spapr: nested: Extend nested_ppc_state for nested PAPR API
Currently, nested_ppc_state stores a certain set of registers and works
with nested_[load|save]_state() for state transfer as reqd for nested-hv API.
Extending these with additional registers state as reqd for nested PAPR API.
Acked-by: Nicholas Piggin <npiggin@gmail.com> Suggested-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Introduce the nested PAPR hcall H_GUEST_CREATE_VCPU which is used to
create and initialize the specified VCPU resource for the previously
created guest. Each guest can have multiple VCPUs upto max 2048.
All VCPUs for a guest gets deallocated on guest delete.
Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Neuling <mikey@neuling.org> Signed-off-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Introduce the nested PAPR hcalls:
- H_GUEST_CREATE which is used to create and allocate resources for
nested guest being created.
- H_GUEST_DELETE which is used to delete and deallocate resources
for the nested guest being deleted. It also supports deleting all nested
guests at once using a deleteAll flag.
Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Neuling <mikey@neuling.org> Signed-off-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Introduce the nested PAPR hcalls:
- H_GUEST_GET_CAPABILITIES which is used to query the capabilities
of the API and the L2 guests it provides.
- H_GUEST_SET_CAPABILITIES which is used to set the Guest API
capabilities that the Host Partition supports and may use.
[amachhiw: support for p9 compat mode and return register bug fixes]
Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Neuling <mikey@neuling.org> Signed-off-by: Amit Machhiwal <amachhiw@linux.vnet.ibm.com> Signed-off-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Adding initial documentation about Nested PAPR API to describe the set
of APIs and its usage. Also talks about the Guest State Buffer elements
and it's format which is used between L0/L1 to communicate L2 state.
Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Neuling <mikey@neuling.org> Signed-off-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
spapr: nested: keep nested-hv related code restricted to its API.
spapr_exit_nested and spapr_get_pate_nested_hv contains code which
is specific to nested-hv API. Isolating code flows based on API
helps extending it to be used with different API as well.
Suggested-by: Nicholas Piggin <npiggin@gmail.com> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
spapr: nested: Introduce SpaprMachineStateNested to store related info.
Currently, nested_ptcr is being used by existing nested-hv API to store
nested guest related info. This need to be organised to extend support
for the nested PAPR API which would need to store additional info
related to nested guests in next series of patches.
Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Neuling <mikey@neuling.org> Signed-off-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
spapr: nested: move nested part of spapr_get_pate into spapr_nested.c
Most of the nested code has already been moved to spapr_nested.c
This logic inside spapr_get_pate is related to nested guests and
better suited for spapr_nested.c, hence moving there.
Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
spapr: nested: register nested-hv api hcalls only for cap-nested-hv
Since cap-nested-hv is an optional capability, it makes sense to register
api specfic hcalls only when respective capability is enabled. This
requires to introduce a new API to unregister hypercalls to maintain
sanity across guest reboot since caps are re-applied across reboots and
re-registeration of hypercalls would hit assert otherwise.
Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
These wrappers call out to handle POWER7 and newer in separate
functions but reduce to the generic case when TARGET_PPC64 is not
defined. It is easy enough to include the switch in the beginning of
the generic functions to branch out to the specific functions and get
rid of these wrappers. This avoids one indirection and entirely
compiles out the switch without TARGET_PPC64.
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
BALATON Zoltan [Tue, 27 Feb 2024 15:08:03 +0000 (16:08 +0100)]
target/ppc: Add gen_exception_err_nip() function
Add gen_exception_err_nip() that does the same as gen_exception_err()
but takes the nip as a parameter to allow specifying it instead of
using the current instruction address then change gen_exception_err()
to use it.
The gen_exception() and gen_exception_nip() functions are similar so
remove code duplication from those too while at it.
Suggested-by: Nicholas Piggin <npiggin@gmail.com> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
BALATON Zoltan [Tue, 27 Feb 2024 15:08:02 +0000 (16:08 +0100)]
target/ppc: Readability improvements in exception handlers
Improve readability by shortening some long comments, removing
comments that state the obvious and dropping some empty lines so they
don't distract when reading the code.
Acked-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
BALATON Zoltan [Tue, 27 Feb 2024 15:08:01 +0000 (16:08 +0100)]
target/ppc: Use env_cpu for cpu_abort in excp_helper
Use the env_cpu function to get the CPUState for cpu_abort. These are
only needed in case of fatal errors so this allows to avoid casting
and storing CPUState in a local variable wnen not needed.
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
This is becuase, no read/write call-backs are registered for
these SPRs. Add support to register generic read/write
functions to these power10 pmu sprs to fix it.
Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Madhavan Srinivasan <maddy@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
This patch introduces XO form instructions into decode tree
specification, for which all the four variations([o][.]) have been
handled with a single pattern. The changes were verified by validating
that the tcg ops generated by those instructions remain the same, which
were captured with the '-d in_asm,op' flag.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Chinmay Rath <rathc@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
BALATON Zoltan [Tue, 20 Feb 2024 23:22:00 +0000 (00:22 +0100)]
docs/system/ppc: Document running Linux on AmigaNG machines
Documentation on how to run Linux on the amigaone, pegasos2 and
sam460ex machines is currently buried in the depths of the qemu-devel
mailing list and in the source code. Let's collect the information in
the QEMU handbook for a one stop solution.
