The realpath() function can return NULL on error, so we need to check
for it to avoid crashing when we try to strstr() into it.
This can happen if we run out of memory, or if /sys/ is not mounted,
among other situations.
Fixes: Coverity 1459913, 1460474
Fixes: ce317be98d ("exec: fetch the alignment of Linux devdax pmem character device nodes")
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Jingqi Liu <jingqi.liu@intel.com>
Message-id: 20210812151525.31456-1-peter.maydell@linaro.org
In the alignment check added to qemu_ram_alloc_from_fd() in commit
ce317be98d, the condition includes a check that 'mr' is not
NULL. This check is unnecessary because we can assume that the
caller always passes us a valid MemoryRegion, and indeed later in the
function we assume mr is not NULL when we pass it to file_ram_alloc()
as new_block->mr. Remove it.
Fixes: Coverity 1459867
Fixes: ce317be98d ("exec: fetch the alignment of Linux devdax pmem character device nodes")
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Jingqi Liu <jingqi.liu@intel.com>
Message-id: 20210812150624.29139-1-peter.maydell@linaro.org
The QEMU_ARCH_VIRTIO_* defines are used only in one file,
qdev-monitor.c. Move them to that file.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-id: 20210730105947.28215-7-peter.maydell@linaro.org
Instead of using an ifdef ladder in arch_init.c (which we then have
to manually update every time we add or remove a target
architecture), have meson.build put "#define QEMU_ARCH QEMU_ARCH_FOO"
in the config-target.h file.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20210730105947.28215-5-peter.maydell@linaro.org
arch_init.c does very little but has a long list of #include lines.
Remove all the unnecessary ones.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20210730105947.28215-4-peter.maydell@linaro.org
The kvm_available() function reports whether KVM support was
compiled into the QEMU binary; it returns the value of the
CONFIG_KVM define.
The only place in the codebase where we use this function is
in qmp_query_kvm(). Now that accelerators are based on QOM
classes we can instead use accel_find("kvm") and remove the
kvm_available() function.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20210730105947.28215-3-peter.maydell@linaro.org
The xen_available() function is used only to produce an error
for some Xen-specific command line options in QEMU binaries where
Xen support was not compiled in: it just returns the value of
the CONFIG_XEN define.
Now that accelerators are QOM classes, we can check for
"does this binary have Xen compiled in" with accel_find("xen"),
and drop the xen_available() function.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20210730105947.28215-2-peter.maydell@linaro.org
The Error ** argument must be NULL, &error_abort, &error_fatal, or a
pointer to a variable containing NULL. Passing an argument of the
latter kind twice without clearing it in between is wrong: if the
first call sets an error, it no longer points to NULL for the second
call.
machine_parse_property_opt() is wrong that way: it passes @errp to
keyval_parse() without checking for failure, then passes it to
keyval_merge(). Harmless, since the only caller passes &error_fatal.
Clean up: drop the parameter, and use &error_fatal directly.
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20210720125408.387910-16-armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
[Rebased, conflict with commit a3c2f12830 resolved]
We did this with scripts/coccinelle/use-error_fatal.cocci before, in
commit 50beeb6809 and 007b06578a. This commit cleans up rarer
variations that don't seem worth matching with Coccinelle.
Cc: Thomas Huth <thuth@redhat.com>
Cc: Cornelia Huck <cornelia.huck@de.ibm.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Juan Quintela <quintela@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Marc-André Lureau <marcandre.lureau@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20210720125408.387910-2-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
When adding RAM_NORESERVE, we forgot to remove the old assertion when
adding the updated one, most probably when reworking the patches or
rebasing. We can easily crash QEMU by adding
-object memory-backend-ram,id=mem0,size=500G,reserve=off
to the QEMU cmdline:
qemu-system-x86_64: ../softmmu/physmem.c:2146: qemu_ram_alloc_internal:
Assertion `(ram_flags & ~(RAM_SHARED | RAM_RESIZEABLE | RAM_PREALLOC))
== 0' failed.
Fix it by removing the old assertion.
Fixes: 8dbe22c686 ("memory: Introduce RAM_NORESERVE and wire it up in qemu_ram_mmap()")
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta@ionos.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-id: 20210805092350.31195-1-david@redhat.com
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
-readconfig is still recording SMP options in QemuOpts instead of
using machine_opts_dict. This means that SMP options from -readconfig
are ignored.
Just stop using QemuOpts for -smp, making it return false for
is_qemuopts_group. Configuration files will merge the values in
machine_opts_dict using the new function machine_merge_property.
At the same time, fix -mem-prealloc which looked at QemuOpts to find the
number of guest CPUs, which it used as the default number of preallocation
threads.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
It will be used to parse smp-opts config groups from configuration
files. The point to note is that it does not steal a reference
from the caller. This is better because this function will be called
from qemu_config_foreach's callback; qemu_config_foreach does not cede
its reference to the qdict to the callback, and wants to free it. To
balance that extra reference, machine_parse_property_opt now needs
a qobject_unref.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
We continue after -smp help:
$ qemu-system-x86_64 -smp help -display none -monitor stdio
smp-opts options:
cores=<num>
cpus=<num>
dies=<num>
maxcpus=<num>
sockets=<num>
threads=<num>
QEMU 6.0.50 monitor - type 'help' for more information
(qemu)
Other options, such as -object help and -device help, don't.
Adjust -smp not to continue either.
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20210720125408.387910-17-armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta@ionos.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Do not instantiate an extra default VGA device if -device virtio-vga-gl
is provided.
Related to commit b36eb8860f ("virtio-gpu:
add virtio-vga-gl")
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20210701062421.721414-1-marcandre.lureau@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Use it to avoid some clang-12 -Watomic-alignment errors,
forcing some structures to be aligned and as a pointer when
we have ensured that the address is aligned.
Tested-by: Cole Robinson <crobinso@redhat.com>
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>
qemu can't start a xen vm after commit d8fb7d0969
"vl: switch -M parsing to keyval" with:
$ ./qemu-system-i386 -M xenfv
Unexpected error in object_property_find_err() at ../qom/object.c:1298:
qemu-system-i386: Property 'xenfv-3.1-machine.accel' not found
Aborted (core dumped)
The default_machine_opts handling doesn't process the legacy machine
options like "accel". Call qemu_apply_legacy_machine_options to provide
the legacy handling.
Signed-off-by: Jason Andryuk <jandryuk@gmail.com>
Message-Id: <20210713021552.19110-1-jandryuk@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Recent GLibC calls sched_getaffinity in code paths related to malloc and
when QEMU blocks access, it sends it off into a bad codepath resulting
in stack exhaustion[1]. The GLibC bug is being fixed[2], but none the
less, GLibC has valid reasons to want to use sched_getaffinity.
It is not unreasonable for code to want to run many resource syscalls
for information gathering, so it is a bit too harsh for QEMU to block
them.
[1] https://bugzilla.redhat.com/show_bug.cgi?id=1975693
[2] https://sourceware.org/pipermail/libc-alpha/2021-June/128271.html
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Acked-by: Eduardo Otubo <otubo@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
MemoryRegion names is cached on first call to memory_region_name(),
so displaying the name is trace events is cheap. Add it for read /
write ops.
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20210307074833.143106-1-f4bug@amsat.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Coverity reports that qemu_parse_config_group is returning without
unrefing the "crumpled" dictionary in case its top level item is a
list. But actually the contract with qemu_record_config_group is
the same as for qemu_parse_config_group itself: if those function
need to stash the dictionary they get, they have to take a reference
themselves (currently this is never the case for either function).
Therefore, just add an unconditional qobject_unref(crumpled) to
qemu_parse_config_group.
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add module_allow_arch() to set the target architecture.
In case a module is limited to some arch verify arches
match and ignore the module if not.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Jose R. Ziviani <jziviani@suse.de>
Message-Id: <20210624103836.2382472-19-kraxel@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Use module database to figure which module adds given QemuOpts group.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Jose R. Ziviani <jziviani@suse.de>
Message-Id: <20210624103836.2382472-17-kraxel@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add script to generate C source with a small
database containing the module meta-data.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Jose R. Ziviani <jziviani@suse.de>
Message-Id: <20210624103836.2382472-4-kraxel@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
We want to separate the two cases whereby we discard ram
- uncoordinated: e.g., virito-balloon
- coordinated: e.g., virtio-mem coordinated via the RamDiscardManager
Reviewed-by: Pankaj Gupta <pankaj.gupta@cloud.ionos.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Alex Williamson <alex.williamson@redhat.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Auger Eric <eric.auger@redhat.com>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: teawater <teawaterz@linux.alibaba.com>
Cc: Marek Kedzierski <mkedzier@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210413095531.25603-12-david@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
We have users in migration context that don't hold the BQL (when
finishing migration). To prepare for further changes, use a dedicated mutex
instead of atomic operations. Keep using qatomic_read ("READ_ONCE") for the
functions that only extract the current state (e.g., used by
virtio-balloon), locking isn't necessary.
While at it, split up the counter into two variables to make it easier
to understand.
Suggested-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta@cloud.ionos.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Alex Williamson <alex.williamson@redhat.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Auger Eric <eric.auger@redhat.com>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: teawater <teawaterz@linux.alibaba.com>
Cc: Marek Kedzierski <mkedzier@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210413095531.25603-11-david@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
In case one wants to create a permanent copy of a MemoryRegionSections,
one needs access to flatview_ref()/flatview_unref(). Instead of exposing
these, let's just add helpers to copy/free a MemoryRegionSection and
properly adjust references.
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Alex Williamson <alex.williamson@redhat.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Auger Eric <eric.auger@redhat.com>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: teawater <teawaterz@linux.alibaba.com>
Cc: Marek Kedzierski <mkedzier@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210413095531.25603-3-david@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
We have some special RAM memory regions (managed by virtio-mem), whereby
the guest agreed to only use selected memory ranges. "unused" parts are
discarded so they won't consume memory - to logically unplug these memory
ranges. Before the VM is allowed to use such logically unplugged memory
again, coordination with the hypervisor is required.
This results in "sparse" mmaps/RAMBlocks/memory regions, whereby only
coordinated parts are valid to be used/accessed by the VM.
In most cases, we don't care about that - e.g., in KVM, we simply have a
single KVM memory slot. However, in case of vfio, registering the
whole region with the kernel results in all pages getting pinned, and
therefore an unexpected high memory consumption - discarding of RAM in
that context is broken.
Let's introduce a way to coordinate discarding/populating memory within a
RAM memory region with such special consumers of RAM memory regions: they
can register as listeners and get updates on memory getting discarded and
populated. Using this machinery, vfio will be able to map only the
currently populated parts, resulting in discarded parts not getting pinned
and not consuming memory.
A RamDiscardManager has to be set for a memory region before it is getting
mapped, and cannot change while the memory region is mapped.
Note: At some point, we might want to let RAMBlock users (esp. vfio used
for nvme://) consume this interface as well. We'll need RAMBlock notifier
calls when a RAMBlock is getting mapped/unmapped (via the corresponding
memory region), so we can properly register a listener there as well.
Reviewed-by: Pankaj Gupta <pankaj.gupta@cloud.ionos.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Alex Williamson <alex.williamson@redhat.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Auger Eric <eric.auger@redhat.com>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: teawater <teawaterz@linux.alibaba.com>
Cc: Marek Kedzierski <mkedzier@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210413095531.25603-2-david@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Make -smp syntactic sugar for a compound property "-machine
smp.{cores,threads,cpu,...}". machine_smp_parse is replaced by the
setter for the property.
numa-test will now cover the new syntax, while other tests
still use -smp.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Switch from QemuOpts to keyval. This enables the introduction
of non-scalar machine properties, and JSON syntax in the future.
For JSON syntax to be supported right now, we would have to
consider what would happen if string-based dictionaries (produced by
-M key=val) were to be merged with strongly-typed dictionaries
(produced by -M {'key': 123}).
The simplest way out is to never enter the situation, and only allow one
-M option when JSON syntax is in use. However, we want options such as
-smp to become syntactic sugar for -M, and this is a problem; as soon
as -smp becomes a shortcut for -M, QEMU would forbid using -M '{....}'
together with -smp. Therefore, allowing JSON syntax right now for -M
would be a forward-compatibility nightmare and it would be impossible
anyway to introduce -M incrementally in tools.
Instead, support for JSON syntax is delayed until after the main
options are converted to QOM compound properties. These include -boot,
-acpitable, -smbios, -m, -semihosting-config, -rtc and -fw_cfg. Once JSON
syntax is introduced, these options will _also_ be forbidden together
with -M '{...}'.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
It's just a wrapper around the -display ...,window-close=off parameter,
and the name "no-quit" is rather confusing compared to "window-close"
(since there are still other means to quit the emulator), so we should
rather tell our users to use the "window-close" parameter instead.
While we're at it, update the documentation to state that
"-no-quit" is available for GTK, too, not only for SDL.
