[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH REPOST v3 00/80] refactor main RAM allocation to use hostmem
From: |
Laurent Vivier |
Subject: |
Re: [PATCH REPOST v3 00/80] refactor main RAM allocation to use hostmem backend |
Date: |
Thu, 23 Jan 2020 15:45:07 +0100 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.3.1 |
Le 23/01/2020 à 12:37, Igor Mammedov a écrit :
>
> v3:
> - due to libvirt not being ready, postpone till 5.1
> * [PATCH v2 82/86] numa: forbid '-numa node, mem' for 5.0 and newer
> machine types
> and depended
> [PATCH v2 86/86] numa: remove deprecated implicit RAM distribution
> between nodes
> - drop as not related "[PATCH v2 85/86] numa: make exit() usage consistent"
> - drop "[PATCH v2 76/86] post conversion default_ram_id cleanup"
> so that default memory-backedend won't be created for boards that do not
> care
> about -m. Which makes -m optin feature. We should decide what do in case
> board doesn't use -m (but that's out of scope of this series)
> - use object_register_sugar_prop() instead of hacking compat props directly
> - simplified/reworked aspeed patches
> - s/RAM_ADDR_FMT/size_to_str()/
> - rename 'ram-memdev' property to 'memory-backend'
> - minor fixes to numa-test
> - fixes for issues noticed during review of
> [PATCH v2 66/86] ppc/{ppc440_bamboo,sam460x}: drop RAM size fixup
>
> v2:
> - fix compile errors on mingw32 host by introducing RAM_ADDR_UFMT [11/86]
> - replace "[PATCH 43/86] hppa: drop RAM size fixup" with alternative
> patches made by Philippe (which effectively do the same thing but other
> way around)
> - ppc440: fix crash and add suggested valid RAM size in error output.
> s/ppc4xx_sdram_adjust/ppc4xx_sdram_prep/ and simplify it by removing
> not necessary nested loop
> - rebase on current master due to new conflicts
>
>
> Series removes ad hoc RAM allocation API
> (memory_region_allocate_system_memory)
> and consolidates it around hostmem backend. It allows to
> * resolve conflicts between global -mem-prealloc and hostmem's "policy"
> option
> fixing premature allocation before binding policy is applied
> * simplify complicated memory allocation routines which had to deal with 2
> ways
> to allocate RAM.
> * it allows to reuse hostmem backends of a choice for main RAM without adding
> extra CLI options to duplicate hostmem features.
> Recent case was -mem-shared, to enable vhost-user on targets that don't
> support hostmem backends [1] (ex: s390)
> * move RAM allocation from individual boards into generic machine code and
> provide them with prepared MemoryRegion.
> * clean up deprecated NUMA features which were tied to the old API (see
> patches)
> - "numa: remove deprecated -mem-path fallback to anonymous RAM"
> - (POSTPONED, waiting on libvirt side) "forbid '-numa node,mem' for 5.0
> and newer machine types"
> - (POSTPONED) "numa: remove deprecated implicit RAM distribution between
> nodes"
>
> Conversion introduces a new machine.memory-backend property and wrapper code
> that
> aliases global -mem-path and -mem-alloc into automatically created hostmem
> backend properties (provided memory-backend was not set explicitly given by
> user).
> And then follows bulk of trivial patches that incrementally convert individual
> boards to using machine.memory-backend provided MemoryRegion.
>
> Board conversion typically involves:
> * providing MachineClass::default_ram_size and MachineClass::default_ram_id
> so generic code could create default backend if user didn't explicitly
> provide
> memory-backend or -m options
> * dropping memory_region_allocate_system_memory() call
> * using convenience MachineState::ram MemoryRegion, which points to
> MemoryRegion
> allocated by ram-memdev
> On top of that for some boards:
> * added missing ram_size checks (typically it were boards with fixed ram
> size)
> * ram_size fixups were replaced by checks and hard errors, forcing user to
> provide correct "-m" values instead of ignoring it and continuing running.
>
> After all boards are converted the old API is removed and memory allocation
> routines are cleaned up.
