| Message ID | 20200217131248.28273-1-gengdongjiu@huawei.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | Add ARMv8 RAS virtualization support in QEMU | expand |
On Mon, 17 Feb 2020 at 13:10, Dongjiu Geng <gengdongjiu@huawei.com> wrote: > > In the ARMv8 platform, the CPU error types includes synchronous external abort(SEA) and SError Interrupt (SEI). If exception happens in guest, host does not know the detailed information of guest, so it is expected that guest can do the recovery. > For example, if an exception happens in a guest user-space application, host does > not know which application encounters errors, only guest knows it. > > For the ARMv8 SEA/SEI, KVM or host kernel delivers SIGBUS to notify userspace. > After user space gets the notification, it will record the CPER into guest GHES > buffer and inject an exception or IRQ to guest. > > In the current implementation, if the type of SIGBUS is BUS_MCEERR_AR, we will > treat it as a synchronous exception, and notify guest with ARMv8 SEA > notification type after recording CPER into guest. Hi; I have reviewed the remaining arm bit of this series (patch 9), and made some comments on patch 1. Still to be reviewed are patches 4, 5, 6, 8: I'm going to assume that Michael or Igor will look at those. thanks -- PMM
On 2020/2/21 22:09, Peter Maydell wrote: > On Mon, 17 Feb 2020 at 13:10, Dongjiu Geng <gengdongjiu@huawei.com> wrote: >> >> In the ARMv8 platform, the CPU error types includes synchronous external abort(SEA) and SError Interrupt (SEI). If exception happens in guest, host does not know the detailed information of guest, so it is expected that guest can do the recovery. >> For example, if an exception happens in a guest user-space application, host does >> not know which application encounters errors, only guest knows it. >> >> For the ARMv8 SEA/SEI, KVM or host kernel delivers SIGBUS to notify userspace. >> After user space gets the notification, it will record the CPER into guest GHES >> buffer and inject an exception or IRQ to guest. >> >> In the current implementation, if the type of SIGBUS is BUS_MCEERR_AR, we will >> treat it as a synchronous exception, and notify guest with ARMv8 SEA >> notification type after recording CPER into guest. > > Hi; I have reviewed the remaining arm bit of this series (patch 9), > and made some comments on patch 1. Still to be reviewed are > patches 4, 5, 6, 8: I'm going to assume that Michael or Igor > will look at those. Thanks very much for Peter's review. Michael/Igor, hope you can review patches 4, 5, 6, 8, thank you very much in advance. > > thanks > -- PMM > > . >
On Mon, 24 Feb 2020 16:37:44 +0800 gengdongjiu <gengdongjiu@huawei.com> wrote: > On 2020/2/21 22:09, Peter Maydell wrote: > > On Mon, 17 Feb 2020 at 13:10, Dongjiu Geng <gengdongjiu@huawei.com> wrote: > >> > >> In the ARMv8 platform, the CPU error types includes synchronous external abort(SEA) and SError Interrupt (SEI). If exception happens in guest, host does not know the detailed information of guest, so it is expected that guest can do the recovery. > >> For example, if an exception happens in a guest user-space application, host does > >> not know which application encounters errors, only guest knows it. > >> > >> For the ARMv8 SEA/SEI, KVM or host kernel delivers SIGBUS to notify userspace. > >> After user space gets the notification, it will record the CPER into guest GHES > >> buffer and inject an exception or IRQ to guest. > >> > >> In the current implementation, if the type of SIGBUS is BUS_MCEERR_AR, we will > >> treat it as a synchronous exception, and notify guest with ARMv8 SEA > >> notification type after recording CPER into guest. > > > > Hi; I have reviewed the remaining arm bit of this series (patch 9), > > and made some comments on patch 1. Still to be reviewed are > > patches 4, 5, 6, 8: I'm going to assume that Michael or Igor > > will look at those. > > Thanks very much for Peter's review. > Michael/Igor, hope you can review patches 4, 5, 6, 8, thank you very much in advance. done > > > > thanks > > -- PMM > > > > . > > >
