Rename qemu_ram_writeback() as qemu_ram_msync() to better
match what it does.
Suggested-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20200508062456.23344-5-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20200508062456.23344-3-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
We usually use '_do_' for internal functions. Rename
memory_region_do_writeback() as memory_region_writeback().
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20200508062456.23344-2-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The only way object_property_add() can fail is when a property with
the same name already exists. Since our property names are all
hardcoded, failure is a programming error, and the appropriate way to
handle it is passing &error_abort.
Same for its variants, except for object_property_add_child(), which
additionally fails when the child already has a parent. Parentage is
also under program control, so this is a programming error, too.
We have a bit over 500 callers. Almost half of them pass
&error_abort, slightly fewer ignore errors, one test case handles
errors, and the remaining few callers pass them to their own callers.
The previous few commits demonstrated once again that ignoring
programming errors is a bad idea.
Of the few ones that pass on errors, several violate the Error API.
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. ich9_pm_add_properties(), sparc32_ledma_realize(),
sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize()
are wrong that way.
When the one appropriate choice of argument is &error_abort, letting
users pick the argument is a bad idea.
Drop parameter @errp and assert the preconditions instead.
There's one exception to "duplicate property name is a programming
error": the way object_property_add() implements the magic (and
undocumented) "automatic arrayification". Don't drop @errp there.
Instead, rename object_property_add() to object_property_try_add(),
and add the obvious wrapper object_property_add().
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20200505152926.18877-15-armbru@redhat.com>
[Two semantic rebase conflicts resolved]
Uses of gchar * in qom/object.h:
* ObjectProperty member @name
Functions that take a property name argument all use char *. Change
the member to match.
* ObjectProperty member @type
Functions that take a property type argument or return it all use
char *. Change the member to match.
* ObjectProperty member @description
Functions that take a property description argument all use char *.
Change the member to match.
* object_resolve_path_component() parameter @part
Path components are property names. Most callers pass char *
arguments. Change the parameter to match. Adjust the few callers
that pass gchar * to pass char *.
* Return value of object_get_canonical_path_component(),
object_get_canonical_path()
Most callers convert their return values right back to char *.
Change the return value to match. Adjust the few callers where that
would add a conversion to gchar * to use char * instead.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20200505152926.18877-3-armbru@redhat.com>
memory_region_init_rom_nomigrate() has the same content than
memory_region_init_ram_shared_nomigrate(), with setting the
readonly mode. The code is easier to review as creating a
readonly ram/shared/nomigrate region.
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Since memory region aliases are neither rom nor ram, they are
described as i/o, which is often incorrect. Return instead the
type of the original region we are aliasing.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Several objects implemented their own uint property getters and setters,
despite them being straightforward (without any checks/validations on
the values themselves) and identical across objects. This makes use of
an enhanced API for object_property_add_uintXX_ptr() which offers
default setters.
Some of these setters used to update the value even if the type visit
failed (eg. because the value being set overflowed over the given type).
The new setter introduces a check for these errors, not updating the
value if an error occurred. The error is propagated.
Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reallocing the ioeventfds[] array each time an element is added is very
expensive as the number of ioeventfds increases. Batch allocate instead
to amortize the cost of realloc.
This patch reduces Linux guest boot times from 362s to 140s when there
are 2 virtio-blk devices with 1 virtqueue and 99 virtio-blk devices with
32 virtqueues.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200218182226.913977-1-stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The preferred way to test whether a trace event is enabled is to
use trace_event_get_state_backends(), because this will give the
correct answer (allowing expensive computations to be skipped)
whether the trace event is compile-time or run-time disabled.
Convert the four old-style direct uses of TRACE_FOO_ENABLED in
memory.c.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20200120151142.18954-3-peter.maydell@linaro.org
Message-Id: <20200120151142.18954-3-peter.maydell@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
We actually want to access the accelerator, not the machine, so
use the current_accel() wrapper instead.
