[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [xen master] x86/crash: Drop manual hooking of exception_table[]
commit e7f147bf4ac725492962a501da72f5ab6be682db Author: Andrew Cooper <andrew.cooper3@xxxxxxxxxx> AuthorDate: Thu Oct 7 14:02:10 2021 +0100 Commit: Andrew Cooper <andrew.cooper3@xxxxxxxxxx> CommitDate: Mon Nov 29 13:53:05 2021 +0000 x86/crash: Drop manual hooking of exception_table[] NMI hooking in the crash path has undergone several revisions since its introduction. What we have now is not sufficiently different from the regular nmi_callback() mechanism to warrant special casing. Use set_nmi_callback() directly, and do away with patching a read-only data structure via a read-write alias. This also means that the vmx_vmexit_handler() can and should call do_nmi() directly, rather than indirecting through the exception table to pick up the crash path hook. Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx> Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx> --- xen/arch/x86/crash.c | 15 ++------------- xen/arch/x86/hvm/vmx/vmx.c | 2 +- xen/arch/x86/traps.c | 5 +++++ 3 files changed, 8 insertions(+), 14 deletions(-) diff --git a/xen/arch/x86/crash.c b/xen/arch/x86/crash.c index 0611b4fb9b..f6264946a6 100644 --- a/xen/arch/x86/crash.c +++ b/xen/arch/x86/crash.c @@ -36,10 +36,8 @@ static unsigned int crashing_cpu; static DEFINE_PER_CPU_READ_MOSTLY(bool, crash_save_done); /* This becomes the NMI handler for non-crashing CPUs, when Xen is crashing. */ -static void noreturn do_nmi_crash(const struct cpu_user_regs *regs) +static int noreturn do_nmi_crash(const struct cpu_user_regs *regs, int cpu) { - unsigned int cpu = smp_processor_id(); - stac(); /* nmi_shootdown_cpus() should ensure that this assertion is correct. */ @@ -136,16 +134,7 @@ static void nmi_shootdown_cpus(void) SYS_DESC_irq_gate, 0, &trap_nop); set_ist(&idt_tables[cpu][TRAP_machine_check], IST_NONE); - /* - * Ideally would be: - * exception_table[TRAP_nmi] = &do_nmi_crash; - * - * but the exception_table is read only. Access it via its directmap - * mappings. - */ - write_atomic((unsigned long *)__va(__pa(&exception_table[TRAP_nmi])), - (unsigned long)&do_nmi_crash); - + set_nmi_callback(do_nmi_crash); smp_send_nmi_allbutself(); msecs = 1000; /* Wait at most a second for the other cpus to stop */ diff --git a/xen/arch/x86/hvm/vmx/vmx.c b/xen/arch/x86/hvm/vmx/vmx.c index d403e2d806..37c31c08b9 100644 --- a/xen/arch/x86/hvm/vmx/vmx.c +++ b/xen/arch/x86/hvm/vmx/vmx.c @@ -3887,7 +3887,7 @@ void vmx_vmexit_handler(struct cpu_user_regs *regs) ((intr_info & INTR_INFO_INTR_TYPE_MASK) == MASK_INSR(X86_EVENTTYPE_NMI, INTR_INFO_INTR_TYPE_MASK)) ) { - exception_table[TRAP_nmi](regs); + do_nmi(regs); enable_nmis(); } break; diff --git a/xen/arch/x86/traps.c b/xen/arch/x86/traps.c index 3bc4cd5e1e..6d7d88c3c4 100644 --- a/xen/arch/x86/traps.c +++ b/xen/arch/x86/traps.c @@ -1883,6 +1883,11 @@ void do_nmi(const struct cpu_user_regs *regs) this_cpu(nmi_count)++; nmi_enter(); + /* + * Think carefully before putting any logic before this point. + * nmi_callback() might be the crash quiesce... + */ + callback = ACCESS_ONCE(nmi_callback); if ( unlikely(callback) && callback(regs, cpu) ) goto out; -- generated by git-patchbot for /home/xen/git/xen.git#master
|
![]() |
Lists.xenproject.org is hosted with RackSpace, monitoring our |