On 2020/2/26 0:59, Igor Mammedov wrote: > On Mon, 24 Feb 2020 16:37:44 +0800 > gengdongjiu <gengdongjiu@huawei.com> wrote: > >> On 2020/2/21 22:09, Peter Maydell wrote: >>> On Mon, 17 Feb 2020 at 13:10, Dongjiu Geng <gengdongjiu@huawei.com> wrote: >>>> >>>> In the ARMv8 platform, the CPU error types includes synchronous external abort(SEA) and SError Interrupt (SEI). If exception happens in guest, host does not know the detailed information of guest, so it is expected that guest can do the recovery. >>>> For example, if an exception happens in a guest user-space application, host does >>>> not know which application encounters errors, only guest knows it. >>>> >>>> For the ARMv8 SEA/SEI, KVM or host kernel delivers SIGBUS to notify userspace. >>>> After user space gets the notification, it will record the CPER into guest GHES >>>> buffer and inject an exception or IRQ to guest. >>>> >>>> In the current implementation, if the type of SIGBUS is BUS_MCEERR_AR, we will >>>> treat it as a synchronous exception, and notify guest with ARMv8 SEA >>>> notification type after recording CPER into guest. >>> >>> Hi; I have reviewed the remaining arm bit of this series (patch 9), >>> and made some comments on patch 1. Still to be reviewed are >>> patches 4, 5, 6, 8: I'm going to assume that Michael or Igor >>> will look at those. >> >> Thanks very much for Peter's review. >> Michael/Igor, hope you can review patches 4, 5, 6, 8, thank you very much in advance. > > done Thanks a lot, do you think whether it is ready to merge if the comments that you mentioned are fixed? If there are some other places that still needs to modify, hope you can pointed it out, so that in next patch series it can be ready to merge. thanks. > >>> >>> thanks >>> -- PMM >>> >>> . >>> >> > > . >
In the ARMv8 platform, the CPU error types includes synchronous external abort(SEA) and SError Interrupt (SEI). If exception happens in guest, host does not know the detailed information of guest, so it is expected that guest can do the recovery. For example, if an exception happens in a guest user-space application, host does not know which application encounters errors, only guest knows it. For the ARMv8 SEA/SEI, KVM or host kernel delivers SIGBUS to notify userspace. After user space gets the notification, it will record the CPER into guest GHES buffer and inject an exception or IRQ to guest. In the current implementation, if the type of SIGBUS is BUS_MCEERR_AR, we will treat it as a synchronous exception, and notify guest with ARMv8 SEA notification type after recording CPER into guest. A) This series of patches are based on Qemu 4.2, which include two parts: 1. Generate APEI/GHES table. 2. Handle the SIGBUS signal, record the CPER in runtime and fill it into guest memory, then notify guest according to the type of SIGBUS. B) The solution was suggested by James(james.morse@arm.com); The APEI part solution was suggested by Laszlo(lersek@redhat.com). Show some discussions in [1]. C) This series of patches have already been tested on ARM64 platform with RAS feature enabled: 1. Show the APEI part verification result in [2]. 2. Show the SIGBUS of BUS_MCEERR_AR handling verification result in [3]. D) Add 'ras' option in command Line to enable guest RAS error recovery feature, For example: KVM model: ./qemu-system-aarch64 --enable-kvm -cpu host --bios QEMU_EFI.fd_new -machine virt,gic-version=3,ras,kernel-irqchip=on -smp 4 -nographic -kernel Image -append "rdinit=/init console=ttyAMA0 mem=512M root=/dev/ram0" -initrd guestfs_new.cpio.gz TCG model: ./qemu-system-aarch64 -cpu cortex-a57 --bios QEMU_EFI.fd_new -machine virt,gic-version=3,ras,kernel-irqchip=on -smp 4 -nographic -kernel Image -append "rdinit=/init