Tested-by: Bernhard Beschow <shentey@gmail.com> Reviewed-by: Bernhard Beschow <shentey@gmail.com> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Co-authored-by: Bernhard Beschow <shentey@gmail.com> Signed-off-by: Bernhard Beschow <shentey@gmail.com> Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
pSeries machines before 3.0 have complex migration back
compatibility code we'd like to get ride of. The last
one is 2.12, which is 6 years old. We just deprecated up
to the 2.11 machine in commit 1392617d35 ("spapr: Tag
pseries-2.1 - 2.11 machines as deprecated").
Take to opportunity to also deprecate the 2.12 machines.
Acked-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Benjamin Gray [Wed, 27 Dec 2023 13:46:50 +0000 (23:46 +1000)]
ppc/spapr: Add pa-features for POWER10 machines
Add POWER10 pa-features entry.
Notably DEXCR and [P]HASHST/[P]HASHCHK instruction support is
advertised. Each DEXCR aspect is allocated a bit in the device tree,
using the 68--71 byte range (inclusive). The functionality of the
[P]HASHST/[P]HASHCHK instructions is separately declared in byte 72,
bit 0 (BE).
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Benjamin Gray <bgray@linux.ibm.com>
[npiggin: reword title and changelog, adjust a few bits] Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Nicholas Piggin [Wed, 27 Dec 2023 13:53:25 +0000 (23:53 +1000)]
ppc/spapr: Adjust ibm,pa-features for POWER9
"MMR" and "SPR SO" are not implemented in POWER9, so clear those bits.
HTM is not set by default, and only later if the cap is set, so remove
the comment that suggests otherwise.
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Nicholas Piggin [Thu, 18 Jan 2024 12:08:16 +0000 (22:08 +1000)]
ppc/spapr: Remove copy-paste from pa-features
TCG does not support copy/paste instructions. Remove it from
ibm,pa-features. This has never been implemented under TCG or
practically usable under KVM, so it won't be missed.
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Nicholas Piggin [Wed, 27 Dec 2023 13:53:25 +0000 (23:53 +1000)]
ppc/spapr|pnv: Remove SAO from pa-features
SAO is a page table attribute that strengthens the memory ordering of
accesses. QEMU with MTTCG does not implement this, so clear it in
ibm,pa-features. This is an obscure feature that has been removed from
POWER10 ISA v3.1, there isn't much concern with removing it.
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Nicholas Piggin [Mon, 15 May 2023 16:19:53 +0000 (02:19 +1000)]
target/ppc: POWER10 does not have transactional memory
POWER10 hardware implements a degenerate transactional memory facility
in POWER8/9 PCR compatibility modes to permit migration from older
CPUs, but POWER10 / ISA v3.1 mode does not support it so the CPU model
should not support it.
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Peter Maydell [Tue, 12 Mar 2024 13:42:57 +0000 (13:42 +0000)]
Merge tag 'net-pull-request' of https://github.com/jasowang/qemu into staging
# -----BEGIN PGP SIGNATURE-----
#
# iQEzBAABCAAdFiEEIV1G9IJGaJ7HfzVi7wSWWzmNYhEFAmXwPUAACgkQ7wSWWzmN
# YhFnIwgAgctDniJwlRxXB01eVlzXz7IulHnpSby07XEJxENSpGB8ufaeE4eK5gJy
# NVK6C2+1EU2vRxm4oIdcvtN4C4/jtRbYYjiSTx7eE4FmSkqshSnR5XCV72LDqG3i
# WbzInjMvYfysmcMXLfrWgxOnVew9WqEzlpEWlc7FfNKnkzBVf+JDztfqCUx0XM7H
# qefw4ImjqQw993QxJpipXC7aEGUyouB0RIBB71FkCa9ihlh9x7W68evbOI/jTn5q
# HWuStgS02sKHjRFliMbdbMY77FNUz4Yroo/GKSvGt64atxkQSJqPNAV+/9n18LNy
# QAH5eK6cXFPOIAaYpADU5kHDVVAFiw==
# =iBdx
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 12 Mar 2024 11:32:16 GMT
# gpg: using RSA key 215D46F48246689EC77F3562EF04965B398D6211
# gpg: Good signature from "Jason Wang (Jason Wang on RedHat) <jasowang@redhat.com>" [marginal]
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg: It is not certain that the signature belongs to the owner.
# Primary key fingerprint: 215D 46F4 8246 689E C77F 3562 EF04 965B 398D 6211
* tag 'net-pull-request' of https://github.com/jasowang/qemu:
ebpf: Updated eBPF program and skeleton.
qmp: Added new command to retrieve eBPF blob.
virtio-net: Added property to load eBPF RSS with fds.
ebpf: Added eBPF initialization by fds.
ebpf: Added eBPF map update through mmap.
