Re: [Qemu-devel] [PATCH v7 00/73] per-CPU locks

From: Richard Henderson <richard.henderson@linaro.org>
To: "Emilio G. Cota" <cota@braap.org>, qemu-devel@nongnu.org
Cc: "Alex Bennée" <alex.bennee@linaro.org>,
	"Aleksandar Markovic" <amarkovic@wavecomp.com>,
	"Alistair Francis" <alistair23@gmail.com>,
	"Andrzej Zaborowski" <balrogg@gmail.com>,
	"Anthony Green" <green@moxielogic.com>,
	"Artyom Tarasenko" <atar4qemu@gmail.com>,
	"Aurelien Jarno" <aurelien@aurel32.net>,
	"Bastian Koppelmann" <kbastian@mail.uni-paderborn.de>,
	"Christian Borntraeger" <borntraeger@de.ibm.com>,
	"Chris Wulff" <crwulff@gmail.com>,
	"Cornelia Huck" <cohuck@redhat.com>,
	"David Gibson" <david@gibson.dropbear.id.au>,
	"David Hildenbrand" <david@redhat.com>,
	"Edgar E. Iglesias" <edgar.iglesias@gmail.com>,
	"Eduardo Habkost" <ehabkost@redhat.com>,
	"Fabien Chouteau" <chouteau@adacore.com>,
	"Guan Xuetao" <gxt@mprc.pku.edu.cn>,
	"James Hogan" <jhogan@kernel.org>,
	"Laurent Vivier" <laurent@vivier.eu>,
	"Marek Vasut" <marex@denx.de>,
	"Mark Cave-Ayland" <mark.cave-ayland@ilande.co.uk>,
	"Max Filippov" <jcmvbkbc@gmail.com>,
	"Michael Walle" <michael@walle.cc>,
	"Palmer Dabbelt" <palmer@sifive.com>,
	"Paolo Bonzini" <pbonzini@redhat.com>,
	"Peter Maydell" <peter.maydell@linaro.org>,
	qemu-arm@nongnu.org, qemu-ppc@nongnu.org, qemu-s390x@nongnu.org,
	"Sagar Karandikar" <sagark@eecs.berkeley.edu>,
	"Stafford Horne" <shorne@gmail.com>
Subject: Re: [Qemu-devel] [PATCH v7 00/73] per-CPU locks
Date: Wed, 6 Mar 2019 11:40:57 -0800	[thread overview]
Message-ID: <2adf8423-458d-27c3-67bf-d4775104ad43@linaro.org> (raw)
In-Reply-To: <20190304181813.8075-1-cota@braap.org>

On 3/4/19 10:17 AM, Emilio G. Cota wrote:
> v6: https://lists.gnu.org/archive/html/qemu-devel/2019-01/msg07650.html
> 
> All patches in the series have reviews now. Thanks everyone!
> 
> I've tested all patches with `make check-qtest -j' for all targets.
> The series is checkpatch-clean (just some warnings about __COVERITY__).
> 
> You can fetch the series from:
>   https://github.com/cota/qemu/tree/cpu-lock-v7

Having fixed the OSX build errors, Peter reported to me a hang on the x86_64
bios-tables-test with an arm32 host.  I have been able to replicate this.

(gdb) info thr
  Id   Target Id                                       Frame
  1    Thread 0xb35a72e0 (LWP 10966) "qemu-system-x86" 0xb606d138 in
pthread_cond_wait@@GLIBC_2.4 () from /usr/lib/libpthread.so.0
  2    Thread 0xb35a3d30 (LWP 10967) "qemu-system-x86" 0xb5fe70fc in syscall ()
   from /usr/lib/libc.so.6
  3    Thread 0xa07fed30 (LWP 10968) "qemu-system-x86" 0xb5fe0688 in poll ()
   from /usr/lib/libc.so.6
* 4    Thread 0x9fdfed30 (LWP 10969) "qemu-system-x86" 0xb606d138 in
pthread_cond_wait@@GLIBC_2.4 () from /usr/lib/libpthread.so.0

(gdb) thr 1
[Switching to thread 1 (Thread 0xb35a72e0 (LWP 10966))]
#0  0xb606d138 in pthread_cond_wait@@GLIBC_2.4 () from /usr/lib/libpthread.so.0
(gdb) where
#0  0xb606d138 in pthread_cond_wait@@GLIBC_2.4 () from /usr/lib/libpthread.so.0
#1  0x009a66a4 in qemu_cond_wait_impl (cond=cond@entry=0x16727b0,
    mutex=0xcad4c8 <qemu_global_mutex>,
    file=file@entry=0xa057f0 "/home/rth/qemu/qemu/cpus-common.c",
    line=line@entry=166) at /home/rth/qemu/qemu/util/qemu-thread-posix.c:161
#2  0x00692378 in run_on_cpu (cpu=cpu@entry=0x16725a8, func=<optimized out>,
    data=...) at /home/rth/qemu/qemu/cpus-common.c:166
#3  0x00604c18 in vapic_enable_tpr_reporting (enable=<optimized out>)
    at /home/rth/qemu/qemu/hw/i386/kvmvapic.c:507
#4  0x00710160 in qdev_reset_one (dev=<optimized out>, opaque=<optimized out>)
    at /home/rth/qemu/qemu/hw/core/qdev.c:250

(gdb) thr 4
[Switching to thread 4 (Thread 0x9fdfed30 (LWP 10969))]
#0  0xb606d138 in pthread_cond_wait@@GLIBC_2.4 () from /usr/lib/libpthread.so.0
(gdb) where
#0  0xb606d138 in pthread_cond_wait@@GLIBC_2.4 () from /usr/lib/libpthread.so.0
#1  0x009a66a4 in qemu_cond_wait_impl (cond=0x1692ed8,
    mutex=0xcad4c8 <qemu_global_mutex>,
    file=file@entry=0x9d0108 "/home/rth/qemu/qemu/cpus.c",
    line=line@entry=1647) at /home/rth/qemu/qemu/util/qemu-thread-posix.c:161
#2  0x00553070 in qemu_tcg_rr_cpu_thread_fn (arg=<optimized out>)
    at /home/rth/qemu/qemu/cpus.c:1647
#3  0x009a5f20 in qemu_thread_start (args=<optimized out>)
    at /home/rth/qemu/qemu/util/qemu-thread-posix.c:502

I had hoped that I would have been able to reproduce this on another host by
forcing rr mode.  But so far no luck.

Thoughts?

r~