答复: [PATCH] ARM64:DMI: Add smbios/dmi support on arm64

[liyi 00215672]

Please see below:

-----邮件原件-----
发件人: Mark Rutland [mailto:mark.rutland at arm.com] 
发送时间: 2014年6月5日 23:34
收件人: Yi Li
抄送: ard.biesheuvel at linaro.org; Catalin Marinas; Will Deacon; Sudeep Holla; steve.capper at linaro.org; Lorenzo Pieralisi; Liviu Dudau; grant.likely at linaro.org; liyi 00215672; linux-kernel at vger.kernel.org; linux-arm-kernel at lists.infradead.org
主题: Re: [PATCH] ARM64:DMI: Add smbios/dmi support on arm64

On Wed, Jun 04, 2014 at 04:01:36PM +0100, Yi Li wrote:
> Hi Mark,
> 
>      Please see the comments below:
> 
> On Wednesday, June 04, 2014 09:32 PM, Mark Rutland wrote:
> > On Tue, Jun 03, 2014 at 04:57:13PM +0100, Yi Li wrote:
> >> Add smbios/dmi support on arm64 system, it depends on
> >> EFI boot.
> > And what exactly does this provide us with?
> >
> > What is exposed through SMBIOS/DMI, and why would I want to enable it?
> Yi: SMBIOS/DMI is one basic spec/feature for server product(like x86 and 
> IA64).
>       Many OEMs/ODMs hope to use ARM64 as server's processor.
>       So we need to support SMBIOS on ARM64.
> 
>      SMBIOS mainly describes some hardware and software information for 
> the system, like BIOS information
>      CPU information,  Memory information ,and so on. please refer to 
> http://www.dmtf.org/standards/smbios

What information does it provide that is useful that we cannot get from
elsewhere?
Yi: As my previous mail, SMBIOS is one basic specification for all server production (x86 and IA64).
  It can provide some SN or UUID of production/motherboard/memory, I think you can not get that from elsewhere.

What use case is enabled by having this support, other than being able
to tick a box that we have a feature we might not even need?
Yi: dmidecode/ lshw tools can be used as use case to test dmi driver is working well or not .
  And SMBIOS is one MUST section in SBBR document, so it is needed. 
> 
> >
> >> Signed-off-by: Yi Li <yi.li at linaro.org>
> >> ---
> >>
> >> Changes since v1:
> >>    -Followed Ard Biesheuvel's suggestion to rebase the patch on
> >>     Matt Fleming's arm64-efi branch.
> >>
> >>   arch/arm64/Kconfig           |   10 ++++++++++
> >>   arch/arm64/include/asm/dmi.h |   28 ++++++++++++++++++++++++++++
> >>   arch/arm64/kernel/setup.c    |    2 ++
> >>   3 files changed, 40 insertions(+)
> >>   create mode 100644 arch/arm64/include/asm/dmi.h
> >>
> >> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> >> index 6c71f12..13ee261 100644
> >> --- a/arch/arm64/Kconfig
> >> +++ b/arch/arm64/Kconfig
> >> @@ -294,6 +294,16 @@ config EFI
> >>   	  allow the kernel to be booted as an EFI application. This
> >>   	  is only useful on systems that have UEFI firmware.
> >>   
> >> +config DMI
> >> +	bool "Enable support for SMBIOS (DMI) tables"
> >> +	depends on EFI
> >> +	default y
> >> +	help
> >> +	  This enables SMBIOS/DMI feature for systems.
> >> +
> >> +	  This option is only useful on systems that have UEFI firmware.
> >> +	  However, even with this option, the resultant kernel should
> >> +	  continue to boot on existing non-UEFI platforms.
> >>   endmenu
> >>   
> >>   menu "Userspace binary formats"
> >> diff --git a/arch/arm64/include/asm/dmi.h b/arch/arm64/include/asm/dmi.h
> >> new file mode 100644
> >> index 0000000..f2198bf
> >> --- /dev/null
> >> +++ b/arch/arm64/include/asm/dmi.h
> >> @@ -0,0 +1,28 @@
> >> +/*
> >> + * arch/arm64/include/asm/dmi.h
> >> + *
> >> + * Copyright (C) 2013 Linaro Limited.
> >> + * Written by: Yi Li (yi.li at linaro.org)
> >> + *
> >> + * based on arch/ia64/include/asm/dmi.h
> >> + *
> >> + * This file is subject to the terms and conditions of the GNU General Public
> >> + * License.  See the file "COPYING" in the main directory of this archive
> >> + * for more details.
> >> + */
> >> +
> >> +
> >> +#ifndef _ASM_DMI_H
> >> +#define _ASM_DMI_H 1
> >> +
> >> +#include <linux/slab.h>
> >> +#include <asm/io.h>
> > Shouldn't that be linux/efi.h?
> >
> > Why do we need asm/io.h?
>      Yi: porting it from IA64 , so the io.h is not needed exactly!
>           but slab.h is must included ,not efi.h (tested by compiling)

Below you refer to efi_lookup_mapped_addr, which (as far as I can tell)
is defined in efi.h. You might be getting lucky with users of
dmi_early_remap and dmi_remap having already incldued efi.h, but you
shouldn't rely on it.

If you make use of something, you should include the header that defines
it.

Yi: yes, correct. It is luck that dmi_scan.c has included efi.h , I will add efi.h in dmi.h later.

Cheers,
Mark.

> >> +
> >> +/* Use efi mappings for DMI */
> >> +#define dmi_early_remap(x, l)	efi_lookup_mapped_addr(x)
> >> +#define dmi_early_unmap(x, l)
> >> +#define dmi_remap(x, l)			efi_lookup_mapped_addr(x)
> >> +#define dmi_unmap(x)
> >> +#define dmi_alloc(l)			kzalloc(l, GFP_ATOMIC)
> >> +
> >> +#endif
> > None of these seem to use anything from io.h directly.
>      Yi: You are right , io.h doesn't need.
> > Cheers,
> > Mark.
> 
>