Avoid unaligned fetch in ladr_match()
e1000e: fix link state on resume
igb: fix link state on resume
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Peter Maydell [Tue, 12 Mar 2024 13:42:49 +0000 (13:42 +0000)]
Merge tag 'hw-misc-20240312' of https://github.com/philmd/qemu into staging
Misc HW patch queue
- Rename hw/ide/ahci-internal.h for consistency (Zoltan)
- More convenient PCI hotplug trace events (Vladimir)
- Short CLI option to add drives for sam460ex machine (Zoltan)
- More missing ERRP_GUARD() macros (Zhao)
- Avoid faulting when unmapped I/O BAR is accessed on SPARC EBUS (Mark)
- Remove unused includes in hw/core/ (Zhao)
- New PCF8574 GPIO over I2C model (Dmitriy)
- Require ObjC on Darwin macOS by default (Peter)
- Corrected "-smp parameter=1" placement in docs/ (Zhao)
* tag 'hw-misc-20240312' of https://github.com/philmd/qemu:
docs/about/deprecated.rst: Move SMP configurations item to system emulator section
meson.build: Always require an objc compiler on macos hosts
hw/gpio: introduce pcf8574 driver
hw/core: Cleanup unused included headers in numa.c
hw/core: Cleanup unused included header in machine-qmp-cmds.c
hw/core: Cleanup unused included headers in cpu-common.c
sun4u: remap ebus BAR0 to use unassigned_io_ops instead of alias to PCI IO space
hw/misc/ivshmem: Fix missing ERRP_GUARD() for error_prepend()
hw/core/qdev-properties-system: Fix missing ERRP_GUARD() for error_prepend()
hw/core/loader-fit: Fix missing ERRP_GUARD() for error_prepend()
hw/ppc/sam460ex: Support short options for adding drives
hw/pci: add some convenient trace-events for pcie and shpc hotplug
hw/ide/ahci: Rename ahci_internal.h to ahci-internal.h
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Peter Maydell [Tue, 12 Mar 2024 11:35:41 +0000 (11:35 +0000)]
Merge tag 'migration-20240311-pull-request' of https://gitlab.com/peterx/qemu into staging
Migration pull request
- Avihai's fix to allow vmstate iterators to not starve for VFIO
- Maksim's fix on additional check on precopy load error
- Fabiano's fix on fdatasync() hang in mapped-ram
- Jonathan's fix on vring cached access over MMIO regions
- Cedric's cleanup patches 1-4 out of his error report series
- Yu's fix for RDMA migration (which used to be broken even for 8.2)
- Anthony's small cleanup/fix on err message
- Steve's patches on privatize migration.h
- Xiang's patchset to enable zero page detections in multifd threads
# -----BEGIN PGP SIGNATURE-----
#
# iIgEABYKADAWIQS5GE3CDMRX2s990ak7X8zN86vXBgUCZe9+uBIccGV0ZXJ4QHJl
# ZGhhdC5jb20ACgkQO1/MzfOr1wamaQD/SvmpMEcuRndT9LPSxzXowAGDZTBpYUfv
# 5XAbx80dS9IBAO8PJJgQJIBHBeacyLBjHP9CsdVtgw5/VW+wCsbfV4AB
# =xavb
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 11 Mar 2024 21:59:20 GMT
# gpg: using EDDSA key B9184DC20CC457DACF7DD1A93B5FCCCDF3ABD706
# gpg: issuer "peterx@redhat.com"
# gpg: Good signature from "Peter Xu <xzpeter@gmail.com>" [marginal]
# gpg: aka "Peter Xu <peterx@redhat.com>" [marginal]
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg: It is not certain that the signature belongs to the owner.
# Primary key fingerprint: B918 4DC2 0CC4 57DA CF7D D1A9 3B5F CCCD F3AB D706
* tag 'migration-20240311-pull-request' of https://gitlab.com/peterx/qemu: (34 commits)
migration/multifd: Add new migration test cases for legacy zero page checking.
migration/multifd: Enable multifd zero page checking by default.
migration/multifd: Implement ram_save_target_page_multifd to handle multifd version of MigrationOps::ram_save_target_page.
migration/multifd: Implement zero page transmission on the multifd thread.
migration/multifd: Add new migration option zero-page-detection.
migration/multifd: Allow clearing of the file_bmap from multifd
migration/multifd: Allow zero pages in file migration
migration: purge MigrationState from public interface
migration: delete unused accessors
migration: privatize colo interfaces
migration: migration_file_set_error
migration: migration_is_device
migration: migration_thread_is_self
migration: export vcpu_dirty_limit_period
migration: export migration_is_running
migration: export migration_is_active
migration: export migration_is_setup_or_active
migration: remove migration.h references
migration: export fewer options
migration: Fix format in error message
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Peter Maydell [Tue, 12 Mar 2024 11:35:35 +0000 (11:35 +0000)]
Merge tag 'pull-target-arm-20240311' of https://git.linaro.org/people/pmaydell/qemu-arm into staging
target-arm queue:
* contrib/elf2dmp: Improve robustness to corrupt input files
* docs: update copyright date to the year 2024
* hw/arm: Deprecate various old Arm machine types
* tag 'pull-target-arm-20240311' of https://git.linaro.org/people/pmaydell/qemu-arm:
docs: update copyright date to the year 2024
contrib/elf2dmp: Ensure phdrs fit in file
contrib/elf2dmp: Clamp QEMU note to file size
contrib/elf2dmp: Use GPtrArray
MAINTAINERS: Add Akihiko Odaki as a elf2dmp reviewer
contrib/elf2dmp: Use rol64() to decode
contrib/elf2dmp: Use lduw_le_p() to read PDB
contrib/elf2dmp: Ensure segment fits in file
contrib/elf2dmp: Always destroy PA space
contrib/elf2dmp: Always check for PA resolution failure
contrib/elf2dmp: Fix error reporting style in main.c
contrib/elf2dmp: Fix error reporting style in qemu_elf.c
contrib/elf2dmp: Fix error reporting style in pdb.c
contrib/elf2dmp: Fix error reporting style in download.c
contrib/elf2dmp: Fix error reporting style in addrspace.c
contrib/elf2dmp: Change pa_space_create() signature
contrib/elf2dmp: Continue even contexts are lacking
contrib/elf2dmp: Assume error by default
contrib/elf2dmp: Remove unnecessary err flags
hw/arm: Deprecate various old Arm machine types
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Updated section name, so libbpf should init/gues proper
program type without specifications during open/load.
