kernel.git/arch/arm64/kernel/smp.c, branch linux-4.19.y

arm64: smp: fix crash_smp_send_stop() behaviour

2020-03-25T07:06:15Z

commit f50b7dacccbab2b9e3ef18f52a6dcc18ed2050b9 upstream. On a system configured to trigger a crash_kexec() reboot, when only one CPU is online and another CPU panics while starting-up, crash_smp_send_stop() will fail to send any STOP message to the other already online core, resulting in fail to freeze and registers not properly saved. Moreover even if the proper messages are sent (case CPUs > 2) it will similarly fail to account for the booting CPU when executing the final stop wait-loop, so potentially resulting in some CPU not been waited for shutdown before rebooting. A tangible effect of this behaviour can be observed when, after a panic with kexec enabled and loaded, on the following reboot triggered by kexec, the cpu that could not be successfully stopped fails to come back online: [ 362.291022] ------------[ cut here ]------------ [ 362.291525] kernel BUG at arch/arm64/kernel/cpufeature.c:886! [ 362.292023] Internal error: Oops - BUG: 0 [#1] PREEMPT SMP [ 362.292400] Modules linked in: [ 362.292970] CPU: 3 PID: 0 Comm: swapper/3 Kdump: loaded Not tainted 5.6.0-rc4-00003-gc780b890948a #105 [ 362.293136] Hardware name: Foundation-v8A (DT) [ 362.293382] pstate: 200001c5 (nzCv dAIF -PAN -UAO) [ 362.294063] pc : has_cpuid_feature+0xf0/0x348 [ 362.294177] lr : verify_local_elf_hwcaps+0x84/0xe8 [ 362.294280] sp : ffff800011b1bf60 [ 362.294362] x29: ffff800011b1bf60 x28: 0000000000000000 [ 362.294534] x27: 0000000000000000 x26: 0000000000000000 [ 362.294631] x25: 0000000000000000 x24: ffff80001189a25c [ 362.294718] x23: 0000000000000000 x22: 0000000000000000 [ 362.294803] x21: ffff8000114aa018 x20: ffff800011156a00 [ 362.294897] x19: ffff800010c944a0 x18: 0000000000000004 [ 362.294987] x17: 0000000000000000 x16: 0000000000000000 [ 362.295073] x15: 00004e53b831ae3c x14: 00004e53b831ae3c [ 362.295165] x13: 0000000000000384 x12: 0000000000000000 [ 362.295251] x11: 0000000000000000 x10: 00400032b5503510 [ 362.295334] x9 : 0000000000000000 x8 : ffff800010c7e204 [ 362.295426] x7 : 00000000410fd0f0 x6 : 0000000000000001 [ 362.295508] x5 : 00000000410fd0f0 x4 : 0000000000000000 [ 362.295592] x3 : 0000000000000000 x2 : ffff8000100939d8 [ 362.295683] x1 : 0000000000180420 x0 : 0000000000180480 [ 362.296011] Call trace: [ 362.296257] has_cpuid_feature+0xf0/0x348 [ 362.296350] verify_local_elf_hwcaps+0x84/0xe8 [ 362.296424] check_local_cpu_capabilities+0x44/0x128 [ 362.296497] secondary_start_kernel+0xf4/0x188 [ 362.296998] Code: 52805001 72a00301 6b01001f 54000ec0 (d4210000) [ 362.298652] SMP: stopping secondary CPUs [ 362.300615] Starting crashdump kernel... [ 362.301168] Bye! [ 0.000000] Booting Linux on physical CPU 0x0000000003 [0x410fd0f0] [ 0.000000] Linux version 5.6.0-rc4-00003-gc780b890948a (crimar01@e120937-lin) (gcc version 8.3.0 (GNU Toolchain for the A-profile Architecture 8.3-2019.03 (arm-rel-8.36))) #105 SMP PREEMPT Fri Mar 6 17:00:42 GMT 2020 [ 0.000000] Machine model: Foundation-v8A [ 0.000000] earlycon: pl11 at MMIO 0x000000001c090000 (options '') [ 0.000000] printk: bootconsole [pl11] enabled ..... [ 0.138024] rcu: Hierarchical SRCU implementation. [ 0.153472] its@2f020000: unable to locate ITS domain [ 0.154078] its@2f020000: Unable to locate ITS domain [ 0.157541] EFI services will not be available. [ 0.175395] smp: Bringing up secondary CPUs ... [ 0.209182] psci: failed to boot CPU1 (-22) [ 0.209377] CPU1: failed to boot: -22 [ 0.274598] Detected PIPT I-cache on CPU2 [ 0.278707] GICv3: CPU2: found redistributor 1 region 0:0x000000002f120000 [ 0.285212] CPU2: Booted secondary processor 0x0000000001 [0x410fd0f0] [ 0.369053] Detected PIPT I-cache on CPU3 [ 0.372947] GICv3: CPU3: found redistributor 2 region 0:0x000000002f140000 [ 0.378664] CPU3: Booted secondary processor 0x0000000002 [0x410fd0f0] [ 0.401707] smp: Brought up 1 node, 3 CPUs [ 0.404057] SMP: Total of 3 processors activated. Make crash_smp_send_stop() account also for the online status of the calling CPU while evaluating how many CPUs are effectively online: this way the right number of STOPs is sent and all other stopped-cores's registers are properly saved. Fixes: 78fd584cdec05 ("arm64: kdump: implement machine_crash_shutdown()") Acked-by: Mark Rutland Signed-off-by: Cristian Marussi Signed-off-by: Will Deacon Signed-off-by: Greg Kroah-Hartman