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200121110349.25842-10-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
"info mtree -f" prints the wrong accelerator name if used with for example
"-machine accel=kvm:tcg". The right thing to do is to fetch the name
from the AccelClass, which will also work nicely once
current_machine->accel stops existing.
Tested-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add an option to trigger memory writeback to sync given memory region
with the corresponding backing store, case one is available.
This extends the support for persistent memory, allowing syncing on-demand.
Signed-off-by: Beata Michalska <beata.michalska@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20191121000843.24844-3-beata.michalska@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
All targets have now migrated away from the old unassigned_access
hook to the new do_transaction_failed hook. This means we can remove
the core-code infrastructure for that hook and the code that calls it.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20191108173732.11816-1-peter.maydell@linaro.org
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20191007143642.301445-6-dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Currently, when a notifier is attempted to be registered and its
flags are not supported (especially the MAP one) by the IOMMU MR,
we generally abruptly exit in the IOMMU code. The failure could be
handled more nicely in the caller and especially in the VFIO code.
So let's allow memory_region_register_iommu_notifier() to fail as
well as notify_flag_changed() callback.
All sites implementing the callback are updated. This patch does
not yet remove the exit(1) in the amd_iommu code.
in SMMUv3 we turn the warning message into an error message saying
that the assigned device would not work properly.
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Pages that we want to track for NOTDIRTY are RAM. We do not
really need to go through the I/O path to handle them.
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The memory_region_tb_read tracepoint is unreachable, since notdirty
is supposed to apply only to writes. The memory_region_tb_write
tracepoint is mis-named, because notdirty is not only used for TB
invalidation. It is also used for e.g. VGA RAM updates and migration.
Replace memory_region_tb_write with memory_notdirty_write_access,
and place it in memory_notdirty_write_prepare where it can catch
all of the instances. Add memory_notdirty_set_dirty to log when
we no longer intercept writes to a page.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
devend_memop can rely on the fact that the result is always either
0 or MO_BSWAP, corresponding respectively to host endianness and
the opposite. Native (target) endianness in turn can be either
the host endianness, in which case MO_BSWAP is only returned for
host-opposite endianness, or the opposite, in which case 0 is only
returned for host endianness.
With this in mind, devend_memop can be compiled as a setcond+shift
for every target. Do this and, while at it, move it to
include/exec/memory.h since !NEED_CPU_H files do not (and should not)
need it.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reorganize watchpoints out of i/o path.
Return host address from probe_write / probe_access.
-----BEGIN PGP SIGNATURE-----
iQFRBAABCgA7FiEEekgeeIaLTbaoWgXAZN846K9+IV8FAl1uiyYdHHJpY2hhcmQu
aGVuZGVyc29uQGxpbmFyby5vcmcACgkQZN846K9+IV8AuwgAnYLQQbL8kjSqzp7q
gRlj0M2SX41ZW3fMkI794RwsljD9Z0QS7YGnpzHolig9XUYrGnip7STrMvlCr/1L
CIMWNHlgitgBMszLqg42/TB+6RxXn+DMX/ShUzTagC6xQhinCIpdEjoLaTKSgeP+
foIyJ2uoJLKOBP8cPTQp8evongtoQIljpsZZ0K8a4sreO1d6ytH+olkuoGiROft+
VoJkA+kNHd9cE+LPCva8UFGu1QE6uCySvhepzOpnvOtK+SXKUm2yLOFGu7RWP1pT
RkE0oRyRnImtg+cViHfUUFogIffFROdL5tuYMQVuqbINeROPUgJPav+R1Nz1P60a
xM2HEw==
=bLLU
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/rth/tags/pull-tcg-20190903' into staging
Allow page table bit to swap endianness.
Reorganize watchpoints out of i/o path.
Return host address from probe_write / probe_access.