Also, added map_flags with explicitly declared BPF_F_MMAPABLE.
Added check for BPF_F_MMAPABLE flag to meson script and
requirements to libbpf version.
Also changed fragmentation flag check - some TCP/UDP packets
may be considered fragmented if DF flag is set.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
Now, the binary objects may be retrieved by id.
It would require for future qmp commands that may require specific
eBPF blob.
Added command "request-ebpf". This command returns
eBPF program encoded base64. The program taken from the
skeleton and essentially is an ELF object that can be
loaded in the future with libbpf.
The reason to use the command to provide the eBPF object
instead of a separate artifact was to avoid issues related
to finding the eBPF itself. eBPF object is an ELF binary
that contains the eBPF program and eBPF map description(BTF).
Overall, eBPF object should contain the program and enough
metadata to create/load eBPF with libbpf. As the eBPF
maps/program should correspond to QEMU, the eBPF can't
be used from different QEMU build.
The first solution was a helper that comes with QEMU
and loads appropriate eBPF objects. And the issue is
to find a proper helper if the system has several
different QEMUs installed and/or built from the source,
which helpers may not be compatible.
Another issue is QEMU updating while there is a running
QEMU instance. With an updated helper, it may not be
possible to hotplug virtio-net device to the already
running QEMU. Overall, requesting the eBPF object from
QEMU itself solves possible failures with acceptable effort.
Links:
[PATCH 3/5] qmp: Added the helper stamp check.
https://lore.kernel.org/all/20230219162100.174318-4-andrew@daynix.com/
Signed-off-by: Andrew Melnychenko <andrew@daynix.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
virtio-net: Added property to load eBPF RSS with fds.
eBPF RSS program and maps may now be passed during initialization.
Initially was implemented for libvirt to launch qemu without permissions,
and initialized eBPF program through the helper.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
It allows using file descriptors of eBPF provided
outside of QEMU.
QEMU may be run without capabilities for eBPF and run
RSS program provided by management tool(g.e. libvirt).
Signed-off-by: Andrew Melnychenko <andrew@daynix.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
Changed eBPF map updates through mmaped array.
Mmaped arrays provide direct access to map data.
It should omit using bpf_map_update_elem() call,
which may require capabilities that are not present.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com> Signed-off-by: Jason Wang <jasowang@redhat.com>
Nick Briggs [Thu, 1 Feb 2024 18:11:17 +0000 (10:11 -0800)]
Avoid unaligned fetch in ladr_match()
There is no guarantee that the PCNetState is allocated such that
csr[8] is allocated on an 8-byte boundary. Since not all hosts are
capable of unaligned fetches the 16-bit elements need to be fetched
individually to avoid a potential fault. Closes issue #2143
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2143 Signed-off-by: Nick Briggs <nicholas.h.briggs@gmail.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Jason Wang <jasowang@redhat.com>
Laurent Vivier [Wed, 24 Jan 2024 10:40:06 +0000 (11:40 +0100)]
e1000e: fix link state on resume
On resume e1000e_vm_state_change() always calls e1000e_autoneg_resume()
that sets link_down to false, and thus activates the link even
if we have disabled it.
The problem can be reproduced starting qemu in paused state (-S) and
then set the link to down. When we resume the machine the link appears
to be up.
Laurent Vivier [Wed, 24 Jan 2024 10:29:03 +0000 (11:29 +0100)]
igb: fix link state on resume
On resume igb_vm_state_change() always calls igb_autoneg_resume()
that sets link_down to false, and thus activates the link even
if we have disabled it.
The problem can be reproduced starting qemu in paused state (-S) and
then set the link to down. When we resume the machine the link appears
to be up.
Zhao Liu [Tue, 12 Mar 2024 07:15:12 +0000 (15:15 +0800)]
docs/about/deprecated.rst: Move SMP configurations item to system emulator section
In the commit 54c4ea8f3ae6 ("hw/core/machine-smp: Deprecate unsupported
'parameter=1' SMP configurations"), the SMP related item is put under
the section "User-mode emulator command line arguments" instead of
"System emulator command line arguments".
-smp is a system emulator command, so move SMP configurations item to
system emulator section.
Signed-off-by: Zhao Liu <zhao1.liu@intel.com> Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-ID: <20240312071512.3283513-1-zhao1.liu@linux.intel.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Peter Maydell [Mon, 11 Mar 2024 13:33:34 +0000 (13:33 +0000)]
meson.build: Always require an objc compiler on macos hosts
We currently only insist that an ObjectiveC compiler is present on
macos hosts if we're building the Cocoa UI. However, since then
we've added some other parts of QEMU which are also written in ObjC:
the coreaudio audio backend, and the vmnet net backend. This means
that if you try to configure QEMU on macos with --disable-cocoa the
build will fail:
../meson.build:3741:13: ERROR: No host machine compiler for 'audio/coreaudio.m'
Since in practice any macos host will have an ObjC compiler
available, rather than trying to gate the compiler detection on an
increasingly complicated list of every bit of QEMU that uses ObjC,
just require it unconditionally on macos hosts.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-ID: <20240311133334.3991537-1-peter.maydell@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
NXP PCF8574 and compatible ICs are simple I2C GPIO expanders.
