On Thu, Nov 21, 2013 at 03:37:55PM +0100, Igor Mammedov wrote: > On Thu, 21 Nov 2013 15:43:27 +0200 > "Michael S. Tsirkin" <m...@redhat.com> wrote: > > > On Thu, Nov 21, 2013 at 02:39:10PM +0100, Igor Mammedov wrote: > > > On Thu, 21 Nov 2013 08:20:56 +0200 > > > "Michael S. Tsirkin" <m...@redhat.com> wrote: > > > > > > > On Thu, Nov 21, 2013 at 03:38:21AM +0100, Igor Mammedov wrote: > > > > > --- > > > > > What's new since v6: > > > > > > > > > > * DIMM device is split to backend and frontend. Therefore following > > > > > command/options were added for supporting it: > > > > > > > > > > For memdev backend: > > > > > CLI: -memdev-add > > > > > Monitor/QMP: memdev-add > > > > > with options: 'id' and 'size' > > > > > For dimm frontend: > > > > > option "size" became readonly, pulling it's size from attached > > > > > backend > > > > > added option "memdev" for specifying backend by 'id' > > > > > > > > > > * Added Q35 support > > > > > * Added support for 32 bit guests > > > > > * build for i386 emulator (not tested) > > > > > > > > OK so a large patchset so did not review yet. One question > > > > due to the dependency on bios honouring etc/reserved-memory-end: is > > > > there some way to detect old BIOS and fail memory hotplug? > > > version negotiation between ASL and hardware could be used to that effect. > > > > > > QEMU could start with present but disabled memory hotplug and if > > > QEMU and BIOS ASL could come in agreement that both support it in > > > compatible way, it could be enabled. > > > > So at the moment there's no negotiation, is there? > Nope, I surely can add it in next iteration.
I'm not sure about this - if we add this we should make it a generic mechanism I think. It does not have to be part of this series. So maybe drop version for now ... > > > > > > > > > > > --- > > > > > > > > > > This series allows to hotplug 'arbitrary' DIMM devices specifying > > > > > size, > > > > > NUMA node mapping (guest side), slot and address where to map it, at > > > > > runtime. > > > > > > > > > > Due to ACPI limitation there is need to specify a number of possible > > > > > DIMM devices. For this task -m option was extended to support > > > > > following format: > > > > > > > > > > -m [mem=]RamSize[,slots=N,maxmem=M] > > > > > > > > > > To allow memory hotplug user must specify a pair of additional > > > > > parameters: > > > > > 'slots' - number of possible increments > > > > > 'maxmem' - max possible total memory size QEMU is allowed to use, > > > > > including RamSize. > > > > > > > > > > minimal monitor command syntax to hotplug DIMM device: > > > > > > > > > > memdev-add id=memX,size=1G > > > > > device_add dimm,id=dimmX,memdev=memX > > > > > > > > > > DIMM device provides following properties that could be used with > > > > > device_add / -device to alter default behavior: > > > > > > > > > > id - unique string identifying device [mandatory] > > > > > slot - number in range [0-slots) [optional], if not specified > > > > > the first free slot is used > > > > > node - NUMA node id [optional] (default: 0) > > > > > size - amount of memory to add, readonly derived from backing > > > > > memdev > > > > > start - guest's physical address where to plug DIMM [optional], > > > > > if not specified the first gap in hotplug memory region > > > > > that fits DIMM is used > > > > > > > > > > -device option could be used for adding potentially hotunplugable > > > > > DIMMs > > > > > and also for specifying hotplugged DIMMs in migration case. > > > > > > > > > > Tested guests: > > > > > - RHEL 6x64 > > > > > - Windows 2012DCx64 > > > > > - Windows 2008DCx64 > > > > > - Windows 2008DCx32 > > > > > > > > > > Known limitations/bugs/TODOs: > > > > > - only hot-add supported > > > > > - max number of supported DIMM devices 255 (due to ACPI object name > > > > > limit), could be increased creating several containers and putting > > > > > DIMMs there. (exercise for future) > > > > > - failed hotplug action consumes 1 slot (device_add doesn't delete > > > > > device properly if realize failed) > > > > > - e820 table doesn't include DIMM devices added with -device / > > > > > (or after reboot devices added with device_add) > > > > > - Windows 2008 remembers DIMM configuration, so if DIMM with other > > > > > start/size is added into the same slot, it refuses to use it > > > > > insisting > > > > > on old mapping. > > > > > > > > > > Series is based on mst's PCI tree and requires following SeaBIOS > > > > > patch: > > > > > http://patchwork.ozlabs.org/patch/292614/ > > > > > on top of patches to load ACPI tables from QEMU > > > > > working SeaBIOS tree for testing is available at: > > > > > https://github.com/imammedo/seabios/commits/memhp-wip > > > > > > > > > > QEMU git tree for testing is available at: > > > > > https://github.com/imammedo/qemu/commits/memory-hotplug-v7 > > > > > > > > > > Example QEMU cmd line: > > > > > qemu-system-x86_64 -enable-kvm -monitor unix:/tmp/mon,server,nowait > > > > > \ > > > > > -m 4096,slots=4,maxmem=8G -L /custome_seabios guest.img > > > > > > > > > > PS: > > > > > Windows guest requires SRAT table for hotplug to work so add extra > > > > > option: > > > > > -numa node > > > > > to QEMU command line. > > > > > > > > > > > > > > > Igor Mammedov (26): > > > > > acpi: factor out common pm_update_sci() into acpi core > > > > > rename pci_hotplug_fn to hotplug_fn and make it available for other > > > > > devices > > > > > pc: add 'etc/reserved-memory-end' fw_cfg interface for SeaBIOS > > > > > vl: convert -m to qemu_opts_parse() > > > > > qapi: add SIZE type parser to string_input_visitor > > > > > get reference to /backend container via qemu_get_backend() > > > > > add memdev backend infrastructure > > > > > dimm: map DimmDevice into DimBus provided address space > > > > > dimm: add busy slot check and slot auto-allocation > > > > > dimm: add busy address check and address auto-allocation > > > > > dimm: add hotplug callback to DimmBus > > > > > acpi: memory hotplug ACPI hardware implementation > > > > > acpi: initialize memory hotplug ACPI PIIX4 hardware > > > > > acpi: piix4: add memory-hotplug-io-base property to piix4_pm > > > > > acpi: ich9: allow guest to clear SCI rised by GPE > > > > > acpi: initialize memory hotplug ACPI ICH9 hardware > > > > > acpi: ich9: add memory-hotplug-io-base property to ich9_pm > > > > > acpi: piix4/ich9: add optional vmstate field for MemHotplugState > > > > > migration > > > > > pc: piix: make PCII440FXState type public > > > > > pc: add memory hotplug 440fx machine > > > > > pc: add memory hotplug Q35 machine > > > > > pc: ACPI BIOS: implement memory hotplug interface > > > > > pc: ACPI BIOS: add ssdt-mem.hex.generated and update > > > > > ssdt-misc.hex.generated > > > > > pc: ACPI BIOS: use enum for defining memory affinity flags > > > > > pc: ACPI BIOS: reserve SRAT entry for hotplug mem hole > > > > > pc: ACPI BIOS: make GPE.3 handle memory hotplug event on PIIX and > > > > > Q35 > > > > > machines > > > > > > > > > > Vasilis Liaskovitis (1): > > > > > dimm: implement dimm device abstraction > > > > > > > > > > backends/Makefile.objs | 2 + > > > > > backends/hostmem.c | 275 +++++++++++ > > > > > backends/hostmem_compat_ram.c | 42 ++ > > > > > default-configs/i386-softmmu.mak | 1 + > > > > > default-configs/x86_64-softmmu.mak | 1 + > > > > > docs/specs/acpi_mem_hotplug.txt | 38 ++ > > > > > hmp-commands.hx | 13 + > > > > > hw/Makefile.objs | 1 + > > > > > hw/acpi/core.c | 189 ++++++++ > > > > > hw/acpi/ich9.c | 57 ++- > > > > > hw/acpi/pcihp.c | 6 +- > > > > > hw/acpi/piix4.c | 91 +++-- > > > > > hw/i386/Makefile.objs | 3 +- > > > > > hw/i386/acpi-build.c | 91 ++++- > > > > > hw/i386/acpi-dsdt.dsl | 5 +- > > > > > hw/i386/pc.c | 90 ++++- > > > > > hw/i386/pc_piix.c | 7 +- > > > > > hw/i386/pc_q35.c | 4 + > > > > > hw/i386/q35-acpi-dsdt.dsl | 5 +- > > > > > hw/i386/ssdt-mem.dsl | 75 +++ > > > > > hw/i386/ssdt-mem.hex.generated | 197 ++++++++ > > > > > hw/i386/ssdt-misc.dsl | 179 ++++++++ > > > > > hw/i386/ssdt-misc.hex.generated | 881 > > > > > +++++++++++++++++++++++++++++++++++- > > > > > hw/mem/Makefile.objs | 1 + > > > > > hw/mem/dimm.c | 321 +++++++++++++ > > > > > hw/pci-host/piix.c | 36 +- > > > > > hw/pci-host/q35.c | 19 +- > > > > > hw/pci/pci.c | 12 +- > > > > > hw/pci/pcie.c | 7 +- > > > > > hw/pci/shpc.c | 9 +- > > > > > include/hw/acpi/acpi.h | 42 ++ > > > > > include/hw/acpi/ich9.h | 6 + > > > > > include/hw/acpi/pcihp.h | 2 +- > > > > > include/hw/acpi/piix4.h | 2 + > > > > > include/hw/i386/pc.h | 33 ++- > > > > > include/hw/mem/dimm.h | 99 ++++ > > > > > include/hw/pci-host/piix.h | 29 ++ > > > > > include/hw/pci-host/q35.h | 1 + > > > > > include/hw/pci/pci.h | 11 +- > > > > > include/hw/pci/pci_bus.h | 2 +- > > > > > include/hw/qdev-core.h | 9 + > > > > > include/sysemu/hostmem.h | 102 +++++ > > > > > include/sysemu/sysemu.h | 3 + > > > > > monitor.c | 1 + > > > > > qapi-schema.json | 18 + > > > > > qapi/string-input-visitor.c | 18 + > > > > > qemu-options.hx | 16 +- > > > > > qmp-commands.hx | 27 ++ > > > > > ui/console.c | 3 +- > > > > > vl.c | 98 ++++- > > > > > 50 files changed, 3040 insertions(+), 140 deletions(-) > > > > > create mode 100644 backends/hostmem.c > > > > > create mode 100644 backends/hostmem_compat_ram.c > > > > > create mode 100644 docs/specs/acpi_mem_hotplug.txt > > > > > create mode 100644 hw/i386/ssdt-mem.dsl > > > > > create mode 100644 hw/i386/ssdt-mem.hex.generated > > > > > create mode 100644 hw/mem/Makefile.objs > > > > > create mode 100644 hw/mem/dimm.c > > > > > create mode 100644 include/hw/mem/dimm.h > > > > > create mode 100644 include/hw/pci-host/piix.h > > > > > create mode 100644 include/sysemu/hostmem.h > >