arm64: smp: fix smp_send_stop() behaviour

2020-03-25T07:06:14Z

commit d0bab0c39e32d39a8c5cddca72e5b4a3059fe050 upstream. On a system with only one CPU online, when another one CPU panics while starting-up, smp_send_stop() will fail to send any STOP message to the other already online core, resulting in a system still responsive and alive at the end of the panic procedure. [ 186.700083] CPU3: shutdown [ 187.075462] CPU2: shutdown [ 187.162869] CPU1: shutdown [ 188.689998] ------------[ cut here ]------------ [ 188.691645] kernel BUG at arch/arm64/kernel/cpufeature.c:886! [ 188.692079] Internal error: Oops - BUG: 0 [#1] PREEMPT SMP [ 188.692444] Modules linked in: [ 188.693031] CPU: 3 PID: 0 Comm: swapper/3 Not tainted 5.6.0-rc4-00001-g338d25c35a98 #104 [ 188.693175] Hardware name: Foundation-v8A (DT) [ 188.693492] pstate: 200001c5 (nzCv dAIF -PAN -UAO) [ 188.694183] pc : has_cpuid_feature+0xf0/0x348 [ 188.694311] lr : verify_local_elf_hwcaps+0x84/0xe8 [ 188.694410] sp : ffff800011b1bf60 [ 188.694536] x29: ffff800011b1bf60 x28: 0000000000000000 [ 188.694707] x27: 0000000000000000 x26: 0000000000000000 [ 188.694801] x25: 0000000000000000 x24: ffff80001189a25c [ 188.694905] x23: 0000000000000000 x22: 0000000000000000 [ 188.694996] x21: ffff8000114aa018 x20: ffff800011156a38 [ 188.695089] x19: ffff800010c944a0 x18: 0000000000000004 [ 188.695187] x17: 0000000000000000 x16: 0000000000000000 [ 188.695280] x15: 0000249dbde5431e x14: 0262cbe497efa1fa [ 188.695371] x13: 0000000000000002 x12: 0000000000002592 [ 188.695472] x11: 0000000000000080 x10: 00400032b5503510 [ 188.695572] x9 : 0000000000000000 x8 : ffff800010c80204 [ 188.695659] x7 : 00000000410fd0f0 x6 : 0000000000000001 [ 188.695750] x5 : 00000000410fd0f0 x4 : 0000000000000000 [ 188.695836] x3 : 0000000000000000 x2 : ffff8000100939d8 [ 188.695919] x1 : 0000000000180420 x0 : 0000000000180480 [ 188.696253] Call trace: [ 188.696410] has_cpuid_feature+0xf0/0x348 [ 188.696504] verify_local_elf_hwcaps+0x84/0xe8 [ 188.696591] check_local_cpu_capabilities+0x44/0x128 [ 188.696666] secondary_start_kernel+0xf4/0x188 [ 188.697150] Code: 52805001 72a00301 6b01001f 54000ec0 (d4210000) [ 188.698639] ---[ end trace 3f12ca47652f7b72 ]--- [ 188.699160] Kernel panic - not syncing: Attempted to kill the idle task! [ 188.699546] Kernel Offset: disabled [ 188.699828] CPU features: 0x00004,20c02008 [ 188.700012] Memory Limit: none [ 188.700538] ---[ end Kernel panic - not syncing: Attempted to kill the idle task! ]--- [root@arch ~]# echo Helo Helo [root@arch ~]# cat /proc/cpuinfo | grep proce processor : 0 Make smp_send_stop() account also for the online status of the calling CPU while evaluating how many CPUs are effectively online: this way, the right number of STOPs is sent, so enforcing a proper freeze of the system at the end of panic even under the above conditions. Fixes: 08e875c16a16c ("arm64: SMP support") Reported-by: Dave Martin Acked-by: Mark Rutland Signed-off-by: Cristian Marussi Signed-off-by: Will Deacon Signed-off-by: Greg Kroah-Hartman