PCF8574 incorporates quasi-bidirectional IO, and simple
communication protocol, when IO read is I2C byte read, and
IO write is I2C byte write. User can think of it as
open-drain port, when line high state is input and line low
state is output.
Signed-off-by: Dmitrii Sharikhin <d.sharikhin@yadro.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <f1552d822276e878d84c01eba2cf2c7c9ebdde00.camel@yadro.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Zhao Liu [Mon, 11 Mar 2024 07:56:21 +0000 (15:56 +0800)]
hw/core: Cleanup unused included headers in numa.c
Remove unused header in numa.c:
* qemu/bitmap.h
* migration/vmstate.h
Note: Though parse_numa_hmat_lb() has the variable named "bitmap_copy",
it doesn't use the normal bitmap ops so that it's safe to exclude
qemu/bitmap.h header.
Tested by "./configure" and then "make".
Signed-off-by: Zhao Liu <zhao1.liu@intel.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20240311075621.3224684-4-zhao1.liu@linux.intel.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Zhao Liu [Mon, 11 Mar 2024 07:56:20 +0000 (15:56 +0800)]
hw/core: Cleanup unused included header in machine-qmp-cmds.c
Remove unused header (qemu/main-loop.h) in machine-qmp-cmds.c.
Tested by "./configure" and then "make".
Signed-off-by: Zhao Liu <zhao1.liu@intel.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20240311075621.3224684-3-zhao1.liu@linux.intel.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20240311075621.3224684-2-zhao1.liu@linux.intel.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Mark Cave-Ayland [Mon, 11 Mar 2024 06:43:45 +0000 (06:43 +0000)]
sun4u: remap ebus BAR0 to use unassigned_io_ops instead of alias to PCI IO space
During kernel startup OpenBSD accesses addresses mapped by BAR0 of the ebus device
but at offsets where no IO devices exist. Before commit 4aa07e8649 ("hw/sparc64/ebus:
Access memory regions via pci_address_space_io()") BAR0 was mapped to legacy IO
space which allows accesses to unmapped devices to succeed, but afterwards these
accesses to unmapped PCI IO space cause a memory fault which prevents OpenBSD from
booting.
Since no devices are mapped at the addresses accessed by OpenBSD, change ebus BAR0
from a PCI IO space alias to an IO memory region using unassigned_io_ops which allows
these accesses to succeed and so allows OpenBSD to boot once again.
Fixes: 4aa07e8649 ("hw/sparc64/ebus: Access memory regions via pci_address_space_io()") Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20240311064345.2531197-1-mark.cave-ayland@ilande.co.uk> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Zhao Liu [Mon, 11 Mar 2024 03:38:09 +0000 (11:38 +0800)]
hw/misc/ivshmem: Fix missing ERRP_GUARD() for error_prepend()
As the comment in qapi/error, passing @errp to error_prepend() requires
ERRP_GUARD():
* = Why, when and how to use ERRP_GUARD() =
*
* Without ERRP_GUARD(), use of the @errp parameter is restricted:
...
* - It should not be passed to error_prepend(), error_vprepend() or
* error_append_hint(), because that doesn't work with &error_fatal.
* ERRP_GUARD() lifts these restrictions.
*
* To use ERRP_GUARD(), add it right at the beginning of the function.
* @errp can then be used without worrying about the argument being
* NULL or &error_fatal.
ERRP_GUARD() could avoid the case when @errp is &error_fatal, the user
can't see this additional information, because exit() happens in
error_setg earlier than information is added [1].
The ivshmem_common_realize() passes @errp to error_prepend(), and as a
DeviceClass.realize method, there are too many possible callers to check
the impact of this defect; it may or may not be harmless. Thus it is
necessary to protect @errp with ERRP_GUARD().
To avoid the issue like [1] said, add missing ERRP_GUARD() at the
beginning of this function.
[1]: Issue description in the commit message of commit ae7c80a7bd73
("error: New macro ERRP_GUARD()").
Cc: Juan Quintela <quintela@trasno.org> Cc: Manos Pitsidianakis <manos.pitsidianakis@linaro.org> Cc: Michael Galaxy <mgalaxy@akamai.com> Cc: Steve Sistare <steven.sistare@oracle.com> Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
Message-ID: <20240311033822.3142585-17-zhao1.liu@linux.intel.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Zhao Liu [Mon, 11 Mar 2024 03:38:08 +0000 (11:38 +0800)]
hw/core/qdev-properties-system: Fix missing ERRP_GUARD() for error_prepend()
As the comment in qapi/error, passing @errp to error_prepend() requires
ERRP_GUARD():
* = Why, when and how to use ERRP_GUARD() =
*
* Without ERRP_GUARD(), use of the @errp parameter is restricted:
...
* - It should not be passed to error_prepend(), error_vprepend() or
* error_append_hint(), because that doesn't work with &error_fatal.
* ERRP_GUARD() lifts these restrictions.
*
* To use ERRP_GUARD(), add it right at the beginning of the function.
* @errp can then be used without worrying about the argument being
* NULL or &error_fatal.
ERRP_GUARD() could avoid the case when @errp is &error_fatal, the user
can't see this additional information, because exit() happens in
error_setg earlier than information is added [1].
