. This will make it easier to add other accelerator
support and reduces some of the stub shenanigans.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Mads Ynddal <mads@ynddal.dk>
Message-Id: <20220929114231.583801-45-alex.bennee@linaro.org>
The support of single-stepping is very much dependent on support from
the accelerator we are using. To avoid special casing in gdbstub move
the probing out to an AccelClass function so future accelerators can
put their code there.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Mads Ynddal <mads@ynddal.dk>
Message-Id: <20220929114231.583801-44-alex.bennee@linaro.org>
Prepare for targets to be able to produce TBs that can
run in more than one virtual context.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The availability of tb->pc will shortly be conditional.
Introduce accessor functions to minimize ifdefs.
Pass around a known pc to places like tcg_gen_code,
where the caller must already have the value.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Wrap the bare TranslationBlock pointer into a structure.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This function has two users, who use it incompatibly.
In tlb_flush_page_by_mmuidx_async_0, when flushing a
single page, we need to flush exactly two pages.
In tlb_flush_range_by_mmuidx_async_0, when flushing a
range of pages, we need to flush N+1 pages.
This avoids double-flushing of jmp cache pages in a range.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Let tb->page_addr[0] contain the address of the first byte of the
translated block, rather than the address of the page containing the
start of the translated block. We need to recover this value anyway
at various points, and it is easier to discard a page offset when it
is not needed, which happens naturally via the existing find_page shift.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Use the pc coming from db->pc_first rather than the TB.
Use the cached host_addr rather than re-computing for the
first page. We still need a separate lookup for the second
page because it won't be computed for DisasContextBase until
the translator actually performs a read from the page.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Bool is more appropriate type for the alloc parameter.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This bitmap is created and discarded immediately.
We gain nothing by its existence.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220822232338.1727934-2-richard.henderson@linaro.org>
Now that we have collected all of the page data into
CPUTLBEntryFull, provide an interface to record that
all in one go, instead of using 4 arguments. This interface
allows CPUTLBEntryFull to be extended without having to
change the number of arguments.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Add an interface to return the CPUTLBEntryFull struct
that goes with the lookup. The result is not intended
to be valid across multiple lookups, so the user must
use the results immediately.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
When PAGE_WRITE_INV is set when calling tlb_set_page,
we immediately set TLB_INVALID_MASK in order to force
tlb_fill to be called on the next lookup. Here in
probe_access_internal, we have just called tlb_fill
and eliminated true misses, thus the lookup must be valid.
This allows us to remove a warning comment from s390x.
There doesn't seem to be a reason to change the code though.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This field is only written, not read; remove it.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This structure will shortly contain more than just
data for accessing MMIO. Rename the 'addr' member
to 'xlat_section' to more clearly indicate its purpose.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Before: 35.912 s ± 0.168 s
After: 35.565 s ± 0.087 s
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220811151413.3350684-5-alex.bennee@linaro.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220923084803.498337-5-clg@kaod.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This is a follow-up for commit 892a4f6a75 ("linux-user: Add partial
support for MADV_DONTNEED"), which added passthrough for anonymous
mappings. File mappings can be handled in a similar manner.
In order to do that, mark pages, for which mmap() was passed through,
with PAGE_PASSTHROUGH, and then allow madvise() passthrough for these
pages. Drop the explicit PAGE_ANON check, since anonymous mappings are
expected to have PAGE_PASSTHROUGH anyway.
Add PAGE_PASSTHROUGH to PAGE_STICKY in order to keep it on mprotect().
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220725125043.43048-1-iii@linux.ibm.com>
Message-Id: <20220906000839.1672934-5-iii@linux.ibm.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Reported by Coverity as CID 1490142. Since the size is constant and the
lifetime is the same as the StatsDescriptors struct, embed the struct
directly instead of using a separate allocation.
Suggested-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The following scenario can happen if QEMU sets more RESET flags while
the KVM_RESET_DIRTY_RINGS ioctl is ongoing on another host CPU:
CPU0 CPU1 CPU2
------------------------ ------------------ ------------------------
fill gfn0
store-rel flags for gfn0
fill gfn1
store-rel flags for gfn1
load-acq flags for gfn0
set RESET for gfn0
load-acq flags for gfn1
set RESET for gfn1
do ioctl! ----------->
ioctl(RESET_RINGS)
fill gfn2
store-rel flags for gfn2
load-acq flags for gfn2
set RESET for gfn2
process gfn0
process gfn1
process gfn2
do ioctl!
etc.
The three load-acquire in CPU0 synchronize with the three store-release
in CPU2, but CPU0 and CPU1 are only synchronized up to gfn1 and CPU1
may miss gfn2's fields other than flags.