# gpg: Signature made Tue 03 Sep 2019 16:47:50 BST
# gpg: using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg: issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [full]
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A 05C0 64DF 38E8 AF7E 215F
* remotes/rth/tags/pull-tcg-20190903: (36 commits)
tcg: Factor out probe_write() logic into probe_access()
tcg: Make probe_write() return a pointer to the host page
s390x/tcg: Pass a size to probe_write() in do_csst()
hppa/tcg: Call probe_write() also for CONFIG_USER_ONLY
mips/tcg: Call probe_write() for CONFIG_USER_ONLY as well
tcg: Enforce single page access in probe_write()
tcg: Factor out CONFIG_USER_ONLY probe_write() from s390x code
s390x/tcg: Fix length calculation in probe_write_access()
s390x/tcg: Use guest_addr_valid() instead of h2g_valid() in probe_write_access()
tcg: Check for watchpoints in probe_write()
cputlb: Handle watchpoints via TLB_WATCHPOINT
cputlb: Remove double-alignment in store_helper
cputlb: Fix size operand for tlb_fill on unaligned store
exec: Factor out cpu_watchpoint_address_matches
cputlb: Fold TLB_RECHECK into TLB_INVALID_MASK
exec: Factor out core logic of check_watchpoint()
exec: Move user-only watchpoint stubs inline
target/sparc: sun4u Invert Endian TTE bit
target/sparc: Add TLB entry with attributes
cputlb: Byte swap memory transaction attribute
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Now that MemOp has been pushed down into the memory API, and
callers are encoding endianness, we can collapse byte swaps
along the I/O path into the accelerator and target independent
adjust_endianness.
Collapsing byte swaps along the I/O path enables additional endian
inversion logic, e.g. SPARC64 Invert Endian TTE bit, with redundant
byte swaps cancelling out.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Suggested-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Tony Nguyen <tony.nguyen@bt.com>
Message-Id: <911ff31af11922a9afba9b7ce128af8b8b80f316.1566466906.git.tony.nguyen@bt.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Preparation for collapsing the two byte swaps adjust_endianness and
handle_bswap into the former.
Call memory_region_dispatch_{read|write} with endianness encoded into
the "MemOp op" operand.
This patch does not change any behaviour as
memory_region_dispatch_{read|write} is yet to handle the endianness.
Once it does handle endianness, callers with byte swaps can collapse
them into adjust_endianness.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Tony Nguyen <tony.nguyen@bt.com>
Message-Id: <8066ab3eb037c0388dfadfe53c5118429dd1de3a.1566466906.git.tony.nguyen@bt.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Convert memory_region_dispatch_{read|write} operand "unsigned size"
into a "MemOp op".
Signed-off-by: Tony Nguyen <tony.nguyen@bt.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <1dd82df5801866743f838f1d046475115a1d32da.1566466906.git.tony.nguyen@bt.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
memory_region_iommu_replay_all is not used. Remove it.
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Message-id: 20190822172350.12008-2-eric.auger@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The old memory_region_{add|clear}_coalescing() has some defects
because they both changed mr->coalesced before updating the regions
using memory_region_update_coalesced_range_as(). Then when the
regions were updated in memory_region_update_coalesced_range_as() the
mr->coalesced will always be either one more or one less. So:
- For memory_region_add_coalescing: it'll always trying to remove the
newly added coalesced region while it shouldn't, and,
- For memory_region_clear_coalescing: when it calls the update there
will be no coalesced ranges on mr->coalesced because they were all
removed before hand so the update will probably do nothing for real.
Let's fix this. Now we've got flat_range_coalesced_io_notify() to
notify a single CoalescedMemoryRange instance change, so use it in the
existing memory_region_update_coalesced_range() logic by only notify
either an addition or deletion. Then we hammer both the
memory_region_{add|clear}_coalescing() to use it.
Fixes: 3ac7d43a6f
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20190820141328.10009-5-peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The has_coalesced_range could potentially be problematic in that it
only works for additions of coalesced mmio ranges but not deletions.
The reason is that has_coalesced_range information can be lost when
the FlatView updates the topology again when the updated region is not
covering the coalesced regions. When that happens, due to
flatrange_equal() is not checking against has_coalesced_range, the new
FlatRange will be seen as the same one as the old and the new
instance (whose has_coalesced_range will be zero) will replace the old
instance (whose has_coalesced_range _could_ be non-zero).
