Star64_linux/arch/x86/kernel
Thomas Gleixner 7d65f9e806 x86/apic: Mark _all_ legacy interrupts when IO/APIC is missing
PIC interrupts do not support affinity setting and they can end up on
any online CPU. Therefore, it's required to mark the associated vectors
as system-wide reserved. Otherwise, the corresponding irq descriptors
are copied to the secondary CPUs but the vectors are not marked as
assigned or reserved. This works correctly for the IO/APIC case.

When the IO/APIC is disabled via config, kernel command line or lack of
enumeration then all legacy interrupts are routed through the PIC, but
nothing marks them as system-wide reserved vectors.

As a consequence, a subsequent allocation on a secondary CPU can result in
allocating one of these vectors, which triggers the BUG() in
apic_update_vector() because the interrupt descriptor slot is not empty.

Imran tried to work around that by marking those interrupts as allocated
when a CPU comes online. But that's wrong in case that the IO/APIC is
available and one of the legacy interrupts, e.g. IRQ0, has been switched to
PIC mode because then marking them as allocated will fail as they are
already marked as system vectors.

Stay consistent and update the legacy vectors after attempting IO/APIC
initialization and mark them as system vectors in case that no IO/APIC is
available.

Fixes: 69cde0004a ("x86/vector: Use matrix allocator for vector assignment")
Reported-by: Imran Khan <imran.f.khan@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20210519233928.2157496-1-imran.f.khan@oracle.com
2021-05-29 11:41:14 +02:00
..
acpi Trivial cleanups and fixes all over the place. 2021-04-26 09:25:47 -07:00
apic x86/apic: Mark _all_ legacy interrupts when IO/APIC is missing 2021-05-29 11:41:14 +02:00
cpu - Enable -Wundef for the compressed kernel build stage 2021-05-16 09:31:06 -07:00
fpu x86: Fix various typos in comments 2021-03-18 15:31:53 +01:00
kprobes - turn the stack canary into a normal __percpu variable on 32-bit which 2021-04-27 17:45:09 -07:00
.gitignore
alternative.c The x86 MM changes in this cycle were: 2021-04-29 11:41:43 -07:00
amd_gart_64.c dma-mapping: split <linux/dma-mapping.h> 2020-10-06 07:07:03 +02:00
amd_nb.c x86: Fix various typos in comments, take #2 2021-03-21 23:50:28 +01:00
aperture_64.c
apm_32.c x86: Fix various typos in comments, take #2 2021-03-21 23:50:28 +01:00
asm-offsets.c x86/paravirt: Switch iret pvops to ALTERNATIVE 2021-03-11 19:58:54 +01:00
asm-offsets_32.c x86/stackprotector/32: Make the canary into a regular percpu variable 2021-03-08 13:19:05 +01:00
asm-offsets_64.c x86/xen: Drop USERGS_SYSRET64 paravirt call 2021-02-10 12:32:07 +01:00
audit_64.c
bootflag.c
check.c
cpuid.c smp: Cleanup smp_call_function*() 2020-11-24 16:47:49 +01:00
crash.c Devicetree updates for v5.13: 2021-04-28 15:50:24 -07:00
crash_core_32.c
crash_core_64.c
crash_dump_32.c x86/crashdump/32: Simplify copy_oldmem_page() 2020-11-24 14:42:09 +01:00
crash_dump_64.c
devicetree.c x86/devicetree: Fix the ioapic interrupt type table 2020-10-28 20:26:24 +01:00
doublefault_32.c x86/stackprotector/32: Make the canary into a regular percpu variable 2021-03-08 13:19:05 +01:00
dumpstack.c - Remove all uses of TIF_IA32 and TIF_X32 and reclaim the two bits in the end 2020-12-14 13:45:26 -08:00
dumpstack_32.c
