public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: "Ni, Ray" <ray.ni@intel.com>
To: "Zeng, Star" <star.zeng@intel.com>,
	"devel@edk2.groups.io" <devel@edk2.groups.io>,
	"lersek@redhat.com" <lersek@redhat.com>
Cc: "Dong, Eric" <eric.dong@intel.com>,
	"Kumar, Chandana C" <chandana.c.kumar@intel.com>
Subject: Re: [edk2-devel] [PATCH] UefiCpuPkg CpuCommonFeaturesLib: Remove CPU generation check
Date: Fri, 17 May 2019 01:04:25 +0000	[thread overview]
Message-ID: <734D49CCEBEEF84792F5B80ED585239D5C145D7D@SHSMSX104.ccr.corp.intel.com> (raw)
In-Reply-To: <0C09AFA07DD0434D9E2A0C6AEB048310402E357B@shsmsx102.ccr.corp.intel.com>

Star,
I think the discussion is about providing the evidence to support removing the generation check.
Not just the benefit of that.

Thanks,
Ray

> -----Original Message-----
> From: Zeng, Star
> Sent: Thursday, May 16, 2019 10:52 PM
> To: devel@edk2.groups.io; lersek@redhat.com
> Cc: Dong, Eric <eric.dong@intel.com>; Ni, Ray <ray.ni@intel.com>; Kumar,
> Chandana C <chandana.c.kumar@intel.com>; Zeng, Star
> <star.zeng@intel.com>
> Subject: RE: [edk2-devel] [PATCH] UefiCpuPkg CpuCommonFeaturesLib:
> Remove CPU generation check
> 
> Laszlo,
> 
> > -----Original Message-----
> > From: devel@edk2.groups.io [mailto:devel@edk2.groups.io] On Behalf Of
> > Laszlo Ersek
> > Sent: Thursday, May 16, 2019 9:06 PM
> > To: Zeng, Star <star.zeng@intel.com>; devel@edk2.groups.io
> > Cc: Dong, Eric <eric.dong@intel.com>; Ni, Ray <ray.ni@intel.com>;
> > Kumar, Chandana C <chandana.c.kumar@intel.com>
> > Subject: Re: [edk2-devel] [PATCH] UefiCpuPkg CpuCommonFeaturesLib:
> > Remove CPU generation check
> >
> > Hi Star,
> >
> > On 05/16/19 12:33, Star Zeng wrote:
> > > BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=1679
> > >
> > > The checking to CpuInfo->CpuIdVersionInfoEcx.Bits.AESNI is enough,
> > > the checking to CPU generation could be removed, then the code could
> > > be reused by more platforms.
> > >
> > > Cc: Laszlo Ersek <lersek@redhat.com>
> > > Cc: Eric Dong <eric.dong@intel.com>
> > > Cc: Ruiyu Ni <ruiyu.ni@intel.com>
> > > Cc: Chandana Kumar <chandana.c.kumar@intel.com>
> > > Signed-off-by: Star Zeng <star.zeng@intel.com>
> > > ---
> > >  UefiCpuPkg/Library/CpuCommonFeaturesLib/Aesni.c | 12 +++---------
> > >  1 file changed, 3 insertions(+), 9 deletions(-)
> > >
> > > diff --git a/UefiCpuPkg/Library/CpuCommonFeaturesLib/Aesni.c
> > > b/UefiCpuPkg/Library/CpuCommonFeaturesLib/Aesni.c
> > > index b79446ba3ca9..4a56eec1b267 100644
> > > --- a/UefiCpuPkg/Library/CpuCommonFeaturesLib/Aesni.c
> > > +++ b/UefiCpuPkg/Library/CpuCommonFeaturesLib/Aesni.c
> > > @@ -57,15 +57,9 @@ AesniSupport (
> > >    MSR_SANDY_BRIDGE_FEATURE_CONFIG_REGISTER
> *MsrFeatureConfig;
> > >
> > >    if (CpuInfo->CpuIdVersionInfoEcx.Bits.AESNI == 1) {
> > > -    if (IS_SANDY_BRIDGE_PROCESSOR (CpuInfo->DisplayFamily, CpuInfo-
> > >DisplayModel) ||
> > > -        IS_SILVERMONT_PROCESSOR (CpuInfo->DisplayFamily, CpuInfo-
> > >DisplayModel) ||
> > > -        IS_XEON_5600_PROCESSOR (CpuInfo->DisplayFamily, CpuInfo-
> > >DisplayModel) ||
> > > -        IS_XEON_E7_PROCESSOR (CpuInfo->DisplayFamily, CpuInfo-
> > >DisplayModel) ||
> > > -        IS_XEON_PHI_PROCESSOR (CpuInfo->DisplayFamily, CpuInfo-
> > >DisplayModel)) {
> > > -      MsrFeatureConfig =
> > (MSR_SANDY_BRIDGE_FEATURE_CONFIG_REGISTER *) ConfigData;
> > > -      ASSERT (MsrFeatureConfig != NULL);
> > > -      MsrFeatureConfig[ProcessorNumber].Uint64 = AsmReadMsr64
> > (MSR_SANDY_BRIDGE_FEATURE_CONFIG);
> > > -    }
> > > +    MsrFeatureConfig =
