From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from linux.microsoft.com (linux.microsoft.com [13.77.154.182]) by mx.groups.io with SMTP id smtpd.web08.411.1631200987678276387 for ; Thu, 09 Sep 2021 08:23:07 -0700 Authentication-Results: mx.groups.io; dkim=fail reason="body hash did not verify" header.i=@linux.microsoft.com header.s=default header.b=SPpTevMl; spf=pass (domain: linux.microsoft.com, ip: 13.77.154.182, mailfrom: mikuback@linux.microsoft.com) Received: from [10.0.0.19] (c-73-27-179-174.hsd1.fl.comcast.net [73.27.179.174]) by linux.microsoft.com (Postfix) with ESMTPSA id C418920B6C51; Thu, 9 Sep 2021 08:23:06 -0700 (PDT) DKIM-Filter: OpenDKIM Filter v2.11.0 linux.microsoft.com C418920B6C51 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.microsoft.com; s=default; t=1631200987; bh=SPuEg2Dj6ejUtvcVDXuWieiqNTClI7dZmeAc5XvStx0=; h=Subject:To:References:From:Date:In-Reply-To:From; b=SPpTevMl3scbGqqolBhrZ/4eV7KKXYz5D+MqNG3hedPgv8lolxJDUoiqXrTzChnN6 IqfGeQAtQGO1/webhgCDAfTPk/1StQBuR/lluEGT4RMo/N8fb1/GPRTIzzRa6wayv4 UtZVsXxxY8fPJAPJ3VAMLxnRGrG/F90xxShAU3xI= Subject: Re: [edk2-devel] [edk2-platforms][PATCH v1 1/1] IntelSiliconPkg: Add BaseSmmAccessLibNull To: "Ni, Ray" , "devel@edk2.groups.io" References: <28423.1628820964257783330@groups.io> <28669d2c-da78-d35f-4ad3-812a70d09ba2@linux.microsoft.com> <66f3e7cf-7593-6c40-b7c3-091db4164c81@linux.microsoft.com> <5566a97f-43cb-5a8f-0942-7f66cc1ff19b@linux.microsoft.com> From: "Michael Kubacki" Message-ID: Date: Thu, 9 Sep 2021 11:23:05 -0400 User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:78.0) Gecko/20100101 Thunderbird/78.13.0 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8; format=flowed Content-Language: en-US Content-Transfer-Encoding: quoted-printable I think it would be nice to have both. That was just an example of where=20 SmmAccessLib could be linked that would cause two instances of the=20 library API to be invoked by common code that is linked to two different=20 modules. Some platforms might not use BoardInitLib that could benefit from the=20 PEIM in IntelSiliconPkg. On 9/9/2021 10:58 AM, Ni, Ray wrote: > I think refactoring it to a PEIM is better. > But I am not sure if having below Bugzilla completed can meet your needs = without refactoring the lib to PEIM. >=20 >> I don't want to get too distracted with the example given, but I >> completely agree that a different library instance should be used for >> pre-memory and post-memory. I think the library interface is too broad >> in scope and that contributes to causing this issue so I filed this BZ >> to request the BoardInitLib API be refactored: >> https://bugzilla.tianocore.org/show_bug.cgi?id=3D3578 >=20 >> -----Original Message----- >> From: devel@edk2.groups.io On Behalf Of Michael K= ubacki >> Sent: Thursday, September 9, 2021 10:54 PM >> To: devel@edk2.groups.io; Ni, Ray >> Subject: Re: [edk2-devel] [edk2-platforms][PATCH v1 1/1] IntelSiliconPkg= : Add BaseSmmAccessLibNull >> >> So you would rather leave it as a library class instead of refactoring >> it to a PEIM? >> >> Again, the problem is it is a library class. So I am asking whether you >> want to treat it as a library class or you are going to refactor it to a >> PEIM. >> >> On 9/9/2021 10:49 AM, Ni, Ray wrote: >>> No, I don't. >>> I still don't think having a NULL SmmAccessLib is a good idea. >>> >>>> -----Original Message----- >>>> From: Michael Kubacki >>>> Sent: Thursday, September 9, 2021 10:12 PM >>>> To: Ni, Ray ; devel@edk2.groups.io >>>> Subject: Re: [edk2-devel] [edk2-platforms][PATCH v1 1/1] IntelSiliconP= kg: Add BaseSmmAccessLibNull >>>> >>>> Ray, >>>> >>>> Do you have plans to do something here? Whether take this patch or >>>> refactor SmmAccessLib to a PEIM? >>>> >>>> Thanks, >>>> Michael >>>> >>>> On 8/20/2021 3:34 PM, Michael Kubacki wrote: >>>>> Since you asked for an example that was just one that I provided. I >>>>> don't think it detracts from the fact that a NULL instance makes sens= e >>>>> if the SmmAccessLib library class exists. The fact that a NULL instan= ce >>>>> could not be allowed to exist is also confusing. >>>>> >>>>> I don't want to get too distracted with the example given, but I >>>>> completely agree that a different library instance should be used for >>>>> pre-memory and post-memory. I think the library interface is too broa= d >>>>> in scope and that contributes to causing this issue so I filed this B= Z >>>>> to request the BoardInitLib API be refactored: >>>>> https://bugzilla.tianocore.org/show_bug.cgi?id=3D3578 >>>>> >>>>> From the other email thread about SmmAccessLib, I think we are on = the >>>>> same page that the library would be better as a PEIM. Is that somethi= ng >>>>> that could be done soon? Or could we have this until that is done? >>>>> >>>>> Thanks, >>>>> Michael >>>>> >>>>> On 8/20/2021 1:33 AM, Ni, Ray wrote: >>>>>> Null SmmAccessLib is confusing to me. Have you evaluated the option: >>>>>> Create two instances of BoardInitLib for pre-mem and post-mem. Pre-m= em >>>>>> one doesn=E2=80=99t link to SmmAccessLib >>>>>> >>>>>>> -----Original Message----- >>>>>>> From: devel@edk2.groups.io On Behalf Of >>>>>>> Michael Kubacki >>>>>>> Sent: Friday, August 20, 2021 12:53 AM >>>>>>> To: devel@edk2.groups.io; Ni, Ray >>>>>>> Subject: Re: [edk2-devel] [edk2-platforms][PATCH v1 1/1] >>>>>>> IntelSiliconPkg: Add BaseSmmAccessLibNull >>>>>>> >>>>>>> I don't understand your argument. >>>>>>> >>>>>>> The library class (SmmAccessLib) that already exists is the abstrac= tion >>>>>>> layer. This is not introducing a new layer of abstraction. It is us= ing >>>>>>> the current layer of abstraction. >>>>>>> >>>>>>> Thanks, >>>>>>> Michael >>>>>>> >>>>>>> On 8/19/2021 5:49 AM, Ni, Ray wrote: >>>>>>>> Michael, >>>>>>>> >>>>>>>> I don=E2=80=99t think scenario #1 is a good reason for NULL instan= ce of >>>>>>>> SmmAccessLib. The root cause is BoardInitLib lib class supports pr= e-mem >>>>>>>> and post-mem board init. >>>>>>>> >>>>>>>> Below solution can avoid NULL SmmAccessLib: >>>>>>>> >>>>>>>> Create two instances of BoardInitLib for pre-mem and post-mem. Pre= -mem >>>>>>>> one doesn=E2=80=99t link to SmmAccessLib. >>>>>>>> >>>>>>>> For scenario #2, if a particular platform doesn=E2=80=99t support = S3, why does >>>>>>>> this platform include the PEIM? >>>>>>>> >>>>>>>> Please understand that I want to avoid introducing more abstractio= n >>>>>>>> layers. >>>>>>>> >>>>>>>> Thanks, >>>>>>>> >>>>>>>> Ray >>>>>>>> >>>>>>>> *From:* Michael Kubacki >>>>>>>> *Sent:* Friday, August 13, 2021 10:16 AM >>>>>>>> *To:* Ni; Ni, Ray ; devel@edk2.groups.io >>>>>>>> *Subject:* Re: [edk2-devel] [edk2-platforms][PATCH v1 1/1] >>>>>>>> IntelSiliconPkg: Add BaseSmmAccessLibNull >>>>>>>> >>>>>>>> Sure. >>>>>>>> >>>>>>>> Scenario #1: >>>>>>>> >>>>>>>> MinPlatformPkg/PlatformInit/PlatformInitPei/PlatformInitPreMem.inf= and >>>>>>>> MinPlatormPkg/PlatformInit/PlatformInitPei/PlatformInitPostMem.inf= both >>>>>>>> link against an instance of BoardInitLib. >>>>>>>> >>>>>>>> Many boards link against a single BoardInitLib instance. See examp= le - >>>>>>>> https://github.com/tianocore/edk2- >>>>>>> >>>> >> platforms/blob/cd4e6b716c7d1bcde94035e7dce14b53a553e103/Platform/Intel/K= abylakeOpenBoardPkg/KabylakeRvp3/OpenB >>>>>>> >>>>>>> oardPkg.dsc#L203 >>>>>>>> >>>>>> >>>> >> platforms/blob/cd4e6b716c7d1bcde94035e7dce14b53a553e103/Platform/Intel/K= abylakeOpenBoardPkg/KabylakeRvp3/OpenB >>>>>>> >>>>>>> oardPkg.dsc#L203> >>>>>>>> >>>>>>>> That BoardInitLib instance may link against SmmAccessLib. >>>>>>>> PlatformInitPreMem may wish to library class override the SmmAcces= sLib >>>>>>>> to the NULL instance while keeping it to non-NULL instance in >>>>>>>> PlatformInitPostMem. >>>>>>>> >>>>>>>> Scenario #2: >>>>>>>> >>>>>>>> A PEIM is built that checks whether the boot mode is S3. If so, it >>>>>>>> calls >>>>>>>> PeiInstallSmmAccessPpi(). A particular platform does not support S= 3, >>>>>>>> therefore, it links BaseSmmAccessLibNull as its library instance f= or >>>>>>>> SmmAccessLib. >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>> >>> >>> >>> >>> >>> >> >> >>=20 >> >=20