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20210630163231.467987-4-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
According to the QAPI schema, there is a "-" and not a "_" between
"window" and "close", and we're also talking about "window-close"
in the long parameter description in qemu-options.hx, so we should
make sure that we rather use the variant with the "-" by default
instead of only allowing the one with the "_" here. The old way
still stays enabled for compatibility, but we deprecate it, so that
we can switch to a QAPIfied parameter one day more easily.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20210630163231.467987-3-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The frame parameter has been removed along with the support for
SDL 1.2.
Fixes: 09bd7ba9f5 ("Remove deprecated -no-frame option")
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20210630163231.467987-2-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Libvirt's "domcapabilities" command has a way to state whether certain
graphic frontends are available in QEMU or not. Originally, libvirt
looked at the "--help" output of the QEMU binary to determine whether
SDL was available or not (by looking for the "-sdl" parameter in the
help text), but since libvirt stopped doing this analysis of the help
text, the detection of SDL is currently broken, see:
https://bugzilla.redhat.com/show_bug.cgi?id=1790902
QEMU should provide a way via the QMP interface instead. A simple way,
without introducing additional commands, is to make the DisplayType
enum entries conditional, so that the enum only contains the entries if
the corresponding CONFIG_xxx switches have been set. This of course
only gives an indication which possibilities have been enabled during
compile-time of QEMU (and does not take into account whether modules
are later available or not for example - for this we'd need a separate
command), but anyway, this should already be good enough for the above
bug ticket, and it's a good idea anyway to make the QMP interface
conditional here, so let's simply do it.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20210615090439.70926-1-thuth@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Let's support RAM_NORESERVE via MAP_NORESERVE on Linux. The flag has no
effect on most shared mappings - except for hugetlbfs and anonymous memory.
Linux man page:
"MAP_NORESERVE: Do not reserve swap space for this mapping. When swap
space is reserved, one has the guarantee that it is possible to modify
the mapping. When swap space is not reserved one might get SIGSEGV
upon a write if no physical memory is available. See also the discussion
of the file /proc/sys/vm/overcommit_memory in proc(5). In kernels before
2.6, this flag had effect only for private writable mappings."
Note that the "guarantee" part is wrong with memory overcommit in Linux.
Also, in Linux hugetlbfs is treated differently - we configure reservation
of huge pages from the pool, not reservation of swap space (huge pages
cannot be swapped).
The rough behavior is [1]:
a) !Hugetlbfs:
1) Without MAP_NORESERVE *or* with memory overcommit under Linux
disabled ("/proc/sys/vm/overcommit_memory == 2"), the following
accounting/reservation happens:
For a file backed map
SHARED or READ-only - 0 cost (the file is the map not swap)
PRIVATE WRITABLE - size of mapping per instance
For an anonymous or /dev/zero map
SHARED - size of mapping
PRIVATE READ-only - 0 cost (but of little use)
PRIVATE WRITABLE - size of mapping per instance
2) With MAP_NORESERVE, no accounting/reservation happens.
b) Hugetlbfs:
1) Without MAP_NORESERVE, huge pages are reserved.
2) With MAP_NORESERVE, no huge pages are reserved.
Note: With "/proc/sys/vm/overcommit_memory == 0", we were already able
to configure it for !hugetlbfs globally; this toggle now allows
configuring it more fine-grained, not for the whole system.
The target use case is virtio-mem, which dynamically exposes memory
inside a large, sparse memory area to the VM.
[1] https://www.kernel.org/doc/Documentation/vm/overcommit-accounting
Reviewed-by: Peter Xu <peterx@redhat.com>
Acked-by: Eduardo Habkost <ehabkost@redhat.com> for memory backend and machine core
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210510114328.21835-10-david@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Let's introduce RAM_NORESERVE, allowing mmap'ing with MAP_NORESERVE. The
new flag has the following semantics:
"
RAM is mmap-ed with MAP_NORESERVE. When set, reserving swap space (or huge
pages if applicable) is skipped: will bail out if not supported. When not
set, the OS will do the reservation, if supported for the memory type.
"
Allow passing it into:
- memory_region_init_ram_nomigrate()
- memory_region_init_resizeable_ram()
- memory_region_init_ram_from_file()
... and teach qemu_ram_mmap() and qemu_anon_ram_alloc() about the flag.
Bail out if the flag is not supported, which is the case right now for
both, POSIX and win32. We will add Linux support next and allow specifying
RAM_NORESERVE via memory backends.