console=ttyAMA0 mem=512M root=/dev/ram0" -initrd guestfs_new.cpio.gz --- Change since v23: 1. fix a warning for uuid Change since v22: 1. Using 1 * KiB instead of 0x400 to define max size of one error block 2. Make the alignment to 8 bytes in bios_linker_loader_alloc() 3. Change "Copyright (c) 2019" to "Copyright (c) 2020" in file header 4. Fix some code style warnings/errors and add some comments in code 5. Address Jonathan's comments to easily support CCIX error injection 6. Add vmstate_ghes_state .subsections in vmstate_acpi_ged Change since v21: 1. Make the user-facing 'ras' option description more clearly to address Peter's comments. 2. Update the doc description in "docs/specs/acpi_hest_ghes.rst" 3. Split HEST/GHES patches to more patches to make the review easily 4. Using source_id to index the location to save the CPER. 5. Optimize and simplify the logic to build HEST/GHES table to address Igor/Michael/Beata comments. 6. make ghes_addr_le a part of GED device. Change since v20: 1. Move some implementation details from acpi_ghes.h to acpi_ghes.c 2. Add the reviewers for the ACPI/APEI/GHES part Change since v19: 1. Fix clang compile error 2. Fix sphinx build error Change since v18: 1. Fix some code-style and typo/grammar problems. 2. Remove no_ras in the VirtMachineClass struct. 3. Convert documentation to rst format. 4. Simplize the code and add comments for some magic value. 5. Move kvm_inject_arm_sea() function into the patch where it's used. 6. Register the reset handler(kvm_unpoison_all()) in the kvm_init() function. Change since v17: 1. Improve some commit messages and comments. 2. Fix some code-style problems. 3. Add a *ras* machine option. 4. Move HEST/GHES related structures and macros into "hw/acpi/acpi_ghes.*". 5. Move HWPoison page functions into "include/sysemu/kvm_int.h". 6. Fix some bugs. 7. Improve the design document. Change since v16: 1. check whether ACPI table is enabled when handling the memory error in the SIGBUS handler. Change since v15: 1. Add a doc-comment in the proper format for 'include/exec/ram_addr.h' 2. Remove write_part_cpustate_to_list() because there is another bug fix patch has been merged "arm: Allow system registers for KVM guests to be changed by QEMU code" 3. Add some comments for kvm_inject_arm_sea() in 'target/arm/kvm64.c' 4. Compare the arm_current_el() return value to 0,1,2,3, not to PSTATE_MODE_* constants. 5. Change the RAS support wasn't introduced before 4.1 QEMU version. 6. Move the no_ras flag patch to begin in this series Change since v14: 1. Remove the BUS_MCEERR_AO handling logic because this asynchronous signal was masked by main thread 2. Address some Igor Mammedov's comments(ACPI part) 1) change the comments for the enum AcpiHestNotifyType definition and remove ditto in patch 1 2) change some patch commit messages and separate "APEI GHES table generation" patch to more patches. 3. Address some peter's comments(arm64 Synchronous External Abort injection) 1) change some code notes 2) using arm_current_el() for current EL 2) use the helper functions for those (syn_data_abort_*). Change since v13: 1. Move the patches that set guest ESR and inject virtual SError out of this series 2. Clean and optimize the APEI part patches 3. Update the commit messages and add some comments for the code Change since v12: 1. Address Paolo's comments to move HWPoisonPage definition to accel/kvm/kvm-all.c 2. Only call kvm_cpu_synchronize_state() when get the BUS_MCEERR_AR signal 3. Only add and enable GPIO-Signal and ARMv8 SEA two hardware error sources 4. Address Michael's comments to not sync SPDX from Linux kernel header file Change since v11: Address James's comments(james.morse@arm.com) 1. Check whether KVM has the capability to to set ESR instead of detecting host CPU RAS capability 2. For SIGBUS_MCEERR_AR SIGBUS, use Synchronous-External-Abort(SEA) notification type for SIGBUS_MCEERR_AO