The counter was originally used to make sure every FlatRange will only
notify once for coalesced_io_{add|del} memory listeners, because each
FlatRange can be used by multiple address spaces, so logically
speaking it could be called multiple times. However we should not
limit that, because memory listeners should will only be registered
with specific address space rather than multiple address spaces.
So let's fix this up by simply removing the whole has_coalesced_range.
Fixes: 3ac7d43a6f
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20190820141328.10009-3-peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
It is a workaround of current KVM's KVM_UNREGISTER_COALESCED_MMIO
interface. The kernel interface only allows to unregister an mmio
device with exactly the zone size when registered, or any smaller zone
that is included in the device mmio zone. It does not support the
userspace to specify a very large zone to remove all the small mmio
devices within the zone covered.
Logically speaking it would be nicer to fix this from KVM side, though
in all cases we still need to coop with old kernels so let's do this.
Fixes: 3ac7d43a6f
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20190820141328.10009-2-peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Removing the update variable and quit earlier if the memory region has
no coalesced range. This prepares for the next patch.
Fixes: 3ac7d43a6f
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20190820141328.10009-4-peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
There is a race between TCG and accesses to the dirty log:
vCPU thread reader thread
----------------------- -----------------------
TLB check -> slow path
notdirty_mem_write
write to RAM
set dirty flag
clear dirty flag
TLB check -> fast path
read memory
write to RAM
Fortunately, in order to fix it, no change is required to the
vCPU thread. However, the reader thread must delay the read after
the vCPU thread has finished the write. This can be approximated
conservatively by run_on_cpu, which waits for the end of the current
translation block.
A similar technique is used by KVM, which has to do a synchronous TLB
flush after doing a test-and-clear of the dirty-page flags.
Reported-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
It is wrong for an entry to have parts out of scope of notifier's range.
assert this condition.
Out of scope mapping/unmapping would cause problem, as in below case:
1. initially there are two notifiers with ranges
0-0xfedfffff, 0xfef00000-0xffffffffffffffff,
IOVAs from 0x3c000000 - 0x3c1fffff is in shadow page table.
2. in vfio, memory_region_register_iommu_notifier() is followed by
memory_region_iommu_replay(), which will first call address space
unmap,
and walk and add back all entries in vtd shadow page table. e.g.
(1) for notifier 0-0xfedfffff,
IOVAs from 0 - 0xffffffff get unmapped,
and IOVAs from 0x3c000000 - 0x3c1fffff get mapped
(2) for notifier 0xfef00000-0xffffffffffffffff
IOVAs from 0 - 0x7fffffffff get unmapped,
but IOVAs from 0x3c000000 - 0x3c1fffff cannot get mapped back.
Cc: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Message-Id: <1561432878-13754-1-git-send-email-yan.y.zhao@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
sysemu/sysemu.h is a rather unfocused dumping ground for stuff related
to the system-emulator. Evidence:
* It's included widely: in my "build everything" tree, changing
sysemu/sysemu.h still triggers a recompile of some 1100 out of 6600
objects (not counting tests and objects that don't depend on
qemu/osdep.h, down from 5400 due to the previous two commits).
* It pulls in more than a dozen additional headers.
Split stuff related to run state management into its own header
sysemu/runstate.h.
Touching sysemu/sysemu.h now recompiles some 850 objects. qemu/uuid.h
also drops from 1100 to 850, and qapi/qapi-types-run-state.h from 4400
to 4200. Touching new sysemu/runstate.h recompiles some 500 objects.
Since I'm touching MAINTAINERS to add sysemu/runstate.h anyway, also
add qemu/main-loop.h.
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20190812052359.30071-30-armbru@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
[Unbreak OS-X build]
In my "build everything" tree, changing hw/qdev-properties.h triggers
a recompile of some 2700 out of 6600 objects (not counting tests and
objects that don't depend on qemu/osdep.h).