>
> git tree for testing:
> https://github.com/imammedo/qemu convert_main_ram_to_memdev_v3
>
> previous rev:
> https://github.com/imammedo/qemu convert_main_ram_to_memdev_v2
> https://lists.nongnu.org/archive/html/qemu-devel/2020-01/msg02960.html
>
>
> Igor Mammedov (77):
> numa: remove deprecated -mem-path fallback to anonymous RAM
> machine: introduce memory-backend property
> machine: alias -mem-path and -mem-prealloc into memory-foo backend
> machine: introduce convenience MachineState::ram
> initialize MachineState::ram in NUMA case
> alpha:dp264: use memdev for RAM
> arm/aspeed: actually check RAM size
> arm/aspeed: use memdev for RAM
> arm/collie: use memdev for RAM
> arm/cubieboard: use memdev for RAM
> arm/digic_boards: use memdev for RAM
> arm/highbank: use memdev for RAM
> arm/imx25_pdk: drop RAM size fixup
> arm/imx25_pdk: use memdev for RAM
> arm/integratorcp: use memdev for RAM
> arm/kzm: drop RAM size fixup
> arm/kzm: use memdev for RAM
> arm/mcimx6ul-evk: use memdev for RAM
> arm/mcimx7d-sabre: use memdev for RAM
> arm/mps2-tz: use memdev for RAM
> arm/mps2: use memdev for RAM
> arm/musicpal: use memdev for RAM
> arm/nseries: use memdev for RAM
> arm/omap_sx1: use memdev for RAM
> arm/palm: use memdev for RAM
> arm/raspi: use memdev for RAM
> arm/sabrelite: use memdev for RAM
> arm/sbsa-ref: use memdev for RAM
> arm/versatilepb: use memdev for RAM
> arm/vexpress: use memdev for RAM
> arm/virt: use memdev for RAM
> arm/xilinx_zynq: drop RAM size fixup
> arm/xilinx_zynq: use memdev for RAM
> arm/xlnx-versal-virt: use memdev for RAM
> arm/xlnx-zcu102: use memdev for RAM
> s390x/s390-virtio-ccw: use memdev for RAM
> null-machine: use memdev for RAM
> cris/axis_dev88: use memdev for RAM
> hppa: use memdev for RAM
> x86/microvm: use memdev for RAM
> x86/pc: use memdev for RAM
> lm32/lm32_boards: use memdev for RAM
> lm32/milkymist: use memdev for RAM
> m68k/an5206: use memdev for RAM
> m68k/mcf5208: use memdev for RAM
> m68k/next-cube: use memdev for RAM
> mips/boston-cube: use memdev for RAM
> mips/mips_fulong2e: drop RAM size fixup
> mips/mips_fulong2e: use memdev for RAM
> mips/mips_jazz: use memdev for RAM
> mips/mips_malta: use memdev for RAM
> mips/mips_mipssim: use memdev for RAM
> mips/mips_r4k: use memdev for RAM
> ppc/e500: drop RAM size fixup
> ppc/e500: use memdev for RAM
> ppc/mac_newworld: use memdev for RAM
> ppc/mac_oldworld: use memdev for RAM
> ppc/pnv: use memdev for RAM
> ppc/ppc405_boards: add RAM size checks
> ppc/ppc405_boards: use memdev for RAM
> ppc/{ppc440_bamboo,sam460ex}: drop RAM size fixup
> ppc/{ppc440_bamboo, sam460ex}: use memdev for RAM
> ppc/prep: use memdev for RAM
> ppc/spapr: use memdev for RAM
> ppc/virtex_ml507: remove unused arguments
> ppc/virtex_ml507: use memdev for RAM
> sparc/leon3: use memdev for RAM
> sparc/sun4m: use memdev for RAM
> sparc/niagara: use memdev for RAM
> remove no longer used memory_region_allocate_system_memory()
> exec: cleanup qemu_minrampagesize()/qemu_maxrampagesize()
> exec: drop bogus mem_path from qemu_ram_alloc_from_fd()
