From mboxrd@z Thu Jan 1 00:00:00 1970 Authentication-Results: mx.groups.io; dkim=missing; spf=pass (domain: redhat.com, ip: 209.132.183.28, mailfrom: lersek@redhat.com) Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28]) by groups.io with SMTP; Thu, 11 Jul 2019 05:29:48 -0700 Received: from smtp.corp.redhat.com (int-mx01.intmail.prod.int.phx2.redhat.com [10.5.11.11]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 9E58985540; Thu, 11 Jul 2019 12:29:47 +0000 (UTC) Received: from lacos-laptop-7.usersys.redhat.com (ovpn-117-231.ams2.redhat.com [10.36.117.231]) by smtp.corp.redhat.com (Postfix) with ESMTP id 0FC77600CD; Thu, 11 Jul 2019 12:29:45 +0000 (UTC) Subject: Re: [edk2-devel] [PATCH v2 2/3] MdePkg/BaseLib.h: Update IA32_CR4 structure for 5-level paging To: "Ni, Ray" , "Kinney, Michael D" , "devel@edk2.groups.io" , "Dong, Eric" Cc: Leif Lindholm , "Gao, Liming" References: <20190703065416.116816-1-ray.ni@intel.com> <20190703065416.116816-3-ray.ni@intel.com> <23baa040-63a2-3d8f-f75c-91b1d2c0e3ac@redhat.com> <734D49CCEBEEF84792F5B80ED585239D5C222153@SHSMSX104.ccr.corp.intel.com> From: "Laszlo Ersek" Message-ID: <47d8961d-8c9b-dc66-1e79-ad6f574d2610@redhat.com> Date: Thu, 11 Jul 2019 14:29:45 +0200 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.9.1 MIME-Version: 1.0 In-Reply-To: <734D49CCEBEEF84792F5B80ED585239D5C222153@SHSMSX104.ccr.corp.intel.com> X-Scanned-By: MIMEDefang 2.79 on 10.5.11.11 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.28]); Thu, 11 Jul 2019 12:29:47 +0000 (UTC) Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit On 07/11/19 05:25, Ni, Ray wrote: > Laszlo, Mike, > Sorry I did violate the process. > I had two assumptions which led me violate the process: > 1. Reviewed-by from UefiCpuPkg maintainers on this IA32_CR4 change is more important > than that from MdePkg maintainers. In another word, I thought if UefiCpuPkg maintainers > agree with this change, MdePkg maintainers should have no concerns. > (It's a wrong assumption. MdePkg maintainers may have some general suggestions, e.g.: > name, location, comments and etc..) > 2. This change is directly from the published white paper and there is no other option regarding > this IA32_CR4 change. > (It's a wrong assumption. MdePkg maintainers may have some general suggestions, e.g.: > name, location, comments and etc..) Both of these assumptions could make perfect sense *to you*, but the rules exist to uphold a single standard for everyone. I didn't expect Mike or Liming to find an issue in the "MdePkg/BaseLib.h" patch. I did expect the process to be followed. I fully support if you and Eric are added to MdePkg as co-maintainers, for content that is closely related to UefiCpuPkg. Then, the assumptions will be codified, and they will be clear to everyone. To explain "single standard" a bit more: sometimes I too submit code for MdePkg. (Right now, I happen to have a series pending review.) If I were cynical, my thinking could go, "well, if other folks can push to MdePkg without MdePkg maintainer approval, so can I". Do you see my problem? Where does it end? A maintainer A-b or R-b on the list is objective; it's a fact. > I agree I should get Reviewed-by tag from MdePkg maintainers. My assumptions are wrong. > > To strictly follow the process, I will: > 1. Post a patch series to revert the 3 patches. > Since this change doesn't break any functionality (does break the process), I will wait for > Reviewed-by from each package maintainer and make sure push the patches after 24h. > 2. After step #1, post a new patch series to add the 3 patches back with Reviewed-by tag from > Eric and Mike, Regression-Tested-by from you. > > Do you think it follows the existing process? Yes, this can work, but I also don't mind if you use git-revert + git-cherry-pick + git-push in this instance (as explained elsewhere), in this particular instance. This is because no further review on the list is necessary (in this particular case), the reverts