The set_chr() passes @errp to error_prepend() without ERRP_GUARD().
As a PropertyInfo.set method, there are too many possible callers to
check the impact of this defect; it may or may not be harmless. Thus it
is necessary to protect @errp with ERRP_GUARD().
To avoid the issue like [1] said, add missing ERRP_GUARD() at the
beginning of this function.
[1]: Issue description in the commit message of commit ae7c80a7bd73
("error: New macro ERRP_GUARD()").
Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: "Daniel P. Berrangé" <berrange@redhat.com Cc: Eduardo Habkost <eduardo@habkost.net> Signed-off-by: Zhao Liu <zhao1.liu@intel.com> Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-ID: <20240311033822.3142585-16-zhao1.liu@linux.intel.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Zhao Liu [Mon, 11 Mar 2024 03:38:07 +0000 (11:38 +0800)]
hw/core/loader-fit: Fix missing ERRP_GUARD() for error_prepend()
As the comment in qapi/error, passing @errp to error_prepend() requires
ERRP_GUARD():
* = Why, when and how to use ERRP_GUARD() =
*
* Without ERRP_GUARD(), use of the @errp parameter is restricted:
...
* - It should not be passed to error_prepend(), error_vprepend() or
* error_append_hint(), because that doesn't work with &error_fatal.
* ERRP_GUARD() lifts these restrictions.
*
* To use ERRP_GUARD(), add it right at the beginning of the function.
* @errp can then be used without worrying about the argument being
* NULL or &error_fatal.
ERRP_GUARD() could avoid the case when @errp is &error_fatal, the user
can't see this additional information, because exit() happens in
error_setg earlier than information is added [1].
In hw/core/loader-fit.c, there are 2 functions passing @errp to
error_prepend() without ERRP_GUARD():
- fit_load_kernel()
- fit_load_fdt()
Their @errp parameters are both the pointers of the local @err virable
in load_fit().
Though they don't cause the issue like [1] said, to follow the
requirement of @errp, add missing ERRP_GUARD() at their beginning.
[1]: Issue description in the commit message of commit ae7c80a7bd73
("error: New macro ERRP_GUARD()").
Cc: Paul Burton <paulburton@kernel.org> Cc: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com> Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
Message-ID: <20240311033822.3142585-15-zhao1.liu@linux.intel.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
BALATON Zoltan [Tue, 5 Mar 2024 22:57:21 +0000 (23:57 +0100)]
hw/ppc/sam460ex: Support short options for adding drives
Having to use -drive if=none,... and -device ide-[cd,hd] is
inconvenient. Add support for shorter convenience options such as
-cdrom and -drive media=disk. Also adjust two nearby comments for code
style.
Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
Message-ID: <20240305225721.E9A404E6005@zero.eik.bme.hu> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
hw/pci: add some convenient trace-events for pcie and shpc hotplug
Add trace-events that may help to debug problems with hotplugging.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20240301154146.761531-2-vsementsov@yandex-team.ru> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Hao Xiang [Mon, 11 Mar 2024 18:00:15 +0000 (18:00 +0000)]
migration/multifd: Add new migration test cases for legacy zero page checking.
Now that zero page checking is done on the multifd sender threads by
default, we still provide an option for backward compatibility. This
change adds a qtest migration test case to set the zero-page-detection
option to "legacy" and run multifd migration with zero page checking on the
migration main thread.
Hao Xiang [Mon, 11 Mar 2024 18:00:14 +0000 (18:00 +0000)]
migration/multifd: Enable multifd zero page checking by default.
1. Set default "zero-page-detection" option to "multifd". Now
zero page checking can be done in the multifd threads and this
becomes the default configuration.
2. Handle migration QEMU9.0 -> QEMU8.2 compatibility. We provide
backward compatibility where zero page checking is done from the
migration main thread.
Hao Xiang [Mon, 11 Mar 2024 18:00:13 +0000 (18:00 +0000)]
migration/multifd: Implement ram_save_target_page_multifd to handle multifd version of MigrationOps::ram_save_target_page.
1. Add a dedicated handler for MigrationOps::ram_save_target_page in
multifd live migration.
2. Refactor ram_save_target_page_legacy so that the legacy and multifd
handlers don't have internal functions calling into each other.
Hao Xiang [Mon, 11 Mar 2024 18:00:12 +0000 (18:00 +0000)]
migration/multifd: Implement zero page transmission on the multifd thread.
1. Add zero_pages field in MultiFDPacket_t.
2. Implements the zero page detection and handling on the multifd
threads for non-compression, zlib and zstd compression backends.
3. Added a new value 'multifd' in ZeroPageDetection enumeration.
4. Adds zero page counters and updates multifd send/receive tracing
format to track the newly added counters.
Hao Xiang [Mon, 11 Mar 2024 18:00:11 +0000 (18:00 +0000)]
migration/multifd: Add new migration option zero-page-detection.
This new parameter controls where the zero page checking is running.
1. If this parameter is set to 'legacy', zero page checking is
done in the migration main thread.
2. If this parameter is set to 'none', zero page checking is disabled.
Fabiano Rosas [Mon, 11 Mar 2024 18:00:09 +0000 (18:00 +0000)]
migration/multifd: Allow zero pages in file migration
Currently, it's an error to have no data pages in the multifd file
migration because zero page detection is done in the migration thread
and zero pages don't reach multifd. This is enforced with the
pages->num assert.