> make mem_path local variable
> hostmem: introduce "prealloc-threads" property
> hostmem: fix strict bind policy
> tests:numa-test: make top level args dynamic and g_autofree(cli)
> cleanups
> tests:numa-test: use explicit memdev to specify node RAM
>
> Philippe Mathieu-Daudé (3):
> hw/hppa/machine: Correctly check the firmware is in PDC range
> hw/hppa/machine: Restrict the total memory size to 3GB
> hw/hppa/machine: Map the PDC memory region with higher priority
>
> hw/alpha/alpha_sys.h | 2 +-
> include/hw/boards.h | 48 +++++----------
> include/hw/misc/aspeed_sdmc.h | 1 +
> include/hw/ppc/ppc4xx.h | 9 ++-
> include/sysemu/hostmem.h | 20 +++++-
> include/sysemu/numa.h | 1 +
> include/sysemu/sysemu.h | 2 -
> backends/hostmem-file.c | 8 ---
> backends/hostmem-memfd.c | 1 -
> backends/hostmem-ram.c | 2 -
> backends/hostmem.c | 53 +++++++++++-----
> exec.c | 66 +++-----------------
> hw/alpha/dp264.c | 3 +-
> hw/alpha/typhoon.c | 8 +--
> hw/arm/aspeed.c | 18 +++---
> hw/arm/collie.c | 17 ++++--
> hw/arm/cubieboard.c | 25 +++-----
> hw/arm/digic_boards.c | 40 ++++++------
> hw/arm/highbank.c | 10 ++-
> hw/arm/imx25_pdk.c | 13 ++--
> hw/arm/integratorcp.c | 9 ++-
> hw/arm/kzm.c | 18 +++---
> hw/arm/mcimx6ul-evk.c | 25 +++-----
> hw/arm/mcimx7d-sabre.c | 25 +++-----
> hw/arm/mps2-tz.c | 15 +++--
> hw/arm/mps2.c | 15 +++--
> hw/arm/musicpal.c | 18 ++++--
> hw/arm/nseries.c | 32 ++++++----
> hw/arm/omap_sx1.c | 20 ++++--
> hw/arm/palm.c | 20 ++++--
> hw/arm/raspi.c | 34 ++++-------
> hw/arm/sabrelite.c | 23 +++----
> hw/arm/sbsa-ref.c | 7 +--
> hw/arm/versatilepb.c | 7 +--
> hw/arm/vexpress.c | 14 ++---
> hw/arm/virt.c | 7 +--
> hw/arm/xilinx_zynq.c | 20 +++---
> hw/arm/xlnx-versal-virt.c | 7 +--
> hw/arm/xlnx-zcu102.c | 7 +--
> hw/core/machine.c | 30 +++++++++
> hw/core/null-machine.c | 8 +--
> hw/core/numa.c | 101 +++++++++----------------------
> hw/cris/axis_dev88.c | 8 +--
> hw/hppa/machine.c | 19 +++---
> hw/i386/microvm.c | 12 ++--
> hw/i386/pc.c | 19 +++---
> hw/lm32/lm32_boards.c | 39 ++++++++----
> hw/lm32/milkymist.c | 21 ++++---
> hw/m68k/an5206.c | 5 +-
> hw/m68k/mcf5208.c | 5 +-
> hw/m68k/next-cube.c | 5 +-
Could you add hw/m68k/q800.c ?
Thanks,
Laurent
- [PATCH REPOST v3 76/80] make mem_path local variable, (continued)
- [PATCH REPOST v3 76/80] make mem_path local variable, Igor Mammedov, 2020/01/23
- [PATCH REPOST v3 77/80] hostmem: introduce "prealloc-threads" property, Igor Mammedov, 2020/01/23
- [PATCH REPOST v3 78/80] hostmem: fix strict bind policy, Igor Mammedov, 2020/01/23
- [PATCH REPOST v3 73/80] remove no longer used memory_region_allocate_system_memory(), Igor Mammedov, 2020/01/23
- [PATCH REPOST v3 79/80] tests:numa-test: make top level args dynamic and g_autofree(cli) cleanups, Igor Mammedov, 2020/01/23
- [PATCH REPOST v3 80/80] tests:numa-test: use explicit memdev to specify node RAM, Igor Mammedov, 2020/01/23
- Re: [PATCH REPOST v3 00/80] refactor main RAM allocation to use hostmem backend, no-reply, 2020/01/23
- Re: [PATCH REPOST v3 00/80] refactor main RAM allocation to use hostmem backend,
Laurent Vivier <=