SIGBUS, use GPIO-Signal notification Address Shannon's comments(for ACPI part): 1. Unify hest_ghes.c and hest_ghes.h license declaration 2. Remove unnecessary including "qmp-commands.h" in hest_ghes.c 3. Unconditionally add guest APEI table based on James's comments(james.morse@arm.com) 4. Add a option to virt machine for migration compatibility. On new virt machine it's on by default while off for old ones, we enabled it since 2.12 5. Refer to the ACPI spec version which introduces Hardware Error Notification first time 6. Add ACPI_HEST_NOTIFY_RESERVED notification type Address Igor's comments(for ACPI part): 1. Add doc patch first which will describe how it's supposed to work between QEMU/firmware/guest OS with expected flows. 2. Move APEI diagrams into doc/spec patch 3. Remove redundant g_malloc in ghes_record_cper() 4. Use build_append_int_noprefix() API to compose whole error status block and whole APEI table, and try to get rid of most structures in patch 1, as they will be left unused after that 5. Reuse something like https://github.com/imammedo/qemu/commit/3d2fd6d13a3ea298d2ee814835495ce6241d085c to build GAS 6. Remove much offsetof() in the function 7. Build independent tables first and only then build dependent tables passing to it pointers to previously build table if necessary. 8. Redefine macro GHES_ACPI_HEST_NOTIFY_RESERVED to ACPI_HEST_ERROR_SOURCE_COUNT to avoid confusion Address Peter Maydell's comments 1. linux-headers is done as a patch of their own created using scripts/update-linux-headers.sh run against a mainline kernel tree 2. Tested whether this patchset builds OK on aarch32 3. Abstract Hwpoison page adding code out properly into a cpu-independent source file from target/i386/kvm.c, such as kvm-all.c 4. Add doc-comment formatted documentation comment for new globally-visible function prototype in a header --- [1]: https://lkml.org/lkml/2017/2/27/246 https://patchwork.kernel.org/patch/9633105/ https://patchwork.kernel.org/patch/9925227/ [2]: Note: the UEFI(QEMU_EFI.fd) is needed if guest want to use ACPI table. After guest boot up, dump the APEI table, then can see the initialized table (1) # iasl -p ./HEST -d /sys/firmware/acpi/tables/HEST (2) # cat HEST.dsl /* * Intel ACPI Component Architecture * AML/ASL+ Disassembler version 20170728 (64-bit version) * Copyright (c) 2000 - 2017 Intel Corporation * * Disassembly of /sys/firmware/acpi/tables/HEST, Mon Sep 5 07:59:17 2016 * * ACPI Data Table [HEST] * * Format: [HexOffset DecimalOffset ByteLength] FieldName : FieldValue */ .................................................................................. [308h 0776 2] Subtable Type : 000A [Generic Hardware Error Source V2] [30Ah 0778 2] Source Id : 0001 [30Ch 0780 2] Related Source Id : FFFF [30Eh 0782 1] Reserved : 00 [30Fh 0783 1] Enabled : 01 [310h 0784 4] Records To Preallocate : 00000001 [314h 0788 4] Max Sections Per Record : 00000001 [318h 0792 4] Max Raw Data Length : 00001000 [31Ch 0796 12] Error Status Address : [Generic Address Structure] [31Ch 0796 1] Space ID : 00 [SystemMemory] [31Dh 0797 1] Bit Width : 40 [31Eh 0798 1] Bit Offset : 00 [31Fh 0799 1] Encoded Access Width : 04 [QWord Access:64] [320h 0800 8] Address : 00000000785D0040 [328h 0808 28] Notify : [Hardware Error Notification Structure] [328h 0808 1] Notify Type : 08 [SEA] [329h 0809 1] Notify Length : 1C [32Ah 0810 2] Configuration Write Enable : 0000 [32Ch 0812 4] PollInterval : 00000000 [330h 0816 4] Vector : 00000000 [334h 0820 4] Polling Threshold Value : 00000000 [338h 0824 4] Polling Threshold Window : 00000000 [33Ch 0828 4] Error Threshold Value : 00000000 [340h 0832 4] Error Threshold Window : 00000000 [344h 0836 4] Error Status Block Length : 00001000 [348h 0840 12] Read Ack Register : [Generic Address Structure] [348h 0840 1] Space ID : 00 [SystemMemory] [349h 0841 1] Bit Width : 40 [34Ah 0842 1] Bit Offset : 00 [34Bh 