> > (MSR_SANDY_BRIDGE_FEATURE_CONFIG_REGISTER *) ConfigData;
> > > +    ASSERT (MsrFeatureConfig != NULL);
> > > +    MsrFeatureConfig[ProcessorNumber].Uint64 = AsmReadMsr64
> > > + (MSR_SANDY_BRIDGE_FEATURE_CONFIG);
> > >      return TRUE;
> > >    }
> > >    return FALSE;
> > >
> >
> > the patch and the bugzilla ticket claim that the AESNI bit's presence
> > in CPUID guarantees that MSR 0x13C is available.
> 
> That is the case we met. The purpose of this patch is to make the code more
> usable.
> 
> >
> > I don't see what guarantees this. According to the latest Intel SDM
> > Vol 4, which I just downloaded (335592-069US, January 2019),
> > MSR_FEATURE_CONFIG is available on the following (DisplayFamily,
> > DisplayModel) pairs:
> >
> > - 06_37H, 06_4AH, 06_4DH, 06_5AH, 06_5DH, 06_5CH, 06_7AH
> > - 06_25H, 06_2CH
> > - 06_2FH
> > - 06_2AH, 06_2DH
> > - 06_57H
> 
> Yes, right.
> 
> Let me show some examples for the generations not in the list above.
> 
> 1. MSR 0x13C is available: our some internal generations are in this case.
> Without the patch, code needs to use function level override method in a
> CpuSpecificFeaturesLib.
>     Status = RegisterCpuFeature (
>                "AESNI",
>                NULL,                                         // Use core function
>                SpecificAesniSupport,                         // Override core function
>                NULL,                                         // Use core function
>                CPU_FEATURE_AESNI,
>                CPU_FEATURE_END
>                );
> With the patch, the function level override will be not needed. The benefit
> of this patch is here.
> 
> 2. MSR 0x13C is not available: let's assume some other MSR will be available
> for the case.
> Without or with the patch, codes both need to use function level override
> method in a CpuSpecificFeaturesLib.
>     Status = RegisterCpuFeature (
>                "AESNI",
>                NULL,                                         // Use core function
>                SpecificAesniSupport,                         // Override core function
>                SpecificAesniInitialize,                         // Override core function
>                CPU_FEATURE_AESNI,
>                CPU_FEATURE_END
>                );
> 
> 
> Thanks,
> Star
> 
> >
> > Which seems to indicate that at least *the approach* of the original
> > code -- i.e. the family/model checking -- is correct. (It's possible
> > that the family/model list has to be extended from time to time, of
> > course.)
> >
> > Anyway, I don't intend to block this patch; OVMF does not use
> > CpuCommonFeaturesLib, so this change cannot regress it. I will let
> > other UefiCpuPkg reviewers decide about this patch.
> >
> > Thanks!
> > Laszlo
> >
> > 


  parent reply	other threads:[~2019-05-17  1:04 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-05-16 10:33 [PATCH] UefiCpuPkg CpuCommonFeaturesLib: Remove CPU generation check Zeng, Star
2019-05-16 13:06 ` Laszlo Ersek
2019-05-16 14:51   ` [edk2-devel] " Zeng, Star
2019-05-17  1:01     ` Dong, Eric
2019-05-17  1:04     ` Ni, Ray [this message]
2019-05-17  3:05       ` Zeng, Star
2019-05-17 12:13         ` Laszlo Ersek
2019-05-17 13:10         ` Ni, Ray
2019-05-18  5:51           ` Zeng, Star

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-list from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=734D49CCEBEEF84792F5B80ED585239D5C145D7D@SHSMSX104.ccr.corp.intel.com \
    --to=devel@edk2.groups.io \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox