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