Re: [PATCH v9 07/13] x86: add multiboot2 protocol support for EFI platforms

From: Daniel Kiper <daniel.kiper@oracle.com>
To: Jan Beulich <JBeulich@suse.com>
Cc: Juergen Gross <JGross@suse.com>,
	sstabellini@kernel.org, Andrew Cooper <andrew.cooper3@citrix.com>,
	cardoe@cardoe.com, pgnet.dev@gmail.com, ning.sun@intel.com,
	julien.grall@arm.com, david.vrabel@citrix.com,
	xen-devel@lists.xenproject.org, qiaowei.ren@intel.com,
	gang.wei@intel.com, fu.wei@linaro.org
Subject: Re: [PATCH v9 07/13] x86: add multiboot2 protocol support for EFI platforms
Date: Wed, 30 Nov 2016 14:45:27 +0100	[thread overview]
Message-ID: <20161130134527.GU18527@olila.local.net-space.pl> (raw)
In-Reply-To: <5837FB6F020000780012214D@prv-mh.provo.novell.com>

On Fri, Nov 25, 2016 at 12:50:55AM -0700, Jan Beulich wrote:
> >>> On 24.11.16 at 22:44, <daniel.kiper@oracle.com> wrote:
> > On Thu, Nov 24, 2016 at 04:08:12AM -0700, Jan Beulich wrote:
> >> >>> On 23.11.16 at 19:52, <andrew.cooper3@citrix.com> wrote:
> >> > On 29/09/16 22:42, Daniel Kiper wrote:
> >
> > [...]
> >
> >> >> +.Lefi_mb2_tsize:
> >> >> +        /* Check Multiboot2 information total size. */
> >> >> +        mov     %ecx,%r8d
> >> >> +        sub     %ebx,%r8d
> >> >> +        cmp     %r8d,MB2_fixed_total_size(%rbx)
> >> >> +        jbe     run_bs
> >> >> +
> >> >> +        /* Are EFI boot services available? */
> >> >> +        cmpl    $MULTIBOOT2_TAG_TYPE_EFI_BS,MB2_tag_type(%rcx)
> >> >> +        jne     .Lefi_mb2_st
> >> >> +
> >> >> +        /*
> >> >> +         * Yes, store that info in skip_realmode variable. I agree that
> >> >> +         * its name is a bit unfortunate in this context, however, by the
> >> >> +         * way we disable real mode and other legacy stuff which should
> >> >> +         * not be run on EFI platforms.
> >> >> +         */
> >> >> +        incb    skip_realmode(%rip)
> >> >
> >> > Always use add/sub 1 in preference to inc and dec.  They are the same
> >> > length to encode in 64bit, and avoids a pipeline stall from a merge of
> >> > the eflags register.
> >>
> >> What you say regarding length not true - add/sub need to encode
> >> the immediate somewhere (even if the operand was a register,
> >> inc/dec would still be smaller than add/sub, just not by as much as
> >> in 32-bit code). And the pipeline stall, afaik, affects only rather old
> >> processors.
> >
> > Intel 64 and IA-32 Architectures Optimization Reference Manual, section
> > 3.5.1.1, Use of the INC and DEC Instructions says nothing about exceptions.
>
> Which by itself is suspicious, as the dependency issue had been
> introduced only in (iirc) Pentium4. And anyway, this is a general

Hmmm... Interesting... It looks that INC/DEC behavior has not changed since
the beginning (why it would change? It would not make sense). Though my oldest
book about 80286, 80386 and i486 CPUS (printed in 1991 and it is based on
eighties Intel documentation) says nothing about such issues. So, I have
a feeling that this issue appeared (or it existed earlier but it was not
documented minor problem) when x86 implementation changed from 100% in
silicon to mostly (???) microcode. IIRC, it happened around Pentium but
I am not sure.

Daniel

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel