Re: [Qemu-devel] [PATCH RFC V2 4/4] Add virtv2 machine that uses GIC-500

From: Pavel Fedin <p.fedin@samsung.com>
To: 'Shlomo Pongratz' <shlomopongratz@gmail.com>
Cc: peter.maydell@linaro.org, qemu-devel@nongnu.org,
	'Shlomo Pongratz' <shlomo.pongratz@huawei.com>
Subject: Re: [Qemu-devel] [PATCH RFC V2 4/4] Add virtv2 machine that uses GIC-500
Date: Thu, 07 May 2015 15:46:20 +0300	[thread overview]
Message-ID: <01aa01d088c3$d187e9d0$7497bd70$@samsung.com> (raw)
In-Reply-To: <CAHzK-V19rG+ZxXZ+JRQ-qVENj938Dg_zszxtHtKsews6NJfvtQ@mail.gmail.com>

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