dumpstack_64.c x86/irq/64: Adjust the per CPU irq stack pointer by 8 2021-02-10 23:34:14 +01:00
e820.c Power management updates for 5.13-rc1 2021-04-26 15:10:25 -07:00
early-quirks.c x86/gpu: Add Alderlake-S stolen memory support 2021-01-26 06:55:09 -08:00
early_printk.c
ebda.c
eisa.c
espfix_64.c
ftrace.c x86: Remove dynamic NOP selection 2021-03-15 16:24:59 +01:00
ftrace_32.S
ftrace_64.S x86/ftrace: Support objtool vmlinux.o validation in ftrace_64.S 2021-01-26 11:33:02 -06:00
head32.c
head64.c x86/sev-es: Rename sev-es.{ch} to sev.{ch} 2021-05-10 07:40:27 +02:00
head_32.S x86/stackprotector/32: Make the canary into a regular percpu variable 2021-03-08 13:19:05 +01:00
head_64.S - Fix the vmlinux size check on 64-bit along with adding useful clarifications on the topic 2020-12-14 13:54:50 -08:00
hpet.c x86/hpet: Use irq_find_matching_fwspec() to find remapping irqdomain 2020-10-28 20:26:28 +01:00
hw_breakpoint.c x86/debug: Prevent data breakpoints on cpu_dr7 2021-02-05 20:13:12 +01:00
i8237.c
i8253.c
i8259.c
idt.c x86: Fix various typos in comments 2021-03-18 15:31:53 +01:00
io_delay.c
ioport.c
irq.c x86: Fix various typos in comments 2021-03-18 15:31:53 +01:00
irq_32.c softirq: Move do_softirq_own_stack() to generic asm header 2021-02-10 23:34:16 +01:00
irq_64.c x86/softirq/64: Inline do_softirq_own_stack() 2021-02-10 23:34:17 +01:00
irq_work.c
irqflags.S x86/pv: Rework arch_local_irq_restore() to not use popf 2021-02-10 12:36:45 +01:00
irqinit.c
itmt.c
jailhouse.c
jump_label.c x86: Remove dynamic NOP selection 2021-03-15 16:24:59 +01:00
kdebugfs.c
kexec-bzimage64.c x86: Use ELF fields defined in 'struct kimage' 2021-03-08 12:06:29 -07:00
kgdb.c x86: Fix various typos in comments 2021-03-18 15:31:53 +01:00
ksysfs.c
kvm.c x86/kvm: Unify kvm_pv_guest_cpu_reboot() with kvm_guest_cpu_offline() 2021-05-07 06:06:11 -04:00
kvmclock.c x86/kvm: Disable all PV features on crash 2021-05-07 06:06:10 -04:00
ldt.c x86/ldt: Use tlb_gather_mmu_fullmm() when freeing LDT page-tables 2021-01-29 20:02:29 +01:00
machine_kexec_32.c
machine_kexec_64.c Devicetree updates for v5.13: 2021-04-28 15:50:24 -07:00
Makefile x86/sev-es: Rename sev-es.{ch} to sev.{ch} 2021-05-10 07:40:27 +02:00
mmconf-fam10h_64.c x86/msr: Rename MSR_K8_SYSCFG to MSR_AMD64_SYSCFG 2021-05-10 07:51:38 +02:00
module.c x86/build: Treat R_386_PLT32 relocation as R_386_PC32 2021-01-28 12:24:06 +01:00
mpparse.c Surgery of the MSI interrupt handling to prepare the support of upcoming 2020-10-12 11:40:41 -07:00
msr.c x86/MSR: Filter MSR writes through X86_IOC_WRMSR_REGS ioctl too 2021-01-27 19:06:47 +01:00
nmi.c x86/sev-es: Rename sev-es.{ch} to sev.{ch} 2021-05-10 07:40:27 +02:00
nmi_selftest.c
paravirt-spinlocks.c x86/paravirt: Add new features for paravirt patching 2021-03-11 19:51:49 +01:00
paravirt.c The x86 MM changes in this cycle were: 2021-04-29 11:41:43 -07:00
pci-dma.c dma-mapping: move dma-debug.h to kernel/dma/ 2020-10-06 07:07:05 +02:00
pci-iommu_table.c x86: Remove definition of DEBUG 2021-01-15 08:23:10 +01:00
pci-swiotlb.c
pcspeaker.c
perf_regs.c - Remove all uses of TIF_IA32 and TIF_X32 and reclaim the two bits in the end 2020-12-14 13:45:26 -08:00
platform-quirks.c
pmem.c
probe_roms.c
process.c x86/process: setup io_threads more like normal user space threads 2021-05-05 17:47:41 -06:00
process.h
process_32.c