Many places including hw/qdev-properties.h (directly or via hw/qdev.h)
actually need only hw/qdev-core.h. Include hw/qdev-core.h there
instead.
hw/qdev.h is actually pointless: all it does is include hw/qdev-core.h
and hw/qdev-properties.h, which in turn includes hw/qdev-core.h.
Replace the remaining uses of hw/qdev.h by hw/qdev-properties.h.
While there, delete a few superfluous inclusions of hw/qdev-core.h.
Touching hw/qdev-properties.h now recompiles some 1200 objects.
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: "Daniel P. Berrangé" <berrange@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20190812052359.30071-22-armbru@redhat.com>
In my "build everything" tree, changing qemu/main-loop.h triggers a
recompile of some 5600 out of 6600 objects (not counting tests and
objects that don't depend on qemu/osdep.h). It includes block/aio.h,
which in turn includes qemu/event_notifier.h, qemu/notify.h,
qemu/processor.h, qemu/qsp.h, qemu/queue.h, qemu/thread-posix.h,
qemu/thread.h, qemu/timer.h, and a few more.
Include qemu/main-loop.h only where it's needed. Touching it now
recompiles only some 1700 objects. For block/aio.h and
qemu/event_notifier.h, these numbers drop from 5600 to 2800. For the
others, they shrink only slightly.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20190812052359.30071-21-armbru@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
TYPE_IOMMU_MEMORY_REGION is a direct subtype of TYPE_MEMORY_REGION.
Its instance struct is IOMMUMemoryRegion, and its first member is a
MemoryRegion. Correct. Its class struct is IOMMUMemoryRegionClass,
and its first member is a DeviceClass. Wrong. Messed up when commit
1221a47467 introduced the QOM type. It even included hw/qdev-core.h
just for that.
TYPE_MEMORY_REGION doesn't bother to define a class struct. This is
fine, it simply defaults to its super-type TYPE_OBJECT's class struct
ObjectClass. Changing IOMMUMemoryRegionClass's first member's type to
ObjectClass would be a minimal fix, if a bit brittle: if
TYPE_MEMORY_REGION ever acquired own class struct, we'd have to update
IOMMUMemoryRegionClass to use it.
Fix it the clean and robust way instead: give TYPE_MEMORY_REGION its
own class struct MemoryRegionClass now, and use it for
IOMMUMemoryRegionClass's first member.
Revert the include of hw/qdev-core.h, and fix the few files that have
come to rely on it.
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20190812052359.30071-5-armbru@redhat.com>
This adds an accelerator name to the "into mtree -f" to tell the user if
a particular memory section is registered with the accelerator;
the primary user for this is KVM and such information is useful
for debugging purposes.
This adds a has_memory() callback to the accelerator class allowing any
accelerator to have a label in that memory tree dump.
Since memory sections are passed to memory listeners and get registered
in accelerators (rather than memory regions), this only prints new labels
for flatviews attached to the system address space.
An example:
Root memory region: system
0000000000000000-0000002fffffffff (prio 0, ram): /objects/mem0 kvm
0000003000000000-0000005fffffffff (prio 0, ram): /objects/mem1 kvm
0000200000000020-000020000000003f (prio 1, i/o): virtio-pci
0000200080000000-000020008000003f (prio 0, i/o): capabilities
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Message-Id: <20190614015237.82463-1-aik@ozlabs.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Introduce a new memory region listener hook log_clear() to allow the
listeners to hook onto the points where the dirty bitmap is cleared by
the bitmap users.
Previously log_sync() contains two operations:
- dirty bitmap collection, and,
- dirty bitmap clear on remote site.
Let's take KVM as example - log_sync() for KVM will first copy the
kernel dirty bitmap to userspace, and at the same time we'll clear the
dirty bitmap there along with re-protecting all the guest pages again.
We add this new log_clear() interface only to split the old log_sync()
into two separated procedures:
- use log_sync() to collect the collection only, and,
- use log_clear() to clear the remote dirty bitmap.