0843 1] Encoded Access Width : 04 [QWord Access:64] [34Ch 0844 8] Address : 00000000785D0098 [354h 0852 8] Read Ack Preserve : 00000000FFFFFFFE [35Ch 0860 8] Read Ack Write : 0000000000000001 ..................................................................................... (3) After a synchronous external abort(SEA) happen, Qemu receive a SIGBUS and filled the CPER into guest GHES memory. For example, according to above table, the address that contains the physical address of a block of memory that holds the error status data is 0x00000000785D0040 (4) the address of error source which is SEA notification type is 0x785d80b0 (qemu) xp /1 0x00000000785D0040 00000000785d0040: 0x785d80b0 (5) check the content of generic error status block and generic error data entry (qemu) xp /100x 0x785d80b0 00000000785d80b0: 0x00000001 0x00000000 0x00000000 0x00000098 00000000785d80c0: 0x00000000 0xa5bc1114 0x4ede6f64 0x833e63b8 00000000785d80d0: 0xb1837ced 0x00000000 0x00000300 0x00000050 00000000785d80e0: 0x00000000 0x00000000 0x00000000 0x00000000 00000000785d80f0: 0x00000000 0x00000000 0x00000000 0x00000000 00000000785d8100: 0x00000000 0x00000000 0x00000000 0x00004002 (6) check the OSPM's ACK value(for example SEA) /* Before OSPM acknowledges the error, check the ACK value */ (qemu) xp /1 0x00000000785D0098 00000000785d00f0: 0x00000000 /* After OSPM acknowledges the error, check the ACK value, it change to 1 from 0 */ (qemu) xp /1 0x00000000785D0098 00000000785d00f0: 0x00000001 [3]: KVM deliver "BUS_MCEERR_AR" to Qemu, Qemu record the guest CPER and inject synchronous external abort to notify guest, then guest do the recovery. [ 1552.516170] Synchronous External Abort: synchronous external abort (0x92000410) at 0x000000003751c6b4 [ 1553.074073] {1}[Hardware Error]: Hardware error from APEI Generic Hardware Error Source: 8 [ 1553.081654] {1}[Hardware Error]: event severity: recoverable [ 1554.034191] {1}[Hardware Error]: Error 0, type: recoverable [ 1554.037934] {1}[Hardware Error]: section_type: memory error [ 1554.513261] {1}[Hardware Error]: physical_address: 0x0000000040fa6000 [ 1554.513944] {1}[Hardware Error]: error_type: 0, unknown [ 1555.041451] Memory failure: 0x40fa6: Killing mca-recover:1296 due to hardware memory corruption [ 1555.373116] Memory failure: 0x40fa6: recovery action for dirty LRU page: Recovered Dongjiu Geng (10): acpi: nvdimm: change NVDIMM_UUID_LE to a common macro hw/arm/virt: Introduce a RAS machine option docs: APEI GHES generation and CPER record description ACPI: Build related register address fields via hardware error fw_cfg blob ACPI: Build Hardware Error Source Table ACPI: Record the Generic Error Status Block address KVM: Move hwpoison page related functions into kvm-all.c ACPI: Record Generic Error Status Block(GESB) table target-arm: kvm64: handle SIGBUS signal from kernel or KVM MAINTAINERS: Add ACPI/HEST/GHES entries MAINTAINERS | 9 + accel/kvm/kvm-all.c | 36 +++ default-configs/arm-softmmu.mak | 1 + docs/specs/acpi_hest_ghes.rst | 110 ++++++++ docs/specs/index.rst | 1 + hw/acpi/Kconfig | 4 + hw/acpi/Makefile.objs | 1 + hw/acpi/aml-build.c | 2 + hw/acpi/generic_event_device.c | 18 ++ hw/acpi/ghes.c | 450 +++++++++++++++++++++++++++++++++ hw/acpi/nvdimm.c | 8 +- hw/arm/virt-acpi-build.c | 17 ++ hw/arm/virt.c | 23 ++ include/hw/acpi/aml-build.h | 1 + include/hw/acpi/generic_event_device.h | 2 + include/hw/acpi/ghes.h | 74 ++++++ include/hw/arm/virt.h | 1 + include/qemu/uuid.h | 5 + include/sysemu/kvm.h | 3 +- include/sysemu/kvm_int.h | 12 + target/arm/cpu.h | 4 + target/arm/helper.c | 2 +- target/arm/internals.h | 5 +- target/arm/kvm64.c | 73 ++++++ target/arm/tlb_helper.c | 2 +- target/i386/cpu.h | 2 + target/i386/kvm.c | 36 --- 27 files changed, 854 insertions(+), 48 deletions(-) create mode 100644 docs/specs/acpi_hest_ghes.rst create mode 100644 hw/acpi/ghes.c create mode 100644 include/hw/acpi/ghes.h