RE: [RFC][PATCH] objtool,x86_64,paravirt: Add pv_ops[] support

[Michael Kelley <mikelley@microsoft.com>]

From: peterz@infradead.org Sent: Monday, August 3, 2020 7:33 AM
> 
> Thomas wanted paramuck vs noinstr, here goes. Very rough, very nasty,
> mostly works.
> 
> It requires call sites are the normal indirect call, and not mangled
> retpoison (slow_down_io() had those), it also requires pv_ops[]
> assignments are single instructions and not laundered through some
> pointless intermediate helper (hyperv).
> 
> It doesn't warn when you get any of that wrong.
> 
> But if you have it all lined up right, it will warn about noinstr
> violations due to paramuck:
> 
> $ ./objtool check -afl ../../defconfig-build/vmlinux.o
> ../../defconfig-build/vmlinux.o: warning: objtool: xen_pv_evtchn_do_upcall()+0x69: return
> with instrumentation enabled
> ../../defconfig-build/vmlinux.o: warning: objtool: exc_double_fault()+0x15: call pv_ops[48]
> from noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: pv_op[48]: xen_read_cr2
> ../../defconfig-build/vmlinux.o: warning: objtool: exc_double_fault()+0x15: call to
> {dynamic}() leaves .noinstr.text section
> ../../defconfig-build/vmlinux.o: warning: objtool: exc_debug()+0x29: call pv_ops[4] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: pv_op[4]: native_get_debugreg
> ../../defconfig-build/vmlinux.o: warning: objtool: exc_debug()+0x29: call to {dynamic}()
> leaves .noinstr.text section
> ../../defconfig-build/vmlinux.o: warning: objtool: noist_exc_debug()+0x26: call pv_ops[4]
> from noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: pv_op[4]: native_get_debugreg
> ../../defconfig-build/vmlinux.o: warning: objtool: noist_exc_debug()+0x26: call to
> {dynamic}() leaves .noinstr.text section
> ../../defconfig-build/vmlinux.o: warning: objtool: exc_nmi()+0x3a: call pv_ops[48] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: pv_op[48]: xen_read_cr2
> ../../defconfig-build/vmlinux.o: warning: objtool: exc_nmi()+0x3a: call to {dynamic}() leaves
> .noinstr.text section
> ../../defconfig-build/vmlinux.o: warning: objtool: mce_setup()+0x80: call pv_ops[25] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: pv_op[25]: native_cpuid
> ../../defconfig-build/vmlinux.o: warning: objtool: mce_setup()+0x80: call to {dynamic}()
> leaves .noinstr.text section
> ../../defconfig-build/vmlinux.o: warning: objtool: do_machine_check()+0x46: call to
> mce_rdmsrl() leaves .noinstr.text section
> ../../defconfig-build/vmlinux.o: warning: objtool: exc_machine_check()+0x24: call
> pv_ops[4] from noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: pv_op[4]: native_get_debugreg
> ../../defconfig-build/vmlinux.o: warning: objtool: exc_machine_check()+0x24: call to
> {dynamic}() leaves .noinstr.text section
> ../../defconfig-build/vmlinux.o: warning: objtool: noist_exc_machine_check()+0x24: call
> pv_ops[4] from noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: pv_op[4]: native_get_debugreg
> ../../defconfig-build/vmlinux.o: warning: objtool: noist_exc_machine_check()+0x24: call to
> {dynamic}() leaves .noinstr.text section
> ../../defconfig-build/vmlinux.o: warning: objtool: exc_page_fault()+0xe: call pv_ops[48]
> from noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: pv_op[48]: xen_read_cr2
> ../../defconfig-build/vmlinux.o: warning: objtool: exc_page_fault()+0xe: call to {dynamic}()
> leaves .noinstr.text section
> ../../defconfig-build/vmlinux.o: warning: objtool: .entry.text+0xf17: call pv_ops[32] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: .entry.text+0x12f1: call pv_ops[35] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: .entry.text+0xf5a: call pv_ops[32] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: .entry.text+0x1002: call pv_ops[32] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: .entry.text+0x1186: call pv_ops[32] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: .entry.text+0x12ba: call pv_ops[32] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: .entry.text+0x12a7: call pv_ops[32] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: .entry.text+0x124c: call pv_ops[32] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: .entry.text+0x14f0: call pv_ops[32] from