With the new interface, the memory listener users will still be able
to decide how to implement the log synchronization procedure, e.g.,
they can still only provide log_sync() method only and put all the two
procedures within log_sync() (that's how the old KVM works before
KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is introduced). However with this
new interface the memory listener users will start to have a chance to
postpone the log clear operation explicitly if the module supports.
That can really benefit users like KVM at least for host kernels that
support KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2.
There are three places that can clear dirty bits in any one of the
dirty bitmap in the ram_list.dirty_memory[3] array:
cpu_physical_memory_snapshot_and_clear_dirty
cpu_physical_memory_test_and_clear_dirty
cpu_physical_memory_sync_dirty_bitmap
Currently we hook directly into each of the functions to notify about
the log_clear().
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20190603065056.25211-7-peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Also we change the 2nd parameter of it to be the relative offset
within the memory region. This is to be used in follow up patches.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20190603065056.25211-6-peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Similar to 9460dee4b2 ("memory: do not touch code dirty bitmap unless
TCG is enabled", 2015-06-05) but for the migration bitmap - we can
skip the MIGRATION bitmap update if migration not enabled.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20190603065056.25211-4-peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
The memory region reference is increased when insert a range
into flatview range array, then decreased by destroy flatview.
If some flat range merged by flatview_simplify, the memory region
reference can not be decreased by destroy flatview any more.
In this case, start virtual machine by the command line:
qemu-system-x86_64
-name guest=ubuntu,debug-threads=on
-machine pc,accel=kvm,usb=off,dump-guest-core=off
-cpu host
-m 16384
-realtime mlock=off
-smp 8,sockets=2,cores=4,threads=1
-object memory-backend-file,id=ram-node0,prealloc=yes,mem-path=/dev/hugepages,share=yes,size=8589934592
-numa node,nodeid=0,cpus=0-3,memdev=ram-node0
-object memory-backend-file,id=ram-node1,prealloc=yes,mem-path=/dev/hugepages,share=yes,size=8589934592
-numa node,nodeid=1,cpus=4-7,memdev=ram-node1
-no-user-config
-nodefaults
-rtc base=utc
-no-shutdown
-boot strict=on
-device piix3-usb-uhci,id=usb,bus=pci.0,addr=0x1.0x2
-device virtio-scsi-pci,id=scsi0,bus=pci.0,addr=0x2
-device virtio-serial-pci,id=virtio-serial0,bus=pci.0,addr=0x3
-drive file=ubuntu.qcow2,format=qcow2,if=none,id=drive-virtio-disk0,cache=none,aio=native
-device virtio-blk-pci,scsi=off,bus=pci.0,addr=0x4,drive=drive-virtio-disk0,id=virtio-disk0,bootindex=1
-chardev pty,id=charserial0
-device isa-serial,chardev=charserial0,id=serial0
-device usb-tablet,id=input0,bus=usb.0,port=1
-vnc 0.0.0.0:0
-device VGA,id=video0,vgamem_mb=16,bus=pci.0,addr=0x5
-device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x6
-msg timestamp=on
And run the script in guest OS:
while true
do
setpci -s 00:06.0 04.b=03
setpci -s 00:06.0 04.b=07
done
I found the reference of node0 HostMemoryBackendFile is a big one.
(gdb) p numa_info[0]->node_memdev->parent.ref
$6 = 1636278
(gdb)
Signed-off-by: King Wang<king.wang@huawei.com>
Message-Id: <20190712065241.11784-1-king.wang@huawei.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Hot-unplugging a PHB with a VFIO device connected to it crashes QEMU:
-device spapr-pci-host-bridge,index=1,id=phb1 \
-device vfio-pci,host=0034:01:00.3,id=vfio0
(qemu) device_del phb1
[ 357.207183] iommu: Removing device 0001:00:00.0 from group 1
[ 360.375523] rpadlpar_io: slot PHB 1 removed
qemu-system-ppc64: memory.c:2742:
do_address_space_destroy: Assertion `QTAILQ_EMPTY(&as->listeners)' failed.