and reapplications are fully mechanical, you just need to make sure to pick up Mike's R-b for the reapplied MdePkg patch (and to keep the tree buildable at every stage). See again my other email with the detailed steps. > Sorry again for this violation. To emphasize -- I wasn't worried that the patch would cause breakage. My point is that all contributors should be held to the same standard. Thanks Laszlo >> -----Original Message----- >> From: Kinney, Michael D >> Sent: Thursday, July 11, 2019 3:39 AM >> To: Laszlo Ersek ; devel@edk2.groups.io; Dong, Eric ; Ni, Ray ; >> Kinney, Michael D >> Cc: Leif Lindholm ; Gao, Liming >> Subject: RE: [edk2-devel] [PATCH v2 2/3] MdePkg/BaseLib.h: Update IA32_CR4 structure for 5-level paging >> >> Laszlo, >> >> I agree with your feedback. Process must be followed. >> >> I also agree that it may make sense to add some more maintainers >> to the MdePkg, especially for some of the content in MdePkg that >> is closely related to the UefiCpuPkg content. >> >> I have reviewed this patch to the BaseLib.h. The new LA57 bit >> added to IA32_CR4 matches the documentation in the white paper >> referenced in the series. >> >> Reviewed-by: Michael D Kinney >> >> Thanks, >> >> Mike >> >> >>> -----Original Message----- >>> From: Laszlo Ersek [mailto:lersek@redhat.com] >>> Sent: Wednesday, July 10, 2019 10:17 AM >>> To: devel@edk2.groups.io; Dong, Eric >>> ; Ni, Ray >>> Cc: Leif Lindholm ; Gao, Liming >>> ; Kinney, Michael D >>> >>> Subject: Re: [edk2-devel] [PATCH v2 2/3] >>> MdePkg/BaseLib.h: Update IA32_CR4 structure for 5-level >>> paging >>> >>> Ray, Eric, >>> >>> (+Liming, +Mike, +Leif) >>> >>> On 07/09/19 03:04, Dong, Eric wrote: >>>> Reviewed-by: Eric Dong >>>> >>>>> -----Original Message----- >>>>> From: Ni, Ray >>>>> Sent: Wednesday, July 3, 2019 2:54 PM >>>>> To: devel@edk2.groups.io >>>>> Cc: Dong, Eric ; Laszlo Ersek >>>>> >>>>> Subject: [PATCH v2 2/3] MdePkg/BaseLib.h: Update >>> IA32_CR4 structure >>>>> for 5-level paging >>>>> >>>>> 5-level paging is documented in white paper: >>>>> >>> https://software.intel.com/sites/default/files/managed/2b >>> /80/5- >>>>> level_paging_white_paper.pdf >>>>> >>>>> Commit f8113e25001e715390127f23e2197252cbd6d1a2 >>>>> changed Cpuid.h already. >>>>> >>>>> This patch updates IA32_CR4 structure to include LA57 >>> field. >>>>> >>>>> Signed-off-by: Ray Ni >>>>> Cc: Eric Dong >>>>> Regression-tested-by: Laszlo Ersek >>>>> --- >>>>> MdePkg/Include/Library/BaseLib.h | 3 ++- >>>>> 1 file changed, 2 insertions(+), 1 deletion(-) >>>>> >>>>> diff --git a/MdePkg/Include/Library/BaseLib.h >>>>> b/MdePkg/Include/Library/BaseLib.h >>>>> index ebd7dd274c..a22bfc9fad 100644 >>>>> --- a/MdePkg/Include/Library/BaseLib.h >>>>> +++ b/MdePkg/Include/Library/BaseLib.h >>>>> @@ -5324,7 +5324,8 @@ typedef union { >>>>> UINT32 OSXMMEXCPT:1; ///< Operating System >>> Support for >>>>> ///< Unmasked SIMD >>> Floating Point >>>>> ///< Exceptions. >>>>> - UINT32 Reserved_0:2; ///< Reserved. >>>>> + UINT32 Reserved_2:1; ///< Reserved. >>>>> + UINT32 LA57:1; ///< Linear Address >>> 57bit. >>>>> UINT32 VMXE:1; ///< VMX Enable >>>>> UINT32 Reserved_1:18; ///< Reserved. >>>>> } Bits; >>> >>> I'm sorry but you will have to revert this patch series >>> immediately. >>> None of the MdePkg maintainers have approved this patch - >>> - commit 7c5010c7f88b. >>> >>> In the first place, Mike and Liming were never CC'd on >>> the patch, so they may not have noticed it, even. >>> >>> The situation is very similar to the recent SM3 crypto >>> series that I had to revert myself. An MdePkg patch was >>> pushed without package owner review. >>> >>> Can you guys please revert this series immediately, >>> without me having to do it? >>> >>> >>> If we think that MdePkg should have more "M" folks, in >>> order to distribute the review load better, then we >>> should address that problem first. Ignoring rules just >>> because that's more convenient is not acceptable. >>> >>> Thanks, >>> Laszlo