We're about to add zero page detection on the multifd thread. Fix the
file_write_ramblock_iov() to stop considering p->iovs_num=0 an error.
Steve Sistare [Mon, 11 Mar 2024 17:48:39 +0000 (10:48 -0700)]
migration: export fewer options
A small number of migration options are accessed by migration clients,
but to see them clients must include all of options.h, which is mostly
for migration core code. migrate_mode() in particular will be needed by
multiple clients.
Refactor the option declarations so clients can see the necessary few via
misc.h, which already exports a portion of the client API.
Peter Maydell [Mon, 11 Mar 2024 18:42:40 +0000 (18:42 +0000)]
Merge tag 'qga-pull-2024-03-11-2' of https://github.com/kostyanf14/qemu into staging
qga-pull-2024-03-11-2
# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEEwsLBCepDxjwUI+uE711egWG6hOcFAmXvMF4ACgkQ711egWG6
# hOd7GQ/9H11bXH5U2HZHAyEv68rCuGxHt57yjy9GfrSGAE7kqkLJ0bHwxdgoj09A
# mmaTOakOEhM5tyrkFYsROde3ta0fAwdFQXyhqpWDHG0ZwDDCAlsNsEuDd541KXbg
# qFTue26BM4EJEwTYy94nEEhOHD+2GgEAuPIsUCF6QDhrq+sBqUts1pH3uUM+E3Sg
# 7HSXaF9O/RbgYR6J8FVA53tvNOP4WgOYZ/SZoFwKYzIOAblcaRgJvVbm600OqQDb
# DUZ96s6HUBVBazKx4t24WwPOkgcvYgp7b8QYj2VVfjRU2IHnFumv9A47KuEdJEUl
# meZlo8TsgfL+uSiYWiRvps1Eo1uoS2M4s4FYGbuOeABYviLr3XhJM7VarNyPp7nf
# lupuyeqydXCic5LFbjSg5gkpaIhFYQ9gANYr6JZjPPjX5G8hmgCeB6YsZT98Ygmi
# yG1np2luapvFGsDCPd8kgNpfTkKhOpKghUnTC3UESReV8mjBGhJHRQTd+lMV7YWJ
# reMRxlqrqNq69lDUnlNuSXOqPJX1qHHGkWuV9pdz641tFvw7vigE07qhvsCr8Y7b
# tAg6oDcw+1uuq2t3nHmfkzic6WTxb4lFmSpShsDu2sRA7MqWE3lCefNoS75R+xaI
# FebdrkjkFxdbazyl6oKXwxkOkAR7rFoTvVKKbG79DVMIBDiu4BE=
# =Xt/n
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 11 Mar 2024 16:25:02 GMT
# gpg: using RSA key C2C2C109EA43C63C1423EB84EF5D5E8161BA84E7
# gpg: Good signature from "Kostiantyn Kostiuk (Upstream PR sign) <kkostiuk@redhat.com>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg: There is no indication that the signature belongs to the owner.
# Primary key fingerprint: C2C2 C109 EA43 C63C 1423 EB84 EF5D 5E81 61BA 84E7
* tag 'qga-pull-2024-03-11-2' of https://github.com/kostyanf14/qemu:
qga-win: Add support of Windows Server 2025 in get-osinfo command
qga/commands-win32: Do not set matrix_lookup_t/win_10_0_t arrays size
qga/commands-win32: Declare const qualifier before type
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Anthony PERARD [Mon, 11 Mar 2024 12:34:39 +0000 (12:34 +0000)]
migration: Fix format in error message
In file_write_ramblock_iov(), "offset" is "uintptr_t" and not
"ram_addr_t". While usually they are both equivalent, this is not the
case with CONFIG_XEN_BACKEND.
Use the right format. This will fix build on 32-bit.
physmem: Fix wrong address in large address_space_read/write_cached_slow()
If the access is bigger than the MemoryRegion supports,
flatview_read/write_continue() will attempt to update the Memory Region.
but the address passed to flatview_translate() is relative to the cache, not
to the FlatView.
On arm/virt with interleaved CXL memory emulation and virtio-blk-pci this
lead to the first part of descriptor being read from the CXL memory and the
second part from PA 0x8 which happens to be a blank region
of a flash chip and all ffs on this particular configuration.
Note this test requires the out of tree ARM support for CXL, but
the problem is more general.
Avoid this by adding new address_space_read_continue_cached()
and address_space_write_continue_cached() which share all the logic
with the flatview versions except for the MemoryRegion lookup which
is unnecessary as the MemoryRegionCache only covers one MemoryRegion.
physmem: Reduce local variable scope in flatview_read/write_continue()
Precursor to factoring out the inner loops for reuse.
Reviewed-by: Peter Xu <peterx@redhat.com> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: David Hildenbrand <david@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Link: https://lore.kernel.org/r/20240307153710.30907-3-Jonathan.Cameron@huawei.com Signed-off-by: Peter Xu <peterx@redhat.com>
physmem: Rename addr1 to more informative mr_addr in flatview_read/write() and similar
The calls to flatview_read/write[_continue]() have parameters addr and
addr1 but the names give no indication of what they are addresses of.
Rename addr1 to mr_addr to reflect that it is the translated address
offset within the MemoryRegion returned by flatview_translate().
