Re: [Xen-devel] [v3 04/12] x86/fsgsbase/64: Enable FSGSBASE instructions in the helper functions

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 26/10/2018 00:11, Andy Lutomirski wrote:
> On Thu, Oct 25, 2018 at 4:09 PM Andrew Cooper <andrew.cooper3@citrix.com> wrote:
>> On 25/10/2018 07:09, Juergen Gross wrote:
>>> On 24/10/2018 21:41, Andrew Cooper wrote:
>>>> On 24/10/18 20:16, Andy Lutomirski wrote:
>>>>> On Tue, Oct 23, 2018 at 11:43 AM Chang S. Bae <chang.seok.bae@intel.com> wrote:
>>>>>> The helper functions will switch on faster accesses to FSBASE and GSBASE
>>>>>> when the FSGSBASE feature is enabled.
>>>>>>
>>>>>> Accessing user GSBASE needs a couple of SWAPGS operations. It is avoidable
>>>>>> if the user GSBASE is saved at kernel entry, being updated as changes, and
>>>>>> restored back at kernel exit. However, it seems to spend more cycles for
>>>>>> savings and restorations. Little or no benefit was measured from
>>>>>> experiments.
>>>>>>
>>>>>> Signed-off-by: Chang S. Bae <chang.seok.bae@intel.com>
>>>>>> Reviewed-by: Andi Kleen <ak@linux.intel.com>
>>>>>> Cc: Any Lutomirski <luto@kernel.org>
>>>>>> Cc: H. Peter Anvin <hpa@zytor.com>
>>>>>> Cc: Thomas Gleixner <tglx@linutronix.de>
>>>>>> Cc: Ingo Molnar <mingo@kernel.org>
>>>>>> Cc: Dave Hansen <dave.hansen@linux.intel.com>
>>>>>> ---
>>>>>>  arch/x86/include/asm/fsgsbase.h | 17 +++----
>>>>>>  arch/x86/kernel/process_64.c    | 82 +++++++++++++++++++++++++++------
>>>>>>  2 files changed, 75 insertions(+), 24 deletions(-)
>>>>>>
>>>>>> diff --git a/arch/x86/include/asm/fsgsbase.h b/arch/x86/include/asm/fsgsbase.h
>>>>>> index b4d4509b786c..e500d771155f 100644
>>>>>> --- a/arch/x86/include/asm/fsgsbase.h
>>>>>> +++ b/arch/x86/include/asm/fsgsbase.h
>>>>>> @@ -57,26 +57,23 @@ static __always_inline void wrgsbase(unsigned long gsbase)
>>>>>>                         : "memory");
>>>>>>  }
>>>>>>
>>>>>> +#include <asm/cpufeature.h>
>>>>>> +
>>>>>>  /* Helper functions for reading/writing FS/GS base */
>>>>>>
>>>>>>  static inline unsigned long x86_fsbase_read_cpu(void)
>>>>>>  {
>>>>>>         unsigned long fsbase;
>>>>>>
>>>>>> -       rdmsrl(MSR_FS_BASE, fsbase);
>>>>>> +       if (static_cpu_has(X86_FEATURE_FSGSBASE))
>>>>>> +               fsbase = rdfsbase();
>>>>>> +       else
>>>>>> +               rdmsrl(MSR_FS_BASE, fsbase);
>>>>>>
>>>>>>         return fsbase;
>>>>>>  }
>>>>>>
>>>>>> -static inline unsigned long x86_gsbase_read_cpu_inactive(void)
>>>>>> -{
>>>>>> -       unsigned long gsbase;
>>>>>> -
>>>>>> -       rdmsrl(MSR_KERNEL_GS_BASE, gsbase);
>>>>>> -
>>>>>> -       return gsbase;
>>>>>> -}
>>>>>> -
>>>>>> +extern unsigned long x86_gsbase_read_cpu_inactive(void);
>>>>>>  extern void x86_fsbase_write_cpu(unsigned long fsbase);
>>>>>>  extern void x86_gsbase_write_cpu_inactive(unsigned long gsbase);
>>>>>>
>>>>>> diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
>>>>>> index 31b4755369f0..fcf18046c3d6 100644
>>>>>> --- a/arch/x86/kernel/process_64.c
>>>>>> +++ b/arch/x86/kernel/process_64.c
>>>>>> @@ -159,6 +159,36 @@ enum which_selector {
>>>>>>         GS
>>>>>>  };
>>>>>>
>>>>>> +/*
>>>>>> + * Interrupts are disabled here. Out of line to be protected from kprobes.
>>>>>> + */
>>>>>> +static noinline __kprobes unsigned long rd_inactive_gsbase(void)
>>>>>> +{
>>>>>> +       unsigned long gsbase, flags;
>>>>>> +
>>>>>> +       local_irq_save(flags);
>>>>>> +       native_swapgs();
>>>>>> +       gsbase = rdgsbase();
>>>>>> +       native_swapgs();
>>>>>> +       local_irq_restore(flags);
>>>>>> +
>>>>>> +       return gsbase;
>>>>>> +}
>>>>> Please fold this into its only caller and make *that* noinline.
>>>>>
>>>>> Also, this function, and its "write" equivalent, will access the
>>>>> *active* gsbase.  So it either needs to be fixed for Xen PV or some
>>>>> clear comment and careful auditing needs to be added to ensure that
>>>>> it's not used on Xen PV.  Or it needs to be renamed
>>>>> native_x86_fsgsbase_... and add paravirt hooks, since Xen PV allows a
>>>>> very efficient but different implementation, I think.  The latter is
>>>>> probably the right solution.
>>>>>
>>>>> (Hi Xen people -- how does CR4.FSGSBASE work on Xen?  Is it always
>>>>> set?  Never set?  Set only if the guest tries to set it?)
>>>> FML.  Seriously - whoever put this code into the hypervisor in the past
>>>> did an atrocious job.  After some experimentation, you're going to be
>>>> sad and I'm declaring this borderline unusable.
>>>>
>>>> Looks like Xen unconditionally enabled CR4.FSGSBASE if it is available.
>>>> Therefore, PV guests can use the instructions, even if the bit is clear
>>>> in vCR4.
>>>>
>>>> The CPUID bits are exposed to guests by default, and Xen will emulate
>>>> vCR4.FSGSBASE being set and cleared.
>>>>
>>>> We don't however emulate swapgs (which is a cpl0 instruction).  The
>>>> guest gets handed a #GP[0] instead.
>>>>
>>>> The Linux WRMSR PVop uses the set_segment_base() hypercall in instead of
>>>> going through the full wrmsr emulation path.
>>>>
>>>> There is no equivalent get hypercall, so the only way I can see of
>>>> getting the value is to actually read MSR_KERNEL_GS_BASE and take the
>>>> full rdmsr emulation path.
>>> Or shadow the value in a percpu variable.
>> Hmm true, so long as no paths try to use native_rd{fs,gs}base() to
>> bypass the PVop.
> But *user* code can change the base.  How is the kernel supposed to
> context-switch the user gsbase?

user code can change the user gs base.

Xen will switch user/kernel base as appropriate on context switch so the
kernel is entered on the kernel gs base.

But you are right - there is no way for Linux to peek at the current
user gs base without reading MSR_GS_SHADOW.  (The user gs base can be
set via a hypercall, but not obtained).

~Andrew