Re: [PATCH 5/5] x86_64: add 64-bit syscall entry point

[Samuel Thibault]

Sergey Bugaev, le mar. 28 févr. 2023 09:39:40 +0300, a ecrit:
> > +                 :                                      \
> > +                 : "c" (regaddr), "a" (low), "d" (high) \
> > +                 : "memory"                             \
> > +        );
> > +}
> 
> Why "memory" here? Can wrmsr clobber unrelated memory?

No, but if you don't put a memory clobber here, the compiler will
optimize the asm away since it's not said to have side effect. Put
another way, the MSR register is some kind of memory.

> > +static inline uint64_t rdmsr(uint32_t regaddr)
> > +{
> > +    uint32_t low, high;
> > +    asm volatile("rdmsr\n"                              \
> > +                 : "=a" (low), "=d" (high)              \
> > +                 : "c" (regaddr)                        \
> > +        );
> > +    return ((uint64_t)high << 32) | low;
> 
> Ditto about spacing -- and does this need volatile? As in, does
> reading from a MSR have side effects that we're interested in,

IIRC it does?

> > diff --git a/i386/include/mach/i386/syscall_sw.h 
> > b/i386/include/mach/i386/syscall_sw.h
> > index 86f6ff2f..20ef7c13 100644
> > --- a/i386/include/mach/i386/syscall_sw.h
> > +++ b/i386/include/mach/i386/syscall_sw.h
> 
> OK, so the x86_64 syscall definition stays in i386/syscall_sw.h, and
> not in a separate x86_64/syscall_sw.h file? That's what I thought. In
> this case, we do want that mach-machine patch in glibc. Samuel, does
> this make sense to you?

Better separate them indeed.

> Predicating on USER32 is not really going to work here.

Partly because of that :)

> it has its own syscall impl, so I need to understand the ABI.
> trap_number is in rax, args on the stack, return value in rax, is that
> right? What's rcx/r10?  Does a syscall preserve other registers? What
> about rflags? Is this the same as Linux does or?..

It's the same as Linux.

Samuel