'as' is the IOMMU address space, which indeed has a listener registered
to by vfio_connect_container() when the VFIO device is realized. This
listener is supposed to be unregistered by vfio_disconnect_container()
when the VFIO device is finalized. Unfortunately, the VFIO device hasn't
reached finalize yet at the time the PHB unrealize function is called,
and address_space_destroy() gets called with the VFIO listener still
being registered.
All regions have just been unmapped from the address space. Listeners
aren't needed anymore at this point. Remove them before destroying the
address space.
The VFIO code will try to remove them _again_ at device finalize,
but it is okay since memory_listener_unregister() is idempotent.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <156110925375.92514.11649846071216864570.stgit@bahia.lan>
Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
[dwg: Correct spelling error pointed out by aik]
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Other accelerators have their own headers: sysemu/hax.h, sysemu/hvf.h,
sysemu/kvm.h, sysemu/whpx.h. Only tcg_enabled() & friends sit in
qemu-common.h. This necessitates inclusion of qemu-common.h into
headers, which is against the rules spelled out in qemu-common.h's
file comment.
Move tcg_enabled() & friends into their own header sysemu/tcg.h, and
adjust #include directives.
Cc: Richard Henderson <rth@twiddle.net>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20190523143508.25387-2-armbru@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
[Rebased with conflicts resolved automatically, except for
accel/tcg/tcg-all.c]
It's never used anywhere.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20190520030839.6795-5-peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The dirty bit is DIRTY_MEMORY_MIGRATION. Correct the comment.
Signed-off-by: Wei Yang <richardw.yang@linux.intel.com>
Message-Id: <20190426020927.25470-1-richardw.yang@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
mtree_info() takes an fprintf()-like callback and a FILE * to pass to
it, and so do its helper functions. Passing around callback and
argument is rather tiresome.
Its only caller hmp_info_mtree() passes monitor_printf() cast to
fprintf_function and the current monitor cast to FILE *.
The type-punning is technically undefined behaviour, but works in
practice. Clean up: drop the callback, and call qemu_printf()
instead.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20190417191805.28198-9-armbru@redhat.com>
Currently, a callback registered through the RAMBlock notifier
is not able to get the memory region type (i.e callback is not
able to use memory_region_is_ram_device function). This is
because mr->ram assignment happens _after_ the memory is allocated
whereas the callback is executed during allocation.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1667249
Suggested-by: Alex Williamson <alex.williamson@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Brijesh Singh <brijesh.singh@amd.com>
Message-Id: <20190204222322.26766-2-brijesh.singh@amd.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Do not add/del coalesced IO ranges in the case where the
same FlatRanges are present in both old and new FlatViews
Fixes: 3ac7d43a6f ("memory: update coalesced_range on transaction_commit")
Signed-off-by: Jagannathan Raman <jag.raman@oracle.com>
Message-Id: <59572a7353830be4b7aa57d79ccb7ad6b72f0dda.1549406119.git.jag.raman@oracle.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Most list head structs need not be given a name. In most cases the
name is given just in case one is going to use QTAILQ_LAST, QTAILQ_PREV
or reverse iteration, but this does not apply to lists of other kinds,
and even for QTAILQ in practice this is only rarely needed. In addition,
we will soon reimplement those macros completely so that they do not
need a name for the head struct. So clean up everything, not giving a
name except in the rare case where it is necessary.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The e1000 driver calls memory_region_add_coalescing but
kvm_coalesce_mmio_region is never called for those regions. The bug
dates back to the introduction of the memory region API; to fix it,
delete and re-add coalesced MMIO ranges when building the FlatViews.
Because coalesced MMIO regions apply to all address spaces, the
has_coalesced_range flag has to be changed into an int.
Fixes: 093bc2cd88 ("Hierarchical memory region API")
Reported-by: Atsushi Nemoto <atsushi.nemoto@sord.co.jp>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