The kernel must be able to cope with invalid values of the fields, and
userspace *will* invoke the ioctl once more. However, once the RESET flag
is cleared on gfn2, it is lost forever, therefore in the above scenario
CPU1 must read the correct value of gfn2's fields.
Therefore RESET must be set with a store-release, that will synchronize
with KVM's load-acquire in CPU1.
Cc: Gavin Shan <gshan@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Cache the translation from guest to host address, so we may
use direct loads when we hit on the primary translation page.
Look up the second translation page only once, during translation.
This obviates another lookup of the second page within tb_gen_code
after translation.
Fixes a bug in that plugin_insn_append should be passed the bytes
in the original memory order, not bswapped by pieces.
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Tested-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Pass these along to translator_loop -- pc may be used instead
of tb->pc, and host_pc is currently unused. Adjust all targets
at one time.
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Tested-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
It was non-obvious to me why we can raise an exception in
the middle of a comparison function, but it works.
While nearby, use TARGET_PAGE_ALIGN instead of open-coding.
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Simplify the implementation of get_page_addr_code_hostp
by reusing the existing probe_access infrastructure.
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Tested-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The base qemu_ram_addr_from_host function is already in
softmmu/physmem.c; move the nofail version to be adjacent.
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Tested-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The function is not used outside of cpu-exec.c. Move it and
its subroutines up in the file, before the first use.
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Tested-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The mmap_lock is held around tb_gen_code. While the comment
is correct that the lock is dropped when tb_gen_code runs out
of memory, the lock is *not* dropped when an exception is
raised reading code for translation.
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Tested-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The current implementation is a no-op, simply returning addr.
This is incorrect, because we ought to be checking the page
permissions for execution.
Make get_page_addr_code inline for both implementations.
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Tested-by: Ilya Leoshkevich <iii@linux.ibm.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The KVM_DIRTY_GFN_F_DIRTY flag ensures that the entry is valid. If
the read of the fields are not ordered after the read of the flag,
QEMU might see stale values.
Cc: Gavin Shan <gshan@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-M none creates a guest without a vCPU, causing the following error:
$ ./qemu-system-x86_64 -qmp stdio -M none -accel kvm
{execute:qmp_capabilities}
{"return": {}}
{execute: query-stats-schemas}
Segmentation fault (core dumped)
Fix it by not querying the vCPU stats if first_cpu is NULL.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
aarch64 stores MTE tags in target_date, and they should be reset by
MADV_DONTNEED.
Signed-off-by: Vitaly Buka <vitalybuka@google.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220711220028.2467290-1-vitalybuka@google.com>
[lv: fix code style issues]
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
The function icount_prepare_for_run() is called with the iothread
unlocked, but it can call icount_notify_aio_contexts() which will
run qemu timer handlers. Those are supposed to be run only with
the iothread lock held, so take the lock while we do that.
Since icount mode runs everything on a single thread anyway,
not holding the lock is likely mostly not going to introduce
races, but it can cause us to trip over assertions that we
do hold the lock, such as the one reported in issue 1130.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1130
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Tested-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
Message-id: 20220801164527.3134765-1-peter.maydell@linaro.org
perror() is designed to append the decoded errno value to a
string. This, however, only makes sense if we called something that
actually sets errno prior to that.
For the callers that check for split irqchip support that is not the
case, and we end up with confusing error messages that end in
"success". Use error_report() instead.
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20220728142446.438177-1-cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Coverity complains that there is a codepath in the query_stats()
function where it can leak the memory pointed to by stats_list. This
can only happen if the caller passes something other than
STATS_TARGET_VM or STATS_TARGET_VCPU as the 'target', which no
callsite does. Enforce this assumption using g_assert_not_reached(),
so that if we have a future bug we hit the assert rather than
silently leaking memory.
Resolves: Coverity CID 1490140
Fixes: cc01a3f4ca ("kvm: Support for querying fd-based stats")
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20220719134853.327059-1-peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Setup a negative feedback system when vCPU thread
handling KVM_EXIT_DIRTY_RING_FULL exit by introducing
throttle_us_per_full field in struct CPUState. Sleep
throttle_us_per_full microseconds to throttle vCPU
if dirtylimit is in service.
Signed-off-by: Hyman Huang(黄勇) <huangy81@chinatelecom.cn>
Reviewed-by: Peter Xu <peterx@redhat.com>
Message-Id: <977e808e03a1cef5151cae75984658b6821be618.1656177590.git.huangy81@chinatelecom.cn>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