> noinstr
> ../../defconfig-build/vmlinux.o: warning: objtool: .entry.text+0x1683: call pv_ops[32] from
> noinstr
> 
> [ funny that, WTH does x86_64 have IRET pv_ops... ]
> 
> Requested-by: Thomas Gleixner <tglx@linutronix.de>
> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> ---
>  arch/x86/include/asm/paravirt.h    |   8 +--
>  arch/x86/kernel/cpu/mshyperv.c     |   7 ---
>  drivers/clocksource/hyperv_timer.c |   7 +++
>  include/asm-generic/mshyperv.h     |   1 -
>  lib/Kconfig.debug                  |   2 +-
>  tools/objtool/Makefile             |   2 +-
>  tools/objtool/arch.h               |   2 +-
>  tools/objtool/arch/x86/decode.c    |  30 ++++++++-
>  tools/objtool/builtin-check.c      |  10 +--
>  tools/objtool/builtin-orc.c        |  10 +--
>  tools/objtool/check.c              | 126 ++++++++++++++++++++++++++++++++++---
>  tools/objtool/elf.h                |   1 +
>  tools/objtool/objtool.c            |  46 ++++++++++++++
>  tools/objtool/objtool.h            |   9 +++
>  14 files changed, 218 insertions(+), 43 deletions(-)
> 
> diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h
> index 3d2afecde50c..db30b35f0d59 100644
> --- a/arch/x86/include/asm/paravirt.h
> +++ b/arch/x86/include/asm/paravirt.h
> @@ -39,11 +39,11 @@ static inline u64 paravirt_steal_clock(int cpu)
>  /* The paravirtualized I/O functions */
>  static inline void slow_down_io(void)
>  {
> -	pv_ops.cpu.io_delay();
> +	PVOP_VCALL0(cpu.io_delay);
>  #ifdef REALLY_SLOW_IO
> -	pv_ops.cpu.io_delay();
> -	pv_ops.cpu.io_delay();
> -	pv_ops.cpu.io_delay();
> +	PVOP_VCALL0(cpu.io_delay);
> +	PVOP_VCALL0(cpu.io_delay);
> +	PVOP_VCALL0(cpu.io_delay);
>  #endif
>  }
> 
> diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
> index af94f05a5c66..31125448b174 100644
> --- a/arch/x86/kernel/cpu/mshyperv.c
> +++ b/arch/x86/kernel/cpu/mshyperv.c
> @@ -361,13 +361,6 @@ static void __init ms_hyperv_init_platform(void)
>  #endif
>  }
> 
> -void hv_setup_sched_clock(void *sched_clock)
> -{
> -#ifdef CONFIG_PARAVIRT
> -	pv_ops.time.sched_clock = sched_clock;
> -#endif
> -}
> -
>  const __initconst struct hypervisor_x86 x86_hyper_ms_hyperv = {
>  	.name			= "Microsoft Hyper-V",
>  	.detect			= ms_hyperv_platform,
> diff --git a/drivers/clocksource/hyperv_timer.c b/drivers/clocksource/hyperv_timer.c
> index 09aa44cb8a91..bed769a84637 100644
> --- a/drivers/clocksource/hyperv_timer.c
> +++ b/drivers/clocksource/hyperv_timer.c
> @@ -418,6 +418,13 @@ static struct clocksource hyperv_cs_msr = {
>  	.flags	= CLOCK_SOURCE_IS_CONTINUOUS,
>  };
> 
> +static __always_inline void hv_setup_sched_clock(void *sched_clock)
> +{
> +#ifdef CONFIG_PARAVIRT
> +	pv_ops.time.sched_clock = sched_clock;
> +#endif
> +}
> +

Moving this function to hyperv_timer.c is problematic because we
want hyperv_timer.c to be architecture independent.  We're in the
process of supporting Linux guests running on Hyper-V on ARM64, and
there's a different implementation of hv_setup_sched_clock() on ARM64.

Michael


>  static bool __init hv_init_tsc_clocksource(void)
>  {
>  	u64		tsc_msr;
> diff --git a/include/asm-generic/mshyperv.h b/include/asm-generic/mshyperv.h
> index 1c4fd950f091..c5edc5e08b94 100644
> --- a/include/asm-generic/mshyperv.h
> +++ b/include/asm-generic/mshyperv.h
> @@ -168,7 +168,6 @@ void hyperv_report_panic_msg(phys_addr_t pa, size_t size);
>  bool hv_is_hyperv_initialized(void);
>  bool hv_is_hibernation_supported(void);
>  void hyperv_cleanup(void);
> -void hv_setup_sched_clock(void *sched_clock);
>  #else /* CONFIG_HYPERV */
>  static inline bool hv_is_hyperv_initialized(void) { return false; }
>  static inline bool hv_is_hibernation_supported(void) { return false; }