arm64: smp: Handle errors reported by the firmware

2019-12-05T08:19:59Z

[ Upstream commit f357b3a7e17af7736d67d8267edc1ed3d1dd9391 ] The __cpu_up() routine ignores the errors reported by the firmware for a CPU bringup operation and looks for the error status set by the booting CPU. If the CPU never entered the kernel, we could end up in assuming stale error status, which otherwise would have been set/cleared appropriately by the booting CPU. Reported-by: Steve Capper Cc: Will Deacon Signed-off-by: Suzuki K Poulose Signed-off-by: Will Deacon Signed-off-by: Sasha Levin

arm64: mm: Prevent mismatched 52-bit VA support

2019-12-05T08:19:59Z

[ Upstream commit a96a33b1ca57dbea4285893dedf290aeb8eb090b ] For cases where there is a mismatch in ARMv8.2-LVA support between CPUs we have to be careful in allowing secondary CPUs to boot if 52-bit virtual addresses have already been enabled on the boot CPU. This patch adds code to the secondary startup path. If the boot CPU has enabled 52-bit VAs then ID_AA64MMFR2_EL1 is checked to see if the secondary can also enable 52-bit support. If not, the secondary is prevented from booting and an error message is displayed indicating why. Technically this patch could be implemented using the cpufeature code when considering 52-bit userspace support. However, we employ low level checks here as the cpufeature code won't be able to run if we have mismatched 52-bit kernel va support. Signed-off-by: Steve Capper Signed-off-by: Will Deacon Signed-off-by: Sasha Levin

arm64: numa: rework ACPI NUMA initialization

2018-07-09T17:21:40Z

Current ACPI ARM64 NUMA initialization code in acpi_numa_gicc_affinity_init() carries out NUMA nodes creation and cpu<->node mappings at the same time in the arch backend so that a single SRAT walk is needed to parse both pieces of information. This implies that the cpu<->node mappings must be stashed in an array (sized NR_CPUS) so that SMP code can later use the stashed values to avoid another SRAT table walk to set-up the early cpu<->node mappings. If the kernel is configured with a NR_CPUS value less than the actual processor entries in the SRAT (and MADT), the logic in acpi_numa_gicc_affinity_init() is broken in that the cpu<->node mapping is only carried out (and stashed for future use) only for a number of SRAT entries up to NR_CPUS, which do not necessarily correspond to the possible cpus detected at SMP initialization in acpi_map_gic_cpu_interface() (ie MADT and SRAT processor entries order is not enforced), which leaves the kernel with broken cpu<->node mappings. Furthermore, given the current ACPI NUMA code parsing logic in acpi_numa_gicc_affinity_init(), PXM domains for CPUs that are not parsed because they exceed NR_CPUS entries are not mapped to NUMA nodes (ie the PXM corresponding node is not created in the kernel) leaving the system with a broken NUMA topology. Rework the ACPI ARM64 NUMA initialization process so that the NUMA nodes creation and cpu<->node mappings are decoupled. cpu<->node mappings are moved to SMP initialization code (where they are needed), at the cost of an extra SRAT walk so that ACPI NUMA mappings can be batched before being applied, fixing current parsing pitfalls. Acked-by: Hanjun Guo Tested-by: John Garry Fixes: d8b47fca8c23 ("arm64, ACPI, NUMA: NUMA support based on SRAT and SLIT") Link: http://lkml.kernel.org/r/1527768879-88161-2-git-send-email-xiexiuqi@huawei.com Reported-by: Xie XiuQi Signed-off-by: Lorenzo Pieralisi Cc: Punit Agrawal Cc: Jonathan Cameron Cc: Will Deacon Cc: Hanjun Guo Cc: Ganapatrao Kulkarni Cc: Jeremy Linton Cc: Catalin Marinas Cc: Xie XiuQi Signed-off-by: Will Deacon

