From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:40070) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YqLCH-0004xY-0g for qemu-devel@nongnu.org; Thu, 07 May 2015 08:46:38 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1YqLC8-0002Xc-Ea for qemu-devel@nongnu.org; Thu, 07 May 2015 08:46:36 -0400 Received: from mailout2.w1.samsung.com ([210.118.77.12]:29891) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YqLC8-0002XP-9i for qemu-devel@nongnu.org; Thu, 07 May 2015 08:46:28 -0400 Received: from eucpsbgm1.samsung.com (unknown [203.254.199.244]) by mailout2.w1.samsung.com (Oracle Communications Messaging Server 7.0.5.31.0 64bit (built May 5 2014)) with ESMTP id <0NNZ0094ACTCQF50@mailout2.w1.samsung.com> for qemu-devel@nongnu.org; Thu, 07 May 2015 13:46:24 +0100 (BST) From: Pavel Fedin References: <1430921082-16779-1-git-send-email-shlomopongratz@gmail.com> <1430921082-16779-5-git-send-email-shlomopongratz@gmail.com> <00bd01d08890$b4d60f20$1e822d60$@samsung.com> <00e201d0889d$7537ab50$5fa701f0$@samsung.com> In-reply-to: Date: Thu, 07 May 2015 15:46:20 +0300 Message-id: <01aa01d088c3$d187e9d0$7497bd70$@samsung.com> MIME-version: 1.0 Content-type: text/plain; charset=UTF-8 Content-transfer-encoding: quoted-printable Content-language: ru Subject: Re: [Qemu-devel] [PATCH RFC V2 4/4] Add virtv2 machine that uses GIC-500 List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: 'Shlomo Pongratz' Cc: peter.maydell@linaro.org, qemu-devel@nongnu.org, 'Shlomo Pongratz' Hello! > The issue is similar to what you have noticed. In this case it's probably not your fault. > The larger the number of smp cores it is more likely that the boot = will get stuck. > I've commuted this patch series as an RFC because of this issue. I have tried to give more testing to it. Up to 4 CPUs worked okay (if = we ignore "press any key" problem which i described). Starting from 5 i = get a hard freeze right after " SMP: Total of N processors activated" = message. However, i get this deadlock also with GICv2, so it's perhaps = not your fault. The second problem is that i did not manage to bring up = more than 8 CPUs. The following happens: --- cut --- Initializing cgroup subsys memory Initializing cgroup subsys hugetlb hw perfevents: no hardware support available EFI services will not be available. CPU1: Booted secondary processor Detected PIPT I-cache on CPU1 CPU1: found redistributor 1 region 0:0x0000000008060000 CPU2: Booted secondary processor Detected PIPT I-cache on CPU2 CPU2: found redistributor 2 region 0:0x0000000008080000 CPU3: Booted secondary processor Detected PIPT I-cache on CPU3 CPU3: found redistributor 3 region 0:0x00000000080a0000 CPU4: Booted secondary processor Detected PIPT I-cache on CPU4 CPU4: found redistributor 4 region 0:0x00000000080c0000 CPU5: Booted secondary processor Detected PIPT I-cache on CPU5 CPU5: found redistributor 5 region 0:0x00000000080e0000 CPU6: Booted secondary processor Detected PIPT I-cache on CPU6 CPU6: found redistributor 6 region 0:0x0000000008100000 CPU7: Booted secondary processor Detected PIPT I-cache on CPU7 CPU7: found redistributor 7 region 0:0x0000000008120000 psci: failed to boot CPU8 (-22) CPU8: failed to boot: -22 psci: failed to boot CPU9 (-22) CPU9: failed to boot: -22 psci: failed to boot CPU10 (-22) CPU10: failed to boot: -22 psci: failed to boot CPU11 (-22) CPU11: failed to boot: -22 psci: failed to boot CPU12 (-22) CPU12: failed to boot: -22 psci: failed to boot CPU13 (-22) CPU13: failed to boot: -22 psci: failed to boot CPU14 (-22) CPU14: failed to boot: -22 psci: failed to boot CPU15 (-22) CPU15: failed to boot: -22 Brought up 8 CPUs SMP: Total of 8 processors activated. --- cut --- But, perhaps it's the fault of my code (i replaced your patch No 4 with = my own implementation, merging the code). I will post this alternative = version soon if you don't mind (need to rebase it on master). My current guest kernel is v3.19.5. I will test on 4.1rc2 soon. Kind regards, Pavel Fedin Expert Engineer Samsung Electronics Research center Russia