The target use case is virtio-mem, which dynamically exposes memory
inside a large, sparse memory area to the VM.
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Acked-by: Eduardo Habkost <ehabkost@redhat.com> for memory backend and machine core
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210510114328.21835-9-david@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Let's pass flags instead of bools to prepare for passing other flags and
update the documentation of qemu_ram_mmap(). Introduce new QEMU_MAP_
flags that abstract the mmap() PROT_ and MAP_ flag handling and simplify
it.
We expose only flags that are currently supported by qemu_ram_mmap().
Maybe, we'll see qemu_mmap() in the future as well that can implement these
flags.
Note: We don't use MAP_ flags as some flags (e.g., MAP_SYNC) are only
defined for some systems and we want to always be able to identify
these flags reliably inside qemu_ram_mmap() -- for example, to properly
warn when some future flags are not available or effective on a system.
Also, this way we can simplify PROT_ handling as well.
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Acked-by: Eduardo Habkost <ehabkost@redhat.com> for memory backend and machine core
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210510114328.21835-8-david@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Let's pass ram_flags to qemu_ram_alloc() and qemu_ram_alloc_internal(),
preparing for passing additional flags.
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Eduardo Habkost <ehabkost@redhat.com> for memory backend and machine core
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210510114328.21835-7-david@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Let's pass in ram flags just like we do with qemu_ram_alloc_from_file(),
to clean up and prepare for more flags.
Simplify the documentation of passed ram flags: Looking at our
documentation of RAM_SHARED and RAM_PMEM is sufficient, no need to be
repetitive.
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Acked-by: Eduardo Habkost <ehabkost@redhat.com> for memory backend and machine core
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210510114328.21835-5-david@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
RAM_SHARED now also properly indicates shared anonymous memory. Let's check
that flag for anonymous memory as well, to restore the proper mapping.
Fixes: 06329ccecf ("mem: add share parameter to memory-backend-ram")
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210406080126.24010-4-david@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
We can create shared anonymous memory via
"-object memory-backend-ram,share=on,..."
which is, for example, required by PVRDMA for mremap() to work.
Shared anonymous memory is weird, though. Instead of MADV_DONTNEED, we
have to use MADV_REMOVE: MADV_DONTNEED will only remove / zap all
relevant page table entries of the current process, the backend storage
will not get removed, resulting in no reduced memory consumption and
a repopulation of previous content on next access.
Shared anonymous memory is internally really just shmem, but without a
fd exposed. As we cannot use fallocate() without the fd to discard the
backing storage, MADV_REMOVE gets the same job done without a fd as
documented in "man 2 madvise". Removing backing storage implicitly
invalidates all page table entries with relevant mappings - an additional
MADV_DONTNEED is not required.
Fixes: 06329ccecf ("mem: add share parameter to memory-backend-ram")
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210406080126.24010-3-david@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Let's drop the "shared" parameter from ram_block_add() and properly
store it in the flags of the ram block instead, such that
qemu_ram_is_shared() properly succeeds on all ram blocks that were mapped
MAP_SHARED.
We'll use this information next to fix some cases with shared anonymous
memory.
Reviewed-by: Igor Kotrasinski <i.kotrasinsk@partner.samsung.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210406080126.24010-2-david@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Based on the description of error_setg(), the local variable err in
qemu_init_subsystems() should be initialized to NULL.
Fixes: efd7ab22fb ("vl: extract qemu_init_subsystems")
Cc: qemu-stable@nongnu.org
Signed-off-by: Peng Liang <liangpeng10@huawei.com>
Message-Id: <20210610131729.3906565-1-liangpeng10@huawei.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Based on the description of error_setg(), the local variable err in
qemu_maybe_daemonize() should be initialized to NULL.
Without fix, the uninitialized *errp triggers assert failure which
doesn't show much valuable information.
Before the fix:
qemu-system-x86_64: ../util/error.c:59: error_setv: Assertion `*errp == NULL' failed.
After fix:
qemu-system-x86_64: cannot create PID file: Cannot open pid file: Permission denied
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Message-Id: <20210610084741.456260-1-zhenzhong.duan@intel.com>
Cc: qemu-stable@nongnu.org
Fixes: 0546c0609c ("vl: split various early command line options to a separate function", 2020-12-10)
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The VM stop process has to flush outstanding I/O and this is a critical
failure scenario that is hard to diagnose. Add a probe point that
records the flush return code.
Reviewed-by: Connor Kuehl <ckuehl@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