arm64: smp: remove cpu and numa topology information when hotplugging out CPU

2018-07-06T12:18:18Z

We already repopulate the information on CPU hotplug-in, so we can safely remove the CPU topology and NUMA cpumap information during CPU hotplug out operation. This will help to provide the correct cpumask for scheduler domains. Cc: Catalin Marinas Cc: Will Deacon Tested-by: Ganapatrao Kulkarni Tested-by: Hanjun Guo Signed-off-by: Sudeep Holla Signed-off-by: Will Deacon

arm64: numa: separate out updates to percpu nodeid and NUMA node cpumap

2018-07-06T12:18:18Z

Currently numa_clear_node removes both cpu information from the NUMA node cpumap as well as the NUMA node id from the cpu. Similarly numa_store_cpu_info updates both percpu nodeid and NUMA cpumap. However we need to retain the numa node id for the cpu and only remove the cpu information from the numa node cpumap during CPU hotplug out. The same can be extended for hotplugging in the CPU. This patch separates out numa_{add,remove}_cpu from numa_clear_node and numa_store_cpu_info. Cc: Catalin Marinas Cc: Will Deacon Reviewed-by: Ganapatrao Kulkarni Tested-by: Ganapatrao Kulkarni Tested-by: Hanjun Guo Signed-off-by: Sudeep Holla Signed-off-by: Will Deacon

arm64: make secondary_start_kernel() notrace

2018-06-19T13:19:24Z

We can't call function trace hook before setup percpu offset. When entering secondary_start_kernel(), percpu offset has not been initialized. So this lead hotplug malfunction. Here is the flow to reproduce this bug: echo 0 > /sys/devices/system/cpu/cpu1/online echo function > /sys/kernel/debug/tracing/current_tracer echo 1 > /sys/kernel/debug/tracing/tracing_on echo 1 > /sys/devices/system/cpu/cpu1/online Acked-by: Mark Rutland Tested-by: Suzuki K Poulose Signed-off-by: Zhizhou Zhang Signed-off-by: Catalin Marinas

arm64: capabilities: Change scope of VHE to Boot CPU feature

2018-03-26T17:01:41Z

We expect all CPUs to be running at the same EL inside the kernel with or without VHE enabled and we have strict checks to ensure that any mismatch triggers a kernel panic. If VHE is enabled, we use the feature based on the boot CPU and all other CPUs should follow. This makes it a perfect candidate for a capability based on the boot CPU, which should be matched by all the CPUs (both when is ON and OFF). This saves us some not-so-pretty hooks and special code, just for verifying the conflict. The patch also makes the VHE capability entry depend on CONFIG_ARM64_VHE. Cc: Marc Zyngier Cc: Will Deacon Reviewed-by: Dave Martin Signed-off-by: Suzuki K Poulose Signed-off-by: Will Deacon

arm64: capabilities: Move errata work around check on boot CPU

2018-03-26T17:01:36Z

We trigger CPU errata work around check on the boot CPU from smp_prepare_boot_cpu() to make sure that we run the checks only after the CPU feature infrastructure is initialised. While this is correct, we can also do this from init_cpu_features() which initilises the infrastructure, and is called only on the Boot CPU. This helps to consolidate the CPU capability handling to cpufeature.c. No functional changes. Cc: Will Deacon Cc: Catalin Marinas Cc: Mark Rutland Reviewed-by: Dave Martin Signed-off-by: Suzuki K Poulose Signed-off-by: Will Deacon