Similarly rename the parameter in address_space_read/write_cached_slow()
Yu Zhang [Wed, 6 Mar 2024 08:06:54 +0000 (09:06 +0100)]
migration/rdma: Fix a memory issue for migration
In commit 3fa9642ff7 change was made to convert the RDMA backend to
accept MigrateAddress struct. However, the assignment of "host" leads
to data corruption on the target host and the failure of migration.
isock->host = rdma->host;
By allocating the memory explicitly for it with g_strdup_printf(), the
issue is fixed and the migration doesn't fail any more.
Fixes: 3fa9642ff7 ("migration: convert rdma backend to accept MigrateAddress") Cc: qemu-stable <qemu-stable@nongnu.org> Cc: Li Zhijian <lizhijian@fujitsu.com> Link: https://lore.kernel.org/r/CAHEcVy4L_D6tuhJ8h=xLR4WaPaprJE3nnxZAEyUnoTrxQ6CF5w@mail.gmail.com Signed-off-by: Yu Zhang <yu.zhang@ionos.com>
[peterx: use g_strdup() instead of g_strdup_printf(), per Zhijian] Signed-off-by: Peter Xu <peterx@redhat.com>
Fabiano Rosas [Tue, 5 Mar 2024 19:56:29 +0000 (16:56 -0300)]
migration/multifd: Don't fsync when closing QIOChannelFile
Commit bc38feddeb ("io: fsync before closing a file channel") added a
fsync/fdatasync at the closing point of the QIOChannelFile to ensure
integrity of the migration stream in case of QEMU crash.
The decision to do the sync at qio_channel_close() was not the best
since that function runs in the main thread and the fsync can cause
QEMU to hang for several minutes, depending on the migration size and
disk speed.
To fix the hang, remove the fsync from qio_channel_file_close().
At this moment, the migration code is the only user of the fsync and
we're taking the tradeoff of not having a sync at all, leaving the
responsibility to the upper layers.
migration: Do not call PRECOPY_NOTIFY_SETUP notifiers in case of error
When commit bd2270608fa0 ("migration/ram.c: add a notifier chain for
precopy") added PRECOPY_NOTIFY_SETUP notifiers at the end of
qemu_savevm_state_setup(), it didn't take into account a possible
error in the loop calling vmstate_save() or .save_setup() handlers.
This will help detect issues regarding I/O channels usage.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Peter Xu <peterx@redhat.com> Signed-off-by: Cédric Le Goater <clg@redhat.com> Link: https://lore.kernel.org/r/20240304122844.1888308-7-clg@redhat.com Signed-off-by: Peter Xu <peterx@redhat.com>
Maksim Davydov [Mon, 4 Mar 2024 14:42:03 +0000 (17:42 +0300)]
migration/ram: add additional check
If a migration stream is broken, the address and flag reading can return
zero. Thus, an irrelevant flag error will be returned instead of EIO.
It can be fixed by additional check after the reading.
Avihai Horon [Mon, 4 Mar 2024 10:53:39 +0000 (12:53 +0200)]
vfio/migration: Add a note about migration rate limiting
VFIO migration buffer size is currently limited to 1MB. Therefore, there
is no need to check if migration rate exceeded, as in the worst case it
will exceed by only 1MB.
However, if the buffer size is later changed to a bigger value,
vfio_save_iterate() should enforce migration rate (similar to migration
RAM code).
Add a note about this in vfio_save_iterate() to serve as a reminder.
Avihai Horon [Mon, 4 Mar 2024 10:53:38 +0000 (12:53 +0200)]
vfio/migration: Refactor vfio_save_state() return value
Currently, vfio_save_state() returns 1 regardless of whether there is
more data to send or not. This was done to prevent a fast changing VFIO
device from potentially blocking other devices from sending their data,
as qemu_savevm_state_iterate() serialized devices.
Now that qemu_savevm_state_iterate() no longer serializes devices, there
is no need for that.
Refactor vfio_save_state() to return 0 if more data is available and 1
if no more data is available.
Avihai Horon [Mon, 4 Mar 2024 10:53:37 +0000 (12:53 +0200)]
migration: Don't serialize devices in qemu_savevm_state_iterate()
Commit 90697be8896c ("live migration: Serialize vmstate saving in stage
2") introduced device serialization in qemu_savevm_state_iterate(). The
rationale behind it was to first complete migration of slower changing
block devices and only then migrate the RAM, to avoid sending fast
changing RAM pages over and over.
This commit was added a long time ago, and while it was useful back
then, it is not the case anymore:
1. Block migration is deprecated, see commit 66db46ca83b8 ("migration:
Deprecate block migration").
2. Today there are other iterative devices besides RAM and block, such
as VFIO, which are registered for migration after RAM. With current
serialization behavior, a fast changing device can block other
devices from sending their data, which may prevent migration from
converging in some cases.
The issue described in item 2 was observed in several VFIO migration
scenarios with switchover-ack capability enabled, where some workload on
the VM prevented RAM from ever reaching a hard zero, thus blocking VFIO
initial pre-copy data from being sent. Hence, destination could not ack
switchover and migration could not converge.
Fix that by not serializing iterative devices in
qemu_savevm_state_iterate().
Note that this still doesn't fully prevent device starvation. As
correctly pointed out by Peter [1], a fast changing device might
constantly consume all allocated bandwidth and block the following
devices. However, this scenario is more likely to happen only if
max-bandwidth is low.
projects / mirror_qemu.git / log