From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28]) by mx.groups.io with SMTP id smtpd.web10.3561.1570696083014757291 for ; Thu, 10 Oct 2019 01:28:03 -0700 Authentication-Results: mx.groups.io; dkim=missing; spf=pass (domain: redhat.com, ip: 209.132.183.28, mailfrom: lersek@redhat.com) Received: from smtp.corp.redhat.com (int-mx02.intmail.prod.int.phx2.redhat.com [10.5.11.12]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 8CEE1309BF1D; Thu, 10 Oct 2019 08:28:02 +0000 (UTC) Received: from lacos-laptop-7.usersys.redhat.com (ovpn-120-48.rdu2.redhat.com [10.10.120.48]) by smtp.corp.redhat.com (Postfix) with ESMTP id 7EF5A60C05; Thu, 10 Oct 2019 08:28:01 +0000 (UTC) Subject: Re: [edk2-devel] [PATCH 1/4] UefiCpuPkg/MpInitLib: honor the platform's boot CPU count in AP detection From: "Laszlo Ersek" To: "Ni, Ray" , "Dong, Eric" , "devel@edk2.groups.io" Cc: Igor Mammedov References: <20191008112714.6215-1-lersek@redhat.com> <20191008112714.6215-2-lersek@redhat.com> <3c535d1a-d977-4543-b419-f4442e4a3cff@redhat.com> <734D49CCEBEEF84792F5B80ED585239D5C31955E@SHSMSX104.ccr.corp.intel.com> <571f8d70-c046-516d-3536-a7d9314fb5ac@redhat.com> Message-ID: Date: Thu, 10 Oct 2019 10:28:00 +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: <571f8d70-c046-516d-3536-a7d9314fb5ac@redhat.com> X-Scanned-By: MIMEDefang 2.79 on 10.5.11.12 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.45]); Thu, 10 Oct 2019 08:28:02 +0000 (UTC) Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit On 10/10/19 09:38, Laszlo Ersek wrote: > On 10/10/19 04:52, Ni, Ray wrote: >> Laszlo, >> Can you add comments in the code you changed to describe the two different behaviors? > > It's described in the DEC file, near the PCD: > > + # nonzero - PcdCpuApInitTimeOutInMicroSeconds is ignored. The initial > + # AP detection finishes when the detected CPU count (BSP > + # plus APs) reaches the value of this PCD.
> > (1) Do you want me to repeat this explanation in the code as well? > > (2) Do you want me to emphasize in the explanation that, if at least one processor fails to check in, the boot is going to block forever, by design? > > (3) Emphasize this in the DEC file only, in the code only, or in both? ... I think I've figured it out, let me send a v2 of just this patch. (The rest of the series (for OvmfPkg) will likely be dropped, due to an alternative proposal from Igor. The UefiCpuPkg facility will remain the same however.) Thanks! Laszlo