process_64.c x86/irq: Sanitize irq stack tracking 2021-02-10 23:34:13 +01:00
ptrace.c x86/ptrace: Clean up PTRACE_GETREGS/PTRACE_PUTREGS regset selection 2021-02-04 12:33:15 +01:00
pvclock.c x86: Fix various typos in comments 2021-03-18 15:31:53 +01:00
quirks.c x86, powerpc: Rename memcpy_mcsafe() to copy_mc_to_{user, kernel}() 2020-10-06 11:18:04 +02:00
reboot.c x86: 2021-02-21 13:31:43 -08:00
reboot_fixups_32.c
relocate_kernel_32.S x86: Fix various typos in comments, take #2 2021-03-21 23:50:28 +01:00
relocate_kernel_64.S x86: Fix various typos in comments, take #2 2021-03-21 23:50:28 +01:00
resource.c
rtc.c
setup.c - turn the stack canary into a normal __percpu variable on 32-bit which 2021-04-27 17:45:09 -07:00
setup_percpu.c x86/stackprotector/32: Make the canary into a regular percpu variable 2021-03-08 13:19:05 +01:00
sev-shared.c x86/sev-es: Invalidate the GHCB after completing VMGEXIT 2021-05-18 07:06:29 +02:00
sev.c x86/sev-es: Use __put_user()/__get_user() for data accesses 2021-05-19 18:45:37 +02:00
sev_verify_cbit.S x86/boot/compressed/64: Check SEV encryption in 64-bit boot-path 2020-10-29 18:06:52 +01:00
signal.c Merge branch 'linus' into x86/cleanups, to resolve conflict 2021-03-21 22:16:08 +01:00
signal_compat.c signal: Deliver all of the siginfo perf data in _perf 2021-05-18 16:20:54 -05:00
smp.c x86: Fix various typos in comments, take #2 2021-03-21 23:50:28 +01:00
smpboot.c x86, sched: Fix the AMD CPPC maximum performance value on certain AMD Ryzen generations 2021-05-13 12:10:24 +02:00
stacktrace.c stacktrace: Move documentation for arch_stack_walk_reliable() to header 2021-03-10 15:52:31 +01:00
static_call.c x86: Remove dynamic NOP selection 2021-03-15 16:24:59 +01:00
step.c entry: Ensure trap after single-step on system call return 2021-02-06 00:21:42 +01:00
sys_ia32.c x86: switch to kernel_clone() 2020-08-20 13:12:58 +02:00
sys_x86_64.c x86/mm: Refine mmap syscall implementation 2021-01-05 19:07:42 +01:00
sysfb.c
sysfb_efi.c x86: Fix various typos in comments 2021-03-18 15:31:53 +01:00
sysfb_simplefb.c
tboot.c x86/boot/tboot: Avoid Wstringop-overread-warning 2021-03-23 00:16:13 +01:00
time.c
tls.c x86/stackprotector/32: Make the canary into a regular percpu variable 2021-03-08 13:19:05 +01:00
tls.h
topology.c x86: Fix various typos in comments 2021-03-18 15:31:53 +01:00
trace_clock.c
tracepoint.c
traps.c - turn the stack canary into a normal __percpu variable on 32-bit which 2021-04-27 17:45:09 -07:00
tsc.c Trivial cleanups and fixes all over the place. 2021-04-26 09:25:47 -07:00
tsc_msr.c Misc fixes and small updates all around the place: 2020-08-15 10:38:03 -07:00
tsc_sync.c x86: Fix various typos in comments, take #2 2021-03-21 23:50:28 +01:00
umip.c - turn the stack canary into a normal __percpu variable on 32-bit which 2021-04-27 17:45:09 -07:00
unwind_frame.c
unwind_guess.c
unwind_orc.c x86/unwind/orc: Silence warnings caused by missing ORC data 2021-03-06 13:09:45 +01:00
uprobes.c x86/uprobes: Convert to insn_decode() 2021-03-15 12:05:03 +01:00
verify_cpu.S
vm86_32.c x86/vm86/32: Remove VM86_SCREEN_BITMAP support 2021-01-21 20:08:53 +01:00
vmlinux.lds.S x86/build: Fix vmlinux size check on 64-bit 2020-10-29 21:54:35 +01:00
vsmp_64.c
x86_init.c x86/apic: Support 15 bits of APIC ID in MSI where available 2020-10-28 20:26:29 +01:00